Characterization of neutralizing epitopes of the respiratory syncytial virus (Pneumoviridae: Orthopneumovirus: Orthopneumovirus hominis) F protein using a panel of monoclonal antibodies and escape mutants

Full Text

Introduction

Respiratory syncytial virus (RSV) is the leading cause of acute lower respiratory tract infections in children, the elderly, and individuals with underlying medical conditions. In children aged 0 to 5 years with lower respiratory tract diseases, RSV-induced pneumonia and bronchiolitis are observed in 28% of cases. Mortality from RSV infections reaches 13–22% in children and 23% among patients over 65 years of age [1, 2].

RSV has a single serotype, which is subdivided into two antigenic (AG) groups (RSV-A and RSV-B). Despite its relatively limited genetic diversity compared to other RNA-containing viruses, RSV does not confer long-term post-infection immunity, even against genetically closely related strains. RSV is characterized by widespread annual seasonal circulation in the human population. RSV-specific antibodies are present in virtually all children and adults; however, their presence does not prevent reinfection. In RSV infection, protective antiviral antibodies are produced against two surface glycoproteins of RSV: the G protein, responsible for viral attachment to the cell, and the F protein, which mediates fusion of the viral envelope with the cell membrane. Neutralizing G-specific antibodies are characterized by narrow specificity (strain-specific), whereas antibodies against the F protein demonstrate cross-reactivity with different RSV groups [3].

The development of virus-neutralizing monoclonal antibodies (mAbs) is a promising approach for the prevention and treatment of RSV infection. Due to the relatively high antigenic stability of the F-glycoprotein across strains, efforts to develop new anti-RSV immunotherapeutics are focused on producing mAbs that specifically interact with this protein. Specific anti-RSV preparations (palivizumab, nirsevimab, motavizumab, suptavumab) are humanized virus-neutralizing mAbs directed against the F protein [4].

Amino acid substitutions in certain epitopes of RSV surface glycoproteins can alter the virus's antigenic properties, allowing it to evade the immune response, which contributes to reinfection and a complicated course of the disease [5, 6]. The clinical course of RSV infection has been shown to depend on the genotype and antigenic group of RSV [7].

The widespread adoption of sequencing technology has facilitated the active identification of genetic changes in the structure of circulating RSV. In contrast, the antigenic variability of the virus has been scarcely studied. In this regard, the identification of the most immunogenic and evolutionarily stable antigenic determinants capable of inducing a protective response to RSV that changes over time is important for the development of effective specific immunobiological drugs. Currently, the only effective method for studying the antigenic properties of RSV is to characterize the virus's interaction with mAbs with known epitope specificity. Neutralizing epitopes of viral proteins are identified by sequencing the sequences of escape mutants (EMs) resistant to the neutralizing action of mAbs.

The aim of this study was to identify neutralizing epitopes of the F-protein of RSV subtype A (RSV-A) using a panel of mAbs with characterized epitope specificity.

Materials and methods

Isolation and cultivation of escape mutants. To produce the EMs, the prototype strain RSV-A Long (isolated in 1956) was used, provided by the National Institute for Medical Research (London). EMs were obtained by selection following the cultivation of various RSV dilutions in MA-104 cell lines (monkey kidney epithelial cells obtained from the Russian Cell Culture Collection of the Research Institute of Cytology of the Russian Academy of Sciences, St. Petersburg, Russia), grown in 96-well culture plates (JetBio-Filtration Co., Ltd., China) in the presence of high concentrations of neutralizing RSV-specific mAbs (ammonium sulfate fraction of ascites fluid from the abdominal cavities of mice). Serum-free alpha-MEM medium (BioloT LLC, St. Petersburg) was used as the culture medium. The mAb/RSV mixture was incubated for 1 hour at 37°C, after which it was added to plates containing a monolayer culture of MA-104 cells (100 μL of the mixture per well). After 5–7 days of cultivation in a CO₂ incubator (Sanyo, Japan) at 37 °C, the infected cell monolayer containing single symplasts was removed from the wells and subjected to repeated incubation with mAbs several (5–9) times. An inverted research microscope MS 700(I) (Ernst Leitz Wetzlar, Germany) was used to visualize the symplasts in the infected cell monolayer. RSV from the final dilution of the last passage was analyzed using whole-genome sequencing.

Microculture immunoenzyme assay (m-ELISA) was used to evaluate the interaction between mAbs and EMs. Monolayer cultures of MA-104 cells, grown in 96-well culture plates (JetBio-Filtration Co., Ltd., China), were infected with RSV Long or EM at a dose of 100 TCID₅₀. After 5–7 days of incubation at +37 °C in a CO₂ incubator (Sanyo, Japan), following the development of pronounced cytopathic effects (CPE) of the virus, the cells in the wells were fixed with 80% chilled acetone (Himmed, Russia) for 20 minutes. MAbs diluted with 5% skim milk (IP Khiba M.A., Russia) in 0.01 M phosphate-buffered saline (PBS-M), pH 7.2 (Biolot, Russia), and incubated for 2 h at +37 °C. After washing the plates with PBS, detection of the virus-bound mAbs was performed by adding a peroxidase-conjugated goat anti-mouse IgG (Sigma, USA), diluted 1 : 5000 in PBS-M. The plates were incubated for 1 hour at +37 °C. The peroxidase reaction was developed by adding a substrate mixture containing 0.1 mg/mL TMB (Sigma, USA) and 0.02% H₂O₂ in acetate-citrate buffer, pH 5.0 (Lenreactive, Russia). The reaction was stopped with 2N H₂SO₄ (Lenreactive, Russia). Optical density was measured on a MultiskanMS photometer (Labsystems, Finland) at a wavelength of 450 nm (OD₄₅₀).

The results were expressed as the percentage reduction in OD₄₅₀ values for the test group compared to the K+ control (OD₄₅₀ values obtained when the wild-type Long virus interacted with each of the antiviral agents). The data were analyzed under conditions of a directly proportional linear relationship between mAb binding and virus levels: OD₄₅₀ at mAb concentrations of 0.01–0.1 μg/mL.

Assessment of the virus-neutralizing activity of monoclonal antibodies against escape mutants. Twofold dilutions of mAbs (50 μL) were combined with equal volumes of virus-containing culture supernatant containing 100 TCID₅₀ of EM. 100 μL of the EM/mAb mixture was incubated for 2 h at 37 °C, after which it was added to wells containing a monolayer of MA-104 cells grown in 96-well culture plates (JetBio-Filtration Co., Ltd., China). The plates were incubated for 5–7 days in a CO₂ incubator until pronounced CPE developed in the K+ wells (100 TCID₅₀ of EM without mAb addition). The end-neutralizing concentration of mAb (μg/mL) was defined as the concentration in wells where at least 50% inhibition of symplastoformation (50% of the cell monolayer preserved) was visually observed compared to the total destruction of the cell monolayer in the K+ wells.

Sequencing of escape mutants. EM RNA was extracted from 250 μL of sample (RSV-infected cells suspended in Alpha-MEM medium) using the Magnoprime FAST-R reagent kit (Nextbio LLC, Russia). Whole-genome amplification was performed using a set of 12 primer pairs developed at the Research Institute of Influenza, with overlapping fragments of 60–160 nucleotides. The average length of the amplicons was set at approximately 1,350 nucleotides [8]. For reverse transcription polymerase chain reaction (RT-PCR), the BioMaster RT-PCR Premium reagent kit (Biolabmix LLC, Russia) was used. Full-length genome sequencing was performed using the IlluminaDnaPrep reagent kit (Illumina, USA). Genomic libraries were sequenced on the Illumina NextSeq2000 platform (Illumina, USA) in PE150 mode. The BWA algorithm was used for sequence alignment. Samtools and Ivar tools were used to generate consensus sequences.

Statistical analysis of the results obtained. To assess differences between two independent samples in terms of a quantitatively measured characteristic, we used the nonparametric Mann–Whitney U test. For multiple comparisons of independent samples within a single dataset, we used one-way ANOVA, followed by the Dunn test for pairwise comparisons between the two groups of data. Null hypotheses were rejected at a significance level of p < 0.05.

Results

Previously, the Laboratory of Diagnostic Biotechnology at the Research Institute of Influenza had developed a panel of 8 mAbs against RSV-A, Long strain: 5H8, 5F3, 7H8, 9E12, 5F8, 12C9, 7B12, and 10G6. All monoclonal antibodies interacted with oligomeric forms of the RSV F protein but not with its primary sequence [9], and possessed neutralizing activity against the immunogenic virus, although they were less effective than palivizumab (PVZ). Repeated passaging resulted in the selection of clones that are resistant to the neutralizing action of homologous mAbs, which were used to obtain the clones (Table 1).

 

Table 1. Neutralizing activity of monoclonal antibodies against escape mutants (EM)

Таблица 1. Нейтрализующая активность моноклональных антител (МКА) по отношению к эскейп-мутантам (ЭМ)

RSV strain Штамм РСВ	Passage of EM Пассаж ЭМ	Final neutralizing concentrations of mabs (μg/mL)* Конечные нейтрализующие концентрации МКА (мкг/мл)*
		5H8	5F3	7H8	9E12	5F8	12C9	7B12	10G6	PVZ
Original strain Long Исходный штамм Long	‒	6.3	1.2	1.8	2.6	2.2	0.5	1.4	0.6	0.1
EM 5Н8	8	> 71.3	> 40.6	> 31.3	> 84.4	> 84.4	6.3	> 91.9	> 103.1	0.6
EM 5F3	9	> 71.3	> 40.6	> 31.3	> 84.4	> 84.4	> 30.5	> 91.9	> 103.1	0.6
EM 7H8	7	> 71.3	> 40.6	> 31.3	> 84.4	84.4	3.1	> 91.9	> 103.1	0.6
EM 9E12	7	> 71.3	> 40.6	> 31.3	> 84.4	> 84.4	1.6	> 91.9	> 103.1	0.6
EM 5F8	5	> 71.3	> 40.6	> 31.3	> 84.4	> 84.4	> 30.5	> 91.9	> 103.1	0.6
EM 12C9	6	> 71.3	> 40.6	7.8	> 84.4	84.4	> 30.5	> 91.9	25.7	0.6
EM 7B12	6	> 71.3	40.6	31.3	> 84.4	15.6	3.1	> 91.9	3.2	0.3
EM 10G6	9	> 71.3	> 40.6	> 31.3	> 84.4	> 84.4	> 30.5	> 91.9	> 103.1	0.6

Note. * ‒ final concentration of mAbs (μg/mL) at which at least 50% inhibition of symplast formation was observed compared to positive control K+. K+ is the CPE observed in the wells of the plates upon addition of 100 TID50 of the virus in the absence of mAbs. Wells shaded in gray indicate the level of interaction of EM with «homologous» mAbs. Cases of partially retained neutralizing activity are highlighted in pink.

Примечание. * ‒ конечная концентрация МКА (мкг/мл), при которой наблюдалось ингибирование симпластообразования не менее 50% по сравнению с К+. К+ ‒ это ЦПД, наблюдаемое в лунках планшетов при добавлении 100 ТИД₅₀ вируса в отсутствие МКА. Серым окрашены ячейки, показывающие уровень взаимодействия ЭМ с «гомологичными» МКА. Жирным шрифтом выделены случаи частично сохраненной нейтрализующей активности.

 

To understand the mechanisms of action of mAbs, a comparative analysis was conducted of their neutralizing activity against the level of reaction with EMs in a micro-ELISA. It is known that the loss of the ability to effectively interact with viral antigens leads to a decrease in the neutralizing activity of antibodies.

This is consistent with the results obtained for mAbs 5H8, 7H8, 9E12, 12C9, and 5F8. A complete loss of neutralizing activity against homologous mutant clones (Table 1) was observed with a marked decrease (to 64%; p < 0.01) in the binding of these mAb clones to EM compared to the unmodified Long virus (Figure).

 

Figure. Reduction in the level of interaction of neutralizing MAbs with EM obtained for them, compared to RSV Long (microcultural ELISA data).

The Y-axis shows the degree of reduction (%) of OD₄₅₀ values in the microcultural ELISA for EM compared to C+. C+ – OD₄₅₀, shown in the microcultural ELISA during the interaction of the Long strain with each of the MAbs. The infection dose for MA-104 cells with viruses was 100 TID₅₀. The results were taken into account under conditions of a directly proportional linear relationship between the MAb and the antigen (OD₄₅₀ at MAb concentrations of 0.01 and 0.1 μg/mL).

Рисунок. Снижение уровня взаимодействия нейтрализующими МКА с ЭМ, полученными к ним, по сравнению с РСВ Long (данные мк-ИФА).

Ось ординат – степень снижения (%) показателей OD₄₅₀ в мк-ИФА для ЭМ по сравнению с К+. К+ – OD₄₅₀, показанные в мк-ИФА при взаимодействии исходного штамма Long с каждым из МКА. Доза заражения клеток МА-104 вирусами – 100 ТИД₅₀. Результаты учитывали в условиях прямо пропорциональной линейной зависимости связи МКА с антигеном (OD₄₅₀ при концентрациях МКА 0,01 и 0,1 мкг/мл).

 

These results indicated negative selection of EMs, as clones that effectively interacted with neutralizing mAbs disappeared from the viral population during passage.

For mAb 5F3, 7B12, and 10G6, the loss of the ability to neutralize the EMs generated using them was accompanied by a slight decrease in the level of response to them (a 0–15% reduction in OD₄₅₀ values; p > 0.05) (Figure). In this case, positive selection of clones with changes in the F-protein structure could be observed; these changes were not critical for binding to the mAbs but prevented the antibodies from exerting neutralizing activity.

It should be noted that the EMs obtained using certain mAbs were, in most cases, also resistant to other, heterologous mAbs (Table 1). Such synchronization of neutralizing activity may result from significant structural and/or conformational changes in the F-protein antigenic sites of all EMs, regardless of their origin. This was evidenced by the results of EM interaction with F-specific mAbs lacking neutralizing activity. Previously, in addition to neutralizing antibodies, 8 F-specific non-neutralizing mAbs were obtained, the reaction pattern of which also depended on the conformation of the F-protein. With all EMs, the binding level in the micro-ELISA of two such mAbs (4F2 and 1H3) was significantly reduced (p < 0.01) by 60–70% compared to Long. It should be emphasized that the target epitopes for the mAbs 4F2 and 1H3 did not overlap and did not compete for binding to the F-protein with the neutralizing mAbs [9].

According to competitive ELISA data, PVZ did not share any common determinants with the target epitopes for our mAbs [9]. However, PVZ neutralized all EMs, albeit at concentrations 3–6 times higher than those required for Long (Table 1). Thus, passaging Long RSV with the obtained mAbs did not lead to the selection of clones with significantly altered interaction sites with PVZ.

To map the neutralizing epitopes of the RSV F protein, whole-genome sequencing of the mAbs was performed. The identified substitutions in the mAbs are shown in Table 2.

 

Table 2. Amino acid substitutions in the proteins of RSV strain Long escape mutants (EM) obtained as a result of selection with F-specific neutralizing mAbs

Таблица 2. Аминокислотные замены в белках эскейп-мутантов (ЭМ) РСВ штамма Long, полученных в результате селекции F-специфичными нейтрализующими моноклональными антителами

Antigenc determinant Антигенные детерминант	EM strain Штамм ЭМ	EM passage Пассаж ЭМ	Amino acid substitutions in EM proteins Аминокислотные замены в белках ЭМ
			F	G	L	N	P	M	M2-2
А	5Н8	8	N276S	S269T* A184P	Нет / No	L201Q	No / Нет	No / Нет	No / Нет
	7H8	7	N276S	S269T	R396G K1532E	No / Нет	L186F	No / Нет	No / Нет
	5F3	9	N276S I530M	S269TF208L	Нет / No	L201Q	No / Нет	No / Нет	No / Нет
	9E12	7	N276S M526L	S269T	K1532E	No / Нет	No / Нет	No / Нет	No / Нет
В	7B12	6	N276S R190S	T108K	I1871V	No / Нет	No / Нет	No / Нет	No / Нет
	12C9	6	N240S R190S	Нет / No	L1175M	No / Нет	I48T	No / Нет	H53Y G71D
	5F8	5	N240S R190S	Нет / No	Y2135H	No / Нет	No / Нет	L177P E178G	No / Нет
	10G6	9	N515K R190S	Нет / No	K1532E N2164D	No / Нет	Q80H	No / Нет	No / Нет

Note. * ‒ the numbering of position 269 in the G protein is given according to the RSV Long sequence deposited in Genbank (KF713490.1). In the Nextstrain database, position 269 corresponds to position 293, since it takes into account the 24-AA insertion observed in the G protein of almost all modern RSV-A. For earlier viruses without this insertion, the absence of the insertion is designated as a deletion. Therefore, this position is also designated as 293.

Примечание. * ‒ нумерация положения 269 в G-белке дана в соответствии с последовательностью РСВ Long в Genbank (KF713490.1). В базе данных Nextstrain 269 положению соответствует положение 293, поскольку учитывается инсерция 24 аминокислот, наблюдаемая в G-белке практически у всех современных РСВ-А. Для более ранних вирусов без данной инсерции отсутствие вставки обозначена как делеция. В этой связи данная позиция также обозначается как 293.

 

Based on the characteristics of the mAbs under study, which we obtained previously [9], as well as the genetic and antigenic characteristics of EM presented above, it can be assumed that the mAbs interacted with two antigenic determinants of the F protein of RSV (provisionally designated as A and B).

Determinant A, with which the neutralizing mAbs 5H8, 7H8, 5F3, and 9E12 interacted, included AC 276. Furthermore, mAbs 5F3 and 9E12 interacted with amino acid residues 530 and 526, respectively.

The mAbs 12C9, 9B12, 10G6, and 5F8 reacted with determinant B. The amino acid residue at position 190 was common to the binding of these 4 mAbs to the F-protein. Position 240 was also critical for the neutralizing activity of mAbs 5F8 and 12C9. The epitope for binding of mAbs 7B12 included the amino acid residue at position 276, which is also characteristic of site A. The amino acid residue at position 515 was part of the target epitope for mAb 10G6.

Interestingly, under the influence of different mAbs, EMs with identical substitutions in the primary sequence of the F protein were obtained. At the same time, in some cases, mAbs of the same specificity reacted differently with the same viruses.

Thus, the target epitopes for mAbs 5H8 and 7H8 had a single amino acid substitution in the F protein–N276S–while those for mAbs 12C9 and 5F8 had identical substitutions, R190S and N240S (Table 2). Despite this, it has previously been shown that mAbs 5H8 and 5F8 (OD450 0.59 ± 0.10 and 0.58 ± 0.09) reacted significantly weaker in the micro-ELISA with the Long immunogenic virus than mAbs 7H8 and 12C9 (OD₄₅₀ 0.91 ± 0.15 and 0.95 ± 0.13; p < 0.01) [9].

Furthermore, these mAb pairs also differed in their neutralizing properties. The reduced affinity of mAbs 5H8 and 5F8 can explain their weaker virus-neutralizing activity against both the Long immunogen virus and EM. The mAb 12C9, unlike mAb 5F8, neutralized EM 5H8, 7H8, and 9E12, although less effectively than the Long RSV. For its part, mAb 7H8, unlike mAb 5H8, exhibited neutralizing activity against EM 12C9 (Table 1).

Whole-genome sequencing of EMs showed that F-specific mAbs facilitated the selection of RSV Long clones with changes in the primary structure not only of the F protein but also of other viral proteins (G, L, N, P, M, M2-2). In certain cases, substitutions in different proteins were associated. Thus, in all 4 EMs, the A-determinants of the F-N276S substitution were detected simultaneously with G-S269T. Of these, two clones (EM 7H8 and 9E12) also had substitutions in the L-K1532E polymerase. In the other two EMs from this group (EM 5H8 and 5F3), changes were detected in the N-protein (N-L201Q) (Table 2).

The Nextstrain database was used to analyze the potential evolutionary significance of the substitutions identified in EM. Approximately 3,000 RSV-A amino acid sequences circulating in the human population from the 1970s to the present were analyzed.

Viruses with 190S and 276S in the F-protein structure, corresponding to the most frequently detected R190S and N276S substitutions in EM, appeared in the mid-1980s. At that time, viruses of the new GA2/ON1 genotype were first detected; the vast majority of RSV-A viruses that have been circulating widely since 2010 to the present belong to this genotype [10]. The proportion of viruses with the 190S and 276S structures was 99.9% and 77.0%, respectively. The R190 position, characteristic of the Long strain (isolated in 1956 and classified as genotype GA1), was detected only in isolated isolates. The N276 position, also characteristic of Long, was detected in 22.3% of viruses circulating from the late 1970s to 2010, i.e., among viruses with genotypes preceding GA2/ON1.

In the G protein, the positions 208F and 269/293S characteristic of the Long virus were detected until the mid-1990s. From 2010 to the present, 208L and 269/293T have been detected in the vast majority of viruses, corresponding to the F208L and S269/293T substitutions in EM.

The amino acid 2135N mutation in the structure of polymerase L first appeared in RSV-A in the mid-1990s, and the proportion of such viruses has now reached 90.3%.

Apart from the aforementioned 5 substitutions in the amino acid residues of the F-, G-, and L-proteins of EM, which consequently turned out to be identical to the structure of the vast majority of modern RSV-A of genotype GA2/ON1, all other changes in the F, G, L, N, P, M, and M2-2 proteins (Table 2) were random in nature and occurred in only 0–0.3% of cases among viruses represented in the Nextstrain database.

Discussion

The genetic diversity of RSV is due to the high variability of the G-protein gene. The variability of the F-protein gene is an order of magnitude lower [11, 12]. Nevertheless, an amino acid polymorphism has been observed in the F-protein sequence, which depends on the time and location of RSV circulation [10, 13]. Despite the low level of genetic variability in the F protein, significant phenotypic variations have been demonstrated in related traits: the efficiency of symplastogenesis, the virus's sensitivity to temperature, and its ability to interact with virus-neutralizing antibodies [14]. This paradox is explained by variations in the conformational structure of the RSV F protein.

The surface F-glycoprotein of RSV is synthesized as an inactive precursor (F0), which is post-translationally cleaved by the protease furin at two polybasic sites separated by 27 amino acids, into two chains, F1 and F2, which remain covalently linked by two disulfide bonds. The mature F protein within the viral particle is a homotrimer (F1xF2) in a metastable pre-fusion (pre-F) conformation. The pre-F form is necessary for viral entry and fusion of the viral envelope with the cell membrane or the membranes of infected and uninfected cells. During these processes, the F protein undergoes a series of conformational changes, leading to the formation of a highly stable structure (post-fusion, post-F). Three heptad repeat (HR) domains interact with each other, facilitating the transition from the pre-F to the post-F form [11].

Five major antigenic sites in the structure of the RSV-A F protein are known to induce the synthesis of neutralizing antibodies: Ø (zero, amino acid sequence 92–96, 195–227); I (amino acid sequence 27–45, 312–318, 379–389); II (AS 254–277); III (amino acid sequence 46–54, 301–311, 345–353, 367–378); IV (amino acid sequence 422–471); V (amino acid sequence 55–61, 146–194, 287–300) [15].

Conformational changes in the F-protein directly influence the characteristics of the antigenic sites and the binding of antibodies to them. Antigenic sites II and IV retain their properties in both the pre-F and post-F structures, whereas antigenic sites Ø and V are present on the surface of the F-protein only in the pre-F form. Antibodies interact more actively with site III in the pre-F conformation, whereas with site I, conversely, in the post-F conformation [3].

The results of this study suggest the presence of two neutralizing antigenic determinants of the F-protein, which we have provisionally designated as determinants A and B and which complement the structure of those already known.

Determinant A included the amino acid residue 276, which is part of antigenic site II, with which many F-specific protective mAbs interact, including PVZ (whose binding epitope is amino acid sequence 262–276). Some authors claimed that the combined K272E and N276S mutations caused resistance to PVZ [16], while others denied this [17]. Position 276 is also part of the epitope with amino acid sequence 255–278, which is responsible for the development of a Th1-mediated anti-RSV protective immune response [18]. Furthermore, determinant A includes positions 526/530 of the transmembrane domain, through which the F protein is incorporated into the membranes of the endoplasmic reticulum and Golgi apparatus during glycosylation as the F protein matures, as well as into the viral envelope and the cellular plasma membrane.

The existence of such an unusual site in the F-protein of RSV is supported by the findings of other researchers, who obtained an EM image of the Long virus showing two substitutions in the F-protein similar to those we identified: in antigenic site II (N268I) and in the transmembrane domain (V533M) [19].

For the second determinant B proposed by us, position 190, which is part of antigenic site V, proved to be decisive. The amino acid residue 190 is important for stabilizing the conformation of the F protein [20] and is part of an epitope on the surface of B lymphocytes (amino acid sequence 176–240) that induces the synthesis of neutralizing antibodies in humans [21]. Furthermore, the following amino acid residues were also identified: 276 from antigenic site II, 515 from the HR2 region, and 240.

Position 515 is part of one of three conserved hydrophobic regions, the heptad repeats (HR): HR2 with the amino acid sequence 474–523. HRs consist of a double helix of α-strands with a coiled-coil formed by seven amino acid residues. HR regions interact with each other, facilitating the transition of the conformationally unstable pre-fusion structure of the F-protein into a stable post-fusion structure necessary for fusion with the cell membrane [3]. Position 240 is not part of any known neutralizing antigenic sites. However, antibodies identified in RSV-infected children interacted with a peptide containing the amino acid sequence 234–287 [21].

It should be noted that virtually all of the sites critical for the response of the analyzed mAbs were part of the determinants (antigenic sites II, V, HR2), where amino acid substitutions could affect the conformation of the F protein and, consequently, the nature of the virus's interaction with the mAbs. At the same time, all the neutralizing mAbs we obtained interacted only with the conformationally organized form of the F protein, but not with its primary sequence [9].

The above suggests that the discrepancies observed in some cases between the primary structure of the F-protein of the obtained EMs and the nature of the reaction of F-specific mAbs with them are a consequence of conformational variations in the target epitopes.

Thus, despite their specificity for epitopes with the same amino acid sequence of the F-protein, mAbs 5H8 and 7H8, as well as 12C9 and 5F8, differed significantly from one another in certain parameters (the level of suppression of RSV synplasty, the ability to compete for binding with neutralizing and non-neutralizing viral epitopes).

Thus, in RSV infection, viruses with identical primary structures can induce antiviral immune responses of varying efficacy, depending on the conformation of the antigenic sites on the F protein.

Other researchers have also observed a lack of correlation between amino acid substitutions in RSV surface glycoproteins and changes in antibody interaction with them. At the same time, a direct dependence of antibody neutralizing activity on the conformation of antigenic sites has been demonstrated [14, 22, 23]. The most effective protective mAbs, including the commercial drugs palivizumab and nirsevimab, are strictly conformation-dependent and primarily target the pre-F form [14].

Interestingly, under the influence of various neutralizing monoclonal antibodies, EMs with identical substitutions in the primary structure of the F protein were obtained.

The selection of RSV with identical mutations in the F protein under the action of different F-specific monoclonal antibodies has also been observed by other authors [19, 23, 24].

An explanation may lie in the existence, alongside the well-known phenomenon of immunodominance (the induction of specific antibody synthesis predominantly against certain antigenic determinants), of a hierarchy in the variability of individual epitopes of the RSV F protein under the influence of external factors.

Whole-genome sequencing revealed that, under the influence of F-specific mAbs, EMs were obtained with substitutions not only in the F protein but also in the G, L, N, P, M, and M2-2 proteins.

All substitutions in the surface glycoprotein G were located in known functionally significant determinants. The amino acid residue 108 is part of a B-cell epitope (amino acid sequence 90–110) identified in RSV-infected mice [25]. The amino acid residue 108 in RSV-A is O-glycosylated [26]. The T108K substitution we identified in EM 7B12 results in the loss of a glycosylation site, which inevitably affected the antigenic characteristics of the epitope.

The substitution at position 184 is located in the central hydrophobic conserved region (AP 156–194), which is involved in protein-protein interactions during the formation of the 3D structure of the G protein. Most G-specific mAb reported in the literature interacted with this region. This amino acid also binds to the site with amino acid sequence 182–186 of the RSV G-protein, which mimics the amino acid sequence of fractalkine (a chemokine of the CX3C family). As a result, RSV acts as a fractalkine antagonist, blocking its CX3CR1 receptors on many immune cells. This led to the inhibition of the influx of CX3CR1+ effector cells to the site of viral infection and the suppression of the RSV-specific protective immune response [27].

Position 208 is critical for interaction with G-specific neutralizing antibodies [28]. Position 269 is part of the B-cell immunodominant determinants described in both animals (amino acid sequence 245–274, 263–298) [25] and in humans (amino acid sequence 265–273) [29].

The viral genomic RNA is encapsulated in a ribonucleoprotein (RNP) complex, which consists of nucleoprotein (N), phosphoprotein (P), and polymerase L. Polymerase L contains three conserved enzymatic domains: RNA-dependent RNA polymerase (RdRp), polyribonucleotidyl transferase (PRNTase/capping domain), methyl transferase (MTase), as well as the C-terminal, most variable domain. Protein P is a cofactor for the polymerase. The regulatory factor M2-2 plays an important role in inhibiting transcription, thereby promoting viral RNA replication. Matrix protein M lines the inner surface of the viral envelope and plays a key role in virion morphogenesis [30].

The altered amino acids at positions 396, 1175, and 1871 in the L protein of EM were located in the RdRp, PRNTase, and MTase domains, respectively, while the amino acid at position 1532 was located in the connector domain (CD), which lacks catalytic activity. Amino acids 2135 and 2164 are located in the C-terminal domain, which regulates MTase enzymatic activity [30].

Information on the structure of antigenic sites in other RSV proteins is extremely limited in the literature.

Interestingly, some recurrent substitutions observed in EM in different proteins were grouped into clusters: F-N276S/G-S269T/N-L201Q (EM 5H8 and 5F3), as well as F-N276S/G-S269T/L-K1532E (EM 7H8 and 9E12).

This study is the first to demonstrate an association of substitutions in RSV surface glycoproteins and intravirion proteins, although the complementarity of substitutions for RSV replicative complex proteins has recently been described. Analysis of currently circulating RSV-A strains allowed the authors to identify an association between certain non-synonymous substitutions in the L protein and specific patterns of changes in the primary sequences of the P and M2-2 proteins. The authors suggested the possibility of a reciprocal influence of changes in the structure of RSV replicative complex proteins (N, P, M2-1, M2-2, and L) on the emergence of variations in the viral genome caused by adaptive changes in polymerase characteristics (its rate and accuracy of function) [12].

The simultaneous substitutions we identified in the sequences of different EM proteins indicate coordinated interaction among RSV proteins, which is consistent with the literature.

Mature RSV particles assemble on the apical surface of cells in the form of thread-like structures (filaments). Experimental evidence has been obtained for the presence of strong bonds between F and G proteins within such filaments. The G protein facilitated the cleavage of the F protein precursor and the conformational stabilization of its mature form, thereby ensuring intercellular transmission of the virus [31]. The G protein, as part of a constructed virus-like particle, enhanced the binding of F-specific mAbs to the F protein [32].

The F protein associated with the filaments has binding sites for proteins of the replicative complex (L, N, P) as well as for M and M2-1 proteins [33]. The specific conformation of the cytoplasmic domain of the F-protein ensures binding to the M, N, and P proteins and their incorporation into viral filaments on the cell surface [34]. In turn, the MTase and CD domains of the L-polymerase, to which the substitution at position 1532 — identified three times in our EM analyses — belongs, can interact with the P protein [12].

Thus, under the influence of immune pressure exerted by neutralizing F-specific monoclonal antibodies, RSV strains with specific changes in this protein may be selected within the human population. In light of data on the structural relationship and conformational correspondence between the surface and internal protein domains of RSV, selected viruses may simultaneously contain substitutions in other proteins (e.g., the G protein) that are complementary to the altered structure of the F protein.

This may revise the widely held postulate regarding the independent evolution of RSV F and G genes [13].

On the other hand, changes in the amino acid sequences of the G protein and/or ribonucleoprotein complex proteins may directly or indirectly alter the conformation of the associated F protein, thereby influencing the nature of the interaction between F-specific antibodies and the virus. This may be another reason for the observed discrepancy between the genetic and antigenic structures of RSV.

Using the Nextstrain database, an analysis was conducted of the potential evolutionary significance of the substitutions identified under the pressure of neutralizing antibodies in the Long virus strain, described in 1956. The results presented showed that the substitutions F-R190S, F-N276S, G-F208L, G-S269/293T, and L-N2135H were similar to those observed for RSV-A, which emerged in the population in the mid-1980s. These changes were evolutionarily fixed in the vast majority of viruses circulating from 2010 to the present and distinguish modern RSV of the GA2/ON1 genotype from viruses that circulated previously.

The presented data suggest that the generation of EMs can be considered a simplified model for determining the direction of RSV evolution under the influence of immune pressure generated by virus-specific neutralizing antibodies.

Conclusion

The widespread prevalence of RSV in the population and the diversity of clinical manifestations of RSV infection underscore the need for a thorough understanding of the pathogenesis of this disease. Currently, the primary focus is on viral genome sequencing, which allows for the analysis of the extent and direction of RSV genetic variability. However, this does not allow for the assessment of antigenic properties and their changes in circulating viruses, as it does not account for the conformation of viral proteins and interactions between various structural components of RSV, which influence the structure of its antigenic sites. Studying the antigenic variability of RSV requires the use of antibodies with known characteristics. In this study, the epitopic specificity of 8 neutralizing mAbs against the F protein of RSV-A was determined as a result of whole-genome sequencing of the obtained mAbs. Two neutralizing antigenic determinants of the F protein have been identified, which complement the structure of those already known. For the first time, a correlation has been demonstrated between substitutions in surface F- and G-glycoproteins and intravirion proteins (M, N, P, L, M2-2) in RSV clones obtained through selection under pressure from F-specific neutralizing mAbs. The use of virus-specific monoclonal antibodies in the analysis of currently circulating RSV will allow for a better understanding of the pathogen's evolution and the nature of the immune response to it, which must be taken into account when developing new antiviral drugs.

About the authors

Vera Z. Kryvitskaya

Smorodintsev Research Institute of Influenza, Ministry of Health of the Russian Federation

Email: vera.krivitskaya@influenza.spb.ru
ORCID iD: 0000-0002-9146-0816

D.B.Sc., Leading Researcher, Laboratory of Risk Factors for Influenza and Acute Respiratory Viral Infections, Department of Etiology and Epidemiology

Russian Federation, 197022, Saint Petersburg

Ekaterina R. Petrova

Smorodintsev Research Institute of Influenza, Ministry of Health of the Russian Federation

Email: ekaterina.petrova@influenza.spb.ru
ORCID iD: 0000-0003-2258-4679

Researcher, Laboratory of Risk Factors for Influenza and Acute Respiratory Viral Infections, Department of Etiology and Epidemiology

Russian Federation, 197022, Saint Petersburg

Evgeniy V. Sorokin

Smorodintsev Research Institute of Influenza, Ministry of Health of the Russian Federation

Email: evgeniy.sorokin@influenza.spb.ru
ORCID iD: 0000-0003-1732-1727

PhD, Head of the Laboratory of Biotechnology of Diagnostic Products, Department of Biotechnology

Russian Federation, 197022, Saint Petersburg

Tatyana R. Tsareva

Smorodintsev Research Institute of Influenza, Ministry of Health of the Russian Federation

Email: tatyana.tsareva@influenza.spb.ru
ORCID iD: 0000-0003-4757-0521

Senior Researcher, Laboratory of Biotechnology of Diagnostic Products, Department of Biotechnology

Russian Federation, 197022, Saint Petersburg

Ksenia S. Komissarova

Smorodintsev Research Institute of Influenza, Ministry of Health of the Russian Federation

Email: kseniya.komissarova@influenza.spb.ru
ORCID iD: 0000-0002-1465-5548

Researcher, Laboratory of Molecular Virology, Department of Etiology and Epidemiology

Russian Federation, 197022, Saint Petersburg

Anna A. Sominina

Smorodintsev Research Institute of Influenza, Ministry of Health of the Russian Federation

Email: anna.sominina@influenza.spb.ru
ORCID iD: 0000-0002-1015-595X

D.B.Sc., Head of Laboratory, Chief Researcher, Laboratory of Risk Factors for Influenza and Acute Respiratory Viral Infections, Department of Etiology and Epidemiology

Russian Federation, 197022, Saint Petersburg

Daria M. Danilenko

Smorodintsev Research Institute of Influenza, Ministry of Health of the Russian Federation

Author for correspondence.
Email: daria.danilenko@influenza.spb.ru
ORCID iD: 0000-0001-6174-0836

PhD, Head of the Department of Etiology and Epidemiology, Leading Researcher, Laboratory of Etiology of Viral Infections, Deputy Director for Research

Russian Federation, 197022, Saint Petersburg

References

Supplementary files