Antigenic Fingerprinting following Primary RSV Infection in Young Children Identifies Novel Antigenic Sites and Reveals Unlinked Evolution of Human Antibody Repertoires to Fusion and Attachment Glycoproteins

Abstract

Respiratory Syncytial Virus (RSV) is the major cause of pneumonia among infants. Here we elucidated the antibody repertoire following primary RSV infection and traced its evolution through adolescence and adulthood. Whole genome-fragment phage display libraries (GFPDL) expressing linear and conformational epitopes in the RSV fusion protein (F) and attachment protein (G) were used for unbiased epitope profiling of infant sera prior to and following RSV infection. F-GFPDL analyses demonstrated modest changes in the anti-F epitope repertoires post-RSV infection, while G-GFPDL analyses revealed 100-fold increase in number of bound phages. The G-reactive epitopes spanned the N- and C-terminus of the G ectodomain, along with increased reactivity to the central conserved domain (CCD). Panels of F and G antigenic sites were synthesized to evaluate sera from young children (<2 yr), adolescents (14-18 yr) and adults (30-45 yr) in SPR real-time kinetics assays. A steady increase in RSV-F epitope repertoires from young children to adults was observed using peptides and F proteins. Importantly, several novel epitopes were identified in pre-fusion F and an immunodominant epitope in the F-p27. In all age groups, antibody binding to pre-fusion F was 2-3 folds higher than to post-fusion form. For RSV-G, antibody responses were high following early RSV infection in children, but declined significantly in adults, using either G proteins or peptides. This study identified unlinked evolution of anti-F and anti G responses and supportive evidence for immune pressure driven evolution of RSV-G. These findings could help development of effective countermeasures including vaccines.

Fig 1. Evaluation of RSV F and G gene fragment phage display library using neutralizing monoclonal antibodies.

(A) A gene fragment library displaying random fragments and spanning the entire F gene was used to map the binding epitope of anti-F monoclonal antibodies, Palivizumab and D25. The schematic on top represents the F protein including the location of previously discovered antigenic regions: ϕ, I, II and IV. The alignments of the displayed fragments on the phages selected by Palivizumab and D25 in the F-GFPDL are shown in blue and red bars, respectively. The thickness of the bars represents the relative frequencies of the bound fragments. Numbers at the bottom represent the amino acid residues of the RSV-F protein. (B) Representation of the Palivizumab and D25 antibody epitopes identified by either X-Ray crystallography or GFPDL within the pre-fusion F protein structure (PDB Id- 4JHW). Red refers to the D25 epitope and blue represents the Palivizumab epitope. (C) A gene fragment library displaying random fragments and spanning the entire G gene was used to map the binding epitope of MAb 131-2G. Schematic represents the G protein including previously identified central conserved domain (CCD). (D) Consensus amino acid sequences identified as the epitope recognized by the monoclonal antibodies using GFPDL or other methods.

Fig 2. Elucidation of antibody epitope profile against the RSV F protein following RSV primary infection in children.

(A) Distribution of phage clones after RSV-F GFPDL affinity selection with sera obtained from children at 9 months (top) or 15–18 months of age (bottom). The amino acid designation is based on the RSV-F protein sequence. Bar location indicates the homology of the displayed RSV-F protein sequence on the phage clones after affinity selection. The thickness of each bar represents the frequencies of repetitively isolated phage inserts (only clones with a frequency of two or more are shown). (B) Antigenic sites within the RSV F protein recognized following primary RSV infection. Previously described antigenic sites (sites ϕ, I, II and IV) are shown above, and the antigenic regions discovered in this study are depicted below the F protein schematic, and are color coded. (C) The number of clones that encoded for each antigenic site before and after RSV infection. (D) Distribution and frequency of phage clones expressing each of the key RSV F antigenic sites. The number of clones that encoded for each antigenic site was divided by the total number of F-GFPDL clones and represented as a percentage before and after RSV infection.

Fig 3. Structural representation of antigenic sites in pre- and post-fusion form of F identified using GFPDL.

The surface structure of the pre-fusion trimeric form of RSV F protein (PDB Id—4ZYP) is shown in white and the antigenic region in a monomer (chain A) is highlighted in red. The surface structure of the RSV F post-fusion trimer (PDB Id—3RRT) is shown in white and the antigenic regions from one monomer are highlighted in blue. The RSV F protein used for crystallography encompasses amino acid residues 26-513del108-134 of the complete 574 a.a. long protein sequence. In addition, amino acids 98–109 are not represented in the structures.

Fig 4. Elucidation of antibody epitope profile against the RSV G protein following RSV primary infection in children.

(A) Distribution of phage clones after RSV-G GFPDL affinity selection with sera obtained from children at 9 months (top) or 15–18 months of age (bottom). The amino acid designation is based on the RSV-G protein sequence. Bar location indicates identified inserts of the phage clones after affinity selection mapped on the RSV-G protein. The thickness of each bar represents the frequencies of repetitively isolated phage inserts (only clones with a frequency of two or more are shown). (B) Antigenic sites within the RSV G protein identified following primary RSV infection using GFPDL analysis are depicted below the G protein and are color coded. CCD motif is shown on top of the F protein schematic. (C) The number of clones that encoded for each antigenic site before and after RSV infection. (D) Frequency and distribution of the antigenic regions are represented as a percentage of the total number of affinity selected phage clones. The number of clones that code for each antigenic site was divided by the total number of phage clones and represented as a percentage before and after RSV infection.

Fig 5. Surface plasmon resonance (SPR) analysis of human sera from of different age group for binding to peptides from different antigenic sites within RSV-F and RSV-G.

Selected peptides of RSV-F and G proteins representing the antigenic sites in Figs 2 and 4 were chemically synthesized and tested for binding against individual sera samples using real time SPR kinetics experiment. Total antibody binding is represented as SPR resonance units (RU). Panels A-C show total antibody binding against the F peptides and panels D-F show total antibody binding to G peptides with serum/plasma samples from children less than 2 years old (A and D), adolescents; 10–14 years old (B and E) and adults 30–45 years old (C and F). Peptides in the X-axis are labeled by RSV-A2-F subunit followed by amino acid sequence (A-C) and for RSV-A2-G protein followed by amino acid sequence (D-F). Statistical significant differences of antibody reactivity between different age groups are depicted with p values.

Fig 6. SPR based analysis of RSV-F and RSV-G purified proteins reactivity with human sera from of different age groups.

(A) Serum samples collected from children (< 2 years), adolescents (10–14 years), and adults (30–45 years), were analyzed for total binding to purified pre-fusion RSV-F (blue) or post-fusion RSV-F (green) protein by SPR. The pre-fusion RSV F (DS-Cav1) was produced in 293F cells and encodes amino acids 26–513Δ110–136. Post-fusion RSV-F protein was obtained from Sino biologicals and encodes amino acids 22–529Δ110–136. (B) Ratio of serum binding to RSV pre-fusion vs post-fusion F protein in different age groups by SPR. (C) Binding to non-glycosylated RSV-G (blue) vs glycosylated RSV-G (green). The RSV-A2-G (67–298) ectodomains were produced in bacterial or mammalian 293F cells. Total antibody binding is represented in SPR resonance units. Individual sample reactivity to different F proteins (A) or to RSV-G proteins (C) are shown by connected lines. Statistical significant differences of antibody reactivity between different age groups or within group are depicted with p values.

Fig 7. Contribution of anti-G antibody response following RSV primary infection in the evolution of RSV viruses.

Accumulation of amino acid variations in RSV protein sequence in human isolates collected from 1979–2013. Protein sequences from RSV subtype A human isolates were obtained through the Virus Pathogen Resource (ViPR, http://www.viprbrc.org). 297 isolates with complete genome sequences and non-ambiguous amino acids were included in the analysis. The sequences were aligned using MUSCLE and a consensus sequence was created based on the alignment. The polymorphisms from that consensus were then scored by calculating the entropy (Hx) using Bioedit. Entropy scores are represented in the Y-axis and the RSV proteins in the X-axis. Alignment of antigenic sites in RSV-G identified using GFPDL approach with sequence variability scores in the RSV-G protein is shown the bottom half.