Culture Medium Reshapes RSV Glycoprotein Sugars and Antibody Recognition

Respiratory syncytial virus, or RSV, remains one of the most consequential pathogens of early childhood and old age, a leading cause of severe lower respiratory tract disease in infants, young children and older adults. Among its surface proteins, the attachment glycoprotein G is extraordinary in the sheer scale of its post-translational modification: it carries four to five N-linked glycosylation sites and roughly thirty to forty O-linked sites, sugar decorations that inflate its apparent molecular weight severalfold beyond what its amino acid sequence alone would predict. These glycans are thought to contribute to antigenic masking and to other functional properties of the protein, and their extent has long been known to vary between laboratory systems. A new study now adds an unexpected variable to that list: the composition of the cell culture medium in which the virus or its proteins are produced.

Researchers at the University of Veterinary Medicine Hannover set out to disentangle two factors that had previously been confounded in studies of RSV G protein heterogeneity. Earlier work had shown that G proteins made in primary human airway epithelial cultures, including air-liquid interface systems, reach an apparent molecular weight of approximately 170 kilodaltons and are more heavily glycosylated than the 55 to 90 kilodalton species typical of transformed lines such as HEp-2 or A549. Other experiments had shown that, within a single complete medium, different continuous cell lines produce G proteins differing in electrophoretic mobility and in reactivity with G-specific monoclonal antibodies. But in those comparisons either the cell line varied while the medium was held constant, or both cell identity and medium changed together. By varying the medium within a single cell background, the new study isolates the contribution of the medium itself from that of the cell.

The rationale for suspecting such an effect is grounded in glycobiology. Culture medium composition can influence the availability of nucleotide sugar donors, the activated sugars that glycosyltransferases in the Golgi apparatus attach to nascent glycoproteins, and can modulate the activity of glycosylation pathways more broadly. Because G protein glycans are likely to shape antigenic masking and functional behaviour of the virus, medium-dependent metabolic effects were a plausible driver of the heterogeneity reported across systems. The question also matters for antibody-based interventions: approved therapeutic monoclonal antibodies against RSV, including palivizumab, nirsevimab and clesrovimab, are all directed at the fusion F protein, yet virion composition and glycoprotein organisation may influence antibody binding and neutralising activity.

To test the principle directly, the team first used a recombinant expression system. A soluble form of the RSV G ectodomain, tagged for purification, was produced in HEK293T cells maintained in three different commercial media, designated A, B and C, which had been selected from an initial screen of nine formulations to represent the extremes and midpoint of the observed range. Purified proteins were analysed by SDS-PAGE and Western blot. The apparent molecular weight of the soluble G protein differed substantially depending on the medium: protein produced in medium A was the smallest at roughly 60 to 90 kilodaltons, medium B yielded an intermediate species of about 100 kilodaltons, and medium C produced the largest form at approximately 100 to 130 kilodaltons.

The critical control came from glycosylation-deficient cells. When the same soluble G protein was expressed in HEK293T cells knocked out for the GALE and GALK1 genes, which are deficient in galactose metabolism, the medium-dependent size differences vanished entirely. In these knockout cells the protein appeared as a lower molecular weight cluster of roughly 15 to 25 kilodaltons, consistent with incompletely processed, non-galactosylated glycans that the knockout does not eliminate. This result demonstrated that the differences in apparent size were caused by glycosylation and that medium composition, not cell-type-specific differences in the glycosylation machinery, defined the extent of that glycosylation. Importantly, the same hierarchy of medium-dependent molecular weights was reproduced during authentic infection: when A549 cells were infected with a reporter RSV strain in the three media and lysed 24 hours later, Western blots showed the same pattern, albeit with smaller absolute differences.

The team then asked whether the medium affects the biological properties of virus preparations. A549 cells were infected with an EGFP-expressing RSV at a multiplicity of one in each medium, with Opti-MEM, the reduced-serum medium B commonly used for RSV propagation, serving as the reference condition. Virus spread kinetics, quantified by integrated GFP fluorescence over 120 hours with a live-cell imaging system, revealed pronounced medium-dependent differences, with spread significantly higher in medium C than in media A and B. More strikingly, when virus stocks produced in each medium were used to infect fresh A549 cells at equal infectious doses and the inoculum was replaced with Opti-MEM after adsorption, the differences persisted: medium C virus spread most, medium B virus showed an intermediate phenotype, and medium A virus remained markedly attenuated. Because the production medium was no longer present during the infection period, and because a difference confined to the initial inoculum would be expected to produce comparable spread curves with a delay rather than a sustained gap, the result points to a stock-intrinsic property that also affects subsequent rounds of infection.

Analyses of the virus stocks themselves added a further layer. When stocks were adjusted to equal total protein concentrations and probed by Western blot, the relative G protein signal normalised to the viral N protein was lower for virus produced in media A and B than for virus produced in medium C, while relative F protein signals did not differ significantly between preparations. The G protein with the highest apparent molecular weight, consistent with the most extensive glycosylation, was thus associated with the highest relative G to N ratio, an association consistent with a link between G protein maturation and its relative abundance in stocks, though the authors are careful to note that this does not establish causality. A comparison of RT-qPCR signals with infectious titres proved revealing: virus produced in medium A yielded higher Cq values than medium C virus, yet its infectious titre was approximately 100-fold lower, indicating that the ratio of infectious to RT-qPCR-detectable viral material differed between preparations.

The consequences extended to antibody recognition. Fluorescently labelled palivizumab, which binds both pre-fusion and post-fusion F, and the pre-F-specific antibodies nirsevimab and clesrovimab were used to stain infected cells cultured in the three media, with signals normalised to the GFP reporter. Binding of all three therapeutic antibodies was significantly lower in cells cultured in media B and C than in medium A, whereas a post-F-specific control antibody showed largely unchanged binding. Because relative F protein signals in the virus stocks did not differ, these binding differences were not accompanied by detectable differences in total F protein abundance, though the analysis did not distinguish F conformations. In neutralisation assays, palivizumab consistently showed the highest IC50 values at approximately 1.5 micrograms per millilitre, with nirsevimab and clesrovimab more potent, and no statistically significant differences in IC50 emerged between viruses produced in media B and C, indicating that the altered binding signals did not translate into measurable differences in neutralising potency under the conditions tested.

The authors conclude that the choice of culture medium is an important variable that should be considered whenever RSV preparations are produced for studies of the virus’s biological and immunological properties. Several caveats frame the findings: the precise compositions of the commercial media are proprietary, precluding identification of the responsible components; medium B is reduced-serum rather than serum-free, so its effects may partly reflect serum content and lower nutrient availability; the study used transformed cell lines and a single reporter virus strain, so whether similar effects occur in primary airway epithelial, air-liquid interface or organoid systems remains open. Notably, such advanced cultures are maintained in differentiation media that differ substantially from those used for continuous lines, raising the possibility that the heavier glycosylation of G reported in primary cultures reflects the medium as well as cell identity. Resolving the exact medium-dependent glycan patterns will require site-resolved glycoproteomics, and the F protein’s glycosylation and conformational distribution merit parallel investigation. For now, the message to the field is clear: two laboratories growing the same virus in the same cells but different media may not be studying the same virus at all.

Subject of Research: Effect of cell culture medium composition on respiratory syncytial virus G protein glycosylation and monoclonal antibody recognition

Article Title: Cell culture medium influences RSV G protein glycosylation and recognition by monoclonal antibodies

Article References: Engling, N., Knittler, M.-C., Tagore, R., Ludlow, M., Osterhaus, A. D., Rimmelzwaan, G. F., & Meineke, R. (2026). Cell culture medium influences RSV G protein glycosylation and recognition by monoclonal antibodies. Virology Journal, 23(1), Article 226. https://doi.org/10.1186/s12985-026-03318-8

Image Credits: AI Generated

DOI: 10.1186/s12985-026-03318-8

Keywords: respiratory syncytial virus, glycosylation, G glycoprotein, cell culture medium, monoclonal antibodies, palivizumab, nirsevimab, clesrovimab, glycobiology, viral infectivity, HEK293T cells, A549 cells