key: cord-0988547-38t82q9t authors: Lamarre, Alain; Yu, Mathilde W. N.; Chagnon, Fanny; Talbot, Pierre J. title: A recombinant single chain antibody neutralizes coronavirus infectivity but only slightly delays lethal infection of mice date: 2005-12-06 journal: Eur J Immunol DOI: 10.1002/eji.1830271245 sha: 9b4d9173559b08225b6e1e642e54fb2ae50fce17 doc_id: 988547 cord_uid: 38t82q9t The variable region genes of a murine anti‐coronavirus monoclonal antibody (mAb) were joined by assembly polymerase chain reaction and expressed in Escherichia coli in a single chain variable fragment (scFv) configuration. After induction of expression, the expected 32‐kDa protein was identified by Western immunoblotting with specific rabbit anti‐idiotype antibodies. The scFv fragments were purified from soluble cytoplasmic preparations by affinity chromatography on nickel agarose, which was possible with an N‐terminal but not with a C‐terminal histidine tag. Purified scFv fragments retained the antigen‐binding properties of the parental antibody, could inhibit its binding to viral antigens with apparently higher efficiency than monovalent antigen‐binding (Fab) fragments, but neutralized viral infectivity with lower efficiency (about sevenfold at a molar level). To evaluate the usefulness of these smaller and less immunogenic molecules in the treatment of viral diseases, mice were treated with purified recombinant scFv fragments and challenged with a lethal viral dose. A small delay in mortality was observed for the scFv‐treated animals. Therefore, even though the scFv could neutralize viral infectivity in vitro, the same quantity of fragments that partially protected mice in the form of Fab only slightly delayed virus‐induced lethality when injected as scFv fragments, probably because of a much faster in vivo clearance: the biologic half‐life was estimated to be about 6 min. Since a scFv derived from a highly neutralizing and protective mAb is only marginally effective in the passive protection of mice from lethal viral infection, the use of such reagents for viral immunotherapy will require strategies to overcome stability limitations. Coronaviruses are members of the Coronaviridae virus family that includes important pathogens of the respiratory, gastrointestinal and neurological systems of humans and various animals [ l , 21. Neurotropic strains of the murine coronavirus MHV can induce neurological disorders in rodents that are similar to multiple sclerosis [3] , providing an excellent animal model for the study of human nervous system diseases and immune protection mechanisms. Passive protection from MHV infection has been achieved by administration of mAb specific for all Received Feb. 24,1997; revised Aug. 18, 1997; four major structural proteins of the virus [4-81. We have recently shown that F(ab'), and Fab fragments of mAb 7-1 OA specific for the viral surface glycoprotein can also neutralize the virus in vitro and protect mice in vivo [9] . The utilization of mAb in the treatment of important viral diseases is an attractive approach because of their wide specificities and potent biological effects. However, their clinical use has been hampered by their immunogenicity in humans [lo] . The development of molecular biology techniques which make it possible to express antibody fragments in bacteria and eukaryotic cells offers the possibility of developing immunological reagents with very high specificity and sensitivity, with even less immunogenicity than antibody fragments obtained by enzymatic digestion [l l-131. To explore the possibility of using antibody fragments expressed in bacteria for the treatment of viral diseases, an scFv was constructed from the sequences of MHV-specific mAb 7-1 OA and its in vitro neutralization and in vivo protection properties were evaluated. The variable regions of the heavy and light chains of mAb 7-10A were amplified by PCR with VH-and V,-specific primers using cDNA synthesized from RNA extracted from 7-10A hybridoma cells (Fig. 1) . Assembly of the variable regions of heavy and light chains of mAb 7-10A was done by splicing with overlap extension [24] . A linker molecule (Gly, Ser), was used to bridge the two chains together in an scFv configuration [25] . The assembly product of the correct size (750 bp) was gel-purified and cloned into the bacterial expression vector PET-22b. The nucleotide and deduced amino acid sequences of scFv 7-1 OA were determined (Fig. 2) Expression of the recombinant scFv was induced with 1 mM IPTG for 18 h at 30 "C and total cellular proteins were analyzed by SDS-PAGE (Fig. 3A) . Coomassie blue staining of induced cells revealed a major protein band of 32 kDa corresponding to the predicted size of the scFv and that was undetectable in uninduced cells. Fractionation of soluble and insoluble material revealed that the majority of the recombinant protein was contained in insoluble inclusion bodies (data not shown). The identity of the scFv was verified by Western immunoblotting with polyclonal anti-idiotype antibodies produced against the parental 7-1 OA antibody [22] . These antibodies reacted with a 32-kDa protein present only in induced cells (Fig. 3B ). Attempts to affinity purify the recombinant scFv expressed in the PET-22b vector by Ni-NTA agarose column chromatography under either non-denaturing or denaturing conditions failed. It is possible that the Cterminal histidine tag was so embedded in the protein core even under denaturing conditions that it was inaccessible to the Ni'+ cations. We subcloned the scFv into the PET-1 6b vector which expresses the histidine tag at the N-terminal end of the protein. Although most of the recombinant scFv was also produced in insoluble inclusion bodies, enough soluble protein was present in cytoplasmic extracts to be purified on the Ni-NTA agarose column, with a yield of about 0.2 mg/l of bacterial culture. Adsorbed proteins were eluted with 60 rnM imidazole and the fractions were analyzed by SDS-PAGE and Western immunoblotting. Coomassie blue staining of the eluted fractions revealed a unique band of 32 kDa ( Fig. 3C) , which was also revealed in Western immunoblotting with the anti-7-1 OA anti-idiotype antibodies (Fig. 30 ). Microtiter plates were coated with 500 ng/well of viral antigen preparations ( . ) or uninfected cell lysates (0). The binding of threefold dilutions of purified scFv fragments was detected using 7-1 OA-specific anti-idiotype antibodies and horseradish peroxidase-labeled goat anti-rabbit IgG antibodies. To verify whether the purified scFv fragments had retained the antigenic specificity of the bivalent parental immunoglobulin, their binding to viral antigen prepara- ) or control antibody (0) were added to the plates and the binding of a fixed concentration of the parental antibody was detected using Fc-specific horseradish peroxidase-labeled antimouse antibodies. tions was tested by ELlSA using the specific anti-7-lOA anti-idiotype antibodies for detection (Fig. 4) . The scFv could indeed bind in a concentration-dependent manner to viral proteins present in infected cell lysates whereas no specific interaction with preparations from uninfected cells was observed. In order to determine the relative affinity of the scFv fragment for antigen, its ability to inhibit the binding of the bivalent natural antibody was examined (Fig. 5) . Fifty percent inhibition of 7-10A binding was achieved with 0.6 mg of purified scFv fragments. In contrast, we have previously shown that the same amount of purified Fab fragments inhibited less then 20 % of the intact antibody binding [9]. The neutralization capacity of the recombinant scFv fragment was evaluated and compared to that of the Fab fragment by incubating 50 PFU of virus with dilutions of purified fragments and determining the residual viral infectivity on murine fibroblast cells (Fig. 6) . The neutralizing titer of scFv fragments (350 x mole) was about sevenfold lower than that of Fab fragments (50 x mole). BALB/c mice were treated with 500 pg of purified scFv or Fab fragments and were challenged 30 min later with 10 LDS0 of infectious MHV. No mice survived the viral infection but a small delay in the mortality of the animals treated with the scFv was observed, which contrasted with the protection of about 70 % of animals treated with Fab fragments (Fig. 7) . Given the relatively efficient in vitro virus-binding properties of scFv compared to Fab fragments, we evaluated whether the limited in vivo protective capacity of the scFv fragments was due to faster clearance. Indeed, we estimated the half-life of scFv 7-10A to be only about 6 min (Fig. 8) . Murine antibodies that neutralize virus infectivity and have the capacity of protecting against viral infection are attractive candidates as potential immunotherapeutic agents. However, their large scale use has been hampered by allergic immune reactions in humans (291 and the difficulty and costs of producing large quantities of antibodies. Recombinant antibody fragments present several advantages over conventional monoclonal antibodies: they are less immunogenic in humans and can be produced in large amounts and at lower costs. These advantages have encouraged the development of a number of genetically engineered virus-specific antibody fragments with neutralizing properties [30-331. As a model for the utilization of recombinant antibodies in the treatment of viral diseases, we tested whether the same antibody engineering technology could be employed for the production of a scFv that could protect from virus infection in a convenient animal model. We report the construction and expression of a scFv rescued from a hybridoma line that secretes anticoronavirus IgG2a mAb which can neutralize virus infection in vitro and protect mice against a normally lethal dose of virus. We show that the location of the histidine tag, either at the Cor N-terminal end of the recombinant protein, may have a major importance for purification by affinity chromatography on a nickel agarose column. Indeed, we have observed that the scFv fragment produced in this study could only be purified when the histidine tag was expressed at the N-terminal end. In contrast, Lake et al. [34] have reported the purification on a nickel agarose column of an anti-insulin scFv with a Cterminal histidine tag. This demonstrates that the con-formation of the particular scFv will determine whether the expression of a Cor N-terminal histidine tag will be accessible to the Ni2+ cations and will allow purification by metal chromatography. The scFv described in the present report showed biological properties similar to Fab fragments obtained by papain digestion. In fact, they exhibited much better inhibition of parental antibody binding to viral antigen than Fab fragments, which is consistent with a higher affinity. Indeed, 50 % inhibition of binding of the parental antibody to viral antigen was achieved with only 0.6 pg of scFv whereas 10 pg of Fab fragments only inhibited 42 % of binding [9]. However, this did not correlate with a better neutralization activity of the SCFV, with molar titers about sevenfold lower than these of Fab fragments. Even with an apparent higher affinity than Fab fragments, the scFv was less effective in the passive protection of animals against lethal viral infection. This was most likely due to a shorter half-life, which we measured to be about 55-fold shorter (6 min The very fast blood clearance of scFv fragments represents an advantage for some clinical uses such as tumor immunotargeting for diagnosis or treatment of cancer but represents a major limitation for their utilization in viral immunotherapy. However, some reports have suggested that the in vivo stability of these small antibody fragments can be significantly prolonged, for example by disulfide stabilization [38, 401 or the identification and introduction of stabilizing mutations [41] . Importantly, the results presented in the current study with a murine coronavirus have very recently been confirmed in another animal model, vesicular stomatitis virus [39]. These authors also concluded that a short half-life of the antibody fragments hampered passive protection of mice against lethal infection and showed that protection required pre-incubation of the challenge virus with antibody fragments. This confirms that monovalent antibody fragments may be able to passively protect against viral infections and emphasize the need to engineer more stable molecules before clinical uses can be envisaged. Male or female, 6-to 7-week-old, MHV-seronegative BALB/c mice (Charles River, St-Constant, Canada) were used in the protection experiments. The neurotropic A59 strain of MHV (MHV-A59) was obtained from the American Type Culture Collection (Rockville, MD), plaque-purified twice, and passaged on DBT cells as described previously [14] . Total cellular RNA was isolated from 7-10A hybridoma cells as described previously [15] . of JKl 5'-CCGllTGAllTCCAGClTGGTGCC-3', JK2: 5'-CCGllTAllTCCAGClTGGTCCC-3', JK4: 5'-CCGTT-TATITCCAACTTGTCCC-3' and JK5: 5'-CCGTITCAGCT-CCAGClTGGTCCC-3') and VK2BACK (5'-GACAlT-GAGCTCACCCAGTCTCCA-3') [18] to amplify VK. The linker DNA was similarly amplified from the plasmid pSW2scD1.3 [19] using primers MO-LINK-BACK and MO-LINK-FOR (complementary to VH1 FOR-2 and VK2BACK respectively). Gel purified VH and VK amplicons (100 ng each) were mixed with 20 ng of the linker DNA fragment encoding the peptide (Gly4Ser), in a 50-pl reaction mixture without primers and cycled 7 times (94 "C for 1 min, 72 "C for 2.5 min) with Vent DNA polymerase (New England Biolabs, Ltd., Mississauga, Canada) to randomly join the fragments, then amplified for 23 cycles (94 "C for 1 rnin, 60 "C for 2 rnin and 72 "C for 2 min) using 25 pmol each of VHlBACK and VK4FOR primers to which Ncol and Not1 restriction sites where appended, respectively. The amplification product was digested with Ncol and Not1 for cloning into the PET-22b vector (Novagen, Inc., Madison, WI) containing a C-terminal histidine tag. The scFv product was also cloned into the PET-16b vector (Novagen) for the expression of an Nterminal histidine tag. The scFv 7-1 OA insert contained in the PET-22b plasmid was amplified using primers VH1 BNDE (5'-GGMlTCCATATGGCCGAGGTCAAGCTGC-3') and and the PCR product was digested with Ndel and Sall and ligated into the Ndel-and Sall-digested PET-16b vector. The ligation products were used to transform the XL1 -blue strain of E. coli (Stratagene, La Jolla, CA). Positive clones were subcloned into E. coli strain BL21 (DE3) (Novagen) for expression. Transformed BL21 (DE3) cells were grown at 37 "C in LB containing 100 pg/ml ampicillin (Boehringer Mannheim Canada, Laval, Canada) until the OD at 600 nm reached 0.6, at which time 1 mM isopropyl-p-Dthiogalactopyranoside (IPTG) was added (Clonetech Laboratories Inc., Palo Alto, CA). Following induction, the cultures were grown for an additional 18 h at 30 "C. Induced cells were centrifuged for 20 min at 10000 x g and resuspended in 1/25 volume of 500 mM NaCl and 20 mM Tris-HCI, pH 7.9. Lysozyme (Boehringer Mannheim) was added to a concentration of 100 pg/ml and the suspension was incubated at 30 "C for 30 min. To shear the DNA, the suspension was sonicated on ice for 1 rnin or until the solution lost viscosity using a Braun-Sonic 2000 sonicator. The lysate was centrifuged at 29 000 x g for 30 min and the supernatant filtered through a 0.45-prn Sterifil-D membrane (Millipore Canada, Nepean, Canada) for column chromatography. The PCR assembly product of scFv 7-1 OA was cloned into the pCRll TA cloning vector (Invitrogen Corporation, San Diego, CA). Nucleotide sequencing was performed on both strands of two PCR products by the dideoxynucleotide chain terminating method [20] using T7 DNA polymerase Ltd., Montreal, Canada) according to the manufacturer's instructions (Pharmacia). A soluble cytoplasmic extract was used for the purification on a Ni-NTA agarose column (Qiagen Inc., Chatsworth, CA). The column was washed with ten volumes of binding buffer (500 mM NaCl and 20 mM Tris-HCI, pH 7.9) and loaded with the prepared cell extract at a flow rate of about 10 column volumes per hour. After loading, the column was washed with binding buffer until the OD at 280 nm reached the base line level, after which time the bound proteins were eluted with binding buffer containing 60 mM imidazole (Sigma-Aldrich, Canada, Ltd., Mississauga, Canada). Elution fractions were analyzed by SDS-PAGE and Western immunoblotting. Fractions containing the purified scFv were pooled and dialyzed against PBS for the biological assays. Induced cell extract preparations and elution fractions were separated by SDS-PAGE [21] and stained with Coomassie blue for direct visualization or electrotransferred onto Hybond-C Extra nitrocellulose membranes (Amersham Searle Corp., Oakville, Canada) for 1 h at 100 V. Membranes were blocked for 1 h with PBS-T and incubated for 90 min with a 2 pg/ml solution of purified anti-7-lOA rabbit antiidiotype antibodies [22] in PBS-T. Membranes were washed five times with PBS-T and incubated for 60 min with horseradish peroxidase-conjugated goat anti-rabbit IgG antibodies (diluted 1/1,000; Kirkegaard & Perry Laboratory, Gaithersburg, MD). Membranes were again washed five times with PBS-T and incubated with hydrogen peroxide (Sigma) and 4-chloro-1 -naphtol (Bio-Rad Laboratories Ltd.). All incubations were performed at room temperature (about 25°C). Each well of 96-well microtiter plates was coated with 500 ng of viral antigen prepared from MHV-A59infected cells, as described previously [23] . After overnight incubation, the remaining binding sites in the wells were blocked with PBS containing 10 % (v/v) FCS and 0.2 % (v/v) Tween-20 for 90 min. Serial threefold dilutions of purified scFv fragments were added to the wells and incubated for 90 min. The wells were washed five times with PBS-T and a 2 pg/ml solution of purified rabbit anti-7-1 OA anti-idiotype antibodies in blocking solution was added and incubated for 90 min. The wells were washed as described above and peroxidase-labeled goat anti-rabbit IgG antibodies (Kirkegaard & Perry Laboratories, Inc) were then added and the plates incubated for another 90 min. The plates were washed five times with PBS-T and the bound peroxidase revealed by incubation with o-phenylenediamine (Sigma) and hydrogen peroxide. The reaction was stopped with 1 N HCI and the absorbance read at 492 nm using an SLT EAR 400 AT plate reader. The wells of 96-well microtiter plates were coated with viral antigen as described above and incubated overnight at room temperature. Serial threefold dilutions of purified scFv fragments were added to the wells and incubated for 90 min, after which the wells were emptied and purified mAb 7-1 OA added without any previous washing. After incubation for 60 min, Fc-specific peroxidase-labeled goat antimouse IgG (ICN) was added and the plates incubated for 90 min. All subsequent steps were performed as described above. The amount of purified mAb 7-1 OA added in this test was determined in a binding ELISA in which an absorbance value at 492 nm of about 1 .O was achieved in the absence of inhibition by antibody fragments. Duplicate serial dilutions of scFv or Fab fragments were incubated with approximately fifty PFU of MHV-A59 for 1 h at 37 "C. The mixtures were transferred onto 12-well plates containing confluent monolayers of DBT cells. After an adsorption period of 1 h at 37"C, the virus-scFv mixtures were removed and cells were overlaid with 1.5 % (w/v) agarose in Earle's minimum essential mediumMank's M199 (1 : 1, v/v) (Gibco Canada, Burlington, Canada) supplemented with 5 % (v/v) FCS. Plates were incubated for 48 h at 37 "C in a humidified atmosphere containing 5 % (v/v) GO2, after which the cells were fixed with formaldehyde and stained with crystal violet. Viral neutralization titers are expressed as the amount of antibody fragments that could neutralize 50 % of input viral infectivity. MHV-seronegative 6-week-old BALB/c mice (Charles River) were injected intraperitoneally with 500 pg of antibody fragments 30 rnin prior to challenge with 5 x 1 O5 PFU (1 0 LD5, , ) of MHV-A59 injected intracerebrally. The purification of scFv 7-10A was as described above except that it was performed in batch, using 50-ml tubes. Two milliliters of Ni-NTA agarose resin were washed three times with 30 ml of binding buffer and mixed with 15 ml of bacterial cell extract for 60 min at 4 "C on a rocker platform. The resin was washed twice with binding buffer and three times with wash buffer (binding buffer containing 30 mM imidazole). The scFv 7-10A were then eluted from the resin with 5 ml of elution buffer (binding buffer containing 1 M imidazole). Purified scFv fragments were analyzed by SDS-PAGE and Western immunoblotting as described above, concentrated with Aquacide II (Calbiochem-Novabiochern Corporation, La Jolla, CA) and dialyzed against PBS. Radioiodination of the purified scFv 7-10A was performed with the lodo-Beads8 radioiodination reagent (Pierce, Rockford, IL). Three beads were washed in 0.1 M phosphate buffer, pH 7.2, dried on filter paper and resuspended in 0.1 ml of the same buffer, to which 5 mCi Na'251 (ICN) were added. After 5 min incubation, 1 mg of purified scFv 7-10A was added and iodination was allowed to proceed for 14 rnin at room temperature. The reaction was stopped by removing the beads. Radioiodinated scFv 7-10A were separated from free iodine by Sephadex G-25 chromatography (Pharmacia) and their purity reverified by SDS-PAGE followed by autoradiography of the dried gel using a Kodak X-Omat AR X-ray film. To measure biological half-life in vivo, two BALB/c mice (Charles River) were injected intravenously with 15 pg of radioiodinated scFv 7-10A. After 2, 4, 8, 15, 30, 60, 90 and 120 min, 60 yl of blood was collected from the retroorbital plexus into heparinized capillary tubes. Plasma samples were analyzed by SDS-PAGE and autoradiography and the 32-kDa scFv bands were quantitated by videodensitometry using an AlphalmagerTM 2000 Documentation and Analysis System with AlphalmagerTM 3.241 software (Applied Innotech, San Leandro, CA). The validity of this densitometric procedure was ascertained using classical laser densitometry (Bio-Rad), with identical results. Immunochemistry of viruses. The basis for serodiagnosis and vaccines Proc. Natl. Acad. Sci Proc. Natl. Acad. Sci Proc. Natl. Acad. Sci Sequences of proteins of immunological interest Proc. Natl. Acad. Sci Proc. Natl. Acad. Sci Proc. Natl. Acad. Sci Centre de recherche en virologie, lnstitut Armand-Frappier Acknowledgment: The authors are grateful to Dr. Greg Winter (Cambridge, U.K.) for providing primer sequences and the pSW2scDl.3 plasmid and to Dr. Jean-FranCois Laliberte and Dr. Christopher D. Richardson for helpful discussions. This work was supported by grant MT-9203 from the Medical Research Council of Canada (MRCC) which also provided studentship support to A. Lamarre. P. Talbot and F. Chagnon acknowledge senior scholarship and studentship support, respectively, from the Fonds de la recherche en sante du Quebec. M. Yu received a studentship from the Fonds pour la formation et I'aide a la recherche du Quebec.