A-Protein from Achromogenic Atypical Aeromonas salmonicida ...

Protein Expression and Purification 16, 396 – 404 (1999) Article ID prep.1999.1054, available online at http://www.idealibrary.com on

A-Protein from Achromogenic Atypical Aeromonas salmonicida: Molecular Cloning, Expression, Purification, and Characterization Sarah Maurice,* Dietland Ha¨dge,† Mara Dekel,‡ Aharon Friedman,§ Arieh Gertler,* ,1 and Oded Shoseyov‡ *Institute of Biochemistry, Food Science and Nutrition, ‡The Kennedy Leigh Centre for Horticulture Research, §Department of Animal Sciences, Faculty of Agriculture, Food and Environmental Quality Sciences, The Hebrew University of Jerusalem, P.O. Box 12, Rechovot 76100, Israel; and †University of Leipzig, Institute for Zoology, Talstrasse 33, Leipzig, Germany

Received January 5, 1999, and in revised form February 15, 1999

Achromogenic atypical Aeromonas salmonicida is the causative agent of goldfish ulcer disease. Virulence of this bacterium is associated with the production of a paracrystalline outer membrane A-layer protein. The species-specific structural gene for the monomeric form of A-protein was cloned into a pET-3d plasmid in order to express and produce a recombinant form of the protein in Escherichia coli BL21(DE3). The induced protein was isolated from inclusion bodies by a simple solubilization-renaturation procedure and purified by ion exchange chromatography on QSepharose to over 95% pure monomeric protein. Recombinant A-protein was compared by biochemical, immunological, and molecular methods with the Aprotein isolated from atypical A. salmonicida bacterial cells by the glycine and the membrane extraction methods. The recombinant form was found to be undistinguishable from the wild type when examined by SDS–PAGE and gel filtration chromatography. The immunological similarity of the protein samples was demonstrated by employing polyclonal and monoclonal antibodies in ELISA and Western blot techniques. All forms of A-protein were found to activate the secretion of tumor necrosis factor a from murine macrophage. To date, this represents the first largescale production of biologically active recombinant A-protein. © 1999 Academic Press Key Words: atypical Aeromonas salmonicida; A-protein; recombinant.

1 To whom correspondence and reprint requests should be addressed. Fax: (972) 89-476-189. E-mail: [email protected].

396

Atypical Aeromonas salmonicida has been isolated from a wide range of cultivated and wild fish species inhabiting fresh water (1,2), brackish water (3), and marine environments (4). Diseases ascribed to this organism are carp erythrodermatitis (5), goldfish ulcer disease (6 – 8), and ulcer disease of flounder (4). Pathogenic isolates of typical and atypical forms of A. salmonicida produce an outer membrane protein (A-layer). The A-layer has been demonstrated to be an important virulence factor. Bacteria that were defective in their ability to form the outer membrane protein were generated by employing the transposon T5 and were found to be avirulent (9). The A-layer has been attributed to the ability of the bacteria to produce disease (10) by protecting the invader against the hosts’ specific and nonspecific defense mechanisms (11,12). The protein structure is defined as a tetragonal paracrystalline array composed of a single protein species (A-protein) of molecular mass (M r) 49,000 dalton. It has been extracted from outer membranes with chaotropic agents (13) and from intact bacterial cells by acid depolymerization (14,15). The structural gene (vapA) 2 controlling production of the surface array protein of typical A. salmonicida has been cloned from a gene bank of DNA (9) and sequenced (16). Analysis of the vapA DNA sequence disclosed an ORF of 1056 bp encoding 502 amino acid residues. Cloning of the entire vapA gene has been carried out in the bacteriophage vector, lgt11 but the clone proved to be unstable. Chu et al. (18) succeeded in preparing a clone containing the entire vapA sequence in the cosmid vector pLA2917. This expression vector 2

Abbreviation used: vap, virulent A-protein. 1046-5928/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

RECOMBINANT A-PROTEIN

lacked a large part of the promotor region and produced low levels of A-protein in E. coli. To date, a permanent clone has not been attained that can efficiently produce a recombinant form of the protein (17). In this paper we report the overexpression, refolding, and purification of A-protein using a recombinant Escherichia coli system. We confirmed that the purified A-protein is immunologically and biochemically homologous to the protein isolated from atypical and typical A. salmonicida bacterial cells. MATERIAL AND METHODS

Bacterial Strains and Culture Conditions Achromogenic atypical Aeromonas salmonicida strain At3-96, isolated from ulcerated goldfish and identified by The Centre for Environment, Fisheries and Aquaculture (CEFAS), Weymouth, UK, was used for cloning of the atypical A-protein coding sequence, vapA atyp . Bacterial stock cultures, maintained at 270°C in 25% glycerol, were routinely grown on Terrific Agar (TA) containing 0.1% Brilliant blue-R (Sigma, MO) or Terrific broth (TB) (21°C, 48 –72 h.). E. coli, strains DH10b, HB101 and BL21(DE3) pLysed S were used as bacterial hosts for cloning and expression of vapA atyp, respectively. Both transformed bacterial strains were maintained in 15% glycerol stock cultures at 270°C and grown with Luria agar (LA), broth (LB), or TB overnight (37°C). Isolation of A-Protein DNA Sequences and Construction of pET-vapA atyp vector The atypical A-protein coding sequences was amplified from double-stranded total genomic DNA isolated from 2 ml of a 48 h TB bacterial culture by the method described in Current Protocols in Molecular Biology (18). Two oligonucleotides, AT1 (59-AAAACCATGGATGTCGTGATTACCCG-39), which contained the nucleotides encoding amino acids 1–5 and a NcoI restriction site (underlined) at the initiation codon, and AT2 (59ATTAGGATCCGGCGCCCTTTATCGACTA-39), a 19-mer complementary to the nucleotides situated upstream from the terminator codon in addition to a 9-mer tail for the creation of a BamHI site (underlined) were employed for PCR amplification of the gene. PCR was performed in 10-ml capillary tubes (Rapid cycler, Idaho Tech, ID), containing a mix of 5 ng DNA template, 10 mM of each primer, 1.25 mM of each dNTP, 2 mM MgCl 2, and 0.5 U Taq pol. The following protocol was used: initial 2 min denaturation (94°C) followed by 30 cycles of 0 min denaturation (94°C), 0 min annealing (55°C), 1 min elongation (72°C), and ending with 4 min elongation at the same temperature. The amplified PCR product was purified by desalting with 1/10 vol 3 M sodium acetate (pH 5.2), followed by an equal

397

volume of phenol/chloroform. Two and one half volumes of 100% ethanol were added to the resultant upper phase followed by incubation (1 h, 270°C) and centrifugation. The pellet was washed twice with 70% ethanol and dissolved in water. The purified PCR insert was digested with restriction enzymes BamHI and NcoI and subcloned into the vector pET 3d, previously linearized with the same enzymes, to form construct pET-vapA atyp. Expression, Refold, and Purification of Recombinant A-Layer Protein (At-R) E. coli strain DH10b was transformed with pETvap atyp by electroporation and transformants were selected on LA plates containing 100 mg/ml ampicillin. pET-vap atyp plasmids were isolated by miniprep (Wizard Plus SV, Promega, WI) from an overnight culture in LB/ampicillin and used for the transformation of E. coli BL21(DE3)pLysS by standard methods (19). A transformed colony. was grown overnight in LB/amp and subcultured into 500 ml of TB. Expression of recombinant A-layer protein was induced at OD 600 0.9 by addition of isopropyl b-D-thiogalactopyranoside (IPTG) to a final concentration of 0.4 mM. Cells were harvested 6 h later by centrifugation (6000g, 10 min) and stored at 220°C for later processing. Production of recombinant A-layer protein was monitored by SDS– PAGE (20) of E. coli total protein after solubilization with 10% SDS. Protein analysis was performed using 4% stacking gel and 12% separation gel under reducing conditions. Gels were stained with Coomassie brilliant blue R-250 and molecular mass compared with standard low molecular mass markers. Pelleted bacteria were suspended in 60 ml Tris–HCl (pH 8.0) with EDTA at a final concentration of 10 mM and lysozyme was added (0.5 mg/ml). The mixture was stirred (30 min, 4°C), followed by sonication and separation of soluble and insoluble fractions by centrifugation (25,000g, 30 min). Sedimented inclusion bodies were suspended in cold distilled water containing 0.1% Triton X-100, sonicated, and centrifuged repeatedly until all impurities were removed. This was followed by two more washes in water not containing Triton-X. Inclusion bodies were suspended by sonication in 20 mM Trizma Base containing 5 mM 2-mercaptoethanol (0.5 mg protein to 1 ml buffer) and the pH was adjusted to 12 with NaOH. The clear refold solution was stirred (2 h, 4°C) and dialyzed (48 h) against 6 3 10 L of 10 mM Tris–HCl (pH 8.0). The resultant solution was loaded onto a Q-Sepharose column (2.6 3 10 cm), previously equilibrated with 10 mM Tris–HCl (pH 8.0, 4°C). Elution was carried out using a discontinuous NaCl gradient in the same buffer at a rate of 80 ml/h, and 5-ml fractions were collected. Protein concentrations were determined by absorbence at 280 nm, and

398

MAURICE ET AL.

FIG. 1. Restriction map of vapA atyp employed for sequencing DNA. Fragments were ligated into pBluescript IIKS1 and universal primers were utilized.

monomer content by gel-filtration chromatography on Superdex 75 HR 10/30 column. Isolation of A-Layer Protein by Glycine Extraction (At-G) A-protein was isolated by a modifided method of Bjørnsdottir et al. (14). Frozen cells were thawed and washed three times in 20 mM Tris–HCl (pH 8.0) and suspended (4 g/100 ml) in cold 0.2 M glycine buffer (pH 2.15, 15 min) with occasional mixing. After centrifugation (25,000g, 15 min), the supernatant was dialyzed against 20 mM Tris–HCl (pH 8.0). The protein solution was then chromatographed on a DEAE-Sepharose CL-6B column (30-ml bed vol) equilibrated with the same buffer. Protein fractions were isolated by a stepwise increase of NaCl in buffer. The major peak, eluted at 0.1 M, consisting of A-protein, was pooled, dialyzed against Na-bicarbonate (0.25 mg/ml), lyophilized, and stored at 220°C. Isolation of A-Layer Protein from Bacterial Cell Membranes (At-M) A-protein was isolated from bacterial cell membranes by a modified method of Phipps et al. (13). Frozen cells were lysed by sonication after incubation with EDTA and lysozyme as in the preparation of At-R. Sodium laurylsarcosinate (SLS) and deoxycholate extractions were performed as described. Solubilization of the final product was carried out in 4 M guanidine– HCl. Purification of At-M was executed directly on a DEAE-Sepharose CL-6B column as in the production of At-G. Determination of N-Terminal Sequence Automated Edman degradation technique was used to determine the amino-terminal protein sequence. Degradations were performed on an ABI Model 470A gas phase sequencer (Foster City, CA), using the standard sequencing cycle. The respective PTH-amine acid derivatives were identified by RP-HPLC analysis, using an AbI Model 120A PTH analyzer fitted with a Brownlee 2.1 mm i.d. PTH-C 18 column.

DNA Sequencing Cloned vapA atyp inserts were extracted from harvested pET plasmids using restriction enzymes XbaI and BamHI, cleaned by gel isolation (BIO 101, CA), and subjected to further cutting according to the map seen in Fig. 1. The isolated fragments XbaI/PstI (;952 bp), PstI/BamHI (;580 bp) and KpnI/KpnI (;741 bp) were agarose gel isolated and ligated into pBluescript IIKS1 which had been cut with the respective restriction enzymes. The resultant ligates were transformed into E. coli HB101 and positive transformants selected as previously described. The presence of the desired insert was monitored by isolation with relevant restriction enzymes. Sequencing was performed at The Laboratory of DNA Sequencing, Hebrew University of Jerusalem, with an ABI 377 prism DNA sequencer and by the dye terminator cycle sequencing method. Antibody Production Antisera against A-protein extracted from typical A. salmonicida (strain 391513) and the recombinant form of A-protein was raised in adult rabbits by multiple intradermal injections of 100 mg of protein in Freund’s complete adjuvant followed by two successive injections of 100 mg in Freund’s incomplete adjuvant at 3-week intervals. The IgG fraction was separated by differential precipitation and salting-out utilizing caprylic acid and ammonium sulfate. Immune serum was diluted 1:4 in 60 mM acetate buffer (pH 4.0) and the pH adjusted to 4.5. Caprylic acid was added (25 ml/ml) and the mixture stirred (30 min) followed by centrifugation (30 min, 10000g). The supernatant was filtered through nytex netting, diluted 10:1 with 10x PBS, and the pH raised to 7.4 with 5.0 N NaOH. The serum was cooled and ammonium sulfate (0.277 g/ml) added with constant mixing (30 min). The purified IgG was separated by centrifugation (500g, 20 min) and dialyzed overnight against 100 vol of PBS. Aliquots were heated (55°C, 20 min) before storing (220°C). A-protein specific monoclonal antibodies (mAbs) were generated according to the lymphocyte hybridoma technique of Ko˜hler and Milstein (21) modified by Wagner et al. (22). Splenocytes of mice immunized with either forma-

399


lin-killed cells of A. salmonicida (isolate A 1-/F216.1/83) or A-protein containing culture filtrate of A. salmonicida (isolate A 1-/TG72/78R) were harvested 4 days after the last booster injection. These were fused with P3x63Ag8.653 myeloma cells. Two hybridoma culture supernatants were selected which represented different antibody epitopes and were designated 9B10(IgM, k) and 4B10 (IgG1, l). Immunological Identification: ELISA and Western Blotting Methods ELISA. A mAb-based sandwich-ELISA was used (23). Antibodies were diluted in phosphate-buffered saline (PBS), 1% Tween in PBS (TPBS), or TPBS containing 10 mg/ml bovine serum albumin (TPBS/BSA). Microplates were incubated overnight (5°C) with 50 ml goat-anti mouse IgG(H1L) antibodies (Dianova, 4 –5 mg/ml PBS). The wells were emptied and 50 ml of mAbs (4-5 mg/ml TPBS) were added (1 h, room temperature (RT)) The plates were washed three times with TPBS and incubated with 50 ml of recombinant or glycineextracted A-protein preparations (5 mg/ml TPBS/BSA; 1 h, RT). The amount of the A-protein in the 50 ml was adjusted not to exceed final OD of 1.7 to ensure linearity. Detection of mAb-bound antigen was carried out after incubating the wells overnight (5°C) with 50 ml of immune sera (diluted 1:1000 TPBS/BSA). Two rabbit immune sera were employed; one developed from formalin-killed cells of the A. salmonicida isolate, A 1-/F 302b/86, and the second from outer membrane fragments of the A. salmonicida isolate, A 1-4/88. Following incubation with 50 ml horseradish peroxidase (POD)labeled goat-anti rabbit IgG(H1L) antibodies (Dianova, 0.1 mg/ml TPBS, 1 h, RT) and washing, the reaction was visualized by incubation (30 min, RT) with 2,2-azino-bis(3-ethyl-benzthiazoline-6-sulfonic acid) diammonium salt (ABTS) (5.5 mg/10 ml 0.1 M sodium acetate, pH 4.2, with 5 ml of 30% H 2O 2). The absorbance was measured at 405 nm by ELISA-microplate reader (EL311, Bio-Tek Instruments, Winooski). Western blot technique. Immunotransblots of SDS– PAGE electrophoresis membrane preparations were prepared by electrophoresing the gels onto nitrocellulose sheets in a Trans-Blot Cell (Biorad, MO). Immunoreactions were carried out with either polyclonal antibodies (1/10,000 dilution) derived against glycine extracted outer membrane protein of typical A. salmonicida (24) or the A-protein-specific mAb 4B10 mentioned above and mAb 1B2 (17). When employing polyclonal antibodies the blots were stained by the immunoperoxidase method (28) and developed with ECL (Amersham, Buckinghamshire, UK). When using mAbs, immunostaining was performed by incubation of the blot with hybridoma supernatant (6 mg mAb/ml TPBS, 1 h, RT), followed by POD-labeled goat anti-

mouse IgG/IgM(H1L) antibodies (Dianova; 0.2 mg/ml TPBS, 2 h at 5°C). Detection was performed via the DAB/H 2O 2 reaction. Bioassay Peritoneal macrophages were elicited from C57BL/6 mice by means of an intraperitoneal injection containing 3.0 ml thioglycolate (Difco, MI). Four days post injection the cells were collected by peritoneal lavage with PBS and washed with Dulbecco’s Modified Eagle’s Medium (DMEM) enriched with 1% each of Pen-Strep, sodium pyruvate, 1 mM L-glutamine, nonessential amino acids as well as 5 3 10 25 M 2-mercaptoethanol. Twelve well culture plates were seeded with 1 ml enriched DMEM containing 0.5 3 10 6 macrophages per well and incubated for 18 h (37°C, 5% CO 2). Twenty-five microliters of samples; recombinant (At-R), glycine extracted (At-G), or membrane-extracted (At-M) A-protein were added to the appropriate wells in the free form or bound to 1-mm latex beads (Sigma) to a final protein concentration of 1 mg protein/ml. PBS and protein-free latex beads were used as controls. One microgram of protein is still below the saturation level (not shown). After further incubation (24 h) the macrophage supernatant was collected and added to fibroblastoid cell line (Balb/c clone 7; REF) sensitive to tumor necrosis factor a (TNFa), which was maintained in the same medium supplemented with 10% fetal calf serum. Fibroblasts were collected from stock cultures by trypsin digestion and 100 ml containing 40,000 cells was added to each well of a 96 flat-bottom cell culture plate. These cells were allowed to reestablish overnight and then 100 ml of each macrophage supernatant sample, diluted 1:2 in DMEM/FCS containing actinomycin D (4 mg/100 ml medium) was added to the appropriate fibroblast wells. Latex beads were removed by centrifugation from samples prior to addition. Treated fibroblasts were subsequently incubated for an additional 18 h and 20 ml of 3-(4,5-dimethylthiaxol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (5 mg/ml PBS) were added to each well. Incubation was continued for 4 h. The supernatant was then carefully drained from the culture plate and 10 ml of DMSO added to each well. The color intensity of each sample well was read at 570 nm (with 630 as a reference filter.) The following fibroblast controls were added in order to develop a formula for determining the cytotoxic effect of each macrophage sample: (A) DMEM with 40 ml/ml Triton X-100 (100% cytotoxicity), (B) DMEM without actinomycin D (maximum growth), and (C) DMEM with actinomycin (inhibition of growth). The percentage of cytotoxicity was calculated from the optical density readings using the formula: Percentage of cytotoxicity 5 100 2

~Sample! 2 ~A! 3 100 ~C!~A!

400

MAURICE ET AL.

FIG. 2. SDS–PAGE of A-protein product after induction with IPTG. The electrophoresis was performed on 12% acrylamide gel by the Laemmli discontinuous system. (Lane 1) Markers (molecular mass: 97.4, 66.2, 55.0, 42.7, 40.0, 31.0, 21.5, 14.4 kDa, from top to bottom); (lane 2) total bacterial protein before induction; (lane 3) total bacterial protein after induction.

RESULTS AND DISCUSSION

Purification and Characterization of A-Protein Previous attempts to express A-protein in recombinant form failed due to the toxic affects of the protein on the host organism (16). The strategy which we employed to

elude the potential lethality is described above. The plasmid pET 3d, used in conjunction with bacterial host E. coli BL21 (DE3) pLysed S, resulted in expressed proteins being stored as randomly folded protein inclusion bodies within the bacterial cell body. Upon induction with IPTG, A-protein was expressed at a level approaching approximately 40% of the total cellular protein (Fig. 2). The expressed protein was found exclusively in inclusion bodies and was recovered from the pelted fraction of broken cells. Several postinduction incubation times were tested and the results indicated that the amount of protein produced after 6 h did not significantly increase (data not shown). The induced protein was isolated from the inclusion bodies by a simple solubilization-renaturation procedure. A-protein is deficient in cysteine moieties, therefore, management of disulfide bonding did not have to be addressed. The refolded recombinant A-protein was further purified by ion exchange chromatography on a Q-Sepharose column and a major peak was eluded with 100 mM NaCl. The results in Fig. 4 were integrated, indicating that the pool of the 100 mM NaCl eluate yields approximately 98% monomeric A-protein with low levels of aggregation. In comparison, isolation of outer membrane protein from A. salmonicida cultures by the glycine extraction method produced a yield of less than 75% monomeric form. The efficiency of recombinant A-protein production over isolation from cultured atypical A. salmonicida by means of the membrane and glycine extraction methods is displayed in Table 1.

FIG. 3. Purification of recombinant A-protein by ion exchange chromatography on a Q-Sepharose column (2.5 3 10 cm) equilibrated with 10 mM Tris–HCl (pH 8.0). Elution was accomplished using a discontinuous NaCl gradient of 50, 100, 200, and 300 mM NaCl in the same buffer at a rate of 80 ml/h. 5-ml fractions were collected. The protein concentration was determined by measuring absorbance at 280 nm.


401

FIG. 4. Purification of recombinant A-protein by ion exchange chromatography on a Q-Sepharose column (2.5 3 10 cm) equilibrated with 10 mM Tris–HCl (pH 8.0). Elution was accomplished using a discontinuous NaCl gradient of 50, 100, 200, and 300 mM NaCl in the same buffer at a rate of 80 ml/h. 5-ml fractions were collected. The protein concentration was determined by measuring absorbance at 280 nm.

In the presence of 2-mercaptoethanol and nondenaturing conditions, the purified recombinant A-protein migrated as a single band corresponding to the monomeric 50-kDa forms isolated by glycine and membrane extraction from strain At3-96 (Fig. 5). Analysis of the three A-protein preparations samples by gel chromatography further demonstrated their similarity and monomeric appearance (Fig. 4). The retention times for these proteins (9.05, 9.06, and 9.07) were determined to be analogous to the apparent molecular mass of 48.1 kDa. N-terminal analysis of the recombinant protein exhibited the amino acid sequence, Met-Asp-Val-Val-IleSer with a yield of 189, 91, 127, 141, 173, and 93 picomoles/run, respectively. When disregarding the methionine that was added as an initiation signal, this sequence is identical to the one determined for typical A. salmonidida strain A450 (Gene Bank Accession No. M646655) (18). Translation of the DNA sequence derived by restriction enzyme mapping (Fig. 1), using the local homology algorithm of Smith and Waterman found in the Wisconsin Package displayed a 92.2% identity over 492 amino acids, when compared with the typical A. salmonicida strain A 450. We assume that this difference results largely, if not solely from the fact that in the present work an achromogenic atypical A. salmonicida strain At3-96, isolated from ulcerated goldfish, was used. The similarity demonstrates that

the primary structure of the A-layer protein is highly conserved in typical and atypical strains. Immunological Characterization The presence of cross-reactive immunological epitopes between glycine extracted A-protein and the recombinant form was studied by ELISA (Fig. 6) and Western blot techniques (Fig. 7). Both A-protein preparations showed analogous high capacities for binding mAbs, 4B10 and 1B2 (Fig. 6). These mAbs recognize SDSresistant epitopes as demonstrated in Fig. 7B. Whereas mab 1B2 binds to an epitope detectable on

TABLE 1 Efficiency of Three Protein Production Methods

Method of production

Incubation time

Recovery (mg protein/ liter bacteria)

Outer membrane extraction Glycine extraction Recombinant technique

5 days 5 days 6h

0.15 0.6 160

FIG. 5. SDS–PAGE analysis of A-layer protein produced by three methods. Electrophoresis of 10-mg samples dissolved in sample buffer containing 2-mercaptoethanol was performed by the Laemmli discontinuous system on 12% acrylaminde gel. (Lane 1) Markers (molecular mass 97.4, 66.2, 45, 31, 21.5, 14.4 kDa from top to bottom); (lane 2) glycine-extracted A-protein; (lane 3) membrane extracted A-protein; (lane 4) recombinant A-protein.

402

MAURICE ET AL.

FIG. 6. Sandwich-ELISA for the detection of the binding capacity of recombinant (■) and glycine extracted (h) A-layer protein using three mAbs (9B10, 4B10, 1B2). After capturing with mAbs, the A-protein antigens were further detected with polyclonal A-protein antibodies: (A) rabbit anti-A. salmonicida whole cell (F 302b/86) antibodies or (B) rabbit anti-A. salmonicida outer-membrane fragment (4/88) antibodies.

monomeric A-protein molecules as well as on the surface-exposed tetrameric A-protein complexes (28), mAb 4B10 is specific for the monomeric protein form exclusively. The third mab, 9B10, recognizes a tetrameric conformation-dependent epitope of the surface-array A-protein and as a result it shows a very weak capacity for binding to both protein preparations. As expected, the 9B10-specific epitope is absent in these samples but that they both posses the same specific antigenic epitopes that are associated with A-protein molecules. Two species of anti-A. salmonicida specific rabbit immune sera were used to elucidate the reaction between mAbs and the A-protein preparations. The high degree of complex formation as viewed in Fig. 6 further demonstrates the immunological similarity between A-proteins isolated from atypical and typical strains of A. salmonicida. When using polyclonal antibodies developed against the outer membrane protein of A. salmonicida in Western blot analysis, a single protein species was displayed (Fig. 7A). In accordance with the ELISA studies, the recombinant and the glycineextracted A-proteinsshowed similar binding behavior to mAbs 4B10 and 1B2 (Fig. 7B). The mAb 9B10 is not suitable for Western blot techniques but would be functional in nondenaturizing methods of analysis such as flow cytometry and dot blot. When mab 4B10 was used in a somewhat higher concentration than mab 1B2, some degradation products and A-protein aggregates were seen in both protein preparations. This is probably due to improper storage during shipping of the protein. Biological Activity There are numerous studies which indicate that the presence of the outer-membrane A-protein layer on A. salmonicida suppresses (25) or enhances (26) phagocytosis by macrophages. In addition, it has

been demonstrated in vitro that peritoneal and head kidney macrophages derived from trout have the potential to internalize and kill A. salmonicida strains lacking the A-layer protein (A2) effectively but not A-layer positive (A1) strains. If the macrophages were activated in vivo, they showed an enhanced ability for internal degradation of virulent A1 species (25,27,28). The interaction of bacterial and recombinant Alayer protein with murine macrophages was studied by utilizing an assay patterned from the model described by Garduno et al. (26) and one modified in the laboratory of Professor A. Friedman of the Hebrew University. Our study was directed at determining the effect of A-protein on intracellular events that occur in primed macrophages. This was accomplished by measuring the cytotoxic product produced by peritoneal macrophages when exposed to proteincoated latex beads. Thioglycolate elicited macrophages exhibited a low level of activation (18% cytotoxicity) that was significantly increased (48% cytotoxicity) in the presence of latex beads. Incubation of the activated macrophages with soluble Aprotein did not result in further activation, demonstrating that the macrophages do not possess intrinsic receptors for this protein nor does the presence of small amounts of polysaccharide contamination account for the results obtained (Fig. 8). Coating of the latex beads with each of the three A-protein products resulted in an increase of cytoxicity (mean 6 SEM) from 48 to 91 6 0.39% for AtR-B, 89.5 6 0.39% for AtG-B and 85 6 0.24% for AtM-B. These results are highly significant by t test (P , 0.0001). The calculated cytotoxic effect of theses proteins was determined to be equivalent to the level of cytotoxicity arrived at when the fibroblastoid cell line was exposed to 1.04, 0.97, and 0.87 pg recombinant TNFa, respectively. There was no significant


403

difference between the two isolated and the recombinant A-proteins. In a parallel experiment, murine peritoneal macrophages were exposed to At3-96 bacterin containing an equivalent amount of protein. The cytotoxic level of the macrophage supernatant reached 97%, which is noticeably similar to the values obtained with the protein-coated beads. In both experiments (At3-96 bacterin and purified A-protein) the presence of A-protein did not interfere with

FIG. 8. Effect of 3 different A-protein preparations on the activation of mouse peritoneal macrophages. Mouse peritoneal macrophage were incubated with 1 mg/ml of either recombinant (AtR), glycine extracted (AtG), or membrane extracted (AtM) A-protein in the free form (■; black) or bound to latex beads ( ; grey) (AtR-B, AtG-B, and AtM-B). Uncoated beads and PBS were used as controls. Toxicity of substances secreted by the macrophages was examined on balb/c clone 7 fibroblastoid cells and determined by the MTT method. The degree of cell survival was computed by optical density read at 570 nm and results were converted into a measurement of cytotoxicity. Results shown are an average of triplicates with standard deviation. Cytoxicity arising from macrophages treated with protein coated beads was significantly greater than that of protein solutions alone (P , 0.0001) and beads alone.

phagocytosis nor was the protein toxic to the macrophages. Using this assay we have shown that murine peritoneal macrophages activated in vivo, using thioglycolate display increased “killing” ability when exposed to cell free A-protein and that the recombinant form of the protein is equally effective as the isolated bacterial forms. CONCLUSIONS

FIG. 7. (A) Western blot of recombinant and glycine extracted A-protein. 1.0 mg of each protein was resolved by SDS–PAGE on 12% acrylamide and blotted onto a nitrocellulose membrane. The blot was incubated with rabbit polyclonal anti A-protein IgG and detected by goat anti-rabbit peroxidase conjugate and ECL. (Lane 1) Glycine extracted protein; (lane 2) recombinant protein. (B) SDS–PAGE of 0.5 mg protein on 10% acrylamide gel under reducing conditions followed by Western blotting. (Lanes 2 and 3) Recombinant A-protein; and (Lanes 4 and 5) glycine-extracted A-protein. Blots were incubated with mAb 4B10 (lanes 2 and 4) and mAb 1B2 (lanes 3 and 5) and detected with goat anti-mouse IgG/IgM peroxidase conjugate and DAB/H 2O 2. (Lane 1) Molecular mass markers (103, 77, 48, 34, 21 kDa, top to bottom).

This paper describes a method for the large-scale production of recombinant monomeric A-layer protein genetically cloned from an achromogenic atypical A. salmonicida source. The protein produced showed a high degree of identity to typical A. salmonicida outer membrane protein isolated from diseased trout and salmon. Components of cell surface proteins are fundamental candidates for inclusion in vaccines as they form the initial contact between host and pathogen. They may also be necessary for subsequent stages of bacterial invasion. The recombinant A-protein can be readily produced in yields suitable for industrial production and thereby can be utilized as an acellular, nontoxic antigenic source for the stimulation of a pro-

404

MAURICE ET AL.

tective immune response in fish. The ability to produce this protein in the recombinant form will allow investigators to initiate strategies for the development of peptide subunit vaccines and directed mutation studies, which will aid in understanding the structurefunction and immunological role that A-protein plays in eliciting protective antibody formation. ACKNOWLEDGMENTS We express our thanks to Dr. I. Lachmann and Dr. U. Wagner (Institute for Zoology, University, Leipzig, Germany), as well as Dr. R. Bjornsdottir (The Foundation for Applied Research at University Tromso), for donating the monoclonal antibodies. We thank Shay Weiss (Dept. of Animal Sciences, Hebrew University of Jerusalem) for adapting the TNFa bioassay for our use.

REFERENCES 1. Håstein, T., Saltveit, S. J., and Roberts, R. J. (1978) Mass mortality among minnows Phoxinus phoxinus (L) in Lake Tveitevatn, Norway, due to an aberrant strain of Aeromonas salmonicida. J. Fish Dis. 1, 241–249. 2. Noga, E. J., and Berkoff, H. A. (1990) Pathological and microbiological features of Aeromonas salmonicida infection in the American eel (Anguilla rostrata). Fish Pathol. 25, 127–132. 3. Wilklund, T. (1990) Atypical Aeromonas salmonicida isolated from ulcers of pike, Esox lucius L. J. Fish Dis. 13, 541–544. 4. Wiklund, T., and Dalsgaard, I. (1995) Atypical Aeromonas salmonicida associated with ulcerated flatfish species in the Baltic Sea and the North Sea. J. Aquat. Anim. Health. 7, 218 –224. 5. Bo¨hm, K. H., Fuhrmann, H., Schlotfeldt, H. J., and Ko¨rting, W. (1986) Aeromonas salmonicida from salmonids and cyprinids– serological and cultural identification. J. Vet. Med. B 33, 777– 783. 6. Elliot, D. G., and Shotts, E. B., Jr. (1980) Aetiology of an ulcerative disease in goldfish Carassius auratus (L.): microbiological examination of diseased fish from seven locations. J. Fish Dis. 3, 133–143. 7. Trust, T. J., Khouri, A. G., Austen, R. A., and Ashburner, L. D. (1980) First isolation in Australia of atypical Aeromonas salmonicida. FEMS Microbiol. Lett. 9, 39 – 42. 8. Whittington, R. J., Gudkovs, N., Carrigan, M. J., Ashburner, L. D., and Thurstan, S. J. (1987) Clinical, microbiological and epidemiological findings in recent outbreaks of goldfish ulcer disease due to atypical Aeromonas salmonicida in Southeastern Australia. J. Fish Dis. 10, 353–362. 9. Belland, R. J., and Trust, T. J. (1987) Cloning of the gene for the surface protein array of Aeromonas salmonicida and evidence linking loss of expression with genetic deletion. J. Bateriol. 169, 4086 – 4091. 10. Ishiguro, E. E., Kay, W. W., Ainsworth, T., Chamberlain, J. B., Buckley, J. T., and Trust, T. J. (1981) Loss of virulence during culture of Aeromonas salmonicida high temperature. J. Bacteriol. 148, 333–340. 11. Munn, C. B., Ishiguro, E. E., Kay, W. W., and Trust, T. J. (1982) Role of surface components in serum resistance of virulent Aeromonas salmonicida. Infect. Immun. 36, 1069 –1075. 12. Phipps, B. M., and Kay, W. W. (1988) Immunoglobulin binding by the regular surface array Aeromonas salmonicida. J. Biol. Chem. 263, 9298 –9303.

13. Phipps, B. M., Trust, T. J., Ishiguro, E. E., and Kay, W. W. (1983) Purification and Characterization of the Cell Surface Virulent A Protein from Aeromonas salmonicida. Biochemistry 22, 2934 – 2939. 14. Bjørnsdottir, R., Eggset, G., Nilsen, R., and Jørgensen, T. O. (1992) The A-layer protein of Aeromonas salmonidida: further characterization and a new isolation procedure. J. Fish Dis. 15, 105–118. 15. Evensberg, D., Van Boxtel, R., Lugtenberg, B., Schurer, F., Blommaert, J., and Bootsma, R. (1982) Cell surface of the fish pathogenic bacterium Aeromonas salmonicida. I. Relationship between autoagglutination and the presence of a major cellenvelope protein. Biochim. Biophys. Acta 684, 241–248. 16. Chu, S., Cavaignac, S., Feutrier, J., Phipps, B. M., Kostrzynska, M., Kay, W. W., and Trust, T. J. (1991) Structure of the tetragonal surface virulence array protein and gene of Aeromonas salmonicida. J. Biol. Chem. 266, 15258 –15265. 17. Noonan, B., and Trust, T. J. (1997) The synthesis, secretion and role in virulence of the paracrystalline surface protein layers of Aeromonas salmonicida and A. hydrophila. FEMS Microbiol. Lett. 154 (1), 1–7. 18. Ausubel, F. M., Brent, R., Kingston, R. G., More D. D., Seidman, J. G., Smith, J. A., and Struhl, J. A. “Current Protocols in Molecular Biology,” Massachusetts General Hospital, Wiley. 19. Sambrook, J., Fritsch E. F., and Maniatis, T. (1989) “Molecular Cloning: A Laboratory Manual,” Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 20. Laemmli, U. K. (1970) Cleavage of structural proteins during the assembley of the head of the bacteriophage T 4. Nature 227, 680 – 685. 21. Ko¨hler, G., and Milstein, C. (1975) Continuous cultures of fused cells secreteing antibody of predefined specificity. Nature 256, 495– 497. 22. Wagner, U., Lachmann, I., Ha¨dge, D., and Drysler, K. (1997) Development and characterization of monoclonal antibodies specific for two different exoproteases of Aeromonas salmonicida. J. Fish Dis. 20, 223–235. 23. Lachmann, I., Wagner, U., Ha¨dge, D., and Dro¨bler, K. (1998) Selective analysis of humoral immune responses to native bacterial antigens using murine monoclonal antibodies in sandwich ELISA. In “Methodology in Fish Diseases Research” (A.C. Barnes, G.A. Davidson, M.P. Hiney, D. McIntosh, Eds.). 24. Maurice, S., Tinman, S., Sakal, E., Gertler, A., and Bejerano, I. (1995) Vaccination of fish using antigen isolated from outer membranes of Aeromonas salmonicida. European Aquacult. Soc. 25. 25. Oliver, G., Eaton, C. A., and Campbell, N. (1986) Interaction between Aeromonas salmonicida and peritoneal macrophages of brook trout (Salvelinus fontinalis). Vet. Immunol. Immunopathol. 12, 223–234. 26. Garduno, R. A., Lee, E. J. Y., and Kay, W. W. (1992) S-LayerMediated Association of Aeromonas salmonicida with Murine Macrophages. Infect. Immun. 60 (10), 4373– 4382. 27. Chung, S., and Secombes, C. J. (1987) Activation of rainbow trout macrophages. J. Fish Biol. 31 (Supplement A) 51–56. 28. Graham, S., Jeffries, A. H., and Secombes, C. J. (1988) A novel assay to detect macrophage bactericidal activity in fish: Factors influencing the killing of Aeromonas salmonicida. J. Fish Dis. 11, 389 –396.