Bacterial expression, purification and functional characterization of a recombinant chimeric Fab derived from murine mAb BCF2 that neutralizes the venom of the scorpion Centruroides noxius hoffmann

Toxicon 43 (2004) 43–51 www.elsevier.com/locate/toxicon

Bacterial expression, purification and functional characterization of a recombinant chimeric Fab derived from murine mAb BCF2 that neutralizes the venom of the scorpion Centruroides noxius hoffmann Barbara Selisko1, Gabriela Cosı´o, Consuelo Garcı´a, Baltazar Becerril, Lourival D. Possani, Eduardo Horjales* Department of Molecular Medicine and Bioprocesses, Institute of Biotechnology/National Autonomous University of Mexico (UNAM), Av. Universidad 2001, Apto. Postal 510-3, Cuernavaca, Morelos 62210, Mexico Received 30 April 2003; revised 6 September 2003; accepted 10 October 2003

Abstract The murine monoclonal antibody BCF2 is able to neutralize the venom of the scorpion Centruroides noxius Hoffmann. A chimeric Fab of BCF2 (chFab-BCF2) comprising the variable regions of murine BCF2 and human constant regions was assembled. chFab-BCF2 was expressed as a soluble and functional protein in the periplasmic space of Escherichia coli. An expression yield of 1 mg/l was reached by combination of late-log-phase induction, rich culture medium, low expression temperature and addition of sucrose (0.3 M) to the culture medium. The addition of sucrose induced secretion of 60% of the protein into the medium. After expression for 23 h, a novel process was used to release the remaining periplasmic protein in situ consisting in the addition of lysozyme and sucrose up to 0.6 M (20%) directly to the culture medium. chFab-BCF2 was recovered by ammonium sulfate precipitation and purified in a single step by affinity chromatography using anti-human antiF(ab0 )2 IgG coupled to Sepharose-proteinG. Pure chFab-BCF2 maintained a similar nanomolar affinity as BCF2 to its cognate antigen, the Naþ-channel-affecting toxin Cn2. Recombinant chFab-BCF2 was able to neutralize Cn2 in vivo even at a molar ratio of 1:1, as well as the whole venom of C. noxius. Thus, it is a promising candidate to be used as a specific and efficient recombinant antidote against scorpion stings. q 2003 Elsevier Ltd. All rights reserved. Keywords: Antibody expression; Centruroides noxius; Chimeric Fab; Neutralization; Scorpion; Toxin

1. Introduction

* Corresponding author. Address: Department of Molecular Recognition and Structural Biology, Institute of Biotechnology, National Autonomous University of Mexico (UNAM), Av. Universidad 2001, Apto. Postal 510-3, Cuernavaca, Morelos 62210, Mexico. Tel.: þ52-777-329-1616; fax: þ52-777-317-2388. E-mail address: [email protected] (E. Horjales). 1 Present address: Architecture et Fonction des Macromole´cules Biologiques (AFMB), CNRS-UMR 6098 et Universite´ AixMarseille I et II, Campus de Luminy, Marseille 13288, France. 0041-0101/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.toxicon.2003.10.015

The murine monoclonal antibody (mAb) BCF2 neutralizes the toxic effect of its cognate antigen Cn2, a toxin from the venom of the Mexican scorpion Centruroides noxius Hoffmann (C. noxius) that affects Naþ channels (Zamudio et al., 1992; Dehesa-Da´vila et al., 1996; Licea et al., 1996; Pintar et al., 1999). BCF2 binds to Cn2 with nanomolar affinity (Zamudio et al., 1992). It inhibits the binding of Cn2 to its target channel in rat brain synapyosomes (Zamudio et al., 1992). Thus, it can be assumed that BCF2 recognizes an epitope that overlaps with, or sterically obstructs

44

B. Selisko et al. / Toxicon 43 (2004) 43–51

the region that binds to the Naþ channel. Additionally, BCF2 was found to neutralize the toxic effect of the whole venom of C. noxius (Licea et al., 1996). Scorpion venom consists of a complex mixture of proteic components (Possani, 1984) that act on excitable cells causing respiratory and circulatory problems (Ismail, 1995). The major toxic agents are Naþ-channel-affecting toxins of around 7 kDa (Possani et al., 1999). It has been proposed that the neutralization of the venom by BCF2 is the result of the neutralization (Selisko et al., 1999). Cn2 shows the highest concentration (6.8% of total protein) of all toxins in the venom of C. noxius and a particularly high toxicity (LD50 0.25 mg/20 g average body weight of mice of the strain CD1) (Licea et al., 1996). The hypothesis of BCF2 being a specific neutralizing agent of Cn2 is supported by the observation that neither the mammal-specific toxins Cn3 and Cn4 nor the arthropod-specific toxin Cn1 of C. noxius (82, 86 and 54% identity to Cn2, respectively), are able to inhibit the binding of BCF2 to Cn2(Selisko et al., 1999). Recent experiments showed that two other arthropodspecific toxins, Cn5 and Cn10 (60 and 53% identity), were also not recognized by BCF2 (Garcı´a et al., 2003). Nevertheless, BCF2 showed some cross-reactivity with components of other venoms of scorpions of the genus Centruroides (Zamudio et al., 1992), which might be close homologues of Cn2. In conclusion, BCF2 is considered a promising candidate for a new generation of antidotes against scorpion stings consisting in highly specific, neutralizing antibodies produced by expression in bacterial cultures. They can be expressed as recombinant antigenbinding fragments (Fab) or single-chain variable fragments (scFv) of the original murine antibodies or, alternatively, their chimeric or humanized versions. The antidote which is currently in use in Mexico, consists of equine, bivalent F(ab0 )2 fragments produced from anti-serum of horses that were hyper-immunized with the soluble fraction of gland macerates from Centruroides scorpions (Calderon-Aranda et al., 1996). The polyclonal antibodies are digested with pepsin and purified to remove irrelevant digestion products and serum proteins. This last generation of scorpion-sting antidote has been proven to be safer and more efficient than earlier preparations. Nevertheless, a recombinant antidote that consists exclusively of relevant (anti-venom) antibodies and even better, of chimeric or humanized molecules, is expected to provide higher efficacy. A lower dose of foreign protein has to be injected, and the risk of causing unwanted immunological reactions will be considerably lower. In this study, we report the expression in Escherichia coli of the chimeric Fab of BCF2 (chFab-BCF2) comprising the variable regions of murine BCF2 and human constant regions. chFab-BCF2 was produced as a soluble and active protein in the periplasm using a rich medium, induction at late log-phase of growth, and expression at low temperature in presence of sucrose. It was purified by ammonium sulfate precipitation followed by affinity chromatography using an

anti-human anti-F(ab0 )2 antibody coupled to protein G on Sepharose. Its relative affinity in comparison to BCF2 was determined, and its neutralizing capacity was tested in vivo.

2. Experimental 2.1. Construction of the chFab-BCF2 expression vector Ventwpolymerase (New England Biolabs) was used for PCR. Restriction enzymes, kinase, phosphatase and T4DNA ligase were obtained from Boehringer or New England Biolabs. If not stated otherwise, standard protocols were used (Sambrook et al., 1989). Gene manipulations were done in E. coli strain XL1 Blue (Stratagene). DNA sequences were obtained using an automated DNA sequencer Perkin– Elmer/Applied Biosystems Model 37718EI. The variable regions of the light and heavy chain of BCF2 were obtained by PCR from lgt11 phages carrying the cDNA of the mAb (Selisko et al., 1999). Oligonucleotides Vk5 (50 AAGTGAGCTCGACATTGTGTTGACCCAATCTCC-30 ) adding a Sac I restriction site and Vk3 (5 0 CGCCCGGTCCGTTTCAGCTCCAGGTTGGT-30 ) adding a Rsr II restriction site were used to amplify the variable region of the light chain. Oligonucleotides VH5 (50 GATCCTCGAGGTTCAGCTGCAACAGTCTGGTCCT G-30 ) and VH3 (50 -CGAGGTCGACGCTGAGGAGACGG TGACTGAGGT-30 ) were used to amplify the variable region of the heavy chain. VH5 added a Xho I restriction site and the first three amino acids Glu – Val – Gln (nucleotides in bold) that were not present in the original clone (Selisko et al., 1999). VH3 contained a Sal I restriction site. The constant regions were obtained by PCR from pComb3TT, (kindly provided by Carlos Barbas III, Scripps-Institute, La Jolla, USA), which contains the human constant regions CH1 of IgG1 and Ck (Barbas et al., 1991). Oligonucleotides CH5 (50 CCCAGGATCCGCGTCGACCAAGGGCCCATCGGTC TTC-30 ) adding a BamH1 and a Sal I site, and CH3 (50 -CACC ACTAGTTTTGTCACAAGATTTG-30 ) containing a Spe I site were used to amplify CH1. Ck5 (50 -GTATCGGTCC GTGATCAGGACAGCAAAGACAGCACC-30 ) adding a Rsr II restriction site and Ck3 (50 -GCGCTCTAGAAACAC TCTCCCCTGTTGAAGC-30 ) adding a Xba I site were used for Ck. The cloning vectors, plasmid pM849 and phagemid pM846, containing the light and heavy chains, respectively, of a chimpanzee Fab fragment were a gift from Luc Aujame (Pasteur Me´rieux Se´rums et Vaccins, Marcy l’Etoile, France). Plasmid pM849 was cut with Rsr II and Xba I. The purified vector was used to clone the previously digested Ck5 – Ck3 PCR product. Vector pM849 containing the Ckregion was digested with Sac I and Rsr II, purified and used to ligate Vk. Phagemid pM846 was cut with BamH1 and Spe I and the digested and purified PCR product CH5 – CH3

B. Selisko et al. / Toxicon 43 (2004) 43–51

ligated into it. The resulting vector was digested with Xho I and Sal I and dephosphorylated to prevent re-ligation of the compatible restriction sites. The VH5– VH3 PCR product was purified and ligated into pM846 containing the CH1region. For the generation of the expression vector, the strain D1210HP (Geoffroy et al., 1994; Sodoyer et al.,1996) was transformed with plasmid pM849-chFab-BCF2-light. Phages were generated from XL1 Blue cells carrying the phagemid pM846-chFab-BCF2-heavy using the helper phage R408 (Stratagene). The phage were precipitated with 2.5 M NaCl/20% PEG 8,000, re-suspended in buffer STE (Sambrook et al., 1989) and the titre was determined on XL1 Blue cells to be 6.3 £ 1011 pfu/ml. An overnight culture of pM849-chFab-BCF2-light in D1210HP was diluted to an absorbance at 600 nm (A600) of 0.01 with 2 £ YT (Sambrook et al., 1989) containing 2% glucose, and grown at 30 8C until an A600 of around 0.6. Phage pM846chFab-BCF2-heavy was added at an infection multiplicity of 10 and left for 1 h at 30 8C without agitation. Subsequently, the culture was transferred to 42 8C for 30 min and then returned to 30 8C for 10 min to recover. It was diluted 10-fold in 2 £ YT containing 2% glucose, kanamycin and carbenicillin and left to grow for 24 h at 30 8C. Samples were taken at different intervals of time to analyze recombinant vectors. The process was also monitored on agar plates with kanamycin, carbenicillin or both antibiotics. Vectors were prepared from the liquid culture and from individual cultures of single colonies obtained from agar plates with both antibiotics. The vector composition was examined by restriction analysis. The recombinant vector (pMrec-chFab-BCF2) was obtained from a liquid culture grown from a single colony that contained only this plasmid. It was digested with Hind III at a single restriction site originating from the plasmid. Subsequently, the band of 7.9 kb was excised from the agarose gel, purified, religated and transformed into the nonsuppressor strain TOPP2 (Stratagene). 2.2. Expression of chFab-BCF2 and recovery of periplasmic proteins (final protocol) An overnight-culture of TOPP2 [pMrec-chFab-BCF2] was grown at 30 8C in 2 £ YT (Sambrook et al., 1989) supplemented with 2% glucose, 80 mg/ml carbenicillin and 25 mg/ml kanamycin. The culture was 100-fold diluted in 2 £ YT supplemented with 0.1% glucose and antibiotics, and grown at 30 8C to an A600 of 1.2. Sucrose (stock 1.8 M in 2 £ YT) was added to a final concentration of 0.3 M (10.3%) and the culture grown again to an A600 of 1.2. IPTG was added to 0.5 mM and the culture was incubated at 23 8C. After 24 h of expression (A600 3.3 – 4.0), sucrose was added to a final concentration of 20%, EDTA to 1 mM and lysozyme to 0.32 mg per ml of culture volume, which corresponds to 19 mg/g of cellular pellet. After 1 h incubation on ice (shaking), spheroblasts were removed by

45

centrifugation at 15,300 £ g (60 min, 4 8C). Proteins were precipitated from the cleared supernatant by adding solid ammonium sulfate to a final concentration of 70% saturation. After 30 min stirring on ice, and overnight precipitation at 4 8C, the protein was recovered by centrifugation at 15,300 £ g (30 min, 4 8C). The mixture was re-suspended in the loading buffer of the affinity column (12.8% buffer A (50 mM citric acid, 500 mM NaCl) and 87.2% buffer B (100 mM Na2HPO4 þ 500 mM NaCl) giving pH 7.4) containing 0.05% NaN3. The sample was dialyzed extensively against loading buffer with 0.05% NaN3. 2.3. Preparation of the affinity column The affinity column was prepared according to Barbas et al. (2001). In brief, Gammabind G Sepharose (Amersham-Pharmacia) was washed five times with 0.01 M sodium phosphate buffer, 0.25 M NaCl, pH 7.6. Two ml of beads were mixed with 4 mg of goat anti-human antiF(ab0 )2 IgG (Pierce, Cat Nr 31132) and incubated under shaking for 1 h at room temperature. The beads were washed twice with 200 mM sodium borate buffer, pH 9.0, and 104 mg dimethylpimelimidate-HCl (Pierce) were added to cross-link the antibody and protein G covalently. The mixture was left for 30 min without shaking at room temperature. Subsequently, the beads were washed with 0.2 M ethanolamine pH 8.0, and left in 20 ml 0.2 M ethanolamine for 2 h at room temperature. Finally, the resin was washed with PBS, pH 7.4 and stored in PBS with 0.05% NaN3 at 4 8C. 2.4. Purification and concentration of chFab-BCF2 Periplasmic protein re-suspended in loading buffer (ca. 30 ml from 500 ml culture with a protein concentration of ca. 6 mg/ml) was cleared by centrifugation at 10,000 £ g for 30 min. The chFab-BCF2 concentration at this stage was around 20 mg/ml. The solution was loaded on the column (bed volume 2 ml) by re-circulation (three times at 1.5 ml/min or overnight at 0.5 ml/min). The column was washed with loading buffer until the absorbance at 280 nm was reduced to base-line level. Elution buffer (89.2% buffer A (see above) and 10.8% buffer B (see above), giving pH 2.3 was applied, fractions of 0.5 ml collected, and 0.1 ml of 1 M Tris, pH 9.0 added immediately to each fraction. The column was washed and re-equilibrated with loading buffer containing 0.05% NaN3. Fractions containing chFab-BCF2 were pooled (concentration ca. 30 mg/ml) and dialyzed against 10 mM Tris pH 7.5 containing 50 mM NaCl, or PBS. Subsequently, the chFab-BCF2 was concentrated to ca. 250 mg/ml using dialysis membranes with a cut-off of 3.5 kDa (Spectrum Laboratories) and solid PEG 8,000 (Hampton Research) as absorbing material outside the bag. A final concentration of 1.5 mg/ml was achieved with

46

B. Selisko et al. / Toxicon 43 (2004) 43–51

Centricon 30 (Millipore), although with a considerable loss of material due to adsorption to the filter membrane. 2.5. Centruroides noxius venom and purification of toxin Cn2 The venom from scorpions of the species C. noxius Hoffmann was obtained by electric stimulation as described (Possani, 1984). The dried venom was dissolved in water and centrifuged. The soluble part was freeze-dried and kept at 2 20 8C. Cn2 was purified by Sephadex G-50 gel filtration and cation-exchange chromatography, as described in Possani, 1984 and Zamudio et al., 1992.

pH 9.2). After washing and blocking, they were incubated for 10 or 20 min at 4 8C with a pre-incubated mixture of equal volumes of serial dilutions of Cn2 (0.5– 8.0 £ 1029 M) and fixed concentrations of BCF2 (3.8 £ 1029 M Ab sites corresponding to 0.28 mg/ml) or chFab-BCF2 (3.4 £ 1029 M Ab sites corresponding to 0.17 mg/ml). The pre-incubation was done for 15 h at 4 8C in PBS containing 0.05% Tween-20 and 0.1% BSA. Antibody bound to Cn2 was detected by incubation (1 h at 37 8C) with 50 ml of goat anti-human antiF(ab0 )2 IgG for chFab-BCF2 (diluted 1:1,500 to final concentration of 26.7 ng per 50 ml) and goat anti-mouse anti-F(ab0 )2 IgG for mAb (diluted 1:1,500), both coupled to horse radish peroxidase. Enzyme activity was measured as described earlier.

2.6. ELISA for estimation of chFab-BCF2 concentration 2.8. In vivo neutralization assays Polyvinyl plates (Costar, Cambridge, MA, USA) were coated overnight at 4 8C with 50 ml of toxin Cn2 (0.3 mg in 20 mM sodium bicarbonate, pH 9.2). After washing three times with 20 mM sodium phosphate buffer, pH 7.8, 0.15 M sodium chloride (PBS solution), the wells were blocked with 200 ml 1% BSA in PBS for 1 h at 37 8C. Plates were washed six times with PBS containing 0.01% Tween-20 (TPBS) and incubated overnight at 4 8C with 50 ml of test solution containing chFab-BCF2 or purified mAb BCF2 (six standard dilutions from 0.5 to 5 ng per 50 ml were included in each plate). Afterwards, the plates were washed with TPBS (2 £ ) and PBS (1 £ ). Antibody bound to Cn2 was detected by incubation (1 h at 37 8C) with 50 ml of goat antihuman anti-F(ab0 )2 IgG (Pierce, Cat Nr 31414, diluted 1:500 to a final concentration of 80 ng per 50 ml) for chFab-BCF2, and goat anti-mouse anti-F(ab0 )2 IgG (Pierce, Cat Nr 31436, diluted 1:500) for mAb, both coupled to horse radish peroxidase. After washing with TPBS (3 £ ) and PBS (3 £ ), the peroxidase activity retained in the wells was detected by adding 100 ml of a solution of 2,20 -azinobis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS, Pierce, 10 mg) and H2O2 (20 ml) in sodium citrate buffer solution (mixture of 10 ml. 50 mM citric acid and 10 ml. 50 mM sodium citrate) and reading absorbance at 405 nm after 20 min using a microplate reader (Bio-Rad, model 550). The amount of chFab-BCF2 in the samples was estimated using BCF2 as a standard considering the difference of molecular mass and valence of the Fab fragment and the mAb. 2.7. Competition ELISA to estimate relative affinity of chFab-BCF2 and BCF2 The concentration of SDS-PAGE-pure chFab-BCF2 and BCF2 was determined by absorbance at 280 nm, using an absorptivity coefficient A0.1% (1 mg/ml) of 1.5 ml mg21 cm21 for both. Plates, reagents and washing conditions were the same as described above. Wells were coated overnight at 4 8C with 50 ml of toxin Cn2 (0.03 mg in 20 mM sodium bicarbonate,

For the in vivo neutralization assays, groups of six albino mice (strain CD1, female, around 4 weeks old and of 18 – 20 g body weight) were used. Toxin Cn2, or C. noxius venom, was mixed with chFab-BCF2, both in PBS, and taken to a final volume of 250 ml. For final concentrations see Table 1. The amount of toxin Cn2 or whole venom, corresponding to one LD50, was adjusted for each mouse according to its weight. Although we used the same strain of mice as in Licea et al. (1996), for the present study, the LD50s of toxin and whole venom were newly determined because the mice were obtained from a different supplier. The new LD50s were 0.195 mg per 20 g mouse body weight for Cn2 and 2.5 mg per 20 g for the whole venom. The mixtures of Cn2 or venom and chFab-BCF2 were incubated for 30 min at room temperature with mild shaking. Thus, the antibody-toxin recognition is performed in vitro, whereas the neutralizing capacity was tested with living animals (in vivo). Each sample was applied subcutaneously, and the percentage of survival in each group was read at 24 h after

Table 1 Protection against toxin Cn2 and whole venom of Centruroides noxius Number of ChFab-BCF2 One LD50 of mice tested (mg) 6 6 6 12 6 6 6

– 6.25 3.75 1.25 – 12.5 6.25

Molar ratio Survival chFab/toxin (%)

Cn2 – Cn2 5:1 Cn2 3:1 Cn2 1:1 Whole venom – Whole venom – Whole venom –

33.3 100 100 91.6 50.0 100 100

Cn2 or whole venom and chFab-BCF2 were pre-incubated for 30 min prior to injection. One LD50 of toxin Cn2 was determined as 0.195 mg/20 g mouse body weight, for subcutaneous injection. For whole venom it was 2.5 mg/20 g. Survival rates were measured 24 h after injection.

B. Selisko et al. / Toxicon 43 (2004) 43–51

injection. A control group was injected only with toxin or venom in PBS. In the treated sample, after 24 h observation the intoxication symptoms (when present) start to disappear and the surviving mice behave normally, as the control noninjected mice, suggesting good health conditions. 3. Results 3.1. Construction of the chFab-BCF2 expression vector The light and heavy chain of the chimeric Fab fragment of BCF2 (chFab-BCF2) were cloned in two different vectors that allow the generation of combinatorial phage-display libraries of Fab fragments and the subsequent direct periplasmic expression of selected fragments [Geoffroy et al., 1994; Sodoyer et al., 1996]. The expression vector containing the light and heavy chains of chFab-BCF2 was obtained by an in vivo recombination process shown in Fig. 1. The vector containing the light and heavy chains of chFab-BCF2 was obtained from an in vivo recombinant process as shown in Fig. 1. The formation of recombinant vector (7.9 kb) was observed in DNA preparations from

47

cells of the liquid culture. Single colonies randomly selected from agar plates with both antibiotics contained the original plasmid as well as the recombinant vector or (rarely and) a vector of ca. 6.5 kb (data not shown). Our attempts to clearly elucidate the nature of the 6.5-kb product were not successful. Since it was observed before applying the heat shock, it might represent a product of recombination and deletion using homologous sequence stretches like pelB or the origins of replication of both vectors. The recombinant vector was stable upon transformation and repeated growth. DNA sequencing was carried out to verify correct recombination, the presence of the promoter regions, the pelB signal peptide sequences and the complete chFabBCF2 light-chain and heavy-chain sequences. The translated peptide sequence of Ck is closest to Kabat data base entry ID 013685 with a single change in amino acid position 203 from Ser to Leu. It was verified that this change originated from vector pComb3TT. A negative effect of this substitution on the function of the recombinant Fab is very unlikely because of its distance to the paratope (see also relative affinity measurements below). After the final cysteine, the light chain contains an extra residue, a phenylalanine, at the C-terminus (see primer Ck3 in Section 2). The CH1 sequence corresponds to Kabat data base entry ID 013487. After residue Cys 225 (Kabat numbering) of the CH1 hinge region the protein continues until Thr 229 and finishes with an additional serine. Cys225 forms a disulfide bridge to the cysteine of the Ck region. The mature recombinant ch-Fab-BCF2 consists of a light and a heavy chain of 24.1 and 24.4 kDa, respectively. 3.2. Expression and purification of chFab-BCF2

Fig. 1. Maps of the vectors. Maps of phagemid pM846-chFab-BCF2heavy, plasmid pM849-chFab-BCF2-light and recombinant phagemid pMrec-chFab-BCF2. Both starting vectors contain the lac promoter (lacZpo), the signal peptide sequence pelB and the recombination sequences, attP in the phagemid and attB in the plasmid. Upon recombination, attL and attR sites are generated in pMrec-chFab-BCF2 phagemid. ColE1 and p15 correspond to origins of replication, f1 to the origin of replication of phage f1. Km-R and Ap-R correspond to the genes conferring kanamycin- or ampicillinresistance, respectively. Vk and Ck represent the variable and constant regions of the kappa light chain. VH and CH indicate the corresponding variable and constant regions of the heavy chain. GIII corresponds to the gene coding for phage protein pIII.

For expression of chFab-BCF2 in TOPP2 E. coli cells, the periplasmic fraction of proteins was recovered. In our initial preparations we applied an osmotic shock by resuspension of the cells in PBS containing 1 M NaCl (using 1/50 of culture volume). The protein was detected by Western blot using an anti-human anti-F(ab0 )2 IgG. After non-reducing SDS-PAGE, the main expression product was a protein of ca. 48 kDa, as shown in Fig. 2A. This band was not present in non-transformed cultures (not shown). The presence of chFab-BCF2 even when no IPTG was added (Fig. 2A, lane 2) can be explained by the fact that the suppression of expression of recombinant proteins under a lacZ promoter by the lacI-product is never complete. Furthermore, it can be appreciated in Fig. 2A (lane 3) that after 23 h of expression the recombinant periplasmic protein is partially released into the culture medium. By ELISA we verified that the recombinant protein was able to recognize Cn2 and should thus represent the chFab-BCF2(data not shown). Moreover, there was no recognition of chFab-BCF2 by goat anti-mouse anti-F(ab0 )2 IgG or of BCF2 by goat antihuman anti-F(ab0 )2 IgG. As can be seen in Fig. 2A, lane 1, there were a considerable number of smaller bands and one band of

48

B. Selisko et al. / Toxicon 43 (2004) 43–51

Fig. 2. Expression and recovery of chFab-BCF2. (A) Western blot of expression of chFab-BCF2 at 30 8C for 23 h. Proteins were separated by SDS-PAGE under non-reducing conditions and transferred onto nitrocellulose. The blot was developed with goat anti-human anti-F(ab0 )2 IgG. Lane 1, 3-fold concentrated sample of periplasmic fraction from an induced culture corresponding to 1.5 ml of culture; lane 2, same as lane 1 but from a non-induced culture; lane 3, 13 ml of supernatant from the induced culture. The positions of molecular mass markers are indicated. (B) SDS-PAGE under non-reducing conditions (coomassie stained). Demonstration of in situ release of periplasmic proteins after expression at 23 8C for 24 h in the presence of 0.3 M sucrose: lane 1, molecular mass markers; lane 2, 10 ml of culture medium after 24 h expression (5fold concentrated, ca. 1 mg/ml protein); lane 3, 10 ml of culture medium after in situ lysozyme treatment (5-fold concentrated, ca. 2.5 mg/ml protein).

higher molecular weight recognized by the anti-human anti-F(ab0 )2 IgG. The smaller bands may represent protease-cleaving products of misfolded light or heavy chains, while the higher-molecular-weight product may be generated by incorrect disulfide-bond formation. One way to avoid the problem of accumulation of misfolded protein is to slow down the expression and favour the rate limiting folding process by lowering the temperature during the expression phase (Wall and Plu¨ckthun, 1995). Indeed, when we used a temperature of 23 8C during expression the additional bands were no longer detected in Western blot (data not shown) and the yield improved. The improvement of the yields was also a goal in our work, since the first approximate expression yields that we obtained using osmotic shock, were as low as 50 mg soluble and active protein from one liter of culture. The yield of chFab-BCF2 was estimated by a semi-quantitative direct ELISA (see Section 2). These results were later confirmed using purified chFab-BCF2. Addition of sucrose to the medium, improved the yield of periplasmic proteins (Kipriyanov et al., 1997). In our case, we observed a further increase in yield and found that 60% of the protein was exported to the medium in the presence of 0.3 M sucrose. Furthermore, 24 h was determined to be the optimum time of expression. In order to reach maximum cell density, we

grew the culture at 30 8C to late log-phase (A600 1.2) before induction using a rich medium. The optimal IPTG concentration was found to be 0.5 mM IPTG. Using this improved protocol, we obtained an expression yield of about 1 mg protein per liter of bacterial culture. As 60% of the protein is exported to the culture medium, we adapted the lysozyme treatment to function as an in situ method. The outer membrane of the cells was made porous by adding lysozyme to the medium (approx. 19 mg lysozyme per g cellular pellet), augmenting the sucrose concentration of the medium to 20% and adding EDTA to 1 mM. In Fig. 2B, lane 2, the electrophoretic profile of the proteins present in the culture medium after 24 h expression at 23 8C is shown. Lane 3 shows the protein profile after the release of the periplasmic proteins into the culture medium. The increase of protein leakage into the culture medium can be readily observed. The concentration of the chFab-BCF2 at this stage is around 5 mg/ml and it is therefore not visible as a pronounced band in a SDS-PAGE-protein profile. Using our semi-quantitative ELISA method we verified the recovery yield of periplasmic protein by opening the outer membrane in situ was similar to the one obtained by the classic lysozyme method, where the periplasmic protein is recovered after first separating cells and culture medium. The purified and concentrated chFab-BCF2 migrates as a single band of 48 kDa in SDS-PAGE under non-reducing conditions (Fig. 3A, lane 1). Thus, the inter-chain disulfide bond is properly formed upon expression. Under reducing conditions two bands are produced that correspond to 28 and 23 kDa (Fig. 3A, lane 4). This difference in migration of heavy and light chain in SDS-PAGE, even though both have a relative molecular mass of ca. 24 kDa, may result from the strong difference of their pIs, which are 8.87 and 4.81 for the heavy and light chains, respectively. In Fig. 3B a native polyacrylamide gel is shown demonstrating that

Fig. 3. Analysis of purified, recombinant chFab-BCF2. (A) SDSPAGE (coomassie stained); lane 1, purified chFab-BCF2 under nonreducing conditions; lanes 2 and 3, molecular mass markers; lane 4, purified chFab-BCF2 under reducing conditions. (B) Native-PAGE (10%) (coomassie stained) of purified chFab-BCF2 and mAb BCF2; lanes 1 and 2, 4 and 2 mg, respectively, of purified chFab-BCF2; lane 3, 4 mg of mAb BCF2.

B. Selisko et al. / Toxicon 43 (2004) 43–51

Fig. 4. Competition ELISA experiments to determine relative affinities of chFab-BCF2 and mAb BCF2. Serial dilutions of Cn2 were incubated with fixed concentrations of chFab-BCF2 or BCF2 for 15 h at 4 8C. The pre-incubated mixtures were applied for the given time periods to ELISA plates with immobilized Cn2. Antibody binding was estimated using anti-F(ab0 )2 antibodies coupled to horse-radish peroxidase (see Section 2). Data points are means of two independent determinations. The data were fitted to logistic dose-response functions according to DeLean et al. (1978).

the chFab-BCF2 does not form oligomers under the applied conditions.

3.3. Functional characterization of the recombinant chFab-BCF2 In direct ELISAs chFab-BCF2 was shown to recognize Cn2, thus demonstrating that the chimeric, recombinant Fab of BCF2 was functional. We then estimated the relative affinity of the chFab-BCF2 in comparison with the parent mAb BCF2 using a competitive ELISA. The inhibition curves are shown in Fig. 4. The data were fitted to logistic dose-response functions (DeLean et al., 1978). The calculated IC50 values are well in the range of 3 £ 1029 M for both BCF2 and chFab-BCF2. The in vivo neutralization capacity of the chimeric recombinant antibody for both the Cn2 toxin and the whole venom was demonstrated (Table 1). Interestingly, chFab-BCF2 was able to neutralize toxin Cn2 over 90% of the challenged mice, even when a molar ratio of 1:1 was used. At higher molar ratios, we observed 100% protection.

4. Discussion In the present study we report the expression and purification of a functional, chimeric Fab that maintains the

49

original apparent affinity of its parent antibody BCF2. Additionally, the chimeric Fab is able to neutralize in vivo the toxic effect of the venom of the scorpion C. noxius Hoffmann. The expression of antibody fragments in E. coli offers facile manipulation, inexpensive culture media, fast growth to high cell density and the possibility to produce folded molecules with the intact disulfide bonds in the periplasmic space. On the other hand, a highly efficient expression of well-folded, functional antibody fragments is difficult to achieve in comparison to other recombinant proteins (Skerra, 1993; Verma et al., 1998). Numerous ways of improvement have been reported, either changing culture conditions such as temperature (Wall and Plu¨ckthun, 1995) and growth medium (Kipriyanov et al., 1997), utilizing fedbatch fermentation (Horn et al., 1996), or introducing specific alterations of the protein sequence as mutations in the framework region of Fv (Forsberg et al., 1997). Nevertheless, for each particular mAb a strategy has to be established to obtain reasonable yields of functional protein (Verma et al., 1998). For chFab-BCF2, a yield of 1 mg of well-folded, active protein per liter of expression culture was achieved by combination of late log-phase induction, low expression temperature, rich culture medium and addition of sucrose during expression. Temperature reduction during the expression phase has been shown to be beneficial for the yield of periplasmic proteins as it seems to favour proper folding of the recombinant protein by slowing down the synthesis and export of nascent chains, thus diminishing accumulation of unfolded intermediates prone to degradation by proteases (reviewed in Donovan et al., 1996). Indeed, when we expressed chFab-BCF2 at 23 8C we observed the disappearance of side-products that might have been folding intermediates or degradation products and aggregates of these misfolded conformations. The addition of sucrose to the culture medium during expression caused a ca. 3-fold increase of the expression yield for chFab-BCF2 compared with low-temperature expression in non-supplemented 2 £ YT. Sucrose is small enough to enter the periplasmic space of E. coli but is not metabolized. The mechanism of action of sucrose is not known. Its diffusion into the periplasmic space causes an increase in osmotic pressure, a possible consequence of which is the enlargement of the plasmatic space and thus the lowering of protein concentration leading to less aggregation (Kiefhaber et al., 1991). A second consequence also resulting in a decrease of protein concentration, is the perturbance of the outer membrane and thus an increased release of the periplasmic proteins into the culture medium (Kipriyanov et al., 1997); this was also observed by us. Additionally, sucrose stabilizes proteins by an indirect effect of preferential hydration (Lee and Timasheff, 1981). The importance of the latter effects for the expression of chFabBCF2 was indicated by results that we obtained (not shown) using high salt concentration (170 mM NaCl instead of 85 mM as in 2 £ YT). This simple change of osmotic

50

B. Selisko et al. / Toxicon 43 (2004) 43–51

pressure did not increase the yield of chFab-BCF2, which is in contrast to other reports (Blackwell and Horgan, 1991; Chopra et al., 1994). Finally, a new in situ treatment of the cells with lysozyme and addition of sucrose up to 20% turned out to be a simple and efficient procedure to recover both secreted and periplasmic protein. This protocol may be useful for the isolation of other proteins produced in the periplasm and partially secreted into the medium. In competitive ELISA we estimated the relative affinity of chFab-BCF2 in comparison to the parent antibody. The calculated inhibition curves (Fig. 4) indicate clearly that the chimeric Fab retained an affinity for Cn2 in the nanomolar range, very close to that of BCF2. Furthermore, one should take into account that in the case of the complete antibody BCF2, avidity effects could contribute to the observed affinity, which is not the case for chFab-BCF2. The observed neutralizing capacity of recombinant chFab-BCF2 towards the whole venom and toxin Cn2 even at a molar ratio of 1:1 is remarkable. Previous data generated in our group had shown that native Fab fragments derived from mAb BCF2 were able to neutralize the Cn2 toxin and the whole venom of C. noxius (Licea et al., 1996). In this work, following a comparable protocol of in vivo neutralization assays, we have demonstrated that recombinant chFab-BCF2 has the capacity of protecting more than 90% of challenged animals even at a molar ratio of 1:1. Given its in vivo effectivity after intoxication with scorpion venom, chFab-BCF2 could represent the first component of a new generation of a recombinant antidote against scorpion stings. It has to be verified if the Fab will be the antibody fragment of choice. Possible differences in the effectiveness of scFv, Fab or F(ab0 )2 antibody fragments in the neutralization of scorpion venoms have not yet been investigated in detail. The most efficient fragment should have a pharmacokinetic profile close to the toxic venom components. A detailed study of the pharmacokinetics of equine F(ab0 )2 and the toxic proteic fraction of the venom of the scorpion Buthus occitamus mordechai in rabbits showed that F(ab0 )2 fragments distribute relatively slowly and reach only half the distribution volume of the venom components (Pepin-Covatta et al., 1996). On the other hand, they efficiently sequestered the venom components into the vascular compartments creating a concentration gradient from peripheral compartments. This was also shown for Centruroides limpidus limpidus venom and equine F(ab0 )2 (Calderon-Aranda et al., 1999). Pharmacokinetic studies of mouse Fab in rats showed that it distributed more rapidly and the distribution volume was 3-fold higher than of mouse F(ab0 )2 and, as expected, it was cleared more rapidly by renal catabolism (Bazin-Redureau et al., 1997). Recently, the production of scFv fragments of neutralizing antibodies against the main toxins from the scorpion Androctonus australis was reported (Mousli et al., 1999; Devaux et al., 2001). The scFv was shown to maintain the high affinity of the parent antibody and its neutralizing capacity (Devaux

et al., 2001). A disadvantage of scFv constructions might be the presence of multimeric forms (Devaux et al., 2001), which we did not observe for chFab-BCF2. A further disadvantage of scFv is the higher instability in comparison to Fab (Arndt et al., 2001). A Fab fragment may therefore be the fragment of choice for the recombinant scorpion antidote. It is remarkable that chFab-BCF2, a single highly specific recombinant antibody fragment, is capable of neutralizing the whole venom of C. noxius, one of the most poisonous scorpions of Mexico. This suggests that a mixture of a few recombinant antibody fragments could be used in the future, as an improved antidote compared to the commercially available antiserum against envenomations caused by Mexican scorpion species of the genus Centruroides. Further experiments to test this possibility are under development, including determination of kinetic parameters (using surface plasmon-resonance methods) and in vivo neutralization assays injecting the antibody after initial envenomation symptoms for both the chFab-BCF2 and the commercially available antiserum.

Acknowledgements We would like to thank Sonia Rojas for her technical assistance during Fab preparation, the laboratories Alejandro Alagon and Carlos Arias/Susana Lopez from the Institute of Biotechnology (IBT) for their help providing certain equipment, and Rene Hernandez of the Unidad de Sı´ntesis y Secuenciacio´n de Macromole´culas (IBT) for DNA sequencing. Furthermore, we are indebted to Sonia Longhi (AFMB-CNRS, Marseille) for critical reading of the manuscript. The work was supported by grants from Direccio´n General de Asuntos del Personal Academico de la UNAM (IN211996) to B.B., from the Howard Hughes Medical Institute (55000574) to L.D.P., from Consejo Nacional de Ciencia y Tecnologı´a Me´xico (3702-N9607 to E.H. and Z-002 to L.D.P., B.B. and E.H.) and from Laboratorios Silanes to B.B., L.D.P. and E.H.

References Arndt, K.M., Mu¨ller, K.M., Plu¨ckthun, A., 2001. Helix-stabilized fv (hsfv) antibody fragments: substituting the constant domains of a fab fragment for a heterodimeric coiled-domain. J. Mol. Biol. 312, 221–228. Barbas, C.F. III, Kang, A.S., Lerner, R.A., Benkovic, S.J., 1991. Assembly of combinatorial antibody libraries on phage surfaces: the gene III site. Proc. Natl Acad. Sci. USA 88, 7978–7982. Barbas, C.F., Burton, D.R., Scott, J.K., Silverman, G.J. (Eds.), 2001. Phage Display: a Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Bazin-Redureau, M.I., Renard, C.B., Scherrmann, J.M., 1997. Pharmacokinetics of heterologous and homologous immunoglobulin G, F(ab0 )2 and Fab after intravenous administration in the rat. J. Pharm. Pharmacol. 49, 277 –281.

B. Selisko et al. / Toxicon 43 (2004) 43–51 Blackwell, J.R., Horgan, R., 1991. A novel strategy for production of a highly expressed recombinant protein in an active form. Fed. Eur. Biochem. Sci. Lett. 295, 10– 12. Calderon-Aranda, E.S., Dehesa-Davila, M., Chavez-Haro, A., Possani, L.D. (Eds.), 1996. Scorpion stings and their treatment in Mexico, Envenomings and their treatments. Fondation Marcel Merieux, Lyon, France, pp. 311–320. Calderon-Aranda, E.S., Riviere, G., Choumet, V., Possani, L.D., Bon, C., 1999. Pharmacokinetics of the toxic fraction of Centruroides limpidus limpidus venom in experimentally envenomed rabbits and effects of immunotherapy with specific F(ab0 )2. Toxicon 37, 771– 782. Chopra, A.K., Brasier, A.R., Das, M., Xu, X.J., Peterson, J.W., 1994. Improved synthesis of Salmonella typhimurium enterotoxin using gene fusion expression systems. Gene 144, 81–85. Dehesa-Da´vila, M., Ramı´rez, A., Zamudio, F.Z., Gurrola-Briones, G., Lie´vano, A., Darszon, A., Possani, L.D., 1996. Structural and functional comparison of toxins from the venom of the scorpion Centruroides infamatus infamatus, Centruroides limpidus limpidus and Centruroides noxius. Comp. Biochem. Physiol. 113B, 331–339. DeLean, A., Munson, P.J., Rodbard, D., 1978. Simultaneous analysis of families of sigmoidal curves: application to bioassay, radioligand assay, and physiological dose-response curves. Am. J. Physiol. 235, E97 –102. Devaux, C., Moreau, E., Goyffon, M., Rochat, H., Billiald, P., 2001. Construction and functional evaluation of a single-chain antibody fragment that neutralizes toxin AahI from the venom of the scorpion Androctonus australis hector. Eur. J. Biochem. 268, 694 –702. Donovan, R.S., Robinson, C.W., Glick, B.R., 1996. Review: optimizing inducer and culture conditions for expression of foreign proteins under the control of the lac promoter. J. Ind. Microbiol. 16, 145–154. Forsberg, G., Forsgren, M., Jaki, M., Norin, M., Sterky, C., Enhorning, A., Larsson, K., Ericsson, M., Bjork, P., 1997. Identification of framework residues in a secreted recombinant antibody fragment that control production level and localization in Escherichia coli. J. Biol. Chem. 272, 12430– 12436. Garcı´a, C., Caldero´n-Aranda, E.S., Anguiano, G.A.V., Becerril, B., Possani, L.D., 2003. Analysis of the immune response induced by a scorpion venom sub-fraction, a pure peptide and a recombinant peptide, against toxin Cn2 of Centruroides noxius Hoffmann. Toxicon 41, 417–427. Geoffroy, F., Sodoyer, R., Aujame, L., 1994. A new phage display system to construct multicombinatorial libraries of very large antibody repertoires. Gene 151, 109– 113. Horn, U., Strittmatter, W., Krebber, A., Knu¨pfer, U., Kujau, M., Wenderoth, R., Mu¨ller, K., Matzku, S., Plu¨ckthun, A., Riesenberg, D., 1996. High volumetric yields of functional dimeric miniantibodies in Escherichia coli, using an optimized expression vector and high-cell-density fermentation under nonlimited growth conditions. Appl. Microbiol. Biotechnol. 46, 524–532. Ismail, M., 1995. The scorpion envenoming syndrome. Toxicon 31, 1071–1076.

51

Kiefhaber, T., Rudolf, R., Ko¨hler, H.H., Bu¨chner, J., 1991. Protein aggregation in vitro and in vivo: a quantitative model of the kinetic competition between folding and aggregation. Bio/ Techol. 9, 825 –829. Kipriyanov, S.M., Moldenhauer, G., Little, M., 1997. High level production of soluble single chain antibodies in small-scale Escherichia coli cultures. J. Immunol. Methods 200, 69–77. Lee, J.C., Timasheff, S.N., 1981. The stabilization of proteins by sucrose. J. Biol. Chem. 256, 7193–7201. Licea, A.F., Becerril, B., Possani, L.D., 1996. Fab fragments of the monoclonal antibody BCF2 are capable of neutralizing the whole soluble venom from the scorpion Centruroides noxius Hoffmann. Toxicon 34, 843 –847. Mousli, M., Devaux, C., Rochat, H., Goyffon, M., Billiald, P., 1999. A recombinant single-chain antibody fragment that neutralizes toxin II from the venom of the scorpion Androctonus australis hector. Fed. Eur. Biochem. Sci. Lett. 442, 168–183. Pepin-Covatta, S., Lutsch, C., Grandgeorge, M., Lang, J., Scherrmann, J.M., 1996. Immunoreactivity and pharmacokinetics of horse anti-scorpion venom F(ab0 )2—scorpion venom interactions. Toxicol. Appl. Pharmacol. 141, 272–277. Pintar, A., Possani, L., Delepierre, M., 1999. Solution structure of toxin 2 from Centruroides noxius Hoffmann, a beta scorpion neurotoxin acting on sodium channels. J. Mol. Biol. 287, 359–367. Possani, L.D., 1984. Structure of scorpion toxins. In: Tu, A.T.T., (Ed.), Handbook of Natural Toxins, Marcel Dekker, New York, pp. 513–550. Possani, L.D., Becerril, B., Delepierre, M., Tytgat, J., 1999. Scorpion toxins specific for Naþ-channels. Eur. J. Biochem. 264, 287–300. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: a Laboratory Manual, second ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Selisko, B., Licea, A.F., Becerril, B., Zamudio, F., Possani, L.D., Horjales, E., 1999. Antibody BCF2 against scorpion toxin Cn2 from Centruroides noxius Hoffmann: primary structure and three-dimensional model as free Fv fragment and complexed with its antigen. Proteins 37, 130–143. Skerra, A., 1993. Bacterial expression of immunoglobulin fragments. Curr. Opt. Immunol. 5, 256– 262. Sodoyer, R., Aujame, L., Geoffroy, F., Pion, C., Peubez, I., Monte`gue, B., Jacquemont, P., J., 1996. Multicombinatorial libraries and combinatorial infection: practical considerations for vector design. In: Kay, B., Winter, J., McCafferty, J. (Eds.), Phage Display of Peptides and Proteins. A Laboratory Manual, Academic Press, San Diego, pp. 207–217. Verma, R., Boleti, E., George, A.J., 1998. Antibody engineering: comparison of bacterial, yeast, insect and mammalian expression systems. J. Immunol. Methods 216, 165– 181. Wall, J.G., Plu¨ckthun, A., 1995. Effects of overexpressing folding modulators on the in vivo folding of heterologous proteins in Escherichia coli. Curr. Opin. Biotechnol. 6, 507– 516. Zamudio, F., Saavedra, R., Martin, B.M., Gurrola-Briones, G., Herion, P., Possani, L.D., 1992. Amino acid sequence and immunological characterization with monoclonal antibodies of two toxins from the venom of the scorpion Centruroides noxius. Eur. J. Biochem. 204, 281–292.