EMERGING CONTAMINANTS IN ECOSYSTEMS:

NEW CHALLENGES FOR WATER REUSE IMPLEMENTATION AND

MECHANISMS OF PERFLUOROCHEMICAL BIOACCUMULATION

A DISSERTATION

SUBMITTED TO THE DEPARTMENT OF

CIVIL AND ENVIRONMENTAL ENGINEERING

AND THE COMMITTEE ON GRADUATE STUDIES

OF STANFORD UNIVERSITY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

Heather Nicole Bischel August 2011

http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/zs920pn1994

© 2011 by Heather Nicole Bischel. All Rights Reserved.

Re-distributed by Stanford University under license with the author.

This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 United States License.

ii

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Richard Luthy, Primary Adviser

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Martin Reinhard

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Laura MacManus-Spencer

Approved for the Stanford University Committee on Graduate Studies.

Patricia J. Gumport, Vice Provost Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file inUniversity Archives.

iii

iv

v

Abstract Technological innovations developed in response to pressing water supply needs in

populated arid regions have led to the recovery of municipal wastewater for beneficial

reuse worldwide. At a time when rapid urbanization, severe droughts, and public

concern introduce complex management challenges for water reuse, the persistence of

residual and byproduct pharmaceutical and industrial chemicals in treated municipal

effluent gives rise to new technological hurdles. Bioaccumulation of synthetic organic

chemicals in environments downstream of wastewater effluent discharge and recycled

water use poses an ecological risk and introduces a potential pathway of human

exposure to these contaminants. This dissertation assesses management challenges for

water reuse implementation in Northern California; identifies opportunities of water

reuse for ecosystem enhancement; explores the bioaccumulation of one class of

persistent and toxic unregulated chemical contaminants, perfluoroalkyl acids (PFAAs);

and evaluates mechanisms of bioaccumulation via an in-depth study of PFAA

interactions with a model serum protein.

Chapter 1 provides background and outlines research objectives governing this

dissertation. In Chapter 2, major factors that influenced the implementation of

nonpotable water reuse in Northern California are presented based on a survey of

program managers. Capturing experiences of managers in urban and peri-urban regions

of California provides context for the historical developments of water reuse and the

sources of barriers to implementation. Results demonstrate that in recent times, water

reuse is driven more often by water supply needs rather than by wastewater discharge

limitations. From a management perspective, economic issues stand as the largest

hindrance to successful project implementation, while negative perceptions of water

reuse less frequently inhibit nonpotable water reuse projects. Analysis conducted in

Chapter 3 indicates that while ecosystem protection goals are frequently drivers of

water reuse programs, few water reuse projects have been implemented in California

explicitly for ecosystem enhancement or wildlife habitat creation. Augmentation of

degraded wetlands with recycled water represents an opportunity for expansion of

inland water reuse programs. However, detection of persistent, unregulated

vi

contaminants in recycled water presents management challenges for these projects.

The ability to accurately predict the bioaccumulative potential of chemicals in

aquatic organisms is an essential component to assessing the human health and

ecological risk of trace micropollutants. The bioaccumulation of PFAAs presents a

particularly intriguing case. Unlike other persistent organic pollutants, PFAAs do not

preferentially accumulate in lipids and fatty tissue but rather in body compartments with

high protein content, including the liver, kidneys, and serum. In Chapter 4, PFAA

concentrations detected in the livers of white sturgeon from the San Francisco Bay are

presented as a brief study on the environmental prevalence and bioaccumulation of

these chemicals. A fugacity-based approach that utilizes protein-water distribution

coefficients (KPW) based on interactions with model proteins is introduced as a useful

parameter to characterize the bioaccumulation and in vivo bioavailability of PFAAs.

Based on this modeling paradigm, noncovalent interactions of long-chain perfluoroalkyl

acids with bovine and human serum albumins (BSA and HSA, respectively) are

characterized in Chapter 5. Results suggest binding through specific high affinity

interactions at low PFAA:albumin mole ratios. In an effort to reduce the

bioaccumulation of PFAAs in humans and wildlife, fluorochemical manufacturers have

recently shifted production to shorter chain-length compounds. In Chapter 6,

associations of perfluoroalkyl carboxylates (PFCAs) with 2 to 12 carbons (C2 – C12) and

perfluoroalkyl sulfonates with 4 to 8 carbons (C4, C6, and C8) with BSA and

physiochemical binding mechanisms are evaluated at physiologically relevant

PFAA:albumin mole ratios and various solution conditions using equilibrium dialysis,

nanoelectrospray ionization mass spectrometry, and fluorescence spectroscopy.

Measured log KPW values for C4 to C12 PFAAs confirm that protein associations as

characterized in this model scenario prove to be greater in magnitude for PFAAs than

lipid-based partitioning coefficients. Association constants determined for

perfluorobutanesulfonate and perfluoropentanoate with BSA are on the order of those

for long-chain PFAAs, suggesting that physiological implications of strong binding to

albumin may be important for short-chain PFAAs.

In the final chapter, conclusions are drawn for research objectives outlined initially,

and future research directions are identified. The presented evaluation of management

vii

challenges for water reuse implementation provides context for issues surrounding

chemicals of emerging concern (CECs) in recycled water. However, uncertainty

regarding bioaccumulation of CECs from recycled water used for direct habitat

enhancement or creation remains a concern. Investigation of mechanisms influencing

PFAA bioaccumulation provides insight into one class of CECs now detected in

sensitive aquatic ecosystems. As the elimination of one unsafe chemical does not

guarantee the safety of its commercial replacements, presented findings further

contribute to ongoing efforts to characterize the physiochemical properties and

anticipated environmental fate of compounds used to replace long-chain perfluorinated

chemicals.

viii

Acknowledgement First, I thank my advisor Richard Luthy for thoughtful mentoring, consistent

guidance, and for always challenging each of his students to keep the big picture in

mind. Dick is living reassurance that kindness and compassion go hand-in-hand with

effectiveness and impact – a wonderful role model to have. A special thanks to Laura

MacManus-Spencer for invaluable cross-coast collaboration (including many hours on

the phone and conference-roommate chats) and for a delightful friendship. I also thank

Martin Reinhard, Jim Leckie, and Buzz Thompson, for joining my dissertation

committee.

One of the most enjoyable components of my work has been the opportunity to

interact with and learn from so many extraordinary people. Many thanks to the entire

Luthy Research Group for plenty of support and encouragement; to: Chris, Laura, and

Pam for helping me get started in the lab; to YeoMyoung and Eunah – organic

chemistry extraordinaires – from whom I learned so much as teachers and teaching

partners; Sarah for office-time advice; Jay for mini-debriefs; Aude for her work and

grounded perspectives on water reuse for ecosystems; Jeanne and Diana for a bit of

tennis; Chinghong, Lilli, Sungwoo, Yuan, and Yongju for research advice with a smile;

and Niveen for summertime chats as I finished up.

I am also tremendously grateful to: Gregory Simon and Tammy Frisby for diving

right in with recycled water field trips and helping me view our work from different

perspectives; Sophie Egan for hours of meticulous work, and contagious enthusiasm;

the many water reuse professionals who volunteered their time to participate in our

project and answer our phone calls; the professors at Cal who helped set me on this path

(Go Bears!) and the ones at Stanford who opened up even more; my best bud (at

Stanford), Liv Walter, for giving me the keypad code and keeping her door unlocked;

my dear friend Blythe Layton for countless hugs and laughs; the LCMS csars of the past

for keeping it running; the EES moles– too numerous to name – for advice and beer;

Royal, of course, who always lent a helpful hand; EFML nerds, especially Liv, Joel,

Sarah, Erin, Jaime, and Yacoub for crossing the great divide for many meals and a few

ix

crafts together; ESW friends, especially Eric, Grace, Kathy, Kristof, Milena, Sophie,

and Julie Chow, for trying to save the world; the San Francisco Bay and the Sierras for

inspiration and fun; my awesome housemates – from Cal Ave. to Cowper; my long-time

friends Gracie, Celia, Ycxia, Edna, Rachel, Christi, Heena, and Kofi and my family –

Dad, Mom, Tyler, Mandy and Drew – for lots of love and support, even from afar.

Finalement, à mon mec, Nico – une joie de vivre depuis mon premier jour à

Stanford. Merci à tous!

Support for my work was provided by the Stanford Woods Institute for the

Environment (Environmental Ventures Project), the National Science Foundation

Graduate Research Fellowship, and the National Defense Science & Engineering

Graduate Fellowships.

"When the well's dry, we know the worth of water."

- Benjamin Franklin (1706-1790), Poor Richard's Almanac.

x

Contents Abstract .......................................................................................................................... v

Acknowledgement ..................................................................................................... viii

Contents ......................................................................................................................... x

List of Tables .............................................................................................................. xii

List of Figures .............................................................................................................. xv

Introduction ................................................................................................................... 1

1.1 Motivation and Background ................................................................................... 1

1.2 Research Objectives and Précis .............................................................................. 6

Management experiences and trends for water reuse implementation in Northern

California ........................................................................................................... 9

2.1 Introduction ............................................................................................................. 9

2.2 Methodology ......................................................................................................... 11

2.3 Analysis of Water Reuse in California ................................................................. 13

2.4 Drivers of Water Reuse Implementation in Northern CA .................................... 16

2.5 Challenges for Water Reuse Implementation in Northern CA ............................. 22

2.6 Significance ........................................................................................................... 28

2.7 Supporting Information ......................................................................................... 30

Water reuse for ecosystem enhancement: Matching opportunity with need ........ 51

3.1 Introduction ........................................................................................................... 51

3.2 Few Existing Examples of Water Reuse for Direct Ecosystem Enhancement in

California .................................................................................................................... 52

3.2 Identifying Opportunities for Ecosystem Enhancement ....................................... 57

3.3 Technological and Management Challenges ........................................................ 63

3.4 Significance ........................................................................................................... 68

xi

Exposure of perfluorinated chemicals to San Francisco Bay white sturgeon and

mechanisms of bioaccumulation ................................................................... 69

4.1 Introduction .......................................................................................................... 69

4.2 Predictive models for PFAA bioaccumulation ..................................................... 71

4.3 Materials and Methods ......................................................................................... 74

4.4 Results and Discussion ......................................................................................... 77

4.5 Significance .......................................................................................................... 83

Investigating binding to a model protein: Noncovalent interactions of long-chain

perfluoroalkyl acids with serum albumin .................................................... 85

5.1 Introduction .......................................................................................................... 85

5.2 Materials and Methods ......................................................................................... 88

5.3 Results and Discussion ......................................................................................... 94

5.4 Significance ........................................................................................................ 105

5.5 Supporting Information ...................................................................................... 107

Strong associations of short-chain perfluoroalkyl acids with serum albumin and

investigation of binding mechanisms .......................................................... 123

6.1 Introduction ........................................................................................................ 123

6.2 Methods .............................................................................................................. 126

6.3 Results and Discussion ....................................................................................... 128

6.4 Significance ........................................................................................................ 139

6.5 Supporting Information ...................................................................................... 142

Conclusions ............................................................................................................... 161

7.1 Summary Conclusions ........................................................................................ 161

7.2 Future Work ........................................................................................................ 166

7.3 Final Thoughts .................................................................................................... 168

References ................................................................................................................. 171

xii

List of Tables Table 2.1. Percent of respondents indicating a specific factor as a Driver or one of the

three Most Important Drivers. .......................................................................... 17

Table 2.2. Percent of respondents indicating a specific factor as a Hindrance or one of

the three Most Important Hindrances. .............................................................. 23

Table 2.3S. Chi square analysis of drivers of program implementation by self-reported

date of implementation. .................................................................................... 40

Table 2.4S. Chi square analysis of hindrances to program implementation by self-

reported date of implementation. ...................................................................... 41

Table 2.5S. Chi square analysis of categorized drivers of program implementation by

self-reported date of implementation. ............................................................... 42

Table 2.6S. Chi square analysis of categorized hindrances to program implementation

by self-reported date of implementation. .......................................................... 43

Table 2.7S. Chi square analysis of drivers of program implementation by self-reported

total annual reclaimed water use. ...................................................................... 44

Table 2.8S. Chi square analysis of hindrances to program implementation sorted by

total annual reclaimed water use. ...................................................................... 45

Table 2.9S. Chi square analysis of categorized drivers of program implementation by

self-reported total annual reclaimed water use. ................................................ 46

Table 2.10S. Chi square analysis of hindrances to program implementation by self-

reported total annual reclaimed water use. ....................................................... 47

Table 2.11S. Representation of recycled water beneficial uses from the 2010 Survey of

Northern California (n = 69) agencies and the State Water Resources Control

Board 2009 Municipal Wastewater Recycling Survey (n = 143). .................... 48

Table 2.12S. Milestones for California water reuse and statewide recycling goals. .... 49

Table 2.13S. San Francisco Bay Area Recycled Water Coalition 2011 Project Summary

(64) .................................................................................................................... 50

Table 3.1. Projects in Northern California utilizing recycled water for ecosystem

enhancement or treatment wetlands for wastewater effluent polishing. ........... 55

xiii

Table 4.1. Analyte primary and secondary transitions monitored, internal standards

(IS), average concentration of MDL samples (with standard deviation of 12

replicates), and calculated MDLs. .................................................................... 76

Table 5.1. Percent bound and log KPW for PFNA binding to 500 µM albumin

determined by equilibrium dialysis and LC-MS/MS. ...................................... 96

Table 5.2. Association constants (Ka) and binding stoichiometries (n) for PFOA and

PFNA binding to BSA determined by equilibrium dialysis. ............................ 98

Table 5.3. Summary of PFCA-albumin association constants (Ka) and binding

stoichiometries over a range of ligand:protein mole ratios ([L]:[P]). ............ 104

Table 5.4S. Association constants (Ka,1) and binding stoichiometries (n1 and n2) for

PFOA binding to 1 µM BSA as determined by equilibrium dialysis for a range

of applied weighting factors. .......................................................................... 114

Table 5.5S. Association constants (Ka,1) and binding stoichiometries (n1 and n2) for

PFNA binding to 1 µM BSA as determined by equilibrium dialysis for a range

of applied weighting factors. .......................................................................... 114

Table 5.6S. Association constants (Ka,1 and Ka,2) and binding stoichiometries (n1 and

n2) for PFOA and PFNA and binding to 1 µM BSA as determined by

equilibrium dialysis using Equation 5.16S with no weighting factor (WF) and a

1/[Free PFAA] weighting factor. .................................................................... 115

Table 5.7S. Association constants (Ka,1) and binding stoichiometries (n1) for PFOA and

PFNA binding to 1 µM BSA as determined by equilibrium dialysis for a subset

of the total data and a one-class binding model. ............................................ 117

Table 5.8S. Estimated association constants calculated from nanoESI-MS results for 50

µM BSA exposed to PFOA (25, 50, and 100 µM). ........................................ 120

Table 5.9S. Estimated association constants calculated from nanoESI-MS results for 50

µM BSA exposed to PFNA (25, 50, and 100 µM). ........................................ 120

Table 5.10S. Estimated association constants calculated from nanoESI-MS results for

50 µM BSA exposed to PFDA (25, 50, and 100 µM). ................................... 121

Table 5.11S. Estimated association constants calculated from nanoESI-MS results for

50 µM BSA exposed to PFOS (25 µM). ........................................................ 121

xiv

Table 6.1. Fraction of perfluoroalkyl acids (PFAAs) bound to 200 µM bovine serum

albumin (BSA) and log protein-water distribution coefficients for PFAAs with

BSA measured over a range of equilibrium free PFAA reservoir concentrations.

......................................................................................................................... 129

Table 6.2S. Dialysis mass balance, reservoir matrix and bovine serum albumin (BSA)

spike recovery results, and liquid chromatography tandem mass spectrometry

(LC-MS/MS) transitions monitored. ............................................................... 146

Table 6.3S. Protein-water distribution coefficients for PFAAs with BSA. ............... 147

Table 6.4S. Measured incremental mass shifts (ΔM) from measured BSA peak (P) to

BSA-PFAA peaks (P + jL) for representative spectra in manuscript Figure 6.2

and Supporting Information Figure 6.7S. ....................................................... 149

Table 6.5S. Average fraction of PFAAs bound to BSA (197 ± 2 µM) for a range of pH

conditions. ....................................................................................................... 152

Table 6.6S. Slope of linear regressions for average number of PFAAs bound to BSA,

!

" , versus pH (6 to 9) in equilibrium dialysis tests. ........................................ 153

xv

List of Figures Figure 1.1. Structures and names of perfluoroalkyl acids (PFAAs) included in this

study. .................................................................................................................. 4

Figure 2.1. Timeline of statewide water recycling goals and production volumes, major

drought periods, and select water recycling laws and policies in California

during the implementation period for survey respondents. .............................. 10

Figure 2.2. A snapshot of water reuse facilities in California from the National

Database of Water Reuse Facilities (Annual Production, reported as Facility

Production Average Annual Actual in million gallons) and the California 2009

Municipal Water Recycling Survey (Annual Reuse, reported as Total Reuse for

2009 in AFY). ................................................................................................... 15

Figure 2.3. Beneficial uses of recycled water in Northern California in 2001 and 2009.

.......................................................................................................................... 18

Figure 2.4. Results of χ2 analyses by implementation date for specific factors indicated

as one of the Three Most Important Drivers (top) or more generally a Driver of

implementation (bottom). ................................................................................. 20

Figure 2.5S. Distribution of survey invitations and responses collected from Northern

California counties. ........................................................................................... 37

Figure 2.6S. Agencies binned by annual recycled water flow (AFY) in 2001 and 2009.

.......................................................................................................................... 38

Figure 2.7S. Representation of total annual reclaimed water deliveries reported from

the 2010 Survey of Northern California agencies (n = 64) and the State Water

Resources Control Board 2009 Municipal Wastewater Recycling Survey (n =

143). .................................................................................................................. 39

Figure 3.1. Distribution of California Rapid Assessment Method (CRAM) overall

wetland scores, included for estuarine (saline and non-saline) and riverine

(confined and non-confined), and wastewater facilities with tertiary treatment

capacity. ............................................................................................................ 59

xvi

Figure 3.2. Response frequencies for 2010 Survey respondents who rated five broad

categories of hindrances to implementation of future water reuse programs for

ecosystem enhancement. ................................................................................... 66

Figure 4.1. Study area. ................................................................................................. 74

Figure 4.2. Measured PFAA concentrations (ng/g ww) in white sturgeon fish livers (n =

15). .................................................................................................................... 78

Figure 4.3. Stable isotopes, fish length, and muscle Hg concentration plotted with white

sturgeon, striped bass, or leopard shark liver PFOS concentrations on the

ordinate. ............................................................................................................ 81

Figure 5.1. Equilibrium dialysis results for PFOA (a) and PFNA (b) where

!

" is the

average number of PFAA molecules bound per albumin. ................................ 97

Figure 5.2. Representative spectra from 2500-4800 m/z for BSA, BSA exposed to

PFOA (50 µM), and BSA exposed to PFNA (50 µM) in 9 mM ammonium

acetate (pH 7). ................................................................................................. 100

Figure 5.3. Mass spectra for the +16 charge state of 50 µM BSA in 9 mM ammonium

acetate (pH 7) with PFOA (left) and PFNA (right). ....................................... 101

Figure 5.4S. Structures and names of perfluoroalkyl acids (PFAAs) used in this study.

......................................................................................................................... 107

Figure 5.5S. Samples taken prior to equilibration in the reservoir from control bags

containing only buffer and the PFAA spike are compared to samples taken from

test bags containing 1 µM BSA with the same PFAA spike. ......................... 110

Figure 5.6S. Reservoir samples taken over time in a PFNA equilibrium dialysis test

indicate equilibrium of the system after 24 hours. .......................................... 111

Figure 5.7S. Reservoir samples taken over time in a PFOA equilibrium dialysis test

suggest equilibrium of the system after approximately 48 hours. .................. 111

Figure 5.8S. Measured total and free PFOA and PFNA concentrations taken at

equilibrium from dialysis bag and reservoir samples, respectively. ............... 112

Figure 5.9S. Standard deviations of triplicate measurements of bound PFAAs (Sbound)

were linearly correlated with free PFAA concentrations, as shown above for

PFNA in 1µM BSA equilibrium dialysis tests. ............................................... 113

xvii

Figure 5.10S. Equilibrium dialysis results for PFOA and PFNA up to a 5:1 ligand to

protein mole ratio and 1 µM BSA where ν is the average number of PFAA

molecules bound per albumin. ........................................................................ 116

Figure 5.11S. Representative mass spectra for the +16 charge state of 50 µM BSA in 9

mM ammonium acetate (pH 7) with PFDA (left, cone voltage = 100V) and

PFOS (right, cone voltage = 130 V). .............................................................. 118

Figure 5.12S. Representative deconvoluted spectrum for 50 µM BSA in 9 mM

ammonium acetate (pH 7) with 100 µM PFOA (cone voltage = 100V) used for

determination of Ka. ....................................................................................... 119

Figure 6.1. Measured BSA-water distribution coefficients (KPW) for perfluoroalkyl

sulfonates (PFSAs, ) and perfluoroalkyl carboxylates (PFCAs, ) with

fluorocarbon tail lengths of 4 to 11. ............................................................... 131

Figure 6.2. Deconvoluted spectra of 50 µM BSA alone or with 50 µM PFPeA, PFHxA,

PFHpA, PFOA, or PFNA. .............................................................................. 132

Figure 6.3. Effect of ionic head group on binding of equivalent chain length PFAAs to

BSA. ............................................................................................................... 135

Figure 6.4. Effect of pH on

!

" , the concentration of PFAA bound to BSA normalized to

the total protein concentration. ....................................................................... 137

Figure 6.5S. Structures and names of perfluoroalkyl acids (PFAAs) included in the

present study. .................................................................................................. 145

Figure 6.6S. Total PFAA analyte concentration in the bound phase (CP, [g bound

PFAA / mL BSA]) versus total aqueous PFAA concentration (CW, [g free PFAA

/ mL water]). ................................................................................................... 147

Figure 6.7S. Representative deconvoluted spectra of 50 µM BSA alone or with 50 µM

TFA, PFPrA, PFBA, PFBS, PFHxS, or PFOS. .............................................. 148

Figure 6.8S. Representative deconvoluted spectra of PFPeA and PFNA (100 µM) with

BSA (50 µM) collected at a 10 V collision energy (left) and representative

spectra of PFNA (100 µM) with BSA (50 µM) at 10, 30, 50 or 70 eV collision

energy (right). ................................................................................................. 150

Figure 6.9S. Fluorescence spectra of BSA at pH 6 (solid line), 7 (long dashed line), 8

(short dashed line), or 9 (dotted line). ............................................................ 151

xviii

Figure 6.10S. Average number of bound perfluoroalkyl carboxylates (PFCAs, ) or

perfluoroalkyl sulfonates (PFSAs, ) per BSA,

!

" (µM PFAAbound / µM BSA),

measured in dialysis bags containing 200 µM BSA in a PFAA-spiked reservoir

at pH 7. ............................................................................................................ 152

Figure 6.11S. Effect of pH on the average number of PFHpA (), PFOA (), PFNA

(), or PFDA () molecules bound to BSA. .................................................. 153

Figure 6.12S. Changes in the fluorescence of BSA with added PFNA (top) or PFOS

(bottom) at pH 6 (), 7 (), 8 (), or 9 (). ................................................... 154

Figure 6.13S. The binding of PFNA (a) and PFOS (b) to BSA, plotted as the degree of

saturation (Y) versus total PFAA concentration, at pH 6 (), 7 (), 8 (), or 9

(). .................................................................................................................. 155

Figure 6.14S. Dependence of estimated binding constant (KHill) on pH for the binding

of PFNA () and PFOS () to BSA. .............................................................. 156

Figure 6.15S. Changes in the fluorescence of BSA with added PFNA (top) or PFOS

(bottom) at 0.21 M (), 0.30 M (), or 0.41 M () ionic strength and pH 7. 157

Figure 6.16S. The binding of PFNA (a) and PFOS (b) to BSA, plotted as the degree of

saturation (Y) versus total PFAA concentration, at 0.21 M (), 0.30 M (), or

0.41 M () ionic strength and pH 7. ............................................................... 158

Figure 6.17S. Dependence of estimated binding constant (KHill) on ionic strength for

the binding of PFNA () and PFOS () to BSA at pH 7. .............................. 159

1

Chapter 1

Introduction

1.1 Motivation and Background

Wastewater recycling, or water reuse, is becoming critically important for improved

water management and ecosystem protection and rehabilitation in semi-arid regions of

the American West as increasing water demands deplete freshwater supplies. Although

the amount of municipal wastewater recovered for the augmentation of water supplies

and pollution abatement is increasing globally (1), untapped wastewater resources

remain abundant, even in water-starved regions. In California alone, where reuse of

municipal wastewater more than doubled from 1970 to 2002, only 10% of available

treated effluent is recycled (2, 3). Water resource opportunities arising from advanced

wastewater treatment technologies are tempered by numerous challenges experienced

by managers during the implementation of recycled water programs in California and

elsewhere. For example, recent decisions to implement new water reuse projects are

complicated by considerable uncertainty regarding the risk of unregulated chemical

contaminants in recycled water (4). Reclaimed wastewater intended for a range of

2

beneficial uses, including agricultural irrigation and habitat enhancement, may contain

detectable quantities of chemicals with largely unknown ecological effects (5, 6).

Water reuse challenges and opportunities. Despite efforts to encourage and

support water reuse programs at the state and federal levels (2), not all projects are

successful. California failed to reach statewide goals for water reuse in 1982, 2000, and

most recently in 2010. Public opposition has led to the suspension or abandonment of

several large water reclamation projects for indirect potable use in California (2, 7),

which remains a small fraction of California’s recycled water portfolio. Whereas

negative public perception stands as a key barrier to overcome for potable reuse, other

challenges associated with nonpotable reuse programs include a mélange of funding,

regulatory, and technical hurdles (8). Pressing water supply concerns related to dramatic

population growth and periodic, yet often severe, regional droughts necessitate a

thorough understanding of past experiences in water reuse implementation to identify

the sources of such failures and to focus efforts on effectively implementing new water

reuse programs. Further, one largely underutilized opportunity for recycled water use is

that for natural system enhancement, in which recycled water may serve as a hydrologic

resource for wetlands, lakes, and streams (9, 10). Natural processes may be harnessed to

remove contaminants of concern in treatment or polishing wetlands and augmented

river or stream systems (11, 12). Wetland treatment projects are attractive as a cost-

saving measure over expensive treatment and distribution facilities (12), though many

unresolved issues remain with respect to water reuse for ecosystem enhancement. New

opportunities to couple reuse of wastewater with the needs of streams and wetlands and

strategies to facilitate implementation of such projects, including quantification of

auxiliary benefits realized via ecosystem enhancement projects, require identification.

Chemicals of emerging concern. Chemicals of emerging concern (CECs) have

come to the attention of water reuse policymakers and regulatory bodies (13).

Addressing issues of CECs is especially acute for sensitive aquatic ecosystems that

serve as receiving waters for wastewater discharge or recycled water used in ecosystem

enhancement projects. Contamination of aquatic systems by chemical micropollutants is

a key environmental challenge facing humanity (14). The remarkable diversity of

structural properties of CECs and attendant uncertainty arising from unknown toxicity

3

mechanisms and thresholds necessitates research that assesses impacts on aquatic life

and human health, explores cost-effective and appropriate treatment technologies, and

identifies benign products and processes (14). As one example of the regulatory and

technological challenges facing the management of CECs, the United States

Environmental Protection Agency (U.S. EPA) identified nearly 26,000 substances as

potential candidates for the 2009 Candidate Contaminant List 3 (CCL3), a list used to

prioritize research and data collection efforts for unregulated contaminants. In the final

CCL3, this universe of contaminants was whittled down to 104 chemicals or chemical

groups and 12 microbial contaminants for their potential to present health risks through

drinking water exposure (15). Two compounds included in the U.S. EPA CCL3,

perfluorooctanoate (PFOA) and perfluorooctanesulfonate (PFOS), are representative of

a family of compounds commonly referred to as perfluorochemicals (PFCs) or

perfluoroalkyl acids (PFAAs, Figure 1.1).

Perfluorinated chemicals. Perfluoroalkyl acids are remarkably stable due to

complete fluorination of the saturated carbon chain. The high-energy carbon-fluorine

bond imparts resistance to photolysis, hydrolysis, microbial degradation, and

metabolism by vertebrates (16, 17). The unique chemical properties of PFAAs have

been capitalized over the past half century in the production of a variety of industrial

and consumer products including textile coatings, food-contact paper, fire-fighting

foams, repellants, paints, and cosmetics (16, 18-20). Amongst this diverse class of

synthetic surfactants, PFOA and PFOS are final degradation products of a host of parent

compounds that include fluorotelomer alcohols (21), n-ethyl sulfonamido ethanols (22,

23), and polyfluoroalkyl phosphate surfactants (19). Today, PFAAs are recognized to be

extremely persistent in the environment as well as globally distributed,

bioaccumulative, and toxic (24-26).

4

Figure 1.1. Structures and names of perfluoroalkyl acids (PFAAs) included in this

study. Compound abbreviations and notations (C2 – C14) are listed. PFSAs and PFCAs

have a fluorocarbon tail length of n + 1 and m + 1, respectively.

Wastewater treatment plants are likely dominant sources of PFAAs to the

environment (27-29). The observed environmental concentrations of PFOS, which

exhibits a range of ecotoxicological endpoints, are generally greater than those of other

PFAAs (30). Subchronic exposure of PFOS to animals causes significant weight loss

coupled with hepatotoxicity and reduction in thyroid hormones and serum cholesterol

(31). PFOA is likely to be carcinogenic, according to a U.S. EPA Science Advisory

Board, by inducing liver adenomas via activation of the peroxisome proliferator-

activated receptor (32). Although the toxicity of PFOA and PFOS is well studied (26,

31), more information concerning the modes of toxic action is needed to quantify risks

of exposures to PFAA mixtures for a range of species (33). Vast differences in the

elimination half-life for a variety of PFAAs are observed between species, with several

years expected for humans (26).

Responding to comprehensive research documenting widespread occurrence of

5

long-chain PFAAs in humans and wildlife (25, 34), the 3M Company eliminated

production of perfluorooctane sulfonyl fluoride (POSF)-based materials, including

PFOA and PFOS. Several companies followed suit, committing to eliminate emissions

of PFOA and related compounds by 2015 (35). However, legacy products remain in

use, and production continues globally for perfluoroalkyl compounds of varying chain

lengths (C4 – C15) (20). Reformulation of product contents to include substitute

compounds and modified PFAAs, such as those based on C4-sulfonyl chemistries,

necessitates research regarding these compounds (35). A shorter half-life in organisms

of perfluorobutanesulfonate (PFBS) likely reduces its bioaccumulation (36, 37) but is

unlikely to affect its persistence. The environmental and human health implications of

increased production and use of short-chain PFAAs requires further study.

Biological accumulation of perfluoroalkyl acids. Although perfluorinated

compounds are detected ubiquitously in organisms, little is known about the

mechanisms of PFAA bioaccumulation and the processes by which PFAAs are

introduced into the aquatic food web. PFAAs exhibit biouptake patterns divergent from

well-characterized hydrophobic organic contaminants. Despite relatively low

hydrophobicity, perfluorinated compounds with greater than seven fluorinated carbons

bioaccumulate and biomagnify in aquatic food webs (18, 25, 38). Rather than

partitioning to adipose tissue, PFAAs are detected predominantly in protein-rich

compartments such as the liver, kidney and blood (39-42). The bioconcentration factor

(BCF) of PFOS, relating ambient water concentrations to measured tissue

concentrations, ranges from approximately 1,000 to more than 5,000 for bluegill and

rainbow trout fish species, depending on the method of determination and organ

considered (30). The structures of perfluoroalkyl carboxylates (PFCAs) and

perfluoroalkyl sulfonates (PFSAs), two homologue groups of PFAAs, resemble those of

fatty acids and hydrocarbon-based detergents, but the perfluorinated tail renders the

compounds both hydrophobic and oleophobic (16, 33). The nature of PFAA structure

and bioaccumulation suggests an importance of protein interactions (43). However, the

sorptive capacity of animal protein is rarely incorporated in biouptake models to

improve estimations of chemical distribution and bioaccumulation of persistent organic

pollutants (44).

6

1.2 Research Objectives and Précis

Amongst a broad array of challenges, negative public perception of health risks,

actions of influential stakeholders in a changing regulatory environment, and limited

availability of financial assistance may affect the implementation of water reuse

programs. High costs, including those for treatment facilities and distribution systems,

may be exacerbated by limited financial and technological capacity to eliminate trace

contaminants. In the face of new knowledge surrounding chemicals of emerging

concern, coupled with advanced analytical techniques to evaluate the presence of

chemicals, future challenges associated with water reuse programs in California may be

different than historical practices and experiences. Work presented in Chapter 2,

Management experiences and trends for water reuse implementation in Northern

California,1 assesses the greater context of water reuse in California and contains

results from a survey of water reuse managers and professionals in Northern California.

Chapter 3, Water reuse for ecosystem enhancement: Matching opportunity with need,

delves into the role of water reuse for ecosystems, describing existing programs and

broadly identifying opportunities for new enhancements. The analysis draws on

responses from the previously discussed survey and complementary databases to outline

specific challenges for habitat enhancement using tertiary treated wastewater. Together,

these chapters seek to provide insight on the following questions:

• What are the major drivers and barriers to water reuse in Northern

California, and how have these factors evolved through time?

• To what extent has water reuse been applied for the direct benefit of

ecosystems, and what major challenges are associated with the

implementation of water reuse for ecosystem enhancement?

Despite an indication that positive environmental impact is a benefit of water reuse

projects, relatively few projects have been implemented for ecosystem enhancement in

California.

1 The results presented in this chapter are submitted as a Research Article by Heather N. Bischel, Gregory Simon, Tammy M. Frisby, and Richard G. Luthy for the journal Environmental Science & Technology.

7

One potential challenge associated with the implementation of water reuse for

ecosystem enhancement is uncertainty regarding the bioaccumulation of unregulated

chemical contaminants. Chapter 4, Exposure of perfluorinated chemicals to San

Francisco Bay white sturgeon and mechanisms of bioaccumulation, sets the stage for

understanding the role of bioaccumulation of chemicals of emerging concern in

ecosystems as well as mechanisms associated with PFAA bioaccumulation. PFAA

concentrations detected in white sturgeon fish livers from organisms in the San

Francisco Bay are presented as a case study. As perfluorinated chemicals receive

increasingly more attention and are more carefully examined for their potential

ecosystem effects, bioaccumulation processes based on biologically relevant

mechanisms are considered. This chapter asks:

• What dominant processes govern the bioaccumulation of PFAAs, and how

can these processes be captured in bioaccumulation models?

As mentioned, PFAAs do not preferentially accumulate in lipids and fatty tissue but

rather in body compartments with high protein content, including the liver, kidneys, and

serum. Such observations bring question to the appropriateness of using octanol-water

partition coefficients (Kow), which are commonly applied for modeling the

bioaccumulation of persistent organic pollutants, to describe the environmental behavior

of PFAAs.

For PFAAs, molecular interactions with proteins likely contribute to PFAA

bioaccumulation mechanisms. As such, quantitatively determined associations between

perfluorinated chemicals and proteins may be useful parameters to more accurately

describe observed bioaccumulation. In Chapters 5 and 6, processes influencing the

bioaccumulation of these unique compounds are evaluated through associations of

PFAAs with proteins, utilizing bovine serum albumin as a model protein. These

chapters seek to address the following questions:

• How do long-chain PFAAs associate with the model protein, serum albumin,

at physiologically relevant PFAA:albumin mole ratios?

• What analytical tools are appropriate for quantitatively determining PFAA-

albumin associations?

• Given a shift in production of fluorinated compounds to shorter-chain length

8

compounds, how will a reduction in perfluoroalkyl chain length affect

protein-water distribution coefficients?

• What physiochemical mechanisms govern interactions of PFAAs with serum

albumin?

In Chapter 5, Investigating associations with a model protein: Noncovalent

interactions of long-chain perfluoroalkyl acids with serum albumin, 1 association

constants (Ka) and binding stoichiometries for PFAA-albumin complexes are quantified

over a range of physiologically relevant PFAA:albumin mole ratios. Binding

interactions between PFAAs with eight to ten perfluoroalkyl carbons and the model

protein bovine serum albumin (BSA) are studied using equilibrium dialysis with liquid

chromatography tandem mass spectrometry and nanoelectrospray ionization mass

spectrometry. Chapter 6, Strong associations of short-chain perfluoroalkyl acids

(PFAAs) with serum albumin and investigation of binding mechanisms2, expands on

the previous chapter to evaluate associations of PFCAs with 2 to 12 carbons (C2 – C12)

and PFSAs with 4 to 8 carbons (C4, C6, and C8) with BSA at physiologically-relevant

PFAA:albumin mole ratios. Protein-water distribution coefficients (KPW) are quantified,

providing interpretation of hydrophobicity, steric hindrances, and electrostatic effects

on interactions with albumin. This work comprises a thorough evaluation of molecular

interactions of PFAAs with albumin using several analytical tools, a wide range of

ligand and substrate concentrations, a series of fluorochemical chain lengths and two

anionic head group moieties, and varied solution conditions. Chapter 7, Conclusions,

contains final remarks on the research objectives as well as a discussion of research

needs.

1 The results presented in this chapter originally appeared as a Research Article in the journal Environmental Science & Technology: (45) Bischel, H. N.; MacManus-Spencer, L. A.; Luthy, R. G. Noncovalent interactions of long-chain perfluoroalkyl acids with serum albumin. Environ. Sci. Technol. 2010, 44 (13), 5263-5269. 2 The results presented in this chapter are in press as a Research Article for the journal Environmental Toxicology & Chemistry by Heather N. Bischel, Laura A. MacManus-Spencer, Chaojie Zhang, and Richard G. Luthy.

9

Chapter 2

Management experiences and trends for

water reuse implementation in Northern

California

2.1 Introduction

California is at the forefront of recycled water use, treating municipal wastewater to

a high enough degree that it can be returned to the water supply for a variety of

beneficial uses including landscape irrigation (46-48), agriculture (49, 50), ecosystem

enhancement (9), industrial cooling and processing (47, 51), groundwater recharge and

indirect potable reuse (51-53). From 1970 to 2002, reuse of municipal wastewater more

than doubled in California from 175,000 acre-ft per year (AFY) to approximately

525,000 AFY. Yet this growth fell short of the state’s goal to reuse 700,000 AFY by

2000 (2, 3). California’s goal to increase reuse by 2 million acre-feet by 2030 over 2002

levels (54) will require a portfolio of projects for a range of beneficial uses. Given

multiple failures to attain statewide recycling goals (Figure 2.1), questions remain as to

10

the sources of such difficulties as well as the feasibility of reaching near-term goals

described in California’s State Water Board Strategic Plan Update of 2008-2012 (55).

Figure 2.1. Timeline of statewide water recycling goals and production volumes, major

drought periods, and select water recycling laws and policies in California during the

implementation period for survey respondents. Refer to the Supporting Information for

a description of major laws and policies.

Despite efforts to encourage and support water reuse programs at the state and

federal levels (e.g., (2) and (54)), not all projects are successful, and nonpotable reuse

projects frequently fall short of planned delivery goals (56, 57). Public opposition has

led to the suspension or abandonment of several large water reclamation projects for

indirect potable reuse in California (2, 7). Considering the promise of recycled water for

augmenting water supplies in the West and pressing water supply concerns related to

dramatic population changes and climate change, assessment of past and current

experiences in water reuse implementation will aid in more effectively promoting,

evaluating, and implementing water reuse. This paper contributes to this task by

evaluating the experiences and perspectives of current water reuse project managers in

Northern California to understand recent developments and major issues confronting

11

recycled water projects in the region.

Specifically, our study reveals the following: (1) In Northern California, water reuse

programs are widely distributed across 48 counties and, though more numerous than

programs in the 10 Southern California counties, are often smaller in the volumes of

reclaimed water delivered annually. This finding highlights the importance of capturing

experiences of managers in rural regions of California, which likely differ from

experiences in highly urbanized centers. (2) Regulatory requirements that limit

discharge played an important role in motivating many water reuse programs in

Northern California. However, a trend away from reuse as a wastewater disposal issue

is documented in Northern California, as water supply and reliability become more

prevalent drivers of water reuse. (3) Although ecosystem enhancement or protection

goals are frequently cited as drivers of water reuse, such goals are rarely the most

important drivers for reuse programs. Few water reuse programs in California have

been implemented for the purpose of ecosystem enhancement. (4) Negative perceptions

of water reuse were not frequently major hindrances to implementation of water reuse

programs in Northern California. Public perception of water reuse may be positively

influenced by a shift in view of recycled water towards that of a valuable resource and

as public knowledge of water supply challenges increases. (5) Economic issues stand as

the largest hindrance to successful project implementation from a management

perspective. In particular, smaller water reuse programs are less frequently incentivized

by federal or state grants and loans, while larger programs have somewhat greater

challenges associated with distribution system (pipeline) costs.

2.2 Methodology

Data sources. Primary data on water reuse agencies, practices, and management

experiences were collected via an online questionnaire of Northern California water

reuse managers conducted for the present study in 2010 (2010 Survey). Additional data

on water reuse agency characteristics were obtained from the California State Water

Resources Control Board (SWRCB) 2001 Water Recycling Survey released in 2002

(2001 Survey, (3)), the National Database of Water Reuse Facilities (National Database,

(58)), and the 2009 California Municipal Wastewater Recycling Survey, a follow-up

12

survey from the SWRCB released in April 2011 (2009 Survey, (59)). Municipal water

recycling agencies in Northern California (defined as the 48 counties northward of the

southern boundaries of Monterey, Kings, Tulare, and Inyo counties) listed on the

National Database and the 2001 Survey were invited to participate in the 2010 Survey.

Fieldwork administration and questionnaire. Data were collected online from

February to April 2010 using electronic surveys sent to general managers or

water/wastewater directors from 134 agencies in 41 Northern California counties using

a distribution list compiled from the SWRCB 2001 Survey and the publicly available

National Database (58). The questionnaire, which is described further in the Supporting

Information, was developed based on case study research, literature review and site

visits at water and wastewater facilities and agencies with programs implemented for

agriculture, landscape irrigation, industrial power plant cooling, and ecosystem

enhancement. Respondents were asked a number of questions related to the drivers and

challenges experienced in implementing their agency’s water reuse program with

additional survey components addressing responses to recent recycled water policy in

California and future expectations for programs in development. Prior to distribution,

survey testing by several consultants and project staff was conducted for usability and

content feedback.

Categorization and statistical tests. The analyses conducted for 2010 Survey

results provide quantitative confirmation of trends that have been previously discussed

and valuable insights into the characteristics of water reuse in Northern California.

Results represent quantitative response data and are supported by qualitative

descriptions of drivers and barriers experienced in program implementation. Chi square

analysis was conducted on two by two contingency tables constructed from frequency

results of specific drivers (Table 2.1) and hindrances (Table 2.2) to program

implementation to assess relationships between categorical variables. For simplicity in

additional analysis and discussion, the list of specific drivers and hindrances was

consolidated into eight and nine categorical variables, respectively. Chi square analysis

was also performed on these data, and categories were used to contextualize qualitative

responses to survey questions (See Tables 2.3S – 2.10S for full results). The

presentation of representative respondent quotations, extracted primarily from responses

13

to two questions – the single most important driver or hindrance to implementation –

provide context for the diversity of experiences evident throughout the results.

Respondent information and survey limitations. A total of 71 distinct agencies, a

53% response rate, are represented by 2010 Survey responses. Because some parent

utilities represent multiple recycled water facilities, a total of 81 unique production

facilities are represented by responses; however, most agencies (83%) represent only

one recycled water production facility, and another 7% represent a unique distribution

facility coupled to a production facility. Respondents consist of internal public agency

managers or utility staff. The survey completion rate was 40% of invited participants.

Therefore, the response fractions reported for each question indicate values for that

particular question. Respondent agencies for the 2010 Survey were distributed widely

across Northern California though survey representation appears somewhat weak for the

number of agricultural programs relative to the 2009 Survey data (Figure 2.5S and

Table 2.11S). The median year of recycled water program implementation, based on

self-reported implementation dates for 56 respondents, was 1991, with the earliest

reported implementation occurring in the early 1960’s.

2.3 Analysis of Water Reuse in California

Recycled water distribution falls short of statewide goals. Figure 2.1 displays a

timeline of statewide water recycling goals and production volumes (2, 3, 13, 57, 59,

60). According to the 2009 Survey results, recycled water used in California in 2001

included 491,992 AFY from municipal facilities, with the additional volume attributed

to private facilities (59). The newest data from the California SWRCB indicates

California municipal wastewater facilities recycled a total of 723,845 AFY in 2009 (59).

This represents an increase of more than 230,000 AFY from levels in 2001, yet once

again falls short of goals for recycling set by the State of California by nearly 300,000

AFY (Figure 2.1 and Table 2.12S). Although the SWRCB 2009 Survey may

underrepresent current reuse volumes due to the low survey response rate, the results

underline a need to identify continuing challenges associated with implementation of

water reuse programs and to evaluate strategies to develop new recycled water

programs and expand existing distribution networks.

14

Northern California context. Our analysis shows that only 20% of the observed

state-wide increase in reuse between 2002 and 2009 occurred in the Northern 48

counties of California, where 120 municipal agencies recycled 127,000 AF in 2002, and

173,000 AF was produced from 143 agencies in 2009. Recycled water programs in

Northern California are generally smaller in volume (median = 347 AFY in 2009) than

programs in the ten Southern California counties (median = 1064 AFY in 2009), where

82 municipal agencies recycled 365,000 AFY of water in 2002, increasing to a total of

551,000 AFY of water in 2009 by 104 agencies (Figure 2.2 and Figure 2.6S). Water

reuse programs are frequent across rural Northern California and agricultural areas in

the Central Valley (Figure 2.2), typically at much lower volumes than urban areas

generating larger volumes of wastewater. Though reuse in Northern California

represents a lesser fraction of overall reuse in the state, challenges associated with the

implementation of smaller, rural programs are important to consider in developing the

total portfolio of state projects. Several larger programs have been implemented over

the last decade in Northern California, and more are likely to be developed in large

urban centers. However, recycled water program size has remained relatively stable on

average in Northern California.

15

Figure 2.2. A snapshot of water reuse facilities in California from the National

Database of Water Reuse Facilities (Annual Production, reported as Facility Production

Average Annual Actual in million gallons) and the California 2009 Municipal Water

Recycling Survey (Annual Reuse, reported as Total Reuse for 2009 in AFY). In the

inset box plot, the boundary of the box indicates the upper and lower quartiles; a line

within the box indicates the median; whiskers above and below the box demarcate 1.5

times the interquartile distance with outlying points also shown.

16

2.4 Drivers of Water Reuse Implementation in Northern CA

Various social, economic, and environmental factors have been identified as drivers

of water reuse by governments and stakeholders globally (2, 8, 56, 61). These driving

forces include: drought, demand due to population and economic growth, wastewater

management, ecological protection, availability near urban areas, and availability of

proven treatment technologies (8, 56). To establish a forum for free-form responses

regarding principal driving forces behind recycled water implementation in Northern

California, respondents first considered the relative importance of several broad

categories of drivers. The fraction of respondents indicating each broad category as a

very important driver or a driver, respectively, was: regulatory requirements (0.59,

0.27), water shortages (0.49, 0.34), economic concerns (0.28, 0.37), recycled water

policy (0.23, 0.49), and influential stakeholders (0.21, 0.33).

To further gauge the extent to which a range of specific factors motivated water

reuse in Northern California, respondents were asked to select factors that drove

program implementation. Amongst a list of 19 specific factors (Table 2.1), 63% of

respondents indicated “wastewater discharge volume requirements” as a driver of

implementation, with 49% of respondents selecting this factor as one of the three most

important drivers. “Water shortages due to reduced supply” was cited as a driver by

65% of respondents and by 42% of respondents as one of the three most important

drivers of implementation. Together, these two factors were cited by 80% of all

respondents. Expressing a common experience for the most important driver of program

implementation, one respondent described that their “initial recycled water program was

established as a wastewater disposal option out of concern for discharge capacity…

Expansions to the recycled water system since 2005 were based on prudent use of water

resources and extending the limited potable supply.”

17

Table 2.1. Percent of respondents indicating a specific factor as a Driver or one of the

three Most Important Drivers. Responses (n = 65) were further categorized as shown

and are sorted from top to bottom by the highest frequency categorized Most Important

Driver.

Categorized Factor

Most Impt. Driver

Driver Specific Factor Most Impt. Driver

Driver

Wastewater discharge requirements

51% 65% wastewater discharge volume requirements

51% 65%

Water supply needs

49%

69% water shortages due to reduced supply

42% 65%

water shortages due to increased demand

17% 42%

seawater intrusion 5% 6%

Local, regional, or state policy and mandates

45%

68% basin plan water quality objectives

25% 43%

regional or local recycled water policy goals or mandates

20% 42%

state recycled water policy goals or mandates

14% 31%

climate change adaptation plans

0% 5%

Institutional control

29%

58% need for reliable water supply 26% 52%

need for increased institutional control of water

3% 20%

Economic/financial incentives

26%

51% availability of federal/state grants or loans

18% 32%

cost of alternative freshwater sources

9% 32%

Ecological goals or requirements

18%

51% ecological protection or enhancement goals

12% 49%

ecological protection or enhancement requirements

6% 20%

Influential stakeholders

11%

34% large volume user(s) 6% 28% citizen initiative 5% 12%

Technological advancements

3% 22% technological advancements 3% 18%

Other 18% 18% other 18% 18%

18

Figure 2.3. Beneficial uses of recycled water in Northern California in 2001 and 2009.

See Supporting Information for a description of categories.

In addition to specific regulatory requirements, state recycled water policy goals or

mandates were selected as a driver by nearly a third (29%) of 2010 Survey respondents

and as one of the three most important drivers by 13% of respondents. Additionally,

24% of respondents selected basin plan water quality objectives as one of the three most

important drivers of implementation. Such objectives may relate to discharge volume

requirements: one respondent who described Basin Plan Water Quality Objectives as

the single most important driver of their program’s implementation stated, “Reducing

our volume discharged to surface water helps us to meet increasingly more stringent

effluent discharge loading requirements.” Drivers of implementation, other than those

shown in Table 2.1, identified by respondents (n = 12 total) often reflected site-specific

conditions including a need for a specific effluent disposal method or location, the cost

of disposal, a water conservation Executive Order, and the need for replacement water.

Notably, “Ecological protection or enhancement goals” were drivers for the

implementation of many programs (49%) but were rarely the most important drivers for

these programs (12%). In 2001 and 2009, only 6-7% of reuse was for natural

system/wildlife enhancement (Figure 2.3).

Controlling wastewater discharge and the role of regulation.

“We needed a method [to] dispose of treated effluent. The only viable

alternative was recycling.”

Results demonstrate that regulatory requirements, such as those limiting discharge

19

of wastewater, have historically played an important role in driving the implementation

of water reuse in Northern California. The California Department of Public Health

establishes state public health criteria for wastewater reclamation via Title 22 for

bacterial quality, treatment types and levels, and facility reliability. Individual Regional

Water Quality Control Boards (RWQCBs) and local water and health agencies may also

develop more stringent policies and programs related to recycled water use (2). In free-

form responses, respondents who cited regulatory requirements as a very important

category of drivers (n = 28) noted a range of specific regulatory pressures that drove the

implementation of their program (see Supporting Information for details). Various

agencies were mandated or recommended to reduce percolation and increase reuse, cap

discharge flows despite population growth, and eliminate point source discharges or

meet dilution requirements in receiving waters during a particular time period (e.g.,

summer months).

Transitioning from wastewater discharge control to recycled water as a

resource.

“The original driver is not the current driver. Currently water supply and

reliability is the most important driver.”

Water shortages are commonly experienced throughout California, with several

severe droughts throughout the period of implementation represented by survey

responses (Figure 2.1). California’s elaborate system of dams, canals, aqueducts,

groundwater basins, and levees mediates the dichotomy between the state’s water

sources and demand centers, where 75% of the state’s precipitation falls north of

Sacramento, and 75% of demand occurs in the population and farming centers to the

south (62). Because of the interconnectedness of water infrastructure in the state and the

dependence of the largest urban centers on imported water, Northern California is not

immune to challenges associated with limited water supplies. The growing awareness

and response to water supply challenges are reflected in agency experiences. Programs

implemented after 1990 were more likely to cite water shortages due to increased

demand as a driver than older programs (p < 0.01) and were somewhat more likely to

indicate water shortages due to reduced supply as a driver (0.1 < p < 0.2, Figure 2.4).

Conversely, wastewater discharge volume requirements were more frequently indicated

20

as one of the three most important drivers of implementation by agencies with reported

implementation dates before 1991 (p < 0.05), suggesting that early implementation of

water reuse in the region was driven more frequently by such regulatory requirements.

Newer recycled water programs were also more likely to cite the need for reliable water

supply as an important driver of implementation (p < 0.01).

Figure 2.4. Results of χ2 analyses by implementation date for specific factors indicated

as one of the Three Most Important Drivers (top) or more generally a Driver of

implementation (bottom).

When expanding on the role of water shortages in driving program implementation

21

(n = 25), respondents frequently cited recycled water as a replacement source for

potable water supplies, where there may be a site-specific water need or shortage (e.g.,

for golf courses or parks) or a cap on freshwater source allocations (e.g., through

externally controlled piped sources). Increased demand without additional supplies was

also evident in several cases of overdraft of groundwater systems leading to degraded

water quality. Water shortages and reliability planning resulting from droughts were

noted separately as driving forces. For example, the 1976-1977 drought was followed

by the adoption of the Policy and Action Plan for Water Reclamation in California by

the SWRCB and subsequent increased funding recycled water planning studies (60).

Such opportunistic funding support strategies may continue to be important to capitalize

on increased incentive for water reuse implementation during periodic drought periods

in the region. Interestingly, only 5% of respondents in the present survey indicated

climate change adaptation plans as a driver of recycled water program implementation.

However, guidance by the California Natural Resources Agency (63) and Department

of Water Resources (5) incorporate recycled water as a drought-proof and sometimes

energy efficient water management strategy to complement climate change adaptation

measures. As these goals filter from state planning to local practices, state policies for

climate change adaptation will likely become more influential in recycled water

implementation.

Although water shortages were not directly an issue during project implementation

for some older projects, anticipated water shortages and need for long-term reliable

sources are now critical issues, especially following the 2007 – 2009 drought in

California. Projects that were implemented initially due to wastewater requirements

may expand or find new benefits of reuse due to water supply challenges. One

respondent illustrated this changing paradigm, stating:

“Fifteen years ago when we started our program, public acceptance was

an issue. People did not understand recycled water, and we spent a lot of

time educating potential customers and marketing recycled water. There

was some 'fear factor' slowing the expansion. However, things have

changed completely with the worsening drought, delta water problems,

climate change awareness, and the public's desire to be 'green' and

22

recycle everything now. We currently cannot get the water out to

customers fast enough.”

Implementation to increase reliability of potable water supplies (e.g., by sustaining

groundwater supplies for drinking), supplement water supply needs, or free up

freshwater entitlements for use elsewhere were described as other drivers of

implementation.

2.5 Challenges for Water Reuse Implementation in Northern

CA

Challenges for water reuse projects include a need for public education, lack of

available funding, recovery of capital costs for dual distribution systems, a need for

improved documentation of economic benefits of water reuse, political support, a need

for additional research for innovative technologies, public perception, flawed or

unevenly applied regulations and standards, and concerns and liability over the

unknown long-term health effects of chemical contaminants (2, 8). When asked to

select factors that hindered program implementation at the respondent’s site from a list

of 20 specific options, 87% of respondents cited financial or economic challenges as

one of the three most important hindrances to water reuse implementation (Table 2.2).

One respondent commenting on the single most important hindrance to implementation

simply stated, “These projects are big ticket items outside the range of a rate base.”

Specific hindrances from the financial or economic challenges list included: availability

of federal/state grants or loans, capital costs for construction of recycling plant facilities,

cost of alternative freshwater sources, costs for pipeline construction, and ongoing

operations & maintenance cost recovery. Together, these factors dominated the

selection of the most important challenges relative to other categories shown in Table

2.2 consistently through time. Despite various sources of policy and financial support

for water reuse in California, lack of sufficient funding may be the main factor

preventing recycling goals from being achieved (59).

23

Table 2.2. Percent of respondents indicating a specific factor as a Hindrance or one of

the three Most Important Hindrances. Responses (n = 54) were further categorized as

shown and are sorted from top to bottom by the highest frequency categorized Most

Important Hindrance.

Categorized Hindrance

Most Impt. Hind.

Hind. Specific Hindrance Most Impt. Hind.

Hind.

Economic/ financial disincentives

87% 94% capital costs for construction of recycling plant facilities

56% 85%

costs for pipeline construction 48% 80% ongoing operations & maintenance cost recovery

26% 61%

availability of federal/state grants or loans

24% 54%

cost of alternative freshwater sources 7% 26% Perceptions and social attitudes

26% 61% perceived human or environmental health risks due to constituents of emerging concern

13% 48%

social attitudes/public perception 13% 33% perception that recycled water will lead to more development

4% 22%

perception that recycled water will reduce property value

4% 6%

Who pays system costs

20% 59% issue of who pays for program capital or operating costs

20% 59%

Regulatory constraints

15% 52% complexities/conflicts of water law and/or regulation

9% 37%

slow regulatory process in permitting 7% 30% Water quality impacts

13% 48% downstream water quality impacts/NPDES constraints

7% 31%

detection of constituents of emerging concern

4% 33%

effluent residuals (e.g., brine) disposal 2% 11% User acceptance

9% 37% user acceptance 9% 37%

Institutional issues

11% 30% institutional coordination 9% 28% loss of projected users 2% 6%

Technical issues/ treatment

7% 31% technical issues/treatment processes 7% 31%

Uncertainty over future recycled water uses

4% 13% uncertainty over future recycled water uses

4% 13%

Other 9% 11% other 9% 11%

24

Challenges in the next most-cited category, public perception and social attitudes,

were indicated as an important hindrance by only 26% of respondents. Specific

hindrances categorized under public perception and social attitudes challenges were:

constituents of emerging concern, perception that recycled water will lead to more

development, perception that recycled water will reduce property value, and the more

general factor of social attitudes/public perception. In addition to those in Table 2.2,

other factors hindering program implementation identified by individual respondents

included soil salinity, lack of seasonal storage, and overcoming opposition from

influential stakeholders.

Economic constraints and financial implications of challenges.

“Generally in the industry and specifically for us, the cost of pipelines is

really the only reason we haven't been recycling more.”

Several examples of recycled water programs in Northern California provide

context for the expected costs of recent treatment facilities and distribution systems. For

16 projects seeking regional federal funding as part of the San Francisco Bay Area

Recycled Water Coalition, the total costs ranged from $220/AF to $3400/AF, with a

$1200/AF median value, assuming a 20-year period for recycled water generated at the

initial project yield (Table 2.13S, (64)). Recycled water deliveries expected for these

projects range from 115 AFY initially to up to 28,000 AFY in the future. A City of Palo

Alto analysis indicates an annualized cost of $2700/AF (over 30 years, in March 2008

dollars) expected for expansion of distribution facilities. This compares with a projected

cost of $1,600/AF by 2015 for wholesale purchase of potable water from the San

Francisco Public Utilities Commission (SFPUC) (65). An earlier phase of the Palo Alto

project completed in 2009 came to approximately $3.4 million/mile of pipeline for

construction base contract of approximately 5 miles of pipeline along US Highway 101

to the neighboring City of Mountain View (66). A project under analysis by the SFPUC

estimates $9.4 million (including a 30% contingency) for approximately 6.5 miles of

pipeline construction costs as part of a $153 million recycled water treatment and

distribution system (67, 68).

2010 Survey respondents were asked to characterize, as quantitatively as possible,

the impact of cited hindrances to implementation in terms of program cost, scope, and

25

timing. Respondents indicated that hindrances led to a change in program cost (n = 9),

reduced program scope (n = 5), delay of implementation (n = 14), project cancellation

(n = 7), or other (n = 1). For a subset of these responses, estimated costs associated with

impacts (n = 21) ranged from $50,000 to almost $100 million per agency. Estimates by

respondents for changes in program cost represented construction cost increases over

time, costs for additional studies, increased staff time, additional testing “beyond

reasonable needs,” “huge” impacts from years of delay, costs to upgrade to tertiary

treatment, costs for new processes, and a combination of changes in program scope,

changes in design, addition of professional consultants or a combination of conveyance

pipes, distribution piping, tanks and pressure stations.

Issues related to institutional coordination were also noted for increasing project

costs. Nearly a third of respondents indicated institutional coordination as a hindrance

to implementation. One example described:

“While water agencies need recycled water to help them with long term

supply issues, they cannot justify the increased costs and thus tend to be

unsupportive. Water agencies are also concerned about loss of revenue

with recycled water projects. If the water agency is not the same as the

recycled water agency (as in our area), implementation of recycled water

projects means a loss of revenue for the water district as customers are

shifted to the recycled water agency. This means that the potable water

agency must raise rates for the remaining customer base, which is very

difficult in today's economic climate.”

Limited role of negative perceptions.

“In 1984, the biggest hindrance was the negative perception by

landowners next to the farms scheduled to receive recycled water today.

Today the biggest hindrance is cost.”

Since the 1970s, a significant amount of research has investigated reasons for public

resistance to recycled water (69, 70). Although public perceptions of risks are identified

as key impediments in the adoption of indirect potable water reuse (71-73), nonpotable

water reuse programs generally receive public support (56). Thus, opposition

surrounding high-profile indirect potable reuse is likely unrepresentative of the

26

landscape of challenges faced by managers of nonpotable reuse programs distributed

throughout California cities and rural areas. A notable contrary case developed when

homeowners actively opposed the use of recycled water for landscape irrigation in

Redwood City, CA (46). While utilities and consultants have developed more

appropriate modes of communicating with the public, some members of the public

remain skeptical about the safety of the practice, especially as projects are proposed in

their community and the likelihood of human contact increases (74-76). Organizational

trust correlates with intended behavior towards using recycled water and may be an area

of further focus for institutional practices to increase public acceptance (77), and

principles of fairness and equity are significant to people’s decision-making (69).

Analyses emphasize the importance of public engagement early during project

conception and continuously throughout planning, design, and construction (2, 56).

The primary drivers of water reuse programs may also influence public opposition

or acceptance. An early public opinion study in California indicated that those who

believed water supply augmentation was necessary in California were somewhat less

likely to be opposed to reclaimed water for drinking than those who did not believe that

water was scarce (70). Consequently, public education efforts to effectively

communicate the need for water reuse are important. In the present study, respondents

who cited wastewater discharge volume requirements as a driver of implementation

were somewhat more likely to also cite a specific factor within the category of public

perceptions and social attitudes as a hindrance (0.2 > p > 0.1). As freshwater supply and

distribution agencies experience increased demands and pressures on existing resources,

greater public awareness of augmentation needs may reduce challenges associated with

public perceptions. Conversely, in communities where the drivers of recycled water are

discharge-based, rather than supply driven, public perception problems may arise more

readily.

“Perceived human or environmental health risks due to constituents of emerging

concern” was cited as a hindrance to implementation by almost half of respondents.

Yet, this factor was not correlated to program implementation date, reminding us that

unknown or unregulated contaminants change in specific definition with time, but have

challenged managers for decades. Concern for residuals in recycled water has been

27

expressed in various forms. In the 1970’s and 80’s, issues of public perception were

difficult to overcome, as recycled water was relatively unfamiliar and long-term safety

of reuse for high-contact uses was unproven. Today, CECs are a topic for technological

research and a source of concern for recycled water managers (13). Noting this issue as

an additional challenge to cost hindrances, one respondent commented that “opponents

are also trying to use the issue of emerging constituents as a way to portray the project

in a negative light.” Public perception of recycled water continues to be an important

non-technical challenge for water reuse implementation, especially with regards to

CECs. However, the present study finds that economic issues, rather than public

perception, stand as the largest hindrance to nonpotable reuse implementation for

Northern California programs.

Responses to recycled water policy. In 2009, the California State Water Resources

Control Board adopted a California Recycled Water Policy “to increase the use of

recycled water from municipal wastewater sources.” Providing statements towards the

beneficial uses of recycled water, the State Water Board “strongly supports recycled

water as a safe alternative to potable water for such approved uses.” Despite the

policy’s stated objectives, whether the water reuse policy will actually accelerate efforts

to develop and maintain new recycled water projects remains unclear. The legislation

itself takes on a hopeful tone by striving for, among other items, increased use of

recycled water “over 2002 levels by at least one million acre-feet per year (AFY) by

2020 and by at least two million AFY by 2030” (54).

Recycled water managers were questioned about their expectations concerning how

the California Recycled Water Policy of 2009 will facilitate or hinder the

implementation of new recycled water programs. Survey responses reveal both support

and trepidation towards the policy, with a greater number of respondents voicing

concern that the policy will hinder project implementation. According to respondents, a

perceived beneficial impact of the policy stems from standardized and consistent

guidance for recycled water projects. For example, the water reuse policy established a

Blue Ribbon Panel for evaluating contaminants of emerging concern that will apply to

all projects across California and also contains language endorsing water reuse under

the California Environmental Quality Act (CEQA) (13). Second, many respondents

28

viewed the policy favorably due to its singular management structure. The

establishment of an overarching permitting process, and of salt and nutrient

management requirements, in particular, drew positive reviews. As one manager put it,

“The standardization of salinity and nutrient management provisions among the various

regional boards should facilitate reuse and make it easier for some projects to get

permitted.” Thus, for water reuse project managers, the provision of administrative,

legal and scientific continuity across state, regional and local agencies was perceived as

the most beneficial aspect of the policy.

Much of the skepticism expressed for the 2009 policy may be traced to funding

issues. A majority of respondents (19 of 30 question responses) felt the policy would

obstruct new projects through onerous regulatory and cost requirements. According to a

number of managers, while statewide project streamlining and standardization is

important, ultimately the fate of projects will depend on adequate funding support. A

common refrain amongst respondents was a concern over added administrative layers

that will arise with new oversight and reporting requirements. In sum, the perceived

presence of additional financial costs and administrative requirements have led nearly 2

of every 3 survey respondents to suggest the 2009 water reuse policy will in some way

hinder new project implementation. From a management perspective, results suggest

that the 2009 policy has done little to alter the perceived drivers and hindrances of water

reuse project implementation for managers in Northern California.

2.6 Significance

A diverse body of responses from the 2010 Survey illuminates a number of

influential drivers of water reuse implementation, including the protection of

ecosystems, meeting wastewater discharge requirements, and needs for water supply

and reliability. We continue to detect manifestations of the intrinsic links between water

supply and quality: threats of long-term diminished water quality (e.g., seawater

intrusion) necessitates new water conservation and reuse measures, while new water

supplies of altered quality may galvanize community opposition. Although water supply

agencies increasingly face challenges associated with population growth and drought,

wastewater agencies have traditionally approached recycled water as an issue of

29

disposal. This push/pull duality that either push implementation forward (via regulatory

requirements for wastewater discharge) or pull agencies into recycled water programs

(by increased demand for water) is apparent. Results provide evidence of changing

perspectives towards recycled water management, from a waste disposal issue towards a

water supply resource opportunity.

Failure to meet statewide reuse goals results largely from lack of sufficient funding

for water recycling, as the cheapest recycled water opportunities have already been

exploited (2). Following three years of drought and recent passage of the Safe, Clean,

and Reliable Drinking Water Supply Act of 2010 by the State of California that

included $1.25 billion general obligation bond proposal for Water Recycling and Water

Conservation, the physical and political climates may be ripe for aggressive

implementation of new water reuse programs, where financially viable, socially

accepted, and technically sound. Yet the legislature’s 2010 decision to postpone the

water bond initiative for at least two years (62) is testament to the realities of financial

limitations for new water infrastructure in California.

Supporting Information Available. Contains (1) methodological details, (2) a

brief description of recycled water policy and regulation in California, and (3)

additional analysis and summary tables.

Acknowledgement. We thank the numerous participants in the project and survey

respondents for their generous donation of time and thoughtful contributions. This work

was funded by the Stanford University Woods Institute for the Environment

Environmental Venture Projects, the Bill Lane Center for the American West, the NSF

Graduate Research Fellowship Program, and the NSF Engineering Research Center for

Re-inventing Urban Water Infrastructure (UrbanWaterERC.org). We especially thank

Sophie Egan for detailed technical assistance.

Publication Information. Reproduced with permission from Environmental

Science & Technology, submitted for publication. Unpublished work copyright 2011

American Chemical Society.

30

2.7 Supporting Information

Description of data sources. Invited and respondent agency locations for the

present study (2010 Survey) are shown by county in Figure 2.5S. The California State

Water Resources Control Board (SWRCB) 2001 Water Recycling Survey (2001

Survey, (3)) compiled data on planned direct reuse of treated municipal wastewater in

California, excluding industrial reuse. The 2009 California Municipal Wastewater

Recycling Survey (2009 Survey) was conducted by the SWRCB Water Recycling

Funding Program (WRFP), compiling volumetric data for municipal wastewater

recycling facilities to determine progress towards goals set in 2008 and 2009 by the

California State Water Board Strategic Plan Update and the Recycled Water Policy.

Although the 2009 Survey represents the best available information on current water

recycling volumes in California, the survey itself was completed initially by only 18%

of agencies invited to participate (118 agencies responding). SWRCB staff collected

additional data through recycled water annual reports, agency websites, telephone

communication, and carry-over of data from the 2001 Survey, assuming volumes

reported in 2001 remained the same in 2009. Updates in the beneficial use categories

noted in the 2009 Survey as compared to the 2001 Survey include: Golf Course

Irrigation was separately quantified from Landscape Irrigation in the 2009 Survey;

Wildlife Habitat & Miscellaneous Enhancement in the 2001 Survey was labeled Natural

Sys. Restoration, Wetlands, Wildlife Habitat in the 2009 Survey; and Wastewater

Treatment Plant uses were not specified in the 2009 survey (included as Other for 2001

Survey results in manuscript Figure 2.3). Recreational Impoundments was 0% in 2009

and <1% in 2001. The 2001 Survey included private agencies, which were excluded in

the 2009 Survey. Annual flows reported in Northern and Southern CA for the 2001 and

2009 Surveys are compared in Figure 2.6S.

The National Database of Water Reuse Facilities (National Database, (58)) is

maintained by the WateReuse Association. The database was initially populated in 2006

and may be updated manually by agency representatives (78). The National Database

was accessed for the California Query in January 2010 for the 2010 Survey. Private

facilities and individuals from the 2001 Survey were excluded from the initial

31

distribution list for the 2010 Survey, and surveys were not sent to several agencies that

requested exclusion when compiling contact information. Additional information on

survey respondent characteristics was collected from the member-accessible portion of

the National Database in July-August 2010. Data for average annual production

volumes (million gallons) reported by water reuse facilities in California were accessed

from the National Database on August 16, 2010. Facility locations for mapping

purposes were obtained from zip codes available in the National Database or from

online searches for the facility location. Data from the 2010 Survey are not intended to

update or compare directly with the 2009 Survey results. 2010 Survey respondent

characteristics are described to assess representation of surveyed agencies compared to

agencies in the sampling frame.

Questionnaire. For the 2010 Survey, the invited participant was asked to complete

the survey or designate an individual familiar with the implementation of the water

reuse program to complete the survey. Coding entries for respondent “Position” shows

that 72% of respondents were directors/managers of the agency/utility (e.g., Public

Works Director, General Manager, Deputy Director, Division Manager), 14% of

respondents were engineers (e.g., Chief Engineer, Senior Engineer), 10% were

operators or other technical positions (e.g., Chief Plant Operator, Operations Director),

and 4% of respondents did not list a position. In the final dataset, one response per

agency was retained. Three agencies submitted two responses, so the more complete

survey was retained. One respondent did not input agency identification information;

responses from this individual were not excluded. In statistical analysis, quantitative

results were omitted for two respondents who indicated that their facilities do not

produce recycled water; these responses were retained for qualitative analysis purposes.

In the questionnaire, reclaimed or recycled water were noted to be synonymous and

defined as: “water that is used more than one time before it passes back into the natural

water cycle, or wastewater that has been treated to a level that allows for its reuse for a

beneficial purpose.” Driving forces and barriers to water reuse program implementation

in the region were examined through several qualitative and quantitative components.

Respondents rated the importance of five broad categories of potential drivers

(economic concerns, regulatory requirements, recycled water policy, water shortages,

32

and influential stakeholder groups) and five broad categories of hindrances (cost

recovery, human health or water quality concerns, institution/management issues,

influential stakeholder groups, and salt and nutrient management) for program

implementation on a three-point scale (Not a Driver/Hindrance, Driver/Hindrance, or

Very Important Driver/Hindrance). Additionally, from lists of specific factors

(manuscript Tables 2.1 and 2.2) respondents selected all factors that were considered

drivers/hindrances of program implementation and three factors considered to be the

most important drivers/hindrances for the implementation of their agency’s recycled

water program. In several cases, respondents marked a specific factor as one of the three

most important drivers/hindrances without selecting this same factor as a

driver/hindrance. In these cases, responses were updated such that all most important

drivers/hindrances were necessarily considered drivers/hindrances. Respondents were

given opportunity to insert other categorical and specific factors. To encourage free-

form descriptive responses and for clarification, respondents were further invited to

describe components of the broad categories previously marked as very important.

Subsequent questions and focus sections on the role of influential stakeholders,

constituents of emerging concern, and ecosystem enhancements will be the subject of

future analysis.

Categorization and statistical tests. Response data were categorized by self-

reported values of total reclaimed water use and date of program implementation such

that an equal number of respondents were in each category. Thus, recycled water

program size was operationally categorized as programs with self-reported total annual

reclaimed water deliveries greater or less than the median value (990 AFY), and the

date of implementation was categorized as either before or after the median date of

implementation, 1991. Analysis was omitted when the frequency in any contingency

table bin was less than five and p > 0.1 or when the frequency in any bin was equal to

zero. When the frequency of a bin was between 1 and five and p < 0.1 for the

association test, the results were listed in parentheses. Chi square analyses are

summarized in Table 2.3S to 2.10S.

33

Characteristics of 2010 Survey respondent agencies. Based on cross checking of

categorizations available in the member-accessible National Database, which lists all

respondent agencies in the 2010 Survey, respondent agencies represent Public Utilities

(45%), Private Utilities (6%), or Special Districts (23%). The remaining 23% of

respondent agencies were not categorized in the National Database. Most agencies

(76%) conduct “Both Production and Distribution” of recycled water, with 7%

categorized as “Only Production” and 7% as “Only Distribution;” 6% of respondents

represent other agencies (that may have been involved by funding, management of

construction, etc.), and 4% (3 agencies) are not categorized. Non-zero total annual

reclaimed water deliveries reported for 63 respondent agencies ranged from 6 AFY to

28,000 AFY, with a median value of 920 AFY. With the inclusion of delivery rates

from the 2009 Survey for three agencies where a value was not entered and data were

available in the SWRCB survey, the median reclaimed water delivery value was 990

AFY. Responses captured a larger flow range than reported in the 2009 Survey and

yielded a larger median annual recycled water volume (Figure 2.7S).

Historically, agricultural water reuse predominated in California, occurring in the

Central Valley in places where farmland was located adjacent to wastewater treatment

facilities (2). In recent years, agricultural reuse volumes have remained relatively stable,

becoming a smaller fraction of total reuse as new industrial and commercial uses are

developed. However, significant population growth, particularly in the Central Valley,

creates challenges for new or increased wastewater discharge in largely agricultural

areas, especially for environments with limited assimilative capacity (55). Beneficial

uses of recycled water reported in the 2009 and 2010 Surveys are displayed in Table

2.11S. Additionally, between 2002 and 2009, fourteen agencies in Northern California

reported reduced volumes, averaging 61% of 2002 levels recycled in 2009, including

two null values reported in 2009 for previously recycling agencies. From 1970 to 1977

and 1977 to 1987 various agencies were dropped from the latter surveys of water reuse

in California due to: (1) discontinued reuse, (2) treatment plants shut down or converted

to a wet weather plant with the construction of regional facilities, (3) changes in

reporting criteria or interpretation, (4) name changes, or (5) inadvertent omission (60).

34

California recycled water policy and regulation. Several milestones in recycled

water policy and regulation have been implemented in California since the Porter

Cologne Act of 1969 that formulated state and regional water quality control regulatory

bodies (e.g., Table 2.12S). As required by Assembly Bill No. 331 passed by the

California Legislature and signed into California law on October 7, 2001, a Recycled

Water Task Force was convened by the California Department of Water Resources

(DWR) in 2002 to address issues related to the impediments or constraints related to

increasing water recycling. Six issues areas were identified: funding/CALFED

coordination; public information, education, and outreach; plumbing code/cross-

connection control; regulations and permitting; economics of water reclamation; and

science and health issues of indirect potable reuse (2). In February 2009, the California

SWRCB adopted a Recycled Water Policy to encourage expanded reuse in California

(54). In a progressive move to address issues of trace contaminants in recycled water,

the policy led to a series of recommendations on chemicals of emerging concern,

published in 2010 (13).

Funds have been made available through state and federal financial assistance and a

series of bond initiatives from 1978 – 2002, totaling close to $132 million in planning

and construction grants and $509 million in low interest loans for water recycling

projects distributed by the State Water Board from 1978 to 2006 (79). An additional

$275 million in construction grants and $11.5 million in planning grants was distributed

by the U.S. Bureau of Reclamation Title XVI program through 2007 in California (79).

The 2003 Recycled Water Task Force estimated an $11 billion dollar investment was

required at the time to reach 2030 water recycling goals (2), and the 2008-2012 State

Water Board Strategic Plan Update estimates $300 million in annual grants and loans

are required (55).

2010 Survey respondents noted specific sources of regulatory requirements,

mandates, or strong recommendations that influenced program implementation,

including: National Pollutant Discharge Elimination System (NPDES) permits,

Regional Water Quality Control Boards (RWQCBs) and the SWRCB, Basin Plans, the

Porter Cologne Act, Waste Discharge Requirements, waste production located outside

of city limits for a treatment system (e.g., for a college campus), a water right

35

requirement to move golf courses off river water to recycled water, construction

approvals for new stand-alone subdivisions requiring production of recycled water, and

California Energy Commission recommendations to use recycled water for power plant

cooling water. Several other responses inserted by respondents as very important drivers

(n = 9), including flow limits on discharges, need for wastewater disposal, and RWQCB

requirements, supported the theme of regulatory requirements as important drivers of

implementation.

Small- versus large-volume recycled water programs. A chi square analysis

indicated that respondents representing the lower 50% of reuse programs by volume

(small-volume reuse programs) were somewhat more likely to cite wastewater

discharge volume requirements as a driver for implementation (0.1 < p < 0.2). At the

same time, small-volume programs were somewhat less likely to indicate the

availability of federal and state grants and loans as a driver of program implementation

(0.1 < p < 0.2) than programs with total annual reclaimed wastewater deliveries greater

than the median reported value. One respondent noted, “In rural foothill communities,

regulations tend to push agencies towards recycling because there are no viable

alternatives.” An effort to coordinate funding proposals was successfully implemented

by the San Francisco Bay Area Recycled Water Coalition (BARWC), a partnership of

17 agencies formed to secure federal funding under Title XVI of the 1992 Reclamation

Wastewater and Groundwater Study & Facilities Act. Typical project costs and the

federal shares of funding for the BARWC are shown in Table 2.13S (64). This

exemplifies the potential of collaborative funding efforts. Such efforts to pool resources

regionally, rather than compete individually for funding, may prove beneficial for rural

communities seeking facility upgrades for implementation of smaller-volume programs.

Respondents representing small-volume programs also provided somewhat greater

recognition of the role of influential stakeholders in driving program implementation,

citing either citizen initiatives or large volume users as drivers of implementation more

frequently than respondents from large volume programs (0.1 < p < 0.2). Larger volume

programs were more likely than smaller volume programs to be implemented due to a

need for reliable water supplies (0.1 < p < 0.2).

36

Impacts of hindrances. Cost impacts estimated by respondents for project delays

caused by cited hindrances specifically ranged from $50,000 to $25,000,000 due to time

delays, increased material costs, California Environmental Quality Act (CEQA)

documentation, staff time to contact interested parties, and for additional storage ponds

required for groundwater recharge. “Delays in customer connections caused by slow

regulatory approval and issues related to customer acceptance” led to lost revenues

estimated at $1,000,000 for one project. Another respondent noted that lack of local

project funding led to the loss of state grant funding and the cancellation of a project.

One respondent described a revenue challenge experienced in transitioning to recycled

water, “Recycled water is also typically discounted by 20 to 25% in Northern California

to encourage customers to connect, so even if the water and recycled water agency is

the same, the agency loses revenue by connecting recycled water customers while also

having to pay for the expensive recycled water infrastructure required.” A contingency table chi square analysis indicated that respondents who identified at

least one negative impact of stated hindrances to implementation were somewhat more

likely to cite costs for pipeline construction as one of the three most important

challenges as compared to those agencies who did not experience significant negative

impacts (p < 0.1). Programs recycling more than the median flow were also somewhat

more likely to cite costs for pipeline construction as a hindrance than smaller projects

(0.1 < p < 0.2), underlining the need for support of distribution system costs as recycled

water receiving sites are located further from recycled water production facilities.

37

Figure 2.5S. Distribution of survey invitations and responses collected from Northern

California counties. The number of respondents from a given county is displayed, with

the number of agencies invited to participate in a county indicated in parentheses.

38

Figure 2.6S. Agencies binned by annual recycled water flow (AFY) in 2001 and 2009.

39

1

10

100

1000

104

105

2010 Survey 2009 Survey

Tota

l R

euse (

AF

Y)

Figure 2.7S. Representation of total annual reclaimed water deliveries reported from

the 2010 Survey of Northern California agencies1 (n = 64) and the State Water

Resources Control Board 2009 Municipal Wastewater Recycling Survey (n = 143).

Both plots exclude zero-volume survey responses.

1 Includes reclaimed water deliveries values from the 2009 Survey for three agencies in which a value was not entered in the 2010 Survey and data was available in the SWRCB 2009 Survey.

40

Table 2.3S. C

hi square analysis of drivers of program im

plementation by self-reported date of im

plementation. O

lder and recent

implem

entation dates were operationally defined as before and after the reported m

edian value (1991), respectively. 1

D

river M

ost Important D

river

n = 54

Statistical tests (df = 1)

n = 54 Statistical tests

(df = 1)

Older

impl.

date, %

Recent

impl.

date, %

χ2

p O

lder im

pl. date, %

Recent

impl.

date, %

χ2

p

need for reliable water supply

17%

39%

8.45 0.004

9%

19%

1.98 0.16

water shortages due to increased dem

and 9%

30%

8.13

0.004 2%

13%

(4.94)

(0.03) w

astewater discharge volum

e requirements

37%

28%

1.98 0.16

33%

19%

4.59 0.03

water shortages due to reduced supply

22%

44%

(8.47) (0.004)

15%

26%

2.23 0.14

basin plan water quality objectives

26%

19%

1.29 0.26

19%

7%

(3.56) (0.06)

ecological protection or enhancement goals

26%

24%

0.13 0.72

7%

4%

regional or local recycled water policy goals or

mandates

19%

19%

0.01 0.93

4%

11%

state recycled water policy goals or m

andates 15%

15%

0.01

0.94 6%

7%

availability of federal/state grants or loans

6%

22%

(6.72) (0.01)

2%

11%

(3.84) (0.05)

cost of alternative freshwater sources

7%

24%

(6.14) (0.01)

2%

7%

1 A

nalysis is omitted w

hen the frequency in any contingency table bin was less than five and p > 0.1 or w

hen the frequency in any bin was equal to

zero. When the frequency of a bin w

as between 1 and five and p < 0.1 for the association test, the results are listed in parentheses.

41

Tab

le 2

.4S.

Chi

squ

are

anal

ysis

of h

indr

ance

s to

pro

gram

impl

emen

tatio

n by

sel

f-re

porte

d da

te o

f im

plem

enta

tion.

Old

er a

nd re

cent

impl

emen

tatio

n da

tes w

ere

oper

atio

nally

def

ined

as b

efor

e an

d af

ter t

he re

porte

d m

edia

n va

lue

(199

1), r

espe

ctiv

ely.

1

H

indr

ance

M

ost I

mpo

rtan

t Hin

dran

ce

n

= 44

St

atis

tical

test

s (d

f = 1

) n

= 44

St

atis

tical

test

s (d

f = 1

)

Old

er

impl

. da

te, %

Rec

ent

impl

. da

te, %

χ

2 p

Old

er

impl

. da

te, %

Rec

ent

impl

. da

te, %

χ

2 p

avai

labi

lity

of fe

dera

l/sta

te g

rant

s or l

oans

20

%

34%

2.

21

0.14

7%

14

%

user

acc

epta

nce

16%

27

%

1.59

0.

21

7%

2%

dow

nstre

am w

ater

qua

lity

impa

cts/

NPD

ES c

onst

rain

ts

11%

20

%

1.19

0.

28

0%

5%

dete

ctio

n of

con

stitu

ents

of e

mer

ging

con

cern

11

%

20%

1.

19

0.28

2%

0%

co

sts f

or p

ipel

ine

cons

truct

ion

39%

48

%

30%

25

%

0.88

0.

35

perc

eive

d hu

man

or e

nviro

nmen

tal h

ealth

risk

s due

to

cons

titue

nts o

f em

ergi

ng c

once

rn

23%

32

%

0.78

0.

38

5%

9%

issu

e of

who

pay

s for

pro

gram

cap

ital o

r ope

ratin

g co

sts

27%

36

%

0.73

0.

39

7%

11%

co

mpl

exiti

es/c

onfli

cts o

f wat

er la

w a

nd/o

r reg

ulat

ion

16%

23

%

0.48

0.

49

5%

2%

capi

tal c

osts

for c

onst

ruct

ion

of re

cycl

ing

plan

t fac

ilitie

s 41

%

45%

30

%

27%

0.

42

0.52

so

cial

atti

tude

s/pu

blic

per

cept

ion

18%

18

%

0.05

0.

82

5%

9%

slow

regu

lato

ry p

roce

ss in

per

mitt

ing

16%

16

%

0.04

0.

84

2%

5%

ongo

ing

oper

atio

ns &

mai

nten

ance

cos

t rec

over

y 30

%

32%

0.

00

0.94

14

%

14%

0.

03

0.85

pe

rcep

tion

that

recy

cled

wat

er w

ill le

ad to

mor

e de

velo

pmen

t 11

%

14%

0.

03

0.86

2%

2%

te

chni

cal i

ssue

s/tre

atm

ent p

roce

sses

5%

23

%

(6.3

8)

(0.0

1)

2%

0%

inst

itutio

nal c

oord

inat

ion

5%

20%

(5

.13)

(0

.02)

2%

7%

ef

fluen

t res

idua

ls (e

.g. b

rine)

dis

posa

l 2%

11

%

(2.6

9)

(0.1

0)

0%

2%

1 Ana

lysi

s is o

mitt

ed w

hen

the

freq

uenc

y in

any

con

tinge

ncy

tabl

e bi

n w

as le

ss th

an fi

ve a

nd p

> 0

.1 o

r whe

n th

e fr

eque

ncy

in a

ny b

in w

as e

qual

to

zero

. Whe

n th

e fr

eque

ncy

of a

bin

was

bet

wee

n 1

and

five

and

p <

0.1

for t

he a

ssoc

iatio

n te

st, t

he re

sults

are

list

ed in

par

enth

eses

.

42

Table 2.5S. C

hi square analysis of categorized drivers of program im

plementation by self-reported date of im

plementation. O

lder and

recent implem

entation dates were operationally defined as before and after the reported m

edian value (1991), respectively. 1

D

river M

ost Important D

river

n = 54

Statistical tests (df = 1)

n = 54 Statistical tests

(df = 1)

O

lder im

pl. date, %

Recent

impl.

date, %

χ2

p O

lder im

pl. date, %

Recent

impl.

date, %

χ2

p

Institutional control 19%

41%

8.39

0.004 11%

20%

1.80

0.18 W

ater supply needs 24%

44%

(7.10)

(0.01) 15%

33%

6.02

0.01 Influential stakeholders

9%

20%

2.77 0.10

6%

2%

Wastew

ater discharge requirements

37%

28%

1.98 0.16

33%

19%

4.59 0.03

Economic/financial incentives

20%

30%

1.35 0.24

7%

17%

Other policy and regulation

33%

33%

0.02 0.88

22%

20%

0.13 0.72

Ecological goals or requirements

26%

26%

0.01 0.91

11%

6%

1 A

nalysis is omitted w

hen the frequency in any contingency table bin was less than five and p > 0.1 or w

hen the frequency in any bin was equal to

zero. When the frequency of a bin w

as between 1 and five and p < 0.1 for the association test, the results are listed in parentheses.

43

Tab

le 2

.6S.

Chi

squ

are

anal

ysis

of c

ateg

oriz

ed h

indr

ance

s to

pro

gram

impl

emen

tatio

n by

sel

f-re

porte

d da

te o

f im

plem

enta

tion.

Old

er

and

rece

nt im

plem

enta

tion

date

s wer

e op

erat

iona

lly d

efin

ed a

s bef

ore

and

afte

r the

repo

rted

med

ian

valu

e (1

991)

, res

pect

ivel

y.1

H

indr

ance

M

ost I

mpo

rtan

t Hin

dran

ce

n

= 44

St

atis

tical

test

s (d

f = 1

) n

= 44

St

atis

tical

test

s (d

f = 1

)

O

lder

im

pl.

date

, %

Rec

ent

impl

. da

te, %

χ

2 p

Old

er

impl

. da

te, %

Rec

ent

impl

. da

te, %

χ

2 p

Wat

er q

ualit

y im

pact

s 14

%

27%

2.

53

0.11

2%

5%

U

ser a

ccep

tanc

e 16

%

27%

1.

59

0.21

7%

2%

Pu

blic

per

cept

ion

and

soci

al a

ttitu

des

36%

32

%

1.19

0.

28

11%

16

%

0.24

0.

62

Econ

omic

/fina

ncia

l dis

ince

ntiv

es

48%

50

%

0.93

0.

33

43%

45

%

Who

pay

s for

pro

gram

cos

ts

27%

36

%

0.73

0.

39

7%

11%

R

egul

ator

y co

nstra

ints

25%

30

%

0.08

0.

78

7%

7%

Tech

nica

l iss

ues/

treat

men

t pro

cess

es

5%

23%

(6

.38)

(0

.01)

2%

0%

In

stitu

tiona

l iss

ues

7%

20%

(3

.42)

(0

.06)

5%

7%

1 A

naly

sis i

s om

itted

whe

n th

e fr

eque

ncy

in a

ny c

ontin

genc

y ta

ble

bin

was

less

than

five

and

p >

0.1

or w

hen

the

freq

uenc

y in

any

bin

was

equ

al to

ze

ro. W

hen

the

freq

uenc

y of

a b

in w

as b

etw

een

1 an

d fiv

e an

d p

< 0.

1 fo

r the

ass

ocia

tion

test

, the

resu

lts a

re li

sted

in p

aren

thes

es.

44

Table 2.7S. C

hi square analysis of drivers of program im

plementation by self-reported total annual reclaim

ed water use. Sm

all and

large volume program

s were operationally defined as agencies w

ith total reclaimed w

ater use less than and greater than the reported

median value, respectively. 1

D

rivers T

hree Most Im

portant Drivers

n = 65

Statistical tests (df=1)

n = 65 Statistical tests

(df=1)

Sm

all V

olume

%

Large

Volum

e %

χ

2 p

Small

Volum

e %

Large

Volum

e %

χ

2 p

availability of federal/state grants or loans 8%

23%

6.79

0.01 3%

14%

5.11

(0.02) need for increased institutional control of w

ater 14%

5%

3.91

(0.05) 2%

2%

large volume user(s)

18%

9%

3.03 0.08

6%

0%

need for reliable water supply

20%

31%

2.60 0.11

11%

15%

0.60 0.44

wastew

ater discharge volume requirem

ents 35%

28%

2.09

0.15 29%

22%

1.87

0.17 regional or local recycled w

ater policy goals or m

andates 23%

15%

1.89

0.17 11%

6%

state recycled water policy goals or m

andates 11%

18%

1.65

0.20 6%

6%

w

ater shortages due to reduced supply 29%

34%

0.37

0.54 20%

20%

0.01

0.92 cost of alternative freshw

ater sources 17%

15%

0.12

0.73 5%

5%

technological advancem

ents 8%

9%

0.08

0.78 2%

2%

ecological protection or enhancem

ent goals 23%

25%

0.02

0.90 5%

8%

basin plan w

ater quality objectives 20%

20%

0.01

0.92 14%

11%

0.42

0.52 w

ater shortages due to increased demand

20%

20%

0.01 0.92

9%

6%

1 A

nalysis is omitted w

hen the frequency in any contingency table bin was less than five and p > 0.1 or w

hen the frequency in any bin was equal to

zero. When the frequency of a bin w

as between 1 and five and p < 0.1 for the association test, the results are listed in parentheses.

45

Tab

le 2

.8S.

Chi

squ

are

anal

ysis

of h

indr

ance

s to

pro

gram

impl

emen

tatio

n so

rted

by to

tal a

nnua

l rec

laim

ed w

ater

use

. Sm

all a

nd la

rge

volu

me

prog

ram

s wer

e op

erat

iona

lly d

efin

ed a

s age

ncie

s with

tota

l rec

laim

ed w

ater

use

less

than

and

gre

ater

than

the

repo

rted

med

ian

valu

e, re

spec

tivel

y.1

H

indr

ance

s T

hree

Mos

t Im

port

ant H

indr

ance

s

n

= 52

St

atis

tical

test

s (d

f = 1

) n

= 52

St

atis

tical

test

s (d

f = 1

)

Sm

all

Vol

ume

%

Lar

ge

Vol

ume

%

χ2

p Sm

all

Vol

ume

%

Lar

ge

Vol

ume

%

χ2

p

cost

of a

ltern

ativ

e fr

eshw

ater

sour

ces

8%

19%

2.

38

(0.1

2)

2%

6%

dete

ctio

n of

con

stitu

ents

of e

mer

ging

con

cern

10

%

21%

2.

07

0.15

0%

4%

sl

ow re

gula

tory

pro

cess

in p

erm

ittin

g 10

%

21%

2.

07

0.15

2%

6%

pe

rcep

tion

that

recy

cled

wat

er w

ill le

ad to

mor

e de

velo

pmen

t 6%

15

%

2.00

(0

.16)

2%

2%

cost

s for

pip

elin

e co

nstru

ctio

n 33

%

46%

1.

72

(0.1

9)

19%

29

%

0.73

0.

39

user

acc

epta

nce

13%

25

%

1.63

0.

20

4%

6%

inst

itutio

nal c

oord

inat

ion

10%

17

%

0.84

0.

36

8%

2%

2.55

(0

.11)

pe

rcei

ved

hum

an o

r env

ironm

enta

l hea

lth ri

sks d

ue to

co

nstit

uent

s of e

mer

ging

con

cern

19

%

27%

0.

36

0.55

6%

6%

capi

tal c

osts

for c

onst

ruct

ion

of re

cycl

ing

plan

t fa

cilit

ies

40%

44

%

29%

27

%

0.82

0.

37

dow

nstre

am w

ater

qua

lity

impa

cts/

NPD

ES c

onst

rain

ts

13%

17

%

0.05

0.

82

4%

2%

ongo

ing

oper

atio

ns &

mai

nten

ance

cos

t rec

over

y 27

%

33%

0.

03

0.86

10

%

17%

0.

84

0.36

is

sue

of w

ho p

ays f

or p

rogr

am c

apita

l or o

pera

ting

cost

s 27

%

33%

0.

03

0.86

10

%

10%

0.

07

0.79

com

plex

ities

/con

flict

s of w

ater

law

and

/or r

egul

atio

n 17

%

19%

0.

02

0.89

6%

4%

so

cial

atti

tude

s/pu

blic

per

cept

ion

15%

17

%

0.01

0.

93

6%

8%

1 Ana

lysi

s is o

mitt

ed w

hen

the

freq

uenc

y in

any

con

tinge

ncy

tabl

e bi

n w

as le

ss th

an fi

ve a

nd p

> 0

.2 o

r whe

n th

e fr

eque

ncy

in a

ny b

in w

as e

qual

to

zero

. Whe

n th

e fr

eque

ncy

of a

bin

was

bet

wee

n 1

and

five

and

p <

0.1

for t

he a

ssoc

iatio

n te

st, t

he re

sults

are

list

ed in

par

enth

eses

.

46

Table 2.9S. C

hi square analysis of categorized drivers of program im

plementation by self-reported total annual reclaim

ed water use.

Small and large volum

e programs w

ere operationally defined as agencies with total reclaim

ed water use less than and greater than the

reported median value, respectively. 1

D

rivers T

hree Most Im

portant Drivers

n = 65

Statistical tests (df = 1)

n = 65 Statistical tests

(df = 1)

Sm

all V

olume

%

Large

Volum

e %

χ

2 p

Small

Volum

e %

Large

Volum

e %

χ

2 p

Wastew

ater discharge requirements

35%

28%

2.09 0.15

29%

22%

1.87 0.17

Influential stakeholders 20%

12%

1.99

0.16 11%

0%

8.09

(0.004) Econom

ic/financial incentives 20%

29%

1.87

0.17 11%

14%

0.26

0.61 O

ther policy and regulation 29%

34%

0.37

0.54 23%

18%

0.74

0.39 Institutional control

26%

31%

0.37 0.54

12%

17%

0.55 0.46

Ecological goals or requirements

23%

26%

0.14 0.71

8%

11%

0.34 0.56

Technological advancements

11%

9%

0.14 0.71

2%

2%

1 A

nalysis is omitted w

hen the frequency in any contingency table bin was less than five and p > 0.1 or w

hen the frequency in any bin was equal to

zero. When the frequency of a bin w

as between 1 and five and p < 0.1 for the association test, the results are listed in parentheses.

47

Tab

le 2

.10S

. Chi

squ

are

anal

ysis

of h

indr

ance

s to

pro

gram

impl

emen

tatio

n by

sel

f-re

porte

d to

tal a

nnua

l rec

laim

ed w

ater

use

. Sm

all

and

larg

e vo

lum

e pr

ogra

ms

wer

e op

erat

iona

lly d

efin

ed a

s ag

enci

es w

ith t

otal

rec

laim

ed w

ater

use

les

s th

an a

nd g

reat

er t

han

the

repo

rted

med

ian

valu

e, re

spec

tivel

y.1

H

indr

ance

s T

hree

Mos

t Im

port

ant H

indr

ance

s

n =

52

Stat

istic

al te

sts (

df =

1)

n =

52

Stat

istic

al te

sts (

df =

1)

Sm

all

Vol

ume

%

Lar

ge

Vol

ume

%

χ2

p Sm

all

Vol

ume

%

Lar

ge

Vol

ume

%

χ2

p

Reg

ulat

ory

cons

train

ts

19%

33

%

1.88

0.

17

6%

10%

U

ser a

ccep

tanc

e 13

%

25%

1.

63

0.20

4%

6%

0.

08

0.77

In

stitu

tiona

l iss

ues

10%

19

%

1.39

0.

24

8%

4%

Publ

ic p

erce

ptio

n an

d so

cial

atti

tude

s 25

%

35%

0.

55

0.46

10

%

15%

0.

41

0.52

W

ater

qua

lity

impa

cts

19%

27

%

0.36

0.

55

8%

4%

1.15

0.

28

Who

pay

s for

pro

gram

cos

ts

27%

33

%

0.03

0.

86

10%

10

%

0.07

0.

79

Tech

nica

l iss

ues/

treat

men

t pro

cess

es

13%

15

%

0.00

0.

96

4%

4%

1 Ana

lysi

s is o

mitt

ed w

hen

the

freq

uenc

y in

any

con

tinge

ncy

tabl

e bi

n w

as le

ss th

an fi

ve a

nd p

> 0

.1 o

r whe

n th

e fr

eque

ncy

in a

ny b

in w

as e

qual

to

zero

. Whe

n th

e fr

eque

ncy

of a

bin

was

bet

wee

n 1

and

five

and

p <

0.1

for t

he a

ssoc

iatio

n te

st, t

he re

sults

are

list

ed in

par

enth

eses

.

48

Table 2.11S. R

epresentation of recycled water beneficial uses from

the 2010 Survey of Northern C

alifornia (n = 69) agencies and the

State Water R

esources Control B

oard 2009 Municipal W

astewater R

ecycling Survey (n = 143). Volum

es for each beneficial use were

not collected in the 2010 Survey.

A

gricultural Irrigation

Landscape

Irrigation1

Industrial

Wildlife

Habitat

Enhance-m

ent. 2

Com

mercial/

Residential

Buildings 3

Groundw

ater R

echarge R

ecreational Im

poundment

Geotherm

al/ E

nergy Production

Other

2009 Total R

euse (AFY

) 4 91,360

22,556 13,975

12,071 7,371

2,500 0

12,665 10,049

2009 Frequency Percent

60.8%

32.9%

7.7%

4.9%

2.8%

0.7%

0.0%

0.7%

9.1%

2010 Frequency Percent

42.0%

52.2%

24.6%

21.7%

7.2%

11.6%

7.2%

N/A

24.6%

1 Landscape Irrigation and G

olf Course Irrigation w

ere combined from

the 2009 Survey results in this table. 2 Labeled as N

atural Sys. Restoration, Wetlands, W

ildlife Habitat in 2009 Survey results.

3 Labeled as Com

mercial in 2009 Survey results.

4 No agencies in N

orthern California indicated Seaw

ater Intrusion Barrier, Surface Water Augm

entation, or Indirect Potable Reuse as beneficial uses in the 2009 Survey results.

49

Table 2.12S. Milestones for California water reuse and statewide recycling goals.1

Year Name (Ref.) Description 1967 California Legislature supports

policies and laws to promote water recycling

Declared that “the state undertake all possible steps to encourage development of water reclamation facilities…to help meet the growing water requirements of the state.” (California Water Code, Section 13512)

1969 Porter-Cologne Water Quality Control Act (62)

Established State Water Resources Control Board and nine Regional Water Quality Control Boards (California Water Code, Division 7).

1974 Water Reuse Law of 1974 Enacted with the mission that “the primary interest of the people of the State in the conservation of all available water resources requires the maximum reuse of reclaimed water in the satisfaction of requirements for beneficial uses of water.” (California Water Code, Section 461)

1977 Policy and Action Plan for Water Reclamation in California and Executive Order B-36-77

Encouraged water reclamation and funding to support legislative directives. Established Office of Water Recycling and goal “to make available an additional 400,000 acre-feet by 1982.”

1991 California Water Recycling Act of 1991

Established statewide goal to recycle 700,000 AFY by 2000 and 1,000,000 AFY by 2010. (California Water Code, Section 13577)

1994 Statement of Support for Water Reclamation

Joint statement signed by State Water Resources Control Board, the USEPA, California Conference of Environmental Health Directors, Department of Water Resources, U.S. Bureau of Reclamation, and Water Reuse Association of California to pursue and develop policies and regulations that reduce constraints and promote water reclamation.

2000 Water Recycling in Landscaping Act

Required local public or private entities that produce or will provide recycled water to notify the local agency and required local agencies in turn to adopt and enforce a recycled water ordinance requiring recycled water use. (Senate Bill 2095)

2001 Recycled Water Task Force (2) Assembly Bill 331 required the Department of Water Resources to convene a Recycled Water Task Force "to identify opportunities/constraints to increase the industrial and commercial use of recycled water."

2006 Assembly Bill 371 (2) Included statement that agencies should take appropriate steps to implement the Recycled Water Task Force recommendations to meet the goal of recycling one million acre-feet per year of water by 2010. Required installation of piping for landscape irrigation if recycled water will be provided.

2008 State Water Board Strategic Plan Update (55)

Stated goal to recycle 1,250,000 AFY by 2015.

2009 Recycled Water Policy (54) Established goal to recycle 1,525,000 AFY by 2020 and 2,525,000 AFY by 2030.

2010 Science Advisory Panel report on Monitoring Strategies for Chemicals of Emerging Concern (CECs) (13)

A Science Advisory Panel was established by the 2009 Recycled Water Policy “to provide guidance for developing monitoring programs that assess potential CEC threats from various water recycling practices.”

1 Includes Water Recycling Laws and Policies summarized in (79) Water Recycling Funding Program Division of Financial Assistance Strategic Plan; California State Water Resources Control Board: Sacramento, 2007. Excludes summary of funding sources (e.g., Bond Laws).

50

Table 2.13S. San Francisco Bay Area Recycled Water Coalition 2011 Project Summary

(64)

Project Yield

(AFY), Project

Yield (AFY), Future

Total Cost

Federal Share of

Cost

Cost per ac-ft1

South Bay Water Recycling Phase 1.d.

2000 3000 $39.2 M $9.8 M $980

South Santa Clara County Recycled Water Project

1790 2440 $28 M $4.2 M $780

Antioch Recycled Water Project 490 850 $12.5 M $0.875 M $1300 South Bay Advanced Recycled Water Treatment Facility

6720 28000 $53 M $5 M $390

Central Contra Costa Sanitary District (CCCSD)-Concord Recycled Water Project

255 255 $7.2 M $1.8 M $1400

Contra Costa County Refinery Recycled Water Project

5600 22500 $25 M $6.25 M $220

Central Redwood City Recycled Water Project

1075 3170 $32 M $8 M $1500

Central Dublin Recycled Water Distribution and Retrofit Project & other projects

215 215 $4.6 M $1.15 M $1100

Delta Diablo Sanitation District (DDSD) Recycled Water Advanced Treatment and Expansion Project

3900 12500 $25 M $6.25 M $320

Dublin San Ramon Services District (DSRSD) Recycled Water Expansion Project

350 3250 23.85 M $5.96 M $3400

Hayward Recycled Water Project 3760 3760 27 M $6.75 M $360 Ironhouse Sanitary District Recycled Water Project

910 1320 26 M $6.5 M $1400

Palo Alto Recycled Water Pipeline Project

1000 1500 33 M $8.25 M $1700

Petaluma Recycled Water Project, Phases 2A, 2B and 3

1610 3280 24 M $6 M $750

Pleasanton Recycled Water Project 440 1840 20 M $5 M $2300 Yountville Recycled Water Project 115 400 3 M $0.75 M $1300

1 Assumes 20-years of Project yield.

51

Chapter 3

Water reuse for ecosystem enhancement:

Matching opportunity with need

3.1 Introduction

Water and wastewater treatment systems were developed during the twentieth

century as two separate systems that served mutually excusive goals of water supply

and protection of the integrity of receiving waters. Upgrades of wastewater treatment

facilities to meet more expansive regulatory requirements improved ambient water

quality. Yet increase in urban water demand has come at the expense of aquatic

ecosystems. Approximately 91% of historical California wetlands, including 85% of

saline wetlands and 92% of freshwater tidal wetlands, have been lost due to

urbanization (80). Urban and peri-urban development, and a traditional emphasis of

engineers on the prevention of floods and disposal of wastewater, has adversely

impacted urban hydrology and damaged aquatic ecosystems. In California, more than

half (62%) of estuarine wetlands exhibit medium to poor health due to modification of

physical structure, including levees and transportation infrastructure that have changed

52

the shape and reduced the size of wetlands (80). In these regions, non-natural tidal and

freshwater hydrology couple with excessive sediment supplies to reduce physical

complexity and wetland health.

Effective management of urban water can benefit aquatic habitat. Given the

availability of tertiary treated recycled water within the San Francisco Bay Area (81)

and potentially throughout California, the question arises whether some portion of

available highly treated recycled water can be used for beneficial wetland enhancement

and creation or stream augmentation. Redesign of urban hydrology in a manner that

enhances existing aquatic habitat has the potential to provide new sources of water

storage, while restoring the integrity and improving the aesthetics of watersheds and the

urban environment.

In Chapter 2, the growth of water reuse in Northern and Southern California was

documented using statewide survey data and ground-truthing to evaluate major trends in

the size, location, and form of projects implemented over the past half-century. The

objectives of the present chapter are to characterize existing and potential cases of water

reuse for natural system enhancement in California and to outline perceived challenges

associated with the implementation of water reuse projects for ecosystem enhancement.

Projects in California in which environmental enhancement drove the project design,

distinct from discharge of highly treated wastewater with incidental environmental

benefit, are identified. Opportunities for new projects are evaluated based on responses

to the previously described survey of water reuse managers (2010 Survey), an

assessment conducted in the San Francisco Bay Area by a regional coalition of

municipalities, and a statewide projection of wetland condition. Lastly, general issues

and challenges associated with wetland creation and enhancement as well as stream

augmentation with recycled water are discussed.

3.2 Few Existing Examples of Water Reuse for Direct

Ecosystem Enhancement in California

Several databases were queried and assessed to compile data on the use of

wastewater for the benefit of ecosystems: the California State Water Resources Control

53

Board (SWRCB) recycled water surveys (3, 59) conducted in 2001 (2001 Survey) and

2009 (2009 Survey), the National Database of Water Reuse Facilities (National

Database, (58)), the Treatment Wetland Database (TWDB), and the 2010 Survey

conducted as part of this research. Brief descriptions of identified projects that are

located in Northern California are given in Table 3.1.

According to the 2009 Survey, water reuse for ecosystem enhancement totaled

27,849 AFY, representing 4% of total wastewater reuse in the state. A total of 17

programs listed either “Wildlife habitat or misc. enhancement” on the 2001 Survey or

“Natural System Restoration, Wetlands, Wildlife Habitat” on the 2009 Survey. Eight of

these projects are located in the northern 48 counties of California, and the remaining

projects are located in the ten southernmost counties. The average size of Northern

California projects listed on the 2009 Survey was 1,700 AFY. The National Database

notes a total of six projects, three of which are in Northern California, with “Natural

System Restoration – Wetlands” as a beneficial use category. Five of the National

Database listings are represented on the 2001 or 2009 Surveys. Other agencies with

wildlife enhancement beneficial uses noted on the 2001 and 2009 Surveys are listed in

the National Database without beneficial use categorizations.

Wetlands ecosystems that serve as polishing for secondary treated wastewater may

also be considered water reclamation systems. Due to their ability to accept large

quantities of effluent, their partly oxic and partly anoxic soils, and resilient aquatic plant

species, wetlands are particularly suitable for wastewater purification (82). The TWDB

contains system descriptions and performance data on pilot and full-scale constructed

wetlands (83). At the time of access (May 2011), the database lists 11 unique systems in

California, with several additional pilot systems: Arcata Treatment Marsh, Gustine,

Hayward, Hemet/San Jacinto, Kelly Farm, La Franchi, Las Gallinas Sanitary District,

Manila Community Treatment Plant, Mt. View Marsh, Richmond, and Sacramento

Demonstration Wetland. Except for the Hemet/San Jacinto, all of these systems are in

Northern California and several overlap with systems identified via the 2001 and 2009

Surveys. Wetlands as treatment systems are an attractive options for small communities

that may be disproportionately affected by the construction, operation, energy, and labor

costs associated with centralized “concrete and steel” water pollution control facilities

54

(84).

Additional data regarding wastewater reuse for ecosystem enhancement were

collected via an online questionnaire of Northern California water reuse managers

(2010 Survey), which was discussed in detail in Chapter 2. If “Wildlife habitat

enhancement” was selected as a direct beneficial use of the treated wastewater in the

initial background section of the survey, respondents were asked to list the type (e.g.,

wetland enhancement or restoration, stream augmentation, freshwater marsh, etc.),

location, and volume of recycled water for existing uses of recycled water from their

agency’s program for ecosystem enhancement purposes. Fourteen respondents indicated

“Wildlife habitat enhancement” as a direct beneficial reuse utilized by their agency, and

seven of these respondents provided further information on the systems. Descriptions

listed in Table 3.1 from the 2010 Survey were included only if the system was also

listed on either the 2001 Survey, the 2009 Survey, the National Database, or the TWDB.

As a result, three project descriptions, the Dow Wetlands Preserve in Antioch supported

with approximately 0.1 MGD by the Delta Diablo Sanitation District, a Moss Landing

research project using 0 to 8 MGY, and habitat enhancement projects coupled to

development near the Sutter Creek Wastewater Treatment Plant in Amador County,

were described by respondents but not included in Table 3.1.

55

Table 3.1. Projects in Northern California utilizing recycled water for ecosystem

enhancement or treatment wetlands for wastewater effluent polishing.

Name Agency Noted in 2001 or

2009 Survey

Noted in 2010 Survey

Noted in TWDB

Arcata Enhance-ment Wetlands

City of Arcata Y2 Y Y A free-surface constructed wetland operating year round with continuous loading of secondary treated chlorinated effluent, the Arcata Enhancement Wetlands became fully operational in 1986 at a cost of just over $500,000 (1986 base year of capital cost). With an approximately 15 ha footprint, three wetland cells (Allen, Gearheart, and Hauser Marshes) operate in series for tertiary treatment of solids, organics, and nutrients as well as for habitat creation/enhancement, recreation, research, and acting as nature preserve (83). Additional references: (9, 85).

Calera Creek Wetlands

City of Pacifica Y N N The Calera Creek wetland restoration was conducted to improve riverine waters and wetland ecosystem function and to create habitat for the threatened California Red-Legged Frog and endangered San Francisco Garter Snake. Pacifica used a Hydro Geomorphic model for planning their wetland restoration projects. The treatment facility utilizes ultraviolet disinfection (83). In 1999, Pacifica delivered an average of 2,020 AFY (1.8 MGD) to existing wetlands (81). The City of Pacifica reported 3,280 AFY of recycled water use for Natural System Restoration, Wetlands, Wildlife Habitat in the SWRCB 2009 Survey (59).

Emily Renzel Wetlands

City of Palo Alto N2 Y N The Emily Renzel Wetlands restoration project in the Palo Alto Baylands comprises a 15-acre freshwater pond that receives 1 to 2 million gallons per day of pumped reclaimed water from the nearby Palo Alto Regional Water Quality Control Plant. When completed in 1992, it was one of only three projects in the State using reclaimed wastewater to develop freshwater marshes for birds. In 1999, Palo Alto delivered an average of 280 AFY (0.25 MGD) to existing wetlands (81). The City of Palo Alto does not report any use of recycled water for ecosystem enhancement on the 2002 or 2009 SWRCB Surveys, but the wetland enhancement project is listed on the National Database of Water Reuse Facilities.

Kelly Farm Santa Rosa N Y Y A small free surface constructed wetland (4 ha; 5 cells) that receives advanced secondary treated effluent, Kelly Farm became fully operational in 1990 (83). The marsh uses about 20 million gallons of water per year from the City of Santa Rosa Laguna Treatment Plant. This treatment facility also supports riparian revegetation projects that may use 20,000 gallons per day in the dry season.1

Hayward Marsh

Union Sanitary District (USD) Y2 Y Y The original two-phase implementation completed in 1980 and 1988 restored nearly 400 acres of the 1800 acres of Hayward shoreline (12). The 5-cell free surface wetland treatment system (~60 ha) receives conventional secondary effluent for year-round operation (83). In 1999, USD delivered an average of 11,000 AFY (10 MGD) to existing wetlands (81). In 2009, USD reported its total recycled water use, 3,493 AFY, as that for natural system restoration, wetlands, or wildlife habitat (59).

Gustine constructed wetlands

City of Gustine N N Y Fully operational in 1988 at a total cost of $882,000, the 9.6 ha free-surface constructed wetland utilizes 24 marsh cells to manipulate hydraulic detention time after receiving effluent from up to 11 oxidation ponds operated in series (83, 84). The City of Gustine only reports water reuse for irrigation in the 2002 or 2009 SWRCB Survey.

1 Additional information from 2010 Survey response. 2 “National System Restoration – Wetlands” beneficial use also listed on the National Database of Water Reuse Facilities.

56

Name Agency Noted in 2001 or

2009 Survey

Noted in 2010 Survey

Noted in TWDB

Las Gallinas, San Rafael

Las Gallinas Sanitary District (LGSD) N N Y The 20-acre free surface constructed freshwater marsh/pond was designed with varying depths and vegetation in a single unit to incorporate different wildlife habitat types (86). An additional 40 acres of storage ponds are used to irrigate pasture. Including land acquisition, the total cost of the reclamation system was $8.6 million, with state and federal Clean Water Grant funds covering 87.5% of the costs. LGSD reports 378 AFY for agricultural irrigation on the 2009 SWRCB Survey but does not indicate reuse for an environmental purpose (59).

La Franchi Santa Rosa N N Y La Franchi became fully operational as a free-surface constructed wetland (0.1 ha) treating low rate pond secondary-treated agricultural and animal waste in 1991 (83, 87). Because this recycling does not occur from a municipal wastewater treatment facility, the listing is not reported on the SWRCB Surveys.

Manila wetlands

Manila Comm. Treatment N N Y The free surface constructed wetland in Manila, CA operates year round with continuous loading after passing through a low-rate pond for secondary treatment in Manila, CA (83).

Moorhen Marsh and McNabney Marshes

Mt. View Sanitary District (MVSD) Y Y Y Moorhen Marsh is a 21-acre constructed wetland that is 100% fresh water and effluent dominated. MVSD cites the marsh as the first to use conventional secondary treated effluent as its primary water source. The adjacent McNabney Marsh (formerly known as Shell Marsh) is estuarine and seasonally saline, in total consisting of 130-acre restored, seasonally tidal wetland (88). The free-surface constructed wetland, reported as 3 cells and 37 ha in the TWDB, became operational in 1974 at a system cost of $90,000 and annual operation and maintenance cost of $20,000 (1978 base year of costs) (83). In 1999, MVSD delivered an average of 1,000 AFY (0.9 MGD) to existing wetlands (81).

Sacramento Demon-stration Wetlands

Sacramento Regional WWTP N N Y Treatment of municipal disinfected secondary effluent via a full scale free surface constructed wetland (8.9 ha) that utilizes 10 cells began in 1994, operating at about 1 MGD (83, 89), The Sacramento Regional County Sanitation District did not report reuse on the SWRCB 2009 Survey (59).

Wetland Enhance-ment / Restoration

Sonoma County Water Agency Y Y N The Sonoma County Water Agency includes four wastewater treatment facilities. Of these, two report use of recycled water for natural system enhancement or restoration on the SWRCB 2009 Survey: the Sonoma Valley Treatment Plant reported 100 AFY for wildlife of 1,600 AFY total reuse in 2009, and the Russian River Treatment Plant reported 90 AFY for wildlife of 150 AFY total reuse in 2009 (59).

Many 2010 Survey respondents considered the major beneficiaries from the

implementation of recycled water programs to include environmental groups in addition

to natural habitats. Examples cited of environmental and public benefits were: less

reliance on Delta Water, stakeholders concerned with protection of water quality in

Clear Lake, restoring water levels at Lake Merced in San Francisco (e.g., Cal Trout and

Natural Heritage Institute), and reduced discharge flows to Monterey Bay. Beneficiaries

cited also included general environmental advocacy groups (e.g., in San Jose/Santa

Clara) as well as natural habitat and recreational users (e.g., birders).

57

3.2 Identifying Opportunities for Ecosystem Enhancement

Although opportunities to reuse reclaimed water may be gleaned by quantification

of water needs for various applications (1), the needs of ecosystems are less practically

quantified. However, the system hydrology and measurable ecosystem characteristics

are intricately linked. Urbanized estuaries tend to have lower wetland health due to

hydrologic and biotic community structures (80). Water source, velocity, flow rate,

renewal rate, and inundation frequency influence the chemical and physical properties,

and thus biological structure, of wetland substrate (84). A major recommendation of the

Surface Water Ambient Monitoring Program (SWAMP) includes the need to increase

the size of estuarine wetlands to reduce the effects of stressors such as terrestrial

predators (80). Water movement through wetlands tends to have positive impacts on the

ecosystem, promoting increased regional production (84). For streams that have

experienced significant flow reductions due to anthropogenic influences, augmentation

using highly treated recycled water may beneficially impact the stream via increased

summer flows, improved water quality, support of healthier riparian areas, lowered

stream temperatures, enhanced fish and wildlife habitat, and improved aesthetics (90).

Based initially on this premise, artificial augmentation of wetlands, including riparian

corridors that have experienced significant flow reductions and altered hydrologic

regimes, represents potential opportunities for habitat restoration.

California assessment to match opportunity with need. Rapid assessment

methods represent a potential cost effective and consistent mechanism to monitor

relative wetland and riparian health, evaluating complex ecological condition using

observable field indicators. The California Rapid Assessment Method (CRAM) was

developed to assess the health of California wetlands along a continuum of conditions

based on attributes and metrics identified from a literature review and selected for

appropriate accuracy, precision, robustness, ease of use, and cost (91). The analysis

assumes that ecosystem condition, and the ability to support wildlife, may be measured

by structural characteristics and increases with complexity and size. Seven wetland

classifications were selected for the CRAM: riverine and riparian, estuarine, lacustrine,

depressional, wet meadows, vernal pools, and playas. The goal of the CRAM

58

assessment is to “provide rapid, scientifically defensible, standardized, cost-effective

assessments of the status and trends in the condition of wetlands and related policies,

programs and projects throughout California.” Subsequent validation of the CRAM

methodology indicates that the score corresponds with multiple independent

assessments of condition for avian diversity, plant community composition, and benthic

macroinvertebrate indices (92).

Four attributes, landscape context, hydrology, physical structure, and biotic

structure, are each characterized semi-quantitatively based on narrative analyses on a

series of metrics that are further assigned an ordinal or interval score relative to a

pristine condition (91). The hydrology attribute incorporates three metrics: water

source, hydroperiod, and hydrologic connectivity (92). External stressors (e.g.,

anthropogenic influences or natural disturbances to the wetland) are documented

separately from the wetland condition. The CRAM score, expressed as percent possible

ranging from 25 to 100 (80), summarizes the condition, or health, of a wetland or

riparian habitat relative to its maximum achievable condition based on a field visit by

two trained individuals. CRAM scores falling between 25 and 44 are indicative of poor

estuarine wetland health while 44 to 63 indicates medium to poor health. Other widely

used wetland assessment methods include the Hydrogeomorphic Method (HGM), the

Index of Biotic Integrity (IBI), and the Habitat Evaluation Procedure (HEP). These

methods are generally much more time and cost intensive and thus are generally

unavailable at a statewide level (91). For the present assessment, CRAM data were

obtained in July and August 2010. The CRAM scores categorize water bodies as

Estuarine Saline, Estuarine Non-saline, Riverine Confined, or Riverine Non-confined.

59

Figure 3.1. Distribution of California Rapid Assessment Method (CRAM) overall

wetland scores, included for estuarine (saline and non-saline) and riverine (confined and

non-confined), and wastewater facilities with tertiary treatment capacity.

Monitoring wetlands on a broad scale provides general information about

opportunities for wetland enhancement using recycled water. For this purpose, the

locations of tertiary treatment facilities were identified for proximity to wetland

ecosystems under stress. In Figure 3.1 the location of California tertiary treatment

facilities is overlaid with CRAM scores (93). As a conservative approach, only

wastewater facilities that currently utilize tertiary treatment, as indicated by the National

Database, are included. Typically tertiary treated recycled water for general purpose

irrigation comprises additional steps of coagulation, filtration and disinfection beyond

60

secondary treated wastewater. According to the National Database, advanced/tertiary

treatment technologies used by water reuse facilities in California include carbon

adsorption, ion exchange, disk filters, media filtration, ultrafiltration, nanofiltration,

reverse osmosis, and chemical precipitation. The treatment facility locations shown in

Figure 3.1 were generated from the zipcode of the facilities. Further refinement with

GIS mapping and system features are underway. Analysis of the location of tertiary

treatment facilities as generated from the zipcode was compared to the location of the

lower 50% of CRAM scores as a proxy for distance to wetlands. There were 27 low-

quality wetland sites within 10 miles of a wastewater treatment plant. Additional

evaluations at an ecoregion level can inform prioritization of restoration projects.

San Francisco Bay regional assessment. Of California’s 44,456 acres of perennial

tidal estuarine wetlands, 77% are located in the San Francisco Bay Estuary (80). The

1999 Bay Area Water Recycling Master Plan (BAWRMP) contains the most

comprehensive assessment of opportunities for wetland and stream augmentation using

tertiary treated recycled water for the San Francisco Bay region (81). The goals of the

Bay Area Regional Water Recycling Program environmental enhancement committee

were to identify potential environmental enhancement projects, use a watershed context

to evaluate environmental issues, and identify and develop action plans to address

regional environmental issues in recycled water implementation. The March 1999

Baylands Ecosystems Habitat Goals Report was utilized to frame the analysis of

potential for ecosystem enhancement using recycled water. The team evaluated 16

potential wetland restoration locations and 13 possible stream augmentation sites for

ecosystem enhancement, totaling of 13,000 AFY and 19,000 AFY, respectively (81).

The evaluation was not comprehensive, and further identification of potential sites

should be conducted, especially as new wetland assessment techniques are developed

and implemented. The application of recycled water for environmental enhancement

requires further investigation to determine the value of this option as well as appropriate

water quality criteria.

The BAWRMP evaluation of wetland sites involved potential site identification

from an initial market assessment, evaluating wetland water demand based on acreage

and a wetland water application rate (10 AF/acre based on wetland biological

61

requirements), and site evaluation of potential benefits and impacts, as well as

implementation strategies, for 16 potential sites (81). Possible site benefits included

habitat, species, and habitat diversity and management benefits, as well as the potential

to intercept non-point source pollution runoff and improve aesthetics. Potential adverse

impacts considered included the conversion of an existing valuable habitat, impact to

existing habitat or special status species, inconsistency with habitat management plans,

possible bioaccumulation of pollutants based on potential design, and impact on

biological resources from pipeline infrastructure.

The BAWRMP goals for 2010 were to develop an additional 11,000 AFY for

streamflow augmentation and 9,500 AFY for wetland enhancement or creation (81).

This was expected at a total cost of approximate $15 million for wetlands and $0.9

million for streams (discounted using 6.875% nominal discount rate; reported in 1997

dollars). Sites included in these goals included stream augmentation projects at San

Francisquito Creek by the City of Palo Alto facility, San Mateo Creek by South Bay

Systems Authority, Pillarcitos Creek by the San Francisco International Airport facility,

and the Guadalupe River by San Jose/Santa Clara (SJSC) among several other sites.

Based on the database analysis, many of the projects identified by the BAWRMP as

potential enhancement sites have likely not been implemented. The relatively large

number of projects expected for environmental uses by 2010 may have been a result of

low costs of additional treatment expected in the modeling scenarios. At the time of the

BAWRMP, the San Jose Coyote Creek study was underway, with note that additional

treatment beyond tertiary filtration may be a conclusion of that study. As will be

described in more detail later, this project was canceled in 2008 following extensive

water quality analysis that showed the presence of perfluorinated chemicals.

Additional opportunities in Northern California. In addition to identifying

existing sites, respondents from the 2010 Survey of Northern California water reuse

facilities were questioned regarding “opportunities to expand the use of recycled water

for restoration or protection of natural environments.” Respondents were asked to list

future opportunities identified to use or expand use of recycled water from their

recycled water program for ecosystem enhancement purposes. Several agencies noted

future opportunities for wetland or salt pond restoration, creation, or expansion:

62

• South Bay wetland creation (1-5 MGD) and bittern ponds habitat restoration

(5-15 MGD), City of San Jose;

• Napa Salt Marsh wetland reclamation (up to approximately 5 MGD), Napa

Sanitation District;

• Duer and Irwin Creek (estimated to use about 20,000 gallons per day each),

City of Santa Rosa;

• Wildlife habitat, Sacramento Regional County Sanitation District;

• Salt ponds restoration/wetland enhancement, Sonoma County Water

Agency;

• Carmel River Lagoon, Carmel Area Wastewater District;

• Possible wetland enhancement, Laurel Pond, Mammoth Community Water

District;

• Expansion of Hayward Marsh (north and south of existing marsh), East Bay

Dischargers Authority;

• Closed-loop constructed wetlands, Northwest Regional Treatment Plant,

Lake County Sanitation District.

Several respondents noted particular ecosystem enhancements expected due to

increased treatment or reduced discharge:

• Churn Creek, City of Shasta Lake (looking to increase treatment levels to

allow more discharge; could increase fish spawning);

• Pajaro Valley sloughs and Monterey Bay National Marine Sanctuary (~7

MGD reduced discharge), Pajaro Valley Water Management agency;

• Reduce nutrients to Monterey Bay, Marina Coast Water District.

Providing additional commentary on this form of ecosystem benefit, one respondent

noted, “I believe that since the run-off from recycled water is in the creeks, it helps

provide a habitat for certain species. So it does enhance the ecosystem. We have never

seen any adverse affects from recycle water run-off and/or use, only good things like

water and lower rates to use it.”

In response to whether future opportunities to use or expand use of recycled water

for ecosystem enhancement purposes, 32 respondents indicated that no future

opportunities have been identified. In space for comment, several of these respondents

63

provided additional input:

• “Restrictive Regional Water Quality Control Board requirements for

discharging recycled water to local creeks due to downstream potable water

recharge diversions would seem to seriously limit or eliminate ecosystem

enhancement opportunities.”

• “In fact we would like to divert a percentage [of] current flow from wildlife

enhancement to industrial use.”

• “Wildlife/habitat enhancement is not a consistent user of water. We have

consistent users who could use the water instead, if we can make it

available.”

• “Unlikely District will be able to expand system due to new state general

order. May not be able to meet new regulatory standards.”

While regulatory requirements may be a consistent driver of program

implementation, such requirements may also limit the use of recycled water for

ecosystem enhancement. This assessment represents a limited patchwork of projects,

and further systematic identification of sites is necessary as part of future work.

3.3 Technological and Management Challenges

The BAWRMP provides a key resource for framing the issues associated with

wetland creation and enhancement using recycled water as the main source and for

conducting a preliminary assessment of potential benefits and adverse impacts of such

projects (81). Key issues for wetland creation and enhancement with recycled water

include flow velocity and erosion, habitat conversion, ecosystem buffers, control of

non-native species, vector control, water quality, flood control, public access, land

ownership/land use impacts, and water supply dedication/environmental management

sustainability. For stream augmentation, general issues include elevated water

temperature, decreased water quality (e.g., for temperature, nutrient load, total dissolved

solids, and trace organics), increased warm-water predator populations, inadvertent salt

marsh conversions, needs in addition to flow augmentation (especially when this is not

the priority obstacle to restoration), and flow timing. San Francisco Bay Area streams

are typically seasonal streams, and the creation of perennial streams could have adverse

64

effects on indigenous fauna. Some streams (e.g., San Mateo Creek) may have been

perennial historically and therefore warrants inclusion as possible augmentation sites.

Water quality. Undisinfected secondary treated effluent is the miminum water

quality requirement for wetlands (81). Wetland systems can further filter particular

matter, reduce suspended solids (SS), remove biochemical oxygen demand (BOD),

remove and store nutrients, and attenuate some trace chemical contaminants (11, 94).

Due to these features and other environmental benefits, as well as escalating costs of

construction, operation, and distribution associated with conventional treatment, interest

in wetlands as wastewater treatment systems has increased (12, 82, 84). However, the

presence of waterborne pathogens, nutrients and other aesthetic issues (e.g., odors)

could preclude the use of effluent in the urban environment such that more advanced

treatment technologies may be necessary for protection of ecosystem health. Further,

despite potential benefits associated with natural attenuation of pharmaceuticals,

personal care products, and other commercial and industrial chemicals of emerging

concern in effluent-dominated wetlands and rivers (6, 11, 95, 96), recalcitrant chemicals

may persist in treatment wetlands and receiving waters. The bioaccumulation of trace

organic contaminants in biota residing in treatment wetlands (97) and effluent-

dominated natural systems (98) is a concern. Research is needed on the most effective

means of treating wastewater prior to reuse for habitat enhancement.

2010 Survey analysis of drivers and challenges for water reuse for ecosystems.

“In most cases, using recycled water to offset potable use is more

desirable than using limited recycled supplies for habitat enhancement.

Therefore the most important barrier is that environmental enhancement

is not the highest and best use.”

When respondents from the 2010 Survey were asked to consider potential drivers

and challenges for future water reuse projects for ecosystem enhancement, several

important management issues were raised and overall concern for funding and

skepticism for practicality arose. Respondents rated the importance of five broad

categories of potential drivers and hindrances for future use of recycled water for

wildlife or habitat enhancement on a three-point scale. The fraction of respondents

indicating each broad category as a very important driver or a driver (n = 36),

65

respectively, was: regulatory requirements (0.61, 0.28), public value of natural

environments (0.44, 0.41), recycled water policy (0.37, 0.40), water shortages (0.33,

0.50), and influential stakeholder groups (0.34, 0.43). Other drivers listed by

respondents were funding and seasonal viability. The response frequencies for rated

hindrances, including cost recovery, water quality concerns, lack of influential

proponents, lack of a policy mandate, and quantification of ecosystem benefits, are

displayed in Figure 3.2. Other challenges inserted by respondents included lack of

available recycled water (or incoming wastewater) and operational reliability.

When asked to describe the most important driver for the implementation of

recycled water programs for wildlife or habitat enhancement (n = 38), many

respondents noted the role of regulatory requirements in reducing discharge, requiring

increased treatment prior to discharge, or increasing the cost of discharge. However,

regulatory allowance for wildlife or habitat enhancement was also viewed as a

limitation. Others expressed concern over the practicality of such projects, lack of

funding, or higher priority uses such as industrial cooling towers and turf irrigation.

Stakeholder support was considered important, though one responded noted, “The issue

is do stakeholders really want environmental improvement or is it just ‘code’ for

stopping growth.”

Augmentation of natural systems with recycled water that is coupled to revenue

generating reuse options may provide sources of treated wastewater for environmental

uses. For example, flow regimes that mimic naturally occurring ephemeral streams may

receive water in winter months while tertiary treated effluent is otherwise used for

irrigation in summer months. For programs using tertiary treatment for irrigation during

the summer months, one respondent noted, “There might be some possibility of using

the water during winter months to wetland storage.” However, concern over reliability

during wet season operations was expressed. Another respondent described, “Our

existing recycled water storage reservoir (175 acre-feet) is home for over 100 species of

birds. Future expansion might use wetlands treatment. The problem with this is that the

wetlands treatment system would work well in the dry season but would likely fail in

the wet season.”

66

Figure 3.2. Response frequencies for 2010 Survey respondents who rated five broad

categories of hindrances to implementation of future water reuse programs for

ecosystem enhancement.

Many respondents listed cost recovery and inability to pay for habitat enhancement

projects as the most important hindrances to the implementation of recycled water

programs for wildlife or habitat enhancement. Reiterating previous specific cost

concerns, one respondent stated that the greatest challenge is “as always, the cost of

building pipelines to take water from the production site to the use site.” Next most

frequently cited hindrances were regulatory restrictions that may be difficult to meet, as

well as the “level of water quality effluent the wetlands oversight will allow.” Concern

was expressed that runoff into receiving waters may be perceived as a discharge. Others

cited a lack of available recycled water (due to full dedication to other uses) or available

land.

Given the cautionary responses and cost recovery concerns for the use of recycled

water for ecosystem enhancement, the technological advancements that are made with

respect to water reuse for ecosystems must be coupled with relevant management

solutions for practical challenges associated with regulatory hurdles and funding.

Management scenarios may require economic justification and incentives for positively

utilizing tertiary treated water in such new ways. One particular area of research need is

67

with regards to ecosystem services of water reuse wetland systems. Ecosystem services

and values of managed wetlands include water storage, flood protection, water quality

improvement, aquifer maintenance, carbon sequestration, recreation, education and

provision of traditional wildlife habitat (e.g., migratory birds) (99-102). A limited

number of life cycle assessments have been conducted to compare the environmental

impacts of wetland treatment systems to conventional wastewater treatment (103), but

quantifying the magnitude, timing, and spatial variation of ecosystem services provided

by managed wetlands remains a challenge (104).

Case study. The Santa Clara Valley Water District (SCVWD) recently explored the

use of recycled water for stream augmentation, a project that was eventually cancelled

in 2008. The San Jose/Santa Clara Water Pollution Control Plant was required to reduce

freshwater discharges to Artesian Slough to protect the endangered California clapper

rail and salt marsh harvest mouse (81). At the same time, regulatory agencies

recognized that removal of all effluent flow could damage existing wetlands. The

discharge restriction led to the creation of the South Bay Water Recycling Program

(81). Recognizing the potential for tertiary treated municipal wastewater to provide

additional water to a river ecosystem, the SCVWD planned to use tertiary-treated

wastewater from the City of San Jose, CA to augment Coyote Creek as a research

demonstration (105). The project was expected to serve as an analysis for stream

enhancement via recycled water augmentation throughout the County (106); however,

the project was immersed in debates over the use of recycled water in degraded urban

streams and lacked clear metrics for evaluating success. Furthermore, this project was

complicated by uncertainty regarding the risk of unregulated emerging contaminants in

reclaimed wastewater. Plumlee et al. (5) showed that in the case of the San Jose/Santa

Clara Water Pollution Control Plant, measurable levels of perfluorooctane sulfonate

(PFOS) in recycled water effluent would increase surface water concentrations of the

receiving waters above concentrations projected to be protective of avian life. Although

emerging contaminants such as PFAAs may be detected in wastewater and recycled

water effluent, these chemicals are not explicitly regulated by Title 22 requirements for

recycled water. Following the reporting of these results to the SCVWD, the project was

cancelled in February 2008.

68

3.4 Significance

An evaluation of the 2001 and 2009 California Municipal Wastewater Recycling

Surveys conducted by the California State Water Resources Control Board, the National

Database of Water Reuse Facilities, the Treatment Wetland Database, and a 2010

survey of recycled water managers reveals that few projects have been implemented in

Northern California for direct natural system enhancement or habitat creation. Despite

identification of potential restoration sites in the San Francisco Bay Area, relatively few

have been developed. Further work towards identifying specific locations and assessing

potential environmental benefits of such projects is needed.

As a relatively new and innovative use of recycled water, wastewater reclamation

for ecosystem enhancement demands further assessment. At the site level, experiments

to measure ecosystem response to restoration projects and to quantify biotic response to

altered or modified flows are needed. Uncertainties concerning water quality and

ecosystem impact are barriers to the use of water reuse for ecosystems. The interplay of

trace organic contaminants and ecosystems presents an important area of research. One

particular challenge associated with widespread use of synthetic chemicals and

discharge of wastewater to natural systems is the potential for bioaccumulation of

persistent organic pollutants downstream of discharges. The SCVWD case provides one

example in which uncertainty about emerging contaminants, and specifically the

bioaccumulative potential of perfluorinated compounds, played an influential role in

project implementation. Issues of trace contaminants in recycled water were recently

addressed by a California Science Advisory Panel (13). However, the objectives of the

panel were to evaluate human and environmental exposures resulting from landscape

irrigation, indirect potable reuse via surface spreading, and indirect potable reuse via

groundwater injection. The analysis excluded other forms of reuse, including

environmental applications of recycled water.

Acknowledgment. We thank Aude Martin and Sophie Egan for contributions to this

project.

69

Chapter 4

Exposure of perfluorinated chemicals to

San Francisco Bay white sturgeon and

mechanisms of bioaccumulation

4.1 Introduction

The ability to accurately predict the bioaccumulation of chemicals in aquatic

organisms is an essential component to assessing the human health and ecological risk

of chemical pollutants. Bioaccumulation is the increase in the concentration of a

substance in an organism from the intake of contaminated water, food, and air and

results from a greater rate of contaminant uptake compared to that of elimination via

metabolism or excretion (107). There is strong evidence from field biomonitoring

studies and food web analyses in marine and freshwater ecosystems that C8-C12

perfluorocarboxylates (PFCAs), as well as C6 and C8 perfluoroalkyl sulfonates (PFSAs),

bioaccumulate and biomagnify in aquatic organisms (25). Although the global

70

distribution of perfluoroalkyl acids (PFAAs) in biota is well documented (24, 25), less

is known about the biological and chemical processes governing the introduction and

propagation of these compounds in food webs. Traditional evaluations of contaminant

impacts on biological systems rarely consider basic physiological and ecological

processes that drive differences in exposures among species and that may explain

within-species contaminant concentration distributions (108). For PFAAs in particular,

further challenges arise in modeling bioaccumulation because the behavior of PFAAs in

the environment is not obviously deduced from their physiochemical properties.

PFAAs are ubiquitous in coastal and marine systems (27, 109-111), where

characteristic salinity gradients may influence PFAA chemistry as well as organism

physiology (112). The relatively high water solubilities, low Henry's constants, and

amphiphilic nature of PFAAs imply the importance of water partitioning in the

environmental fate of these compounds (16, 113). In rainbow trout, direct uptake of

aqueous PFAAs in freshwater conditions exceeds that of dietary accumulation (114,

115). However, relatively low environmental concentrations of PFAAs suggest the

importance of considering bioaccumulation in aggregate, as accumulation from both

water and food characteristically contribute to chronic exposures. The importance of

PFAA uptake from sediments has been demonstrated in field biomonitoring work and

an assessment of uptake in a freshwater benthic organism (110, 116). In sediments

collected from throughout the San Francisco Bay region, PFAAs were present at low

ng/g concentrations (27).

Study objectives. This chapter commences with a description of traditional

bioaccumulation models and shortcomings of such approaches for capturing the

environmental behavior of PFAAs, providing justification for the elucidation of

mechanistic PFAA bioaccumulation parameters in subsequent chapters. Secondly, an

inter-species comparison of PFAA concentrations in white sturgeon from San Francisco

Bay is presented. Quantitative ecological parameters including trophic position and

foraging location are utilized to evaluate influences of ecological processes on PFAA

liver tissue concentrations. Stable isotope ratios for nitrogen (δ15N) and carbon (δ13C)

act as naturally occurring intrinsic tracers by providing integrated measures of trophic

relationships and feeding locations along a salinity gradient (108).

71

4.2 Predictive models for PFAA bioaccumulation

Predictive approaches for quantifying the extent of bioaccumulation in organisms in

aquatic food webs are important as organisms may be exploited as monitors of

environmental contamination, and consumption of contaminated organisms can result in

high dosage exposures of toxic chemicals. The quantity of a chemical in a target organ

or tissue is controlled by contaminant uptake and retention from environmental

exposures and determines the organism’s toxic response (107). Existing estimation

methods for the bioaccumulation of organic substances in aquatic organisms are

categorized in a tiered predictive approach:

Tier 1: A simplistic correlation whereby the bioconcentration factor is a

function of the compound’s octanol-water partition coefficient;

Tier 2: A mass balance model for bioaccumulation in which uptake and loss

processes are empirically quantified in an organism at steady state;

Tier 3: A detailed prediction of biomagnification in a food chain involving fish

and air-breathing animals.

The simplest Tier 1 approach, used for initial screening purposes, relies on empirical

correlations of bioconcentration with octanol-water partition coefficients (KOW). In this

case, the bioconcentration factor (BCF) is defined as the ratio of the total chemical

concentration in an organism to the total chemical concentration in water. This method

is particularly useful because BCFs are expensive to measure and estimations from

octanol-water partition coefficients are rapidly applied. However, for ionic species,

traditional linear or bilinear relations between log BCF and log KOW do not hold (117).

Meylan et al. (1999) developed a more detailed prediction of bioconcentration factors

from a database of 84 ionic compounds, including 41 carboxylic acids and 37 sulfonic

acids and salts (117):

log KOW < 5, log BCF = 0.50

log KOW 5 to 6, log BCF = 0.75

log KOW 6 to 7, log BCF = 1.75

log KOW 7 to 9, log BCF = 1.00

log KOW > 9, log BCF = 0.50

72

The above relationships exclude compounds with long alkyl chains (≥11 carbons)

because five chemical species with long alkyl chains in the database exhibited log KOW

values ranging from 1.2 to 1.78 with higher corresponding log BCF values, averaging

1.85.

Log BCF values for PFOA, PFDA, perfluoroundecanoate (PFUnA),

perfluorododecanoate (PFDoA), perfluorotetradecanoate (PFTA) (C8 to C14 PFCAs)

and perfluorohexanoate (PFHxS) and PFOS (C6 and C8 PFSAs) ranged from 0.6 to 4.36

for rainbow trout carcass (114). Experimentally determined values for the n-octanol

distribution coefficient (log D), defined as the partitioning between water and n-octanol

for the total concentration of protonated and deprotonated forms of the PFAA, are 1.92

for PFOA, 2.57 ± 0.07 for PFNA, 2.90 ± 0.10 for PFDA, and 2.45 ± 0.08 for PFOS (17,

118). Calculations for log KOW, estimating partitioning of the protonated species, vary

widely for each PFAA, ranging from < 1 to 6.3 for PFOS alone. SPARC 2009 software

estimates log KOW values of 2.91 to 6.38 for C4 to C10 PFCAs and 4.67 for PFOS (17).

Based on the correlations above, log KOW likely underestimates bioconcentration factors

for perfluoroalkyl acids. Further, for equivalent perfluoroalkyl chain length, PFSAs

generally have higher bioconcentration factors than PFCAs (37, 114, 119), a feature lost

in the generalized relationship of log BCF with log KOW.

The mechanism of PFAA accumulation is likely different than lipid partitioning,

which is common to neutral hydrophobic organic contaminants (HOCs). Rather, PFAAs

may be proteinophilic (113), exhibiting bioaccumulative properties similar to fatty acids

(109). However, proteins have not been incorporated into PFAA bioaccumulation

models, and little is known about the sorptive capacity of proteins as a biological

compartment (44). Whereas KOW can be used to predict the accumulation and

environmental fate of HOCs, the amphiphilic nature of PFAA anionic surfactants

renders such descriptors unsuitable for evaluating the biological fate of PFAAs (25). In

fact, the limited utility of a lipid-normalization paradigm was noted due to artificially

high biota-sediment accumulation factors (BSAF) determined for PFAAs taken up by

the aquatic oligochaete, Lumbriculus variegatus (116). In general, PFAA

bioaccumulative potentials may be underestimated by calculations involving log KOW

(18).

73

Perfluoroalkyl carboxylates exhibit similar lipophilicity relative to equivalent chain

length sulfonates (17, 120). Differences in the electron withdrawing nature of the

sulfonate and carboxylate groups may alter the hydrophobicity of the perfluoroalkyl

chain, such that electrostatic and hydrophobic components of free energy may yield

overall equivalent values for PFCA and PFSA compounds with the same perfluoroalkyl

chain length (17). However, an increased electrostatic contribution to the Gibbs free

energy may yield more favorable values for proteinophilic partitioning for PFSAs

relative to equivalent perfluoroalkyl chain length PFCAs (17).

Equilibrium partitioning approach. Protein may represent a significant fraction of

tissue content; typical models do not account for the body burden of chemicals

attributed to this fraction. Whereas L. variegatus has 12.2 ± 1.6% lipid content by dry

weight, the organisms have a higher protein content by dry weight at 47.4 ± 8.3% (121).

Considering lipid, protein, and water fractions of organism tissues, the concentration of

a compound in an organism based on fugacity capacity theory may be assumed to be a

linear function of the contributing tissue constituents (44):

Corg = flip Clip + fprot Cprot + fw Cw (4.1)

where flip, fprot, and fw are the fractions of lipid, protein, and water relative to the whole

body tissue, and Clip, Cprot, and Cw are the respective concentrations of PFAAs in lipid,

protein, and water in the organism. In the case of PFAAs, protein-rich tissue may be the

dominant compartment for PFAA partitioning. The total sorptive capacity for the

organism tissue for PFAAs may be largely influenced by the sorptive capacity of

protein tissue. Thus, a protein-water partition coefficient (KPW), analogous to partition

coefficients between two bulk solvents (i.e., KOW), should be incorporated into Equation

4.1 to more accurately estimate a bioconcentration factor using aqueous concentrations

(Caq):

Corg = flip KOW Caq + fprot KPW Caq + fw Cw (4.2)

Bovine serum albumin (BSA) was previously utilized as a model protein for

establishing KPW for a series of HOCs in the above relationships (44).

74

Figure 4.1. Study area. Fish samples were collected from north San Francisco Bay,

including San Pablo Bay and Suisun Bay. Figure from Stewart et al. (108).

4.3 Materials and Methods

Sample collection. In the fall and early winter of 1999-2000, researchers from the

US Geological Survey (USGS) collected fish samples from Suisun Bay and San Pablo

Bay located in the San Francisco Bay estuary (Figure 4.1). This region, seaward of the

Sacramento-San Joaquin River system, is a part of the migration corridor for the large

fish species assessed in this study. Field sampling and fish tissue sample preparation

and storage by corroborating laboratories are further described by Stewart et al. (108).

Analyses in this chapter focus on white sturgeon liver samples (n = 15). Tissue

concentrations are also measured in striped bass (n = 1) and leopard shark (n = 1) liver

samples for comparison of physiological parameters. Sample collection occurring

during a time when anadromous fish species, such as striped bass (December 1999

collection) and white sturgeon (January 2000 collection), were likely to have spent

several weeks or months feeding in the region. Additional data provided by Robin

Stewart (USGS) for these samples included tissue type (muscle or liver), general sample

collection location, fish length, muscle tissue carbon isotope tissue (δ13C), muscle and

liver tissue nitrogen isotope data (δ15N), and selenium and mercury muscle tissue

75

concentrations (108). White sturgeon liver samples were 52 ± 11% protein and 24 ±

12% fat by dry weight (Anresco, Inc. analysis, Methods AOAC 992.15 and 960.39, 18th

ed.). Liver and muscle tissue samples were received by Christopher Higgins of Stanford

University from the USGS and frozen in sample containers at -4°C until analysis.

Extraction. Samples were analyzed for a series of perfluorinated chemicals using a

method modified from Stevenson et al. (122). Samples were transferred to -80˚C at least

one night before processing. Tissues were freeze dried at -80°C and homogenized with

a mortar and pestle. Ground, freeze-dried liver tissue (~150 mg) was extracted with

acetonitrile (3 × 5 mL) in a 50-mL polypropylene tube. For each addition of acetonitrile,

the tube was vortexed, sonicated in a heated bath (60°C, 10 min), centrifuged at 3100

rpm (10 min), and transferred via glass pipette to 15-mL glass tubes. Extract was

concentrated (N2 concentrator at 40 °C), supplemented with an 80-µL aliquot of glacial

acetic acid (1% by volume), and brought to 8 mL with acetonitrile. For purification, 1.8

mL of extract was added to ENVI-Carb (25-50 mg) in a microcentrifuge tube, which

was vortexed and then centrifuged for 30 minutes at 14,000 rpm. Extract (1.2 mL) was

transferred to a second microcentrifuge tube and centrifuged for 30 minutes at 14,000

rpm before final analysis.

LC-MS/MS analysis. High-performance liquid chromatography tandem mass

spectrometry (LC-MS/MS) as reported by Higgins et al. (27) was used to determine

PFAA concentrations. Perfluorooctanoic acid (PFOA, 96%) and perfluorodecanoic acid

(PFDA, 98%), were from Aldrich Chemical Co. (Milwaukee, WI). Potassium

perfluorooctane sulfonate (PFOS, 98%) was from Fluka through Sigma-Aldrich (St.

Louis, MO). Perfluorononanoic acid (PFNA, 97%) was from Sigma-Aldrich (St. Louis,

MO). Mass labeled internal standards [13C5] PFNA, [13C2] PFDA, [13C2] PFOS, N-

deuterioethylperfluoro-1-octanesulfonamidoacetic acid ([D5]–N-EtFOSAA) were from

Wellington Laboratories (Guelph, ON, Canada), and [13C2] PFOA was from Perkin-

Elmer Life Sciences (Boston, MA). Two mass transitions were monitored for each

analyte (Table 4.1). Analyte-dependent mass-labeled internal standards (IS) spiked

immediately prior to final LC-MS/MS analysis were used for peak verification and

normalization. Quantification was achieved using a 1/x weighted standard calibration

curve (8-12 points) for the primary mass transition and confirmed by the secondary

76

transition. Calibration standards were prepared from a mixed stock solution in 70:30

methanol/aqueous ammonium acetate (0.01%) and were run at the beginning and end of

each LC-MS/MS sample batch. The stock solution contained all analytes, and analyte

concentrations were corrected for impurities. The instrument LOQ was determined as

the lowest calibration curve point with a signal to noise ratio greater than 30:1, an

accuracy between 70 and 130%, and a peak area at least twice that of the largest blank

peak for that sample batch.

Table 4.1. Analyte primary and secondary transitions monitored, internal standards

(IS), average concentration of MDL samples (with standard deviation of 12 replicates),

and calculated MDLs.

Analyte Primary

trans. (m/z)

Secondary trans. (m/z)

IS IS trans. (m/z)

Conc. (ng/g dw)

MDL (ng/g ww)

PFOA 413 > 369 413 > 169 [13C2] PFOA 415 > 370 7.8 ± 0.5 1.3 PFNA 463 > 419 463 > 219 [13C5] PFNA 468 > 423 10. ± 3 6.9 PFDA 513 > 469 513 > 219 [13C2] PFDA 515 > 470 2.5 ± 0.2 0.4 PFOS 499 > 99 499 > 80 [13C2] PFOS 503 > 99 47 ± 2 6.4 PFDS 599 > 99 599 > 80 [D5] N-

EtFOSAA 589 > 419 5.7 ± 0.3 0.81

Quality assurance and data analysis. Solvent blanks were run every six samples

to monitor instrument background. To monitor ion suppression and enhancement,

matrix spike (MS %) recoveries were determined for each extract by performing a

second analysis of each extract, spiked with a known concentration of PFAA analyte.

Tissue concentrations are reported for MS % recoveries that were between 70 and

130%. Tissue samples were analyzed in triplicate; relative standard deviations of PFOS

replicates averaged 5% for white sturgeon samples. In several instances, tissue PFAA

concentrations were at or close to the LOQ, yielding at least one replicate above and

one below the LOQ. In these cases, the tissue concentration was reported as the average

of the measured concentration for replicates above the LOQ and the LOQ for

concentrations below the LOQ. The method detection limits (MDL) determined from

the extraction of replicate fish liver tissue samples (n = 12) are displayed in Table 4.1.

The MDL was calculated as the product of the relative standard deviation of sample

replicates and the student’s t-statistic (99% confidence level) for 11 degrees of freedom.

77

PFOA was spiked into dried tissue whereas other values were determined from analytes

already present in the extracted tissue. A least-squares regression linear fit with

corresponding coefficient of determination (R2) was calculated and displayed for each

correlation plot. Data were analyzed in Microsoft Office Excel (Microsoft Corporation;

Redmond, WA) and Kaleidagraph (Synergy Software Systems; Dubai, United Arab

Emirates).

4.4 Results and Discussion

White sturgeon PFAA tissues concentrations in the San Francisco Bay. White

sturgeon liver PFAA concentrations (ng/g wet weight) are displayed in Figure 4.2.

PFOS was detected in 14 of 15 samples, ranging in concentration from 14 ng/g ww to

180 ng/g ww. PFOS was also detected in the striped bass (83 ng/g ww) and leopard

shark (37 ng/g ww) samples. PFOS was below the LOQ in four white sturgeon muscle

tissue samples also analyzed as well as in side-by-side extracted blanks. For

comparison, concentrations of PFOS were 180-680 ng/g ww in livers of polar bears

from Alaska, up to 2570 ng/mL in blood plasma of bald eagles less than 200 days old,

and as high as 300 ng/g ww in fish (24). PFDS was detected in 13 of 15 white sturgeon

fish liver samples, ranging from 4.1 ng/g ww to 16.8 ng/g ww. PFDA was detected in 6

samples (0.9 – 8.2 ng/g ww), though five additional samples exhibited low PFDA MS%

recoveries and are thus not included. PFNA was greater than the LOQ in 12 of 15 white

sturgeon fish liver samples, ranging from 2.2 ng/g ww to 20.1 ng/g ww. However, 10 of

these samples were less than the MDL. PFNA concentrations above the LOQ, but not

necessarily above the MDL, are displayed in Figure 4.2. PFOA concentrations were

near or below the LOQ for all samples analyzed. Additionally, because PFOA was

detected in one of five blank samples (at a level near the LOQ), tissue concentrations

are not reported for this analyte. Few samples exhibited PFOA concentrations higher

than the LOQ in a global study of birds, fish, and marine mammals (24). PFOS

exhibited relatively high concentrations with wide variability; thus further analysis

regarding correlations with physiological and ecological parameters is limited to

comparisons with PFOS.

78

Figure 4.2. Measured PFAA concentrations (ng/g ww) in white sturgeon fish livers (n =

15). The tissue samples were archival specimens from animals taken from North San

Francisco Bay in December 1999 and January 2000. The boundary of the box indicates

the 25th and 75th percentile; a line within the box marks the median; whiskers above

and below the box indicate the maximum and minimum concentrations; outlying points

(>1.5 times the upper quartile) are shown as open circles. Number of detects for each

PFAA are also shown; samples with concentrations below the LOQ were excluded from

the plot.

An ecological perspective on contaminant variability: Influence of trophic level

and feeding location. Stable isotope ratios for nitrogen (δ15N) and carbon (δ13C) act as

naturally occurring intrinsic tracers by providing integrated measures of trophic

relationships and feeding locations along a salinity gradient. Isotope results are

presented as deviations from standard reference materials, where: δX = [Rsample/Rstandard

– 1] × 103. Here, X is 13C or 15N and R is 13C/12C or 15N/14N. Because 15N becomes

enriched with increasing trophic level (by 2.5 – 5% between prey and predator) without

varying along a salinity gradient in lower trophic level organisms (bivalves and

zooplankton), this ratio can serve as a quantitative measure of trophic position (108). A

higher δ15N for an individual within a given species may indicate consumption of higher

trophic level biota, due to the introduction of more prey-predator trophic steps from

79

baseline organism nitrogen signatures to the higher trophic position of the individual

under consideration.

Figure 4.3a displays PFOS white sturgeon concentrations with trophic position,

measured by muscle and liver δ15N. White sturgeon have enriched δ15N over lower

trophic organisms, such as the clams that are found as a dominant food items in

sturgeon digestive tracts (108). A wide range of PFOS concentration occurs with little

variability in δ15N. Notably, the highest liver PFOS concentration corresponds to the

highest trophic position white sturgeon. PFOS concentrations increase with trophic

position, as concentrations in predatory animals exceeded concentrations in their diets

(24). The high trophic position of this individual may result from increased

consumption of higher trophic level biota through a piscivorous dietary pattern rather

than more typical clam-based consumption. Additionally, this sturgeon was second to

the smallest, by length, of all white sturgeon considered, and PFOS concentrations

decreased with increasing fish length (Figure 4.3c). Although Martin et al. suggest that

the half-life for PFAAs in trout may be much longer for a larger fish of the same species

(37), growth dilution can be an important determinant of concentration for a substance

with slow uptake or clearance rates (107). The striped bass and leopard shark samples

are included in Figure 4.3 and, as expected, do not reflect the length trend for white

sturgeon. The high trophic level of the striped bass individual corresponds with a

relatively high liver PFOS concentration, whereas the leopard shark PFOS

concentration falls below the average PFOS concentration despite its slightly higher

trophic position.

Stable carbon isotope ratios can identify contributions of different foods in a diet by

tracking distinct isotopic signatures of food types. In estuaries, δ13C shows little to no

enrichment with trophic level but is enriched in algae with increasing salinities due to

the influence of δ13C in dissolved inorganic carbon that is incorporated into algae. As

these distinct isotopic signatures are incorporated into the base of the food web, the δ13C

of consumers will reflect their predominant foraging location, as determined by the

salinity gradient (108). Figure 4.3b displays the relationship between PFOS

concentration and foraging location, increasing on the ordinate from the freshwater

eastern reaches of the estuary to the more saline Suisun Bay. When excluding the high

80

trophic position outlier, PFOS liver concentrations decrease with increasing salinity.

Thus, fish spending a greater time feeding in more saline environments appear to

exhibit lower PFOS liver concentrations. Little is known regarding the affect of changes

in salinity on PFAA accumulation and toxicity, however an increase in distribution

coefficient with increasing water salinity suggests that long-chain PFAAs may “salt-

out” onto particles. This led to an increase in bioaccumulation in Pacific oysters

(Crassostrea gigas), filter-feeding bivalves that accumulate contaminants through

ingestion of contaminated particles (112). For organisms in which aqueous uptake is

more important than dietary accumulation, such as rainbow trout in which the blood-

water interface of gills is a major route of uptake and clearance (37), a decrease in

accumulation with increased salinity may be postulated. Additionally, freshwater

sources of PFAAs, such as wastewater treatment plant effluent (27), likely influence site

specific accumulation of PFAAs.

The relationship of total mercury concentration (organic and inorganic Hg) with

PFOS liver concentrations is displayed in Figure 4.3d. With the exception of the

sturgeon outlier, there appears to be a slight inverse relationship between these two

contaminant concentrations. Although both mercury (in its methylated form) and PFOS

are organic contaminants, PFOS sorption behavior does not typically follow the

paradigm of hydrophobic, lipophilic organic contaminants because of its surfactant

properties. The weak relation between Hg and PFOS is confounded by analysis of

different fish organs, with potentially different sorption properties. The physiochemical

behavior of PFAAs cannot be expected to predictably mimic other organic

contaminants (43). The influence of the unique PFAA chemical properties on

mechanisms controlling bioaccumulation requires further study. No significant

relationship was evident between selenium and PFOS concentrations.

81

Figure 4.3. Stable isotopes, fish length, and muscle Hg concentration plotted with white

sturgeon, striped bass, or leopard shark liver PFOS concentrations on the ordinate. (a)

Muscle and liver δ15N serve as measures of organism trophic position. A high trophic

position outlier was excluded from the linear regression. (b) δ13C represents integrated

measure of organism feeding location, indicating foraging location along a salinity

gradient. Liver PFOS concentration decreases with increasing salinity gradient. (c) An

increase in fish length corresponded to a decrease in liver PFOS concentration. (d) With

the exception of the high trophic position sturgeon outlier, PFOS liver concentration

(ng/g ww) yields and inverse relationship with total muscle mercury concentration (µg/g

dw).

Additional factors contributing to PFAA bioaccumulation. In addition to fish

size, trophic position, and foraging location, other factors not considered in this analysis

influence the exposure and retention of PFOS in white sturgeon tissue. For example,

82

fish age and sex may influence elimination rates of contaminants (37). PFAA source

locations in relation to foraging location are important; PFOS concentrations in

relatively industrialized regions may be several times greater than those in isolated areas

(24). Further, although PFOS is metabolically inert (37), precursors to this compound

are not – yielding additional sources of PFOS and other PFAAs in organisms. The role

of precursor compounds as sources of PFAAs has been studied extensively. N-ethyl

perfluorooctane sulfonamidoethanol (N- EtFOSE), produced directly and attached to

phosphate esters in paper coatings, degrades to N-ethyl perfluorooctane sulfonamido

acetic acid (N-EtFOSAA) in wastewater treatment processes (22). PFOS is the terminal

metabolite of this microbial degradation pathway. N-EtFOSE may also be stripped to

the atmosphere from treatment facilities (22) and oxidized to PFCAs and PFSAs (123).

N-EtFOSAA has been detected in natural San Francisco Bay sediments at levels often

exceeding PFOS (27) and oxidizes to perfluorooctane sulfonamide (FOSA) and PFOA

in hydroxyl-mediated photolysis (23). Recent studies elucidate the biotransformation

from such biologically labile precursor compounds to PFAAs as terminal metabolites.

N-EtFOSAA appears to undergo biotransformation to PFOS in an aquatic oligochaete

(116), likely contributing to organism PFOS body burden. PFCAs may form from

fluorotelomer alcohols (FTOHs) via biotransformation in rats and other organisms (124,

125), oxidation in the atmosphere (21), and indirect photolysis (126). 8:2 FTOH (127)

and fluorotelomer acrylates (128) biotransform to PFOA in rainbow trout.

Depuration and elimination rates of PFAAs from fish vary with PFAA chain length

(a proxy for hydrophobicity) as well as head group type (sulfonates or carboxylates)

(37). Such chemical properties may influence the extent of accumulation within a

certain organ, with blood concentrations in rainbow trout exhibiting higher

concentrations than the kidney and liver (114). Chemical properties influence the

interaction of PFAAs with gill membranes, an important site for elimination of

contaminants in fish (37). Controls of contaminant elimination on a cellular level may

contribute to PFOS bioaccumulation. For example, organic anion transporters play a

role in pharmacokinetics of PFAAs (129). Additionally, greatest inhibition of cellular

efflux transporter activity, which serves as a first line of defense against toxic

compounds (130), occurred with exposure to the longer chain acids, PFNA and PFDA,

83

for the marine mussel Mytilus californianus (122).

4.5 Significance

Perfluorinated compounds are bioaccumulative and ubiquitous among

environmental samples. In order to better understand the accumulation and variability of

concentrations measured within a species collected over a limited spatial and temporal

range, ecological and physiological processes must be considered. Wide variability of

concentrations of perfluorooctane sulfonate can occur within a single species type.

PFOS concentrations decreased for sturgeon feeding primarily in more saline

environments. Although correlations such as those presented are useful in developing an

assessment of the fate of PFOS in an ecosystem, a complete understanding of the

ecological and physiological diversity that influences contaminant concentrations

requires analysis of mechanistic processes such as elimination and uptake rates,

compound specific properties, and ecosystem dynamics such as contaminant sources

and transport processes. Even simplistic Tier 1 screening measures for evaluating the

bioaccumulative potential of new chemicals, a necessity for effective decision-making,

generally do not incorporate expected bioaccumulation mechanisms relevant to PFAAs.

In the following two chapters, the binding of perfluoroalkyl acids to a model protein,

BSA, is explored. In Chapter 5, commonly utilized dissociation constants relating free

chemical concentrations to protein-bound concentrations are determined for PFOA and

PFNA. Analytical methods are applied over a wide range of concentrations to assess the

contribution of different binding regimes at varied concentrations and for comparison to

literature values. In Chapter 6, a protein-water partition coefficient is quantified for C5 –

C10 PFCAs as well as C4, C6, and C8 PFSAs, and physiochemical mechanisms of

interactions are explored.

Acknowledgment. This work was conducted while funded by the National Defense

Science and Engineering Graduate Fellowship. Thanks to Christopher P. Higgins and

Laura A. MacManus-Spencer for laboratory guidance and feedback. We thank Robin

Stewart for provision of tissue samples and insights on analysis and interpretation.

84

85

Chapter 5

Investigating binding to a model protein:

Noncovalent interactions of long-chain

perfluoroalkyl acids with serum

albumin1

5.1 Introduction

Used throughout the past half-century in a variety of industrial and commercial

applications, perfluoroalkyl acids (PFAAs) are a class of environmentally persistent

anionic surfactants detected globally in water, air, sediment, and biota (20, 25). Field

1 Reproduced (with modifications) with permission from Bischel, H. N.; MacManus-Spencer, L. A.; Luthy, R. G. Noncovalent interactions of long-chain perfluoroalkyl acids with serum albumin. Environ. Sci. Technol. 2010, 44 (13), 5263-5269. Copyright 2010 American Chemical Society.

86

and laboratory studies indicate that perfluorooctanesulfonate (PFOS) and

perfluorocarboxylates (PFCAs) with greater than seven fluorinated carbons

bioaccumulate and biomagnify in aquatic food webs (25, 37, 113). Tissue distribution

studies show PFAA concentrations are greatest in body compartments high in protein

content, such as the liver, kidney and blood of organisms (24, 40). Typical

concentrations of perfluorooctanoate (PFOA) and PFOS in the serum of non-

occupationally exposed humans are 4 – 7 and 25 – 46 ng/mL, respectively (131). The

half-life of PFOA in serum varies widely by species and sex and is considered long (3.1

– 4.4 years) in human blood (132). Species-specific differences in PFAA distribution

patterns and retention may be influenced by active uptake by organic anion transporters

(41). Studies assessing PFAA-protein interactions (42, 133-138) may shed light on the

tissue distribution patterns, bioaccumulation, and in vivo bioavailability of these

chemicals.

Protein binding. The binding of PFAAs to proteins was first reported in the 1950s,

when PFAAs were investigated for their ability to aid in protein precipitation (139). In

the 1960s, organofluorine compounds were first detected in human blood serum (140).

Serum albumin, the most abundant protein in blood plasma (35 – 50 g/L) (141), binds a

variety of endogenous and exogenous ligands including fatty acids, amino acids, metals,

and pharmaceuticals (142) and was reported as the major binding protein for PFOA in

blood (42). PFAA-protein interactions result from the unique surfactant nature of

PFAAs. The highly hydrophobic perfluorocarbon tail paired with a strongly polar

carboxylate or sulfonate head group resembles the structure of fatty acids and facilitates

both hydrophobic and ionic interactions with proteins. In fact, PFOA binds to liver- and

kidney-fatty acid binding proteins (135), and PFAAs may interfere with the normal

binding of fatty acids or other endogenous ligands to liver-fatty acid binding protein

(136). However, the rigidity of the perfluorocarbon tail differs from the relatively more

fluid hydrocarbon tail (143), limiting extrapolation of fatty acid – albumin binding

results to their fluorinated counterparts.

Studies of PFAA-albumin interactions using spectroscopic methods (137, 144-146),

electrophoresis (144), 19F NMR (42, 137), dialysis (139, 147), and surface tension (144,

145) report a wide range of association constants (102 – 106 M-1) with most values

87

suggesting relatively weak binding (<104 M-1). However, PFAAs are highly bound to

proteins in rat, monkey, and human plasma (138), suggesting stronger, specific binding.

Many studies have used relatively high PFAA:albumin mole ratios; results from such

studies are difficult to extrapolate to physiological protein and substrate concentrations

and may explain weaker observed binding.

In this study, bovine serum albumin (BSA) serves as a model protein for

characterizing PFAA-protein interactions. BSA is widely used as a model protein in

evaluating protein-ligand interactions, as the sequences of human and bovine serum

albumins are highly conserved (44, 142, 148). Here association constants (Ka) and

stoichiometries for PFAA-albumin complexes are quantified over a range of

PFAA:albumin mole ratios via equilibrium dialysis and automated nanoelectrospray

ionization mass spectrometry (nanoESI-MS), with the specific goal of providing

quantitative binding data at low ligand:protein mole ratios. Additional tests are

performed with human serum albumin (HSA) for comparison to results for PFAA-BSA

interactions. Quantification of PFCA-BSA association constants at physiologically

relevant PFAA:protein mole ratios using equilibrium dialysis, the most

thermodynamically sound and straightforward protein-ligand analysis technique (149),

has not been reported. Prior equilibrium dialysis studies (139, 147) reporting PFOA-

albumin affinities at relatively high PFAA:albumin mole ratios (>3) used less sensitive

and selective analytical methods (e.g. surface tension and spectrophotometry) to

measure PFOA concentrations.

Analysis via automated nanoESI-MS provides complementary information about

PFAA-BSA interactions at low PFAA:albumin mole ratios and is explored as a

technique to more rapidly evaluate the binding affinities and stoichiometries.

Automated nanoESI-MS overcomes some disadvantages of conventional ESI-MS such

as irreproducibility due to non-automated sample introduction. This method, with

nL/min flow rates, is gentler than conventional ESI-MS as complexes are transferred

from solution to the gas phase, reducing disruption of intact protein-ligand complexes

(150). Analysis of PFOA with rat serum albumin (RSA) (42) and rat-specific proteins

(135) via electrospray ionization mass spectrometry (ESI-MS) demonstrated intact

complexes. However, challenges of noisy mass spectra due to additional protein adducts

88

(135) and limits in instrument sensitivity and reproducibility (133) were reported in

conventional ESI-MS PFAA-protein binding studies.

Although PFAAs are consistently detected in protein-rich tissues, this concept has

yet to be incorporated into PFAA bioaccumulation models. Models utilizing a lipid

partitioning paradigm may underestimate the bioaccumulative potential of PFAAs

(113), although this remains to be further tested. PFAA-protein association constants

may thus contribute to improved predictive models for the bioaccumulation of these

chemicals.

5.2 Materials and Methods

Materials. Essentially fatty acid free human serum albumin (HSA, 96%), Cohn

Fraction V protease free, essentially γ-globulin free bovine serum albumin (BSA, 99%)

and ammonium acetate were from Sigma-Aldrich, Inc. (St. Louis, MO). Fraction V

fatty acid-free BSA (99.9%) was from EMD Biosciences, Inc. Standards of

perfluorooctanoic acid (PFOA, 96%) and perfluorodecanoic acid (PFDA, 98%) were

from Aldrich Chemical Co. (Milwaukee, WI). Perfluorononanoic acid (PFNA, 97%)

and potassium perfluorooctane sulfonate (PFOS, 98%) were from Fluka through Sigma-

Aldrich (St. Louis, MO). The internal standard [13C5] PFNA was from Wellington

Laboratories (Guelph, ON, Canada) and [13C2] PFOA was from Perkin-Elmer Life

Sciences (Boston, MA). Internal standards had purities greater than 98%, as reported by

the suppliers.

Equilibrium dialysis. Purity-corrected stock solutions of PFOA and PFNA were

prepared in 50 mM sodium phosphate buffer (pH 7.4) in polypropylene containers. A

sonicating bath (~38 °C) was used to assist in the dissolution of PFAAs without the use

of an organic co-solvent. Stock solutions of BSA were prepared fresh daily in the

sodium phosphate buffer for dialysis experiments. Spectra/Pore dialysis membrane

tubing with 6000-8000 Da molecular weight cutoff (Spectrum Laboratories, Inc.,

Rancho Domingo, CA) was cut into 7.5-cm pieces, soaked in deionized water for 30

minutes, and rinsed with deionized water followed by Milli-Q water. Dialysis tests were

prepared with either 500 µM albumin (human or bovine) exposed to a single

concentration of PFNA or 1µM BSA exposed to a range of PFOA or PFNA

89

concentrations. For all tests, a known volume of BSA or HSA solution was added to the

dialysis bag, spiked with a PFAA solution prepared in the same buffer, and brought to 3

mL or 10 mL. Molecular weights of 66430 Da (151) and 66248 Da (152) were used to

calculate the final concentrations of BSA and HSA, respectively. The dialysis bag,

which is impermeable to the protein and its complexes but freely permeable to PFAAs,

was equilibrated in 300 or 500 mL of the sodium phosphate buffer in polypropylene

dialysis reservoirs at laboratory temperatures (approximately 21 °C). Controls were

prepared using a buffer-only solution in dialysis bags with a PFAA spike, and blanks

were prepared using a BSA solution with no PFAA spike. Free and bound PFAA

concentrations were determined via liquid chromatography tandem mass spectrometry

(LC-MS/MS). Details on sample preparation and LC-MS/MS analysis are available in

the Supporting Information. Instrumentation and operating parameters were previously

reported (27).

Dialysis bag and reservoir sample preparation. For 500 µM albumin tests,

dialysis reservoirs (300µL of 50 mM sodium phosphate, pH 7.4) were prepared in

triplicate or quadruplicate with albumin in dialysis bags exposed to a single

concentration of PFNA. Dialysis bags were equilibrated in separate reservoirs for 96

hours prior to sampling. PFNA concentrations were measured at equilibrium both inside

the dialysis bag and in the external reservoir. Extraction of PFNA from albumin

samples taken from dialysis bags was performed using a method modified from

Stevenson et al. (122). For triplicate 500 µL samples from dialysis bags, albumin was

precipitated from solution and PFNA extracted with acidified acetonitrile (1% v/v

glacial acetic acid) in a 15-mL polypropylene tube. After addition of the acidified

acetonitrile (9.5 mL), the tube was vortexed (30 sec), sonicated (60 °C, 10 min), and

centrifuged (3000 rpm, 10 min). Extracts were purified using a dispersed sorbent

(ENVICarb, 25-50 mg) by vortexing 1.8 mL of extract in polypropylene

microcentrifuge tubes containing the sorbent followed by centrifugation (14000 rcf, 10

min). Samples were further diluted in acetonitrile as needed. Glass HPLC vials

contained the acetonitrile extract (200 µL), HPLC grade water (200 µL), 1:1 v/v

methanol:buffer (100 µL), and an internal standard prepared in HPLC grade water that

was spiked prior to analysis (100 µL). Spike/recovery experiments (n = 7) were used to

90

determine the efficiency of the extraction procedure. The averages of PFNA recoveries

were 90.2% from 1µM BSA and 80.5% from 500µM, and average PFOA recovery was

81.8% from 1µM BSA. Reported concentrations were not corrected for the extraction

efficiencies. Matrix-matched calibration curves were prepared with PFAA standards in

a 1:1 methanol:buffer solution.

For 1 µM dialysis tests and control reservoirs, PFOA or PFNA concentrations were

measured in triplicate at the initiation of the test (0 hours) and at equilibrium both inside

the dialysis bag and in the external reservoir (48 hours). In order to ensure equilibrium

was reached in test reservoirs, the final bag and reservoir concentrations were required

to be equal in control experiments conducted on the same day. Dialysis bag samples

(500 µL each) were prepared using the above method for a subset of the PFNA tests

(when free PFNA was greater than 330 µM) or added to an equal volume of methanol

in polypropylene microcentrifuge tubes, vortexed, and diluted into the analytical range

of the LC-MS/MS when necessary using a 1:1 methanol:buffer solution. Data were

restricted to those with a PFAA mass balance of 70-130%, as calculated using initial

and final samples from the dialysis bag and external reservoir. A comparison of BSA

tests and buffer controls spiked with PFAAs in the dialysis bag showed agreement in

determined concentrations (See Supporting Information, Figure 5.5S), suggesting

limited matrix effects on the signal from this method. All reservoir samples were

prepared in 1:1 methanol:buffer, and samples were stored at 4°C until analysis.

Standard 12-point calibration curves were also prepared in 1:1 (v/v) methanol:buffer

and all samples and standards were spiked prior to analysis with a mass-labeled internal

standard prepared in HPLC grade water.

LC-MS/MS analysis. Equilibrium dialysis samples were analyzed for PFOA or

PFNA via liquid chromatography tandem mass spectrometry (LC-MS/MS). Samples

and standards were injected (40 µL) onto a 40 mm x 2.1 mm Targa Sprite C18 column

(5-µm particle size, Higgins Analytical, Mountain View, CA) equipped with a C18

guard column (Higgins Analytical) and analyzed via LC-MS/MS using instrumentation,

chromatography conditions, and negative electrospray ionization multiple reaction

monitoring (MRM) mode operating parameters previously described (27). Analyte

transitions and internal standards used to monitor and quantify each analyte are reported

91

by Higgins et al. (116). Blanks were injected every 3-6 samples to monitor carry-over,

and standards were run before and after samples to monitor instrument drift. Initial

eluent conditions were 35% methanol:65% 2 mM aqueous ammonium acetate.

Methanol was ramped to 100% over 7.5 min, held at 100% for 2.5 min, reverted to 35%

over 0.5 min, and held until 13 min at a flow rate of 0.25 mL/min. Optima-grade

methanol was from Fisher Scientific (Fair Lawn, New Jersey). A VICI Cheminert

system (Valco Instruments Co., Inc.) was employed for the first 4 minutes to divert

buffer salts away from the source needle. Data were processed using Analyst software

version 1.4.2. The detection limit was defined as the lowest calibration standard for the

transition used for quantitation with 70 to 130% accuracy, at least twice the peak area of

blanks, and a signal-to-noise ratio greater than 3:1. Detection was confirmed by the

second monitored transition.

Nanoelectrospray ionization mass spectrometry. Analysis of noncovalent PFAA-

BSA interactions was conducted by automated nanoelectrospray ionization mass

spectrometry (nanoESI-MS) for the PFAAs shown in Figure 5.4S. A 9 mM ammonium

acetate buffer (pH 7) was selected to minimize interfering adducts during electrospray

ionization. Individual stock solutions of PFAAs (1 mM) were prepared with the

ammonium acetate buffer in polypropylene bottles. A sonicating bath (~38°C) was used

to assist in the dissolution of PFAAs. Stock solutions of BSA (100 µM) were prepared

fresh daily in 9 mM ammonium acetate (pH 7) at room temperature and dialyzed

overnight, with exchange of the same buffer in an external reservoir. BSA solutions

were transferred to polypropylene microcentrifuge tubes, spiked with a PFAA in buffer,

and diluted with buffer to a final BSA concentration of 50 µM. PFAA-BSA solutions

were prepared over a range of ligand:protein mole ratios (0:1, 0.1:1, 0.5:1, 1:1, 2:1, 4:1,

and 8:1) and were allowed to equilibrate for one or more hours before same-day

analysis. Samples tested after approximately one month indicated no change in

equilibrium.

A Waters Micromass Q-Tof API-US quadrupole time-of-flight mass spectrometer

(Micromass, Milford, MA) equipped with an Advion Triversa Nanomate nano-

electrospray robot (Advion BioSystems, Inc. Ithaca, NY) was used for nanoESI-MS

analysis. Samples in the ammonium acetate buffer were infused through the ESI Chip (5

92

µm) allowing nL/min flow rates. Optimal conditions of instrument pressures and

voltages were determined by maximizing intensities of peaks due to PFAA-BSA

complexes while maintaining peak resolution and ion current. Typical operating

conditions for the Q-Tof and nanomate in positive-ion mode were: spray voltage, 1.88

kV; sample pressure, 0.5 psi; extraction cone voltage, 6.0 V; source temperature,

100°C; desolvation temperature, 50°C; desolvation gas flow rate, 80 L/hr. Argon was

the collision gas. Backing (2-3 mbar), Penning (~10-5 mbar), and Tof pressures (~10-7

mbar) were adjusted each test day for detection of protein-ligand complexes. In titration

experiments, mass spectra were acquired for 4 min at cone voltages of 100 V or 130 V

and the collision energy was set at 10 V. Mass signals were collected over the scan

range m/z 1000–5000. Data were processed using MassLynx software version 4.1 from

Waters. Mathematical deconvolution of multiply-charged ion spectra of native BSA

using MaxEnt produced an accurate mass of BSA, which was consistent with the

theoretical molecular weight of BSA (66.4 kDa). For quantifying association constants,

deconvoluted spectra acquired with 20 iterations using MaxEnt were smoothed, and a

10-channel center was applied. Spectra were integrated from the expected mass of the

protein or ligand-bound protein to the next expected mass with a signal to noise ratio

greater than 2:1. Unless otherwise stated, statistical comparisons were conducted using

a student’s two-tailed t-test assuming equal variance.

Binding model. The binding of ligand, L, to binding site, j, on protein, P, can be

described by a series of stepwise equilibria for a total of n binding sites:

P + L! PL1

PL1+ L! PL

2

!

PLj!1 + L" PL

j (5.1)

!

PLn!1 + L" PL

n

The general form for stoichiometric binding constants describing binding sites that

act independently from each other is given by:

93

Ka, j=

[PLj]

[PLj!1][L]

(5.2)

The average number of bound ligands per protein molecule,

!

" , can be expressed in

terms of the association constant, Ka, the free ligand concentration, [L], and the total

number of binding sites on the protein, n, yielding (153):

! =

j( Ka, j)[L]

j

1

j

!j=0

n

"

1+ j( Ka, j)[L]

j

1

j

!j=0

n

" (5.3)

The full form of this equation can be simplified by considering classes of binding

sites with similar affinities, as described by Scatchard (154). For ionic surfactants such

as fatty acids binding to water-soluble proteins, high and low affinity binding sites may

be considered, with high-affinity interactions dominant at low [L]:[P] mole ratios (141,

155, 156). Thus, for two classes of binding sites, the equation becomes:

! =n1K

a,1[L]

1+Ka,1[L]

+n2K

a,2[L]

1+Ka,2[L]

(5.4)

where

!

" is equivalent to the ratio of the bound PFAA concentration to the total protein

concentration. Over a narrow range of low [L]:[P] mole ratios, where high affinity

binding dominates, data may be best described with a one-class model. For equilibrium

dialysis results, a nonlinear curve fit based on a Levenberg-Marquardt algorithm that

iteratively minimizes the sum of the squared error (Chisq) between the original data and

the calculated fit was applied to results for bound PFAAs using Kaleidagraph software

(See Supporting Information, Tables 5.4S – 5.7S). Fits were allowed an error of 0.1%,

and initial guess values in the iterations were unity for n1 and Ka,1.

Binding theory applied to nanoESI-MS results. For a series of stepwise

equilibria, with a total of n binding sites, the total protein ([P]o) and total ligand ([L]

o)

concentrations can be expressed as:

[P]o= [P]+[PL]+[PL

2]+...+[PL

n] (5.5)

[L]o= [L]+[PL]+ 2[PL

2]+...+ n[PL

n] (5.6)

The total protein can be expressed in terms of R, the ratio of complex to free protein:

94

Rj=[PL

j]

[P] (5.7)

yielding:

[P]o= [P]+ R

1[P]+ R

2[P]+...+ R

n[P] (5.8)

The free protein concentration can then be written as:

[P]=[P]

o

1+ R1+ R

2+...+ R

n

(5.9)

The free ligand concentration [L] can be expressed as:

[L]= [L]o! R

1[P]! 2R

2[P]!...! nR

n[P] (5.10)

= [L]o![P]

o(R

1+ 2R

2+...+ nR

n)

1+ R1+ R

2+...+ R

n

(5.11)

The stepwise association constants can then be expressed as:

Ka, j=

[PLj]

[PLj!1][L]

=Rj

Rj!1[L]

=Rj

Rj!1([L]o !

[P]o(R

1+ 2R

2+...+ nR

n)

1+ R1+ R

2+...+ R

n

)

(5.12)

The association constant for the case of one bound ligand (j = 1) is:

Ka,1=

[PL]

[P]([L]o![PL])

=R1

[L]o![P]

oR1

1+ R1

(5.13)

5.3 Results and Discussion

Equilibrium dialysis: PFNA-albumin binding at low [L]:[P] mole ratios.

Equilibrium dialysis provides direct measurement of free and bound PFAA

concentrations, minimizing uncertainty from indirect methods such as fluorescence

spectroscopy (137). To quantitatively assess the binding of PFNA to albumin at

physiologically relevant [L]:[P] mole ratios, equilibrium dialysis was conducted with

500 µM HSA or BSA and 0.24 ± 0.08 µM or 0.32 ± 0.07 µM PFNA, respectively, in 50

mM sodium phosphate buffer. Based on measured equilibrium concentrations, at low

[L]:[P] mole ratios (10-3 – 10-4) greater than 99.9% of PFNA was bound to both HSA

and BSA (Table 5.1). The percent bound was also high for BSA dialysis tests in 50 mM

sodium phosphate buffer with an additional 9 g/L NaCl (representative of physiological

95

salinity) although the high salt concentration slightly reduced the fraction of bound

PFNA. Further investigation into the effects of ionic strength and the presence of

competing ligands, such as fatty acids, is needed.

Debruyn and Gobas (44) utilize BSA and HSA binding data to assess the sorptive

capacity of animal protein for a range of neutral hydrophobic organic contaminants

(HOCs) by calculating distribution coefficients between BSA or HSA and water.

Though representation of PFAA-protein binding by a distribution coefficient may be

acceptable only at very low solute concentrations when the binding isotherm tends

towards linearity, the calculation of such a distribution coefficient may be useful in

PFAA bioaccumulation models. The ratio of analyte concentration in the bound phase

to that in the aqueous phase (KPW) is determined by the protein concentration, [P], the

fraction bound to protein (fbound), and the partial specific volume of protein in aqueous

solution (ρalbumin):

KPW =CP

CW

=fbound

!albumin ![P](1" fbound ) (5.14)

Log KPW values for the low PFNA:albumin mole ratio tests are presented in Table 5.1.

A partial specific volume of 0.733 ml/g was used in calculations for BSA and HSA

(141). The measured value of KPW for PFNA is greater than all protein distribution

coefficients presented by Debruyn and Gobas, where log KPW ranged from less than 0.1

to 3.5 (44). Log KPW for PFNA is greater than an experimentally determined octanol-

water distribution coefficient (log KOW = 2.57 ± 0.07) (118), highlighting the potential

influence of interactions with nonlipid materials on PFNA distribution in organism

tissues. Log KPW for neutral HOCs was also generally greater than Log KOW for less

lipophillic compounds (log KOW < 2) (44). Confirming results obtained for PFNA with

essentially fatty acid free BSA (99.93 ± 0.01% bound for a 6 ± 1 × 10-4 [L]:[P] mole

ratio, corresponding to a log KPW of 4.80 ± 0.08) suggest that trace fatty acids in the

essentially γ-globulin free BSA used in this and subsequent tests had little influence on

the binding affinities.

96

Table 5.1. Percent bound and log KPW for PFNA binding to 500 µM albumin

determined by equilibrium dialysis and LC-MS/MS.1

Albumin Conditions Percent Bound Log KPW [Ligand]: [Protein]

mole ratio HSA 50 mM sodium

phosphate, pH 7.4 99.95 ± 0.01% 4.93 ± 0.05 5 ± 1 × 10-4

BSA 50 mM sodium phosphate, pH 7.4

99.92 ± 0.01% 4.74 ± 0.05 6 ± 1 × 10-4

BSA 50 mM sodium phosphate, pH 7.4; 9 g/L NaCl

99.89 ± 0.01% 4.56 ± 0.05 10 ± 1 × 10-4

Equilibrium dialysis: PFAA-BSA binding over a wide range of [L]:[P] mole

ratios. Equilibrium dialysis was used to quantify free and albumin-bound PFAAs in an

equilibrated system over a wide range of PFNA and PFOA concentrations and 1 µM

BSA; [L]:[P] mole ratios ranged from 0.02 to 120 in these experiments. Concentrations

in all initial reservoir samples (0 hours) were below the detection limit, and a null value

was used for mass balance calculations. Initial dialysis bag PFAA concentrations ranged

from 1.6 µM to 2700 µM prior to equilibration in reservoirs. Dialysis bags reached

equilibrium with external reservoirs within 48 hours (see Supporting Information,

Figures 5.6S and 5.7S), when final samples were taken. The average relative standard

deviations for final bag and reservoir concentrations ranged from 4 – 6% for PFOA and

PFNA. Initial bag concentrations, which generally required additional dilution of

samples prior to analysis, had higher relative standard deviations (9% for PFOA and

11% for PFNA) than those for final bag concentrations. Average mass balance results

for control experiments (94%, n = 5 for PFOA and 107%, n = 7 for PFNA) indicate that

PFAAs do not significantly bind to the dialysis membrane or reservoir vessels.

Equilibrium dialysis results for PFOA and PFNA with 1 µM BSA are displayed in

Figure 5.1 along with 500 µM BSA and HSA results for PFNA. Although anionic

1 Errors represent 95% confidence intervals and were calculated from the root mean squared error of all results conducted at a 500 µM albumin concentration (n = 13). Means and error calculations for Log KPW were performed on the log-transformed data.

97

surfactants may cause protein denaturation, this is not expected to occur over the

concentration range tested (157), as reservoir concentrations were well below the

critical micelle concentrations (CMCs) of PFOA and PFNA (8.7 – 10.5 mM and 2.8 –

5.6 mM, respectively) (16). Total PFOA and PFNA concentrations, representing the

sum of free and bound PFAA concentrations, were measured inside dialysis bags at

equilibrium. Final bag concentrations were greater than reservoir concentrations for all

tests, indicating that PFAAs were bound to BSA and that PFAA-BSA complexes were

retained in the dialysis bags. Osmotic dilution of the retentate was assumed to be

negligible. Bound PFAA concentrations were calculated from concentrations measured

inside the dialysis bag in excess of free concentrations.

Figure 5.1. Equilibrium dialysis results for PFOA (a) and PFNA (b) where

!

" is the

average number of PFAA molecules bound per albumin. Data represent averages of

triplicate measurements from each test reservoir or dialysis bag. PFOA and PFNA data

from experiments conducted with 1 µM BSA were fit using Equation 5.4.

Association constants and binding stoichiometries for PFOA- and PFNA-BSA

complexes determined via equilibrium dialysis with 1 µM BSA are reported in Table

5.2. Data were fit both over the full range of test concentrations using Equation 5.4 and

up to a 5:1 PFAA:BSA mole ratio using a one-class binding model. Further details of

the fitting approaches and results are available in the Supporting Information. To reduce

the number of parameters being simultaneously solved in Equation 5.4 and thus reduce

the error in the solved parameters, association constants of 630 M-1 and 8000 M-1 were

98

employed for Ka,2 for PFOA and PFNA, respectively. These values were determined

under the same buffer conditions but at higher PFAA:albumin mole ratios (15 – 200)

using 19F NMR (137). An increased error was observed at higher reservoir

concentrations when samples were diluted into the analytical range of the LC-MS/MS

(see Supporting Information, Figure 5.9S). Due to a strong influence of the two highest-

concentration PFOA data points, which also indicated weak binding at a high mole

ratio, these data points were excluded from the fit presented in Table 5.2 for the PFOA

two-class model. The effects of applying various weighting factors for the full PFOA

and PFNA data sets were tested and yielded similar results to those in Table 5.2. The

primary association constants determined are similar for PFOA and PFNA, on the order

of 106 M-1 with binding stoichiometries of one to four or five. BSA had 150 ± 20

secondary binding sites for PFOA and 31 ± 2 secondary binding sites for PFNA. A

modest effect of albumin concentration on calculated PFAA association constants, in

which increased protein concentration decreased affinities, has been previously

observed (137). This effect was not evaluated in detail in the present study. However,

applying a one-class binding model to the results at a single PFNA:albumin mole ratio

with 500 µM BSA or HSA, and assuming n = 3, yields similar PFNA-albumin

association constants (on the order of 106 M-1) to those from 1 µM albumin tests (Table

5.3).

Table 5.2. Association constants (Ka) and binding stoichiometries (n) for PFOA and

PFNA binding to BSA determined by equilibrium dialysis.

Compound Ka,1 (M-1) n1 R2 [Ligand]: [Protein]

range Model

PFOA 0.20 (± 0.14) × 106 4.3 (± 2.0) 0.772 0.04 – 5 One-class PFOA 1.4 (± 1.9) × 106 1.4 (± 0.5) 0.913 0.04 – 70 Two-class PFNA 1.1 (± 0.2) × 106 4.6 (± 0.3) 0.943 0.02 – 5 One-class PFNA 3.3 (± 3.2) × 106 2.9 (± 0.7) 0.953 0.02 – 120 Two-class

NanoESI-MS results. In nanoESI-MS experiments, soft ionization maintained

protein-ligand complexes in the gas phase such that PFAA-BSA complexes were

distinctly observed relative to free BSA at a 0.5 ligand:protein mole ratio and greater.

99

ESI-MS is widely used to study noncovalent interactions of small molecules with

proteins (150, 158, 159). During the transfer of ions from the liquid to gas phase, even

weak interactions are largely preserved (159). Exposure of BSA to PFAAs causes a

shift in the observed m/z depending on the number of bound PFAA molecules. The

charge state of peaks and number of bound ligands is determined by:

m

z=MWBSA + j !MWPFAA + i !H[ ]

i+

i (5.15)

where MWBSA and MWPFAA are the molecular weights of BSA and the PFAA,

respectively, j is the number of PFAAs bound per BSA molecule, and i is the charge

state of the protein in the gas phase, produced in positive ion mode. Results show that

multiple PFAAs are bound per BSA molecule at a 1:1 PFAA:BSA mole ratio, with

peaks present at j = 1 and j = 2. This pattern is repeated over a range of charge states

and is most intense at i = 16 and 17 (Figure 5.2).

Representative nanoESI-MS spectra for the most intense charge state (+16) of 50

µM BSA exposed to a range of PFOA and PFNA concentrations are shown in Figure

5.3. Representative spectra for PFDA-BSA and PFOS-BSA complexes are included in

the Supporting Information (Figure 5.11S). Deconvolution of results yields spectra with

peaks at m = MWBSA + j × MWPFAA (e.g., see the Supporting Information, Figure

5.12S). Results for PFOA and PFNA suggest a maximum of eight PFAA molecules

bound per BSA molecule at a 4:1 PFAA:BSA mole ratio. At higher ligand

concentrations (8:1 PFAA:BSA, data not shown), the number of bound PFAAs detected

continues to increase. Han et al. detected up to six binding sites at an 8:1 mole ratio of

PFOA to rat serum albumin using conventional ESI-MS (42). The use of nanoESI-MS

here may have reduced disruption of PFAA-albumin complexes during transfer to the

gas phase, resulting in enhanced detection of intact complexes. At 1:1 and 2:1

PFAA:BSA mole ratios, up to two to four bound PFOA and PFNA molecules were

detected. Binding observed by nanoESI-MS was not categorized into primary or

secondary binding classes, although equilibrium dialysis results suggest one to four or

five primary binding sites may be occupied at similar concentration ratios. Overlap of

ligand-bound protein charge states limits the utility of nanoESI-MS results at elevated

ligand concentrations.

100

Figure 5.2. Representative spectra from 2500 – 4800 m/z for BSA, BSA exposed to

PFOA (50 µM), and BSA exposed to PFNA (50 µM) in 9 mM ammonium acetate (pH

7). BSA concentrations are constant at 50 µM. Individual charge states are indicated for

the multiply charged BSA or PFAA-BSA complex.

ESI-MS may also be used to quantify protein-ligand association constants.

Association constants determined via ESI-MS are in good agreement with solution-

based methods for a variety of protein-ligand complexes with affinities ranging from

102 to 106 M-1 (150, 158, 160). However, limitations exist in extrapolating results from

gas-phase measurements to solution-phase behavior. In particular, obtaining high

resolution for proteins in the high-mass range is difficult, and nonspecific binding may

be observed at high ligand:protein mole ratios (158) as solution-phase equilibrium

changes during the evaporation and droplet fission processes (159). The detection of

solution phase specific and non-specific binding is desired, while binding that may

occur during transfer of molecules to the gas phase is not.

101

Figure 5.3. Mass spectra for the +16 charge state of 50 µM BSA in 9 mM ammonium

acetate (pH 7) with PFOA (left) and PFNA (right). The number of PFAA molecules

associated with BSA and the mole ratio of PFAA to BSA concentrations, [L]:[P], are

indicated.

Mathematically deconvoluted spectra of native and ligand-bound BSA were used to

quantitatively analyze PFAA-protein complexes. Because the ligand is small compared

to the protein (tested PFAAs are less than 1% of the molecular weight of BSA) the

surface properties of the ligand-bound and free protein are expected to be similar. As

such, the ionization and detection efficiencies of the free and ligand-bound proteins are

assumed to be similar, so the concentration ratio of complex to free protein, R, is

102

equivalent to the abundance ratio of bound and free protein in the gas phase (161).

Deconvoluted spectra were used to determine the abundance ratio and calculate Rj for

each bound complex (Equation 5.7). Rj was in turn used to calculate stepwise

association constants (Equation 5.12).

Applying Equation 5.12 to deconvoluted spectra for 0.5, 1, and 2 PFOA:BSA,

PFNA:BSA, and PFDA:BSA mole ratios yields association constants for three binding

sites reported in the Supporting Information (Tables 5.8S – 5.11S). For PFOS,

decreased resolution limited calculation of association constants at higher concentration

ratios; only the 0.5 PFOS:BSA results are displayed. Calculated PFOA-, PFNA-, and

PFDA-BSA association constants (Ka,1 – Ka,3) range from 0.9 × 104 M-1 to 5.1 × 105

M-1, 1.4 × 104 M-1 to 3.9 × 105 M-1, and 1.3 × 104 M-1 to 4.5 × 104 M-1, respectively. For

PFOA and PFNA, Ka,1, Ka,2, and Ka,3 are not statistically different from one another.

The relative similarity among the three Ka values suggests a single binding class.

Additionally, Ka,1 results for PFOA, PFNA, and PFDA collected at the 0.5 PFAA:BSA

mole ratio were not statistically different. Two tests were conducted with cone voltage

at 100 V and one test at 130 V. Data collected at a cone voltage of 130 V yielded

statistically lower Ka,1 and Ka,2 values for each PFCA-BSA as compared to data

collected at 100 V (one-tailed t-test, p < 0.1). A higher cone voltage was used to

improve peak resolution, but resulting in-source dissociation of bound ligands could

lead to calculation of artificially low association constants. In future studies, solution

additives may be used to stabilize complexes for prevention of in-source dissociation

(161).

Results collected at 100 V to limit in-source dissociation and low mole ratios to

limit nonspecific binding, which may occur as ligand concentrations increase, are most

appropriate in this study for comparison to solution-based results. The average Ka,1

values for data collected at 100 V and a 0.5 mole ratio are 1.3 × 105 M-1 for PFOA and

2.6 × 105 M-1 for PFNA. Although these results must be interpreted with caution given

the wide range of values, results are within an order of magnitude of association

constants calculated via equilibrium dialysis.

Binding regimes. Binding of PFOA and PFNA at concentrations spanning several

orders of magnitude and 1 µM BSA was fit well by a two binding class equation,

103

whereas a one-class equation was applicable for a narrower range of results at lower

PFAA concentrations. A companion study investigating the binding strength of PFCAs

with BSA and HSA using fluorescence spectroscopy and 19F NMR over a range of

PFCA concentrations (100 nM – 2 mM) suggests PFCA-BSA association constants of

~105 M-1 and 102 M-1 for primary and secondary binding sites, respectively (137).

Taken together with available literature data and results presented here (Table 5.3), the

data sets suggest two major binding regimes: strong specific associations at low

PFCA:albumin mole ratios and weaker nonspecific associations at higher mole ratios.

Specific interactions may be similar to those proposed for fatty acids. Albumin is

known to have a highly flexible conformation in which hydrophobic “pockets” hold the

hydrocarbon tails of fatty acids, and charged residues contribute an electrostatic

component to the binding (155). Although albumin has a net negative charge in

solution, it generally has a greater affinity for small, negatively charged hydrophobic

molecules (142).

PFOA-HSA interactions studied via zeta-potential measurements and ion selective

electrodes also demonstrate specific interactions at low ligand concentration, where

almost all PFOA molecules are bound to HSA (156). Initial binding to high affinity

sites, where Gibbs energies of interaction are at a minimum, stabilizes the HSA

structure (144). The interaction was more favorable at lower concentrations (~1 mM),

and increased at saturation (>10 mM PFOA with 20 µM HSA), where the hydrophobic

effect predominated (157). Using electrophoretic mobility, Blanco et al. showed binding

to high-energy sites for low PFOA concentrations with three proteins of varying size

and alpha-helix contents (162). This interaction was more favorable than that for

sodium caprylate, such that PFOA binding may be more energetically favorable than

the hydrogenated counterpart, which is less hydrophobic and less surface active than the

fluorinated surfactant.

104

Table 5.3. Summary of PFCA-albumin association constants (Ka) and binding

stoichiometries over a range of ligand:protein mole ratios ([L]:[P]).

Ligand Protein Method Ka (M-1) n [L]:[P] range Ref. PFOA BSA Dialysis 1.4 (± 1.9) × 106 1.4 ± 0.5 0.04 – 70 This

paper.1 BSA NanoESI-

MS 1.3 × 105 Up to 8 0.1 – 4 This

paper. HSA ζ-

potential2 2.4 × 104 - ~1.5 (157)

RSA Micro-SEC3

2.8 (± 0.6) × 103 7.8 ± 1.5 1 – 4 (42)

HSA Micro-SEC

2.6 (± 0.3) × 103 7.2 ± 1.3 1 – 4 (42)

HSA Dialysis4 3.12 × 104 13 ~4 – 30 (147) BSA 19F NMR 6.3 (± 0.8) × 102 - 72 – 200 (137) RSA 19F NMR 3.4 (± 1.2) × 103 - 8 – 400 (42) BSA Dialysis 3.2 × 102 85 ~3 - 110 (139) HSA ISE5 1.44 × 105 1414 ± 28 ~180-520 (156)6 HSA ISE 3.17 × 104 2565 ± 52 ~520-1100 (156)6

PFNA HSA Dialysis 2.1 ± 0.3 × 106 - 5 ± 1 × 10-4 This paper.7

BSA Dialysis 1.4 ± 0.4 × 106 - 6 ± 1 × 10-4 This paper.7

BSA Dialysis 3.3 (± 3.2) × 106 2.9 ± 0.7 0.02 – 120 This paper.1

BSA NanoESI-MS

2.6 × 105 Up to 8 0.1 – 4 This paper.

BSA 19F NMR 8 (± 5) × 103 - 15 – 100 (137) HSA 19F NMR 2 (± 3) × 104 - 8 – 32 (137)

1 Primary association constant listed 2 Electrophoretic mobility measured at pH 10; Ka calculated from ΔG = -25 kJ/mol 3 Micro size exclusion chromatography 4 Conducted at 37 °C 5 Ion-selective electrode 6 Hill binding constants reported for two binding regimes with positive cooperativity 7 Ka calculated for one PFAA:albumin mole ratio assuming n = 3

105

Results presented here demonstrate that PFNA is highly bound to BSA (>99%) at

low [L]:[P] mole ratios (< 10-3). Vanden Heuvel et al. (134) determined that 80% of 100

µM PFDA remained bound to 80 µM BSA after 60 minutes of extensive extraction with

organic solvents. This was attributed to covalent binding of the carboxylate head group

to protein sulfhydryl groups. Isolated albumins normally contain 0.5 – 0.7 moles of free

SH per mole of protein molecule (141). For comparison to prior results, at a 1.25 mole

ratio of free PFOA and PFNA to BSA (80 µM) and using equilibrium dialysis values

for a one-class binding model reported in Table 5.3, we calculate that greater than 98%

of PFAAs are bound to albumin.

5.4 Significance

A standard solution-based method (equilibrium dialysis) was compared with a

modern mass spectrometric approach (nanoESI-MS), providing complementary

information about the strength of PFAA-protein binding interactions and number of

binding sites at low ligand:protein mole ratios. Results presented, together with

previously published data, suggest stronger specific associations at low PFAA:albumin

mole ratios and weaker nonspecific associations at higher mole ratios. Equilibrium

dialysis yields primary association constants of ~106 M-1 for PFOA and PFNA, for a

class of one to five high affinity binding sites. A high protein-water partition coefficient

for PFNA (log KPW > 4) relative to neutral HOCs supports the characterization of

specific binding at low ligand concentrations.

NanoESI-MS is a useful technique for more rapidly characterizing PFAA-protein

interactions. However, a wide range of calculated association constants and sensitivity

of complexes to instrument conditions limit the utility of nanoESI-MS as a fully

quantitative method. Stoichiometry values obtained from mass spectrometry

demonstrate up to eight bound PFAAs per BSA molecule at a 4:1 mole ratio. Binding

constants from nanoESI-MS experiments are on the order of 105 M-1 for both PFOA and

PFNA, lower but in qualitative agreement with solution-based values determined via

equilibrium dialysis.

Because Kow may underestimate the bioaccumulative potential of PFAAs, a serum

protein association constant or protein-water distribution coefficient may be useful in

106

characterizing the bioaccumulative potential and in vivo bioavailability of long-chain

PFAAs. However, as proportions of various proteins vary among species and in time,

and likely also have different affinities for PFAAs, further analysis is required to test

the ability of protein partitioning to enhance perfluoroalkyl acid bioaccumulation

models.

Supporting Information. Contains (1) analytical and experimental details, (2)

results of fitting approaches, and (3) additional and summary results from nanoESI-MS.

Acknowledgment. This work was supported by the National Defense Science and

Engineering Graduate Fellowship, the Stanford University UPS Foundation and Woods

Institute for the Environment, and the National Science Foundation Graduate Research

Fellowship Program. We thank Pavel Aronov and Allis Chien from the Stanford

University Mass Spectrometry Laboratory.

107

5.5 Supporting Information

Figure 5.4S. Structures and names of perfluoroalkyl acids (PFAAs) used in this study.

Equilibrium dialysis results fitting approach. As previously described, a

nonlinear curve fit was applied to equilibrium dialysis results using the two-class

binding equation:

! =n1K

a,1[L]

1+Ka,1[L]

+n2K

a,2[L]

1+Ka,2[L]

(5.16S)

where

!

" is average number of bound ligands per protein molecule, L is the free ligand

concentration, Ka,1 and Ka,2 and are the association constants and n1 and n2 are the total

number of binding sites for each class of binding sites. For results presented in Tables

5.4S and 5.5S, association constants of 8000 M-1 for PFNA and 630 M-1 for PFOA were

inserted for Ka,2 in Equation 5.16S. Initial guess values for n2 were 30 and 100 for

PFNA and PFOA data, respectively. A range of initial guess values for n2 was tested for

the non-weighted PFNA and PFOA data fits in Tables 5.4S and 5.5S and did not

influence the fit in these cases. Parameters determined for Equation 5.16S without

insertion of values for Ka,2 showed greater error on determined binding stoichiometries

and association constants for n2 and Ka,2 and are presented in Table 5.6S. For these fits,

initial guess values were Ka,2 = 630 M-1 and n2 = 100 for PFOA and Ka,2 = 8000 M-1 and

n2 = 30 for PFNA. The results for PFAA bound to albumin are obtained by subtracting

the measured free (reservoir) PFAA concentrations from the corresponding total (final

bag) concentrations, displayed in Figure 5.8S. The error of the bound concentrations is

thus a propagation of error of both the free PFAA concentration and the total PFAA

concentration. The standard deviations of bound concentrations (Sbound) may be

calculated from the standard deviation of the reservoir triplicate samples for a single

108

point (Sfree) and the standard deviation of the final bag samples (Stotal) as:

Sbound = (Sfree )2+ (Stotal )

2 (5.17S)

These deviations were linearly correlated with the free and total PFAA

concentrations, such that at higher concentrations, the error of measured values

increases. This correlation is displayed in Figure 5.9S for measured free PFNA

concentrations in tests with 1 µM BSA. Consequently, a weighting factor may be

applied using the Kaleidagraph software when fitting the data to Equation 5.16S.

However, because standard deviations were determined only from triplicate samples at

each point, weighting bound concentrations in the fitted equation by the standard

deviation for that point was not performed. Alternatively, fits were tested with

weighting factors inversely proportional to the free, total, or bound measured PFAA

concentrations or the square of these values. Although fit parameters for PFOA (Table

5.4S) and PFNA (Table 5.5S) changed with different weighting factors applied, results

were consistently of the same order of magnitude for primary association constants and

number of primary binding sites. For these fits, initial guess values for n2 were 100 for

PFOA and 30 for PFNA, unless otherwise noted. In some cases for PFOA the fit did not

converge or yield physiologically relevant results for a fitted parameter, so the initial

guess was modified and data refit. Due to the large standard deviation and strong

influence on the fitted parameters of measurements at the two highest free PFOA

concentrations, these two data points were excluded from the non-weighted PFOA fits

presented in Tables 5.4S and 5.6S. Non-weighted fits for the full PFOA dataset did not

yield physiologically relevant results.

PFOA and PFNA data sets span several orders of magnitude, and as expected, are

not fit well by a model that represents only one binding class (including only Ka,1 and

n1). However, a subset of the experimental results (data for PFAA:albumin mole ratios

less than 1 or 5) was generally fit well by a one class model (Figure 5.10S), although

several outliers reduced the R2 for the fit of the non-weighted PFOA data. Fits utilizing

a one-class model yielded primary association constants similar to those obtained with a

two-class model applied to the full PFAA:albumin mole ratio range (Table 5.7S). Errors

in Tables 5.4S through 5.7S represent the standard error calculated by the Kaleidagraph

software for each parameter.

109

Statistical comparisons for nanoESI-MS results. Statistical comparisons of

results presented in Tables 5.8S – 5.11S were conducted using a student’s two-tailed t-

test assuming equal variance. PFOA and PFNA, Ka,1, Ka,2, and Ka,3 values are not

statistically different from one another (p > 0.1). For PFDA, pooled results for Ka,1 from

all exposure concentrations collected at 100 V are significantly greater than Ka,2 (p <

0.1) and Ka,3 (p < 0.05), indicating a somewhat stronger first binding site. However, all

PFDA-BSA measured affinities are on the order of 104 M-1, and Ka,2 and Ka,3 are not

statistically different. Ka,1 for PFDA was significantly less than Ka,1 for PFOS (p < 0.05)

although both values are on the order of 104 M-1. Pooled results for Ka,1 from data

collected at three concentration ratios and 100 V were compared between each PFAA.

Results for PFDA Ka,1 were significantly less than that for PFOA (p < 0.1) , PFNA (p <

0.05), and PFOS (p < 0.05). No other comparisons for Ka,1 determined at 100 V were

statistically significant. Further, Ka,1 results for PFOA, PFNA, and PFDA collected at

the 0.5 PFAA:BSA mole ratio and 100 V or 130 V were not statistically different. For

PFDA and PFOS, the averages of Ka,1 calculated at a PFAA:BSA mole ratio of 0.5 and

100 V are 3.5 × 104 M-1 and 7.4 × 104 M-1, respectively.

110

Figure 5.5S. Samples taken prior to equilibration in the reservoir from control bags

containing only buffer and the PFAA spike are compared to samples taken from test

bags containing 1 µM BSA with the same PFAA spike. Points fall along the 1:1 line

(plotted), indicating minimal effects from the BSA matrix in LC-MS/MS sample

analysis. The average relative standard deviation of initial bag samples from tests was

9% for PFOA and 11% for PFNA. Error bars represent 95% confidence intervals for

triplicate samples from the same dialysis bag.

111

Figure 5.6S. Reservoir samples taken over time in a PFNA equilibrium dialysis test

indicate equilibrium of the system after 24 hours. At equilibrium, control bag and

reservoir sample concentrations were also equivalent (data not shown). Error bars

represent 95% confidence intervals for triplicate samples from the same reservoir.

Figure 5.7S. Reservoir samples taken over time in a PFOA equilibrium dialysis test

suggest equilibrium of the system after approximately 48 hours. Samples were taken

after 48 hours for all PFOA tests. Error bars represent 95% confidence intervals for

triplicate samples from the same reservoir.

112

0.01

0.1

1

10

100

1000

0.01 0.1 1 10 100

Tota

l [P

FO

A] (!

M)

Free [PFOA] (!M)

0.1

1

10

100

1000

0.01 0.1 1 10 100

Tota

l [P

FN

A] (!

M)

Free [PFNA] (!M)

Figure 5.8S. Measured total and free PFOA and PFNA concentrations taken at

equilibrium from dialysis bag and reservoir samples, respectively. Results for bound

PFAA are obtained by subtracting the measured free PFAA concentrations from the

corresponding total concentrations.

113

Figure 5.9S. Standard deviations of triplicate measurements of bound PFAAs (Sbound)

were linearly correlated with free PFAA concentrations, as shown above for PFNA in

1µM BSA equilibrium dialysis tests. Consequently, a weighting factor may be

employed to account for larger error at higher measured concentrations.

114

Table 5.4S. Association constants (Ka,1) and binding stoichiometries (n1 and n2) for

PFOA binding to 1 µM BSA as determined by equilibrium dialysis for a range of

applied weighting factors. In these fits, 0.00063 µM-1 was inserted for Ka,2 in Equation

5.16S. Fits did not converge or yield physiologically relevant results for the non-

weighted or 1/

!

" -weighted full PFOA dataset.

Weighting Factor Ka,1 (µM-1) n1 n2 R2 Chisq

None 1.4 ± 1.9 1.4 ± 0.5 153 ± 18 0.913 14.25 1/[Free PFAA] 1.9 ± 2.2 1.3 ± 0.8 110 ± 81 0.869 1.74 1/[Total PFAA] 1.8 ± 3.3 1.3 ± 1.1 106 ± 87 0.834 1.41 1/[Free PFAA]2 2.4 ± 2.4 1.1 ± 1.0 160 ± 514 0.946 0.76 1/[Total PFAA]2 2.5 ± 4.8 1.1 ± 1.6 135 ± 622 0.919 0.22 1/

!

" 2 2.6 ± 2.0 1.1 ± 0.4 43 ± 32 0.765 4.69

Table 5.5S. Association constants (Ka,1) and binding stoichiometries (n1 and n2) for

PFNA binding to 1 µM BSA as determined by equilibrium dialysis for a range of

applied weighting factors. In these fits, 0.008 µM-1 was inserted for Ka,2 in Equation

5.16S.

Weighting Factor Ka,1 (µM-1) n1 n2 R2 Chisq

None 3.3 ± 3.2 2.9 ± 0.7 31.0 ± 2.2 0.953 41.82 1/[Free PFAA] 1.9 ± 1.0 3.5 ± 1.0 27.9 ± 10.0 0.938 4.77 1/[Total PFAA] 1.7 ± 1.8 3.6 ± 1.7 27.2 ± 11.9 0.955 1.47 1/

!

" 2.0 ± 1.7 3.3 ± 1.1 27.9 ± 5.3 0.959 4.63 1/[Free PFAA]2 2.4 ± 0.9 3.0 ± 1.1 39.8 ± 54.2 0.766 34.09 1/[Total PFAA]2 1.6 ± 2.4 3.6 ± 4.0 27.5 ± 85.5 0.918 0.55 1/

!

" 2 1.6 ± 1.5 3.5 ± 2.1 25.0 ± 16.3 0.946 1.11

115

Table 5.6S. Association constants (Ka,1 and Ka,2) and binding stoichiometries (n1 and

n2) for PFOA and PFNA and binding to 1 µM BSA as determined by equilibrium

dialysis using Equation 5.16S with no weighting factor (WF) and a 1/[Free PFAA]

weighting factor.

Com-pound WF Ka,1

(µM-1) n1 Ka,2 (µM-1) n2 R2 Chi-sq

PFOA None 3.0 ± 5.1 1.0 ± 0.5 6.1 ± 7.8 × 10-03 22 ± 20 0.917 13.7 PFOA 1/[Free

PFAA] 3.1 ± 6.8 0.9 ± 1.4 3 ± 11 × 10-02 7 ± 11 0.891 1.4

PFNA None 1.7 ± 1.1 3.9 ± 0.8 3 ± 23 × 10-04 400 ± 3100 0.968 28.6 PFNA 1/[Free

PFAA] 1.7 ± 0.9 3.8 ± 1.2 1 ± 14 × 10-03 100 ± 1200 0.940 4.6

116

Figure 5.10S. Equilibrium dialysis results for PFOA and PFNA up to a 5:1 ligand to

protein mole ratio and 1 µM BSA where ν is the average number of PFAA molecules

bound per albumin. Data represent average of triplicate measurements from each test

reservoir or dialysis bag. PFOA and PFNA data were fit using a one-class binding

model with no weighting factor (top) and with a 1/[Free PFAA] weighting factor

(bottom).

117

Table 5.7S. Association constants (Ka,1) and binding stoichiometries (n1) for PFOA and

PFNA binding to 1 µM BSA as determined by equilibrium dialysis for a subset of the

total data and a one-class binding model.

Compound Weighting Factor Ka,1 (µM-1) n1 R2 Chisq

[Ligand]: [Protein]

range PFOA None 1.1 ± 0.2 2.1 ± 0.2 0.988 0.01 0.04 – 1 PFOA None 0.20 ± 0.14 4.2 ± 1.9 0.767 1.8 0.04 – 5 PFOA 1/[Free

PFAA] 1.7 ± 1.7 1.5 ± 0.8 0.894 0.9 0.04 – 5

PFNA None 1.3 ± 0.8 4.6 ± 1.5 0.900 1.1 0.02 – 1 PFNA None 1.2 ± 0.2 4.6 ± 0.3 0.948 1.7 0.02 – 5 PFNA 1/[Free

PFAA] 1.6 ± 0.7 4.2 ± 1.0 0.930 3.8 0.04 – 5

118

Figure 5.11S. Representative mass spectra for the +16 charge state of 50 µM BSA in 9

mM ammonium acetate (pH 7) with PFDA (left, cone voltage = 100V) and PFOS (right,

cone voltage = 130 V). The m/z for free BSA and PFAA-BSA peaks and the mole ratio

of PFAA to BSA concentrations, [L]:[P], are indicated.

119

Figure 5.12S. Representative deconvoluted spectrum for 50 µM BSA in 9 mM

ammonium acetate (pH 7) with 100 µM PFOA (cone voltage = 100V) used for

determination of Ka. Peaks correspond to the number of bound PFOA molecules (j), as

indicated.

120

Table 5.8S. Estimated association constants calculated from nanoESI-MS results for 50

µM BSA exposed to PFOA (25, 50, and 100 µM).

[PFOA]: [BSA] 0.5 1 2

Cone Voltage Ka,1 (M-1) Ka,1 (M-1) Ka,2 (M-1) Ka,1 (M-1) Ka,2 (M-1) Ka,3 (M-1)

100 1.9 × 105 5.1 × 105 3.1 × 105 1.4 × 105 1.2 × 105 6.5 × 104 100 6.3 × 104 1.5 × 105 1.0 × 105 4.8 × 104 4.3 × 104 3.5 × 104 130 3.0 × 104 1.3 × 104 1.4 × 104 8.9 × 103 1.0 × 104 1.2 × 104

Table 5.9S. Estimated association constants calculated from nanoESI-MS results for 50

µM BSA exposed to PFNA (25, 50, and 100 µM).

[PFNA]: [BSA] 0.5 1 2

Cone Voltage Ka,1 (M-1) Ka,1 (M-1) Ka,2 (M-1) Ka,1 (M-1) Ka,2 (M-1) Ka,3 (M-1)

100 3.8 × 105 1.3 × 105 9.6 × 104 1.8 × 105 1.4 × 105 7.6 × 104 100 1.4 × 105 2.6 × 105 1.7 × 105 4.0 × 104 4.2 × 104 3.5 × 104 130 4.9 × 104 1.7 × 104 1.8 × 104 1.4 × 104 1.8 × 104 1.6 × 104

121

Table 5.10S. Estimated association constants calculated from nanoESI-MS results for

50 µM BSA exposed to PFDA (25, 50, and 100 µM).

[PFDA]: [BSA] 0.5 1 2

Cone Voltage Ka,1 (M-1) Ka,1 (M-1) Ka,2 (M-1) Ka,1 (M-1) Ka,2 (M-1) Ka,3 (M-1)

100 4.0 × 104 2.9 × 104 1.8 × 104 2.4 × 104 1.7 × 104 1.6 × 104 100 3.1 × 104 4.5 × 104 2.9 × 104 4.3 × 104 3.3 × 104 2.0 × 104 130 1.3 × 104 1.7 × 104 1.6 × 104 1.4 × 104 1.3 × 104 1.3 × 104

Table 5.11S. Estimated association constants calculated from nanoESI-MS results for

50 µM BSA exposed to PFOS (25 µM). Poor resolution limited quantitative analysis of

PFOS at 50 and 100 µM.

[PFOS]: [BSA] 0.5

Cone Voltage Ka,1 (M-1)

100 9.6 × 104 100 5.2 × 104 130 8.9 × 104

122

123

Chapter 6

Strong associations of short-chain

perfluoroalkyl acids with serum albumin

and investigation of binding mechanisms

6.1 Introduction

The unique chemical properties of perfluoroalkyl acids (PFAAs), a class of stable

anionic surfactants, have been capitalized upon since the 1940s in the production of a

variety of industrial and consumer products (20). In response to comprehensive research

documenting the environmental persistence and widespread occurrence of PFAAs in

humans and wildlife (25, 34), 3M Company voluntarily eliminated perfluorooctane

sulfonyl fluoride (POSF)-based materials including perfluorooctanoate (PFOA) and

perfluorooctane sulfonate (PFOS) from production. Eight companies subsequently

committed to eliminating emissions of PFOA and related compounds by 2015 as part of

a U.S. Environmental Production Agency PFOA Stewardship Program (35).

124

Internationally, however, perfluoroalkyl compounds of varying chain lengths (C4 to C15)

are still manufactured, and legacy products remain in use globally (20). The phase-out

of PFOS has been accompanied by a shift in production to shorter-chain length

compounds, including those based on C4-sulfonyl chemistries (35). The apparently

efficient clearance of perfluorobutane sulfonate (PFBS), observed in several organisms,

reduces this compound’s bioaccumulative potential (36, 37).

The structures of two homologue groups, perfluoroalkyl carboxylates (PFCAs) and

perfluoroalkyl sulfonates (PFSAs), resemble those of fatty acids and hydrocarbon-based

detergents (Figure 6.5S), but the high-energy carbon-fluorine bond imparts resistance to

hydrolysis, photolysis, microbial degradation, and metabolism by vertebrates and

renders the perfluoroalkyl tail both hydrophobic and oleophobic (16, 33). Rather than

partitioning to adipose tissue, PFAAs are detected predominantly in protein-rich

compartments such as the liver, kidney and blood (39-42), and PFOS concentrations

have been shown to positively correlate with tissue and fluid protein content (39).

Recent field monitoring data for C7 to C14 PFCAs demonstrate a curvilinear

relationship between protein-normalized trophic magnification factors in a marine

mammalian food web and protein-water distribution coefficients (KPW) estimated from

a relationship between octanol-water partition coefficients and affinities for bovine

serum albumin (BSA) (43). KPW may be incorporated into predictive models to improve

estimations of chemical distribution and bioaccumulation of organic contaminants (44).

A globular protein consisting of 583 amino acid residues, BSA (66430 Da) is widely

utilized as a model protein in biophysical, biochemical, and physicochemical studies, as

BSA binding sites accommodate a wide variety of endogenous and exogenous ligands.

Albumin is located to some degree in every fluid of the body, accounting for 60% of

total serum proteins at concentrations of 35 to 50 g/L and exhibiting extravascular

concentrations of 10 to 30 g/L in the skin, muscle, liver, gut and subcutaneous

compartment (151). Despite a proposal that PFAA biomagnification patterns may

correlate with binding to serum proteins, empirical data on the protein binding behavior

of PFAAs are limited, especially for short-chain PFAAs. Chen and Guo (146) measured

displacement of fluorescent probes on human serum albumin (HSA) to calculate

association constants for PFBS and the C4 PFCA on the order of 106 M-1. However,

125

other commonly utilized fluorescence methods may not be applicable to the study of

short-chain PFAAs with albumin (163, 164).

A wide-range of association constants (102 to 106 M-1) for C8 and C9 PFAAs with

rat (42), human (42, 137, 146, 147, 156, 157, 163), and bovine (45, 137, 164) albumins

are reported. Primary association constants for PFOA and perfluorononanoate (PFNA)

with BSA suggest binding through specific high affinity interactions. High log KPW

values (>4) were determined for PFOA and PFNA at physiological PFAA:albumin

mole ratios (45). Site-specific binding of PFOA and perfluorohexanoate (PFHxA) to

albumin has the potential to disrupt endogenous functions via displacement of the fatty

acid, oleate (165), which exhibits 107 to 108 M-1 association constants with albumin for

its first five binding sites (166). However, C4 to C10 PFAAs did not displace steroid

hormones from avian serum proteins at physiological concentrations (133). In addition

to active uptake of PFAAs by transporter systems (129), PFAA-albumin binding may

play a role in elimination kinetics (165). Binding of small molecules to albumin controls

their free concentrations in blood and influences the duration of action, as generally

only the unbound fraction can exit the vascular component. Interactions of PFAAs with

other biomolecular targets including liver fatty acid binding proteins may also play a

role in PFAA bioaccumulation (167) and affect the pharmacokinetics of fatty acids or

other endogenous ligands (136).

The objective of the present study was to examine the interactions of a series PFCAs

(C5 to C12) and even chain length PFSAs (C4 to C8) with BSA. The fraction of PFAAs

bound to BSA and protein-water distribution coefficients are determined at

physiologically relevant ligand:protein mole ratios using equilibrium dialysis and liquid

chromatography tandem mass spectrometry (LC-MS/MS), a direct approach that allows

evaluation of binding parameters at low ligand concentrations. Further analyses via

dialysis, nanoelectrospray ionization mass spectrometry (nanoESI-MS), and

fluorescence spectroscopy offer insights into the mechanisms of PFAA-albumin

interactions by evaluating various solution- and chemical structure-specific parameters

potentially affecting the binding of PFAAs to albumin. In particular, the influence of

fluorocarbon carbon chain length, ionic head group, and solution pH and ionic strength

are assessed.

126

6.2 Methods

Equilibrium dialysis. A purity-corrected equimolar stock solution of PFCAs (C5 to

C12, and C14) and PFSAs (C4, C6, and C8) and a separate concentrated solution of PFNA

were prepared in HPLC water and diluted in 50 mM sodium phosphate buffer in

polypropylene containers without the use of an organic co-solvent. BSA solutions were

prepared fresh daily in glass volumetric flasks using 50 mM sodium phosphate buffer at

the desired pH and concentrations measured using absorbance readings from a

NanoDrop-1000 spectrophotometer (NanoDrop Technologies, Rockland, DE, USA).

Polypropylene dialysis reservoirs containing the same buffers (250 to 500 mL) and

spiked with PFAAs were sampled in triplicate to determine initial (0 h) and final (120

h) PFAA concentrations. The buffer system was selected for consistency with previous

analyses (45, 137). Dialysis reservoir pH and temperature measurements were taken

with an Orion 5 Star Multimeter (Thermo Fisher Scientific, Waltham, MA, USA).

Dialysis bags containing 2 mL of BSA solution were secured with dialysis clips, added

to reservoirs, and equilibrated for 120 h at room temperature (21.2 ± 0.4 °C) prior to

sampling. Six reservoirs were prepared at pH 7.0 with a range of mixed PFAA

concentrations and one dialysis bag each ([BSA] = 199 ± 4 µM) and were sampled in

triplicate at equilibrium to determine KPW values. In these tests, individual free PFAA

concentrations at equilibrium ranged from 0.009 to 0.7 µM (Table 6.1), such that total

free PFAA concentrations ranged from 0.8 to 5 µM. Additional dialysis reservoirs

prepared at pH 6.1, 7.0, 8.0, and 8.9 with 2.8 ± 0.6 µM total PFAAs each contained

three dialysis bags ([BSA] = 197 ± 2 µM) that were sampled in duplicate for PFAAs.

Over the pH range studied, the PFAAs tested were in ionized form in solution (168);

any changes in protonation state upon binding to BSA are not detectable by this

method. Unless otherwise stated, error bars represent one standard deviation. Control

dialysis bags containing a buffer-only solution were equilibrated in a PFAA-spiked

reservoir to confirm equilibration and diffusion of PFAAs through the dialysis

membrane. Control results indicated that perfluorotetradecanoic acid (PFTA) does not

diffuse freely into the dialysis bag, so results for PFTA were excluded.

Details of post-dialysis sample preparation, PFAA analysis via LC-MS/MS, and

127

mass balance results are available in the Supporting Information. Samples and standards

were analyzed using a Shimadzu LC system (LC10ADvp pumps controlled by an

SCL10Avp controller, Columbia, MD, USA) coupled to a Sciex API 3000 triple

quadrupole mass spectrometer (MDS Sciex, Ontario, Canada) operating in negative

electrospray ionization multiple reaction monitoring (MRM) mode. Analyte separation

was achieved using a 40 mm x 2.1 mm Targa Sprite C18 column (5-µm particle size,

Higgins Analytical, Mountain View, CA, USA) with a C18 guard column (Higgins

Analytical).

Nanoelectrospray ionization mass spectrometry. Sample preparation, as well as

instrument parameters and conditions, were previously described (45). Briefly,

individual PFAAs prepared in 9 mM ammonium acetate (pH 7) at room temperature

were added to BSA (50 µM) in polypropylene microcentrifuge tubes, brought to a 1:1

or 2:1 PFAA:BSA mole ratio, and equilibrated for one or more hours before same-day

analysis. Samples were analyzed using an Advion Triversa Nanomate nano-electrospray

robot (Advion BioSystems, Ithaca, NY, USA) coupled with a Waters Micromass Q-Tof

API-US quadrupole time-of-flight mass spectrometer (Micromass, Milford, MA, USA).

Instrument gas pressure and voltage settings were selected to maximize peak intensities

while maintaining resolution and ion current. For static collision energy tests, mass

spectra were acquired for 4 min with cone voltage set at 100 V and laboratory scale

collision energies set to 10 eV for singly charged ions. In duplicate collision induced

dissociation (CID) tests for each tested PFAA-BSA complex, the collision energy was

ramped from 10 to 90 eV, with spectra acquired for 2 min at each setting. Additional

details are available in the Supporting Information.

Fluorescence spectroscopy. PFAA stock solutions were prepared in 50-mL

polypropylene centrifuge tubes (Corning, NY, USA) in HPLC grade water and were

sonicated at 35 °C to dissolve. Solutions of BSA were prepared in glass volumetric

flasks in 50 mM sodium phosphate buffer (pH 6.1, 7.0, 8.1, or 9.1) immediately prior to

each experiment. A stock solution of sodium chloride (3.31 M) was prepared in HPLC

grade water as needed. Fluorescence titrations were conducted for PFNA or PFOS with

BSA in solutions buffered at pH 6.1, 7.0, 8.1, or 9.1 without added sodium chloride and

at pH 7.0 with added sodium chloride to achieve an ionic strength of 0.21, 0.30, or 0.41

128

M. Test solutions (14) were prepared in 15-mL polypropylene centrifuge tubes and

allowed to equilibrate overnight at 4 °C. The concentration of BSA was constant at 4

µM, a concentration that was selected, along with instrumental parameters, to maximize

both sensitivity and resolution. One sample was prepared with no added PFAA, while

the other 13 contained a range of PFAA:BSA mole ratios (0.1:1 through 60:1).

Emission scans were collected in a 1-cm quartz cuvette (Starna Cells; Atascadero, CA,

USA) using a PTI QuantaMaster spectrofluorometer (Photon Technology International;

Birmingham, NJ, USA). The excitation wavelength was 295 nm, and the emission was

recorded from 305 to 450 nm with a step size of 1 nm and an integration time of 0.1 s.

All monochromator slit widths were 4 nm.

6.3 Results and Discussion

Effect of chain length on binding of PFAAs to BSA. Dialysis experiments allow

direct measurement of PFAA-BSA binding through partitioning of PFAAs between an

external reservoir and BSA-containing dialysis bags. Measured concentrations of C2 to

C12 PFCAs and C4 to C8 PFSAs in the dialysis bag relative to the external reservoir

indicate free movement of PFAAs through the dialysis membrane and retention of

PFAA-BSA bound complexes in the dialysis bag. At a BSA concentration of 200 µM

(13 g/L) and a range of PFAA mixture concentrations, PFCAs and PFSAs with four to

ten fluorinated carbons were highly bound (>95%) to BSA across the range of free

PFAA concentrations tested (Table 6.1). This was consistent even for the short-chain

compounds PFBS, PFPeA, and PFHxA, for which reduced affinity for BSA was

expected. The lower fraction bound for PFDoA relative to other compounds in the

mixed reservoir could reflect a decrease in affinity at this longer chain length; however

the relatively large error associated with these results and the mass balance calculations

limits full interpretation. In a separate reservoir spiked only with PFNA, 98.2% of

PFNA was bound to BSA at equilibrium, a result in close agreement with that obtained

in the mixed PFAA experiments.

129

Table 6.1. Fraction of perfluoroalkyl acids (PFAAs) bound to 200 µM bovine serum

albumin (BSA) and log protein-water distribution coefficients for PFAAs with BSA

measured over a range of equilibrium free PFAA reservoir concentrations. For the

fraction of PFAA bound to BSA in each tested reservoir, errors represent 95%

confidence values; standard errors for log KPW are from regressions performed using

Kaleidagraph software.

Analyte Reservoir

[PFAA] Range (µM)1

Log KPW Fraction Bound

perfluorobutanesulfonate (PFBS)

0.010 to 0.57 3.86 ± 0.07 99.0 ± 0.5%

perfluorohexanesulfonate (PFHxS)

0.012 to 0.27 4.3 ± 0.1 99.2 ± 0.5%

perfluorooctanesulfonate (PFOS)

0.013 to 0.26 4.1 ± 0.1 99.1 ± 0.4%

perfluoropentanoate (PFPeA)

0.024 to 0.56 3.40 ±0.02 96.6 ± 0.8%

perfluorohexanoate (PFHxA)

0.009 to 0.43 4.05 ± 0.02 99.2 ± 0.3%

perfluoroheptanoate (PFHpA)

0.054 to 0.30 4.23 ± 0.08 99.3 ± 0.1%

perfluorooctanoate (PFOA)

0.023 to 0.32 4.14 ± 0.04 99.1 ± 0.2%

perfluorononanoate (PFNA)

0.064 to 0.46 4.05 ± 0.08 98.9 ± 0.3%

perfluorodecanoate (PFDA)

0.016 to 0.47 3.86 ± 0.08 98 ± 1%

perfluoroundecanoate (PFUnA)

0.026 to 0.38 3.7 ± 0.2 95 ± 3%

perfluorododecanoate (PFDoA)

0.031 to 0.68 3.3 ± 0.1 80 ± 10%

Protein-water distribution coefficients were determined for each PFAA by a linear

regression of the ratio of analyte concentration in the bound phase (Cp) to that in the

aqueous phase (Cw) (Figure 6.1). This ratio may be represented by:

KPW=CP

CW

=fbound

ralbumin

![P](1" fbound

) (6.1)

1 n = 6; n = 5 for PFHpA

130

where [P] is the protein concentration (g/mL), fbound is the fraction bound to protein, and

ρalbumin is the partial specific volume of protein in aqueous solution (0.733 mL/g, (141)).

KPW thus has units of [g bound PFAA / mL BSA]/[g free PFAA / mL water] and

represents the distribution of the combined anionic and neutral forms of PFAAs

between BSA and the aqueous buffer. At the low PFAA concentrations tested, the

binding isotherm for total PFAAs, represented by the sum of the concentrations of

tested PFAAs, is linear (Figure 6.6S). Values of log KPW for individual PFAAs range

from 3.3 to 4.3 (Table 6.1). The highest Cw value for PFBS, PFPeA, and PFHxA (Cw >

0.4 µM) was excluded in the KPW regressions because these binding isotherms were

nonlinear at higher concentrations. All data were included in nonlinear regressions

performed to determine association constants for the short-chain PFAAs. Data obtained

at relatively higher PFAA aqueous concentrations here and in previous work (45)

demonstrate an expected nonlinear binding relationship, and illustrate that application

of KPW should be limited to a narrow concentration range where binding may be

approximated as linear. At the PFAA:BSA ratios tested, log KPW results were

consistently greater than or at the high end of the range of BSA-water distribution

coefficients compiled for traditional hydrophobic organic contaminants (log KPW = 0.09

to 3.5) (44).

An increase in KPW with increasing chain length was observed for PFCAs with 4-6

fluorinated carbons. For PFCAs with greater than 6 fluorinated carbons, KPW values

generally decreased, though the differences between 6 and 7, 7 and 8, and 9 and 10

fluorinated carbons were not significant. Kelly et al. (43) estimated KPW values for

PFCAs with BSA based on a generalized BSA-ligand relationship with octanol-water

partition coefficients (KOW). Reported log KPW values increased linearly from 2.0 for

PFHpA to 5.0 for PFDoA; however, details of the KPW calculation were not provided.

The trend predicts an increase in affinity for BSA with increased hydrophobicity of the

perfluoroalkyl tail but does not account for increased steric hindrance associated with

longer fluorocarbon tails. Increased rigidity of the fluorocarbon tail and changes in

molecular geometry that begin at C8 and become more pronounced for PFCAs with ten

carbons and longer may influence PFAA partitioning and binding behavior (165, 169).

Hydrophobic binding cavities in BSA may have limited ability to accommodate these

131

larger, less flexible ligands.

0

1x104

2x104

3x104

2 4 6 8 10 12

PFSAs

PFCAs

KP

W

Number of fluorinated carbons

Figure 6.1. Measured BSA-water distribution coefficients (KPW) for perfluoroalkyl

sulfonates (PFSAs, ) and perfluoroalkyl carboxylates (PFCAs, ) with fluorocarbon

tail lengths of 4 to 11. Error bars are standard errors for regressions performed in

Kaleidagraph software.

Direct binding observed by nanoESI-MS. NanoESI-MS, employing a soft

ionization technique, is used here to confirm direct binding of PFCAs (C2 – C9) and

PFSAs (C4 – C8) to BSA and determine stoichiometries of binding at 1:1

PFAA:albumin mole ratios. ESI-MS has been widely used to study proteins in their

native conformation and non-covalent interactions of protein-ligand complexes

preserved in the gas phase (170), although to our knowledge this is the first reported

detection of short-chain PFCA and PFSA complexes with BSA by nanoESI-MS.

Several studies have analyzed intact PFOA- and PFNA-protein complexes via ESI-MS

to elucidate the stoichiometry of binding (42, 45, 133, 135, 167). Although caution must

be taken when quantifying gas-phase affinities using this method (45) for comparison to

solution-based results, the technique has been increasingly used to determine

stoichiometries of specific binding for protein-ligand complexes.

132

Figure 6.2. Deconvoluted spectra of 50 µM BSA alone or with 50 µM PFPeA, PFHxA,

PFHpA, PFOA, or PFNA. The location of free BSA (P) and successive bound PFNA

ligands (P+L), (P+2L), and (P+3L) are denoted on the uppermost spectrum. The peak of

each first bound PFAA used to confirm the expected incremental mass shift (ΔM) from

the free BSA peak is indicated by a star (*).

Representative deconvoluted mass spectra for 1:1 mole ratio mixtures of BSA and

several PFAAs are displayed in Figure 6.2 and Figure 6.7S. Deconvoluted spectra

exhibit distinct peaks for at least one bound PFAA per BSA. Despite somewhat broad

peaks, the mass increment of noncovalent PFAA-BSA complex peaks (P + jL) from the

initial protein peak (P) generally corresponds well with the theoretical molecular mass

of the ligands (ΔM, Table 6.4S). Protein-bound complexes were also maintained for C2-

C4 PFCAs in the gas phase, with an observable change in the native BSA spectrum even

with the addition of the C2 PFCA. Peaks corresponding to two bound PFPrA and PFBA

(P + 2L) were visible as shoulders at the expected BSA-2(PFAA) complex mass on the

133

broader peaks. Adduct ions in the protein spectra, which may be partially attributed to

contamination of the protein with low molecular weight cations (e.g., sodium) via

contact with laboratory glassware, may mask the binding of small or weakly bound

ligands and also limit evaluation of binding at low ligand:albumin mole ratios. For

higher molecular weight PFAAs, the pattern of native BSA was preserved with the

addition of successive PFAA ligands leading to a repetition of the native BSA peak

shape at intervals corresponding to the mass of each bound PFAA. Evidence of multiple

PFAAs bound per BSA molecule was distinctly visible for PFNA and the C4-C8 PFSAs

(2-3 PFAAs bound at a 1:1 mole ratio). A rough calculation previously indicated that

one PFOS binds per albumin molecule (133). However, we present clear evidence of

multiple associations of PFOS with a given albumin molecule at a 1:1 PFOS+BSA mole

ratio. Binding sites detected in this manner likely represent both strong and weak

associations at the relatively high ligand:albumin mole ratio tested. Lower numbers of

C4 and shorter PFCAs bound at equivalent mole ratios may indicate that secondary

binding sites or nonspecific interactions for these compounds, experienced at higher

ligand concentrations, are weaker than comparable interactions of longer-chain PFAAs

with BSA.

Because the precise binding location for various chain-length PFAAs on albumin

has not been established, the full albumin molecule was used in this study to establish

specific binding of the full range of PFCAs and PFSAs to BSA, especially for short-

chain PFAAs, without selecting for particular domains on the protein. In three

homologous domains, albumin contains seven fatty acid (FA) binding sites including

four low affinity sites (Sites 1, 3, 6, and 7) and three high affinity sites (Sites 2, 4, and

5) that are likely candidates for PFAA binding (146). Existing evidence suggests that

PFHxA binds to Sudlow’s drug binding site II, which overlaps with FA Sites 3 and 4,

whereas PFOA preferentially binds to Sudlow’s drug binding site I at FA Site 7 (165).

Use of truncated albumin (e.g., recombinant HSA domain II (171)) in future work could

improve nanoESI-MS mass resolution and, if coupled with newly developed

competitive binding nanoESI-MS techniques (171), may provide further information on

site-specific associations.

Effect of ionic head group on PFAA-BSA binding. The protein-water distribution

134

coefficient determined for PFBS was significantly greater than that for the equivalent

fluoroalkyl chain length carboxylate, PFPeA. This corresponded to an association

constant for PFBS (Ka = 7.0 ± 3.6 ×106 M-1) that was more than three times greater than

that for PFPeA (Ka = 2.0 ± 0.8 ×106 M-1). Association constants (ratio of bound to free

reactants) were calculated using a one-class binding model that relates the ratio of the

bound PFAA concentration to the total protein concentration,

!

" , to the association

constant, Ka, the free ligand concentration, and the total number of binding sites on the

protein (45). At the concentrations tested, PFBS and PFPeA exhibited non-linear

isotherms, where a greater fraction of PFBS was bound to BSA relative to PFPeA over

the full range tested in dialysis experiments (Figure 6.3).

NanoESI-MS was exploited to further evaluate the effect of the ionic head group on

PFAA-BSA binding. The stability of two pairs of compounds of equal fluorinated alkyl

chain length (4 and 8) and a sulfonate or carboxylate moiety was observed by collision-

induced dissociation (CID) at 2:1 PFAA:BSA mole ratios. Collision-induced

dissociation provides a relatively rapid way to probe the stability of binding in the gas

phase. Studies have suggested a link between the energy required for dissociation of

gas-phase complexes and solution-phase binding constants, although the relative

stability of protein-ligand complexes in the gas phase and in solution influences the

interpretation of such measurements (172). For example, although binding in a

hydrophobic pocket may minimize the effects of water on protein-ligand binding in

solution (173), hydrophobic interactions may not persist in the gas phase, and

desolvation of the protein may have unknown influences on ligand interactions (172).

135

Figure 6.3. Effect of ionic head group on binding of equivalent chain length PFAAs to

BSA. In equilibrium dialysis tests (a), PFBS () exhibits higher affinity for BSA than

PFPeA () across the full concentration range tested. Data are displayed as the ratio of

the bound PFAA concentration to the total protein concentration (

!

" ) versus the free

PFAA concentration and were fit with a one-class binding model to determine Ka (R2 =

0.906 and 0.973 for PFBS and PFPeA, respectively). The measured fraction of PFAAs

bound to BSA in nanoESI spectra (b) decreased with increasing collision energy for

PFCAs ( and ) at 40 to 60 eV, whereas PFSA-BSA ( and ) complexes did not

dissociate over the range of collision energies.

A markedly different result was observed for PFBS and PFOS compared to PFPeA

and PFNA: the PFSAs did not dissociate from BSA over the range of tested collision

energies (CE), whereas PFCAs of identical chain lengths dissociated from BSA

between 40 to 60 eV (Figure 6.3). Representative deconvoluted mass spectra for PFPeA

and PFNA at the 2:1 PFAA:BSA mole ratio and spectra demonstrating dissociation of

PFNA with increasing CE are displayed in Figure 6.8S. The larger size of the sulfonate

moiety relative to the carboxylate may play a role in explaining the different behavior of

the PFSAs and PFCAs. PFAAs may interact with positive residues on BSA, including

lysine (pKa 10.53) and arginine (pKa 12.48) (174), and electrostatic interaction between

two oppositely charged molecules may be enhanced in the gas phase and become

relatively difficult to disrupt when complexes are desolvated (172). Additionally, BSA

and BSA-PFAA complexes were monitored in positive ionization mode, so the

136

dissociation of PFAAs in an anionic form is not expected. Although both PFCAs and

PFSAs are expected to be charged at physiological pH in solution, the carboxylate

anions may be more amenable to the capture of a proton as compared to the sulfonate

prior to dissociation from BSA in the ESI source and collision cell. Solution-phase

methods may be more appropriate to provide insight into the potential contribution of

electrostatic interactions to PFAA-BSA binding.

Effect of pH on PFAA-BSA binding. The contribution of electrostatic interactions

to protein-ligand binding is often assessed by changing the pH of the solution (175).

However, the analysis and interpretation of the influence of pH on ligand interactions

with BSA is complicated by the fact that BSA undergoes conformational changes with

changes in pH, potentially altering binding site characteristics. Conformational changes

can occur through the folding and unfolding of the tertiary and secondary structure of

the protein. In the case of BSA, successive multi-step, reversible transformations occur

with increasing pH, from fast (F) to normal (N) near pH 4, and to basic (B) and aged

(A) forms near pH 8 and 10, respectively. At near neutral pH (pH 7.4), albumin has a

net charge of -17 and is heart-shaped in the N form (151, 176). Over the pH range tested

(pH 6 to 9), BSA begins in the N form at lower pH and transitions to the B form. An

analysis of the native fluorescence of BSA reveals a decrease in fluorescence intensity

with increasing pH and a slight blue shift at pH 9 (Figure 6.9S). These spectroscopic

changes confirm the conformational changes in BSA as a function of pH.

137

Figure 6.4. Effect of pH on

!

" , the concentration of PFAA bound to BSA normalized to

the total protein concentration. Binding of PFBS (), PFPeA (), and PFHxA () to

BSA (a) decreased with increasing pH while binding of PFHxS (), PFOS (), PFUnA

(), and PFDoA () (b) increased with pH.

For all tested PFAAs, equilibrium binding to BSA was high across all tested pH

levels, ranging from 85% bound for PFPeA at pH 9 to greater than 99% for several

compounds and pH conditions (Table 6.5S). Similar to KPW results described earlier, at

pH 7 the average number of PFCA molecules bound per BSA molecule (

!

" ) increases

for the C4 to C6 compounds and decreases above C8 (Figure 6.10S). There was no

increase observed in bound PFSAs with increasing chain length at pH 7. A plot of the

138

dialysis bag concentration vs. reservoir pH reveals an overall increase in binding of

PFBS, PFPeA, and PFHxA to BSA with pH, whereas a negative relationship was

obtained for several longer chained PFAAs: PFHxS, PFOS, PFUnA, and PFDoA

(Figure 6.4 and Table 6.6S). Little to no overall trend in binding with pH was observed

for PFHpA, PFOA, PFNA, and PFDA (Figure 6.11S). The opposite trend in

!

" observed

for short- and long-chain PFAAs with pH may result from different conformational

changes experienced from binding at different locations on the protein.

Fluorescence analysis was used to further probe the binding of PFNA and PFOS –

two PFAAs with equivalent chain length but different ionic head groups – to BSA. The

binding of PFAAs to serum albumin evokes changes in the protein’s native fluorescence

(Figure 6.12S); these spectroscopic changes may be used to estimate binding constants

and stoichiometries, as recently described (163). For both PFAAs, the estimated binding

constant, KHill, determined from a plot of the degree of saturation (Y) vs. total PFAA

concentration (Figure 6.13S), increases from pH 6 to pH 7, but no further change is

observed at higher pH (Figure 6.14S). This trend is similar to that obtained for these

PFAAs by equilibrium dialysis (Figures 6.4 and 6.11S); however, as the fluorescence

data are obtained at higher PFAA:BSA mole ratios, these data are not directly

comparable. Given the changes in native fluorescence of BSA as a function of pH, the

differences in binding affinity observed here are likely due to conformational changes in

the protein. However, electrostatic changes could also play a role.

One would expect enhanced electrostatic interactions with decreasing pH as

repulsion from negatively charged amino acid residues is reduced; however, BSA

supports a more compact conformation as it approaches its isoelectric point (~pH 4.8),

which may reduce accessibility of surfactants to the hydrophobic cavities of the binding

sites (148). Gelamo et al. (148) observed a lower binding constant for sodium dodecyl

sulfate (SDS) with BSA at pH 5 as compared to pH 7 and pH 9. The compact structure

may have a more pronounced effect on binding of the longer, more hydrophobic

PFAAs, while electrostatic interactions may play a larger role in binding for short-chain

PFAAs. The difference in responses of PFAA-BSA affinity with pH suggests that short-

and long-chain PFAAs may bind at different locations on albumin.

To specifically investigate the role of electrostatics, the binding of PFNA and PFOS

139

to BSA as a function of ionic strength was explored through fluorescence titrations by

adding varying concentrations of sodium chloride to the titration solutions. The results

are shown in Figures 6.15S and 6.16S. For both PFAAs, there was no difference in

estimated binding affinity over a physiologically relevant ionic strength range (Figure

6.17S). This result supports the hypothesis that the observed changes in PFAA-BSA

binding with changing pH result primarily from conformational changes in the protein

and not changes in electrostatics.

6.4 Significance

In recent attempts to include protein associations in bioaccumulation models, which

traditionally approximate organism or tissue sorptive capacity using only lipid content,

researchers utilized BSA-ligand relationships to report log KPW values for a range of

hydrophobic organic contaminants (44). Empirical data for log KPW of a series of PFAA

cogeners are reported in the present study, yielding relatively large values from 3.3 to

4.3. Considered analogous to octanol-water partition coefficients, such parameters may

be useful to more accurately predict chemical accumulation and distribution. However,

use of KPW to describe binding between macromolecules and small molecule ligands

should be conducted with caution and limited to a narrow concentration range over

which binding may be approximated as linear.

Results from the present study indicate increased binding with chain length for C4 to

C6 PFCAs. Above C8, KPW decreased with increasing chain length. Additionally,

differences in binding of PFBS and PFOS relative to PFPeA and PFNA, respectively,

observed via dialysis and collision induced dissociation indicate an electrostatic

component to interactions with BSA. Fluorescence results for PFOS and PFNA suggest

that these affects are minimal in solution. A number of studies have illustrated an

important role of the hydrophobic driving force on PFAA environmental partitioning.

However, increased rigidity associated with long-chain PFCAs may contribute to the

observed nonlinear relationship of KPW with the fluorocarbon tail length.

Prior studies suggest that the binding of short-chain PFAAs may be different in

nature than that of long-chain PFAAs (133, 165). PFOA and PFHxA may bind

primarily to different sites on HSA, demonstrating that an increase in the fluorinated

140

carbon tail length by only two units may have a substantial effect on the nature of

PFCA-albumin interactions (165). Fluorescence studies indicate that long-chain PFAAs

have a greater influence on albumin conformation than short-chain PFAAs (e.g., PFBS,

PFPeA, and PFHxA) (163, 164), and the Trp binding site on HSA, which overlaps with

or exists in the hydrophobic cavity of Fatty Acid Site 7 in Subdomain IIA, may have a

preference for long-chain PFAAs (146). We find strong binding of short-chain PFAAs

to BSA, suggesting that reduced hydrophobicity and steric hindrances of short-chain

PFCAs and PFBS, which may limit observable conformational changes in albumin by

fluorescence methods, do not correspond to low affinity for albumin. In the present

study, pH-induced changes in binding affinity observed via equilibrium dialysis support

evidence that short- and long-chain PFAAs bind at different locations on BSA.

Fluorescence titrations suggest that the observed pH dependence of binding is due to

conformational changes in the protein.

An effort to reduce the bioaccumulation of PFAAs in humans and wildlife has led to

shifts in fluorochemical production to shorter chain-length compounds. Association

constants obtained for PFBS and PFPeA are useful to compare amongst various PFAAs

and with other exogenous and endogenous ligands in blood. Results in the present study

indicate that short-chain PFAAs bind strongly to BSA at low PFAA:albumin mole

ratios, suggesting that physiological implications of strong binding to albumin may be

important for short-chain PFAAs. However, these results contrast with limited evidence

that short-chain PFAAs are less bioaccumulative than long-chain PFAAs (36, 37),

highlighting a need for additional protein association measures for bioaccumulation

modeling beyond a single-protein KPW. Further research is needed to investigate the

binding of short-chain perfluorinated compounds to a multitude of potential protein and

transporter targets and to further understand the influence of such interactions on

biological uptake, retention, and functioning.

Supporting Information. Contains (1) additional analytical and experimental

details, (2) Figures 6.5S – 6.17S and (3) Tables 6.2S – 6.6S.

Acknowledgment. The present study was supported by the National Science

141

Foundation Graduate Research Fellowship Program, the Stanford University Woods

Institute for the Environment, and Union College. We thank Pavel Aronov from the

Stanford University Mass Spectrometry Laboratory for nanoESI-MS instrument

operation and technical input.

Publication Information. Reproduced with minor modifications from Bischel, H.

N.; MacManus-Spencer, L. A.; Zhang, C.; Luthy, R. G. Strong associations of short-

chain perfluoroalkyl acids with serum albumin and investigation of binding

mechanisms. Environmental Toxicology & Chemistry, Copyright © 2011 Society of

Toxicology and Chemistry, Wiley-Blackwell Publisher.

142

6.5 Supporting Information

Materials. Fraction V fatty acid-free bovine serum albumin (BSA, 99.9%) was

from EMD Biosciences, Inc. Perfluoropropanoic acid (PFPrA, 97%), perfluorobutanoic

acid (PFBA, 97%), perfluoropentanoic acid (PFPeA, 97%), perfluorononanoic acid

(PFNA, 97%), potassium perfluorobutane sulfonate (PFBS, 98%), and potassium

perfluorohexane sulfonate (PFHxS, >98%) were from Sigma-Aldrich (St. Louis, MO).

Perfluoroheptanoic acid (PFHpA, 99%), perfluorooctanoic acid (PFOA, 96%),

perfluorodecanoic acid (PFDA, 98%), perfluoroundecanoic acid (PFUnA, 95%),

perfluorododecanoic acid (PFDoA, 95%), and perfluorotetradecanoic acid (PFTA, 97%)

were from Aldrich Chemical Co. (Milwaukee, WI, USA). Triflouroacetic acid (TFA,

>99.5%), perfluorohexanoic acid (PFHxA, >97%), and potassium perfluorooctane

sulfonate (PFOS, 98%) were from Fluka through Sigma-Aldrich (St. Louis, MO, USA).

Mass labeled internal standards [13C5] PFNA, [13C2] PFDA, [13C2] PFOS, N-

deuterioethylperfluoro-1-octanesulfonamidoacetic acid ([D5]–N-EtFOSAA) were from

Wellington Laboratories (Guelph, ON, Canada), and [13C2] PFOA was from Perkin-

Elmer Life Sciences (Boston, MA, USA). Labeled internal standards had purities

greater than 98%, as reported by the suppliers. Structures of perfluoroalkyl acids

(PFAAs) included in this study are displayed in Figure 6.5S. Spectra/Pore dialysis

membrane tubing (6000 to 8000 Da molecular weight cutoff), polypropylene dialysis

reservoirs and clips were from Spectrum Laboratories (Rancho Domingo, CA, USA).

Data were analyzed in Microsoft Office Excel (Microsoft Corporation; Redmond, WA,

USA) and Kaleidagraph (Synergy Software Systems; Dubai, United Arab Emirates).

Post-dialysis sample preparation and liquid chromatography tandem mass

spectrometry (LC-MS/MS) analysis. Triplicate reservoir samples (0.5 mL) were

added to an equal volume of methanol in HPLC vials or polypropylene microcentrifuge

tubes for further dilution with 1:1 methanol:buffer. Reservoir samples were analyzed by

LC-MS/MS following the addition of 7:3 v/v methanol:1% aqueous NH4OH (100 µL)

and a mixed internal standard solution prepared in HPLC grade water (100 µL).

Dialysis bag samples (0.5 mL or 0.2 mL) taken at equilibrium into 15-mL

polypropylene centrifuge tubes were extracted using acetonitrile (1% v/v glacial acetic

143

acid) as previously described (45), using 9.5 mL acidified acetonitrile for 0.5 mL

samples and 3.8 mL for 0.2 mL samples. Samples were vortexed (30 s), sonicated (10

min, 60 °C), and centrifuged (15 min, 3000 rpm). An aliquot from each extraction (1.8

mL) was transferred to a polypropylene microcentrifuge tube containing ENVICarb

(Supelco, Bellefonte, PA, USA), vortexed, and centrifuged (30 min, 14000 rcf).

Additional dilutions were performed as needed to bring the expected sample

concentration into the range of the LC-MS/MS calibration standards. High pressure

liquid chromatography (HPLC) vials contained the acetonitrile extract (200 µL), HPLC

grade water (200 µL), 7:3 v/v methanol:1% aqueous NH4OH (50 µL), and internal

standard in HPLC grade water (45 µL). Reservoir samples and dialysis bag extracts

were stored at 4°C until analysis. Reservoir and BSA matrix spike recovery results are

shown in Table 6.2S. Matrix-matched calibration standards were prepared with PFAA

stock solutions in 7:3 v/v methanol:1% aqueous NH4OH. Further details of LC-MS/MS

analysis are available elsewhere (45). Mass transitions monitored for quantitation and

confirmation are shown in Table 6.2S.

The average mass balance of PFAAs in pH 7.0 dialysis reservoirs (n = 6) was

between 80% and 130% for all compounds (Table 6.2S). The mass balance on PFTA

includes only initial and final reservoir samples (n = 7), as PFTA did not diffuse into

dialysis bags. A decrease in measured BSA concentrations was observed between initial

and final time points, but no BSA was detected in external reservoirs; the decrease may

have been due to osmotic dilution or sorption to the apparatus. Therefore, BSA

concentrations measured at equilibrium were used in calculations. BSA was

equilibrated separately in blank reservoirs. Occasional low levels of several PFAAs

were detected in either blank reservoir or blank dialysis bag measurements. The lowest

reported test dialysis PFAA concentrations were always more than twice that detected

in a blank. For reservoir matrix spike recoveries, a known standard was spiked into 50

mM sodium phosphate buffer at each pH (n = 6) and compared to the standard, which

was prepared in pH 7 buffer. For BSA spike recoveries, freshly prepared BSA (200 µL,

pH 7) was spiked with a mixed PFAA standard (n = 12) and extracted using the

previously described procedure. The average reservoir and BSA matrix spike recovery

results were between 87 to 123% and 95 to 101%, respectively (Table 6.2S). All PFAAs

144

were below the detection limit for side-by-side extractions of blank BSA and sodium

phosphate buffer solutions.

Nanoelectrospray ionization mass spectrometry analysis. BSA prepared in 9 mM

ammonium acetate (pH 7) at room temperature was dialyzed overnight to aid in

removal of salts prior to PFAA exposure. Mass signals were collected over the scan

range m/z 1000 to 5000. The three most intense charge states (+15 to +17) occurred

from m/z 3800 to 4700 for free BSA and BSA-PFAA complexes. Multiply-charged

mass spectra were mathematically deconvoluted using m/z 3800 to 4700 and MaxEnt in

MassLynx software version 4.1 from Waters. Spectra were deconvoluted with 10

iterations using MaxEnt and smoothed. For CID integrations, a 10-channel center was

applied to deconvoluted results, and free BSA was operationally defined as the

integrated spectral area up to a mass of 66690 Da, which is less than the expected mass

of the first-bound PFPeA. The fraction of bound PFAA was then calculated using the

integrated area above the expected mass of the first-bound tested PFAA relative to the

free BSA area.

145

Figure 6.5S. Structures and names of perfluoroalkyl acids (PFAAs) included in the

present study. Compound abbreviations and notations adopted for reference in the

manuscript (C2 to C14) are listed. PFSAs and PFCAs have a fluorocarbon tail length of n

+ 1 and m + 1, respectively.

146

Table 6.2S. Dialysis mass balance, reservoir matrix and bovine serum albumin (BSA)

spike recovery results, and liquid chromatography tandem mass spectrometry (LC-

MS/MS) transitions monitored. Errors represent 95% confidence intervals.

Analyte

Test Reservoir

Mass Balance

Reservoir matrix spike

recovery

BSA spike

recovery

Primary transition monitored

(m/z)

Secondary transition monitored

(m/z)

Internal standard

Internal standard transition monitored

(m/z) PFBS 90 ± 16% 103 ± 21% 95 ± 3% 299 > 80 299 > 99 [13C2] PFOS 503 > 99 PFHxS 87 ± 13% 122 ± 24% 97 ± 5% 399 > 80 399 > 99 [13C2] PFOS 503 > 99 PFOS 95 ± 27% 120 ± 24% 99 ± 6% 499 > 80 499 > 99 [13C2] PFOS 503 > 99 PFPeA 103 ± 19% 87 ± 24% 101 ± 5% 263 > 219 263 > 69 [13C2] PFOA 415 > 370 PFHxA 119 ± 19% 98 ± 23% 95 ± 4% 313 > 269 313 > 119 [13C2] PFOA 415 > 370 PFHpA 107 ± 19% 103 ± 21% 97 ± 4% 363 > 319 363 > 169 [13C2] PFOA 415 > 370 PFOA 103 ± 23% 108 ± 10% 94 ± 3% 413 > 369 413 > 169 [13C2] PFOA 415 > 370 PFNA 107 ± 43% 112 ± 21% 97 ± 4% 463 > 419 463 > 169 [13C5] PFNA 468 > 423 PFDA 106 ± 18% 93 ± 11% 99 ± 3% 513 > 469 513 > 219 [13C2] PFDA 515 > 470 PFUnA 99 ± 14% 105 ± 13% 101 ± 5% 563 > 519 563 > 269 [D5] N-

EtFOSAA1 589 > 419

PFDoA 129 ± 58% 123 ± 21% 93 ± 5% 613 > 569 613 > 319 [D5] N-EtFOSAA

589 > 419

PFTA 92 ± 31% 119 ± 14% 98 ± 7% 713 > 669 713 > 169 [D5]–N-EtFOSAA

589 > 419

1 N-deuterioethylperfluoro-1-octanesulfonamidoacetic acid

147

Figure 6.6S. Total PFAA analyte concentration in the bound phase (CP, [g bound

PFAA / mL BSA]) versus total aqueous PFAA concentration (CW, [g free PFAA / mL

water]). The slope of the linear regression (R2 = 0.923) yields an apparent PFAA-BSA

distribution coefficient (log KPW = 3.92 ± 0.06).

Table 6.3S. Protein-water distribution coefficients for PFAAs with BSA. Standard

errors are from regressions performed using Kaleidagraph software.

Analyte Number of Fluorinated

Carbons log KPW R2

PFBS 4 3.86 ± 0.07 0.921 PFHxS 6 4.3 ± 0.1 0.707 PFOS 8 4.1 ± 0.1 0.782 PFPeA 4 3.40 ±0.02 0.994 PFHxA 5 4.05 ± 0.02 0.993 PFHpA 6 4.23 ± 0.08 0.918 PFOA 7 4.14 ± 0.04 0.973 PFNA 8 4.05 ± 0.08 0.904 PFDA 9 3.86 ± 0.08 0.893 PFUnA 10 3.7 ± 0.2 0.651 PFDoA 11 3.3 ± 0.1 0.738

148

Figure 6.7S. Representative deconvoluted spectra of 50 µM BSA alone or with 50 µM

TFA, PFPrA, PFBA, PFBS, PFHxS, or PFOS.

149

Table 6.4S. Measured incremental mass shifts (ΔM) from measured BSA peak (P) to

BSA-PFAA peaks (P + jL) for representative spectra in manuscript Figure 6.2 and

Supporting Information Figure 6.7S.

Ligand Theoretical

ligand molecular

weight (g/mole)

Measured incremental mass (ΔM)

[(P+L)-P] [(P+2L)-P]/2 [(P+3L)-P]/3

PFBS 299 299 301 3001 PFHxS 399 400 400 399 PFOS 499 500 500 5102 TFA 113 117 ND ND PFPrA 163 164 1631 ND PFBA 213 213 2131 ND PFPeA 263 265 2651 ND PFHxA 313 314 313 ND PFHpA 363 364 363 ND PFOA 413 414 413 ND PFNA 463 463 463 4641

1 Peak visible as a shoulder to broad spectrum. 2 A fourth shoulder peak identified for PFOS at ΔM = 489 mass units.

150

Figure 6.8S. Representative deconvoluted spectra of PFPeA and PFNA (100 µM) with

BSA (50 µM) collected at a 10 V collision energy (left) and representative spectra of

PFNA (100 µM) with BSA (50 µM) at 10, 30, 50 or 70 eV collision energy (right).

Charge states are displayed for raw spectral results.

151

Figure 6.9S. Fluorescence spectra of BSA at pH 6 (solid line), 7 (long dashed line), 8

(short dashed line), or 9 (dotted line). (a) Raw data; (b) Data normalized to the

maximum intensity in each spectrum.

152

Table 6.5S. Average fraction of PFAAs bound to BSA (197 ± 2 µM) for a range of pH

conditions. Errors represent standard deviations of results from triplicate dialysis bags

after 120 hours equilibration.

Analyte pH 6 pH 7 pH 8 pH 9 PFBS 98.2 (± 0.2)% 97.2 (± 0.9)% 96.6 (± 0.5)% 95.6 (± 0.7)% PFHxS 99.3 (± 0.1)% 99.3 (± 0.2)% 99.0 (± 0.1)% 99.2 (± 0.3)% PFOS 99.10 (± 0.04)% 99.3 (± 0.1)% 98.6 (± 0.1)% 99.1 (± 0.2)% PFPeA 92 (± 1)% 91 (± 2)% 86 (± 2)% 84 (± 1)% PFHxA 97.8 (± 0.3)% 97.5 (± 0.5)% 97.00 (± 0.03)% 95.6 (± 0.3)% PFHpA 98.9 (± 0.1)% 99.1 (± 0.2)% 98.9 (± 0.1)% 98.4 (± 0.2)% PFOA 99.0 (± 0.1)% 98.9 (± 0.2)% 98.6 (± 0.1)% 99.0 (± 0.1)% PFNA 98.87 (± 0.05)% 98.9 (± 0.2)% 97.8 (± 0.4)% 99.2 (± 0.1)% PFDA 97.7 (± 0.2)% 98.9 (± 0.1)% 98.2 (± 0.3)% 98.2 (± 0.1)% PFUnA 96.4 (± 0.3)% 98.6 (± 0.2)% 96.9 (± 0.3)% 97.9 (± 0.3)% PFDoA 94.2 (± 0.4)% 96.7 (± 0.9)% 90 (± 2)% 96.2 (± 0.1)%

Figure 6.10S. Average number of bound perfluoroalkyl carboxylates (PFCAs, ) or

perfluoroalkyl sulfonates (PFSAs, ) per BSA,

!

" (µM PFAAbound / µM BSA),

measured in dialysis bags containing 200 µM BSA in a PFAA-spiked reservoir at pH 7.

Results illustrate a chain-length dependence of PFCA binding to BSA.

153

Figure 6.11S. Effect of pH on the average number of PFHpA (), PFOA (), PFNA

(), or PFDA () molecules bound to BSA.

Table 6.6S. Slope of linear regressions for average number of PFAAs bound to BSA,

!

" , versus pH (6 to 9) in equilibrium dialysis tests.

Analyte Slope R2 PFBS -0.005 0.639 PFHxS 0.0097 0.855 PFOS 0.0146 0.924 PFPeA -0.0061 0.965 PFHxA -0.0052 0.887 PFHpA 0.0293 0.001 PFOA 0.1563 0.009 PFNA 0.7645 0.055 PFDA 1.2569 0.272 PFUnA 0.0089 0.570 PFDoA 0.0041 0.592

154

Figure 6.12S. Changes in the fluorescence of BSA with added PFNA (top) or PFOS

(bottom) at pH 6 (), 7 (), 8 (), or 9 (). (a) and (c): Both PFNA and PFOS cause a

dose-dependent blue shift in the wavelength of maximum emission of BSA; there is no

significant difference over the range of tested pH. (b) and (d): Both PFNA and PFOS

cause a dose-dependent decrease in the fluorescence emission of BSA; the extent of this

decrease is diminished as the pH increases.

155

Figure 6.13S. The binding of PFNA (a) and PFOS (b) to BSA, plotted as the degree of

saturation (Y) versus total PFAA concentration, at pH 6 (), 7 (), 8 (), or 9 (). The

degree of saturation was calculated as by Hebert and MacManus-Spencer (163).

156

Figure 6.14S. Dependence of estimated binding constant (KHill) on pH for the binding

of PFNA () and PFOS () to BSA.

157

Figure 6.15S. Changes in the fluorescence of BSA with added PFNA (top) or PFOS

(bottom) at 0.21 M (), 0.30 M (), or 0.41 M () ionic strength and pH 7. (a) and (c):

Both PFNA and PFOS cause a dose-dependent blue shift in the wavelength of

maximum emission of BSA; there is no significant difference over the range of tested

ionic strength. (b) and (d): Both PFNA and PFOS cause a dose-dependent decrease in

the fluorescence emission of BSA; there is no significant difference over the range of

tested ionic strength.

158

Figure 6.16S. The binding of PFNA (a) and PFOS (b) to BSA, plotted as the degree of

saturation (Y) versus total PFAA concentration, at 0.21 M (), 0.30 M (), or 0.41 M

() ionic strength and pH 7. The degree of saturation was calculated as by Hebert and

MacManus-Spencer (163).

159

Figure 6.17S. Dependence of estimated binding constant (KHill) on ionic strength for

the binding of PFNA () and PFOS () to BSA at pH 7.

160

161

Chapter 7

Conclusions

7.1 Summary Conclusions

In this thesis, drivers and hindrances for water reuse implementation in Northern

California were assessed, and opportunities of water reuse for natural system

enhancement were identified (Chapters 2-3). Subsequently, the bioaccumulation of

perfluoroalkyl acids (PFAAs), focusing on behavior of binding to proteins as a

proposed biological partitioning parameter, was evaluated (Chapters 4-6). In response to

the research questions outlined in Chapter 1, the following conclusions regarding these

topics can be drawn:

What are the major drivers and barriers to water reuse in Northern

California, and how have these factors evolved through time? Despite growth of

water reuse throughout California, the state has failed to meet recycling goals

established over the past several decades. In Chapter 2, major factors that influenced the

implementation of water-recycling projects in the region were presented based on a

survey of water reuse program managers and facility representatives in Northern

162

California. The study revealed that regulatory requirements limiting discharge played an

important role in motivating many water reuse programs in the region. However, a trend

away from reuse as a disposal issue was documented, as water supply and reliability

become more prevalent drivers of water reuse. Respondents cited economic challenges

as the greatest barrier to successful project implementation. In particular, managers of

smaller water reuse programs more frequently experienced challenges in acquiring

grants and loans, while larger programs had somewhat greater challenges associated

with distribution system costs. Issues of cost recovery were also expressed as barriers to

implementation of water reuse for ecosystem enhancement. Negative perceptions of

water reuse were not frequently major hindrances to implementation of water reuse

programs in the region. Public perception of water reuse may be positively influenced

by a shift in view of recycled water as a valuable resource and as public knowledge of

water supply challenges increase. However, almost half of survey respondents cited

perceived human or environmental health risks due to constituents of emerging concern

as a hindrance to recycled water program implementation. Today, trace chemicals

detected in effluent and receiving waters represent a technological challenge and a

source of concern for recycled water managers.

To what extent has water reuse been applied for the direct benefit of

ecosystems, and what major challenges are associated with the implementation of

water reuse for ecosystem enhancement? Although ecosystem enhancement or

protection goals were frequently cited as drivers of water reuse, such goals were rarely

the most important drivers for program implementation. A survey of databases and

input from managers in Northern California indicates that few water reuse programs in

California have been implemented for the explicit purpose of ecosystem enhancement.

A past effort by the Bay Area Water Recycling Program to identify potential new

programs in the San Francisco Bay region framed important issues to evaluate potential

wetland or stream augmentation sites, but this analysis was far from comprehensive.

Additionally, relatively little progress has been made towards implementation of water

reuse for ecosystem enhancement over the past decade in California. A newly

developed and validated rapid assessment method for California wetlands represents a

potential tool for identifying opportunities for water reuse for natural system

163

enhancement. This method does not evaluate the capacity of wetlands to improve water

quality; rather, it indicates physical and biological attributes that link to the ecosystem’s

ability to support flora and fauna. However, if the links between hydrologic regimes and

wetland condition indicators are well established, the assessments may be utilized to

characterize opportunities for ecosystem enhancement using tertiary treated wastewater.

Amongst a range of challenges for implementing these types of projects, understanding

the bioaccumulation of chemicals of emerging concern represents a particular research

need. Because of their unique chemical properties, extreme environmental persistence,

and elusive bioaccumulation mechanisms, perfluoroalkyl acids were selected for further

analysis.

What dominant processes govern the bioaccumulation of perfluoroalkyl acids

(PFAAs), and how can these processes be captured in bioaccumulation models?

PFAA concentrations detected in white sturgeon fish livers from organisms in the San

Francisco Bay, presented in Chapter 4, contribute to literature on the widespread

detection of PFAAs in the environment and relative dominance of

perfluorooctanesulfonate (PFOS) in biological samples. PFOS was detected in 14 of 15

white sturgeon fish livers, ranging in concentration from 14 ng/g ww to 180 ng/g ww.

Correlations of fish liver concentrations with organism stable isotope ratios demonstrate

the importance of ecological approaches to understanding biomagnification from varied

food sources. However, even simplistic Tier 1 screening measures for evaluating the

bioaccumulative potential of new chemicals, a necessity for effective decision-making,

generally do not incorporate expected bioaccumulation mechanisms relevant to PFAAs.

Traditional models based on legacy hydrophobic organic contaminants correlate

bioconcentration factors with octanol-water partition coefficients (KOW). These models

are insufficient for capturing the nature of biological uptake of ionic species such as

PFAAs. Because interactions of PFAAs with tissue and serum proteins likely contribute

to their tissue distribution and bioaccumulation patterns, an empirical fugacity approach

that incorporates protein binding, considered analogous to KOW, may improve predictive

measures. Protein-water distribution coefficients (KPW) based on ligand associations

with bovine serum albumin (BSA) are proposed as biologically relevant parameters to

describe the environmental behavior of PFAAs.

164

How do long-chain perfluoroalkyl acids interact with the model protein, serum

albumin, at physiologically relevant PFAA:albumin mole ratios? Previous studies

report a wide range of association constants between perfluorooctanoate (PFOA) and

albumin, with most values suggesting relatively weak binding (<104 M-1), likely

resulting from the use of high PFAA:albumin mole ratios. In Chapter 5, association

constants (Ka) and binding stoichiometries for PFAA-albumin complexes were

quantified over a wide range of PFAA:albumin mole ratios. Primary association

constants for PFOA or perfluorononanoate (PFNA) with BSA determined via

equilibrium dialysis are on the order of 106 M-1 with one to three primary binding sites,

comparable to the affinity of fatty acids with albumin. PFNA was greater than 99.9%

bound to BSA or human serum albumin (HSA) at a physiological PFAA:albumin mole

ratio. Nanoelectrospray ionization mass spectrometry (nanoESI-MS) reveals PFAA-

BSA complexes with up to eight occupied binding sites at a 4:1 PFAA:albumin mole

ratio. Association constants estimated by nanoESI-MS are on the order of 105 M-1 for

PFOA and PFNA and 104 M-1 for perfluorodecanoate and PFOS.

What analytical tools are appropriate for quantitatively determining PFAA-

BSA associations? Work presented in Chapters 5 allows comparison of a standard

solution-based method (equilibrium dialysis) with a modern mass spectrometric

approach (automated nanoESI-MS). Along with fluorescence spectroscopy techniques

applied in Chapter 6, and further investigation conducted by MacManus-Spencer et al.

(137), these methods provide complementary information about the strength of PFAA-

albumin binding interactions, the number of binding sites at low ligand:protein mole

ratios, and physiochemical mechanisms of interactions. As evident in previous studies

of PFAA-albumin interactions that utilized spectroscopic methods, electrophoresis, 19F

NMR, and surface tension, limitations associated with analytical methods often require

high PFAA concentrations in experimental analysis. Equilibrium dialysis, a

thermodynamically sound and straightforward protein-ligand analysis technique,

produces reliable data for PFAAs with less than 12 perfluoroalkyl carbons. When

coupled with sensitive detection methods, achievable via liquid chromatography tandem

mass spectrometry (LC-MS/MS), dialysis can yield binding parameters at

physiologically relevant PFAA:BSA mole ratios. NanoESI-MS is a useful technique for

165

rapid characterization PFAA-protein interactions and may provide utility for screening

small molecule interactions with proteins. However, challenges with spectral resolution,

sensitivity of complexes to instrument conditions, and questions regarding physiological

relevance of a gas-phase approach limit the utility of nanoESI-MS as a fully

quantitative method for characterizing PFAA-BSA interactions. Fluorescence

spectroscopy allows for analysis of a wide range of PFAA:albumin mole ratios, and,

though an indirect method, offers insights into conformational changes in the protein

that occur upon ligand binding and with changes in solution conditions.

How will a reduction in perfluoroalkyl chain length affect protein-water

distribution coefficients? A general shift in commercial production of PFAAs from

long- to short-chain chemicals is underway, yet empirical data on the environmental

behavior of short-chain PFAAs are limited. In Chapter 6, associations of perfluoroalkyl

carboxylates (PFCAs) with 2 to 12 carbons (C2 – C12) and perfluoroalkyl sulfonates

(PFSAs) with 4 to 8 carbons (C4, C6, and C8) with BSA are evaluated at low

PFAA:albumin mole ratios and various solution conditions using equilibrium dialysis,

nanoelectrospray ionization mass spectrometry, and fluorescence spectroscopy. Log

KPW values for C4 to C12 PFAAs range from 3.3 to 4.3, greater in magnitude than

octanol-water distribution coefficients for PFAAs. Corresponding association constants

determined for perfluorobutanesulfonate and perfluoropentanoate with BSA are high

(Ka ~106 M-1), and the C4-sulfonate exhibits increased affinity relative to the equivalent

chain-length PFCA. Association constants determined for perfluorobutanesulfonate and

perfluoropentanoate with BSA (Ka ~ 106 M-1) are on the order of those for long-chain

PFAAs, suggesting that physiological implications of strong binding to albumin may be

important for short-chain PFAAs.

What physiochemical mechanisms govern interactions of perfluoroalkyl acids

with serum albumin? Results in Chapters 5 and 6 suggest binding through specific

high affinity interactions at low PFAA:albumin mole ratios. Affinity for BSA increases

with PFAA hydrophobicity but decreases from the C8 to C12 PFCAs, likely due to steric

hindrances associated with longer and more rigid perfluoroalkyl chains. Differences in

binding of the C4-sulfonate relative to the C5-carboxylate, observed via dialysis and

collision induced dissociation, indicate an electrostatic component to interactions with

166

BSA. PFAA-BSA fluorescence titrations conducted at varying pH and ionic strength

support evidence that an observed dependence of binding on pH is due to

conformational changes in the protein. For strongly associating ligands such as PFAAs,

use of KPW to describe binding to albumin should be conducted with caution and limited

to a narrow concentration range over which binding may be approximated as linear.

Additionally, although a serum protein association constant may be a useful parameter

to contribute to the characterization of PFAA bioaccumulative potential, the parameter

has limitations. Serum albumin represents one amongst a multitude of protein targets

for in vivo perfluoroalkyl binding. Additional protein targets or mixtures require

identification to more fully capture the observed bioaccumulation trends of PFAAs.

7.2 Future Work

The work conducted in this dissertation suggests several lines of future research,

including but not limited to, those discussed below.

Identifying needs of water and wastewater agencies, evaluating and quantifying

success, and recommendations for institutional arrangements.

Research presented in Chapters 2 and 3 comprise a regional assessment of water

reuse opportunities and challenges for existing water recycling facilities. As a practical

approach, we inferred that the challenges overcome by implemented programs

represents a minimum of potential challenges facing those who have not yet

implemented programs. However, as a follow-up step, wastewater and water agencies

that do not produce and utilize recycled water should be approached to better

understand the needs of these agencies. In particular, wastewater facilities with potential

couplings to wetland creation or enhancement opportunities may be a first cohort to

further assess interest, potential, and agency needs for ecosystem enhancement using

recycled water. Secondly, additional metrics to evaluate the degree of success for a

recycled water program require development. In the present study, all operational

programs were equally weighted in the analysis. An assessment of the degree of success

would facilitate identification of strategies for particularly successful programs as well

as lessons from those that faced more significant hurdles. Finally, surprisingly little

167

information is available regarding effective pricing structures and institutional

mechanisms to collaboratively develop and operate recycled water programs. Yet, as

shown in this work, economic hindrances frequently limit reuse development and

expansion. Case study analysis of the particular scenarios in which a recycled water

system is limited due to potential revenue reductions for a water purveyor would be

valuable. Future economic analyses should be coupled to a tight feedback loop to

potential recycled water producers and distributors. The existing National Database of

Water Reuse Facilities, which does not include pricing structures or funding

mechanisms, is a practical choice for compiling and sharing such data.

Ecosystem services of wetlands that reclaim tertiary treated wastewater.

Though the needs of our ecosystems, and their responses to human perturbations,

are inherently difficult to quantify, estimating the ecological capital generated from the

implementation of water reuse for ecosystem enhancement in terms that relate to human

activities and investment may lead to management decisions that yield overall greater

environmental and human benefit. In addition to creating recreational opportunities and

improving urban aesthetic, wetlands have the potential to provide water storage

capacity, treat for bulk water quality (e.g., biochemical oxygen demand), remove

nutrients, sequester (or emit) greenhouse gases including carbon dioxide and nitrous

oxide, and attenuate organic contaminants through photolytic and phytotransformation

as well as sorption and redox processes. In order to aid managers in evaluating the

economic potential for various natural system water reuse options, these processes

require quantification and comparison with more traditional recycled water uses through

life cycle assessment approaches.

In situ bioaccumulation and passive sampling of micropollutants in water reuse

systems.

In order to conduct broad assessments of ecosystem services generated from

recycled water wetlands, in situ characterization of micropollutant attenuation and

bioaccumulation should be conducted. Application of passive samplers in wetland water

reuse systems could yield design insights for physical structures and vegetative

168

communities that promote natural attenuation of trace contaminants. Today, many

apolar, persistent organic pollutants are monitored using passive samplers that

accumulate hydrophobic organic contaminants preferentially compared to water.

Passive samplers may also be biomimetic, emulating the body burden of biota. The

Polar Organic Chemical Integrative Sampler (POCIS) has also been applied for a range

of pharmaceuticals and personal care products, though not commonly for perfluorinated

compounds. The properties of PFAA binding to albumins or other proteins may be

exploited to develop new passive sampling devices for these anionic contaminants.

Bioaccumulation measurement and modeling for persistent, ionic organic pollutants

A pollutant-specific study of biological accumulation processes, such as that for

PFAAs, provides insights into additional needs for bioaccumulation approaches for

emerging chemical contaminants. However, with approximately seven new chemicals

on the horizon each day in the United States alone, decisions to accept or reject

compounds for widespread use require rapid analytical approaches. New screening

methods are especially necessary for compounds that fall outside the range of

predictability of log KOW-bioconcentration relationships (e.g., large, ionizable chemicals

that may metabolize or concentrate in specific organism body compartments). In the

present study, dialysis measurements were augmented by nanoESI as a rapid technique

to analyze small-molecule binding to proteins. However the nanoESI technique is

neither universally applicable nor fully physiologically relevant. Development of

additional techniques that rapidly capture relevant mechanisms of bioaccumulation is

needed.

7.3 Final Thoughts

For the design of our chemicals and engineered systems, the best models are those

that mimic natural systems. This applies at all scales and requires continuous attention

and adaptation. Full-cycle processes that return used water to the environment in

planned ways have potential to meet both human and ecosystem needs. Chemicals that

provide products and services to improve quality of life can be designed to readily

transform to benign products. As one of thousands of chemicals developed and utilized

169

for the betterment of society, perfluorooctanoate exemplifies the all too frequent story

of an unregulated chemical discovered to have negative environmental and human

health effects. Only due to cumulative results of incremental research, leading to public

concern and eventual collaborative industrial and governmental action, is the chemical

phased out of production. It is unfortunate that eliminating the use of one unsafe

chemical does not prevent substitution with another that is equally bad or worse. As

technologies evolve and new policies are developed, moving away from reactionary

measures towards preemptive approaches that limit widespread production and use of

potentially harmful chemicals, prior to their environmental detection, is imperative.

170

171

References (1) Levine, A. D.; Asano, T. Recovering sustainable water from wastewater.

Environ. Sci. Technol. 2004, 38 (11), 201A-208A.

(2) Water Recycling 2030: Recommendations of California's Recycled Water Task

Force; Department of Water Resources: Sacramento, CA, 2003.

(3) California Statewide Recycled Water Survey in 2001; California State Water

Resources Control Board Sacramento, California, 2002.

(4) Sedlak, D. L.; Gray, J. L.; Pinkston, K. E. Understanding microcontaminants in

recycled water. Environ. Sci. Technol. 2000, 34 (23), 509A-515A.

(5) Plumlee, M. H.; Larabee, J.; Reinhard, M. Perfluorochemicals in water reuse.

Chemosphere 2008, 72 (10), 1541-1547.

(6) Gross, B.; Montgomery-Brown, J.; Naumann, A.; Reinhard, M. Occurrence and

fate of pharmaceuticals and alkylphenol ethoxylate metabolites in an effluent-

dominated river and wetland. Environ. Toxicol. Chem. 2004, 23 (9), 2074-2083.

(7) Mills, R. A.; Karajeh, F.; Hultquist, R. H. California's Task Force evaluation of

issues confronting water reuse. Water Sci. Technol. 2004, 50 (2), 301-308.

(8) Miller, G. W. Integrated concepts in water reuse: Managing global water needs.

Desalination 2006, 187 (1-3), 65-75.

(9) Gearheart, R. A. Use of constructed wetlands to treat domestic waste-water, City

of Arcata, California. Water Sci. Technol. 1992, 26 (7-8), 1625-1637.

(10) Smith, B. R. Re-thinking wastewater landscapes: Combining innovative

strategies to address tomorrow's urban wastewater treatment challenges. Water

Sci. Technol. 2009, 60 (6), 1465-1473.

(11) Gurr, C. J.; Reinhard, M. Harnessing natural attenuation of pharmaceuticals and

hormones in rivers. Environ. Sci. Technol. 2006, 40 (9), 2872-2876.

(12) USEPA Wetlands from wastewater: The Hayward Marsh EPA832-R-93-005h;

1993.

(13) Drewes, J. E.; Anderson, P.; Denslow, N.; Olivieri, A.; Schlenk, D.; Snyder, S.

Monitoring strategies for chemicals of emerging concern (CECs) in recycled

172

water: Recommendations of a Science Advisory Panel; California State Water

Resources Control Board: Sacramento, CA, April 15, 2010.

(14) Schwarzenbach, R. P.; Escher, B. I.; Fenner, K.; Hofstetter, T. B.; Johnson, C.

A.; von Gunten, U.; Wehrli, B. The challenge of micropollutants in aquatic

systems. Science 2006, 313 (5790), 1072-1077.

(15) USEPA Drinking Water Contaminant Candidate List 3; Federal Register 74

(194), 51850-51862; 2009.

(16) Kissa, E. Fluorinated Surfactants and Repellents, Marcel Dekker: New York,

USA, 2001; Vol. 97.

(17) Rayne, S.; Forest, K. Perfluoroalkyl sulfonic and carboxylic acids: A critical

review of physicochemical properties, levels and patterns in waters and

wastewaters, and treatment methods. J. Environ. Sci. Health Part A-

Toxic/Hazard. Subst. Environ. Eng. 2009, 44 (12), 1145-1199.

(18) Simcik, M. F. Global transport and fate of perfluorochemicals. J. Environ.

Monit. 2005, 7 (8), 759-763.

(19) D'Eon, J. C.; Mabury, S. A. Production of perfluorinated carboxylic acids

(PFCAs) from the biotransformation of polyfluoroalkyl phosphate surfactants

(PAPS): Exploring routes of human contamination. Environ. Sci. Technol. 2007,

41 (13), 4799-4805.

(20) Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H. Sources, fate

and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40 (1), 32-

44.

(21) Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.;

Sulbaek Andersen, M. P.; Wallington, T. J. Degradation of fluorotelomer

alcohols: A likely atmospheric source of perfluorinated carboxylic acids.

Environ. Sci. Technol. 2004, 38 (12), 3316-3321.

(22) Rhoads, K. R.; Janssen, E. M. L.; Luthy, R. G.; Criddle, C. S. Aerobic

biotransformation and fate of N-ethyl perfluorooctane sulfonamidoethanol (N-

EtFOSE) in activated sludge. Environ. Sci. Technol. 2008, 42 (8), 2873-2878.

(23) Plumlee, M. H.; McNeill, K.; Reinhard, M. Indirect photolysis of

perfluorochemicals: Hydroxyl radical-initiated oxidation of N-ethyl

173

perfluorooctane sulfonamido acetate (N-EtFOSAA) and other

perfluoroalkanesulfonamides. Environ. Sci. Technol. 2009, 43 (10), 3662-3668.

(24) Giesy, J. P.; Kannan, K. Global distribution of perfluorooctane sulfonate in

wildlife. Environ. Sci. Technol. 2001, 35 (7), 1339-42.

(25) Houde, M.; Martin, J. W.; Letcher, R. J.; Solomon, K. R.; Muir, D. C. G.

Biological monitoring of polyfluoroalkyl substances: A review. Environ. Sci.

Technol. 2006, 40 (11), 3463-3473.

(26) Lau, C.; Anitole, K.; Hodes, C.; Lai, D.; Pfahles-Hutchens, A.; Seed, J.

Perfluoroalkyl acids: A review of monitoring and toxicological findings.

Toxicol. Sci. 2007, 99 (2), 366-394.

(27) Higgins, C. P.; Field, J. A.; Criddle, C. S.; Luthy, R. G. Quantitative

determination of perfluorochemicals in sediments and domestic sludge. Environ.

Sci. Technol. 2005, 39 (11), 3946-3956.

(28) Schultz, M. M.; Barofsky, D. F.; Field, J. A. Quantitative determination of

fluorinated alkyl substances by large-volume-injection liquid chromatography

tandem mass spectrometry - Characterization of municipal wastewaters.

Environ. Sci. Technol. 2006, 40 (1), 289-295.

(29) Schultz, M. M.; Barofsky, D. F.; Field, J. A. Fluorinated alkyl surfactants.

Environmental Engineering Science 2003, 20 (5), 487-501.

(30) Giesy, J. P.; Naile, J. E.; Khim, J. S.; Jones, P. D.; Newsted, J. L., Aquatic

toxicology of perfluorinated chemicals. In Reviews of Environmental

Contamination and Toxicology, Vol 202, Springer: New York, 2010; Vol. 202,

pp 1-52.

(31) Lau, C.; Butenhoff, J. L.; Rogers, J. M. The developmental toxicity of

perfluoroalkyl acids and their derivatives. Toxicol. Appl. Pharmacol. 2004, 198

(2), 231-241.

(32) USEPA SAB Review of EPA's Draft Risk Assessment of Potential Human Health

Effects Associated with PFOA and Its Salts; 2006.

(33) Giesy, J. P.; Kannan, K. Perfluorochemical surfactants in the environment.

Environ. Sci. Technol. 2002, 36 (7), 146A-152A.

174

(34) Calafat, A. M.; Wong, L. Y.; Kuklenyik, Z.; Reidy, J. A.; Needham, L. L.

Polyfluoroalkyl chemicals in the U.S. population: Data from the National Health

and Nutrition Examination Survey (NHANES) 2003-2004 and comparisons with

NHANES 1999-2000. Environ. Health Perspect. 2007, 115 (11), 1596-1602.

(35) Ritter, S. K. Fluorochemicals go short. Chem. Eng. News 2010, 88 (5), 12-17.

(36) Olsen, G. W.; Chang, S.-C.; Noker, P. E.; Gorman, G. S.; Ehresman, D. J.;

Lieder, P. H.; Butenhoff, J. L. A comparison of the pharmacokinetics of

perfluorobutanesulfonate (PFBS) in rats, monkeys, and humans. Toxicology

2009, 256 (1-2), 65-74.

(37) Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G. Dietary

accumulation of perfluorinated acids in juvenile rainbow trout (Oncorhynchus

mykiss). Environ. Toxicol. Chem. 2003, 22 (1), 189-195.

(38) Martin, J. W.; Whittle, D. M.; Muir, D. C. G.; Mabury, S. A. Perfluoroalkyl

contaminants in a food web from Lake Ontario. Environ. Sci. Technol. 2004, 38

(20), 5379-5385.

(39) Hoff, P. T.; Van de Vijver, K.; Van Dongen, W.; Esmans, E. L.; Blust, R.; De

Coen, W. M. Perfluorooctane sulfonic acid in bib (Trisopterus luscus) and plaice

(Pleuronectes platessa) from the Western Scheldt and the Belgian North Sea:

Distribution and biochemical effects. Environ. Toxicol. Chem. 2003, 22 (3), 608-

614.

(40) Vanden Heuvel, J. P.; Kuslikis, B. I.; Vanrafelghem, M. J.; Peterson, R. E.

Tissue distribution, metabolism, and elimination of perfluorooctanoic acid in

male and female rats. J. Biochem. Toxicol. 1991, 6 (2), 83-92.

(41) Andersen, M. E.; Butenhoff, J. L.; Chang, S. C.; Farrar, D. G.; Kennedy, G. L.;

Lau, C.; Olsen, G. W.; Seed, J.; Wallacekj, K. B. Perfluoroalkyl acids and

related chemistries - Toxicokinetics and modes of action. Toxicol. Sci. 2008, 102

(1), 3-14.

(42) Han, X.; Snow, T. A.; Kemper, R. A.; Jepson, G. W. Binding of

perfluorooctanoic acid to rat and human plasma proteins. Chem. Res. Toxicol.

2003, 16 (6), 775-781.

175

(43) Kelly, B. C.; Ikonomou, M. G.; Blair, J. D.; Surridge, B.; Hoover, D.; Grace, R.;

Gobas, F. Perfluoroalkyl contaminants in an arctic marine food web: Trophic

magnification and wildlife exposure. Environ. Sci. Technol. 2009, 43 (11), 4037-

4043.

(44) Debruyn, A. M. H.; Gobas, F. A. P. The sorptive capacity of animal protein.

Environ. Toxicol. Chem. 2007, 26 (9), 1803-1808.

(45) Bischel, H. N.; MacManus-Spencer, L. A.; Luthy, R. G. Noncovalent

interactions of long-chain perfluoroalkyl acids with serum albumin. Environ.

Sci. Technol. 2010, 44 (13), 5263-5269.

(46) Ingram, P. C.; Young, V. J.; Millan, M.; Chang, C.; Tabucchi, T. From

controversy to consensus: The Redwood City recycled water experience.

Desalination 2006, 187, 179-190.

(47) Rosenblum, E. Selection and implementation of nonpotable water recycling in

"Silicon Valley" (San Jose area) California. Water Sci. Technol. 1999, 40 (4-5),

51-57.

(48) Hermanowicz, S. W.; Diaz, E. S.; Coe, J. Prospects, problems and pitfalls of

urban water reuse: A case study. Water Sci. Technol. 2001, 43 (10), 9-16.

(49) Sheikh, B.; Cort, R. P.; Kirkpatrick, W. R.; Jaques, R. S.; Asano, T. Monterey

waste-water reclamation study for agriculture. Res. J. Water Pollut. C. 1990, 62

(3), 216-226.

(50) Nakamoto, M., Watsonville recycled water project story: Preserving agriculture

in the Pajaro Valley. In WateReuse California Annual Conference, San Diego,

CA, 2010.

(51) Levine, B.; Reich, K.; Sheilds, P.; Suffet, I. H.; Lazarova, V. Water quality

assessment for indirect potable reuse: A new methodology for controlling trace

organic compounds at the West Basin Water Recycling Plant (California, USA).

Water Sci. Technol. 2001, 43 (10), 249-257.

(52) Everest, W. R.; Watson, I. C.; MacLain, D. Groundwater reclamation by

innovative desalting in Orange County, California. Desalination 1998, 117 (1-

3), 197-202.

176

(53) Asano, T.; Cotruvo, J. A. Groundwater recharge with reclaimed municipal

wastewater: Health and regulatory considerations. Water Res. 2004, 38 (8),

1941-1951.

(54) Recycled Water Policy; State Water Resources Control Board: Sacramento, CA,

2009;

http://www.swrcb.ca.gov/water_issues/programs/water_recycling_policy/docs/r

ecycledwaterpolicy_approved.pdf.

(55) Strategic Plan Update 2008-2012; California State Water Resources Control

Board: Sacramento, September 2, 2008.

(56) Mantovani, P.; Abu-Orf, M.; O’Connor, T. In Nonpotable water reuse

management practices, Bridging the Gap: Meeting the World's Water and

Environmental Resources Challenges, Proceedings of World Water and

Environmental Resources Congress, Orlando, Florida, 2001; Phelps, D.; Sehlke,

G., Eds. ASCE: Orlando, Florida, 2001.

(57) Mills, R. A.; Asano, T. A retrospective assessment of water reclamation

projects. Water Sci. Technol. 1996, 33 (10-11), 59-70.

(58) National Database of Water Reuse Facilities. In WateReuse Foundation: 2008.

(59) California State Municipal Wastewater Recycling Survey in 2009; California

State Water Resources Control Board: Sacramento, CA, April, 2011.

(60) California Municipal Wastewater Reclamation in 1987; California State Water

Resources Control Board: Sacramento, California, June, 1990.

(61) Higgins, J.; Warnken, J.; Sherman, P. P.; Teasdale, P. R. Survey of users and

providers of recycled water: Quality concerns and directions for applied

research. Water Res. 2002, 36 (20), 5045-5056.

(62) Hanak, E.; Lund, J.; Dinar, A.; Gray, B.; Howitt, R.; Mount, J.; Moyle, P.;

Thompson, B. B. Managing California's water: From conflict to reconciliation,

Public Policy Institute of California: San Francisco, CA, 2011.

(63) 2009 California Climate Adaptation Strategy: A report to the Governor of the

State of California in Response to Executive Order S-13-2008; California

Natural Resources Agency: Sacramento, CA, 2009.

177

(64) San Francisco Bay Area Recycled Water Coalition Project Summary; San

Francisco Bay Area Recycled Water Coalition: Antioch, CA, March 3, 2011;

http://www.barwc.org/.

(65) Recycled Water Facility Plan; City of Palo Alto: Palo Alto, CA, December,

2008.

(66) MV/Moffet area recycled water pipeline project: Project completion report;

BACWA: City of Palo Alto, October, 2009.

(67) San Francisco Westside Recycled Water Project Alternatives Workshop Series:

Evaluation of Alternatives; San Francisco Public Utilities Commission: San

Francisco, CA, 2011.

(68) Munoz, C. San Francisco Public Utilities Commission, San Francisco, CA.

Personal communication, 2011.

(69) Po, M.; Kaercher, J. D.; Nancarrow, B. E. N. Literature review of factors

influencing public perceptions of water reuse; 2003.

(70) Bruvold, W. H.; Ward, P. C. Using reclaimed wastewater: Public opinion. J.

Water Pollut. Con. F. 1972, 44 (9), 1690-1696.

(71) Communication principles and practice, public Perception and message

effectiveness; WateReuse Foundation: Alexandria, VA, 2008.

(72) Hurlimann, A.; Dolnicar, S. When public opposition defeats alternative water

projects - The case of Toowoomba Australia. Water Res. 2010, 44 (1), 287-297.

(73) Marks, J. S. Taking the public seriously: The case of potable and non potable

reuse. Desalination 2006, 187 (1-3), 137-147.

(74) Hartley, T. W. Public perception and participation in water reuse. Desalination

2006, 187 (1-3), 115-126.

(75) Dolnicar, S.; Hurlimann, A.; Grun, B. What affects public acceptance of

recycled and desalinated water? Water Res. 2011, 45 (2), 933-943.

(76) Bruvold, W. H. Public-opinion on water reuse options. J. Water Pollut. Con. F.

1988, 60 (1), 45-49.

(77) Nancarrow, B. E.; Leviston, Z.; Po, M.; Porter, N. B.; Tucker, D. I. What drives

communities' decisions and behaviours in the reuse of wastewater. Water Sci.

Technol. 2008, 57 (4), 485-491.

178

(78) Leiby, V. M.; Carpenter, G. Project to conduct survey research to obtain

information/data from all water recycling facilities in California; WateReuse

Foundation: Alexandria, VA, 2008.

(79) Water Recycling Funding Program Division of Financial Assistance Strategic

Plan; California State Water Resources Control Board: Sacramento, 2007.

(80) Sutula, M.; Collins, J. N.; Wiskind, A.; Roberts, C.; Solek, C.; Pearce, S.; Clark,

R.; Fetscher, A. E.; Grosso, C.; O’Connor, K.; Robinson, A.; Clark, C.; Rey, K.;

Morrissette, S.; Eicher, A.; Pasquinelli, R.; May, M.; Ritter, K. Status of

perennial estuarine wetlands in the State of California: Final report to the

Surface Water Ambient Monitoring Program State Water Resources Control

Board; Southern California Coastal Water Research Program: September, 2008.

(81) San Francisco Bay Area Regional Recycling Program Recycled Water Master

Plan; Bay Area Regional Water Recycling Program: 1999.

(82) Verhoeven, J. T. A.; Meuleman, A. F. M. Wetlands for wastewater treatment:

Opportunities and limitations. Ecol. Eng. 1999, 12 (1-2), 5-12.

(83) Constructed treatment wetland system description and performance database.

http://firehole.humboldt.edu/wetland/twdb.html (Accessed May 24, 2011).

(84) USEPA Design manual: Constructed wetlands and aquatic plan systems for

municipal wastewater treatment; EPA/625/1-88/022; 1988.

(85) Gearheart, R. A. In The use of free surface constructed wetland as an alternative

process treatment train to meet unrestricted water reclamation standards, 1999;

Pergamon-Elsevier Science Ltd: 1999; pp 375-382.

(86) USEPA Wetlands as a part of reuse and disposal: Las Gallinas Valley Sanitary

District; EPA832-R-93-005h; 1993.

(87) Cronk, J. K. Constructed wetlands to treat wastewater from dairy and swine

operations: A review. Agric. Ecosyst. Environ. 1996, 58 (2-3), 97-114.

(88) Anderson, S. G.; Blades, A. T.; Klassen, J.; Kebarle, P. Determination of ion-

ligand bond-energies and ion fragmentation energies of electrospray-produced

ions by collision-induced dissociation threshold measurements. Int. J. Mass

Spectom. Ion Process. 1995, 141 (3), 217-228.

179

(89) Crites, R. W.; Dombeck, G. D.; Watson, R. C.; Williams, C. R. Removal of

metals and ammonia in constructed wetlands. Water Environ. Res. 1997, 69 (2),

132-135.

(90) Ponce, V. M.; Lindquist, D. S. Management of baseflow augmentation: A

review. Water Resour. Bull. 1990, 26 (2), 259-268.

(91) Sutula, M. A.; Stein, E. D.; Collins, J. N.; Fetscher, A. E.; Clark, R. A practical

guide for the development of a wetland assessment method: The California

experience. J. Am. Water Resour. Assoc. 2006, 42 (1), 157-175.

(92) Stein, E. D.; Fetscher, A. E.; Clark, R. P.; Wiskind, A.; Grenier, J. L.; Sutula,

M.; Collins, J. N.; Grosso, C. Validation of a wetland rapid assessment method:

Use of EPA's Level 1-2-3 Framework for method testing and refinement.

Wetlands 2009, 29 (2), 648-665.

(93) Collins, J. N., E.D. Stein, M. Sutula, R. Clark, A.E. Fetscher, L. Grenier, C.

Grosso, and A. Wiskind California Rapid Assessment Method (CRAM) for

Wetlands and Riparian Areas (website). www.cramwetlands.org (Accessed

August 2010).

(94) Greenway, M.; Simpson, J. S. Artificial wetlands for wastewater treatment,

water reuse and wildlife in Queensland, Australia. Water Sci. Technol. 1996, 33

(10-11), 221-229.

(95) Conkle, J. L.; White, J. R.; Metcalfe, C. D. Reduction of pharmaceutically active

compounds by a lagoon wetland wastewater treatment system in Southeast

Louisiana. Chemosphere 2008, 73 (11), 1741-1748.

(96) Fono, L. J.; Kolodziej, E. P.; Sedlak, D. L. Attenuation of wastewater-derived

contaminants in an effluent-dominated river. Environ. Sci. Technol. 2006, 40

(23), 7257-7262.

(97) Barber, L. B.; Keefe, S. H.; Antweiler, R. C.; Taylor, H. E.; Wass, R. D.

Accumulation of contaminants in fish from wastewater treatment wetlands.

Environ. Sci. Technol. 2006, 40 (2), 603-611.

(98) Wang, Y. W.; Li, X. M.; Li, A.; Wang, T.; Zhang, Q. H.; Wang, P.; Fu, J. J.;

Jiang, G. B. Effect of municipal sewage treatment plant effluent on

180

bioaccumulation of polychlorinated biphenyls and polybrominated diphenyl

ethers in the recipient water. Environ. Sci. Technol. 2007, 41 (17), 6026-6032.

(99) Euliss, N. H.; Smith, L. M.; Wilcox, D. A.; Brwaine, B. A. Linking ecosystem

processes with wetland management goals: Charting a course for a sustainable

future. Wetlands 2008, 28 (3), 553-562.

(100) Gustavson, K.; Kennedy, E. Approaching wetland valuation in Canada.

Wetlands 2010, 30 (6), 1065-1076.

(101) Lane, C. R.; D'Amico, E. Calculating the ecosystem service of water storage in

isolated wetlands using LiDAR in North Central Florida, USA. Wetlands 2010,

30 (5), 967-977.

(102) Engle, V. D. Estimating the provision of ecosystem services by Gulf of Mexico

coastal wetlands. Wetlands 2011, 31 (1), 179-193.

(103) Fuchs, V. J.; Mihelcic, J. R.; Gierke, J. S. Life cycle assessment of vertical and

horizontal flow constructed wetlands for wastewater treatment considering

nitrogen and carbon greenhouse gas emissions. Water Res. 2011, 45 (5), 2073-

2081.

(104) Brauman, K. A.; Daily, G. C.; Duarte, T. K.; Mooney, H. A. The nature and

value of ecosystem services: An overview highlighting hydrologic services.

Annu. Rev. Environ. Resour. 2007, 32, 67-98.

(105) Hoehn, E.; Plumlee, M. H.; Reinhard, M. Natural attenuation potential of

downwelling streams for perfluorochemicals and other emerging contaminants.

Water Science and Technology 2007, 56 (11), 59-64.

(106) Reinhard, M.; Plumlee, M. H.; Gurr, C. J.; Ashktorab, H.; John, P.; Larabee, J.

Attenuation of emerging contaminants in streams augmented with recycled

water: Draft final project report; Water Reuse Foundation: Alexandria, VA,

2006.

(107) Mackay, D.; Fraser, A. Bioaccumulation of persistent organic chemicals:

Mechanisms and models. Environmental Pollution 2000, 110 (3), 375-391.

(108) Stewart, A. R.; Luoma, S. N.; Schlekat, C. E.; Doblin, M. A.; Hieb, K. A. Food

web pathway determines how selenium affects aquatic ecosystems: A San

Francisco Bay case study. Environ. Sci. Technol. 2004, 38 (17), 4519-4526.

181

(109) de Vos, M. G.; Huijbregts, M. A. J.; van den Heuvel-Greve, M. J.; Vethaak, A.

D.; de Vijver, K. I. V.; Leonards, P. E. G.; van Leeuwen, S. P. J.; de Voogt, P.;

Hendriks, A. J. Accumulation of perfluorooctane sulfonate (PFOS) in the food

chain of the Western Scheldt estuary: Comparing field measurements with

kinetic modeling. Chemosphere 2008, 70 (10), 1766-1773.

(110) Nakata, H.; Kannan, K.; Nasu, T.; Cho, H. S.; Sinclair, E.; Takemura, A.

Perfluorinated contaminants in sediments and aquatic organisms collected from

shallow water and tidal flat areas of the Ariake Sea, Japan: Environmental fate

of perfluorooctane sulfonate in aquatic ecosystems. Environ. Sci. Technol. 2006,

40 (16), 4916-4921.

(111) So, M. K.; Taniyasu, S.; Yamashita, N.; Giesy, J. P.; Zheng, J.; Fang, Z.; Im, S.

H.; Lam, P. K. S. Perfluorinated compounds in coastal waters of Hong Kong,

South China, and Korea. Environ. Sci. Technol. 2004, 38 (15), 4056-4063.

(112) Jeon, J.; Kannan, K.; Lim, H. K.; Moon, H. B.; Ra, J. S.; Kim, S. D.

Bioaccumulation of perfluorochemicals in Pacific oyster under different salinity

gradients. Environ. Sci. Technol. 2010, 44 (7), 2695-2701.

(113) Conder, J. M.; Hoke, R. A.; De Wolf, W.; Russell, M. H.; Buck, R. C. Are

PFCAs bioaccumulative? A critical review and comparison with regulatory

lipophilic compounds. Environ. Sci. Technol. 2008, 42 (4), 995-1003.

(114) Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G. Bioconcentration

and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus

mykiss). Environ. Toxicol. Chem. 2003, 22 (1), 196-204.

(115) Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C.

G.; Mabury, S. A. Identification of long-chain perfluorinated acids in biota from

the Canadian Arctic. Environ. Sci. Technol. 2004, 38 (2), 373-380.

(116) Higgins, C. P.; McLeod, P. B.; MacManus-Spencer, L. A.; Luthy, R. G.

Bioaccumulation of perfluorochemicals in sediments by the aquatic oligochaete

Lumbriculus variegatus. Environ. Sci. Technol. 2007, 41 (13), 4600-4606.

(117) Meylan, W. M.; Howard, P. H.; Boethling, R. S.; Aronson, D.; Printup, H.;

Gouchie, S. Improved method for estimating bioconcentration/bioaccumulation

182

factor from octanol/water partition coefficient. Environ. Toxicol. Chem. 1999,

18 (4), 664-672.

(118) Jing, P.; Rodgers, P. J.; Amemiya, S. High lipophilicity of perfluoroalkyl

carboxylate and sulfonate: Implications for their membrane permeability. J. Am.

Chem. Soc. 2009, 131 (6), 2290-2296.

(119) Rayne, S.; Forest, K.; Friesen, K. J. Estimated bioconcentration factors (BCFs)

for the C4 through C8 perfluorinated alkylsulfonic acid (PFSA) and

alkylcarboxylic acid (PFCA) congeners. J. Environ. Sci. Health Part A-

Toxic/Hazard. Subst. Environ. Eng. 2009, 44 (6), 598-604.

(120) Higgins, C. P.; Luthy, R. G. Sorption of perfluorinated surfactants on sediments.

Environ. Sci. Technol. 2006, 40 (23), 7251-7256.

(121) Mount, D. R.; Highland, T. L.; Mattson, V. R.; Dawson, T. D.; Lott, K. G.;

Ingersoll, C. G. Use of the oligochaete, Lumbriculus Variegatus, as a prey

organism for toxicant exposure of fish through diet. Environ. Toxicol. Chem.

2006, 25 (10), 2760-2767.

(122) Stevenson, C. N.; MacManus-Spencer, L. A.; Luckenbach, T.; Luthy, R. G.;

Epel, D. New perspectives on perfluorochemical ecotoxicology: Inhibition and

induction of an efflux transporter in the marine mussel, Mytilus californianus.

Environ. Sci. Technol. 2006, 40 (17), 5580-5585.

(123) D'Eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A. Atmospheric

chemistry of N-methyl perfluorobutane sulfonamidoethanol,

C4F9SO2N(CH3)CH2CH2OH: Kinetics and mechanism of reaction with OH.

Environ. Sci. Technol. 2006, 40 (6), 1862-1868.

(124) Fasano, W. J.; Carpenter, S. C.; Gannon, S. A.; Snow, T. A.; Stadler, J. C.;

Kennedy, G. L.; Buck, R. C.; Korzeniowski, S. H.; Hinderliter, P. M.; Kemper,

R. A. Absorption, distribution, metabolism, and elimination of 8-2 fluorotelomer

alcohol in the rat. Toxicol. Sci. 2006, 91 (2), 341-355.

(125) Nabb, D. L.; Szostek, B.; Himmelstein, M. W.; Mawn, M. P.; Gargas, M. L.;

Sweeney, L. M.; Stadler, J. C.; Buck, R. C.; Fasano, W. J. In vitro metabolism

of 8-2 fluorotelomer alcohol: Interspecies comparisons and metabolic pathway

refinement. Toxicol. Sci. 2007, 100 (2), 333-344.

183

(126) Gauthier, S. A.; Mabury, S. A. Aqueous photolysis of 8 : 2 fluorotelomer

alcohol. Environ. Toxicol. Chem. 2005, 24 (8), 1837-1846.

(127) Butt, C. M.; Muir, D. C. G.; Mabury, S. A. Elucidating the pathways of poly-

and perfluorinated acid formation in Rainbow Trout. Environ. Sci. Technol.

2010, 44 (13), 4973-4980.

(128) Butt, C. M.; Muir, D. C. G.; Mabury, S. A. Biotransformation of the 8:2

fluorotelomer acrylate in Rainbow Trout. 1. In vivo dietary exposure. Environ.

Toxicol. Chem. 2010, 29 (12), 2726-2735.

(129) Kudo, N.; Katakura, M.; Sato, Y.; Kawashima, Y. Sex hormone-regulated renal

transport of perfluorooctanoic acid. Chem.-Biol. Interact. 2002, 139 (3), 301-

316.

(130) Epel, D.; Luckenbach, T.; Stevenson, C. N.; Macmanus-Spencer, L. A.;

Hamdoun, A.; Smital, T. Efflux transporters: Newly appreciated roles in

protection against pollutants. Environ. Sci. Technol. 2008, 42 (11), 3914-3920.

(131) Olsen, G. W.; Huang, H. Y.; Helzlsouer, K. J.; Hansen, K. J.; Butenhoff, J. L.;

Mandel, J. H. Historical comparison of perfluorooctanesulfonate,

perfluorooctanoate, and other fluorochemicals in human blood. Environ. Health

Perspect. 2005, 113 (5), 539-545.

(132) Olsen, G. W.; Burris, J. M.; Ehresman, D. J.; Froehlich, J. W.; Seacat, A. M.;

Butenhoff, J. L.; Zobel, L. R. Half-life of serum elimination of

perfluorooctanesulfonate, perfluorohexanesulfonate, and perfluorooctanoate in

retired fluorochemical production workers. Environ. Health Perspect. 2007, 115

(9), 1298-1305.

(133) Jones, P. D.; Hu, W. Y.; De Coen, W.; Newsted, J. L.; Giesy, J. P. Binding of

perfluorinated fatty acids to serum proteins. Environ. Toxicol. Chem. 2003, 22

(11), 2639-2649.

(134) Vanden Heuvel, J. P.; Kuslikis, B. I.; Peterson, R. E. Covalent binding of

perfluorinated fatty-acids to proteins in the plasma, liver and testes of rats.

Chem.-Biol. Interact. 1992, 82 (3), 317-328.

184

(135) Han, X.; Hinderliter, P. M.; Snow, T. A.; Jepson, G. W. Binding of

perfluorooctanoic acid to rat liver-form and kidney-form alpha 2u-globulins.

Drug Chem. Toxicol. 2004, 27 (4), 341-360.

(136) Luebker, D. J.; Hansen, K. J.; Bass, N. M.; Butenhoff, J. L.; Seacat, A. M.

Interactions of fluorochemicals with rat liver fatty acid-binding protein.

Toxicology 2002, 176 (3), 175-185.

(137) MacManus-Spencer, L. A.; Tse, M. L.; Hebert, P. C.; Bischel, H. N.; Luthy, R.

G. Binding of perfluorocarboxylates to serum albumin: A comparison of

analytical methods. Anal. Chem. 2010, 82 (3), 974-981.

(138) Kerstner-Wood, C.; Coward, L.; Gorman, G. Protein binding of

perfluorohexane sulfonate, perfluoroooctane sulfonate and perfluorooctanoate

to plasma (human, rat, and monkey), and various human-derived plasma protein

fractions; U.S. EPA Docket AR-226-1354; U.S. Environmental Protection

Agency: Washington, DC, 2003.

(139) Klevens, H. B.; Ellenbogen, E. Protein-fluoroacid interaction - Bovine Serum

Albumin perfluoro-octanoic acid. Discuss. Faraday Soc. 1954, (18), 277-288.

(140) Taves, D. R. Evidence that there are two forms of fluoride in human serum.

Nature 1968, 217 (5133), 1050.

(141) Peters, T. Serum-albumin. Adv. Protein Chem. 1985, 37, 161-245.

(142) He, X. M.; Carter, D. C. Atomic structure and chemistry of human serum

albumin. Nature 1992, 358 (6383), 209-215.

(143) Barton, S. W.; Goudot, A.; Bouloussa, O.; Rondelez, F.; Lin, B. H.; Novak, F.;

Acero, A.; Rice, S. A. Structural transitions in a monolayer of fluorinated

amphiphile molecules. J. Chem. Phys. 1992, 96 (2), 1343-1351.

(144) Messina, P.; Prieto, G.; Ruso, J. M.; Sarmiento, F. Conformational changes in

human serum albumin induced by sodium perfluorooctanoate in aqueous

solutions. J. Phys. Chem. B 2005, 109 (32), 15566-15573.

(145) Messina, P.; Prieto, G.; Dodero, V.; Cabrerizo-Vilchez, M. A.; Maldonado-

Valderrama, J.; Ruso, J. M.; Sarmiento, F. Surface characterization of human

serum albumin and sodium perfluorooctanoate mixed solutions by pendant drop

tensiometry and circular dichroism. Biopolymers 2006, 82 (3), 261-271.

185

(146) Chen, Y. M.; Guo, L. H. Fluorescence study on site-specific binding of

perfluoroalkyl acids to human serum albumin. Arch. Toxicol. 2009, 83 (3), 255-

261.

(147) Wu, L. L.; Gao, H. W.; Gao, N. Y.; Chen, F. F.; Chen, L. Interaction of

perfluorooctanoic acid with human serum albumin. BMC Struct. Biol. 2009, 9,

7.

(148) Gelamo, E. L.; Silva, C.; Imasato, H.; Tabak, M. Interaction of bovine (BSA)

and human (HSA) serum albumins with ionic surfactants: Spectroscopy and

modelling. Biochim. Biophys. Acta-Protein Struct. Molec. Enzym. 2002, 1594

(1), 84-99.

(149) Magdassi, S. Surface activity of proteins: Chemical and physicochemical

modifications, Marcel Dekker, Inc.: New York, 1996.

(150) Zhang, S.; Van Pelt, C. K.; Wilson, D. B. Quantitative determination of

noncovalent binding interactions using automated nanoelectrospray mass

spectrometry. Anal. Chem. 2003, 75 (13), 3010-3018.

(151) Peters, T. J. All About Albumin: Biochemistry, Genetics, and Medical

Applications, Academic Press: San Diego, CA, USA, 1996.

(152) Putnam, F. W., Ed. The Plasma Proteins, 2nd ed.; New York, 1975.

(153) Bordbar, A. K.; Saboury, A. A.; Housaindokht, M. R.; Moosavi-Movahedi, A.

A. Statistical effects of the binding of ionic surfactant to protein. J. Colloid

Interface Sci. 1997, 192, 415-419.

(154) Scatchard, G. The attractions of proteins for small molecules and ions. Ann. N.

Y. Acad. Sci. 1949, 51 (4), 660-672.

(155) Spector, A. A. Fatty acid binding to plasma albumin. J. Lipid Res. 1975, 16

(165-179).

(156) Messina, P.; Prieto, G.; Dodero, V.; Ruso, J. M.; Schulz, P.; Sarmiento, F.

Ultraviolet-circular dichroism spectroscopy and potentiometric study of the

interaction between human serum albumin and sodium perfluorooctanoate.

Biopolymers 2005, 79 (6), 300-309.

186

(157) Sabin, J.; Prieto, G.; Gonzalez-Perez, A.; Ruso, J. M.; Sarmiento, F. Effects of

fluorinated and hydrogenated surfactants on human serum albumin at different

pHs. Biomacromolecules 2006, 7 (1), 176-182.

(158) De Vriendt, K.; Sandra, K.; Desmet, T.; Nerinckx, W.; Van Beeumen, J.;

Devreese, B. Evaluation of automated nano-electrospray mass spectrometry in

the determination of non-covalent protein-ligand complexes. Rapid Commun.

Mass Spectrom. 2004, 18 (24), 3061-3067.

(159) Sundqvist, G.; Benkestock, K.; Roeraade, J. Investigation of multiple binding

sites on ribonuclease A using nano-electrospray ionization mass spectrometry.

Rapid Commun. Mass Spectrom. 2005, 19 (8), 1011-1016.

(160) Veros, C. T.; Oldham, N. J. Quantitative determination of lysozyme-ligand

binding in the solution and gas phases by electrospray ionisation mass

spectrometry. Rapid Commun. Mass Spectrom. 2007, 21 (21), 3505-3510.

(161) Sun, J. X.; Kitova, E. N.; Klassen, J. S. Method for stabilizing protein-ligand

complexes in nanoelectrospray ionization mass spectrometry. Anal. Chem. 2007,

79 (2), 416-425.

(162) Blanco, E.; Ruso, J. M.; Prieto, G.; Sarmiento, F. On relationships between

surfactant type and globular proteins interactions in solution. J. Colloid Interface

Sci. 2007, 316 (1), 37-42.

(163) Hebert, P. C.; MacManus-Spencer, L. A. Development of a fluorescence model

for the binding of medium- to long-chain perfluoroalkyl acids to human serum

albumin through a mechanistic evaluation of spectroscopic evidence. Anal.

Chem. 2010, 82 (15), 6463-6471.

(164) Qin, P. F.; Liu, R. T.; Pan, X. R.; Fang, X. Y.; Mou, Y. Impact of carbon chain

length on binding of perfluoroalkyl acids to bovine serum albumin determined

by spectroscopic methods. J. Agric. Food Chem. 2010, 58 (9), 5561-5567.

(165) D'Eon, J. C.; Simpson, A. J.; Kumar, R.; Baer, A. J.; Mabury, S. A. Determining

the molecular interactions of perfluorinated carboxylic acids with human sera

and isolated human serum albumin using nuclear magnetic resonance

spectroscopy. Environ. Toxicol. Chem. 2010, 29 (8), 1678-1688.

187

(166) Spector, A. A. Fatty-acid binding to plasma albumin. J. Lipid Res. 1975, 16 (3),

165-179.

(167) Woodcroft, M. W.; Ellis, D. A.; Rafferty, S. P.; Burns, D. C.; March, R. E.;

Stock, N. L.; Trumpour, K. S.; Yee, J.; Munro, K. Experimental characterization

of the mechanism of perfluorocarboxylic acids' liver protein bioaccumulation:

The key role of the neutral species. Environ. Toxicol. Chem. 2010, 29 (8), 1669-

1677.

(168) Rayne, S.; Forest, K.; Friesen, K. J. Computational approaches may

underestimate pKa values of longer-chain perfluorinated carboxylic acids:

Implications for assessing environmental and biological effects. J. Environ. Sci.

Health Part A-Toxic/Hazard. Subst. Environ. Eng. 2009, 44 (4), 317-326.

(169) Ellis, D. A.; Denkenberger, K. A.; Burrow, T. E.; Mabury, S. A. The use of F-19

NMR to interpret the structural properties of perfluorocarboxylate acids: A

possible correlation with their environmental disposition. J Phys Chem A 2004,

108 (46), 10099-10106.

(170) Griffiths, W. J.; Johnsson, A. P.; Liu, S.; Rai, S. K.; Wang, Y. Electrospray and

tandem mass spectrometry in biochemistry. Biochem. J. 2001, 355, 545-561.

(171) Benkestock, K.; Edlund, P. O.; Roeraade, J. Electrospray ionization mass

spectrometry as a tool for determination of drug binding sites to human serum

albumin by noncovalent interaction. Rapid Commun. Mass Spectrom. 2005, 19

(12), 1637-1643.

(172) Yin, S.; Xie, Y. M.; Loo, J. A. Mass spectrometry of protein-ligand complexes:

Enhanced gas-phase stability of ribonuclease-nucleotide complexes. J. Am. Soc.

Mass Spectrom. 2008, 19 (8), 1199-1208.

(173) Vinokur, N.; Ryzhov, V. Using collision-induced dissociation with corrections

for the ion number of degrees of freedom for quick comparisons of relative

bonding strength. J Mass Spectrom 2004, 39 (11), 1268-1274.

(174) Zhang, X.; Chen, L.; Fei, X. C.; Ma, Y. S.; Gao, H. W. Binding of PFOS to

serum albumin and DNA: Insight into the molecular toxicity of

perfluorochemicals. BMC Mol. Biol. 2009, 10 (1), 16.

188

(175) Kun, R.; Kis, L.; Dekany, I. Hydrophobization of bovine serum albumin with

cationic surfactants with different hydrophobic chain length. Colloid Surf. B-

Biointerfaces 79 (1), 61-68.

(176) Kun, R.; Szekeres, M.; Dekany, I. Isothermal titration calorimetric studies of the

pH induced conformational changes of bovine serum albumin. J. Therm. Anal.

Calorim. 2009, 96 (3), 1009-1017.