Species-Specific Conservation of Linear Antigenic Sites on Vaccinia Virus A27 Protein Homologs of Orthopoxviruses

The vaccinia virus (VACV) A27 protein and its homologs, which are found in a large number of members of the genus Orthopoxvirus (OPXV), are targets of viral neutralization by host antibodies. We have mapped six binding sites (epitopes #1A: aa 32–39, #1B: aa 28–33, #1C: aa 26–31, #1D: 28–34, #4: aa 9–14, and #5: aa 68–71) of A27 specific monoclonal antibodies (mAbs) using peptide arrays. MAbs recognizing epitopes #1A–D and #4 neutralized VACV Elstree in a complement dependent way (50% plaque-reduction: 12.5–200 µg/mL). Fusion of VACV at low pH was blocked through inhibition of epitope #1A. To determine the sequence variability of the six antigenic sites, 391 sequences of A27 protein homologs available were compared. Epitopes #4 and #5 were conserved among most of the OPXVs, while the sequential epitope complex #1A–D was more variable and, therefore, responsible for species-specific epitope characteristics. The accurate and reliable mapping of defined epitopes on immuno-protective proteins such as the A27 of VACV enables phylogenetic studies and insights into OPXV evolution as well as to pave the way to the development of safer vaccines and chemical or biological antivirals.


Introduction
The genus Orthopoxvirus (OPXV) contains a group of large and closely related DNA viruses within the family Poxviridae, encompassing viruses that replicate in the cytoplasm of vertebrate or invertebrate cells [1,2]. Vaccinia virus (VACV), the prototype of the genus, was applied as the vaccine against the related Variola virus (VARV). This vaccination campaign led to the eradication of smallpox [3,4].

Plaque Reduction Test
The neutralization potency of six A27-specific mAbs was tested by plaque reduction test (PRT) against VACV Elstree as reference strain. Purified antibodies were diluted with MEM (PAN-BIOTECH, Aidenbach, Germany) and adjusted to a concentration of 400 µg/mL. A volume of 125 µL of the antibody preparations was titrated in two-fold serial dilutions on 96-well microplates containing 125 µL/well MEM supplemented with 2.5% FCS to avoid antibody coating. After antibody titration, one dilution series received 1% sterile human complement (Sigma Aldrich, Taufkirchen, Germany) per well, the other remained free of complement. Then, 100 pfu (125 µL) of VACV Elstree was added to each well. As plaque-forming control, 250 µL MEM/well with or without 0.5% human complement, containing 100 pfu VACV Elstree was used. The virus negative control was 250 µL MEM/well alone with or without 0.5% human complement. After incubation of the 96-well microplates at 37 • C for one hour, the mixtures were transferred to 24-well plates containing a confluent monolayer of Vero cells. After incubation at 37 • C for one hour, the supernatants were poured out and replaced by 0.5 mL MEM containing 2.5% FCS and 0.5% methyl cellulose (Sigma Aldrich, Taufkirchen, Germany). The plates were then incubated at 37 • C for 48 h, before the cells being fixed and stained with a solution containing 25% formaldehyde, 8.5% ethanol and 1.5% crystal violet. The plaques were counted by visual inspection while illuminated. Neutralization was determined as ≥50% plaque reduction compared to the virus control. Each PRT was performed in triplicates

Inhibition of Cell Fusion and Syncytium Formation
Cell fusion experiments were performed as described before [31,67,70]. Confluent BS-C-1 monolayers cultured in MEM with 2.5% FCS in 24-well plates (1 mL/well) were infected with 100 pfu/well VACV WR for 1 h at 37 • C, washed twice and incubated either with warm medium alone or with warm MEM containing purified mAbs (200 µg/mL). Then, 24 h post infection, the cells were incubated for 3 min at 37 • C at pH 4.8 with warm fusion buffer (phosphate-buffered saline with 10 mM 2-(N-morpholino)ethanesulfonic acid and 10 mM HEPES). The cells were washed twice with warm MEM (treated for two min at 37 • C). Afterwards, warm medium (MEM + 1% FCS), with or without mAbs (200 µg/mL), was added again. The cells were incubated for 4 h at 37 • C and then observed by phase-contrast microscopy. An indicator for cell fusion was the formation of syncytia, which are large, structure-less, fused cell areas [54].

Binding Affinities of the Monoclonal Antibodies (mAbs) in Indirect Enzyme-Linked Immunosorbent Assays (ELISAs)
For quantification of the binding affinities of mAbs to different OPXVs, an indirect enzyme-linked immunosorbent assay (ELISA) was applied; 96-well microplates were coated with 1 µg/mL of the VACV strains Bern, CVA, Elstree, IHD-J, Copenhagen wild type (WT), Copenhagen host range (HR), R325, TT, the neuro-vaccinia virus strains Hagen, Levaditi and Munich1, the modified VACV Ankara (MVA) as well as the OPXV strains camelpox virus (CMLV) CP1, cowpox virus (CPXV) KR2 Brighton, mousepox virus (ectromelia; ECTV) Munich 1, and MPXV Copenhagen in carbonate/bicarbonate buffer (pH 9.6; 100 µL/well). After blocking with 2% skimmed milk and 10% fetal calf serum in PBS, purified mAbs adjusted to a concentration of 50 µg/mL were titrated in two-fold serial dilutions (100 µL/well). Incubation was performed at 37 • C for 1 h. After five washing steps with PBS, peroxidase conjugated goat anti-mouse IgG (whole molecule; Sigma Aldrich, Taufkirchen, Germany) produced in goats was added to the 96-well microplate in a working dilution of 1:2000 (100 µL/well) and incubated at 37 • C for 1 h. Thereafter, the plate was washed five times with PBS again, before the developing solution (3, 3 , 5, 5 tetramethylbenzidine; Abcam, Cambridge, UK) was added (100 µL/well). The reaction was stopped by 1 N hydrochloric acid (50 µL/well). The OD-values were measured by a photometric plate reader (TECAN Sunrise plate reader with the Magellan complete software, Männedorf, Switzerland) at a wavelength of 450 nm. Affinity was calculated from the average absorption of the triplicates using Michaelis-Menten kinetics [71,72] and the program GraphPad Prism version 7.00 for Mac (La Jolla, CA, USA).

Epitope Mapping by SPOT Synthesis on Cellulose Membranes
The whole A27 protein of VACV Copenhagen [52,56] representing 110 amino acids [56], was directly synthesized stepwise on derivatized cellulose membranes through 101 decapeptides with an offset of one aa (9 aa overlap). The synthesis on derivatized cellulose membranes using Fmoc-protected amino acid pentafluorophenyl or /V-hyroxyoxo-dihydro-benzotriazine esters and the screening were performed according to the method described before [73] and the manufacturer of the SPOTs kit (Cambridge-Research Biochemicals, ICI, representative in Germany IC-Chemikalien, Carl-Zeiss-Ring 15, Ismaning).

Epitope Mapping by Microarray Scanning Chips
An OPXV microarray chip was designed as depicted in Figure S1. 15-mer peptides overlapping by 12 amino acids (3 aa offset) were synthesized via SPOT synthesis on a cellulose membrane [73], passed through the SC 2 process [74] and spotted onto microscope glass slides. The chip contained eight identical arrays of 521 peptides each ( Figure S1A). A total of 475 of those overlapping peptides represented the entire amino acid sequences of A27, D8, H3, L1, A33, and B5 proteins of VACV Western Reserve ( Figure S1B, GenBank accession number: AY243312.1). Forty-six peptides were amino acid variations of VACV A27 and D8 proteins to the corresponding homologs of other OPXVs (Table S1). In addition, 10 cellulose-conjugated biotin spots served as a positive control and orientation for the SPOT Calling Program. The OPXV microarray chip was designed to screen four samples simultaneously. Therefore, each peptide was printed eight times to obtain technical replicates, which could be divided into four identical sub-arrays using an adhesive chamber (SecureSeal, Sigma-Aldrich Co. LLC, USA). In order to obtain equal antibody concentrations of 2 µg/µL per chamber, protein concentrations were measured using a NanoDrop ND-1000 Spectrophotometer. The screening procedure with the microarray chip was performed as previously reported [75].

Fine Mapping of the VACV A27 Epitopes by Microarray Analysis
Similar mapping results were obtained when using the OPXV microarray chip imprinted with 521 pentdecapeptides with 12 aa overlap. Epitope #1A (mAb 5B4/2F2) was only one aa longer compared to the SPOTs membrane and was, therefore, directed to the sequence region aa 31-KREAIVKAD-39. Epitopes #1B (mAb 2C11/1B4), #1C (mAb 3F5/2D5) and #1D (mAb 1D5/2D11) were all assigned to the aa region 28-PEAKRE-33. For epitope #1B, the microarray chip and the SPOTs membrane yielded identical results. The epitope #1D was mapped to the same region, but only one aa shorter on the microarray chip. Epitope #4 (mAb 2G8/1E4) was allocated to aa 7-PGDDDLAIPATE-18 and, therefore, by 6 aa longer compared to results from the SPOTs membrane. MAb 5B1/1A11 (epitope #5), however, did not react with any of the peptides on the chip, although the target sequences detected on the SPOTs membranes were present in the microarray spots no. 20-23 and 493-496 (Tables S1 and S2 and Figure S3). In the following investigations, we refer, therefore, to the epitope locations provided by the SPOTs membrane, because they were regarded to be more accurate due to the shorter aa offset of one aa compared to three aa in the microarrays.

Fine Mapping of the VACV A27 Epitopes by Microarray Analysis
Similar mapping results were obtained when using the OPXV microarray chip imprinted with 521 pentdecapeptides with 12 aa overlap. Epitope #1A (mAb 5B4/2F2) was only one aa longer compared to the SPOTs membrane and was, therefore, directed to the sequence region aa 31-KREAIVKAD-39. Epitopes #1B (mAb 2C11/1B4), #1C (mAb 3F5/2D5) and #1D (mAb 1D5/2D11) were all assigned to the aa region 28-PEAKRE-33. For epitope #1B, the microarray chip and the SPOTs membrane yielded identical results. The epitope #1D was mapped to the same region, but only one aa shorter on the microarray chip. Epitope #4 (mAb 2G8/1E4) was allocated to aa 7-PGDDDLAIPATE-18 and, therefore, by 6 aa longer compared to results from the SPOTs membrane. MAb 5B1/1A11 (epitope #5), however, did not react with any of the peptides on the chip, although the target sequences detected on the SPOTs membranes were present in the microarray spots no. 20-23 and 493-496 (Tables S1 and S2 and Figure S3). In the following investigations, we refer, therefore, to the epitope locations provided by the SPOTs membrane, because they were regarded to be more accurate due to the shorter aa offset of one aa compared to three aa in the microarrays.

Inhibition of Cell Fusion
A27 was initially designated as the fusion protein [54,79,80]. However, more recent evidence indicates that there is not only one fusion protein in the envelope of IMV, but rather a fusion complex consisting of at least 11 proteins [36,37,81]. Evidence now suggests that the A27 protein is not integrated into the fusion complex [36,82]. Other investigations reported a second fusion complex consisting of A17 and A27 [65], where the fusion event of VACV WR at pH 4.8 was inhibited by anti-A27 mAbs. Therefore, we retested this effect using three epitope-mapped anti-A27 mAbs from our collection to cover the entire target region. Fusion of infected BS-C-1 cells was indicated by the formation of large areas of fused cells, rather than separate individual cells ( Figure 3A). Fusion was inhibited by the mAb 5B4/2F2 directed to epitope #1A (aa 32-39) ( Figure 3B). The mAb 3F5/2D5 against epitope #1C (aa 26-31) was binding upstream of the mAb 5B4/2F2 and was not able to block cell fusion ( Figure 3C). The same was observed for mAb 5B1/2G6 binding to the C-terminal epitope #5 (aa 68-71) ( Figure 3D). improved in the presence of complement. The mAbs 3F5/2D5 and 1D5/1E10 neutralized the VACV Elstree only in the presence of 1% complement, while no neutralization was observed with the mAb 5B1/2G6.

Inhibition of Cell Fusion
A27 was initially designated as the fusion protein [54,79,80]. However, more recent evidence indicates that there is not only one fusion protein in the envelope of IMV, but rather a fusion complex consisting of at least 11 proteins [36,37,81]. Evidence now suggests that the A27 protein is not integrated into the fusion complex [36,82]. Other investigations reported a second fusion complex consisting of A17 and A27 [65], where the fusion event of VACV WR at pH 4.8 was inhibited by anti-A27 mAbs. Therefore, we retested this effect using three epitope-mapped anti-A27 mAbs from our collection to cover the entire target region. Fusion of infected BS-C-1 cells was indicated by the formation of large areas of fused cells, rather than separate individual cells ( Figure 3A). Fusion was inhibited by the mAb 5B4/2F2 directed to epitope #1A (aa 32-39) ( Figure 3B). The mAb 3F5/2D5 against epitope #1C (aa 26-31) was binding upstream of the mAb 5B4/2F2 and was not able to block cell fusion ( Figure 3C). The same was observed for mAb 5B1/2G6 binding to the C-terminal epitope #5 (aa 68-71) ( Figure 3D).

Binding Affinities of the mAbs to Various Variants of OPXVs
Binding affinities of the purified mAbs to the six A27 epitopes detected in VACV Elstree were determined by indirect ELISAs on microplates coated with the purified reference strains VACV-MVA, VACV, CPXV KR2 Brighton, CMLV CP1, ECTV Munich 1, and MPXV Copenhagen. The binding curves were determined in triplicates for each virus strain. In case of the VACV strains, with the exception of VACV MVA, all data were calculated as mean values. VACV MVA was presented

Binding Affinities of the mAbs to Various Variants of OPXVs
Binding affinities of the purified mAbs to the six A27 epitopes detected in VACV Elstree were determined by indirect ELISAs on microplates coated with the purified reference strains VACV-MVA, VACV, CPXV KR2 Brighton, CMLV CP1, ECTV Munich 1, and MPXV Copenhagen. The binding curves were determined in triplicates for each virus strain. In case of the VACV strains, with the exception of VACV MVA, all data were calculated as mean values. VACV MVA was presented alone in order to compare affinity data directly to other VACV strains ( Figure 4). All mAbs directed to epitope complex #1 showed strong binding activity to VACV, CPXV and CMLV, but did not react with or bound only weakly to ECTV and MPXV. In all VACV strains, the mAb 5B4/2F2 bound to its epitope #1A equally well. There was no difference in the amino acid sequence of the respective epitope. An 11.5-23-fold decrease in binding activity was observed with CPXV KR2 Brighton and CMLV CP1. Responsible for this finding were obviously the aa exchanges D39E in CPXV and V36I in CMLV. In ECTV Munich 1 and MPXV Copenhagen, the epitope #1A could not be detected, apparently due to aa exchanges R32H and I35T in ECTV and D39Y in MPXV. Epitope #1B was detected by the mAb 2C11/1B4 in VACVs, CPXV and CMLV with a similar affinity, whereas aa exchanges A30D and R32H in ECTV and A30T in MPXV caused the loss of the mAb reaction. Epitope #1C was also detected equally well in VACVs, CPXV and CMLV by the corresponding mAb 3F5/2D5. In ECTV, the kinetics of the mAb were reduced 25-to 53-fold according to the aa exchange A30D. In MPXV, the epitope was only very weakly detectable. The mAb 1D5/2D11 against epitope #1D, which is only one aa longer than epitope #1B (A at position 34), reacted equally well with VACVs, CPXV and CMLV. Despite the aa exchanges A30D and R32H in ECTV, which were also present, the mAb detected the epitope with 2.6 to 7.6-fold weaker affinity compared to VACVs, CPXV and CMLV. Even in MPXV, the epitope #1D was detected by the mAb, albeit with a 4.6-(ECTV) to 34.8-fold (VACVs) weaker intensity. In contrast to the heterogeneous species-specific binding behavior of mAbs directed to the epitope complex #1A-D, the mAbs targeting epitopes #4 and #5 showed the same strong binding activities to all OPXVs tested. V max -and K m -values were in the same range.

Species-Specific Epitope Conservation and Variation among the OPXV Members
A total of 391 amino acid sequences of the OPXV A27 protein homologs from the GenBank were analyzed with respect to species-specific conservation or variation of the six sequential antigenic sites mapped. Epitope #4, located at the N-terminus of the A27 protein between aa residues 9-14 (9-DDDLAI-14), is highly conserved within the genus OPXV (Tables S3-S5). This motif was found in 372 of the 391 analyzed sequences.

Discussion
The A27 protein is immunogenic and highly conserved within the members of OPXV [32,61,65]. In this study, six antigenic sites on the A27 protein (epitope #4: aa region 9-14, epitope complex #1A-D: between aa 26 and 39 and epitope #5: aa region 68-71) were mapped. With respect to epitopes #4 and #1A-D, the mapping with the SPOT synthesis and microarray chip showed similar results with only few amino acids divergence. By contrast, epitope #5 could not be detected when the OPXV microarray chip was employed. Although the granularity of membranes is higher with an offset of one, compared to an offset of three in the microarrays, also the resolution of the microarray slide is sufficient to detect epitope #5 with a length of four spots. One reason may be degradation of the respective antibodies over time, as a time lag >5 years between the former spot measurement and the more recent microarrays existed. Another reason may be the use of different side chain protection groups. For production of membranes, Cys(acetyl-aminomethyl) was used, whereas for peptides in the microarray assay Cys(triphenylmethyl) was used that was deprotected during slide preparation via trifluoroacetic acid (TFA). As a cysteine residue is involved in binding, a formation of disulfide bonds or incomplete Cys deprotection may have altered antibody binding properties, thereby leading to different signals in the microarrays compared to the membranes [83]. The OPXV microarray chip, however, was used to screen for additional species-specific epitope variations by aa exchanges according to the GenBank database entries.
Epitope #4 was conserved among all OPXVs and nearly genus-specific, because the main motif 9-DDDLAI-14 was found in 372/391 database entries. In 13 OPXV strains, the N-terminal sequences could not be assessed due to truncations in the sequences uploaded from GenBank. In 3/26 BPXVs, the motif was 9-DDDLAT-14, whereas all three SkPXVs contain the motif 9-DDDMAI-14. The changed aa sequence (L12M) was represented by spot 479 of the OPXV microarray chip and mAb 2G8/1E4 reacted with this spot (Table S7). In 2D predictions of the secondary structure of the A27 protein, a β-turn was evident in this area with a high antigenicity. This was expected as in previously published data epitopes often were identified in the region of β-turns [84][85][86][87]. Predictions on hydrophilicity [88] and surface probability [89] did not show any special features for this region of the A27 protein. Nevertheless, the mAb 2G8/1E4 against epitope #4 showed equally good binding affinities to all OPXV reference strains tested. Its neutralization capacity could be enhanced by the addition of complement. However, in previous studies the mAb 2G8/1E4 showed no neutralizing abilities against the tested ECTV M1 [40]. The discrepancy is caused by another test setup including another OPXV strain (VACV Elstree instead of ECTV M1) and different cells (Vero cells instead of MA-cells) as well as by a 20-fold higher initial concentration of the antibody. The same was true for mAbs 3F5/2D5 and 1D5/2D11 (epitope complex #1 as mentioned further below). Epitope #5 was the most highly conserved one in A27 as the motif 68-IEKC-71 was present in 389/391 aa sequences. In all three SkPXVs and 1/61 VACV sequences the epitope was shifted downstream to 93-IEKC-96. In 1/134 CPXV strains, the motif was 68-IEKY-71 while in another CPXV the epitope was missing because the C-terminus was truncated. The mAb 5B1/2G6 against epitope #5 was not neutralizing. However, the antigenic site is located in a functionally very important area within the C-terminus of the A27 protein. In this hydrophilic region, the two cysteines at positions 71 and 72 are responsible for formation of disulfide bonds and, therefore, play an important role for a functionally active trimeric A27 structure [52,79]. In 2D predictions of the secondary structure, a β-sheet (aa 58-83) followed by two β-turns (aa 70-75) was evident in this constant area. Two β-turns led generally to a high antigenicity [85]. Predictions on hydrophilicity [88] and surface probability [89] did not reveal any special features. The mAb 5B1/2G6 showed similar binding affinities to all OPXV reference strains tested. This was expected because of the high sequence conservation of the targeted epitope among OPXV.
The most important antigenic region of the A27 protein was confined by aa 26-39. This has already been known from a previous investigation that identified functional domains in the A27 envelope protein [90]. In our present study, a complex of four closely related epitopes (#1A-D) could be allocated to this region. The narrow location of these epitopes has already been predicted previously from data obtained with two overlapping oligopeptides [91] and from quantitative competitive ELISAs performed with purified mAbs and viruses [40]. Two of the four mAbs binding to these epitopes were neutralizing in vitro. In this study, the four mAbs could enhance virus inhibition after adding complement. Other authors identified this region also as a strong target for binding of mAbs [42]. An epitope with a larger extension (aa 21-40) comprised the area of the epitope complex #1A-D completely (aa [26][27][28][29][30][31][32][33][34][35][36][37][38][39]. However, those mAbs directed to this epitope region only neutralized in the presence of complement. According to 2-D structure predictions and published data [52,91,92], the A27 region with the four epitopes #1A-D was classified as hydrophilic. Between aa residues 25 and 45 a hypervariable structure region was found. In case of VACV and VARV, it started with an α-helix up to aa 40, followed by two β-turns. In CPXV and CMLV, however, the α-helix changed at aa 37 into three and four β-turns, respectively. MPXV showed two β-turns at aa residues 25-34 followed by an α-helix up to aa 39 and three β-turns. In ECTV, the β-sheet structure was found up to aa 30, followed by an α-helix up to aa 37 and three β-turns. These highly variable structural conditions led to a significant species-specific difference in the overall structure of the investigated A27 proteins. Thus, the proteins of the species VACV, VARV and CMLV had a more linear form, while the proteins of the species MPXV, CPXV and ECTV were folded to a larger extent. Considering the epitope complex #1A-D, the aa main motif was 26-KKPEAKREAIVKAD-39. Based on GenBank database entries for 391 A27 protein sequences, this motif in the complete form (aa 26-39) was found in 210/391 OPXVs (68/69 VARV, 59/61 VACV, 26/26 BPXV, 2/2 HSPV, 2/2 RPXV, 51/134 CPXV, 2/3 TaPXV). The database entries for the 391 A27 protein sequences also indicated that this region could be defined as a very variable area with a lot of aa exchanges and structural differences. Affinity experiments showed, that the binding of the four mAbs to their respective targets was different and obviously dependent on aa exchanges. The epitopes #1A and #1B were completely absent in MPXV and ECTV. Especially in case of MPXV, three aa exchanges led to the motif variation 26-KNPETKREAIVKAY-39, independent of the geographic distribution of isolates. In ECTVs, three different aa exchanges in comparison to VACV led to the motif to 26-KKPEDKHEATVKAD-39. In both OPXV genera, these aa exchanges were absolutely species-specific (57/57 MPXV; 14/14 ECTV). To investigate the direct influence of the exchanged amino acids on the binding of the corresponding mAbs, the epitope #1A was re-synthesized on a SPOTs membrane in the unchanged (VACV) and changed (MPXV and ECTV) design. The aa exchanges led unequivocally to the loss of mAb binding to its epitope #1A ( Figure S4) which was confirmed by lack of binding to spots 488 and 489 on the OPXV microarray chip (Table S7). The variations between aa 26 and 39 led also to a significant structural change of the A27 protein homologs of MPXV and ECTV. The change of the aspartic acid at position 39 of MPXV to tyrosine, containing a benzene ring, was mainly responsible for the loss of the mAb reactivity. The binding site of the main immunogenic epitope #1A was defined as an octapeptide of 31-REAIVKAD-39, when the three decapeptides on the SPOTs membrane with the highest spot intensity (11)(12)(13) were taken for epitope determination ( Figure S2). However, in all seven decapeptides (No. [10][11][12][13][14][15][16], even those with weaker reactions, used for the evaluation, it became clear that the tetrapeptide 35-IVKA-38 was the most important factor for binding of the mAb 5B4/2F2 (epitope 1A). It was apparent, that the VACV tetrapeptide 35-IVKA-38 was present in the CMLV variation 35-IIKA-38. If the CMLV specificity is referred to the whole region of epitope complex #1A-D, database analysis will reveal that the aa exchange V36I is unique for all 18/18 CMLVs, independent of their geographical origin (Africa, Asia), thereby leading to the two motifs 26-KKPEAKREAIIKAD-39 (17/18) (Table S7), whereas the epitope #1A was not detected. In VPXV and SkPXV (spots 491 and 492), the epitope complex #1A-D could not be detected by any of the mAbs.
Because of the fact that the sequence differences in the A27 region, representing the epitope complex #1A-D, were species-specifically conserved, the Old World OPXVs, such as VARV, VACV, HSPV, RPXV, BPXV, ECTV, MPXV, CMLV, and TaPXV, as well as the New World OPXVs, like SkPXV, RCNV and VPXV, revealed a monophyletic character. The sequence variations in this area, however, were not species-specifically conserved in CPXVs, which is why this group was regarded as polyphyletic. This taxonomic arrangement was concordant with previous investigations, where CPXVs were classified into different clades based on whole genome analysis [93][94][95]. According to the most recent findings, CPXVs were divided into four clades, CPXV-like 1, CPXV-like 2, VACV-like and VARV-like [93]. In our study, we could identify seven CPXV variants when referring only to the amino acid sequences of the epitope complex #1A-D ( Figure 5).
The A27 protein was formerly categorized as a fusion protein [32,52,54] and believed to mediate the direct fusion of virus and cytoplasm membranes ("fusion from without") [52,79]. Hitherto, A27 is not settled to be a part of a fusion complex consisting of at least 11 different proteins (A16, A21, A28, F9, G3, G9, H2, J5, L1, L5 and O3), being conserved in all OPXVs [36,37,81]. Still, there is also evidence in the literature that the A27 and A17 proteins form a second fusion complex [65], which was assigned to fusion proteins type I. Typical for type I viral fusion proteins is the presence of a coiled-coil structure [59], which is, beside the A27, also seen in influenza virus HA2 [96], Ebola GP2 [97,98] and HIV gp41 [99]. The authors suggested that the A17-A27-complex is transported to the cell membrane during viral replication and mediates fusion of the infected cells ("fusion from within"), meaning that A17 is the membrane-anchoring domain with the fusion peptide (aa 18-34) and A27 is responsible for the oligomerization as well as the membrane-attachment [65]. A27 binds to the GAG heparan sulfate of neighboring cells. This binding is mediated through the aa residues 26-KKPE-29 [34,57,61,62], resulting in an accumulation of cells in the immediate vicinity [65]. In several studies, mAbs against the A27 protein were able to block the "fusion from within" in a model described previously [32,52,54]. Therefore, we used this model to test inhibition of the fusion by three anti-A27-mAbs from our collection, whose antigenic sites were exactly mapped. Through low-pH treatment [54,100], we were able to induce fusion of VACV infected cells. This "fusion from within" was indicated by the formation of large and structureless fused cell areas known as syncytia [54]. The mAb 3F5/2D5 directed to epitope #1C (aa 26-KKPEAK-31) was not able to block the fusion, although the GAG binding site being inside its epitope. Acid-induced syncytia were formed. By adding the mAb 5B4/2F2 directed to epitope #1A (aa 32-REAIVKAD-39) and binding just more to the C-terminus of the mAb 3F5/2D5, fusion could be inhibited. The reason for the inhibition is not clear at the moment. However, at least a steric hindrance of the mAb could be ascertained. Moreover, antibodies may directly interfere with interactions by occupying binding sites or sterically hindering binding sites in close proximity. In addition, antibody binding affects protein conformation, and different antibodies have different effects on protein conformation that may alter distant interacting sites. The mAb 5B1/2G6 binding to the C-terminal epitope #5 (aa 68-71) failed to block the fusion by showing polykaryon formation, too. The epitope of this mAb is directly related to the binding site (aa 71-72) of the A26 fusion suppressor protein to the A27 protein, but there was no direct influence on the fusion event.
In summary, we mapped six antigenic sites on the A27 protein of VACV. This enabled us to interpret species-specific epitope variations and conservations of various OPXVs to gain an impression of their phylogenic relationships. To elucidate structure function relationships in more detail, co-crystallization might be helpful for future investigations. Moreover, the data on antigenic sites for cross-reacting or monospecific neutralizing antibodies are of high relevance for target directed screening of human immunoglobulin libraries to generate specifically engineered human recombinant antibodies, which might help in controlling any future outbreak of zoonotic orthopoxviruses.
Supplementary Materials: The following are available online at http://www.mdpi.com/1999-4915/11/6/493/s1: Figure S1: Layout of the poxvirus microarray chip. A: The chip contains eight arrays arranged in the same manner. B: Arrangement of spots on each array: yellow is peptides representing A27 antigen of Vaccinia virus Western Reserve; red, D8; blue, H3; grey, L1; green, A33; orange, B5. The lower part of the array contains peptide sequence variations of A27 (lilac) and D8 (pink) antigens of the other orthopoxviruses. Sites of the sample deposition depicted by the plus symbol. The number serves as the batch identification and letters represent the chip name. C: Amino acid sequence of the first three peptides covering A27, corresponding to spots #1-#3 in B. Figure S2: The 14 kDa A27 protein of VACV Copenhagen was synthesized on a SPOTs membrane as 101 decapeptides with 9 aa overlap to cover the whole sequence of 110 aa. By immunodetection with the mAb 5B4/2F2 seven spots were identified (No. [10][11][12][13][14][15][16] to carry the target sequences spots 11-13 showed the strongest signals. The eight amino acids the three peptides had in common were represented to the sequence region 32-REAIVKAD-39. The minimal sequence essential for binding is 35-IVKA-38 and marked in red. Figure S3: The OPXV microarray is based on 521 15-mer peptides overlapping by 12 aa. These peptides were spotted by SPOT technique on a chip. After immunodetection, the responding spots with the highest common intensity were chosen and outlined. Epitope #1A (mAb 5B4/2F2) was directed to a sequence region aa 31-KREAIVKAD-39. Epitopes #1B (mAb 2C11/1B4), #1C (mAb 3F5/2D5) and #1D (mAb 1D5/2D11) were all assigned to the aa region 28-PEAKRE-33. Epitope #4 (mAb 2G8/1E4) was allocated to aa 7-PGDDDLAIPATE-18. MAb 5B1/1A11 (epitope #5), however, did not react with any of the peptides on the chip. Figure S4: Immunological detection of ECTV-and MPXV-specific aa exchanges of the epitope #1A sequence on SPOTs membrane by the mAb 5B4/2F2. Spot 89: VACV epitope #1A (32-REAIVKAD-39). Spot 90: MPXV epitope #1A (32-REAIVKAY-39). Spot 91: ECTV epitope #1A (32-HEATVKAD-39). The aa exchanges led unequivocally to the loss of the binding of the mAb 5B4/2F2. The weak response on spot 91 was also observed with the secondary antibody used for detection. Table S1: Reactivity of six A27-specific mAbs on the OPXV microarray chip. The amino acids marked in red are species-specific sequence variations to the VACV WR or epitope sequence exchanges. *Accession number is representative for several sequences of this type. All GenBank accession numbers can be found in Table S4 immunodetection, the responding spots with the highest common intensity were chosen and outlined. Epitope #1A (mAb 5B4/2F2) was directed to a sequence region aa 31-KREAIVKAD-39. Epitopes #1B (mAb 2C11/1B4), #1C (mAb 3F5/2D5) and #1D (mAb 1D5/2D11) were all assigned to the aa region 28-PEAKRE-33. Epitope #4 (mAb 2G8/1E4) was allocated to aa 7-PGDDDLAIPATE-18. MAb 5B1/1A11 (epitope #5), however, did not react with any of the peptides on the chip. Figure S4: Immunological detection of ECTV-and MPXV-specific aa exchanges of the epitope #1A sequence on SPOTs membrane by the mAb 5B4/2F2. Spot 89: VACV epitope #1A (32-REAIVKAD-39). Spot 90: MPXV epitope #1A (32-REAIVKAY-39). Spot 91: ECTV epitope #1A (32-HEATVKAD-39). The aa exchanges led unequivocally to the loss of the binding of the mAb 5B4/2F2. The weak response on spot 91 was also observed with the secondary antibody used for detection. Table S1: Reactivity of six A27-specific mAbs on the OPXV microarray chip. The amino acids marked in red are species-specific sequence variations to the VACV WR or epitope sequence exchanges. *Accession number is representative for several sequences of this type. All GenBank accession numbers can be found in Table S4. Different shades of grey represent strength of the fluorescence intensity. n.d.: not detected Weak reaction Strong reaction Table S2: Detection of the six A27 antigenic sites by the corresponding anti-OPXV-mAbs via SPOTs-membrane and OPXV peptide microarray chip. The matches within the epitope sequence are highlighted and underlined. Table S3: Mapping of epitope #4 based on 391 complete and partial amino acid sequences from the NCBI GenBank database. Differences within the epitope sequence are highlighted. Table S4: 391 amino acid sequences of the OPXV A27 protein homologs available so far in the GenBank, being analyzed with respect to speciesspecific conservation or variation of the six sequential antigenic sites mapped. Table S5: Nucleotide sequences of the OPXV A27 protein homologs available so far in the GenBank, showing mutations which are crucial for amino acid exchanges as well as silent mutations. Table S6: Mapping of epitope #5 based on 391 complete and partial amino acid sequences from the NCBI GenBank database. Differences within the epitope sequence are highlighted. Table S7: Mapping of epitope complex #1A-D based on 391 complete and partial amino acid sequences from the NCBI GenBank database. Differences within the epitope sequence are highlighted.  Table S2: Detection of the six A27 antigenic sites by the corresponding anti-OPXV-mAbs via SPOTs-membrane and OPXV peptide microarray chip. The matches within the epitope sequence are highlighted and underlined. Table S3: Mapping of epitope #4 based on 391 complete and partial amino acid sequences from the NCBI GenBank database. Differences within the epitope sequence are highlighted. Table S4: 391 amino acid sequences of the OPXV A27 protein homologs available so far in the GenBank, being analyzed with respect to species-specific conservation or variation of the six sequential antigenic sites mapped. Table S5: Nucleotide sequences of the OPXV A27 protein homologs available so far in the GenBank, showing mutations which are crucial for amino acid exchanges as well as silent mutations. Table S6: Mapping of epitope #5 based on 391 complete and partial amino acid sequences from the NCBI GenBank database. Differences within the epitope sequence are highlighted. Table S7: Mapping of epitope complex #1A-D based on 391 complete and partial amino acid sequences from the NCBI GenBank database. Differences within the epitope sequence are highlighted.