bio degradation of thiomersal containing effluents by a mercury resistance pseudomonas putida strain

8/6/2019 Bio Degradation of Thiomersal Containing Effluents by a Mercury Resistance Pseudomonas Putida Strain

1/12

Water Research 39 (2005) 35113522

Biodegradation of thiomersal containing efuents by amercury resistant Pseudomonas putida strain

Raquel Fortunato, Joa o G. Crespo, M.A.M. Reis

REQUIMTE/CQFB, Department of Chemistry, FCT, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal

Received 21 January 2005; received in revised form 7 June 2005; accepted 16 June 2005Available online 1 August 2005

Abstract

Thiomersal, a toxic organomercurial with a strong bactericidal effect, is the most widely used preservative in vaccineproduction. As a result, vaccine production wastewaters are frequently polluted with thiomersal concentrations abovethe European limit for mercury efuent discharges for which there is, presently, no remediation technology available.This work proposes a biotechnological process for the remediation of vaccine production wastewaters based on thebiological degradation of thiomersal to metallic mercury, under aerobic conditions, by a mercury resistant bacterialstrain. The kinetics of thiomersal degradation by a pure culture of Pseudomonas putida spi3 was rstly investigated inbatch reactors using a thiomersal amended mineral medium. Subsequently, a continuous stirred tank reactor fed withthe same medium was operated at a dilution rate of 0.05 h

1, and the bioreactor performance and robustness wasevaluated when exposed to thiomersal shock loads. In a second stage, the bioreactor was fed directly with a real vaccinewastewater contaminated with thiomersal and the culture ability to grow in the wastewater and remediate it was

evaluated for dilution rates ranging from 0.022 to 0.1 h

1.r 2005 Elsevier Ltd. All rights reserved.

Keywords: Thiomersal; Vaccine wastewaters; Organic mercury; Wastewater treatment; Bioremediation; Pseudomonas

1. Introduction

Thiomersal is a toxic organomercurial with a strongbactericidal effect, which has been used routinely as anadditive to biological products and cosmetics since the1930s. In vaccine production, specically, thiomersal isthe preservative most widely used to maintain a sterileproduction line and to prevent bacterial growth in thecell culture, media and/or in the nal container ( Keithand Walters, 1992 ). Thiomersal contains 49.6% mercury(w/w) and is metabolised in the human body tothiosalycilic acid and ethylmercury (mainly excreted in

the faeces as inorganic mercury) ( Pichichero et al., 2002 ;Magos, 2003 ).

The wastewaters that result from vaccine productionprocesses are polluted with thiomersal concentrationsranging from 25 to 50 mg/L, well above the Europeanlimit for mercury efuent discharges of 0.05mg Hg/L,which is equivalent to 0.1 mg/L thiomersal. Since there ispresently no remediation technology available fororganomercurials this wastewater is, in most cases,delivered to municipal waste treatment plants where thehigh mercury content in the dried activated sludgeprevents its disposal in the environment.

The conventional processes for mercury removalinvolve physical and chemical approaches such as theuse of ion exchange columns (e.g. for HgCl 2 removal),activated carbon adsorbents and chemical precipitation

ARTICLE IN PRESS

www.elsevier.com/locate/watres

0043-1354/$- see front matter r 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.watres.2005.06.012

Corresponding author. Tel./fax: +351 212 948 385.E-mail address: [email protected] (M.A.M. Reis).
http://www.elsevier.com/locate/watreshttp://www.elsevier.com/locate/watres


2/12


3/12

fed with a thiomersal amended mineral medium wasoperated and the bioreactor performance and robustnesswere evaluated when exposed to thiomersal shock loads.In a second stage, the bioreactor was fed directly with avaccine production wastewater contaminated withthiomersal and the culture ability to grow in thewastewater and remediate it was evaluated for differentconditions of reactor operation.

2. Experimental

2.1. Microbial culture and culture media

The P. putida spi3 strain used was isolated fromsediments of the Spittelwasser River by the MolecularMicrobial Ecology Group of GBF-Braunschweig, Ger-many as described in von Canstein et al. (1999) .

In the studies with the thiomersal contaminatedsynthetic wastewater (thiomersal amended mineral med-ium) a mineral medium (M9: 3.7 g/L Na 2HPO 4, 3g/LKH 2PO 4 , 5 g/L NaCl and 1 g/L NH 4Cl) supplementedwith thiomersal and with a concentrated micronutrientssolution (7.23 g/L MgCl 2.6H 2O, 1 g/L CaCO 3, 0.72g/LZnSO 4.7H 2O, 0.42 g/L MnSO 4.H 2O, 0.125 g/L Cu-SO 4.5H 2O, 0.14g/L CoSO 4.7H 2O, 0.06g/L MgSO 4.7-H 2O, 2.5 g/L FeSO 4.7H 2O and 0.01g/L H 3BO 3) wasused. The medium pH was adjusted to pH 7 with 1 MNaOH and glucose was used as carbon source. Glucoseconcentration was chosen accordingly to the aim of theexperiment (see below Section 2.3 and 2.4).

In the studies with the vaccine wastewater, athiomersal contaminated vaccine efuent kindly sup-plied by a multinational vaccine company was used.Prior to operation, the efuent was sterilized byultraltration using a hollow bre polysulfone mem-brane module with a molecular weight cut-off of 500kDa. The thiomersal concentration in the efuentvaried between 41.4 and 48.4mg/L. The total mercuryconcentration was found to vary between 20.7 and24.2 mg/L, meaning that all mercury present in thevaccine wastewater was in the form of thiomersal. Thedissolved organic carbon in the efuent was 128 mg/L.The sterile efuent was supplemented with the concen-trated micronutrients solution described above (400 mLof solution to 1 L of medium). Glucose was used ascarbon source. The bioreactor pH was kept constant atpH 7 (by addition of hydrochloric acid solution 0.5 M).

2.2. Batch reactor operation

A Biostat MD reactor (Braun, Germany) was used forthe batch kinetics studies and the culture was grownaerobically, at 25 1 C, in the mineral medium describedabove. Thiomersal was added to the culture medium, in

an initial concentration of 25 mg/L, immediately before

inoculation. Glucose was added to a nal concentrationof 2.5 g/L. In order to study the microbial culturekinetics in the absence of thiomersal, batch studies werealso carried out in a manner similar to that describedabove, without adding thiomersal to the medium.

A respirometer, through which the culture mediumwas continuously recirculated, was coupled to thereactor. The volumetric oxygen uptake rates (OUR v)were measured, taking into account the decay of thedissolved oxygen concentration in the respirometer,when the recirculation pump was stopped. The observedoxygen/substrate yields ( Y O2/S ) were calculated as:Y O2=s R OUR vtdt=DS , where DS (g/L) is the differ-ence between the glucose concentration at the beginningand at the end of the experiment, OUR v (mg O 2L

1 min1 ) is the volumetric oxygen uptake rate;

R OUR vtdt was calculated as the area below the curvethat describes the variation of the measured OUR v

versus time.For both continuous and discontinuous operation,

the stirring rate, bioreactor temperature, pH anddissolved oxygen concentration were continuously mea-sured and controlled at 150 rpm, 25 1 C, pH 7 and 5mgO2 /L (aeration with compressed air), respectively.

2.3. Continuous reactor operation

Two different procedures were used: (a) the bioreactorwas fed with thiomersal amended mineral medium and(b) the bioreactor was fed with a vaccine efuentcontaminated with thiomersal.

In the rst operation mode, the bioreactor wascontinuously fed with mineral medium, supplementedwith a thiomersal concentration of 192 mg/L and aglucose concentration of 1g/L. The system was operatedat a dilution rate ( D) equal to 0.05 h

1 and the reactorworking volume was kept constant and equal to 2 L,using an overow device. Thiomersal shock loads wereapplied to the system at steady state (thiomersalconcentration at steady state always below 5 mg/L).After the rst thiomersal pulse (additional concentrationof 25 mg/L in the bioreactor), the system was allowed toreturn to steady state, after which a new thiomersalpulse was applied (additional concentration of 50 mg/Lin the bioreactor).

In the second operation mode, the bioreactor wascontinuously fed with the vaccine production efuent. Asolution with 3.6 g/L of glucose was added separately tothe bioreactor at a constant ow rate of 0.18mL/min.The reactor was operated at four different dilution rates(0.022, 0.03, 0.05, 0.1h

1) by adjusting the efuent owrate and keeping the reactor working volume constantand equal to 1 L, using an overow device. Chemicalanalysis of the wastewater showed that the averageconcentration of ammonia in the efuent was 13 mg/L.

In order to verify whether this value was limiting,

ARTICLE IN PRESS

R. Fortunato et al. / Water Research 39 (2005) 35113522 3513


4/12

preliminary experiments were carried out for D 0:03 h 1 with and without the addition of an externalammonia source. For the other dilution rates tested(D 0:022, 0.05, 0.1 h

1) no external ammonia sourcewas added to the system. The operation conditions used

are depicted in Table 1 .In order to test the efcacy of using a carbon lter for

nal polishing, 200mL of mercury containing solutions(HgCl 2) in the concentration range 0.150mg/L werecontacted, for 48h at room temperature, with 0.005g of activated carbon (Norit GAC 1240Plus, A-10128). Aftercontact, the mercury concentration in solution wasdetermined by inductively coupled plasma spectroanalysis.

2.4. Analytical methods

The cell density in the bioreactor was determined byoptical density (OD) measurements at 600 nm andcompared with a OD versus cell dry weight calibrationcurve.

The thiomersal concentration was determined byHPLC, using a reverse phase column (Nucleosil 1005C18, MN) connected to a UV detector (Merck Hitachi)set at a wavelength of 250 nm. The mobile phase usedwas a methanol/10 mM phosphate buffer (pH6) solution(55:45) at a ow rate of 0.4mL/min. The detection limitfor thiomersal was 0.8 mg/L. The total mercury contentof the samples was measured by inductively coupledplasma spectroanalysis (ICP-JYultima 238). The detec-tion limit was 0.05 mg Hg/L. In order to verify whethermercury was adsorbed by the cells, cell pellets werefreeze-dried and digested with 1.5 mL of nitric acid(Fluka, Germany) for 24 h at 42 1 C (according to themethod described by Drapeau et al., 1983 ) The digestedsamples were diluted 1:2 with de-ionised water and thetotal mercury concentration was measured by induc-tively coupled plasma spectroanalysis (ICP). The volu-metric amount of mercury adsorbed was calculatedbased on the measured mercury concentration and thecell dry weight in the analysed samples.

The glucose concentration, in the experiments withsynthetic media, was determined by HPLC (Merck

Hitachi), at 301

C, using a Aminex HPX-87H column

and a refractive index detector (Merck Hitachi). Themobile phase used was a 0.01 N solution of sulphuricacid at a ow rate of 0.2mL/min. The detection limitwas 5 mg/L. In the experiments with the vaccine efuent,the glucose concentration was determined using an

enzymatic method (Glucose HK, Sigma Diagnostics).The detection limit was 5 mg/L.

The ammonia concentration in the bioreactor wasdetermined using an ammonia gas sensing combinationelectrode (ThermoOrion 9512). The detection limit was0.5mg N/L.

A gas-chromatograph coupled to a mass spectrometer(GS/MS Fison MD800) equipped with a molecular sievecolumn (GS-Molesieve 30 m 0.541 mm) was used forethane detection.

The detection of metallic mercury in the bioreactorheadspace was carried out by cold vapour atomicadsorption spectroscopy (CV-AAS), using a Thermoelectron corporation Spectrometer S Series 711483 v 1.26at a selected wavelength of 253.7 nm. During themeasurements, the gaseous outow of the bioreactorwas connected directly to the detector and a 200 mL/minow rate of carrier gas (N 2) was used.

3. Results and discussion

3.1. Batch reactor operation

In order to evaluate the ability of the P. putida spi3strain to degrade thiomersal and its potential applica-tion in the remediation of thiomersal-contaminatedefuents the kinetics of thiomersal biodegradation wereinitially evaluated in a batch reactor (experiment A).The initial thiomersal concentration used was 25 mg/L,which is within the range of the expected thiomersalconcentration in the vaccine production efuent. Glu-cose was used as carbon source (2.5g/L) and themedium was supplemented with a concentrated micro-nutrient solution (200 mL micronutrients to 1 L of medium). Parallel kinetic experiments, with the same

initial glucose concentration but in the absence of

ARTICLE IN PRESS

Table 1Operation conditions (dilution rates, inow rates and concentrations) for the continuous bioreactor fed with vaccine productionefuent

D s (h1) F thiomersal

(mgL1 h

1)F glucose(mgL

1 h1)

F NH4 (mgN L

1 h1)

[thiomersal] in(mg/L)

[glucosel] i n(g/L)

[NH4] in(mgN/L)

0.022 0.60 25.2 0.22 25.20 1.16 12.00.03 0.93 38.9 0.31 30.97 1.30 10.380.05 1.55 38.9 0.35 33.10 0.83 7.450.1 3.60 38.9 1.25 36.86 0.40 12.81

R. Fortunato et al. / Water Research 39 (2005) 351135223514


5/12

thiomersal, were also carried out (experiment B). Theresults obtained are presented in Fig. 1(a) and (b) .

The rst observation is that the microbial culture isable to grow both in the presence and in the absence of thiomersal. Additionally, a decrease in the thiomersalconcentration was observed in experiment A (see Fig.1(a) ) and, after 3 h of operation, thiomersal was nolonger detected in the medium, suggesting that it hadbeen degraded by the microbial culture.

The degradation of thiomersal under abiotic condi-tions has been reported in the literature, namely whenexposed to sunlight ( The Merck Index, 1996 ), inaqueous solutions (like ophthalmic solutions) over longstorage periods ( Reader and Lines, 1983 ) and in thepresence of NaCl ( Caraballo et al., 1993 ) (which ispresent in relatively high concentrations in the culturemedium used). Thus, the hypothesis that the observeddecrease in the thiomersal concentration was not biotic,

but caused by spontaneous thiomersal degradation inthe growth medium used, could not be overruled.Therefore, to clarify this point, two blank experimentswere carried out. Thiomersal was added to the mediumunder the same operating conditions (150 rpm stirring,5 mg O 2/L, pH 7) but no microorganisms were added to

the reactor. Two reactors were operated simultaneouslyunder sterile conditions: one protected from light andthe other exposed to it. The results obtained showedthere was no signicant chemical degradation of thiomersal in the culture medium (less than 3%variation, which is within the HPLC average error)during observation time (24 h). Therefore, it wasconcluded that thiomersal biodegradation by P. Putidaspi 3 is the mechanism responsible for the observeddecrease in the thiomersal concentration.

The discontinuity in growth observed between 9 and10.5h was due to a limitation of micronutrients.Addition of a pulse of micronutrients (at 10.5h) to themedia caused a re-start of cell growth that ended onlywhen glucose was totally consumed. Therefore, it waspossible to conclude that the initial micronutrientsconcentration used was growth limiting, for the cellconcentration present in the reactor. As a result, in all

further experiments, the concentrated micronutrientssolution was added to the medium in the proportion of 400 mL micronutrients to 1 L of medium.

As can be observed in Fig. 1(a) , thiomersal is degradedduring the rst 3 h, after which there is a clear increase inthe growth rate. This behaviour is similar to thatobserved by Chang and Hong (1995) for the detoxica-tion of Hg 2+ from low-inoculum batch cultures of P.aeruginosa and would be expected considering thatthiomersal degradation also occurs through a microbialdetoxication mechanism. In this case, bacteria transformthe highly toxic thiomersal in a less toxic mercury form(Hg 0), which diffuses out of the cell, thus allowing thebacteria to grow using glucose as carbon source. As aresult, as long as the initial cell concentration is highenough to reduce thiomersal to sub-toxic levels before allthe cells are damaged, the maximum specic growth rate(mmax ) should be independent of the thiomersal concen-tration. In fact, as can be observed in Table 2 , similarspecic growth rates were obtained in the presence and inthe absence of thiomersal (0.40 and 0.41h

1, respec-tively). The maximum specic growth rates were calcu-lated as the maximum of the function m f t, using themeasured data (biomass concentration ( X ) versus time(t)) acquired in the period between 3 and 8.5 h. Theobserved growth yield in glucose ( Y x=s) was calculated forthe same time interval by plotting the experimentalbiomass concentration of each sample versus the glucoseconcentration of the same and calculating the slope of thestraight line obtained.

The observation that more oxygen was used inexperiment A than in experiment B, for the sameamount of substrate consumption (see the observedyield oxygen substrate Y O2/S ), suggests that the meta-bolic ux of glucose is different in the presence and inthe absence of thiomersal. In fact, the energy require-ments are higher (more oxygen is consumed) when

thiomersal is present in the medium.

ARTICLE IN PRESS

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 42 6 8 10 12 14 16

Operation time (h)

0 42 6 8 10 12 14 16

Operation time (h)

0

5

10

15

20

25

[ t h i o m e r s a l

] ( m g /

l )

OUR spglucosethiomersalcell dry weight

[ c e l

l d r y w e i g h

t ] , [

g l u c o s e ]

( g / l )

O U R

s p ( m g

0 2 / g

c e l l . m

i n )

0.0

0.5

1.0

1.5

2.0

2.5

3.0

[ c e l

l d r y w e i g h t

] , [ g l u c o s e

] ( g / l )

O U R

s p ( m g

0 2 / g

c e l l . m

i n )

(a)

(b)

Fig. 1. Evolution of the cell dry weight, glucose, specic oxygenuptake rate (OURsp) and thiomersal concentrations in a batchculture of P. putida spi3 (a) Experiment Athiomersal present inthe medium, (b) Experiment Bno thiomersal present in themedium. The arrows indicate the pulse addition of micronu-

trients.



6/12

Thiomersal and total mercury analysis of the super-natant samples showed that all mercury in the mediumwas in the form of thiomersal. Additionally, it wasobserved that the thiomersal and total mercury molaramounts decreased at the same rate and there was no

mercury accumulation in the medium. The questionraised at this point was whether the metallic mercuryformed remained inside the cells, or adhered to them, orwas released to the medium. In order to clarify this, cellsamples were collected along the experiment and the cellpellets were freeze-dried and digested with nitric acid for24h at 42 1 C. The total mercury concentration in thedigested samples and in the supernatant was measuredby ICP. The results obtained showed that, for a cellconcentration of 0.37g/L and a mercury concentrationin the supernatant of 0.2 mg Hg/L, the nal amountof mercury accumulated in the cells was 0.8 mg Hg /g

celldryweight. When we compare the total amount of

mercury transformed by the cells (calculated from theamount of degraded thiomersal: 25 mg/L thiomersalwhich is equivalent to 12.5mg Hg/L) with the amount of mercury accumulated by the cells (0.3 7 0.03mg Hg/L),the latter corresponds to only 2.5% of the former. Thisresult suggests that mercury does not signicantlyaccumulate in the biomass. Analysis of the bioreactorheadspace by cold vapour atomic adsorption spectro-scopy (CV-AAS) showed that metallic mercury isvolatilised to the headspace and from the headspace tothe bioreactor gaseous outow. The results obtained byCV-AAS corroborate those obtained by Fehr andWagner-Do bler (2000) and suggest that under homo-geneous culturing conditions (provided by a goodmixing) and strong aeration, the metallic mercuryformed is mainly volatilised from the medium and maybe recovered and re-utilised.

The volumetric thiomersal degradation rate ( r thiomersal ),calculated as the rst derivative of the function thatdescribes the decrease in the thiomersal concentration forthe initial points (from 0 to 1.5 h) obtained in experimentA was equal to 9.4 7 0.4mgL

1 h1 (initial glucose

concentration of 2.5 g/L). Replication of the experimentsproduced analogous results, and the average rate of

thiomersal degradation for a glucose and thiomersal

initial concentrations of 2.5 and 25 mg/L, respectively,was 7.9 7 0.6mgL

1 h1. The fact that the volumetric

thiomersal degradation rate is similar to that obtainedwhen the initial glucose concentration used was 1 g/L(r thiomersal 7.87 1.3mgL

1 h1), suggests that the rate

of thiomersal degradation is independent of the concen-tration of glucose used, for the same initial biomassconcentration, in the range of concentrations tested. Thisdoes not mean, however, that carbon source addition tothe medium is not necessary. The mechanism of Hg (II)detoxication, and hence that of thiomersal, requires asteady expression of the mer operon ( Chen and Chang,2000 ) and is therefore strongly dependent upon bacterialgrowth. As a result, maintaining a certain amount of viable and active cells is a necessary condition formaximizing thiomersal removal.

3.2. Continuous reactor operation

3.2.1. Biodegradation of thiomersal in a CSTR fed with asynthetic wastewaters

The thiomersal amended mineral medium was used toevaluate the culture behaviour, in continuous operationmode, for well-dened medium conditions. The thio-mersal concentration fed to the bioreactor was deliber-ately set to a value four times higher than that in theactual efuent. Therefore, it was possible to make surethat the residual thiomersal concentration was abovezero (thiomersal not limiting for the detoxicationprocess). The dilution rate ( D 0:05 h 1) was abouteight times lower than the maximum growth rateobtained in the batch studies ( mmax 0:4 h

1). Thethiomersal and glucose inow rates ( F thiomersal andF glucose ) were 9.6 and 51.5mgL

1 h1, respectively.

The system was initiated in a batch mode and switchedto continuous during the cells exponential growth phase.The results obtained are presented in Fig. 2 .

Steady state was attained after approximately 120 h of operation, i.e. after six hydraulic retention times (seeFig. 2 ). The average biomass concentration reached atsteady state was 1.03 7 0.09 g/L and the glucose con-centration in the bioreactor outlet was 0.40 7 0.04g/L.The biomass formation rate was equal to 0.053 7

0.004 g L 1 h1 and the glucose consumption rate was0.033 7 0.005 g L

1 h1 , which leads to an observed

growth yield in glucose ( Y x=s) of 1.6 7 0.1 gcell /gglucose .This growth yield is about 2.8 times higher than theaverage Y x=s obtained in the batch studies (see Table 2 ).The difference may be attributed to the fact that, whenthe bioreactor is operated in a continuous mode, theinitial micronutrient concentration in the medium ishigher, and their supply to the medium is continuous.Additionally, one should bear in mind that, when abioreactor is operated in a continuous mode there is aconstant removal of possible growth inhibitors and less

glucose is used for cell maintenance.

ARTICLE IN PRESS

Table 2Kinetic and stoichiometric parameters (maximum specicgrowth rate, observed growth yield in glucose and observedyield oxygen substrate) for a P. putida spi3 grown in thepresence (experiment A) and absence of thiomersal (experimentB)

Parameter Experiment A Experiment B

mmax (h1) 0.407 0.03 0.41 7 0.02

Y x=S gx =gS 0.57 7 0.03 0.63 7 0.02Y O2 =S gO2 =gS 0.30 7 0.02 0.18 7 0.01



7/12

The results also show that P. putida spi 3 was able toreduce the high thiomersal inow concentration (192mg/L) to a thiomersal outlet concentration below 5 mg/L.The volumetric and specic thiomersal degradation ratesat steady state were equal to 9.45 7 0.10mgL

1 h1 and

9.17 7 0.95m gthiomersal gcell1 h

1, respectively. The presenceof ethane in the bioreactor gaseous outow was detectedwhen a preliminary analysis of the gaseous phase wasperformed by gas chromatography, and suggests that themechanism of thiomersal degradation is, as mentioned inthe introduction, probably similar to that described formethyl mercury ( Begley et al., 1986 ).

In order to evaluate the capacity of the culture to dealwith shock loads, two different thiomersal pulses(instantaneous addition thiomersal to the bioreactor)were applied to the system at steady state (see Fig. 3 ).

After the rst pulse, the microorganisms were able todegrade the excess of thiomersal and return to the

thiomersal steady-state residual concentration within arelatively short time ( E 4 h). However, in the second

pulse, although is clear that the system also responded tothe thiomersal shock load and degraded part of theexcess thiomersal it did not return to the thiomersalsteady-state residual concentration within the observa-tion time.

When we plot the instantaneous thiomersal degrada-tion rate versus the thiomersal concentration for bothpulses (see Fig. 4 ) a decrease in the degradation rate isobserved as the thiomersal concentration decreases. Thedegradation rate was calculated from the massbalance: thiomersal in thiomersal out +thiomersal consumed thiomersal accumulated , where thiomersal accumulated(mgL

1 h1) was calculated as the 1st derivative of the

function that describes the experimental variation of thethiomersal concentration during the pulse ( Fig. 3 ). Thesystem response ( rthiomersal Versus thiomersal concentra-tion) after the rst pulse can be described (in a over-simplied way considering that rthiomersal depends only of

the thiomersal concentration) by the mathematical func-tion : r thiomersal k thiomersal n . Two lumped parameters

ARTICLE IN PRESS

0

20

40

60

80

100

326 328 330 332 334 336 338 3400

4

8

12

16

20

206 208 210 212 214 216 218 220

[ t h i o m e r s a

l ] ( m g

/ l )

Operation time (h) Operation time (h)

Pulse 1 Pulse 2

Fig. 3. Evolution of the thiomersal concentration, after thiomersal pulses, in a continuous culture of P. putida spi3 ( D 0:05 h 1) fed

with a synthetic wastewater (a) 20 mg/L thiomersal (b) 80 mg/L thiomersal.

0

5

10

15

20

25

30

0 20 40 60 80 100

[thiomersal] (mg/l)

r t h i o m e r s a

l ( m g

l - 1 h -

1 )

Pulse 1

Pulse 2

Fig. 4. Variation in thiomersal degradation rate with thethiomersal concentration in the two pulses experiment (Pulse120mg/L thiomersal and Pulse 280 mg/L thiomersal).

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0 60 120 180 240 300 360

Operation time (h)

0

5

10

15

20

25

30

35

40

[ t h i o m e r s a

l ] ( m g /

l )

[ c e l

l d r y w e i g h

t ] , [

g l u c o s e ]

( g / l ) thiomersalglucosecell dry weight

Fig. 2. Evolution of the cell dry weight, glucose and thiomersalconcentrations in a continuous culture of P. putida spi3 fed witha synthetic wastewater at a dilution rate D 0:05 h 1 .



8/12

are obtained: k 7.85 7 0.12h1, n 0.282 7 0.006, r2

0:99. The reaction order obtained ( n) is clearly differentfrom 1, suggesting that there is a complex dependencebetween the thiomersal concentration and the degradationrate. In fact, if a kinetic model was to be proposed,probably the contribution of other metabolic specieswould have to be considered.

Interestingly, the same mathematical equation de-scribes adequately the evolution of r thiomersal withthiomersal concentration in the beginning of pulse 2(see Fig. 4 , higher thiomersal concentrations, closedsymbols). However, it no longer describes the systemresponse as the thiomersal concentration decreases overtime. In fact, for concentrations lower than 60mg/L, thedegradation rate is lower in the second pulse experimentthan in the rst pulse, for equal thiomersal concentra-tions. Moreover, the difference between degradationrates is higher for the lower thiomersal concentration

region. These results suggest that, after a certain time of exposure to the pollutant (longer in pulse 2) there is adecrease in the thiomersal detoxication efciency,which may possibly be attributable to a partial andgradual loss of activity and, ultimately, cell viability.Chang and Law (1998) also observed a decrease in theextent of detoxication after 16 cycles of operation of afed-batch reactor for HgCl 2 removal.

The system response after the thiomersal peakssuggests that P. putida spi 3 possesses, to a certaindegree, the ability to adjust to the concentration of thiomersal in the inow and partially or totally degradethe excess thiomersal. In that case, the observation thatthe thiomersal concentration at steady state is abovezero may be attributed to the fact that, for the amountof biomass present, the steady-state concentrationreached (3 mg/L) is not toxic enough, to require theneed to induce the mer operon in more cells to eliminateit. Nevertheless, it was possible, using P. Putida spi 3, todetoxify a thiomersal amended mineral medium with aremoval efciency higher than 98%. The next step wasto evaluate whether the microbial culture was able togrow in real vaccine production wastewater andbiodegrade the thiomersal present in this efuent.

3.2.2. Biodegradation of thiomersal in a CSTR fed withvaccine wastewater

The thiomersal inlet concentration fed to the bioreactorvaried between 25 and 37 mg/L (see Table 1 ). A solutionwith 3.6 g/L of glucose was added separately to thebioreactor at a constant ow rate of 0.18mL/min, exceptfor D 0:1 h 1 . In the latter case, due to experimentalproblems with the glucose addition pump, there was acertain oscillation in the glucose-feeding rate that,however, had no signicant effect on the dilution rate.

The results obtained (data not shown) in thepreliminary experiments for D 0:03 h 1 with and

without the addition of an external ammonia source

showed that the microbial culture was able, for bothcases, to grow in the wastewater and biodegrade thethiomersal inow concentration of 31 mg/L to valuesbelow the detection limit for thiomersal (0.8 mg/L).Although in the absence of an external ammonia sourcethere was a large decrease in the biomass steady-stateconcentration (from 0.52 to 0.19g/L), thiomersal wasalways the limiting substrate in respect to the biomassconcentration. Additionally, removal efciencies of more than 99% were observed in both cases. Therefore,it was decided to operate the bioreactor withoutsupplementing the efuent with an external ammoniasource, thus improving the process economic feasibility.The results for D 0:022, 0.03, 0.05 and 0.1 h

1 aredepicted in Fig. 5 .

The rst observation is that thiomersal was neverdetected at the bioreactor outlet (detection lim-it 0.8mg/L), suggesting that, similarly to what had

been observed in preliminary experiments, thiomersalwas again limiting in respect to the biomass concentra-tion. As a result, the rate of thiomersal consumption wasassumed equal to the thiomersal-feeding rate. In orderto comply with the European Unions limit for mercurydischarge, the total mercury concentration in the efuentmust be below 50 mg Hg/L. Therefore, the total mercuryresidual concentration in the bioreactor was alsomeasured.

As can be observed in Table 3 , different thiomersalvolumetric degradation rates were obtained, for thesame cell concentration ( X 0:14 g=L), when thedilution rate was equal to 0.02 and 0.05h

1. This resultsuggest that, under the conditions tested, and as pointedout before, the thiomersal degradation rate is indepen-dent of the biomass concentration, if there are enoughcells in the medium transcribing the mer operon,necessary for thiomersal detoxication. This nding isin agreement with those described by Chang and Law(1998) , for the detoxication of HgCl 2 using a P.aeruginosa strain. They observed an increase in thevolumetric mercury detoxication rate (from 1 to1.94mg Hg L

1 h1), in chemostat operation, when they

increased the dilution rate (and consequently themercury-feeding rate) from 0.18 to 0.32 h

1, withoutobserving any alterations in the steady-state biomassconcentration.

As expected, the glucose residual concentrationincreased as the bioreactor was operated at higherdilution rates glucose D0:05 h14 glucose D0:03 h14glucose D0:022 h 1, with the exception of the valueobtained for D 0:1 h 1 . The signicantly lower re-sidual glucose concentration obtained for this dilutionrate was attributed to oscillations in the glucose-feedingrate due to experimental problems with the glucoseaddition pump.

The measured ammonia concentration in the bioreactor

outlet was always below the detection limit (detection

ARTICLE IN PRESS



9/12

limit 0.8mg N/L). As a result, cell growth was limitedby the ammonia concentration in the bioreactor. The rateof ammonia addition depended both on the dilution rateused and on the ammonia concentration in the wastewater(which varied slightly for the different efuent stock used).Despite the ammonia limitation thiomersal was comple-

tely removed.

Finally, from the observation of the residual mercuryconcentrations ( mg Hg/L) obtained for each operatingcondition tested, it becomes clear that it was possible, byadjusting the dilution rate (see Fig. 6 ), to reducethe mercury concentration in the bioreactor outlet([Hg] residual ) to values below the European limit for

mercury wastewaters discharges (50 mg Hg/L).

ARTICLE IN PRESS

Table 3Steady-state biomass and glucose concentrations, thiomersal consumption rate (volumetric and specic), total mercury residualconcentration and percentage of mercury removal in a CSTR fed with a vaccine production wastewater for different dilution rates

D (h1) r thiomersal (mgL

1 h1) X (g/L) r thiomersal sp (mgh

1 g1cell )

0.022 0.63 0.14 7 0.01 4.500.03 0.93 0.19 7 0.02 4.890.05 1.55 0.14 7 0.004 11.10.1 3.60 0.18 7 0.004 20.0

D (h1) Residual glucose (mg L

1) Residual Hg ( mg Hg L1) % Hg removal

0.022 31.4 7 9.6 35.3 7 0.5 99.83

0.03 55.2 7 14 1227 23 99.500.05 202.5 7 6.2 177 7 16 99.020.1 30.3 7 8.4 284 7 23 98.64

0 40 80 120 160 200 240 280 3200

60

120

180

240

300

360

D = 0.03 h -1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

110 120 130 140 150 160 170 1800

60

120

180

240

300

360

[ c e l

l d r y

w e i g h

t ] ( g / l )

0.0

0.1

0.2

0.3

0.4

0.5

0.6

[ c e l

l d r y

w e i g h

t ] ( g / l )

0.0

0.1

0.2

0.3

0.4

0.5

0.6

[ c e l

l d r y w e i g h

t ] ( g / l )

0.0

0.1

0.2

0.3

0.4

0.5

0.6

[ c e l

l d r y w e i g h

t ] ( g / l )

Operation time (h) Operation time (h)

Operation time (h)Operation time (h)

D=0.022 h -1

[ H g ]

( g / l

) [ g l u c o s e

] ( m g /

l )

0

60

120

180

240

300

360

[ H g ]

( g /

l ) [ g l u c o s e

] ( m g /

l )

[ H g ]

( g / l

) [ g l u c o s e

] ( m g /

l )

0

60

120

180

240

300

360

[ H g ]

( g /

l ) [ g l u c o s e

] ( m g /

l )

0 10 20 30 40 50 60

D = 0.1 h -1

0 20 40 60 80 100 120 140

D = 0.05 h -1

glucosetotal mercurycell dry weight

(a) (b)

(c) (d)

Fig. 5. Evolution of cell dry weight, total mercury and glucose concentrations in a continuous culture of P. putida spi3 fed with vaccinewastewater: (a) D 0:022h 1 , (b) D 0:03 h 1 , (c) D 0:05 h 1 , (d) D 0:1 h 1 .



10/12

The major disadvantage of operating the bioreactor ata lower dilution rate is the decrease in the volume of wastewater processed. Therefore, a compromise must bereached between the amount of wastewater treated perreactor volume and the desired wastewater quality at thebioreactor outlet.

A potentially interesting alternative would be toinclude a carbon adsorption column at the reactoroutow for polishing, similarly to what was proposed byWagner-Do bler (2003) for the bioremediation of HgCl 2containing wastewaters. In order to evaluate thispossibility, preliminary adsorption experiments using aNorit GAC-activated carbon were carried out. 200ml of mercury containing solutions (HgCl 2) in the concentra-tion range 0.150mg/L were contacted, for 48 h at roomtemperature, with 0.005 g of activated carbon. Mercurywas no longer detected in solution (total mercuryconcentration below the detection limit 5 mg/L) after

contact. These results showed that the carbon used has amercury removal capacity of at least 2000 mg Hg/gcarbon and may be used at the reactor outow forefuent polishing, thus allowing operation under higherdilution rates. Although it might be possible to use onlya carbon adsorption column to treat the efuent, thistype of approach would concentrate the pollutant,requiring a further treatment of the carbon. Thebiological thiomersal detoxication process is muchmore advantageous as in this cases the CHg bond iscleaved. The utilisation of an adsorption process should

therefore only be considered as a nal polishing step,in a situation where the operating costs, necessaryto attain mercury levels below the recommended ones,are too high.

The comparison of the bioreactor performance whenfed with the vaccine wastewater with that observed whenthe bioreactor was fed with the synthetic wastewater isalso worth of notice. Table 4 displays the resultsobtained for D 0:05 h 1 .

As mentioned above, no additional ammonia sourcewas added to the vaccine wastewater for D 0:05 h 1

and growth was limited by the ammonia concentrationfed to the bioreactor, while in the synthetic wastewaterammonia was clearly in excess. As a result, a muchhigher steady state cell concentration was reached whenthe synthetic wastewater was used. Furthermore, thethiomersal inlet concentration was 192 mg/L in thesynthetic wastewater, while in the case of the vaccine

wastewater (thiomersal inlet concentration equal to33.1mg/L), thiomersal was clearly limiting for thedetoxication process, which accounts for the lowervolumetric thiomersal degradation rate obtained in thelatter case. The values obtained for the specicthiomersal degradation rate are quite similar. In fact,when calculating the error associated with r thiomersal sp ,for the vaccine wastewater, one must take into accountnot only the errors associated with the biomassconcentration, but also those associated with thecalculation of the thiomersal degradation rate. As nothiomersal was detected in the reactor outlet, r thiomersalwas assumed equal to the thiomersal-feeding rate. Theerror associated with this value is due to variations in theow rate of the peristaltic pump used and was found tobe less than 5% for the considered ow rates. As aresult, the specic thiomersal degradation rate in thecase of the vaccine wastewater would be expected to bewithin the range 11.1 7 0.6mgh

1 g1cell .

4. Conclusions

The results obtained showed that the microbialculture used is capable of degrading the thiomersalpresent in a thiomersal amended mineral medium, bothin batch and continuous culture. When operated

ARTICLE IN PRESS

0

60

120

180

240

300

360

0.00 0.02 0.04 0.06 0.08 0.10 0.12

D (h -1

[ H g ]

r e s i

d u a l

( g

/ l )

Fig. 6. Variation of residual mercury concentration at thebioreactor outlet as a function of the dilution rate.

Table 4Thiomersal and glucose feeding rates, steady state biomass concentration and thiomersal consumptions rates (volumetric and specic)in two CSTR, one fed with a synthetic wastewater and the other with vaccine efuent, for D 0.05h

1

Feed F thiomersal(mgL

1 h1)

[thiomersal] in(mg/L)

F glucose(mgL

1 h1)

r thiomersal(mgL

1 h1)

r thiomersal sp(mgh

1 g1cell )

X (g/L)

Synthetic wastewater 9.6 192 51.5 9.5 7 0.1 9.2 7 1.0 1.03 7 0.04

Vaccine efuent 1.55 33.1 38.9 1.55 11.1 0.14 7 0.004



11/12

continuously, and despite a reduction in the detoxica-tion efciency, the bioreactor was able to deal with thetransient thiomersal shock loads applied and to degradepartially or totally the excess contaminant. The resultsobtained suggest that the thiomersal degradationprocess by P. putida spi3 is able to deal with uctuationsin the thiomersal inow concentration.

When the bioreactor was fed, in a continuous mode,with a real vaccine efuent contaminated with athiomersal concentration of 31 mg/L, the microbialculture was able to grow in the efuent, without theaddition of an external ammonia source. Thiomersalwas not detected in the bioreactor outow (thiomersaldetection limit equal to 0.8 mg/L) and mercury removalefciencies higher than 98.5% were observed. Addition-ally, it was possible, by adjusting the dilution rate, toobtain a wastewater with a mercury concentration(35.3 mg/L for D 0:022h 1) below the European limit

(50 mg/L) for mercury efuent discharges.In conclusion, the results obtained suggest that there

is a potential for the utilisation of P. putida spi3 in theremediation of thiomersal-contaminated wastewatersprovided that the bioreactor is coupled to an effectivesystem for the recovery of metallic mercury from thevapour outow stream.

Acknowledgements

R. Fortunato acknowledges Fundac - a o para a Ciencia ea Tecnologia (FCT, Portugal) for the research scholarshipPRAXIS XXI/BD/21618/99. The authors acknowledgefunding from the European Commission through projectno QLK3-1999-01213. The authors acknowledge Dr. Jose -Lu s Capelo-Mart nez for the cold vapour atomic adsorp-tion spectroscopy analysis of Hg 0.

References

Albiston, J., Turchi, C.S., Stewart, R.M., Broderick, T.E., 1999.Field Testing of Regenerable Sorbents for Recovery of Mercury from Flue Gas. ADA Technologies Inc., Houston,TX Presented at the Spring AIChE Meeting.

Barbieri, P., Bestetti, G., Reniero, D., Galli, E., 1996. Mercuryresistance in aromatic compounds degrading Pseudomonasstrains. FEMS Microbiol. Ecol. 20, 185.

Barkay, T., Miller, S.M., Summers, A.O., 2003. Bacterialmercury resistance from atoms to ecosystems. FEMSMicrobiol. Rev. 27 (23), 355.

Begley, T.P., Walts, A.E., Walsh, C.T., 1986. Mechanisticstudies of a protonolytic organomercurial cleaving enzyme:bacterial organomercurial lyase. Biochemistry 25 (22), 7192.

Caraballo, I., Rabasco, A.M., Ferna ndez-Are valo, M.,1993. Study of thimerosal degradation mechanism. Int.

J. Pharmaceut. 89, 213.

Chang, J.S., Hong, J., 1995. Estimation of mercury detoxica-tion from low-inoculum batch cultures of Pseudomonasaeruginosa PU21 (Rip64). J. Biotechnol. 42, 85.

Chang, J.S., Law, W.S., 1998. Development of microbialmercury detoxication processes using mercury-hyper re-sistant strain of Pseudomonas aeruginosa PU21. Biotechnol.

Bioeng. 57 (4), 462.Chang, J.S., Chao, Y.P., Law, W.S., 1998. Repeated fed-batch

operations for microbial detoxication of mercury usingwild-type and recombinant mercury-resistant bacteria. J.Biotechnol. 64, 219.

Chen, B.Y., Chang, J.S., 2000. Characterization and theoreticalanalysis on toxicological threshold of mercuric ions toPseudomonas aeruginosa PU21 (Rip 64). Bioprocess Eng.23, 675.

Drapeau, A.J., Laurence, R.A., Saint-Germain, G., Harbec,P.S., Lambert, N.G., 1983. Bioaccumulation of heavymetals by microorganisms. Sci. Tech. Eau 16, 359.

Elferink, J.G.R., 1999. A versatile sulphydryl reagent, calciummobilizer, and cell function-modulating agent. GeneralPharmacol. 33, 1.

European Commission, 2001. Joint Research Centre, IPTS,European IPPC Bureau. Reference Document on BestAvailable Techniques in the Chlor-Alkali ManufacturingIndustry.

Fehr, W., Wagner-Do bler, I., 2000. Microbial degradation of anorganic mercury compound (thiomersal). In: Proceedings of the Biotechnology for Environmental Applications Meeting.

Keith, L.H., Walters, D.B., 1992. The National ToxicologyPrograms Chemical Data Compendium. Lewis Publishers,Boca Raton, FL.

Magos, L., 2003. Neurotoxic character of thimerosal and theallometric extrapolation of adult clearance half-time to

infants. J. Appl. Toxicol. 23, 263.Misra, T.K., 1992. Bacterial resistances to inorganic mercurysalts and organomercurials. Plasmid 27, 4.

Nelson, J.H., Hendrix, J.L., Milosavjevic, E., 1987. Use of thiourea and thioacetamide for separation and recovery of heavy metals from mineral treatment wastewaters. In:Patterson, J.W., Passino, R. (Eds.), Metal Speciation andRecovery.pp. 119.

Noyes, O.R., Hamdy, M.K., Muse, L.A., 1976. Control of mercury pollution. Toxicol. Environ. Health 1, 409.

Pichichero, M.E., Cernichiari, E., Lopreiato, J., Treanor, J.,2002. Mercury concentrations and metabolism in infantsreceiving vaccines containing thiomersal: a descriptivestudy. Lancet 360, 1737.

Pinney, R.J., 1978. Survival of plasmid containing strains of E.coli , Pseudomonas aeruginosa and Staphylococcus aureus inphenylmercuric nitrate and thiomersal. J. Pharmacy Phar-macol. 30, 228.

Reader, M.J., Lines, C.B., 1983. Decomposition of thimerosalin aqueous solutions and its determination by high-performance liquid chromatography. J. Pharmaceut. Sci.72 (12), 1406.

Schiering, N., Kabsch, W., Moore, M.J., Distefano, M.D.,Walsh, C.T., Pai, E.F., 1991. Structure of the detoxicationcatalyst mercuric ion reductase from Bacillus sp. strainRC607. Nature 352 (6331), 168.

The Merck Index, 1996. Safety MSDS data for thiomersal. 12th

ed. vol. 1. pp. 1590, #9451.

ARTICLE IN PRESS



12/12

von Canstein, H., Li, Y., Timmis, K.N., Deckwer, W.D.,Wagner-Do bler, I., 1999. Removal of mercury fromchloralki electrolysis wastewater by a mercury resistantPseudomonas putida strain. Appl. Environ. Microbiol. 65(12), 5279.

Wagner-Do bler, I., 2003. Pilot plant for bioremediation of

mercury-containing industrial wastewater. Appl. Microbiol.Biotechnol. 62, 124.

Wagner-Do bler, I., Canstein, H., Li, Y., Timmis, K.N.,Deckwer, W.D., 2000a. Removal of mercury from chemicalwastewater by microorganisms in technical scale. Environ.Sci. Technol. 34, 628.

Wagner-Do bler, I., Lu nsdorf, H., Lu bbehu sen, T., Canstein,H., Li, Y., 2000b. Structure and species composition of

mercury-reducing biolms. Appl. Environ. Microbiol. 66(10), 4559.

ARTICLE IN PRESS


bio degradation of thiomersal containing effluents by a mercury resistance pseudomonas putida strain

Documents