Lincomycin-induced over-expression of mature recombinant cholera toxin B subunit and the holotoxin in Escherichia coli

Protein Expression and Purification 67 (2009) 96–103

Contents lists available at ScienceDirect

Protein Expression and Purification journal homepage: www.elsevier.com/locate/yprep

Lincomycin-induced over-expression of mature recombinant cholera toxin B subunit and the holotoxin in Escherichia coli Hideyuki Arimitsu a,*, Kentaro Tsukamoto a, Sadayuki Ochi a, Keiko Sasaki a, Michio Kato a, Koki Taniguchi b, Keiji Oguma c, Takao Tsuji a a

Department of Microbiology, Fujita Health University, School of Medicine, Toyoake, Aichi 470-1192, Japan Department of Virology and Parasitology, Fujita Health University, School of Medicine, Toyoake, Aichi 470-1192, Japan c Department of Bacteriology, Okayama University Graduate School of Medicine and Dentistry, Okayama 700-8558, Japan b

a r t i c l e

i n f o

Article history: Received 22 January 2009 and in revised form 21 April 2009 Available online 3 May 2009 Keywords: Cholera toxin Heat-labile enterotoxin B subunit Shine-Dalgarno sequence Lincomycin Lactose operon

a b s t r a c t Cholera toxin (CT) B subunit (CTB) was overproduced using a novel expression system in Escherichia coli. An expression plasmid was constructed by inserting the gene encoding the full-length CTB and the ShineDalgarno (SD) sequence derived from CTB or from the heat-labile enterotoxin B subunit (LTB) of enterotoxigenic E. coli into the lacZa gene fragment in the pBluescript SK(+) vector. The E. coli strain MV1184 was transformed with each plasmid and then cultured in CAYE broth containing lincomycin. Recombinant CTB (rCTB) was purified from each cell extract. rCTB was overproduced in both transformants without obvious toxicity and was structurally and biologically identical to that of CT purified from Vibrio cholerae, indicating that the original SD and CTB signal sequences were also sufficient to express rCTB in E. coli. Lincomycin-induced rCTB expression was inhibited by mutating the lac promoter, suggesting that lincomycin affects the lactose operon. Based on these findings, we constructed a plasmid that contained the wild-type CT operon and successfully overproduced CT (rCT) using the same procedure for rCTB. Although rCT had an intact A subunit, the amino-terminal modifications and biological properties of the A and B subunits of rCT were identical to those of CT. These results suggest that this novel rCTB over-expression system would also be useful to generate both wild-type and mutant CT proteins that will facilitate further studies on the characteristics of CT, such as mucosal adjuvant activity. Ó 2009 Elsevier Inc. All rights reserved.

Introduction Cholera toxin (CT)1 is produced by Vibrio cholerae and is a causative factor of diarrhea, which can severely affect children and the elderly. CT shares many properties with heat-labile enterotoxin (LT) produced by enterotoxigenic Escherichia coli (ETEC). Both toxins are composed of one A subunit (CTA or LTA) and five B subunits (CTB or LTB), each of which share approximately 80% identity at both the nucleotide and amino acid levels [1], and cross-react serologically. The A subunit (28 kDa) induces toxicity (diarrhea) by activating the cellular adenylate cyclase through the ADP-ribosylation of Gsa protein. The B subunit (11.6 kDa) forms a pentameric structure and binds to GM1 (Galb1-3GalNAcb1-4(NeuAca2–3)Galb1–4Glcb1–1ceramide) strongly and GD1b (Galb1-3GalNAcb1-4(NeuAca2-8NeuAca2-3)Galb1-4Glcb1-1ceramide) weakly [2]. Because of its association with gangliosides, the B subunit is often used as a tracer for lipid rafts on the cell surface. Furthermore, the attenuated holo* Corresponding author. Fax: +81 562 93 4003. E-mail address: [email protected] (H. Arimitsu). 1 Abbreviations used: CT, Cholera toxin; PBS, phosphate-buffered saline; SDS–PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis. 1046-5928/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.pep.2009.04.011

toxin containing a mutation in the active center of the A subunit, and its B subunits (without A subunit) exhibit powerful mucosal adjuvant activity when administered orally or intranasally with vaccine antigens [1,3]. Many recombinant technologies have been employed to overexpress holotoxins and their B subunits. However, it is difficult to express the wild-type CT operon in E. coli due to toxicity [4,5], despite the high homology with the LT operon. In addition, although some researchers have established expression systems for recombinant CTB (rCTB) by changing the promoter and replacing the signal sequence derived from CTB with that from another protein [6– 8], these expression products were produced in an insoluble fraction or contained additional amino acid residues at the amino-terminal end [7,8]. Lincomycin was originally identified as an antibiotic that prevents protein synthesis in gram-positive bacteria by inhibiting peptidyltransferase activity on the 50S ribosomal subunit [9]. However, lincomycin has been also shown to increase the production of LT in wild-type ETEC and CT in V. cholerae strains [10,11]. Previously, we constructed an rLTB expression plasmid that produced LTB when introduced into E. coli MV1184 [12]. This plasmid was constructed by inserting the full-length LTB gene and three

H. Arimitsu et al. / Protein Expression and Purification 67 (2009) 96–103

base pairs of its Shine-Dalgarno (SD) sequence located at the upstream region. Protein expression was induced by adding lincomycin to the culture medium (unpublished data), but the mechanism of this induction is still unknown. In this study, we investigated whether rCTB could be expressed as a mature form in E. coli by replacing the entire structural gene of LTB in our expression plasmid with that of CTB to focus on the significance of the SD sequence derived from LTB, which was successfully used to over-express the LTB protein in E. coli. However, the three base pairs of the SD sequence of CTB [13] (shown in Fig. 1) were identical with those of LTB, leading to the hypothesis that differences in the SD sequence of LTB and CTB do not affect protein expression. Therefore, we constructed another plasmid containing

97

the SD sequence of CTB. Finally, we report the over-expression of wild-type CT based on the findings obtained with this novel CTB expression system. Materials and methods Construction of expression plasmids Total genomic DNA was isolated from the V. cholerae classical O1 strain 569B using the G NOME DNA purification kit (BIO 101) and then used as a template for PCR. To amplify the CTB and CT holotoxin genes (ctb and ct, respectively), PCR was conducted using the following gene-specific primer sets: CTB(EcoRI)-f and CTB(Hin-

Fig. 1. Schematic structure of the rCTB and rCT expression plasmids. (a) The full-length ctb containing the SD sequence of LTB upstream of the structural gene (ctb) was inserted into the lacZa gene fragment using the restriction sites EcoRI and HindIII (pBSK-CTB). pBSK-CTB2 was prepared by replacing the SD sequence of LTB in pBSK-CTB with that of CTB by site-directed mutagenesis. (b) A full-length CT operon (from the SD sequence of CTA) was inserted into the lacZa gene fragment using the restriction sites EcoRI and HindIII (pBSK-CT). The deduced amino acid sequences of CTA and CTB are shown below and above the nucleotide sequences, respectively. *Stop codon. The underlined SD sequences are from a previous report (Yamamoto et al. [13]). P lac: lac promoter.

98

H. Arimitsu et al. / Protein Expression and Purification 67 (2009) 96–103

dIII)-r; and CTA(EcoRI)-f and CTB(HindIII)-r, respectively. The amplified DNA product was digested with EcoRI and HindIII and then cloned into the lacZa gene fragment in pBluescript II SK (+) that was digested with the same enzymes. The plasmids were named pBSK-CTB and pBSK-CT, respectively. The following mutant plasmids were generated by site-directed mutagenesis using the QuikChange II Site-Directed Mutagenesis kit (Stratagene) and pBSK-CTB as a template: pBSK-CTB2 with a SD sequence derived from CTB, pBSK(-35m)-CTB and pBSK(-10m)-CTB with non-functional mutations in the lac promoter, and pBSK(DT3)-CTB without T3 promoter. Two mutant plasmids encoding holotoxins, which were generated by replacing glutamic acid at position 110 or 112 in CTA with aspartic acid (E110D) or glutamine (E112Q), respectively, were also prepared by site-directed mutagenesis using pBSK-CT as a template. All primer sets used in this study are listed in Table 1, and the plasmid maps for pBSK-CTB and -CTB2 and pBSK-CT are shown in Fig. 1. Expression and purification of rCTB and rCT The constructed plasmids were introduced into the E. coli strains MV1184 (ara, D(lac-proAB), rpsL, thi (/80lacZDM15), D(srlrecA) 306::Tn10 (tetr)/F’[traD36, proAB+, lacIq, lacZDM15]). Each transformant was cultured in 3 ml of Luria–Bertani (LB) broth containing 50 lg ml1 (final concentration) ampicillin overnight at 37°C. Next, 10 ll of culture was inoculated into 4 ml of CAYE broth (2% casamino acids, 0.6% yeast extract, 0.25% NaCl, 0.871% K2HPO4, 0.25% Glucose) containing a 0.1% (vol/vol) trace salt solution (5% MgSO4, 0.5% MnCl2 and 0.5% FeCl3) [14], with or without 90 lg ml1 of lincomycin (Pfizer) and then cultured at 37 °C. For protein purification, cells from the 3-ml culture of LB broth were inoculated into 1 l of CAYE medium containing lincomycin and cultured for an additional 48 h at 30 °C. The cells were collected by centrifugation (7600g, 20 min) and then sonicated in TEAN buffer (50 mM Tris, 1 mM EDTA, 3 mM sodium azide, 200 mM NaCl, pH 6.8). After centrifugation (15,000g, 90 min), the supernatant was applied to a 2-ml column of immobilized D-galactose gel (Pierce), according to the procedure described by Uesaka et al. [15]. To eliminate minor contaminants, the eluent from the D-galactose gel was dialyzed against a 10 mM sodium phosphate buffer (pH 7.0) and then applied to a 2-ml CM-Sephadex C-50 column equilibrated

Table 1 Primers used in this study. Primer name

Sequence (50 –30 )

CTB (EcoRI)-fa CTA (EcoRI)-fb CTB (HindIII)-ra,b CTB(SD)-fc CTB(SD)-rc lacZ(-35)down-fd lacZ(-35)down-rd lacZ(-10)down-fd lacZ(-10)down-rd T3 delete-fe T3 delete-re CT(E110D)-ff CT(E110D)-rf CT(E112Q)-ff CT(E112Q)-rf

GGAATTCGGGATGAATTATGATTAAATTAAAATTTGGTG GGAATTCAGGGAGCATTATATGGTAAAGATAATATTTG CCCAAGCTTGGGTTAATTTGCCATACTAATTGCGG ATCCCCCGGGCTGCAGGAATTAAGGATGAATTATG AATTTAATCATAATTCATCCTTAATTCCTGCAGCCCG CATTAGGCACCCCAGGCCCCTCACTTTATGCTTC GAAGCATAAAGTGAGGGGCCTGGGGTGCCTAATG GCTTCCGGCTCGCCCGTTGTGTGGAATTGTG CACAATTCCACACAACGGGCGAGCCGGAAGC GATATCGCGCGCTTGGCGTAATCATGG GATATCACAAAAGCTGGAGCTCCACCG CTCATCCAGATGATCAAGAAGTTTCTGCTTTAGG CCTAAAGCAGAAACTTCTTGATCATCTGGATGAG CTCATCCAGATGAACAACAAGTTTCTGCTTTAGG CCTAAAGCAGAAACTTGTTGTTCATCTGGATGAG

The sequences underlined were EcoRI and HindIII recognition sites, respectively. a Used for cloning the CTB gene. b Used for cloning the CT holotoxin gene. c Used for changing the SD sequence of LTB to that of CTB in pBSK-CTB. d Used for mutation in the lac promoter of pBSK-CTB. e Used for deletion of the T3 promoter of pBSK-CTB. f Used for preparation of the mutant CT.

with the same buffer. The adsorbed proteins were eluted with stepwise increases in the NaCl concentration (0.1, 0.2, 0.3 and 0.5 M), and CTB eluted with 0.2 M NaCl (data not shown). For the rCT holotoxin, the same expression and purification steps were employed up to the D-galactose gel column chromatography described above. In order to isolate the rCT holotoxin from the free CTB pentamer (e.g. rCTB) and minor contaminants, the eluent from the D-galactose gel was dialyzed against a 10 mM sodium phosphate buffer (pH 7.0) and then applied to a 1-ml column of hydroxyapatite (Bio-Rad). The adsorbed proteins were initially eluted with the same buffer containing 1 M NaCl, and a successive second elution was performed with a 0.4 M sodium phosphate buffer (pH 7.0). Protein concentrations were determined with the DC protein assay reagent (Bio-Rad) using bovine serum albumin as a standard. Trypsin treatment of rCT rCT was mixed with tosylphenylalanyl chloromethyl ketonetreated bovine pancreatic trypsin (TPCK-trypsin) (Worthington Biochemical) at an rCT to trypsin ratio of 50:1, and the mixture was incubated for 1 h at 37°C. Purification of CT from V. cholerae strain 569B The culture supernatant of V. cholerae 569B was obtained following the procedure described by Finkelstein et al. [16]. CT was purified using an immobilized D-galactose gel column, as described above. Preparation of antisera To generate an antiserum that detects CTB in bacterial cell lysates, a rabbit (New Zealand White, 13 weeks of age, female) was immunized with purified rCTB. Initially, 500 lg of rCTB was emulsified with Freund’s complete adjuvant and injected subcutaneously. Three weeks later, the rabbit was given a second subcutaneous immunization of 500 lg of rCTB with Freund’s incomplete adjuvant. After three additional weeks, a final booster injection of 100 lg of rCTB was administered intravenously without adjuvant. Two weeks later, a whole blood sample was collected from the carotid artery under anaesthesia with pentobarbital sodium, and the antiserum was collected by centrifugation at 3000g for 20 min. This experiment was conducted following the Guidelines for the Management of Laboratory Animals at Fujita Health University. Preparation of bacterial lysates The transformants were collected by centrifugation at 7600g for 10 min, and the cell pellet was resuspended in an appropriate volume of phosphate-buffered saline pH 7.4 (PBS). The cell suspension corresponding to 3  109 cells was centrifuged again, and the cell pellet was disrupted in 100 ll of SDS–PAGE sample buffer and then boiled for 5 min. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS– PAGE) and Western blot analysis Aliquots (5 ll) of each bacterial cell lysate (corresponding to 1.5  108 cells) described above or 5 lg of purified protein were loaded onto a 15% polyacrylamide gel. The gel was stained with Coomassie brilliant blue (CBB)-R250 or electroblotted onto a nitrocellulose membrane using the iBlotTM gel transfer system (Invitrogen). The membrane was incubated with blocking buffer (PBS containing 5% (wt/vol) skim milk) for 1.5 h, and then with anti-CTB

99

H. Arimitsu et al. / Protein Expression and Purification 67 (2009) 96–103

rabbit serum diluted 2000-fold for 1 h. After washing, the membrane was incubated with HRP-labeled anti-rabbit Ig porcine Ig (Dako) diluted 10,000-fold for 1 h and specific bands were detected by the ECL system (GE Healthcare). All reactions were carried out at room temperature, and the membrane was washed with PBS containing 0.05% (vol/vol) Tween 20 three times for 5 min prior to each reaction. Each antibody was diluted with blocking buffer.

tions ranging from 10 pg ml1 to 1000 ng ml1. As a vehicle control, rCTB was added to the chamber at a final concentration of 1000 ng ml1. After incubating overnight at 37 °C, the chamber slides were stained with Giemsa solution, and the total number of the cells including spindle-shaped cells in random areas was counted. The concentration of each toxin that caused elongation in 50% of the cells (EC50) was calculated using Graphpad Prism ver. 5.

Determination of the amino (N)-terminal amino acid sequence Results rCTB and rCT were separated by electrophoresis on a 15% polyacrylamide gel and then electroblotted onto a polyvinylidene difluoride membrane (Millipore) using a semidry blotting apparatus (Bio-Rad). After staining the membrane with CBB-R250, each band was excised and sequenced with a pulsed-liquid phase protein sequencer (model 477-A; Applied Biosystems). Surface plasmon resonance (SPR) One millimolar of ganglioside GM1, GD1b, Asialo GM1 (Galb1– 3GalNAcb1–4 Galb1–4Glcb1–1ceramide) and GT1b (NeuAca2– 3Galb1–3GalNAcb1–4 (NeuAca2–8NeuAca2–3) Galb1–4Glcb1– 1ceramide) (Sigma) in a chloroform/methanol mixture (1:1 by volume) was mixed with 10 mM of 1-palmitoyl-2-oleoyl-sn-glycero3-phosphocholine (POPC) (Avanti Polar Lipids) in chloroform at a molar ratio of 1:100. After each lipid mixture was dried to a thin film using an evaporator under nitrogen gas, 500 ll of PBS was added and the mixture was hydrated by vortexing to yield a 10 mM solution with respect to the liposome concentration. After five rounds of freezing with liquid nitrogen followed by thawing at 37 °C, each suspension was extruded through a 50-nm filter 20 times. As a negative control, a liposome consisting of POPC alone was prepared. The binding of rCTB, rCT and CT to immobilized gangliosides was analyzed with a Biacore2000 (GE Healthcare) and each liposome was immobilized on an L1 sensor chip. Following a 5-min injection of 40 mM octylglucoside at a flow rate of 5 ll min1, each liposome was injected at a flow rate of 2 ll min1 for 30 min. This resulted in an increase in the baseline of 7300–8250 RU. The lipid bilayer was then washed with 100 ll of 50 mM NaOH at a flow rate of 100 ll min1. Finally, nonspecific binding sites were blocked by injecting bovine serum albumin (0.1 mg ml1) for 5 min at a flow rate of 5 ll min1. The immobilized L1 sensor chip with each liposome was primed with 10 mM HEPES buffer containing 0.15 M NaCl pH 7.4 (HBS-N). Various concentrations (15.6–125 nM) of analytes (rCTB, rCT and CT) were injected for 2 min at a flow rate of 30 ll min1. After dissociation of the analyte with HBS-N for 2 min, the chip was washed with a 1-min injection of 100 mM HCl at a flow rate of 30 ll min1 to remove the residual analytes and then replaced with HBS-N for use in subsequent cycles. Sensorgrams of nonspecific binding to the POPC alone (flow cell 1) were subtracted from those of binding to the POPC containing each ganglioside (flow cells 2–4), and kinetic constants (KD) were calculated using BIAevaluation ver.4.1. Chinese hamster ovary (CHO) cell elongation assay A morphological analysis of CHO cells using rCT (nicked by TPCK-trypsin) and CT was conducted according to the procedure described by Guerrant et al. [17]. Briefly, CHO cells were cultured in F-12 medium containing 10% fetal calf serum (FCS) and 100 lg ml1 of gentamycin. After passaging the cells three times by trypsinization, a suspension of 1000 cells in 0.4 ml of the same medium containing 1% FCS was added to each chamber of an eightchamber culture slide (Lab-Tek products). Diluted rCT or CT was immediately added in serial 10-fold dilutions at final concentra-

rCTB expression in E. coli is induced by lincomycin To investigate whether rCTB was expressed in E. coli, E. coli strain MV1184 was transformed with pBSK-CTB and then cultured in CAYE broth with or without lincomycin. As shown in Fig. 2, the transformant cultured in CAYE broth containing lincomycin expressed rCTB after 24 h with increased expression after 48 h. On the other hand, the transformant cultured in CAYE broth without lincomycin did not express rCTB under any of the conditions, suggesting that the expression of rCTB was strongly regulated by lincomycin. Purification of rCTB from the soluble fraction of cell lysates Expressed rCTB was purified using two types of column chromatography. It was previously reported that rCTB was collected from an insoluble fraction of E. coli [7,8]. In our expression system, the amount of rCTB that was purified from the soluble fraction of the transformant was also low (2.0 mg per liter of culture after purification by galactose gel chromatography) when the cells were cultured for 48 h at 37 °C, suggesting that the expressed rCTB formed inclusion bodies. Therefore, transformants that were used to purify rCTB were cultivated at 30 °C. As shown in Table 2, the yield of rCTB from the extract derived from the pBSK-CTB transformant was significantly greater. In addition, the rCTB yield from the extract derived from the transformant with pBSK-CTB2, which had the SD sequence derived from CTB, was the same as that with pBSK-CTB (0.8% and 0.9% of the total extract, respectively).

Time after induction 24 hours

48 hours

LM Fig. 2. Detection of rCTB in lysates from E. coli transformed with pBSK-CTB by Western blot analysis using an anti-CTB rabbit serum. The transformants were cultured in CAYE broth with (+) or without () lincomycin (LM) and harvested after 24 and 48 h. Each lane was loaded with total protein lysates corresponding to 1.5  108 cells.

Table 2 Yield of rCTB from cells transformed with each plasmid. Total protein (mg)

Recovery (%)

Transformed with pBSK-CTB Cell extract D-galactose gel eluent CM-Sephadex eluent

2155.0 23.3 19.5

100.0 1.1 0.9

Transformed with pBSK-CTB2 Cell extract D-galactose gel eluent CM-Sephadex eluent

1758.0 15.6 13.4

100.0 0.9 0.8

H. Arimitsu et al. / Protein Expression and Purification 67 (2009) 96–103

M 1 2

The mobility of rCTB (lane 1) by SDS–PAGE gave an approximate size of 11.6 kDa and was identical to that of the CTB in CT (lane 2) (Fig. 3). The N-terminal amino acid sequence of rCTB, Thr-Pro-GlnAsn-Ile-Thr-Asp-Leu-Cys-Ala, was also identical to that of CTB [18].

TB T3 ) -C

K(

promoter expressed rCTB when grown in CAYE broth containing lincomycin, but neither of the lac promoter mutants (pBSK(35m)-CTB and pBSK(-10m)-CTB) expressed rCTB (Fig. 4) under any of the conditions, suggesting that lincomycin affects the lactose operon.

Lincomycin induces rCTB expression by activating the lac promoter

Construction of the rCT over-expression system

It had been unclear whether the lac promoter (P lac, in Fig. 1) or the T3 promoter upstream of ctb in pBSK-CTB was responsible for lincomycin-induced CTB expression. Therefore, two mutants with a non-functional mutation in the -35 or -10 region of the lac promoter and a mutant without T3 promoter in pBSK-CTB were prepared. The pBSK(DT3)-CTB transformant that lacked the T3

Absorbance at OD 280nm

pB S

Fig. 4. Detection of rCTB in the lysates of E. coli (1.5  108 cells) transformed with pBSK-CTB (normal lac and T3 promoters), pBSK(-35m)-CTB and pBSK(-10m)-CTB (mutated lac promoter), and pBSK(DT3)-CTB (T3 promoter completely deleted) by Western blot analysis using an anti-CTB rabbit serum. All of the transformants were cultured in CAYE broth with (+) or without () 90 lg ml1 of lincomycin (LM) for 48 h at 37 °C.

Fig. 3. SDS–PAGE profile of rCTB purified from cell lysates of the MV1184 strain transformed with pBSK-CTB. Lanes M: molecular size marker, 1: rCTB, 2: CT from V. cholerae 569B.

(a)

-10 m) -C

TB

pB S

CTB

6.5

K(

pB S

14.4

-35 m) -C

KCT B

CTA 21.5

TB

- + - + - + - +

LM

31

pB S

(kDa) 66.2 45

K(

100

Based on the findings described above, we speculated that lincomycin-induced overproduction of rCTB might be applicable to CT, which has been reported to be deleterious to E. coli harboring the wild-type CT gene [5].

2.5 2 1.5 1

4

1 st elution

0.5

5

6 7 8 9 11 17 18

2nd elution

0 1

6

11

16

21

Fraction number (2ml/tube)

(b)

(kDa)

N-terminal amino acid sequence

66.2 45 31

Intact CTA

(NDDKLYRADS)

Nicked CTA (NDDKLYRADS)

21.5 14.4

CTB

6.5

1

2

(TPQNITDLCA)

3

Fig. 5. (a) Pattern of rCT separation by hydroxyapatite chromatography. Adsorbed proteins were initially eluted with a 10 mM sodium phosphate buffer containing 1 M NaCl (1st elution), and then eluted with a 0.4 M sodium phosphate buffer (2nd elution). SDS–PAGE profiles of arbitrary fractions are shown above the graph. (b) SDS–PAGE profile of rCT. Five micrograms of toxin treated with (lane 2) or without (lane 1) trypsin and CT derived from V. cholerae (lane 3) were separated on a 15% polyacrylamide gel.

101

H. Arimitsu et al. / Protein Expression and Purification 67 (2009) 96–103

We constructed a pBSK-CT transformant as shown in Fig. 1b. rCT was expressed comparably to rCTB in E. coli without any obvious toxic effects on cell growth. rCT was initially purified with a Dgalactose gel, and then isolated using hydroxyapatite column chromatography, as shown in Fig. 5a. rCT holotoxin eluted with the first elution, while the other contaminants (mainly free CTB) eluted with the second elution. The SDS–PAGE profiles of rCT and CT are shown in Fig. 5b. The size of CTA in CT (lane 3) was approximately 24 kDa due to nicking Table 3 Yield of rCT holotoxin from cells transformed with pBSK-CT.

Cell extract D-galactose

gel eluent Hydroxyapatite eluent

by bacterial proteases, while that of CTA in rCT (lane 1) was intact (approximately 28 kDa). A product with the same band sizes as CT was obtained by treating rCT with TPCK-trypsin (lane 2). In addition, the N-terminal amino acid sequences of the A and B subunits of rCT were identical to those of CT. The rCT yield was 15.6 mg per liter of culture in CAYE medium (Table 3). The same results were also obtained with two attenuated CT mutants that were generated by site-directed mutagenesis using pBSK-CT as a template. The final amount of each mutant toxin after purification from one liter of culture was 12.8 mg (E110D) and 13.5 mg (E112Q) (data not shown). rCT and rCTB have similar binding properties to gangliosides

Total protein (%)

Recovery (%)

1760.5 24.2 15.6

100.0 1.4 0.9

To confirm whether the biological property of CTB in rCT was identical with that in CT, the ability of both toxins to bind to gangliosides was analyzed by SPR. As shown in the sensorgrams

CT Response units (RU)

600 500

rCT 700 600

GM1

500 400

400 300

300 200

200 100

100 0

0 -100

Response units (RU)

600 500

-50

0

50

100

150

200

250

600

300 200

300 200

100

100

-100 -50

Response units (RU)

0

50

100

150

200

250

100

150

200

250

0

50

100

150

200

250

100

150

200

250

Asialo GM1

500 400

400 300

300 200

200 100

100 0

0 -100 -50

0

50

100

150

200

250

GT1b

0

50

GT1b

125nM 62.5nM 31.3nM 15.6nM

500 400

400 300

300 200

200 100 -50 -100

-100 -50

700 600

600

0

50

GD1b

0 -100 -50 700 600

Asialo GM1

500

500

0

500 400

0

Response units (RU)

-100 -50 700

GD1b

400

600

GM1

100 0 0

50

100

150

Time(sec)

200

250

-100 -50

0

50

100

150

200

250

Time(sec)

Fig. 6. Sensorgrams of rCT and CT (15.6–125 nM) for POPC liposomes containing 1% gangliosides. Each sensorgram was obtained by subtracting the data when HBS-N alone (0 nM analyte) was injected.

102

H. Arimitsu et al. / Protein Expression and Purification 67 (2009) 96–103

(Fig. 6), both toxins bound to GM1 and GD1b in a dose-dependent (15.6–125 nM) manner, but did not bind to Asialo GM1 and GT1b, which were similar to the binding properties of CT from V. cholerae as reported by Fukuta et al. [2]. The KD value calculated from each sensorgram showed that both toxins bound to GM1 much stronger than to GD1b (Table 4). The same results were obtained when rCTB was analyzed (data not shown).

(Fig. 7). rCT induced morphological changes in CHO cells with an EC50 of 21.9 ng ml1, which was similar to that of CT (42.3 ng ml1). On the other hand, the EC50s of the two mutant CTs (E110D and E112Q) were both over 1000 ng ml1. rCTB, which was used as a negative control, did not induce any changes, even at 1000 ng ml1 and was similar to untreated cells. Discussion

Comparison of the toxic activity of rCT with that of CT from V. cholerae To confirm whether the biological activity of CTA in rCT was identical with that in CT, the elongation of CHO cells was examined

Table 4 Kinetic constants calculated from the fit to the 1:1 binding model of sensorgrams obtained from the binding of rCT and CT (from V. cholerae). Analyte

Ligand

ka (M1s1)

kd (s1)

KD (M)

Chi2

rCT

GM1 GD1b

2.08  105 2.35  105

1.46  104 2.12  103

7.00  1010 9.02  109

1.33 1.47

CT

GM1 GD1b

4.12  105 2.79  105

3.09  104 1.30  103

7.50  1010 4.66  109

2.77 1.14

ka, association rate constant; Kd, dissociation rate constant; KD, kinetic constant (kd/ ka).

To over-express rCTB in E. coli, researchers have used various promoters and replaced the signal sequence of ctb with that of other proteins. Such approaches may be based on the speculation that the original signal sequence of ctb is not suitable for high protein expression and/or translocation of CTB into the periplasm in E. coli. In this study, however, we succeeded in establishing an overexpression system for rCTB from soluble fractions of E. coli harboring expression plasmids that encode the full-length ctb (including its signal sequence) with the SD sequence derived from ltb in the upstream region. The N-terminal amino acid sequence of the purified rCTB was identical to that of the B subunit of CT and rCTB bound to gangliosides with similar affinities. In addition, although we initially used the SD sequence derived from ltb because of the homogeneity with the host cell, sufficient amounts of rCTB were

Fig. 7. Elongation activity of rCT in CHO cells. CHO cells were incubated with the indicated proteins in F-12 medium containing 1% fetal calf serum overnight at 37 °C.

H. Arimitsu et al. / Protein Expression and Purification 67 (2009) 96–103

obtained when the SD sequence of CTB was used instead of that from LTB, suggesting that the differences in the SD and signal sequences between E. coli and V. cholerae were not important factors in CTB expression. Consequently, we also speculated that the wildtype CT operon, which has been difficult to express in E. coli, could also be expressed using this system. As expected, rCT was expressed in E. coli in the same manner as rCTB. Jobling et al. reported that bacterial cells transformed with a plasmid harboring the wildtype CT operon or the CTB gene alone had inhibited growth and underwent considerable cell lysis after induction by IPTG [5]. To solve these problems, they constructed an expression system for rCT in which the native signal sequences of both CTA and CTB were replaced with that of LTIIb. In our expression system, the growth inhibition of transformants in CAYE broth containing lincomycin was well within several fold compared to that in the same medium without lincomycin, suggesting that the wild-type CT operon worked well in E. coli without any obvious toxicity. Furthermore, we were able to isolate rCT from the free CTB pentamer (cholera genoid) using a simple procedure with hydroxyapatite column chromatography. Only the CT holotoxin eluted with the first elution, and preliminary experiments confirmed that free CTB did not elute in the first elution using purified rCTB (data not shown). This method of separation has advantages of simplicity and rapidity. In these expression systems, lincomycin induces protein expression. It has been reported that lincomycin, which was originally identified as bacteriostatic compound and a potent inhibitor of protein synthesis in gram-positive bacteria [9], increased the production of LT and CT in wild-type ETEC and V. cholerae, respectively [10,11]. Although the mechanism by which lincomycin induces the opposite effect on protein synthesis has been unknown for the past 30 years, it became clear that lincomycin affected the lactose operon in the rCTB expression system as a result of site-directed mutagenesis analyses. These findings led to the speculation that the same effect was obtained by induction of the lac promoter with IPTG, which has been extensively used in expression systems. The transformants induced rCTB expression in CAYE broth, but not in LB broth as a standard medium when IPTG was added to each culture at final concentration of 0.1 or 1.0 mM in log-phase proliferation (absorbance at 600 nm was between 0.5 and 1.0) and the cultures were further incubated for 48 h at 30 °C (data not shown). Furthermore, lincomycin did not induce rCTB expression in LB broth (data not shown), suggesting that the ability of lincomycin to induce protein expression was optimized by the media ingredients. It was reported that regulation of the expression of the LT operon was associated with the sequence of the structural gene itself [19]. The length of ctb inserted into the vector was identical in size (375 bp) and the nucleotide sequence was similar to that of ltb, suggesting that the structural gene contributes to this effect. However, the same expression profiles were obtained when the gene encoding CTB was replaced with the HA1 gene (894 bp), a subcomponent of hemagglutinin in Clostridium botulinum (data not shown) [20], suggesting that this expression system was not affected by the length or sequence of the coding gene. The precise pharmacological effect of lincomycin on protein synthesis through the lactose operon has been unclear, both at the transcriptional and translational levels. In summary, novel systems for rCTB and rCT expression induced by lincomycin were established. Lincomycin can induce protein expression when added to CAYE broth at the beginning of the culture and is superior to IPTG, which generally needs to be added in

103

the log phase of proliferation. In addition, these procedures will be useful to generate both wild-type and mutant CT proteins, leading to facilitate further studies on the characteristics of CT, such as mucosal adjuvant activity. Acknowledgment This work was supported by grants from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. References [1] J. Sanchez, J. Holmgren, Cholera toxin structure, gene regulation and pathophysiological and immunological aspects, Cell. Mol. Life Sci. 65 (2008) 1347–1360. [2] S. Fukuta, J.L. Magnani, E.M. Twiddy, R.K. Holmes, V. Ginsburg, Comparison of the carbohydrate-binding specificities of cholera toxin and Escherichia coli heat-labile enterotoxins LTh-I, LT-IIa, and LT-IIb, Infect. Immun. 56 (1988) 1748–1753. [3] M. Pizza, M.M. Giuliani, M.R. Fontana, E. Monaci, G. Douce, G. Dougan, K.H. Mills, R. Rappuoli, G. Del Giudice, Mucosal vaccines: non toxic derivatives of LT and CT as mucosal adjuvants, Vaccine 19 (2001) 2534–2541. [4] M.L. Gennaro, P.J. Greenaway, D.A. Broadbent, The expression of biologically active cholera toxin in Escherichia coli, Nucleic Acids Res. 10 (1982) 4883–4890. [5] M.G. Jobling, L.M. Palmer, J.L. Erbe, R.K. Holmes, Construction and characterization of versatile cloning vectors for efficient delivery of native foreign proteins to the periplasm of Escherichia coli, Plasmid 38 (1997) 158– 173. [6] P. Slos, D. Speck, N. Accart, H.V. Kolbe, D. Schubnel, B. Bouchon, R. Bischoff, M.P. Kieny, Recombinant cholera toxin B subunit in Escherichia coli: high-level secretion, purification, and characterization, Protein Expr. Purif. 5 (1994) 518– 526. [7] A.P. Areas, M.L. Oliveira, C.R. Ramos, M.E. Sbrogio-Almeida, I. Raw, P.L. Ho, Synthesis of cholera toxin B subunit gene: cloning and expression of a functional 6His-tagged protein in Escherichia coli, Protein Expr. Purif. 25 (2002) 481–487. [8] C. L’Hoir, A. Renard, J.A. Martial, Expression in Escherichia coli of two mutated genes encoding the cholera toxin B subunit, Gene 89 (1990) 47–52. [9] J. Spizek, T. Rezanka, Lincomycin, clindamycin and their applications, Appl. Microbiol. Biotechnol. 64 (2004) 455–464. [10] M. Levner, F.P. Wiener, B.A. Rubin, Induction of Escherichia coli and Vibrio cholerae enterotoxins by an inhibitor of protein synthesis, Infect. Immun. 15 (1977) 132–137. [11] M.H. Levner, C. Urbano, B.A. Rubin, Lincomycin increases synthetic rate and periplasmic pool size for cholera toxin, J. Bacteriol. 143 (1980) 441–447. [12] T. Tsuji, M. Kato, Y. Kato, H. Kawase, S. Imamura, A. Miyama, Construction of plasmids useful for production of the B subunit of cholera toxin from Vibrio cholerae or a heat-labile enterotoxin from enterotoxigenic Escherichia coli, Eur. J. Epidemiol. 10 (1994) 393–398. [13] T. Yamamoto, A. Suyama, N. Mori, T. Yokota, A. Wada, Gene expression in the polycistronic operons of Escherichia coli heat-labile toxin and cholera toxin: a new model of translational control, FEBS Lett. 181 (1985) 377–380. [14] K. Sasaki, M. Kato, T. Takahashi, S. Ochi, Y. Ichinose, K. Shiraki, Y. Asano, M. Iwanaga, T. Tsuji, Live varicella vaccine polarizes the mucosal adjuvant action of cholera toxin or its B subunit on specific Th1-type helper T cells with a single nasal coadministration in mice, J. Med. Virol. 70 (2003) 329–335. [15] Y. Uesaka, Y. Otsuka, Z. Lin, S. Yamasaki, J. Yamaoka, H. Kurazono, Y. Takeda, Simple method of purification of Escherichia coli heat-labile enterotoxin and cholera toxin using immobilized galactose, Microb. Pathog. 16 (1994) 71–76. [16] R.A. Finkelstein, J.J. LoSpalluto, Pathogenesis of experimental cholera. Preparation and isolation of choleragen and choleragenoid, J. Exp. Med. 130 (1969) 185–202. [17] R.L. Guerrant, L.L. Brunton, T.C. Schnaitman, L.I. Rebhun, A.G. Gilman, Cyclic adenosine monophosphate and alteration of Chinese hamster ovary cell morphology: a rapid, sensitive in vitro assay for the enterotoxins of Vibrio cholerae and Escherichia coli, Infect. Immun. 10 (1974) 320–327. [18] C.Y. Lai, Determination of the primary structure of cholera toxin B subunit, J. Biol. Chem. 252 (1977) 7249–7256. [19] J. Yang, M. Tauschek, R. Strugnell, R.M. Robins-Browne, The H-NS protein represses transcription of the eltAB operon, which encodes heat-labile enterotoxin in enterotoxigenic Escherichia coli, by binding to regions downstream of the promoter, Microbiology 151 (2005) 1199–1208. [20] H. Arimitsu, Y. Sakaguchi, J.C. Lee, S. Ochi, K. Tsukamoto, Y. Yamamoto, S. Ma, T. Tsuji, K. Oguma, Molecular properties of each subcomponent in Clostridium botulinum type B haemagglutinin complex, Microb. Pathog. 45 (2008) 142– 149.