Functional Display of ForeignProtein on Surface of Escherichia coli UsingN-Terminal Domain of Ice Nucleation Protein

Lin Li,1,2 Dong Gyun Kang,1,3 Hyung Joon Cha1,3

1Division of Molecular and Life Sciences, Pohang University of Science andTechnology, Pohang 790-784, Korea; telephone: +82-54-279-2280;fax: +82-54-279-2699; e-mail: hjcha@postech.ac.kr2Key Laboratory of Agricultural Microbiology, School of Life Science andTechnology, Huazhong Agricultural University, Wuhan 430070, P.R. China3Department of Chemical Engineering, Pohang University of Science andTechnology, Pohang 790-784, Korea

Received 29 April 2003; accepted 29 September 2003

Published online 10 December 2003 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/bit.10892

Abstract: We investigated the ability of the N-terminal do-main of InaK, an ice nucleation protein from Pseudomonassyringae KCTC1832, to act as an anchoring motif for thedisplay of foreign proteins on the Escherichia coli cell sur-face. Total expression level and surface display efficiencyof green fluorescent protein (GFP) was compared followingtheir fusion with either the N-terminal domain of InaK (InaK-N), orwith the known truncated InaK containing bothN- andC-terminal domains (InaK-NC). We report that the InaK-N/GFP fusion protein showed a similar cell surface display effi-ciency (f50%) as InaK-NC/GFP, demonstrating that theInaK N-terminal region alone can direct translocation of for-eign proteins to the cell surface and can be employed as apotential cell surface display motif. Moreover, InaK-N/GFPshowed the highest levels of total expression and surfacedisplay based on unit cell density. InaK-Nwas also success-ful in directing cell surface display of organophosphorushydrolase (OPH), confirming its ability to act as a displaymotif. B 2004 Wiley Periodicals, Inc.

Keywords: cell surface display; ice nucleation protein; N-terminal domain; green fluorescent protein; organophos-phorus hydrolase; Escherichia coli

INTRODUCTION

Bacterial cell surface display of heterologous proteins can

be useful in procedures such as development of live vac-

cines and multiple antigen antisera (Lee et al., 2000), con-

struction and screening of protein libraries (Georgiou et al.,

1997), whole cell bioconversion and biocatalysis (Richins

et al., 1997), and development of environmental bioadsorb-

ents (Sousa et al., 1996). Various display systems have been

developed with different anchoring motifs. For example,

outer membrane proteins (Sousa et al., 1996; Richins et al.,

1997; Mejare et al., 1998), lipoproteins (Francisco et al.,

1993; Taylor et al., 1990), subunits of cellular appendix

proteins (Newton et al., 1995; Sleytr and Sara, 1997), and

some secretory proteins (Jung et al., 1998a; Kornacker and

Pugsley, 1990) have been used in Gram-negative bacteria,

while staphylococcal protein A has been used in Gram-pos-

itive bacteria (Hansson et al., 1992), and S-layer protein in

archaeobacteria (Sleytr and Sara, 1997). One such anchoring

motif is the glycosylphosphatidylinositol (GPI)-anchored se-

cretory protein, ice nucleation protein (INP), that is generated

from Pseudomonas syringae and some other Gram-negative

strains. INP’s advantages include stable expression and outer

membrane translocation and modulatable length of internal

repeating units (Jung et al., 1998b).

INP is a membrane-bound protein which confers on host

cells the ability to nucleate crystallization in supercooled

water (Mararitis and Bassi, 1991). It has potential biologi-

cal applications in many fields, including the food industry,

spray-ice technology, and other biotechnological processes

(Mararitis and Bassi, 1991; Cochet and Widehem, 2000). It

is composed of three domains structurally distinguished as

the N-terminal domain (191 amino acids, 15% of the pro-

tein), which is the portion most responsible for targeting to

the cell surface, the C-terminal domain (49 amino acids, 4%

of the protein), and the central domain, composed of repeats

comprising an 8-, 16-, and 48-residue periodicity that acts as

a template for ice crystal formation (Kozloff et al., 1991;

Schimid et al., 1997). Of the several identified INPs, two

have been used as anchoring motifs for display of foreign

proteins on host cell surfaces, namely, InaK fromP. syringae

KCTC1832 (Jung et al., 1998a) and InaV from P. syringae

INA5 (Schimid et al., 1997). This was achieved using either

full-length sequences (Jung et al., 1998a; Lee et al., 2000) or

B 2004 Wiley Periodicals, Inc.

L. Li and D.G. Kang contributed equally to this work.

Correspondence to: Hyung Joon Cha Korea Science and Engineering

Foundation (KOSEF) through the Center for Traditional Microorganism

Resources (TMR); Brain Korea 21 program issued from the Ministry of

Education, Korea

truncated portions containing only N- and C-domains (INP-

NC) (Shimazu et al., 2001a), INP-NC with five additional

internal repeating units (Jung et al., 1998b; Lee et al., 2000),

or N-domain (INP-N) with two additional internal repeating

units (Lee et al., 2000). Since full-length INP is quite large

(1,200–1,500 amino acid residues), functional truncated INP

moleculesmay serve as better anchoringmotifs to carry large

heterologous proteins. Therefore, experiments investigating

the properties of functional truncated INP motifs are impor-

tant for rapid progress in the field of cell surface display.

It is not yet clear which INP domain is essential for

anchoring to the outer cell membrane, since no clear signal

(targeting) sequence has been identified, and the secretion

mechanism for INP is still unknown (Georgiou et al., 1997;

Schmid et al., 1997). Truncated InaK consisting of the N-

and C-terminal domains plus the first two and last three

internal repeating subunits was able to direct the expressed

CMCase fusion protein to the cell surface without apparent

loss of enzyme activity, whereas enzyme activity was lost

when full-length InaK was used (Jung et al., 1998b).

However, the similar InaK-based INP-NC/OPH (organo-

phosphorus hydrolase) fusion protein showed poor surface

localization when compared to Lpp-OmpA or an InaV-based

INP-NC/OPH fusion (Shimazu et al., 2001a).

In the present work, we investigated whether a truncated

INP fragment consisting of only the N-terminal domain

could serve as a potential anchoring motif. A truncated InaK

N-terminal domain (InaK-N) containing the first two inter-

nal repeating units was previously shown to be capable of

displaying viral antigens on the surface of Salmonella sp.

(Lee et al., 2000). However, this was a just minor portion of

their entire work; it did not compare the efficiency and levels

of cell surface display by portions of the InaK-N motif. In

addition, the fusion protein in the previous report did not

have the predicted size on Western blot analysis, suggesting

that the construct could be flawed. Therefore, we sought

to specifically investigate the anchoring ability of the N-ter-

minal region of InaK by comparing the cell surface display

efficiency of InaK-N with that of InaK containing both the

N- and C-terminal domains (InaK-NC).

In these studies, the truncated InaK proteins were fused to

versatile reporter green fluorescent protein (GFP) as a target

model foreign protein, because this noninvasive biological

marker is quite stable and durable in various physical and

chemical processes (Chalfie et al., 1994), and its fluores-

cence is easily quantifiable (Cha et al., 2000). GFP has been

successfully displayed on the E. coli cell surface using Lpp-

OmpA hybrid motif (Shi and Su, 2001). In addition, we also

fused OPH to the truncated InaK proteins to confirm surface

display ability of InaK-N. OPH from Flavobacterium sp.

(Mulbry and Karns, 1989) is a homodimeric organophos-

photriesterase that can degrade a broad spectrum of toxic

organophosphates (Grimsley et al., 1997) that are widely

used in many pesticides and chemical nerve agents (Donar-

ski et al., 1989). This enzyme can hydrolyze various phos-

phorus-ester bonds including P-O, P-F, P-CN, and P-S bonds

(Lai et al., 1995).

MATERIALS AND METHODS

Bacterial Strains, Plasmids, and Culture Condition

Escherichia coli TOP10 [F� mcrA �(mrr-hsdRMS-mcrBC)

&80lacZ�M15� lacX74 deoR recA1 araD139�(ara-leu)

7697 ga1U ga1K rpsL (StrR) endA 1 nupG]

(Invitrogen, Carlsbad, CA) was used for constructing

recombinant plasmids. Escherichia coli W3110

[F�E�mcrAmcrB IN(ornD-ornE)] was used as a host for

surface display of InaK. Plasmid pMPL003 (unpubl.

plasmid), bearing a full-length inaK gene that amplified

from genomic DNA of P. syringae KCTC1832, was used as

a template for further PCR manipulation. Plasmid pGFPuv

(Clontech, Palo Alto, CA), carrying the gfpuv gene,

provided the gene source for construction of gfp deriva-

tives. Plasmid pTG (Kang et al., 2002a), which can express

cytosolic GFP, was used as a negative control in either

expression or localization assays. Plasmid pTO (Kang et al.,

2002b), which carries the opd gene for expression of

cytosolic OPH, was also used for gene source or as a nega-

tive control for confirmation experiments. Plasmid

pTrcHisC (Invitrogen) was used as a parent vector for

construction of truncated inaK hybrids.

Recombinant strains bearing plasmids were inoculated

in M9 medium (12.8 g/L Na2HPO4.7H2O, 3 g/L KH2PO4,

0.5 g/L NaCl, 1 g/L NH4Cl, 3 mg/L CaCl2, 1 mM MgSO4)

containing 0.5% (wt/vol) glucose and 50 Ag/mL of

ampicillin at the final concentration. Cells were cultured in

250-mL Erlenmeyer flasks with a 50-mL working volume at

250 rpm and 37jC. When the cultures were grown to an

optical density of 0.6 (at 600 nm, OD600), 0.05, 0.1, 0.5, or

1 mM (as final concentration) isopropyl-h-D-thiogalacto-pyranoside (IPTG) (Sigma, St. Louis, MO) was added to the

culture broth for induction of recombinant protein expres-

sion. After IPTG induction, cells were grown at 250 rpm and

25jC for 24 h.

Plasmid Construction

Recombinant plasmids harboring various inaK hybrids were

constructed as follows (Fig. 1). To construct full-length

inaK-gfp fusion, PCR was performed to generate a 3,559 bp

of inaK encoding fragment without termination codon by

using primers inaK-N 5V and inaK-C 3V from the plasmid

pMPL003 as a template. The amplified fragment was di-

gested with NcoI and BglII, and then inserted into the same

sites of plasmid vector pTrcHisC, resulting in plasmid

pMPL006. A 724 bp of gfpuv fragment amplified with prim-

ers gfpuv 5V and gfpuv 3V from pGFPuv and digestedwithBglII

and HindIII was further introduced into the same sites of

pMPL006 to give plasmid pINPF-GFP, which harbors ‘‘full-

length inaK-gfpuv’’ fusion under the control of trc promoter.

Primers inaK-N 5V and inaK-N 3Vwere used to amplify theN-

terminal domain of inaK (inaK-N) from the pINPF-GFP, and

primers inaK-C 5V and gfpuv 3V were used to amplify the

fusion fragment with C-terminal domain (inaK-C) and gfpuv

LI ET AL.: CELL SURFACE DISPLAY USING N-DOMAIN OF INP 215

coding sequence from the plasmid pINPF-GFP. The 552 bp

of inaK-N amplified fragment was cleaved with NcoI and

Bgl II and ligated into the same enzyme-digested pTrcHisC,

resulting in pINPN. Then the combinations of ‘‘inaK-N/

gfpuv’’ and ‘‘inaK-NC/gfpuv’’ were carried out by intro-

ducing a 724 bp of BglII andHindIII-digested fragment from

the pINPF-GFP into the same sites of pINPN; or the

amplified ‘‘inaK-C/gfpuv’’ fragment digested with BamHI

and HindIII into the BglII and HindIII digested-pINPN. The

resulting plasmids were designated pINPN-GFP and

pINPNC-GFP, respectively. The plasmids INPN-OPH and

INPNC-OPH were constructed by introducing a 1,020 bp

fragmentwhichwas amplifiedwith primers opd 5V and opd 3Vfrom the pTO and digested with BglII and HindIII into the

sample site of pINPN-GFP and pINPNC-GFP, respectively.

All primers used are listed in Table I.

Analytical Assays

Cell density (OD600) was measured at 600 nm on a UV/VIS

spectrophotometer (UV-1601PC; Shimadzu, Kyoto Japan).

After 24 h culture upon IPTG induction, cells were harvested

and diluted to unit cell density (OD600 = 1) with PBS buffer

(pH 7.5) and the similarly dilutedE. coli recipients were used

as background references. GFP fluorescence intensity was

determined using a fluorescence spectrophotometer (RF-

5301PC, Shimadzu) at an excitation of 395 nm and emission

of 509 nm. OPH activity was measured by following the

increase in absorbance of p-nitrophenol from the hydrolysis

of substrate (1 mM Paraoxon (Sigma)) at 400 nm (e400 =

17,000 M�1cm�1) using the UV/VIS spectrophotometer.

One unit of OPH activity was defined as lmoles Paraoxon

hydrolyzed per min (Caldwell et al., 1991).

Cell Fractionation

Cell fractionation was performed according to the method

described in Shi and Su (2001). Cells harboring inaK-gfpuvhybrids were induced with 0.1 mM IPTG and cultured at 25C

for 24 h. Harvested cells were diluted to set as unit cell

density (OD600 = 1), washed, and resuspended in PBS buffer

containing 1mMEDTAand lysozyme at 10 Ag/mL.After 2 h

Figure 1. Gene maps of recombinant plasmids harboring truncated (A) inaK/gfpuv and (B) inaK/opd fusion constructs. Plasmid pTrcHisC was used as a

parent vector for constructing these fusions. Ptrc, trc promoter; inaK, ice nucleation protein gene; inaK-N, N-terminal domain of inaK; inaK-C, C-terminal

domain of inaK; gfpuv, UV-optimized green fluorescent protein gene; opd, organophosphorus hydrolase gene; term, termination sequence.

Table I. Primers used for construction of the recombinant plasmids.

Primer Restriction site Sequence (5V!3V)

inp-N 5V NcoI TGCTGCCATGGCTCTCGACAAGGCGTTGG

inp-N 3V BglII TAAGATCTGGTCTGCAAATTCTGCGGCGTCGTCACCGG

inp-C 5V BamHI GCAGGATCCAGACTCTGGGACGGGAAG

inp-C 3V BglII GAAGATCTTACCTCTATCCAGTCATCGTCCTCG

gfpuv 5V BglII GCAGATCTAGTAAAGGAGAAGAACTTTTC

gfuuv 3V HindIII CGAAGCTTTCATTATTTGTAGAGCTCATC

opd 5V BglII GGAGATCTGGATCGATCGGCACAGGC

opd 3V HindIII GGAAGCTTTCATGACGCCCGCAAGGTCG

216 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 85, NO. 2, JANUARY 20, 2004

incubation, cell suspension was treated with an ultrasound

sonication at 30 sec � 2 cycles. To obtain total membrane

fraction, whole cell lysate was pelleted by centrifugation

at 39,000 rpm for 1 h using an ultracentrifuge (Optima LE-

80K; Beckman, Fullerton, CA). The supernatant was

regarded as soluble cytoplasmic fraction. For further outer

membrane fractionation, the pellet (total membrane fraction)

was resuspended with PBS buffer containing 0.01 mM

MgCl2 and 2% Triton X-100 for solubilizing inner mem-

brane and incubated at room temperature for 30 min, and

then the outer membrane fraction was repelleted by ultra-

centrifugation. Equal volumes of each fractionated sample

were saved for further analyses; GFP fluorescence intensity,

OPH activity, and Western blotting.

Western Blot Analysis

An equal volume of each fraction (cytoplasmic and outer

membrane) of the cells containing inaK/gfpuv hybrids were

mixed with SDS sample buffer (10% sodium dodecyl sulfate

(SDS), 10% h-mercaptanol, 0.3M Tris-HCl (pH 6.8), 0.05%

bromophenol blue, 50% glycerol), boiled for 5 min, and

resolved by 12.5% (wt/vol) SDS-polyacrylamide gel electro-

phoresis (SDS-PAGE), followed by electrophoretic transfer

to Hybond-PVDF membranes (Amersham Pharmacia Bio-

tech, Buckinghamshire, UK) with transfer buffer (48 mM

Tris-HCl, 39 mM glycine, 20%methanol, pH 9.2) by using a

Trans-Blot SD Cell (Bio-Rad, Hercules, CA) at 15V for

30 min. After blocking for 1 h in TBS buffer (20 mM Tris-

HCl, 500 mM NaCl, pH 7.5) containing 5% (wt/vol) nonfat

dry milk, the membrane was then incubated for 1.5 h at room

temperature in antibody solution (1% (wt/vol) nonfat dry

milk in TTBS (TBS with 0.05% Tween-20)) containing

monoclonal anti-GFP antibody (1:1,000 vol/vol) (Roche,

Basel, Switzerland) and probed with secondary antimouse

IgG conjugated with alkaline phosphatase (1:1,000 vol/vol)

(Sigma). After successive washing with TTBS and TBS,

FAST Fast Red TR/Napthol AS-MX (Sigma) was added to

detect and the reaction was quenched with distilled water.

Immunolabeling Assays

For immunolabeling examination of intact cells, cells har-

boring inaK-gfpuv hybrids were induced with 0.1 mM IPTG

and cultured at 25C for 24 h, then harvested and washed

4 times with PBS buffer. Cells were then incubated at 4C

overnight in anti-GFP antibody solution and in secondary

antimouse antibody solution for 2 h. After washing three

times with PBS, the cells were also incubated in FAST so-

lution for 10 min to let the color develop.

Protease Accessibility Assays

Cells harboring inaK-gfpuv hybrids were induced with

0.1 mM IPTG and cultured at 25C for 24 h. Cells were

then harvested and washed three times with PBS buffer, then

adjusted to OD600 of 10. Pronase (f4 units/mg; Sigma) was

added to a final concentration of 2 mg/mL. Cell suspensions

were incubated at 37C and GFP fluorescence intensities of

pronase-treated samples were measured at each 1-h interval.

RESULTS AND DISCUSSION

Construction and Expression of InaK/GFP Fusions

Recombinant plasmids pINPN-GFP and pINPNC-GFP

that encode InaK-N/GFP and InaK-NC/GFP, respectively,

were constructed (Fig. 1). The binary and tripartite

hybrids were created under the trc promoter-operator region

andwere modulated at the whole encoding frame; all hybrids

presented with predicted sizes. All transformed strains were

found to be fluorescent upon UV illumination, or by fluores-

cence microscopic observation, easily indicating that GFP

variants were functionally expressed when fused to trun-

cated InaK (data not shown).

Specific expression levels of each GFP sample were

determined by monitoring whole cell fluorescence intensity

based on unit cell density (OD600 = 1). Fluorescence

was present in the exponential phase, and increased as

cells reached the stationary phase (data not shown),

possibly reflecting the time required for full translocation

of InaK/GFP fusions to the cell surface (Shimazu et al.,

2001b). Therefore, fluorescence intensities of cells collected

at the very late stationary phase (24 h) were used for analysis.

When 1 mM IPTG was used for induction, both truncated

InaK/GFP fusion proteins exhibited much lower specific

GFP fluorescence than that generated by control cytosolic

GFP expression constructs (Fig. 2). In the context of ex-

pression and secretion of a protein, secretion is generally the

limiting step; thus, a high transcription rate can block the

translocation pathway and cause growth inhibition (Rodri-

gue et al., 1999; Shi and Su, 2001). High doses (1 mM) of

IPTG will likely induce a high transcription rate; however,

the large amounts of GFP protein thus produced may not be

Figure 2. Total GFP expression levels in the cells harboring control cy-

tosolic GFP, InaK-N/GFP, and InaK-NC/GFP fusions under induction with

two different concentrations (0.1 and 1 mM) of IPTG. Cells were grown in

M9 media at 25jC for 24 h upon IPTG induction.

LI ET AL.: CELL SURFACE DISPLAY USING N-DOMAIN OF INP 217

efficiently translocated onto the cell surface. This results in

low total fluorescence, as the whole-cell fluorescence of

cytosolic GFP is much lower than that of surface-displayed

GFP due to the barrier effect of the cell membrane. Actually,

we found that densities of both surface-displayed cells were

much lower under the 1 mM IPTG condition than those

under 0.1 mM (data not shown). However, when cells are

treated with low (0.1 mM) doses of IPTG, it is more likely

that transcription and secretion will be balanced, allowing

efficient display of GFP proteins and high whole-cell fluo-

rescence. Also, it was reported that secretion might increase

the total expression level (Miksch et al., 1997); by emptying

the intracellular compartments, secretion desaturates them

of the secreted protein, thus allowing more synthesis. This

can be a possible explanation for higher fluorescence of sur-

face-displayed GFP expression than cytosolic GFP expres-

sion driven by the same promoter under 0.1 mM IPTG.

Because low levels of GFP whole-cell fluorescence were

observed in the 1 mM IPTG condition, we tested the effect of

final IPTG concentrations of 1.0, 0.5, 0.1, and 0.01 mM.

Under these conditions, the final fluorescence intensity of

the control cytosolic GFP construct was reduced, while the

two InaK/GFP fusion constructs showed increased GFP

surface display, with the highest whole-cell fluorescence

observed following treatment with 0.1 mM IPTG (data not

shown). For both fusion proteins, >3-fold higher GFP fluo-

rescence intensities were obtained using 0.1 mM IPTG

compared to using 1 mM IPTG (Fig. 2). Interestingly, cells

expressing the InaK-N/GFP fusion protein showed the high-

est fluorescence, which was 1.8-fold greater than cells ex-

pressing the InaK-NC/GFP fusion protein (Fig. 2).

Approximately 50% of total fluorescence was found in the

total membrane fraction of cells expressing InaK-N/GFP and

InaK-NC/GFP (Fig. 3). However, the control strain express-

ing the cytosolic GFP chromophore showed about 5% of

total fluorescence in the total membrane fraction. These

differences between cell surface-displayed and cellular-ex-

pressed strains suggest correct localization of fusion proteins

on the outer membrane of host cells. Importantly, InaK-N/

GFP showed almost identical cell surface display efficiency

as InaK-NC/GFP, demonstrating that the InaK N-terminal

domain alone can act as an anchoring motif to direct

translocation of target heterologous proteins to the host cell

surface. Also note that because the fusion constructs had

similar cell surface display efficiencies for GFP, we could

use whole-cell fluorescence intensity alone to measure and

compare surface display capabilities of anchoring motifs.

It was reported that InaK-N with two repeating subunits

(perhaps functionally equivalent to the N-terminal domain

alone) allowed the display of foreign viral antigens on the

surface of Samonella cells (Lee et al., 2000). However, this

previous work did not compare the efficiency and levels of

cell surface display by portions of the InaK-N motif, and the

fusion protein in the previous report was much smaller than

expected onWestern blot analysis, suggesting that the InaK-

N fusion construct might be flawed.

Surface Localization Analysis of InaK-N/GFP Fusion

For successful use of InaK derivatives as anchoring motifs

for target heterologous proteins, it is essential to verify the

location and activity of expressed foreign proteins, and to

determine cell surface display efficiencies of InaK/GFP

fusion proteins. In order to achieve this, subcellular frac-

tionation was performed to separate equal volumes of the

outer membrane and soluble cytoplasmic fractions. These

fractions were then subjected to Western blot analysis

with an anti-GFP antibody (Fig. 4). Most InaK/GFP proteins

were found in the outer membrane fraction for both fusion

constructs, while almost all GFPs from control ‘‘cytosolic

expression’’ cells were found in the soluble cytoplasmic frac-

tion. Immunoblotting showed InaK-N/GFP bands as thicker

than InaK-NC/GFP bands because total expression of the

former was much higher (recall Fig. 2). While the previously

reported InaK-N fusion protein was smaller than predicted

(Lee et al., 2000), the molecular weight of the InaK-N/GFP

fusion protein in our work was almost identical to the pre-

Figure 3. Percentage of total membrane fractions in GFP fluorescence

intensities of the cells harboring control cytosolic GFP, InaK-N/GFP, and

InaK-NC/GFP fusions. Cells were grown in M9 media at 25jC for 24 h

upon 0.1 mM IPTG induction. Analyses were based on unit cell density

(OD600 = 1).

Figure 4. Western blot analysis for cell fractionations of the cells

harboring control cytosolic GFP, InaK-N/GFP, and InaK-NC/GFP fusions.

CP, cytoplasmic fraction; OM, outer membrane fraction. Cells of unit

density (OD600 = 1) were fractionated into equal volumes of cytoplasmic

and outer membrane fractions, and an equal volume of each fraction was

loaded into each lane. Anti-GFP antibody was used for the assay.

218 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 85, NO. 2, JANUARY 20, 2004

dicted 46 kDa, demonstrating the proper construction of our

InaK-N fusion. However, InaK-NC/GFP showed a higher

molecularweight than the predicted 51kDaand the reason for

this is currently unknown. The fusion proteins detected in the

cytoplasmic fractions showed signs of extensive proteolysis

(Fig.4).However,when fusionproteinsweredisplayedon the

cell surface they showed relatively good structural stability.

These data suggest better cell surface display efficiencies

might be achieved if fusion proteins could be protected from

proteolysis in the cytoplasm. Interestingly, for InaK-N/GFP,

the smallest degradation product band was the same size as

GFPalone(f27kDa),while thesmallestdegradationproduct

ofInaK-NC/GFPwasabout5kDabigger, suggestingitmaybe

GFP conjugated with the C-terminal domain. From these

results, we can surmise that the N-terminal of InaK is sus-

ceptible to protease attack and/or is structurally unstable,

while the C-terminal is not. We also note that degradation of

the N-terminal region resulted in free GFP being detected in

theoutermembrane fractionof cells expressing InaK-N/GFP,

but such free GFP was not detected in the outer membrane

fraction of cells expressing InaK-NC/GFP (Fig. 4).

Immunolabeling assays can indicate specific interactions

between a target antigen protein and an antibody. When

added exogenously to intact transformed cells, anti-GFP

antibodies cannot diffuse through the cell membrane; there-

fore, the location of fusion proteins on the cell surface can be

revealed. We observed that cells expressing either InaK

hybrid stained very red (positive), with the strongest signal

observed for InaK-N/GFP, the best surface-displayed prod-

uct (Fig. 5). In contrast, only weak staining was observed in

the cytosol-expressing control strain. That there was any

staining in these control cells may be due to trace amounts of

GFP escaping any broken cells, or to nonspecific binding of

the anti-GFP antibody. These results confirm that both InaK

constructs worked efficiently as cell surface display motifs.

Macromolecules cannot penetrate the outer membrane,

and GFP is resistant to many common proteases except

pronase, which is a mixture of broad-specificity proteinases

(Bokman and Ward, 1981). Therefore, a pronase accessi-

bility assay can be used to provide evidence for the surface

localization of GFP (Shi and Su, 2001). Accordingly, we

performed a pronase accessibility assay to confirm the cell

surface display of GFP by the two fusion constructs (Fig. 6).

While GFP fluorescence intensity decreased only slightly

(about 10%) in pronase-treated control cells (which ex-

pressed GFP in the cytosol), the intensity decreased more

than 40% in pronase-treated cells expressing either of the

InaK fusions (Fig. 6). This confirms that both truncated InaK

motifs are capable of inducing cell surface display of GFP.

From the similar decline profiles of GFP fluorescence in

cells harboring the two InaK/GFP hybrids, we could also

confirm that both truncated InaK motifs have similar cell

surface display efficiencies (recall Fig. 3).

Among the three distinct INP structural domains, it was

proposed that the N-terminal domain is responsible for

targeting to the outer membrane via a glycosylphosphati-

dylinositol (GPI) anchor (Kozloff et al., 1991), or by

phosphatidylethanolamine, which has been proposed to be

more involved than phosphatidylinositol (Palaiomylitou

et al., 1998). The internal repeating domain acts as a tem-

plate for ice nucleation formation, and also possibly for

targeting to the membrane through O-glycan linkage with

sugar residues (Kozloff et al., 1991). The C-terminal region

is highly hydrophilic and extends out of the membrane, and a

proposed hydrophobic tether permits energetically easy mi-

gration through hydrophobic cell membranes (Kozloff et al.,

1991). We have successfully demonstrated in this study that

the truncated InaK fragment containing only the N-termi-

nal domain can really direct cell surface localization of target

proteinswith similar efficiency as the known InaK-NCmotif,

suggesting that a secretion signal might exist in the N- ter-

minal domain.

Notably, GFP expression levels were greatest for the

InaK-N/GFP fusion, and this was reflected in the amount

of GFP displayed on the cell surface, which was signifi-

Figure 5. Immunolabeling analysis of the intact cells harboring control

cytosolic GFP, InaK-N/GFP, and InaK-NC/GFP fusions. Cells were grown

in M9 media at 25jC for 24 h upon 0.1 mM IPTG induction. Anti-GFP

antibody was used for the analysis.

Figure 6. Pronase accessibility analysis for GFP fluorescence intensity

of the intact cells harboring control cytosolic GFP, InaK-N/GFP, and

InaK-NC/GFP fusions. Cells were grown in M9 media at 25jC for 24 h

upon 0.1 mM IPTG induction. Cells of 10 OD600 were incubated with

2 mg/mL pronase at 37jC.

LI ET AL.: CELL SURFACE DISPLAY USING N-DOMAIN OF INP 219

cantly higher than that for the InaK-NC/GFP. However,

the InaK N-terminal region showed susceptibility to pro-

teolysis, while the C-terminal region was more stable. This

suggests the InaK C-terminal domain might play a role in

protecting INP from proteolytic degradation during self-

assembly. Detailed investigation of the role of the C-

terminal domain is underway.

Cell Surface Display of OPH Using InaK-N

To confirm the ability of InaK-N to act as a display motif,

we employed OPH as another target protein. Cells were

transformed with plasmids in order to express OPH(cyto-

solic expression control), InaK-N/OPH and InaK-NC/OPH,

andOPHactivity in the totalmembrane fractionwas assayed.

Both truncated InaK/OPH fusion samples exhibited much

higher OPH activities (f8–9-fold) than the cytosolic control

sample (Fig. 7). These data confirmed that the InaK N-

terminal domain alone can successfully and efficiently direct

translocation of foreign proteins to the cell surface. However,

OPH activity was not greater for the InaK-N/OPH construct

compared to the InaK-NC/OPH fusion protein, unlike for

InaK/GFP fusion proteins (recall Fig. 2). We might surmise

that this was due to a homodimeric property of OPH; in order

to obtain full biological activity, OPH should be formed

dimer and C-terminal domain might facilitate OPH dimer

formation on the cell surface.

CONCLUSION

We constructed and compared the ability of two types of

truncated InaK to direct cell surface display of two target

foreign proteins, GFP and OPH. The N-terminal domain of

InaK directed cell surface display of GFP with efficiencies

similar to the InaK anchoring motif containing both N- and

C-terminals. Indeed, based on unit cell density, total ex-

pression and surface-displayed levels of GFP were greater

for InaK-N than for InaK-NC. These data suggest a secretion

signal for InaK might be in the N-terminal domain sequence.

In conclusion, truncated InaK containing only the N-termi-

nal domain can be successfully employed as a cell surface

display motif.

References

Bokman SH, Ward WW. 1981. Renaturation of Aequorea green fluores-

cent protein. Biochem Biophys Res Commun 101:1372–1380.

Caldwell SR, Newcomb JR, Schlecht KA, Raushel FM. 1991. Limits of dif-

fusion in the hydrolysis of substrates by the phophotriesterase from

Pseudomonas diminuta. Biochemistry 30:7438–7444.

Cha HJ, Wu CF, Valdes JJ, Rao G, Bentley WE. 2000. Observations of

green fluorescent protein as a fusion partner in genetically engineered

Escherichia coli: monitoring protein expression and solubility. Bio-

technol Bioeng 67:565–574.

Chalfie M, Tu Y, Euskirchen G, Ward WW, Prasher DC. 1994. Green fluo-

rescent protein as a marker for gene expression. Science 263:802–805.

Cochet N, Widehem O. 2000. Ice crystallization by Pseudomonas syringae.

Appl Microbiol Biotechnol 54:153–161.

Donarski WJ, Dumas DP, Heitmeyer DP, Lewis VE, Raushel FM. 1989.

Structure-activity relationship in the hydrolysis of substrates by the

phosphotriesterase from Pseudomonas diminuta. Biochemistry 28:

4650–4655.

Francisco JA, Campbell R, Iverson BL, Georgiou G. 1993. Production and

fluorescence-activated cell sorting of Escherichia coli expressing a

functional antibody fragment on the external surface. Proc Natl Acad

Sci USA 90:10444–10448.

Georgiou G, Stathopoulos C, Daugherty PS, Nayak AR, Iverson BL, Cur-

tiss R. 1997. Display of heterologous proteins on the surface of micro-

organisms: from the screening of combinatorial libraries to live re-

combinant vaccines. Nat Biotechnol 15:29–34.

Grimsley JK, Scholtz JM, Pace CN, Wild JR. 1997. Organophosphorus

hydrolase is a remarkably stable enzyme that unfolds through a homo-

dimeric intermediate. Biochemistry 36:14366–14374.

Hansson M, Stahl S, Nguyen TN, Bachi T, Robert A, Binz H, Sjolander A,

Uhlen M. 1992. Expression of recombinant proteins on the surface of

the coagulase-negative bacteria Staphylococcus xylosus. J Bacteriol

174:4239–4245.

Jung HC, Lebeault CM, Pan JG. 1998a. Surface display of Zymomonas

mobilis levansucrase by using the ice-nucleation protein of Pseudom-

onas syringae. Nat Biotechnol 16:576–580.

Jung HC, Park JH, Park SH, Lebeault CM, Pan JG. 1998b. Expression

of carboxymethyl cellulase on the surface of Escherichia coli using

Pseudomonas syringae ice nucleation protein. Enzyme Microb Tech-

nol 22:348–354.

Kang DG, Kim YK, Cha HJ. 2002a. Comparison of green fluorescent pro-

tein expression in two industrial Escherichia coli strains, BL21 and

W3110, under co-expression of bacterial hemoglobin. Appl Microbiol

Biotechnol 59:523–528.

Kang DG, Kim JYH, Cha HJ. 2002b. Enhanced detoxification of organo-

phosphates using recombinant Escherichia coli with co-expression of

organophosphorus hydrolase and bacterial hemoglobin. Biotechnol

Lett 24:879–883.

Kornacker MG, Pugsley AP. 1990. The normally periplasmic enzyme

h-lactamase is specifically and efficiently translocated through the

Escherichia coli outer membrane when it is fused to the cell-surface

enzyme pullulanase. Mol Microbiol 4:1101–1109.

Kozloff LM, Turner MA, Arellano F. 1991. Formation of bacterial mem-

brane ice-nucleating lipoglycoprotein complexes. J Bacteriol 173:

6528–6536.

LaiK, StolowichNJ,Wild JR. 1995.Characterization of P-S bondhydrolysis

in organophosphorothioate pesticides by organophosphorus hydrolase.

Arch Biochem Biophys 318:59–64.

Lee JS, Shin KS, Pan JG, Kim CJ. 2000. Surface-displayed viral antigens on

Salmonella carrier vaccine. Nat Biotechnol 18:645–648.

Figure 7. OPH activities in total membrane fractions of the cells har-

boring control cytosolic OPH, InaK-N/OPH, and InaK-NC/OPH fusions.

Cells were grown in M9 media at 25jC for 24 h upon 0.1 mM IPTG

induction.

220 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 85, NO. 2, JANUARY 20, 2004

Mararitis A, Bassi AS. 1991. Principles and biotechnological applications

of bacterial ice nucleation. Crit Rev Biotechnol 11:277–295.

Mejare M, Ljung S, Bulow L. 1998. Selection of cadmium specific hexa-

peptide and their expression as OmpA fusion proteins in Escherichia

coli. Protein Eng 11:489–494.

Miksch G, Fiedler E, Dobrowolski P, Friehs K. 1997. The kil gene of the

ColE1 plasmid of Escherichia coli controlled by a growth-phase-de-

pendent promoter mediates the secretion of a heterologous periplasmic

protein during the stationary phase. Arch Microbiol 167:143–150.

Mulbry WW, Karns JS. 1989. Parathion hydrolase specified by the Fla-

vobacterium opd gene: relationship between the gene and protein. J

Bacteriol 171:6740–6746.

Newton SMC, Joys TM, Anderson SA, Kennedy RC, Hovi ME, Stocker

BAD. 1995. Expression and immunogenicity of an 18-residue epitope

of HIV1 gp41 inserted in the flagellar protein of a Salmonella live

vaccine. Res Microbiol 146:203–216.

Palaiomylitou MA, Kalimanis A, Koukkou AI, Drainas C, Anastasso-

poulos E, Panopoulos NJ, Ekateriniadou LV, Kyriakidis DA.

1998. Phospholipid analysis and fractional reconstitution of the

ice nucleation protein activity purified from Escherichia coli

overexpressing the inaZ gene of Pseudomonas syringae. Cryobiology

37:67–76.

Richins RD, Kaneva I, Mulchandani A, Chen W. 1997. Biodegradation of

organophosphorus pesticides by surface-expressed organophosphorus

hydrolase. Nat Biotechnol 15:984–987.

Rodrigue A, Chanal A, Beck K, Muller M, Wu L. 1999. Co-translocation

of a periplasmic enzyme complex by a hitchhiker mechanism through

the bacterial Tat pathway. J Biol Chem 274:13223–13228.

Schmid D, Pridmore D, Capitani G, Battistutta R, Neeser JR, Jann A.

1997. Molecular organization of the ice nucleation protein InaV from

Pseudomonas syringae. FEBS Lett 414:590–594.

Shi H, Su WW. 2001. Display of green fluorescent protein on Escherichia

coli cell surface. Enzyme Microb Technol 28:25–34.

Shimazu M, Mulchandani A, Chen W. 2001a. Cell surface display of or-

ganophosphorus hydrolase using ice nucleation protein. Biotechnol

Prog 17:76–80.

Shimazu M, Mulchandani A, Chen W. 2001b. Simultaneous degradation

of organophosphorus pesticides and p-nitrophenol by a genetically

engineered Moraxella sp. with surface-expressed organophosphorus

hydrolase. Biotechnol Bioeng 76:318–324.

Sleytr UB, Sara M. 1997. Bacterial and archaeal s-layer proteins: structure-

function relationship and their biotechnological applications. Trends

Biotechnol 15:20–26.

Sousa C, Cebolla A, de Lorenzo V. 1996. Enhanced metalloadsorption

of bacterial cells displaying poly-his peptides. Nat Biotechnol 14:

1017–1020.

Taylor IM, Harrison JL, Timmis KN, O’Connor CD. 1990. The TraT li-

poprotein as a vehicle for the transport of foreign antigenic determinants

to the cell surface of Escherichia coli K12: structure-function relation-

ships in the TraT protein. Mol Microbiol 4:1259–1268.

LI ET AL.: CELL SURFACE DISPLAY USING N-DOMAIN OF INP 221