Introduction

Influenza A virus (IAV) is a highly contagious respiratory pathogen responsible for annual epidemics and sporadic pandemics, causing significant morbidity worldwide1,2. Due to the simultaneous expression of the avian-like α-2,3 and human-like α-2,6 sialic acid (SA) receptor, pigs can serve as intermediate hosts between birds and humans, facilitating the adaptation of avian influenza viruses (AIV)3,4. The Eurasian avian-like (EA) clade 1 C H1N1 swine influenza viruses (SIVs) originated from avian sources in 1979, subsequently spreading through pig populations in Europe and Asia and causing sporadic human infections5,6. Over prolonged evolution, the receptor binding preferences of EA H1N1 SIV have switched from the α-2,3 SA receptor to the dual binding of α-2,3 SA and α-2,6 SA receptor, and some recent isolates of EA H1N1 SIV have demonstrated greater propensity to bind to α-2,6 SA receptor7,8. Moreover, after the global spread of the 2009 H1N1 pandemic and rapid reverse-zoonosis of this virus back into pigs, reassortants between EA H1N1 SIV with the H1N1pdm09, arose widely, some of which exhibit effective transmission through respiratory droplets in ferrets9,10,11, placing these viruses as potential instigators of the next influenza pandemic12,13.

The emergence of an antigenically novel virus capable of efficiently infecting and transmitting between humans creates a scenario where EA H1N1 SIV could cause a pandemic14,15,16. During the evolution of EA H1N1 SIV, adaptive mutations have arisen to overcome barriers between species, and specific viral substitutions have been identified as being critical for the transmission of the virus within mammalian populations17,18. The E190D and D225E substitutions in HA are well characterized as altering sialoglycan specificity19,20,21. Viruses with 225E in HA replicated faster than those with 225 G due to differences in assembly and budding efficiency, possibly because the HA 225 mutation alters the salt bridge structure between amino acids D225 and K222, resulting in the receptor binding switch18,22. The complex interplay between virus and host factors determines the adaptation of EA H1N1 SIV to different host species23,24. However, the critical factor that contributes to HA 225E adaptation remains unknown. Therefore, identifying the host factors required for EA H1N1 SIV infection and exploring the potential host factors that drive adaptive substitutions contribute to elucidating the mechanisms underlying interspecies transmission and facilitate the development of targeted interventions aimed at disrupting its transmission pathways.

In this work, using a CRISPR/Cas9-based high-throughput loss-of-function screening approach, we are able to provide a functional genomics resource critical for understanding the host factors involved in EA H1N1 SIV infection in pigs. Here, vitamin K-dependent gamma-carboxylase (GGCX) is identified for the first time as an essential host factor for EA H1N1 SIV infection. Our results show that GGCX catalyzes the carboxylation modification of viral HA protein, crucial for progeny virus binding to α-2,6 SA receptors. These findings suggest that GGCX is a major underlying driver of D225E substitution of EA H1N1 SIV and catalyzes the carboxylation modification of viral HA 225E, which determines the receptor binding preference of the EA H1N1 SIV and may serve as a potential target to prevent EA H1N1 SIV replication and cross-species transmission.

Results

GGCX is required for efficient EA H1N1 SIV replication as identified by an unbiased genome-wide CRISPR knockout screen in pig cells

To identify the host factors necessary for EA H1N1 SIV infection, we conducted a genome-wide CRISPR screen in Cas9-expressing porcine kidney-15 cells (PK-15-Cas9) following an established procedure (Fig. 1a)25. Both parental cells and cells expressing version 1.0 of the PigGeCKO single guide RNA (sgRNA) libraries were repeatedly infected with EA H1N1 SIV (A/Swine/HuBei/221/2016, HuB/H1N1) at a multiplicity of infection (MOI) of 0.01 and performed live/dead screens. Following four rounds of screening, we subjected the surviving cells from the second, third, and fourth rounds of challenge to high-throughput sequencing, then analyzed and ranked candidate genes using the model-based analysis of the genome-wide CRISPR/Cas9 knockout (MAGeCK) program26. Enrichment of 14 genes (RBM6, SLC35A1, GGCX, ADD1, LOC100516036, GPCPD1, CMAS, ST3GAL4, etc) was observed in the second, third, and fourth rounds (Supplementary Data 1 and Supplementary Fig. 1).

Fig. 1: Genome-wide CRISPR screen identifies GGCX as EA H1N1 SIV host-dependent factor in porcine kidney cells.
figure 1

a A schematic diagram of the genome-wide CRISPR screening process in PK-15 cells. b–e Effects of GGCX-KO on the replication of influenza virus strains namely, (b) HuB/H1N1, (c) HuN/H1N1, (d) PR8/H1N1, and (e) SH13/H9N2. GGCX-KO and WT PK-15 cells were infected with IAV strains at a MOI of 0.1, and the viral titers at indicated time points were determined by TCID50. SLC35A1 is an IAV host-dependent factor, and SLC35A1-KO cells were used as a positive control. f Restoration of GGCX promotes influenza virus replication. Exogenous GGCX (Flag-GGCX) and Flag negative control were transfected into GGCX-KO PK-15 cells or WT PK-15 cells. The transfected cells were infected with HuB/H1N1 at an MOI of 0.1, viral titers at 24 hpi were determined by TCID50, and corresponding protein expression was detected by western blot. The error bar in panels (b–f) indicates the standard deviation. The data shown in panels (b–f) are means ± SD (n  =  3 biologically independent experiments). Statistical analysis was performed using an unpaired, two-tailed Student’s t test. (ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001). The statistical analysis in panel (b) is between the GGCX-KO and Control. Exact P-values are available in the Source Data. Source data are provided as a Source Data file.

Full size image

To explore the role of host-mediated post-translational modifications (PTMs) in EA H1N1 SIV pathogenesis, an integral membrane protein, GGCX, was selected as it had been previously reported to catalyze the post-translational carboxylation of several proteins that convert specific peptide-bound glutamate (Glu) to γ-carboxyglutamate (Gla)27,28,29. We employed CRISPR/Cas9 technology to successfully knock out the endogenous GGCX gene in PK-15 cells shown by the lack of GGCX protein expression and genomic base deletion (Supplementary Fig. 2a, b). GGCX deficiency did not appear to affect cell viability (Supplementary Fig. 2c). Interestingly, when compared with a well-known gene, SLC35A1, that is involved in the synthesis of sialic acid receptors and has been identified in many genome CRISPR screens26,30,31, the ability of GGCX-knockout (KO) cells to inhibit HuB/H1N1 infectivity is the same as in SLC35A1-KO cells, suggesting that GGCX is key for efficient EA H1N1 SIV infection (Fig. 1b). We further evaluated the role of GGCX in the infection of other IAV strains and observed significantly reduced virus titers in GGCX-deficient cells infected with human-isolated EA H1N1 SIV A/Hunan/42443/2015 (HuN/H1N1), human-origin A/Puerto Rico/8/1934 (PR8/H1N1), or avian-origin A/chicken/Shanghai/SC197/2013 (SH13/H9N2) IAV strains (Fig. 1c–e). Furthermore, the inhibitory effect of GGCX-deficient cells was more pronounced for EA H1N1 SIV (HuB/H1N1) than for human-origin (HuN/H1N1 and PR8/H1N1) and avian-origin (SH13/H1N1) strains (Fig. 1b–e). We then evaluated whether rescue or overexpression of GGCX could restore or promote viral infection. Our results show that overexpression of porcine GGCX in GGCX-deficient or WT cells can restore or promote HuB/H1N1 infection (Fig. 1f). Though PK-15 cells were capable of being infected with multiple IAV strains, it was not the natural susceptible cells of IAVs. To simulate natural infection conditions and test the species specificity of GGCX, GGCX knockout cells were also generated in newborn pig trachea (NPTr) and human A549 cells (Supplementary Fig. 3a, b). It was revealed that knockout of GGCX also significantly decreased the replication of multiple IAV strains in NTPr and A549 cells (Supplementary Fig. 3c–h). Meanwhile, the role of GGCX in IAV replication in avian cells was also examined by using siRNAs, and GGCX was similarly required for IAV infection in duck embryonic fibroblasts (DEF) cells (Supplementary Fig. 3i). Thus, the results showed that GGCX promoted the infection of EA H1N1 SIV and other subtypes of IAV in vitro.

To further investigate the role of GGCX in EA H1N1 SIV replication in vivo, chemically cholesterol-conjugated and 2’-OME-modified siRNA targeting GGCX (si-GGCX) and negative control siRNA (si-NC) were nasally instilled into 6-week-old female BALB/c mice, and the mice were challenged with HuN/H1N1 (Fig. 2a). The knockdown efficiency of the chemically modified siRNA in mice was assessed (Fig. 2b). Mouse weight loss and survival were monitored daily for 14 days post-challenge. GGCX knockdown mice exhibited slightly attenuated infection as measured by reduced weight loss and increased survival rates compared to siRNA control mice (Fig. 2c, d). We also observed a significant decrease in viral titers in the lungs of GGCX knockdown mice compared to the control mice (Fig. 2e). Histopathological analysis of the infected control siRNA-treated mice’s lungs demonstrated moderate to severe bronchiolar necrosis, pulmonary edema, and inflammatory cell infiltration (Fig. 2f). Conversely, the examination of the infected GGCX knockdown mice’s lungs revealed a noteworthy reduction in the infiltration of lymphoid tissue compared to control mice (Fig. 2f). Meanwhile, weaker viral nucleoprotein (NP) antigen signals were detected in the lungs of GGCX knockdown mice compared to control mice (Fig. 2g). These results demonstrated that GGCX knockdown in the lungs of mice significantly contributed to a protective effect against EA H1N1 SIV challenge. Taken together, GGCX was identified as a host-dependent factor for EA H1N1 SIV infection by an unbiased genome-wide CRISPR-Cas9 loss-of-function screen.

Fig. 2: GGCX promoted the infection of EA H1N1 SIV in vivo.
figure 2

a Schematic of siRNA treatment and HuN/H1N1 strain challenge in an experimental mouse model. Mice were treated with siRNA the day before and after the viral challenge, and monitored for 14 days. b Western blot analysis of GGCX protein expression in GGCX-siRNA treated mice. c Weight loss of HuN/H1N1 strain infected mice after siRNA treatments. Mice with a body weight loss of more than 30% were euthanized according to the ethical principles of animal welfare. Each treatment group had ten mice (n  =  10) per group. d Mortality of HuN/H1N1 strain-infected mice after siRNA treatments. Each treatment group had ten mice (n  =  10) per group. e Virus titers in the lungs of infected mice 3 days (left) and 5 days (right) after infection. Each treatment group had three mice (n  =  3) per group. f Hematoxylin and eosin (H&E) staining of pathological lesions in the lungs of GGCX knockdown mice infected with HuN/H1N1 strain at 3 and 5 days post-challenge. Scale bars, 200 μm. g Immunofluorescence staining of lung sections from GGCX knockdown mice infected with HuN/H1N1 strain at 3 and 5 days post-challenge. The viral NP antigen was stained red, and the nucleus was stained blue. Scale bars, 200 μm. The images in panels (f and g) are representative of three independent experiments. The error bar in panels (c and e) indicates the standard deviation. The data shown in panels c and e are means ± SD (n = 10 for c and n  =  3 for e biologically independent experiments). Statistical analysis was performed using an unpaired, two-tailed Student’s t test. (**P < 0.01; ***P < 0.001). Exact P-values are available in the Source Data. Source data are provided as a Source Data file.

Full size image

GGCX catalyzes carboxylation modification of viral hemagglutinin

To explore the phase of the EA H1N1 SIV infection cycle in which GGCX was implicated, the cellular distributions of the viral nucleoprotein (NP) in infected WT and GGCX-KO cells were visualized. It was revealed that there were no obvious differences in viral NP expression and distribution until 9 hpi, and GGCX knockout resulted in reduced NP expression at 12 hpi (Supplementary Fig. 4a), indicating that GGCX was involved in the late stage of EA H1N1 SIV infection. As previous reports established GGCX as an integral membrane protein catalyzing the post-translational carboxylation of several vitamin K-dependent (VKD) proteins27,28,29, we hypothesized that GGCX participated in EA H1N1 SIV infection through post-translational carboxylation modification of viral proteins. To identify the viral protein targeted by GGCX-mediated carboxylation, we infected WT PK-15 cells with the HuB/H1N1 strain and conducted co-immunoprecipitation (co-IP) experiments using anti-Gla antibodies. The results confirmed that the carboxylation modification of viral HA protein by immunoprecipitating carboxylated proteins (Fig. 3a). To investigate the effect of GGCX knockout on the carboxylation modification of viral HA, GGCX-KO and WT cells were infected with HuB/H1N1, and carboxylation modified viral HA proteins were precipitated using anti-Gla antibodies and detected by western blot assay. The results indicated that GGCX depletion resulted in decreased carboxylation modification of viral HA (Fig. 3b). To further validate the interaction between GGCX and HA, exogenous Flag-GGCX and pSP72-HA were transfected into the PK-15 cells, and co-IP experiment using anti-Flag antibody confirmed the interaction between exogenous GGCX and viral HA proteins (Fig. 3c). Moreover, co-IP experiments using anti-HA (Fig. 3d) or anti-GGCX (Fig. 3e) antibodies with WT cells infected with HuB/H1N1 strain also revealed the interaction between endogenous GGCX and viral HA proteins.

Fig. 3: GGCX catalyzed the carboxylation modification of viral HA protein.
figure 3

a Viral HA protein was modified for carboxylation. WT cells were infected with HuB/H1N1 strain, and cell lysates were prepared at 9 hpi for co-IP assays using anti-Gla or control IgG antibody, followed by western blot analysis to detect the co-precipitated viral proteins. b Effects of GGCX knockout on the carboxylation modification of HuB/H1N1 HA protein. GGCX-KO and WT cells were infected with HuB/H1N1. At 9 hpi, cell lysates were prepared for co-IP assays using anti-Gla or control IgG antibodies, followed by western blot analysis to detect co-precipitated proteins. BGLAP was used as a positive control. c Interaction between exogenous GGCX and viral HA. WT PK-15 cells were transfected with Flag-GGCX and pSP72-HA, and cell lysates were prepared at 24 h post-transfection for co-IP assays using anti-Flag, followed by western blot analysis to detect co-immunoprecipitated proteins. d, e Interaction between endogenous GGCX and viral HA. WT cells were infected with HuB/H1N1 strain, and cell lysates were prepared at 24 hpi for co-IP assays using (d) anti-HA or (e) anti-GGCX with corresponding control IgG antibody, followed by western blot analysis to detect co-immunoprecipitated proteins. f Effects of GGCX-KO on different viral HA carboxylation modifications. GGCX-KO and WT cells were infected with three types of IAV strains. Cell lysates were prepared at 9 hpi for co-IP assays using anti-Gla or control IgG antibodies, followed by western blot analysis to detect co-precipitated proteins. BGLAP was used as a positive control. The images in (a–f) are representative of three independent experiments. Source data are provided as a Source Data file.

Full size image

Interestingly, GGCX-KO resulted in reduced virus replication, with varying inhibitory effects observed among different IAV strains (Fig. 1b–e). We investigated whether this phenomenon was associated with the degree of carboxylation modification of viral protein mediated by GGCX. GGCX-KO and WT cells were infected with HuB/H1N1, PR8/H1N1, and SH13/H9N2, respectively, and carboxylation-modified viral HA proteins were precipitated and detected by western blot assay. The results indicated the presence of carboxylated HA proteins in all three types of IAV-infected cells, with reduced expression in GGCX-KO cells, except in the SH13/H9N2-infected cells (Fig. 3f). Moreover, the carboxylation modification levels of HuB/H1N1 HA protein decreased more than that of PR8/H1N1, attributed to GGCX knockout (Fig. 3f). As the inhibitory effect of GGCX-deficient cells was more pronounced for EA H1N1 SIV (HuB/H1N1) than for human-origin PR8/H1N1 and avian-origin SH13/H9N2 (Fig. 1b–e), it was suggested that reducing viral infection was synchronous with the reduction of HA carboxylation modification for three IAV strains. These results collectively suggest that GGCX catalyzes the carboxylation modification of viral HA, potentially playing a critical role in EA H1N1 SIV infection.

Carboxylation modification of viral HA by GGCX promotes receptor binding activity of progeny virus to host cells

GGCX promotes EA H1N1 SIV infection and catalyzes the carboxylation modification of viral HA protein, which is involved in attachment to host cells. Therefore, carboxylation modification of HA may determine the receptor binding activity of progeny viral HA. To test our hypothesis, GGCX-KO and WT PK-15 cells were infected with HuB/H1N1 and treated with Oseltamivir (OSTV) or DMSO. It was revealed that there were no obvious differences in viral NP expression between virus-infected GGCX-KO and WT PK-15 cells in a single lifecycle of EA H1N1 SIV infection when treated with OSTV (Supplementary Fig. 4b–d). These results indicated that GGCX affected the next round of infection of EA H1N1 SIV. Therefore, progeny viruses propagated in WT cells (Gla virus) or carboxylation modification insufficient GGCX-KO cells (Glu virus) were quantified by absolute quantitative real-time PCR, then the equivalent progeny Gla and Glu viruses were used to respectively infect WT and GGCX-KO PK-15 cells and viral HA binding activities were visualized (Fig. 4a). We found that the progeny viruses from virus-infected WT and GGCX-KO cells showed similar quantities (Supplementary Fig. 4e). However, Glu viruses of HuB/H1N1 and PR8/H1N1 had a lower binding activity compared to the Gla viruses in both WT and GGCX-KO PK-15 cells (Fig. 4b, c). In contrast, the progeny Gla and Glu viruses of SH13/H9N2 showed identical binding activities (Fig. 4b, c). Viral binding activities determined by flow cytometry were consistent with those obtained by confocal microscopy (Fig. 4d). To determine whether human GGCX shared a similar role in progeny IAV binding, progeny Gla, and Glu viruses respectively propagated from IAV-infected WT and GGCX-KO A549 cells were used to respectively infect the WT and GGCX-KO A549 cells, and viral HA binding activities were visualized. It was found that the progeny Glu viruses of HuN/H1N1 and PR8/H1N1 also had a lower binding activity compared to the Gla viruses in both WT and GGCX-KO A549 cells (Supplementary Fig. 5). Taken together, we found HA binding activity was directly regulated by GGCX-mediated carboxylation modification.

Fig. 4: Hemagglutinin carboxylation modification by GGCX was required for progeny virus attachment in PK-15 cells.
figure 4

a Flowchart to determine the effects of GGCX-KO on progeny virus attachment. b–d Determination of binding activities of progeny Gla or Glu viruses. Equal amounts of progeny Gla or Glu viruses of HuB/H1N1, PR8/H1N1, and SH13/H9N2 were used to infect the WT and GGCX-KO PK-15 cells, respectively, followed by incubation with anti-influenza virus HA protein antibody. The HA proteins were then analyzed by (b) confocal microscopy (the red and blue fluorescence respectively indicated the HA protein and nucleus) and (c) average fluorescence intensity analysis (each group had n = 20 cells) as shown in (b) and (d) flow cytometry. Scale bars, 10 μm. e Sialic acid receptor binding preferences of progeny Gla or Glu virus. Equal amounts of progeny Gla or Glu viruses of HuB/H1N1, PR8/H1N1, and SH13/H9N2 were used to incubate with biotinylated sugar mimics of SA receptors, respectively, and ELISA assays were performed to determine receptor binding activities. The images in panels b and d are representative of three independent experiments. The error bar in panel (e) indicates the standard deviation. The data shown in panels (c and e) are means ± SD (n = 20 for c and n  =  3 for e biologically independent experiments). Statistical analysis was performed using an unpaired, two-tailed Student’s t test. (ns, P > 0.05; ****P < 0.0001). Exact P-values are available in the Source Data. Source data are provided as a Source Data file.

Full size image

To assess which type of SA receptor binding activity of HA was affected in GGCX-KO cells, ELISA-based assays were used to investigate the receptor binding profiles of progeny Gla and Glu viruses. We observed that the progeny HuB/H1N1 and PR8/H1N1 viruses preferentially bound to α-2,6 SA receptors, and progeny Glu viruses of HuB/H1N1 and PR8/H1N1 had lower binding activity than the Gla viruses (Fig. 4e). In contrast, progeny SH13/H9N2 viruses preferentially bound to α-2,3 SA receptors and were not affected by GGCX-KO (Fig. 4e). Taken together, knocking out GGCX revealed different inhibitory effects on viral binding to different SA receptor types, and the binding activity of HA to the α-2, 6 SA receptors was regulated by GGCX-mediated carboxylation modification, indicating the critical role of GGCX in determining the receptor binding preferences of EA H1N1 SIV.

GGCX-mediated carboxylation modification of viral HA 225E promotes its binding activity to α-2, 6 SA receptor

Since GGCX regulates the receptor binding activity of the progeny virus, we aimed to identify the carboxylation modification sites of viral HA catalyzed by GGCX. Thus, GGCX-KO and WT PK-15 cells were infected with the HuB/H1N1, and viral HA proteins were precipitated using anti-HA antibodies, followed by liquid chromatography-tandem mass spectrometry (LC-MS) analysis (Fig. 5a). The results revealed that, in addition to the 9 identical carboxylation modification sites (24, 106, 115, 118, 216, 246, 399, 408, and 427) identified in both HA proteins expressed in virus-infected GGCX-KO and WT cells, 13 distinct carboxylation modification sites (37, 97, 175, 225, 341, 387, 404, 433, 435, 450, 494, 501, and 502) were exclusively identified in HA proteins expressed in virus-infected WT cells (Supplementary Data 2). This suggests that these sites may play a crucial role in HA binding activity, which could be compromised by GGCX-KO (Fig. 5b). As GGCX-KO cells exhibited different effects on binding activity of progeny virus of EA H1N1 SIV (HuB/H1N1) and AIV (SH13/H9N2), the conservation of the 13 carboxylation sites was analyzed among the EA H1N1 SIV and different AIV subtypes. Six glutamic acid (E) carboxylation sites (97, 175, 225, 387, 450, and 494) were selected for further study, due to the glycine (G) or aspartic acid (D) substitutions at the corresponding carboxylation sites in AIV subtypes (Fig. 5c and Supplementary Fig. 6). Carboxylation site mutant and WT pseudoviruses were generated, and their binding activities were evaluated through a luciferase assay. The results indicated that the mutant HA E225A inhibited the binding activities of pseudoviruses (Fig. 5d), highlighting the critical role of the carboxylation modification of HA 225E in determining virus binding activities.

Fig. 5: GGCX-mediated Carboxylation modification of viral HA 225E promoted its binding activity to α-2, 6 SA receptors.
figure 5

a Flowchart showing the identification of carboxylation modification sites of viral HA by LC-MS. b Venn diagram of identified carboxylation modification sites in viral HA proteins expressed in infected GGCX KO and WT cells. c The Venn diagram analysis of the conservation of glutamic acid (E) at the carboxylation modification sites in EA H1N1 SIV and different AIV subtypes. d Identification of key carboxylation modification sites of viral HA. HEK293T cells were transfected with psPAX2, pLenti-luc, and WT or mutant viral HA plasmids to generate pseudoviruses. The pseudoviruses were then used to infect WT PK-15 cells and luciferase assays were performed. Protein expression levels were evaluated by western blot. e, f Binding activities of HA 225E and mutant viruses. Equal amounts of HA 225E, 225D, and 225 G viruses were each used to infect the WT PK-15 cells and incubated on ice for 1 h, followed by incubation with an anti-influenza virus HA protein antibody. Then, the HA proteins were analyzed by (e) confocal microscopy and (f) average fluorescence intensity analysis (each group had n = 20 cells) as shown in (e). The viral HA was stained red, and the nucleus was stained blue. Scale bars, 10 μm. g Sialic acid receptor binding preferences of HA 225E and mutant viruses. Equal amounts of HA 225E, 225D, and 225 G viruses were used to incubate biotinylated sugar mimics of sialic acid receptors, and ELISA assays were performed to determine the receptor binding activities. The images in panel e are representative of three independent experiments. The error bar in panels (d and g) indicates the standard deviation. The data shown in panels (d, f, and g) are means ± SD (n  =  3 for (d and g), and n = 20 for f biologically independent experiments). Statistical analysis was performed using an unpaired, two-tailed Student’s t test. (ns, P > 0.05; ****P < 0.0001). Exact P-values are available in the Source Data. Source data are provided as a Source Data file.

Full size image

To explore the impact of carboxylation modification of HA 225E on receptor binding activity of EA H1N1 SIV, we introduced HA E225D or E225G substitutions into the HuN/H1N1 virus, generating two corresponding mutant viruses. Equal amounts of mutant and WT viruses were then used to infect PK-15 cells, and the binding activities were compared through confocal microscopy. The results demonstrated that mutant viruses with HA 225D and 225 G exhibited lower binding activities than the WT HA 225E virus (Fig. 5e, f). In addition, we observed that the binding activities of the two mutant viruses to α-2,6 SA receptor were decreased compared to the WT viruses, while binding activities to α-2,3 SA receptor showed no differences between the different virus types (Fig. 5g). Collectively, these results indicate that GGCX-mediated 225E carboxylation modification can regulate the binding activities of viral HA to α-2,6 SA receptor.

GGCX mediates the HA 225E substitution of EA H1N1 SIV

The EA H1N1 SIVs originated from Eurasian avian H1N1, and the viral HA 225E site has been reported to be involved in HA receptor binding sites and determines the switch of EA H1N1 SIV receptor binding specificity from α-2,3 SA receptor to α-2,6 SA receptor32,33. When analyzing the conservation of HA 225E in H1N1 AIV and EA H1N1 SIV, we found that HA 225E was more conserved in EA H1N1 SIV than in H1N1 AIV (Fig. 6a). Furthermore, the conservatism of HA 225E in EA H1N1 SIV increased over time while that of HA 225 G decreased significantly (Fig. 6b). This suggests that HA 225E plays a key role in determining the evolutionary adaptation of EA H1N1 SIV.

Fig. 6: GGCX catalyzed the substitution adaption of EA H1N1 SIV HA 225E.
figure 6

a Conservation analysis of the HA 225E site in H1N1 AIV and EA H1N1 SIV. b Conservation change analysis of the EA H1N1 SIV HA 225E site over time. c Virus growth kinetics curves of HA 225E and mutant viruses in WT and GGCX-KO PK-15 cells. HA 225E, 225D, and 225 G viruses were used to infect WT and GGCX-KO PK-15 cells, respectively, at MOI = 0.1, and the virus titers at the indicated time points were determined by TCID50. Error bars in panel (c) indicate the standard deviation. The data shown in panel (c) is means ± SD (n  =  3 biologically independent experiments). Statistical analysis was performed using an unpaired, two-tailed Student’s t test. (*P < 0.05; **P < 0.01). The statistical analysis is between the 225E WT cell and 225D WT cells. Exact P-values are available in the Source Data. d–f EA H1N1 SIV passage and sequencing. HA 225E, 225D, and 225 G viruses were used to infect the (d) WT and (e) GGCX-KO PK-15 or (f) NPTr cells, respectively, at MOI = 0.1 for 48 h, the progeny viruses were collected and sequenced to infect the WT and GGCX-KO cells, respectively, and the viruses from each round in PK-15 cells and the fifth round of NPTr cells were subjected to be sequenced. g The EA H1N1 SIVs that naturally bear HA 225D or 225 G passage and sequencing. The naturally bearing HA 225D (WZ/H1N1) or 225 G (SD/H1N1) viruses were used to infect the WT and GGCX-KO PK-15 or NPTr cells, respectively, at MOI = 0.1 for 48 h. The progeny viruses were collected and sequenced to infect the WT and GGCX-KO cells, respectively, and the viruses from the fifth round were subjected to be sequenced. Source data are provided as a Source Data file.

Full size image

Since the GGCX-catalyzed carboxylation of HA 225E determines the receptor binding activity of EA H1N1 SIV to the α-2,6 SA receptor, GGCX might be responsible for the evolutionary adaptation of EA H1N1 SIV HA G225E. To further clarify the role of GGCX in the evolutionary adaptation of EA H1N1 SIV HA 225E, WT HA 225E, and mutant viruses were infected into GGCX-KO and WT PK-15 cells, respectively, and viral growth curves were documented. The result showed that the titers of the proliferating HA 225E virus from the WT PK-15 cells were significantly higher than those of the mutant virus. However, the titer of the virus proliferating from the GGCX-KO cells was almost the same as that of the mutant virus (Fig. 6c). Meanwhile, the progeny viruses infecting GGCX-KO and WT PK-15 cells were sequenced sequentially after serial passage. It was found that the HA 225D and HA 225 G mutant viruses proliferating from WT PK-15 cells gradually reverted to HA 225E, whereas the viruses proliferating from GGCX-KO PK-15 cells remained unchanged (Fig. 6d, e), suggesting that the substitution of HA 225E was catalyzed by GGCX. The progeny viruses from virus-infected GGCX-KO and WT NPTr cells were also sequenced sequentially after serial passage. The results revealed that the HA 225D and HA 225 G mutant viruses proliferated from WT NPTr cells can also revert to HA 225E in the fifth round, whereas the viruses proliferated from GGCX-KO NPTr cells remained unchanged (Fig. 6f). To further confirm the role of GGCX in the evolutionary adaptation of EA H1N1 SIV, the natural HA 225D (A/Wenzhou/1248/2019 (WZ/H1N1)) and 225 G (A/Shandong/0112/2021 (SD/H1N1)) bearing viruses were utilized to respectively infect the WT and GGCX-KO PK-15 or NPTr cells, and the progeny viruses were also sequenced sequentially after serial passage. It was revealed that the natural HA 225D and 225 G bearing viruses proliferated from WT PK-15 and NPTr cells can partially revert to HA 225E in the fifth round, whereas the viruses proliferated from GGCX-KO PK-15 and NPTr cells remained unchanged (Fig. 6g). Taken together, GGCX is a host factor required for HA 225E carboxylation modification and mediates the HA 225E substitution of EA H1N1 SIV.

Discussion

Over an extended evolutionary timeframe, EA H1N1 SIVs have gradually accumulated increased affinity for binding to α-2,6 SA receptor, facilitated in part by the D225E substitution in the viral HA. This substitution has the potential to contribute to the emergence of the next influenza pandemic. However, the precise factor driving this evolutionary substitution is poorly understood. Our investigation identified that GGCX promotes EA H1N1 SIV infection through a genome-scale CRISPR screen conducted in porcine kidney cells. The research uncovered that GGCX-mediated post-translational carboxylation modification of viral HA played a critical role in enabling progeny EA H1N1 SIV to bind to α-2, 6 SA receptor (Fig. 7). Moreover, GGCX emerged as the determinant for the D225E substitution, actively promoting the evolutionary adaptation of EA H1N1 SIV. Collectively, GGCX plays a pivotal role in catalyzing the carboxylation modification of viral HA 225E, thereby fostering the receptor binding preferences of EA H1N1 SIV to the α-2, 6 SA receptor and promoting the evolutionary adaption of EA H1N1 SIV.

Fig. 7: Working model for the regulation of GGCX in the receptor binding activity of progeny EA H1N1 SIV.
figure 7

In wild-type (WT) cells expressing γ-glutamyl carboxylase (GGCX), the haemagglutinin (HA) protein of Eurasian avian-like H1N1 swine influenza virus carrying the 225E amino acid undergoes carboxylation modification (Gla) by GGCX in the endoplasmic reticulum. The newly synthesized and modified viral HA proteins then assemble into progeny viruses (Gla virus) and bind to the α-2,6 SA receptor on the cell membrane, triggering the next round of infection. In contrast, GGCX knockout (KO) cells exhibit defective carboxylation modification, leading to the production of progeny viruses with insufficient carboxylation modification (Glu viruses), which reduces the efficiency of progeny Glu viruses to bind to the α-2,6 SA receptor on the cell membrane, thereby preventing subsequent infections.

Full size image

In contrast to previous genome-scale CRISPR screens primarily focused on the isolated human or avian influenza virus in their hosts, our study initiated a genome-scale CRISPR screen in porcine kidney cells to identify the host-dependency factors necessary for EA H1N1 SIV infection26,30,31,34,35,36,37,38,39,40. Shared hits with other studies highlighted the enrichment of host factors involved in SA receptor biosynthesis and related glycosylation pathways, including SLC35A1, CMAS, ST3GAL4, and ALG5, emphasizing the critical role of SA receptor biosynthesis in multiple IAV strains30,41. In addition, unique genes specific to the EA H1N1 SIV strain (HuB/H1N1) and its corresponding host were identified, including the GGCX, indicating their specific involvement in EA H1N1 SIV infection. GGCX was identified as essential for EA H1N1 SIV infection and was confirmed to be responsible for the HA 225E substitution through its previously unreported carboxylation modification function on the viral HA. Similar to other post-translational modifications such as glycosylation, methylation, and acetylation42,43,44, carboxylation modification of viral HA can alter the structure and functions of the target HA protein, influencing the formation of the salt bridge structure between amino acids D225/E225 and K222, as well as virion assembly and budding efficiency18,22. This may explain the switch in the SA receptor binding preferences of the EA H1N1 SIV from α-2,3 SA to the α-2,6 SA receptor. In addition, the protein structures of the viral HA protein, with and without carboxylation modification, would be analyzed to clarify the mechanism by which HA carboxylation modification alters the SA receptor binding preference and determine the evolutionary adaptation of EA H1N1 SIV.

Our study showed that the carboxylation modification insufficient progeny Glu virus had lower α-2,6 SA receptor binding activities than the Gla virus. Carboxylated modification of HA 225E was identified as a key regulator of progeny virus binding activities. However, the decreased binding activity observed in the 225E mutant was not sufficient to compensate for the KO effect of GGCX, suggesting the presence of other unidentified carboxylation modification sites. GGCX plays a critical role in the vitamin K cycle, which facilitates the γ-carboxylation and recycling of VK via GGCX and vitamin K epoxide reductase (VKOR), respectively27,28,29. Therefore, carboxylation modification of viral HA may undergo dynamic cyclic changes, with carboxylation modification sites varying at different stages of viral HA function. To investigate additional functional carboxylation modification sites of HA, HA proteins expressed in infected cells at different post-infection time points are collected and subjected to LC-MS analysis. In addition, potential carboxylation modification sites located in the receptor binding domain of HA will be validated by structural analysis. Overall, the identification of functional carboxylation modification sites in HA highlights the importance of viruses with substitutions at these sites, which may have a higher propensity to bind to α-2,6 SA receptor and pose a potential threat to human health.

GGCX is required for infection of several IAV strains, including EA H1N1 SIV and AIV. Interestingly, carboxylation modification of the AIV (SH13) HA protein and progeny virus binding activities were not affected by GGCX-KO, suggesting that GGCX-regulated IAV infection involves alternative mechanisms. This observation suggests that carboxylation modification of other Gla proteins catalyzed by GGCX may also contribute to IAV replication. In addition to the HA protein, the viral M1 and NP proteins were also found to undergo carboxylation modification with potential regulation by GGCX, indicating a universal effect on IAV replication45,46. Furthermore, host proteins that undergo carboxylation modification, such as coagulation factors, osteocalcin, and matrix Gla proteins, have been implicated in viral protein cleavage, autophagy, and immune pathways47,48,49, providing an alternative explanation for the shown up of GGCX in the screen and the inhibition of IAV replication. It was found that knockout of GGCX exhibited limited effects on immune responses (Supplementary Fig. 7), as well as illustrated by the results that there was no discernible effect on the initial round of viral infection (Supplementary Fig. 4). Our data suggest that GGCX plays a dominant role as an underlying driver of the HA D225E substitution in EA H1N1 SIV, catalyzing the carboxylation modification of viral HA 225E and ultimately determining the receptor binding preference of the virus. In conclusion, our study highlights the multifaceted role of GGCX in the regulation of IAV replication and suggests that it may serve as a potential target for the development of IAV therapeutics.

In conclusion, we used CRISPR/Cas9-based high-throughput loss-of-function screening to identify cellular factors involved in EA H1N1 SIV infection. Next, we found that GGCX catalyzes the carboxylation modification of the viral HA protein. This modification is essential to promote the binding capacity of progeny EA H1N1 SIV to α-2, 6 SA receptor. In particular, the GGCX-catalyzed carboxylation modification was found to be responsible for the substitution of HA 225E during the evolutionary adaptation of EA H1N1 SIV. These findings provide valuable insights into the mechanisms underlying the receptor binding preferences adaptation and the potential evolutionary adaptation of EA H1N1 SIV. They have important implications for the development of interventions aimed at disrupting transmission pathways and mitigating the risk of future influenza pandemics.

Methods

Ethics statement

Approval for the animal experiments carried out in this study was obtained from the Committee on the Ethics of Animal Experiments at Huazhong Agricultural University (No. HZAUMO-2023-0286).

Cells

Human embryonic kidney 293 T cells (HEK293T, Cat# CRL-3216), Madin-Darby canine kidney (MDCK, Cat# CCL-34) cells, Porcine Kidney-15 (PK-15, Cat# CCL-33) cells, Adenocarcinoma human alveolar basal epithelial cells (A549, Cat# CCL-185), and Duck embryonic fibroblasts (DEF, Cat# CCL-141) were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Newborn pig trachea (NPTr) cells were preserved in our laboratory. Stably expressing Cas9 PK-15 (PK-15-Cas9), NPTr (NPTr-Cas9), and A549 (A549-Cas9) cells were established through puromycin or blasticidin screening. All cell lines were cultured at 37 °C in a 5% CO2 humidified atmosphere using RPMI 1640 (SH30809.01, HyClone, USA), Minimum Essential Medium (MEM) (SH30265.01, HyClone, USA), or Dulbecco’s modified Eagle’s medium (DMEM) (SH30243.01, HyClone, USA) supplemented with 10% fetal bovine serum (FBS) (FSP500, ExCell, China).

Viruses and reverse genetics

The IAVs used in this study were A/Swine/HuBei/221/2016 (HuB/H1N1), A/Wenzhou/1248/2019 (WZ/H1N1), A/Shandong/0112/2021 (SD/H1N1), A/Puerto Rico/8/1934 (PR8/H1N1), A/chicken/Shanghai/SC197/2013 (SH13/H9N2), A/duck/Sheyang/1/2005 (YS/H5N1), and A/Hunan/42443/2015 (HuN/H1N1). A/Wenzhou/1248/2019 (WZ/H1N1) and A/Shandong/0112/2021 (SD/H1N1) are gifted from Honglei Sun (China Agricultural University). Recombinant viruses were generated in the genetic background of A/Hunan/42443/2015 (HuN/H1N1) using an eight-plasmid reverse genetic system50. All other viruses were amplified using 10-day-old embryonic chicken eggs and titrated by determining TCID50 values on MDCK cells. All experiments with A/chicken/Shanghai/SC197/2013 (SH13/H9N2) and A/duck/Sheyang/1/2005 (YS/H5N1) viruses were performed in an animal biosafety level 3 laboratory at Huazhong Agricultural University.

Plasmids

Lentiviruses were produced using the lenti-sgRNA-EGFP and lenti-guide-puro vectors, along with the pMD2.G and psPAX2 plasmids. Pseudoviruses were generated using the pLenti-luc, a generous contribution from Dr Rui Luo of Huazhong Agricultural University. For the construction of the lentiviral sgRNA vector, paired sgRNA oligonucleotides (50 μM per oligo) were annealed and cloned into lenti-sgRNA-EGFP or lenti-guide-puro vector, which was linearized with BbsI or BsmBI (R3539 and R0739, NEB, USA). The p3xFlag-GGCX (Flag-GGCX) and pSP72-HA were respectively constructed by cloning the full-length cDNA into the p3×Flag (Flag) or pSP72 vector, digested with HindIII/XbaI or PmeI/MluI, respectively. Eight segments of A/Hunan/42443/2015 (HuN/H1N1) were inserted into the pHW2000 vector, and mutant HA genes targeting amid acid 225 were generated by PCR-based site-directed mutagenesis, confirmed by sequencing.

Antibodies and reagents

The antibodies and reagents used in the study were as follows: Rabbit anti-GGCX (16209-1-AP, Proteintech, China); mouse anti-Gla (3570, Biomedica, Canada), mouse anti-Flag tag (F1804, Sigma-Aldrich, USA); rabbit anti-BGLAP (A6205, ABclonal, China); mouse anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (CB100127, California Bioscience, USA); rabbit anti-IAV NP and HA (GTX125989 and GTX127357, GeneTex, USA); horseradish peroxidase-conjugated anti-mouse and anti-rabbit (BF03001 and BF03008, Beijing Biodragon Immunotechnologies, China); goat anti-Cy3 anti-rabbit IgG (H + L) (AS007, ABclonal, China), and 4′,6′-diamidino-2-phenylindole (1:5000) (C1002, Beyotime, China).

Genome-scale CRISPR screening in pig

To conduct CRISPR screening, approximately 6 × 107 genome-scale PK-15 mutant cell libraries underwent infection with A/Swine/HuBei/221/2016 (HuB/H1N1) at an MOI of 0.01 in DMEM devoid of FBS. The cells were then incubated at 37 °C and 5% CO2. Following a 1.5 h incubation, the initial inoculum was replaced with fresh DMEM supplemented with 2.5% BSA (Cat# A4161, Sigma-Aldrich), 0.25 μg/mL TPCK (Cat# 4370285, Sigma-Aldrich), and 1% penicillin-streptomycin (Cat# P4333, Sigma-Aldrich). Surviving cells were collected 10 days post-infection and expanded for subsequent rounds of infection. In the subsequent rounds of infections, the survived cells from the last round of infection were repeatedly infected with the WT HuB/H1N1 at a MOI of 0.1. After four rounds of screening, high throughput sequencing was applied to the surviving cells from the second, third, and fourth rounds of challenge, followed by the analysis of candidate genes.

Generation of GGCX knockout cell line using CRISPR/Cas9

Individual sgRNA constructs targeting GGCX were generated and incorporated into the lenti-sgRNA-EGFP or lenti-guide-puro vector. Lentiviruses were produced following established protocols51. These lentiviruses were then transduced into PK-15-Cas9, NPTr-Cas9, or A549-Cas9 cells, respectively. Transduced cells were selected using fluorescence-activated cell sorting (FACS) or 1.5 μg/mL puromycin. Monoclonal cells were obtained through the limiting dilution method and subsequently expanded. Confirmation of GGCX-KO cells was achieved through Sanger sequencing and western blot analysis.

Cell viability assay

To examine the impact of GGCX-KO on cellular proliferation, the viability of GGCX-KO cells and WT cells was measured through CCK-8 activity, following the manufacturer’s instructions51. Briefly, cells were seeded onto 96-well plates, and their viabilities were measured at 12-, 24-, and 36 h post-seeding. CCK-8 reagent (CK04-500T, Dojindo Molecular Technologies, Japan) was applied to each well. and the subsequent measurement of absorbance at 450 nm was conducted using a microplate reader after a 1-hour incubation at 37 °C in the dark.

Transfection

Transfections were performed using Lipofectamine 2000 (11668019, Invitrogen, USA) according to the manufacturer’s instructions. Briefly, plasmids or siRNAs and Lipofectamine were diluted to equal volumes with Opti-MEM and incubated for 5 min at room temperature. The diluted Lipofectamine and plasmids or siRNAs were mixed and incubated for 20 min at room temperature. The mixture was added to cells and incubated for 6 h, and the medium was then replaced with fresh medium supplemented with 10% FBS.

Virus infection and titration

To evaluate the impact of GGCX-KO on viral replication, negative control, and GGCX-KO cells were independently seeded in triplicate within 12-well plates. For influenza A virus (IAV) infection, cells underwent two times washes with DMEM, followed by incubation with diluted virus at the MOI of 0.01 or 0.1 for 1 h. Subsequently, cells were again washed twice with DMEM and replenished with fresh infection medium (DMEM supplemented with 0.2 μg/mL TPCK-treated trypsin (T1426, Sigma-Aldrich, USA)). Supernatants were collected at designated time points post-infection, and viral supernatants were serially diluted with DMEM. Eight replicates of each dilution were added to the wells, and the 50% tissue culture infectious dose (TCID50) was calculated using the Reed-Muench method 72 h after infection52.

Mouse models for GGCX knockdown

For an in-depth exploration of GGCX’s role in EA H1N1 SIV infection in vivo, we synthesized cholesterol-conjugated and 2’-OME-modified si-GGCX or si-NC (GenePharma, China) and administered them nasally to 6-week-old BALB/c female SPF mice on days 0 and 239. Subsequently, the mice were either challenged with 30 pfu of HuN/H1N1 or mock-infected on day 1. Daily monitoring of body weight loss and survival occurred over 2 weeks post-infection (n = 10). Mice exceeding a 30% loss in initial body weight were humanely euthanized. On 3 and 5 days post-challenge, a subset of mice from each group (n = 3) underwent anesthesia and sacrifice and their lungs were either homogenized and/or fixed in 4% formaldehyde. The homogenized lung samples were utilized for assessing gene expression as well as virus titers. The fixed mouse lung samples were used for hematoxylin & eosin (H&E) and immunofluorescence staining for histopathological analysis.

Western blot and immunoprecipitation

For immunoprecipitation, GGCX-KO and WT PK-15 cells were infected with the IAV strains at 0.01 MOI for 9 or 24 h. Cells were washed with cold phosphate-buffered saline (PBS) and lysed with NP-40 lysis buffer (P0013F, Beyotime, China) containing protease inhibitor cocktail (04693132001, Roche, Switzerland). Cell lysates were incubated overnight at 4 °C with Dynabeads (Sc-2003, Santa Cruz Biotechnology, USA) conjugated with antibodies against either the γ-carboxyglutamyl (Gla) residues or control IgG antibodies. Protein-antibody-Dynabeads complexes were washed three times with NP-40 lysis buffer and analyzed by western blot.

Indirect immunofluorescence assay and confocal microscopy

WT and GGCX-KO PK-15 cells were incubated with HuB/H1N1 at an MOI of 0.1 for 1 h. The cells were fixed with 4% paraformaldehyde (PFA) for 10 min at indicated time points, followed by incubating with 1% (wt/vol) bovine serum albumin (BSA) for 1 h at room temperature. Samples were then incubated with the anti-NP antibody (GTX125989, GeneTex, USA) for 2 h at room temperature, followed by incubation with the Cy3-conjugated secondary antibody for 1 h. The nuclei were stained with DAPI for 10 min at room temperature. Images were acquired using a confocal microscope (LSM880, Zeiss, Germany).

Absolute quantitative real-time PCR

Viral RNA was extracted from cell suspensions using TRIzol Regent (15596018, Invitrogen, USA) according to the manufacturer’s protocol. Extracted viral RNAs were used as a template to generate cDNA using reverse transcriptase (RK20403, ABclonal, China). Quantitative real-time PCR (qRT-PCR) (ABI Vii7A, USA) was performed using SYBR GREEN (RK21203, ABclonal, China). The constructed plasmid expressing the full length of viral NP (pcDNA3.1-NP) was used as a standard to generate a standard curve. The amount of viral RNA was calculated according to the formula provided by the standard curve53.

Receptor binding activity assay

Serial dilutions of Neu5Acα2-3Galβ1-4Glcβ-sp4-PAA-biot (0060-BP, GlycoNZ, New Zealand) and Neu5Acα2-6Galβ1-4GlcNAcβ-sp3 (0997-BP, GlycoNZ, New Zealand) were applied to pre-streptavidin-coated high-capacity plates (15,500, Thermo Scientific, USA) and incubated at 4 °C for overnight54. Subsequently, the plates were washed three times with PBS, followed by incubation with 2% PBSA for 1 h at room temperature and three additional PBS washes. Diluted influenza virus was then added to the plates and allowed to incubate at 4 °C overnight. After five washes with PBST, the plates were incubated with chicken anti-influenza virus serum for 4 hours at 4 °C, washed with PBST, and incubated with HRP rabbit anti-chicken (IgG) (H + L) (AS030, ABclonal, USA). TMB substrate (CW0050S, CWBIO, China) was added to the plates to react for 20 min at room temperature in the dark and stopped with 0.5 M H2SO4. OD values were recorded at 450 nm wavelength using a multimode reader (EnVision).

HA binding assays

Cell surface binding of HA was performed with quantified progeny Gla and Glu virus. Briefly, for PK-15 cells, WT and GGCX-KO PK-15 cells were respectively incubated with progeny Gla and Glu virus of HuB/H1N1, PR8/H1N1, and SH13/H9N2 at an MOI of 5 for 1 h on ice. For A549 cells, WT and GGCX-KO A549 cells were incubated with progeny Gla and Glu virus of HuN/H1N1 and PR8/H1N1 at an MOI of 5 for 1 h ice. The unbound virus was washed with PBS. Cells were then fixed with 4% formaldehyde for 10 min and incubated with 1% (wt/vol) bovine serum albumin (BSA) for 1 h at room temperature. The amount of bound viral HA was measured by cell surface staining for viral HA with a polyclonal anti-HA antibody followed by a second antibody conjugated to anti-Cy3 goat anti-rabbit IgG (H + L) (AS007, ABclonal, China) and observed by confocal microscopy analysis.

Flow cytometry

For PK-15 cells, WT and GGCX-KO PK-15 cells were infected with titrated progeny Gla and Glu virus of HuB/H1N1, PR8/H1N1, and SH13/H9N2 at an MOI of 5 and incubated on ice for 1 h. For A549 cells, WT and GGCX-KO A549 cells were incubated with progeny Gla and Glu virus of HuN/H1N1 and PR8/H1N1 at an MOI of 5 and incubated on ice for 1 h. The unbound virus was washed with PBS. Subsequently, the cells were collected, fixed with 4% paraformaldehyde for 10 minutes, and washed three times with PBS. After a 2-hour incubation with 1% PBSA, the cells were incubated with anti-influenza virus HA (GTX127357, GeneTex, USA) for an additional 2 h. Staining of the cells was achieved using Cy3-conjugated goat anti-rabbit IgG (H + L) (5230–0359, SeraCare, USA), and positive cells were analyzed through Cytoflex-LX.

Liquid chromatography-tandem mass spectrometry

GGCX-KO and WT cells were infected with the HuB/H1N1 strain virus at MOI of 1 for 9 h. Subsequently, cells were washed with cold PBS and lysed using NP-40 lysis buffer (P0013F, Beyotime, China) supplemented with a protease inhibitor cocktail (04693132001, Roche, Switzerland). The resulting cell lysates were each incubated overnight at 4 °C with Dynabeads (Sc-2003, Santa Cruz Biotechnology, USA) pre-conjugated with antibodies against the viral HA protein (GTX127357, GeneTex, USA). Complexes of protein-antibody-Dynabeads were washed three times with NP-40 lysis buffer and then subjected to SDS-PAGE. The gel bands corresponding to the molecular weight of viral HA were collected and digested by chymotrypsin and Trypsin & ASP-N, subsequently the digested proteins were analyzed by LC-MS (Biotechpack, China). The glutamic acid (E) sites which added a corresponding carboxyl molecular weight (44 Da) were identified as the carboxylation modification sites.

Pseudoviruses packaged and assayed for luciferase activity

Pseudoviruses were produced by co-transfecting HEK293T cells with psPAX2, pLenti-luc and plasmids encoding either the WT or the carboxylation site mutant (97, 175, 225, 387, 450 and 494 E/A) viral HA, using Lipofectamine™ 2000 (11668019, Invitrogen, USA)55. Supernatants were harvested at 60- and 72 h post-transfection and filtered through a 0.45 μm filter. Wild-type PK-15 cells were seeded (cell density of 20%) in 12-well plates and incubated in 1 mL media containing pseudoviruses for transduction. Following a 1 h incubation on ice, the transduced cells were replenished with fresh media lysed, and the luciferase activities of the pseudoviruses were gauged using a luciferase assay system (E1501, Promega, USA).

Virus passage and sequencing

GGCX-KO and WT PK-15 and NPTr cells were respectively infected with HA 225E, 225D, and 225 G HuN/H1N1 viruses or WZ/H1N1 and SD/H1N1 viruses for 48 h. The supernatants containing the progeny viruses were collected, and these viruses were subjected to sequential rounds of infection on GGCX-KO and WT cells for an additional four cycles each. Viral RNAs obtained from these five rounds were extracted using TRIzol Regent (15596018, Invitrogen, USA). The extracted viral RNAs served as templates for cDNA synthesis using reverse transcriptase (RK20403, ABclonal, China). Viral HA segments were PCR-amplified and cloned into pMD-18T vector (6011, TAKARA Beijing, China) or pEASY®-Blunt vector (CB101-01, TransGen, China), respectively. Subsequently, 10 bacterial colonies were randomly selected for sequencing (Tsingke, China), and the proportions of HA amino acids 225E, 225D, and 225 G were analyzed.

Statistical analysis

All measurements were taken in triplicate, and the presented data are outcomes from at least two separate experiments. The results are shown as the mean ± standard deviation of the triplicate determinations. Statistical significance was ascertained by computing P-values using the paired two-tailed Student’s t test (ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.