RSV G protein specific antibodies

Abstract

The disclosure relates to human isolated, synthetic or recombinant antibodies or functional parts thereof, specific for the RSV G protein. Antibodies specific for the RSV G protein are particularly suitable for counteracting RSV and symptoms, such as inflammation, resulting from an RSV infection. The disclosure further relates to the use of such RSV G-specific antibodies for diagnosis of an RSV infection and as a medicament and/or prophylactic agent for, at least in part, treating or alleviating symptoms of a Respiratory Syncytial Virus infection and/or a Respiratory Syncytial Virus related disorder.

Claims

1. A composition comprising: a combination of AT44, AT42 and AT33; wherein said AT44, said AT42 and said AT33 are each isolated, synthetic or recombinant antibodies, or antigen binding fragments thereof; said AT44 comprises SEQ ID NO: 2, SEQ ID NO: 20, SEQ ID NO: 38, SEQ ID NO: 56, SEQ ID NO: 74 and SEQ ID NO: 92; said AT42 comprises SEQ ID NO:10, SEQ ID NO: 28, SEQ ID NO: 46, SEQ ID NO: 64, SEQ ID NO: 82 and SEQ ID NO: 100; and said AT33 comprises SEQ ID NO: 3, SEQ ID NO: 21, SEQ ID NO: 39, SEQ ID NO: 57, SEQ ID NO: 75 and SEQ ID NO: 93.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1. Amino acid sequence of RSV G protein of two subtypes A (Ga) A2 and Long strain and of three subtypes B (Gb) viruses from the USA (the B1 strain), Turkey and Uruguay (SEQ ID NOS 289-293, respectively). * indicate conserved/identical amino acid residues.

(2) FIG. 2. Screening B cell supernatants for specificity to the RSV F and G protein. RSV A2 infected, PKH2 Green Fluorescent labeled HEp2 cells were mixed with PFA fixed RSV G expressing VERO cells and incubated with 20 cell/well B cell culture supernatant. IgG antibodies that bound the HEp2 or VERO cells were detected with a mouse anti-human IgG-PE. Antibodies present in the B cell culture supernatant that bound the RSV infected HEp2 cells (population 1) but not the G expressing VERO cells (population 2) presumably recognize the RSV F protein. Antibodies recognizing both cell lines most likely recognize the RSV G protein.

(3) FIG. 3. Neutralization of RSV A2 by anti-RSV G protein specific human monoclonal antibodies in the absence and presence of complement. (A) Before RSV A2 was administered to HEp2 cells in 96-well plates, the virus was co-incubated with purified, recombinant AT32, AT33, AT40, AT42, AT44, AT46 and AT47 at 27 μg/ml. When combinations of 3 mAbs were tested the final concentration was 27 μg/ml, thus 9 μg/ml of each antibody. (B) Monoclonal B cell supernatant with anti-RSV G specific antibodies were first incubated with RSV virus for 1 hour at 37° C. before 10% rabbit serum complement was added for another hour in the presence of HEp2 cells. Cells were washed and cultured for 2 more days in normal culture medium.

(4) FIG. 4. Enhanced neutralization of RSV A2 virus by combinations of anti-RSV G and F protein specific antibodies. (A) Increasing amounts of RSV F specific antibodies and a fixed amount (500 ng/ml) of anti-RSV G antibody were co-incubated with 25 PFU of RSV A2 virus for 1 hour at 37° C. Subsequently, the virus antibody mixture was added to 20,000 HEp2 cells in solution in 96-well flat-bottom culture plates. After two days, the number of infected foci was determined. Antibody virus combinations were tested at least three times; FIG. 4B shows the average increase compared to the F antibody alone.

(5) FIG. 5. 3D3 antibody-binding competition for the conserved domain on the RSV G protein. To analyze if the anti-RSV G antibodies bind within the conserved region of the RSV G protein we performed antibody competition assays. The antibodies were compared to 3D3, which binds the epitope HFEVFNFVP (aa 164-172, FIG. 1, US patent application US 2010-0285022 and Collarini et al. J Immunol (2009) 183: 6338-6345). 3D3 was directly labeled with ALEXA Fluor 647 (Molecular Probes) and antibody competition was determined by incubation of RSV-infected HEp2 cells with an increasing dose of the non-labeled antibody before the labeled antibody was added at a standard concentration. In addition, the assay was also performed by simultaneously incubation of the labeled and non-labeled antibodies, in general no differences between the two methods was detected. Shown in FIG. 5 is the average binding of ALEXA Fluor 647 labeled 3D3 antibody relative to the control of three separate experiments.

(6) FIG. 6. Western Blots showing binding of RSV G-specific antibodies to RSV A2 supernatant (denatured), detected with anti-human IgG IR-dye.

(7) FIG. 7. (A) Example of recombinant full-length RSV Ga protein binding to human anti-RSV G antibodies captured on a SPR anti-human IgG chip. (B) Graph summarizes the binding of RSV G antibodies to biotinilated 12-mer peptides that were coupled to streptavidin coated on an IBIS SPR chip. The peptide library spans the amino acid sequence 149 to 199 of the RSV A2 strain. (C) Amino acid epitope recognize by the antibodies 3D3, AT40, AT44, 131-2G and AT32.

DETAILED DESCRIPTION

Examples

Example 1: Generation of Human Monoclonal Antibodies Against the RSV G Protein by Transduction of Human Peripheral Blood Memory, IgG+ B Cells by BCL6 and Bcl-xL

(8) Materials and Methods

(9) B Cell Isolation

(10) B cells were obtained from PBMCs from 40 to 50 ml Peripheral blood of three healthy adult volunteers by density gradient separation using Lymphoprep (Axis-Shield PoC, Oslo, Norway) and CD22 MACS microbeads (Miltenyi Biotech, Bergisch Gladbach, Germany). IgG memory B cells were isolated as CD19+CD3−CD27+IgM−IgA− population by FACSAria (Becton Dickinson, San Jose, Calif., USA).

(11) The following mAbs against the human molecules CD3 (SK7), CD19 (SJ25C1), CD27 (O323; eBioscience), IgA (F(ab)2; DAKO Glostrup Denmark), IgD (IA6-2), IgG (G18-145), IgM (G20-127) (BD), Ig-kappa (F(ab)2; DAKO, G20-193), and Ig-lambda (F(ab)2; JDC12, DAKO) were directly labeled with fluorescein isothiocyanate (FITC), phycoerythrin (PE), phycoerythrin cyanine 5, (PE-Cy5), allophycocyanin (APC), phycoerythrin-indotricarbocyanine (PE-Cy7) or allophycocyanin-indotricarbocyanine (APC-Cy7) and were purchased from BD-Pharmingen (San Diego, Calif.) unless otherwise indicated. Stained cells were analyzed on an LSRII or FACSCanto (BD) and flow cytometry data were processed using FlowJo software (Tree Star, Ashland, Oreg., USA).

(12) Retroviral Transduction

(13) Use of the BCL6 and Bcl-xL retroviral construct has been described previously (Kwakkenbos et al. Generation of stable monoclonal antibody-producing B cell receptor-positive human memory B cells by genetic programming. Nature Medicine (2010) vol. 16 (1) pp. 123-8). Briefly, cDNAs encoding human BCL6, Bcl-xL and EGFP were cloned into the LZRS retroviral vector and retrovirus was generated by transfection Phoenix packaging cells (Shvarts et al. A senescence rescue screen identifies BCL6 as an inhibitor of anti-proliferative p19(ARF)-p53 signaling. Genes Dev (2002) 16:681-686). After enrichment (by ficoll density gradient and high speed cell sorting (FACSAria, BD)) and activation of human peripheral memory B cells on CD40L-L cells in the presence of rmIL-21, the cells were transduced. (Diehl et al. STAT3-mediated up-regulation of BLIMP1 is coordinated with BCL6 down-regulation to control human plasma cell differentiation. J Immunol (2008) 180(7):4805-15). Transduced cells express EGFP and can be sorted to enrich for cells that besides EGFP will express BCL6 and Bcl-xL.

(14) B Cell Culture and Screening of Anti-RSV G Protein Specific B Cells

(15) After 4 days from transduction, GFP positive cells were sorted by FACSAria, plated at 20 cells per well in ten 96-well flat-bottom tissue culture-treated plates per donor. After 14 days in culture, B cells and supernatants were harvested. B cells were frozen and supernatants were tested for binding capacity to RSV A2 virus infected HEp-2 cell. In brief, HEp-2 cell culture monolayers were infected with RSV A2 virus at a MOI of 2-3. The infected HEp-2 cells were harvested 48 hours after infection. Cells were stained with PKH2 Green Fluorescent Cell Linker Kit (Sigma-Aldrich, St. Louis, Mo., USA). In addition, B cell supernatants were screened simultaneously on paraformaldehyde (PFA) fixed RSV G protein transduced VERO cells (kindly provided by Myra Widjojoatmodjo, NVI, Bilthoven, The Netherlands) (FIG. 2). A mixture of 2.5E4 PKH2 stained RSV A2 virus infected HEp-2 cell and 2.5E4 RSV G protein expressing VERO cells were incubated for 1 hour at 4° C. with 100 μl of supernatant. Cells were washed once with IMDM supplemented with 1% FBS. IgGs binding to the target cells were detected with PE labeled anti-human IgG (SouthernBiotech, Birmingham, Ala., USA).

(16) Double positive cells were plated at 1 cell per well in 96-well flat-bottom tissue culture-treated plates by FACSAria to obtain single clones. B cells were maintained in standard culture medium containing IMDM (Invitrogen), 8% FBS (HyClone) and penicillin/streptomycin (Roche) and were co-cultured on irradiated (50Gy) mouse L cell fibroblasts stably expressing CD40L (CD40L-L cells, 10E5 cells/ml) and recombinant mouse IL-21 (25 ng/ml, R&D systems, Minneapolis, Minn., USA). After 14 days in culture, supernatants were collected to test binding capacity to A2 virus infected HEp-2 cell by FACS. Table 4 shows an overview and some characteristics of the final 17 B cell clones of which recombinant antibodies were generated. Table 5 shows binding of the antibodies to Hep2 cells infected with RSV A2, RSV X (both subtype A viruses) and RSV 2007-2 (an RSV subtype B virus).

(17) Cloning of Anti-RSV G Monoclonal Antibodies

(18) Total RNA was isolated from approximately 5E5 monoclonal B cells with TRIzol® (Invitrogen). cDNA was generated and subjected to PCR to produce heavy and light chain fragments using 1U AmpliTaq Gold DNA polymerase (Applied Biosystems Inc. Foster City, Calif., USA). PCR products were run on agarose gels, purified and cloned into the pCR2.1 TA cloning vector according to manufacturers' recommendations (Invitrogen). Sequence analysis was performed using BigDye Terminator chemistry (Applied Biosystems Inc.) and Vector-NTI software (Invitrogen). To rule out reverse transcriptase and/or DNA polymerase induced mutations, several independent cDNA conversions and PCR reactions were performed and individually cloned and sequence analyzed.

(19) IgG ELISA

(20) Plates were coated with either anti-human IgG Fc-fragment (Jackson ImmunoResearch Laboratories, Bar Harbor, Me., USA) at 10 μg/ml in PBS for 1 hour at 37° C. or o/n at 4° C. and washed in ELISA wash buffer (PBS, 0.5% TWEEN®-20). 4% Protifar (Nutricia, Zoetermeer, The Netherlands) in PBS was used as blocking agent, before serial dilution of cell culture supernatants and enzyme-conjugated detection Abs were added (dilutions 1:2500 for HRP-conjugated anti-IgG (Jackson ImmunoResearch Laboratories, Inc.). TMB substrate/stop solution (Biosource, Carlsbad, Calif., USA) was used for development of the ELISAs.

Example 2: Functional Testing of 17 Unique, Fully Human Anti-RSV G Protein Specific Antibodies RSV Culture and Neutralization Assay

(21) An RSV A2 virus stock was obtained from supernatant of 3 day infected HEp2 cells maintained in standard culture medium. Supernatants were centrifuged and filtered (0.22 μM filter, Millipore). Subsequently aliquots were snap-frozen, stored in liquid nitrogen and virus titer was determined by standard TCID50 and PFU assay on adherent HEp2 cells. For neutralization assays 10E4 HEp2 cells were seeded in flat-bottom 96-well plates (Costar, Schiphol-Rijk, Netherlands) in standard culture medium. The next day 100TCID50 of RSV A2 and B cell culture supernatant were pre-incubated in the absence or presence of 10% rabbit complement serum (Sigma-Aldrich) before being added in triplicate to HEp2 cells for 1 hour at 37° C. After two days, cells were fixed with 80% acetone and stained with polyclonal anti-RSV-HRP (Biodesign, Kennebunk, Me., USA). 3-Amino-9-ethylcarbazole (AEC) was added for detection and visualization of RSV plaques by light microscopy (plaques were counted). In addition, RSV infected cells could also be stained with polyclonal goat anti-RSV directly labeled with-Alexa Fluor 647 (Molecular Probes). Fluorescent signal was detected with and analyzed by the automated fluorescent microscoop (Operetta, Perkin Elmer). Palivizumab (Medlmmune, Gaithersburg, Md., USA) and D25 (WO 2008/147196) were used as positive control for RSV neutralization.

(22) Results

(23) RSV A2 neutralization experiments with antibodies derived from monoclonal B cell cultures did not result in neutralization in the absence of rabbit serum complement. In general antibody IgG concentrations in B cell supernatant vary between 600 and 2000 ng/ml, which could be too low. When we used increased concentrations of recombinant, purified monoclonal antibodies we did found that AT44 and AT47 could reduce virus infection (FIG. 3a, top panels). AT40, AT33 and AT42 did so only partially. This effect was not seen for the other 12 anti-RSV G antibodies (not shown). More interestingly, we found that combinations of anti-RSV G antibodies were able to neutralize the virus up to 50-60% without the addition of complement (FIG. 3a bottom panels).

(24) Besides the direct neutralization we could identify a large group (9 out of 17) of monoclonal antibodies that neutralized RSV when virus and B cell culture supernatant were co-incubated with 10% rabbit serum complement thereby inducing complement dependent cytotoxicity (CDC) (FIG. 3b). IC50 values were between 10 and 325 ng/ml.

(25) Not all antibodies did broadly recognize RSV-A and RSV-B strains. Depicted in Table 5 is the binding of antibodies to HEp2 cells infected with the RSV A2, RSV-X (subtype A) and a RSV-2007-2 strain of the B subtype (also summarized in Table 4).

Example 3: Synergistic Effect of RSV G Protein Specific Antibodies on the Neutralizing Capacity of Anti-RSV F Antibodies

(26) The role of the G protein on the surface of the RS virus is thought to be associated with target cell attachment. But also for other (unknown) process and mechanisms the G protein could be important, for example, the stabilization of the F protein trimer. It has been shown that the two proteins form a complex (Low et al., The RSV F and G glycoproteins interact to form a complex on the surface of infected cells, Biochemical and Biophysical Research Communications (2008) 366(2):308-13) and it has been shown that an anti-RSV G and RSV-F antibody in vivo can reduce virus titers in mice (Haynes et al., Therapeutic Monoclonal Antibody Treatment Targeting Respiratory Syncytial Virus (RSV) G Protein Mediates Viral Clearance and Reduces the Pathogenesis of RSV Infection in BALB/c Mice, J. Infect. Dis. (2009) 200(3):439-47). Without being bound by theory, antibodies directed against RSV G may influence the interaction between F and G and thereby induce 1) destabilization of the F trimer or 2) expose epitopes on the F trimer that become better accessible for anti-F antibodies; in either situation the F trimer may unfold to its post-fusion state and thereby become non-functional.

(27) To study this we incubated RSV with increasing doses of anti-F antibodies e.g. D25, AM14 and palivizumab and with G specific antibodies (increasing concentrations or fixed at 500 ng/ml). As shown in FIG. 4a we observed that recombinant purified AT46 and AT32 did enhance the neutralizing capacity of AM14 and D25 but less of palivizumab. The synergistic effect was mainly seen at lower concentrations of anti-F antibody, the effect was consistent (the data shown is an average of three or more experiments) and the synergistic effect enhanced neutralization of the F antibodies by a factor 2 (FIG. 4b). Thus the G specific antibodies may induce changes in the presentation and/or stability of the F protein making the F protein more susceptible to neutralization by F specific antibodies.

Example 4: Direct Labeling of Purified Antibodies to Determine Antibody-Binding Competition by FACS

(28) The RSV G protein can bind to the CX3C chemokine receptor 1 (CX3CR1) also named fractalkine receptor or G-protein coupled receptor 13 (GPR13). CX3CR1 is expressed on multiple cell lineages (NK cells, monocytes, Th1 CD4+ T cells and CD8+ T cells, mast cells and B cells. The ligand for CX3CR1, CX3CL1 induces adhesion of leukocytes when the chemokine is expressed as a membrane-anchored protein whereas the soluble form of CX3CL1 induces chemotaxis of leukocytes. The RSV G protein contains a conserved epitope (CWAIC residue 182 to 186, FIG. 1) that mimics the CX3CR1 binding epitope of CX3CL1. Antibodies exist that bind RSV G within the larger conserved domain (aa 169 to 191) and thereby (partially) compete with binding to CX3CR1 (Mekseepralard et al. Protection of mice against Human respiratory syncytial virus by wild-type and aglycosyl mouse-human chimaeric IgG antibodies to subgroup-conserved epitopes on the G glycoprotein. J. Gen. Virol. (2006) 87(Pt 5):1267-73). To analyze if the anti-RSV G antibodies disclosed herein bind similar epitopes we performed antibody competition assays. Antibodies disclosed herein were compared to 3D3 from Trellis Bioscience which binds the epitope HFEVFNFVP (aa 164-172, FIG. 1, U.S. patent application US 2010-0285022, and Collarini et al. Potent high-affinity antibodies for treatment and prophylaxis of respiratory syncytial virus derived from B cells of infected patients. J. Immunol. (2009) 183: 6338-6345). 3D3 was directly labeled with ALEXA Fluor 647 (Molecular Probes) and antibody competition was determined by incubation of RSV-infected HEp2 cells with an increasing dose of the non-labeled antibody before the labeled antibody was added at a standard concentration. In addition, the assay was also performed by simultaneous incubation of the labeled and non-labeled antibodies, in general no differences between the two methods were detected. Shown in FIG. 5 is the average binding of ALEXA Fluor 647 labeled 3D3 antibody relative to the control of three separate experiments. Binding of the 3D3 antibody can be out-competed by itself and antibodies that bind a similar or proximal epitope like the mouse antibody 131-2G (epitope HFEVF). Of the antibodies that bind RSV Ga and Gb (left panel), AT40, AT44 and AT34 strongly reduced 3D3 binding, suggesting that they compete for the same or proximal epitopes. AT42, AT45 and AT49 only partially compete with 3D3, which may suggest that they recognize different epitopes but may sterically hinder 3D3 from efficient binding. The antibody AT46 did not interfere with 3D3 binding. The right panel indicates competition of Ga specific antibodies with 3D3. None of the antibodies (AT32, 33, 36, 37, 39, 43, 50 and 51) did interfere with 3D3 binding, which indicates that they all recognize different RSV Ga specific epitopes.

Example 5: Binding of Anti-RSV G Antibodies in ELISA, SPR and WB

(29) Antibodies, especially human antibodies, which contain relatively long variable domains (CDR regions), often recognize non-linear structures within their putative target. These non-linear structures can be disrupted by standard purification methods, which, for example, include denaturing compounds like TWEEN®. Our B cell technology is utmost suitable to screen for antibodies that recognize these non-linear structures since the method allows for functional screening of antibodies. However, this implies that not all antibodies discovered will recognize its putative target in standard binding assays like western blot (WB), surface plasma resonance (SPR) or ELISA. Besides AT46, AT42, AT43 and AT47 all antibodies gave clear signals in the ELISA (Table 6).

(30) For the RSV ELISA 2 ml of 1% TRITON X®-100 in PBS was added to a cell pellet containing RSV infected Hep2 cells. The lysed cells were mixed thoroughly and kept for 5′ at RT before 10 ml of ice cold PBS was added. The mixture was homogenized using a syringe with needle and cleared though a 0.22 μm filter (Millipore) or centrifuged at 5,000 rpm at 4° C. for 5′. Subsequently, the lysate was dialyzed against 1 L PBS overnight at 4° C. After dialyzation, 0.05% NaN3 was added and samples were stored at 4° C. until use, or stored at −80° C. for long-term storage.

(31) ELISA plates were coated with a lysate of RSV infected HEp-2 cells in PBS for 1 hour at 37° C. or o/n at 4° C. and washed in ELISA wash buffer (PBS, 0.5% TWEEN®-20). Plates were blocked by incubation with 4% milk in PBS, before the anti-RSV antibodies or polyclonal goat anti-RSV (Biodesign) in combination with enzyme-conjugated anti-IgG antibodies were added (dilutions 1:2500 for HRP-conjugated anti-IgG (Jackson). TMB substrate/stop solution (Biosource) was used for development of the ELISAs.

(32) To confirm antibody binding to RSV G, western blots were prepared which were loaded with denatured and boiled supernatants of RSV A2 infected HEp2 cells. These supernatants contain relatively high amounts of the secreted form of RSV G. Summarized in Table 6 and shown in FIG. 6 is a western blot on which antibodies that recognize RSV Ga only (AT32, 33, 35, 36, 37, 39, 50 and 51), bind relative strong to the RSV G protein. From the same group of RSV Ga only binding antibodies, AT47 and AT43 only weakly bind RSV G. From the panel of antibodies that recognize RSV Ga and Gb on infected cells only AT40, 44, 45, 49 and 34 recognize RSV Ga by western blot. AT46 and AT42 do not bind.

(33) In addition, we generated surface plasmon resonance (SPR) data with the IBIS MX96 instrument (Krishnamoorthy et al., Electrokinetic label-free screening chip: a marriage of multiplexing and high throughput analysis using surface plasmon resonance imaging, Lab Chip (2010) 10(8):986-90; van Beers et al., Mapping of citrullinated fibrinogen B-cell epitopes in rheumatoid arthritis by imaging surface plasmon resonance, Arthritis Research & Therapy (2010) 12(6):R219; Krishnamoorthy et al., Electrokinetic lab-on-a-biochip for multi-ligand/multi-analyte biosensing, Anal. Chem. (2010) 82(10):4145-50; de Lau et al., Lgr5 homologues associate with Wnt receptors and mediate R-spondin signaling, Nature (2011) 476(7360):293-7). The rate- and affinity constant of the antibodies were determined using a similar method as described in de Lau et al., Nature (2011).

(34) Briefly, pre-activated SPR sensor chips (IBIS Technologies, Hengelo, Netherlands) were coated with an array of anti-human IgG specific spots (goat anti-human IgG, polyclonal, Fc-specific, Jackson ImmunoResearch Laboratories, Bar Harbor, Me., USA) using a continuous flow microspotter (CFM) (Wasatch Microfluidics, Salt Lake City, Utah, USA). After preparing the sensor chip, the chip was placed in the instrument and treated with RSV-G-specific human IgG. Each anti-IgG spot in the array thus captured a decreasing amount of IgG. After measuring a new baseline for each spot, purified RSV-Ga or Gb protein (Sino Biologics, Beijing, China) was injected to determine label-free surface plasmon resonance (Table 7a, 7b and FIG. 7a).

(35) Kinetic parameters were calculated using Sprint 1.6.8.0 (IBIS Technologies, Hengelo, Netherlands) and Scrubber2 software (BioLogic software, Campbell, Australia). Results between experiments are comparable, the antibodies form stable complexes with the RSV G protein and similar k.sub.a, k.sub.d and K.sub.D were generated. In contrast to the WB and ELISA data we do find binding of the AT42 antibody to RSV Ga and RSV Gb protein in the IBIS SPR (Tables 7a and 7b). AT46 did not bind to either the recombinant or denatured form of the protein, probably because AT46 binds to a conformational epitope, which is not present in the recombinant and denatured form of the G protein because their conformation differs from that of the protein expressed on the surface of the RS virus.

Example 6: Determination of the Epitope of Anti-RSV-G Antibodies

(36) In addition, we performed studies to precisely determine the epitopes recognized by several anti-RSV G antibodies, according to the disclosure. Therefore, we generated 40 peptides containing a 5′ biotin molecule plus a spacer followed by 12-successive amino acids, spanning the amino acid domain 149 to 199 of the RSV A2 G protein. This domain contains the conserved region, which is also recognized by the 131-2G and 3D3 antibody (FHFEVFNFV) and the cysteine rich domain forming the fractakine binding epitope (CWAIC). To detect binding to the peptides we obtained streptavidin-coated sensor chips (IBIS Technologies, Hengelo, Netherlands), on which the biotin labeled peptides were spotted using the CFM. Subsequently, the antibodies were run one by one over the chip at four different concentrations, after each run the chip was regenerated. Since the peptides were still present after regeneration, this indicated that the immobilized strepativin-biotin/peptides complexes were very stable.

(37) FIG. 7b shows the maximum response observed in the SPR instrument when antibodies recognized a certain peptide (1 to 40, as depicted below). The height of the signal is influenced by the affinity of the antibody for the peptide, the concentration of the antibody, the amount of peptide immobilized and the conformation/polarity of the peptide on the sensor chips (polarity of the FHFEVFNF is low). Together we can conclude that the epitope recognized 3D3, 131-2G, AT40 and AT44 are in close proximity of each other (FIGS. 7b and 7c). Antibodies AT42, AT46 and AT49 did not recognize any of the captured peptides on the chip (not shown), indicating that the epitope of these antibodies is at least partly located outside the amino acid domain 149 to 199 of the G protein or that these antibodies recognize a conformation not present when the peptides are captured on the chip.

(38) The domain described for the 131-2G antibody; HFEVF (Tripp et al. CX3C chemokine mimicry by respiratory syncytial virus G glycoprotein, Nat. Immunol., 2001) could be confirmed by us. Regarding 3D3 we find that the antibody binds to the residues FHFEVFNF as core residues and FHFEVFNFV as the complete epitope. The epitope published for 3D3 is HFEVFNFVP (Collarini et al., Potent high-affinity antibodies for treatment and prophylaxis of respiratory syncytial virus derived from B cells of infected patients, J. Immunol., 2009), we, however, find one more residue at the beginning, (F163) which is necessary.

(39) AT40 and AT44 both start at residue 165F. AT40's epitope then continues till residue F170, making the epitope consist of FEVFNF. AT44 needs at least residue E166 till F170, making the complete epitope ranging from EVFNF. To our current knowledge these antibody epitopes have never been described before. Antibody AT32, which only binds to RSV subtype A viruses did bind to the more distal epitope RIPNK (position 188 to 192), an epitope located just after the fractalkine binding site.

(40) TABLE-US-00009  1 KQRQNKPPSKPN  2 QRQNKPPSKPNN  3 RQNKPPSKPNND  4 QNKPPSKPNNDF  5 NKPPSKPNNDFH  6 KPPSKPNNDFHF  7 PPSKPNNDFHFE  8 PSKPNNDFHFEV  9 SKPNNDFHFEVF 10 KPNNDFHFEVFN 11 PNNDFHFEVFNF 12 NNDFHFEVFNFV 13 NDFHFEVFNFVP 14 DFHFEVFNFVPC 15 FHFEVFNFVPCS 16 HFEVFNFVPCSI 17 FEVFNFVPCSIC 18 EVFNFVPCSICS 19 VFNFVPCSICSN 20 FNFVPCSICSNN 21 NFVPCSICSNNP 22 FVPCSICSNNPT 23 VPCSICSNNPTC 24 PCSICSNNPTCW 25 CSICSNNPTCWA 26 SICSNNPTCWAI 27 ICSNNPTCWAIC 28 CSNNPTCWAICK 29 SNNPTCWAICKR 30 NNPTCWAICKRI 31 NPTCWAICKRIP 32 PTCWAICKRIPN 33 TCWAICKRIPNK 34 CWAICKRIPNKK 35 WAICKRIPNKKP 36 AICKRIPNKKPG 37 ICKRIPNKKPGK 38 CKRIPNKKPGKK 39 KRIPNKKPGKKT 40 RIPNKKPGKKTT