Introduction

Herpes simplex virus type 1 (HSV-1), a ubiquitous pathogen, causes various diseases in humans, ranging from the common mucocutaneous lesions to the severe life-threatening encephalitis1. Once infected, HSV-1 establishes latent infections in peripheral neurons and occasionally reactivates to cause recurrent lesions2. HSV-1 affects a broad range of hosts3 and exclusively induces cell death via its cytopathic effect in vitro. It employs two modes to transmit from virus-producing cells to uninfected cells. Cell-to-cell infection occurs in the reactivation process. In this mode, viral particles transmit through a direct interaction between infected and neighboring uninfected cells4. In cell-free transmission, HSV-1 spreads to new hosts and often between cells within the same host, after which extracellular viral particles released from virus-producing cells enter the target cells. The attachment of the virus particle to the cell surface is the initial step in the cell-free infection of HSV-1. Evidence has demonstrated that HSV-1 glycoproteins (gC and/or gB) on the viral envelope interact with heparan sulfate (HepS) on the cell surface for attachment5. Subsequently, attachment to HepS enhances the interaction between gD and one of its receptors (i.e., nectin-1, HVEM, or 3-O-sulfated HepS)6,7,8, inducing a conformational change in gD, and recruiting gB, gH and gL9,10. This tetrameric complex of glycoproteins mediates viral entry into the cells via receptor-mediated fusion between the viral envelope and the host cell11,12. Specifically, four viral glycoproteins (gB, gD, gH, and gL) are required and considered sufficient for viral entry into host cells13. However, the host factors involved in this process have not been fully understood.

HepS is a linear polysaccharide ubiquitously expressed on the cell surface and in the extracellular matrix. HepS chains covalently bound to proteins are known as HepS proteoglycans (HSPGs)14. These chains are heavily modified via sulfation at various positions on their sugar residues, imparting an overall high negative charge and creating binding sites for different molecules. Physiologically, HepS plays a vital role in regulating various cellular functions, such as cell growth, adhesion, angiogenesis, and blood coagulation, through interactions between HSPGs and their partner molecules15. Furthermore, HepS is utilized by various viruses for efficient infection, owing to its molecular diversity. In addition to HSV-1, studies have identified HepS as a factor that facilitates the binding and cell entry of numerous viruses, including HSV-216, human coronavirus (HCoV) OC4317, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)17,18, Ebola virus19, human immunodeficiency virus (HIV)20, and hepatitis B virus (HBV)21. However, the factors involved in regulating HepS biosynthesis are currently not completely understood.

Therefore, in this study, we performed a genome-wide CRISPR screen for HSV-1 host factors and highlighted the importance of HepS in HSV-1 infection. As a result, we identified 3′-phosphoadenosine 5′-phosphosulfate synthase 1 (PAPSS1) as a critical factor for HSV-1 infection. The genetic ablation of PAPSS1 was sufficient to abolish HepS expression in HAP1 cells and consequently reduce the binding of various pathogenic viruses. Moreover, while the single KO of PAPSS1 slightly affected HepS biosynthesis in some PAPSS2-expressing cell lines, the double KO of PAPSS1 and PAPSS2 resulted in reduced HepS expression and higher resistance against HSV-1 infection in human retinal pigment epithelial-1 (RPE-1) cells, indicating a redundant role of PAPSS1 and PAPSS2 in HepS biosynthesis.

Results

A genome-wide CRISPR screen identifies host factors for HSV-1 infection

To identify the genes required for HSV-1 infection, we performed a genome-wide CRISPR screen using a human near-haploid cell line (HAP1) (Fig. 1a). We used haploid cells because of their higher efficiency in generating loss-of-functional mutations relative to diploid cells22. First, we established HAP1 cells that stably expressed Cas9 (HAP1/Cas9). Then, we transduced these cells using a pooled lentivirus, encoding single-guide RNAs (sgRNAs) and targeting 19,052 genes (6 sgRNA/gene), 1864 microRNA (miRNA) (4 sgRNA/miRNA), and 1000 nontargeting control sgRNAs at a multiplicity of infection (MOI) of 0.3. The transduction step ensured that each cell generally expressed only one sgRNA. After the mutagenesis by lentiviral transduction and antibiotic selection for a week, we infected 20 million cells with HSV-1 at an MOI of 0.1 and selected the surviving cells. Subsequently, the survivor cells were subjected to a second challenge with HSV-1 at two weeks after the first HSV-1 challenge, to enable the enrichment of cells resistant to HSV-1 infection. The cells were expanded further for 2 weeks, and the genomic DNA of the survivor cells was extracted and subjected to sequencing. As shown in Fig. 1b, sgRNAs in the virus-infected group were selected from <3% of the total sgRNAs in the duplicate screening step, and then each gene was ranked as either the most enriched sgRNA or the second-most enriched sgRNA23 (Fig. 1c). The known HSV-1 entry receptor NECTIN-1, which interacts with HSV-1 glycoprotein D6, was ranked at the top in our enrichment analysis, demonstrating the technical quality of the screening process. Additionally, seven genes in the HepS biosynthesis pathway (XYLT2, B4GALT7, B3GAT3, EXTL3, EXT1, EXT2, and SLC35B2) were markedly enriched (Fig. 1c). While these six former genes encode enzymes that catalyze the HepS backbone formation14 (Fig. 1d), SLC35B2 serves as the transporter of adenosine 3′-phospho 5′-phosphosulfate (PAPS), the universal sulfuryl donor for sulfation. PAPS was previously reported to be involved in HepS biosynthesis24,25. We focused on the genes supported by ≥2 sgRNAs with >150 RPM for further validations since a sgRNA targeting EXT2 showed the lowest abundance (150.5 RPM) among these HepS-related genes (Fig. 1c; Supplementary Table 1). In addition to NECTIN1 and HepS-related genes, our screen identified three candidate genes (i.e., IRF2BPL, PAPSS1, and VANGL2) and a candidate miRNA (MIR4647) (Fig. 1c).

Fig. 1: Genome-wide CRISPR screen to identify genes involved in HSV-1 infection.
figure 1

a Schematic representation of the CRISPR-based KO screening process. HAP1/Cas9 cells were mutagenized with a pooled lentiviral single-guide RNAs (sgRNA) library containing 121,449 unique sgRNAs targeting 19,052 genes and 1864 miRNAs. Mutant cells were infected with HSV-1, following which the viral infection-resistant cells were harvested. The control groups were mock-infected with PBS and cultured for the same period. The abundance of each sgRNA in the control and selected groups was determined through next-generation sequencing. This screening was performed in duplicate at an estimated coverage of 150 cells/sgRNA. b Abundance of sgRNAs in the mock-infected control and HSV-1-infected groups. c Scatter plot showing the most enriched sgRNA (log-transformed RPM: X-axis) versus the second-most enriched sgRNA (log-transformed RPM: Y-axis) for each gene following selection by HSV-1 infection. The size of the plots indicates the number of unique sgRNA detected with >150 RPM. Genes supported by ≥2sgRNAs with >150 RPM (red lines) were selected for further validation. d Overview of HepS biosynthesis. Genes supported by two or more sgRNA with over 150 RPM on our screen are highlighted in red font. e Validation of candidate genes. One or two KO HAP1 clones of the indicated genes were obtained by the CRISPR-Cas9 system. The wildtype (WT), nontargeting control (Ctrl), and KO clones were infected with HSV-1at a MOI of 3, and the cell viabilities were measured at 48 hpi via the MTS assay. The results are presented as means ± SEM of three independent experiments. Asterisks, p < 0.05; Double asterisks, p < 0.01; n.s. not significant.

Full size image

We performed validation experiments for NECTIN1, several HepS-related genes (including XYLT2 and EXT2), and three additional candidate genes. Among the HepS-related genes, we selected XYLT2 and EXT2 for validation since XYLT2 initiates the biosynthesis of glycosaminoglycan chains26, and EXT2 is involved in the later step of chain formation27 (Fig. 1d). Then, we generated one or two HAP1 knockout (KO) clones for each gene via the CRISPR-Cas9 approach using several sgRNAs (Supplementary Fig. 1a). Later, these clones were infected with HSV-1, and their viabilities were measured at 48 h post-infection (hpi). Results showed that the cytopathic effects of HSV-1 infection were significantly suppressed in the KO clones of NECTIN1, XYLT2, EXT2, and PAPSS1 compared with the wild type (WT) and nontargeting control cells (Fig. 1e). We also confirmed this validation through HSV-1 infection at a different MOI (Supplementary Fig. 1b). We observed that the perturbation of MIR4647 by a miRNA inhibitor did not suppress the cytopathic effects of HSV-1 infection (Supplementary Fig. 2). Of note, the genomic locus of MIR4647 is overlapped in the 3′-UTR of SLC35B228, suggesting the possibility that sgRNAs targeting MIR4647 reduce the expression of SLC35B2.

Therefore, our screening process collectively identified NECTIN1, PAPSS1, and a series of genes in the HepS biosynthesis pathway as important for HSV-1 infection.

XYLT2 and EXT2 enhance the binding of HSV-1 to the cell surface through HepS expression

Next, to investigate the role of genes that catalyze the sugar chain formation of HepS in HSV-1 infection, two KO clones of XYLT2 and EXT2 each and their rescued clones were obtained using HAP1 cells (Fig. 2a). Fluorescence-activated cell sorting (FACS) analysis with an anti-HepS antibody revealed that HepS expression was downregulated to background levels in both the XYLT2- and EXT2-KO clones and the same extent in the heparinase-treated cells (Figs. 2b, c). Furthermore, treatment with an antagonist to HepS, Surfen, inhibited HSV-1-infection-mediated cell death (Fig. 2d). Although the XYLT2- and EXT2-KO clones demonstrated higher survivability against HSV-1 infection than the control clone at 48 hpi (Fig. 2e), the complementation of the lost gene canceled the resistance (Fig. 2e). Consistent with these findings, progeny virus production was also reduced in the KO clones by 1–2 log fold compared with the control clone (Fig. 2f). Subsequently, since HepS involves the binding of HSV-1 on the cell surface12, we performed a binding assay using XYLT2- and EXT2-KO HAP1 cells. The levels of viral DNA on the cell surface were dramatically decreased by 30–200 fold in both XYLT2- and EXT2-KO clones (Fig. 2g). To further investigate the roles of these genes in HSV-1 infection using epithelial cells, which are HSV-1 natural targets, we obtained XYLT2- and EXT2-KO clones using RPE-1 cells (Supplementary Fig. 3a). Similar to HAP1 cells, XYLT2- and EXT2-KO RPE-1 cells showed reduced HepS expression and higher resistance against HSV-1 infections than the control clone (Fig. 2h and Supplementary Fig. 3b, c). These findings confirmed that HepS enhances the attachment of HSV-1 particles on the cell surface, assisting the cell-free infection of HSV-1.

Fig. 2: XYLT2 and EXT2 as essential genes for HSV-1 binding and infection.
figure 2

a Western blotting analysis confirming the absence or complementation of XYLT2, the EXT2 expression in KO, and rescued clones of HAP1 cells. b, c HepS expression analysis by FACS. After treatment with or without heparinase for 1 h at 37 °C, the cells were stained using an anti-HepS antibody. The graph summarized the mean fluorescent intensity (MFI) of the HepS expression analysis of three independent experiments (c). Results are presented as means ± SEM. Viability of the nontargeting control clone (d) and the XYLT2- and EXT2-KO clones (e) of HAP1 cells after HSV-1 infection. The cells were either pretreated with 5 μM surfen hydrate or DMSO and infected with HSV-1 at a MOI of 9. The cell viabilities were measured via the MTS assay at 48 hpi. The results are presented as means ± SEM of three independent experiments. f Progeny virus production of HSV-1 in the control, XYLT2-KO, and EXT2-KO clones of HAP1 cells. Following surfen hydrate or DMSO treatment, the cells were infected with HSV-1 at a MOI of 1. The progeny viruses were collected at 48 hpi and titrated via the plaque assay. The results are presented as means ± SEM of three independent experiments. g Binding of HSV-1 in the control, XYLT2-KO, and EXT2-KO clones of HAP1 cells. After surfen hydrate or DMSO treatment, the cells were adsorbed with HSV-1 at an MOI of 50 for 1 h and 4 °C. The nontargeting control cells that were mock-infected with PBS were used as a negative control. After removing the residual viruses, the viral DNA was extracted and quantified using real-time PCR. The results are presented as means ± SEM of three independent experiments and shown as the relative DNA level of the control cells. h Viability of the nontargeting control clone, XYLT2-KO, and EXT2-KO human retinal pigment epithelial-1 cell clones after HSV-1 infection. The cells were infected with HSV-1 at an MOI of 9. Then, the viabilities were measured via the MTS assay at 48 hpi. The results are presented as means ± SEM of three independent experiments. Asterisks, p < 0.05; Double asterisks, p < 0.01; n.s. not significant; N.D. not detected.

Full size image

HepS is important for HSV-1 cell-to-cell infection

In addition to cell-free infections, we investigated whether HepS contributed to the other modes of HSV-1 transmission, i.e., cell-to-cell infection. XYLT2 initiates sugar chain formation by transferring one xylose to the serine residues of core proteins, followed by the conjugation of two galactose molecules and one glucuronic acid (Fig. 1d). These first four sugars are shared by other glycosaminoglycans: chondroitin sulfate and dermatan sulfate26. Therefore, to ensure that only HepS expression was affected, an EXT2-KO clone and its rescue clone in RPE-1 cells were generated using the CRISPR-Cas9 system (Fig. 3a). Here we employed RPE-1 cells to accurately calculate the plaque areas because the plaque formation in HAP1 cells by HSV-1 infections showed an obscure margin. The reduced HepS expression of EXT2-KO RPE-1 cells was recovered by complementing with EXT2 (Figs. 3b, c). Subsequently, RPE-1 WT cells infected with HSV-1 were used as an inoculate, after which these inoculates (30 cells/well) were cocultured with a pre-seeded EXT2-KO clone or nontargeting control clone. Neutralizing antibodies against HSV-1 were added to the media to prevent cell-free spread. The plaque sizes for the EXT2-KO clone were decreased compared with that of the nontargeting control clone at 72 hpi (Figs. 3d, e), whereas the number of plaques was almost the same (Figs. 3d–f). Furthermore, the complementation of EXT2 recovered the plaque size (Figs. 3d, e). These findings propose that HepS contributes to both the cell-free and cell-to-cell infection of HSV-1.

Fig. 3: HepS is important for HSV-1 cell-to-cell infection.
figure 3

a Western blotting confirming the absence or complementation of EXT2 expression in the KO and complemented clones of RPE-1 cells. b, c HepS expression analysis by FACS. The graph summarizes the MFI of the HepS expression analysis of three independent experiments (c). The results are presented as means ± SEM. e–f HSV-1 cell-to-cell spreading assay. Each clone was inoculated with 30 cells/well of wildtype RPE-1, which pre-infected with HSV-1 and then incubated in a media containing 5 mg/mL pooled human IgG (d). The areas of 10 individual plaques were measured (e), and the number of plaques was counted (f). The results are presented as means ± SEM of three independent experiments and shown as the relative size (e) and the relative number of plaques (f). Double asterisks, p < 0.01; n.s. not significant.

Full size image

Knockout of PAPSS1 in HAP1 cells impairs the biosynthesis of HepS and reduces the attachment of HSV-1 on the cell surface

Our screening identified PAPSS1 as a host factor of HSV-1 infection (Fig. 1). PAPSS1 encodes an enzyme that synthesizes PAPS. PAPS is synthesized by PAPSS1 in the cytosol or nucleus, then transported into the lumen of the Golgi apparatus, where sulfation of various substrates, including HepS, occurs29,30 (Fig. 4a). The transporter of PAPS is encoded by SLC35B2, which was also highly enriched in our screen (Fig. 1), and it was previously considered essential in HepS biosynthesis in HAP1 cells and 293 T cells24,25. Thus, we hypothesized that PAPSS1 is also required for HepS biosynthesis. To test this, we obtained two PAPSS1-KO clones of HAP1 cells and the rescued clones (Fig. 4b). We analyzed HepS expression by FACS analysis. As shown in Figs. 4c, d, FACS analyses revealed drastically diminished background-level HepS expression in PAPSS1-KO clones, which was rescued by the complementation of PAPSS1. Furthermore, HepS chain analysis by high-performance liquid chromatography (HPLC) demonstrated that although the total amount of HepS disaccharides was reduced by approximately 30%, the sulfation degree was decreased approximately by 20-fold in PAPSS1-KO cells (Fig. 4e, Supplementary Fig. 4, and Supplementary Table 2). Consequently, PAPSS1-KO cells exhibited a robust resistance against HSV-1 infection, presenting a >90% survival rate at 48 hpi (Fig. 4f). The knockout of PAPSS1 decreased the binding of virus particles on the cell surface (Fig. 4g), causing a reduction in progeny virus production by 2–3 log fold (Fig. 4h).

Fig. 4: PAPSS1 is an important gene for HepS biosynthesis.
figure 4

a Schematic representation showing the contribution of PAPSS1 and SLC35B2 to the sulfation of substrates within the Golgi apparatus. b Western blotting confirming the absence or complementation of PAPSS1 expression in the KO and rescued clones of HAP1 cells. c, d HepS expression analysis by FACS. The graph summarizes the MFI of HepS expression analysis of three independent experiments (d). The results are presented as means ± SEM. e HepS disaccharide analysis by HPLC. The total amount and the sulfation degree of HepS disaccharides of the WT, nontargeting Ctrl, and PAPSS1-KO clones of HAP1 cells were measured. The results are presented as means ± SEM of four independent experiments. f Viability of PAPSS1-KO clones of HAP1 cells after HSV-1 infection. The cells were infected with HSV-1 at a MOI of 9. Their cell viabilities were measured using the MTS assay at 48 hpi. The results are presented as means ± SEM of three independent experiments. g Binding of HSV-1 in the PAPSS1-KO clones. The cells were adsorbed with HSV-1 at an MOI of 50 for 1 h at 4 °C. The nontargeting control cells that were mock-infected with PBS were used as a negative control. After removing the residual viruses, the viral DNA was extracted and quantified using qPCR. The results are presented as means ± SEM of three independent experiments and are shown as the relative DNA level compared with that of the control. h Progeny virus production of HSV-1 in the PAPSS1-KO clones. The cells were infected with HSV-1 at an MOI of 1. The progeny viruses were collected at 48 hpi and titrated using the plaque assay. The results are presented as means ± SEM of three independent experiments. i Binding of the indicated viruses in the control and the PAPSS1-KO clones. The cells were adsorbed with the virus for 1 h at 4 °C. The non-targeting control cells that were mock-infected with PBS were used as a negative control. After removing the residual viruses, the viral DNA (HSV-2 and hepatitis B virus) or RNA (human coronavirus OC43) was extracted. The viral genome was quantified using qPCR. The results are presented as means ± SEM of three independent experiments and shown as relative DNA levels compared with that of the control. Asterisks, p < 0.05; Double asterisks, p < 0.01; N.D. not detected; n.s. not significant.

Full size image

We further examined whether PAPSS1 contributes to the binding of other viruses that are already known to use HepS for binding16,17,21. Likewise, for HSV-1, the binding of HSV-2, HBV, and HCoV OC43 was significantly reduced in the PAPSS1-KO clones compared with the control cells (Fig. 4i). Hence, our cumulative findings suggested that PAPSS1 influences HepS biosynthesis, contributing to viral infections by assisting the binding of virus particles.

PAPSS1 influence on HepS biosynthesis depends on PAPSS2 expression

We further investigated whether PAPSS1 was necessary for HSV-1 infection and HepS biosynthesis in RPE-1 cells. As shown in Fig. 5a, PAPSS 1-KO RPE-1 cells showed significantly, but weaker resistance against HSV-1 infection than XYLT2- and EXT2-KO clones (Fig. 2h). Consistent with this, FACS analysis presented a slight reduction in the HepS expression of PAPSS1-KO RPE-1 cell clones (Figs. 5b, c). To evaluate the impact of PAPSS1-KO on HepS biosynthesis in other cell lines, we additionally obtained PAPSS1-KO clones using human gastric adenocarcinoma (AGS) cells and lung adenocarcinoma A549 cells (Supplementary Fig. 5). The PAPSS1-KO clones of AGS cells showed reduced HepS expression to the background level, similar to HAP1 cells. However, the PAPSS1-KO clones of A549 cells showed a limited reduction in HepS expression, similar to RPE-1 cells (Figs. 5b, c). This limited effect of PAPSS1-KO on HepS biosynthesis implied the expression of a paralog enzyme, PAPSS2. When we investigated the level of PAPSS2 expression, it correlated with the limited reduction of HepS expression by PAPSS1-KO (Fig. 5d). To assess the functional redundancy between PAPSS1 and PAPSS2, we generated PAPSS2-KO clones and double KO (DKO) clones of PAPSS1 and PAPSS2 using RPE-1 cells (Fig. 5e). Similar to single PAPSS1-KO clones, single PAPSS2-KO clones showed a limited reduction of HepS expression in RPE-1 cells and little or no gain of resistance to HSV-1 infection. However, DKO clones showed a strong decrease in HepS expression and robust resistance to HSV-1 infections (Figs. 5f–h). These findings propose the redundant role of PAPSS1 and PAPSS2 in the HepS biosynthesis of RPE-1 cells.

Fig. 5: PAPSS1-dependency of HepS biosynthesis differs between cell lines due to the PAPSS2 expression.
figure 5

a Cell viability of PAPSS1- KO of RPE-1 cell clones after HSV-1 infection. The cells were infected with HSV-1 at a MOI of 9. Their cell viabilities were measured using the MTS assay at 48 hpi. The results are presented as means ± SEM of three independent experiments. b, c HepS expression analysis of PAPSS1-KO clones of RPE-1, AGS, and A549 cells by FACS (b). The graph summarizes the MFI of the HepS expression analysis of three independent experiments (c). The results are presented as means ± SEM. d Western blotting showing the expression of PAPSS1 and PAPSS2 in HAP1, RPE-1, AGS, and A549 cells. The arrowhead indicates the cross-reactive band of PAPSS1. e Western blotting confirming the absence of PAPSS1 and/or PAPSS2 expression in the KO clones of RPE-1 cells. The arrowhead indicates the cross-reactive band of PAPSS1. f, g HepS expression analysis of PAPSS2-KO and PAPSS1- and PAPSS2-double KO (DKO) clones of RPE-1 cells by FACS (f). The graph summarized the MFI of HepS expression analysis of three independent experiments (g). The results are presented as means ± SEM. h Viability of PAPSS2-KO and DKO clones of RPE-1 cells after HSV-1 infection. The cells were infected with HSV-1 at an MOI of 9. The cell viabilities were measured by the MTS assay at 48 hpi. The results are presented as means ± SEM of three independent experiments. Double asterisks, p < 0.01; n.s. not significant.

Full size image

Discussion

Loss-of-function genetic screening has identified various host factors required for viral infection31,32. This screening strategy involves the isolation of mutations that render host cells resistant to viral infection using a knockdown or knockout mutant library. Compared with previous gene knockdown approaches, such as RNAi, the complete ablation of gene expression using a haploid screen or CRISPR screen produces clearer phenotypes on virus infection, enabling the identification of critical host factors for viral infection. This approach has been applied previously to uncover host factors involved in viral entry. Haploid screens via insertional mutagenesis discovered essential receptors for the Ebola virus33 and Lassa virus34,35. Recently, CRISPR screening identified host factors that regulate SARS-CoV-217,36,37, common cold coronaviruses17,37, HIV38, adeno-associated virus39, flavivirus40, and Zika virus41. In addition to identifying the entry receptor of the Lassa virus, Jae et al. revealed the gene networks encoding for enzymes and accessory factors involved in the O-linked glycosylation of α-dystroglycan35. Likewise, we also performed a genome-wide CRISPR screen for HSV-1 host factors and revealed a set of genes involved in HepS biosynthesis, including a viral entry receptor, NECTIN1. The biosynthesis of HepS has also been reported as a sequential multistep process that occurs in the Golgi apparatus (Fig. 1d)42. Our results showed that knocking out XYLT2 and EXT2 caused a severe defect in HepS expression on the cell surface (Fig. 2). XYLT2 is a glycosyltransferase that initiates the biosynthesis of glycosaminoglycan chains (both HepS and chondroitin sulfate/dermatan sulfate) in proteoglycans26, and EXT2 is involved in heparan sulfate chain elongation27. In contrast to the genes related to the backbone formation of HepS, no other sulfotransferases were enriched in our screening process, possibly because of their functional redundancy.

Additionally, all the host factors identified in our screening were involved in viral attachment and entry, the earliest stage of viral infection. This bias might be caused by the long selection period of our screening. The cells lacking host factors engaged in the later stages of infections would not allow viruses to replicate or mature, but allow to enter the cell. The cytopathic and cytocidal effects are caused even by the entry of UV-inactivated HSV-143,44 and replication-defective mutants45. Alterenatively, the cells lacking host factors engaged in the later stages of infections could not propagate sufficiently or underwent apoptosis for harboring viruses46,47. Therefore, the cells allowed viruses to enter the cell might be eliminated during the selection.

Although we demonstrated that HepS contributed to both the cell-free and cell-to-cell infection of HSV-1 (Figs. 2 and 3), how HepS is involved in the cell-to-cell infection of HSV-1 remains unclear. During vaccinia virus infection, the viral factor, F11L, and a vaccinia growth factor contribute to the cell-to-cell spread of the vaccinia virus via the activation of RhoA- and EGFR-mediated cell motility, respectively48,49,50. Cell motility is essential for the dissemination of various viruses48,51,52,53. Recently, it was reported that the interaction between HSV-1 glycoprotein E and prohibin-1 activates the MAPK/ERK pathway, promoting cell-to-cell spread54. This signaling pathway also promotes cell motility55,56. Similarly, HepS on the cell surface can serve as an anchor to facilitate the endocytosis of growth factors, subsequently inducing rapid and efficient motility57,58,59. These findings invoke a model in which HSV-1 uses HepS to promote cell motility, increasing cell-to-cell contact with the neighboring uninfected cells. However, further studies are warranted to determine how HepS contributes to HSV-1 cell-to-cell transmission, for instance, whether the motility of the infected cells is affected by HepS expression on their cell surface and/or by the MAPK/ERK pathway via HepS-dependent endocytosis.

The HepS chain’s modification is complex, occurring via sulfation. As a result, its molecular diversity generates binding sites for multiple protein ligands60,61. This study identified PAPSS1 as a critical factor of HepS biosynthesis in HAP1 cells (Fig. 4). PAPSS1 produces the universal sulfate donor PAPS required for the sulfation of polysaccharides. We demonstrated a diminished HepS expression of PAPSS1-KO clones through FACS analysis. Consistent with this, HepS disaccharide analysis also revealed a decrease in the level of HepS disaccharides and entire sulfation, resulting in reduced viral attachment on the cell surface in HAP1 cells (Fig. 4). In humans, PAPS, the sole source of sulfate, is synthesized from adenosine 5-prime triphosphate (ATP) and inorganic sulfate by two enzymes; PAPSS1 and PAPSS2. While PAPSS1 is ubiquitously expressed in human adult tissues, PAPSS2 shows a more restricted expression pattern62,63. In this study, KO analysis of these enzymes showed PAPSS1 and PAPSS2 functions to be redundant during HepS biosynthesis in RPE-1 cells (Fig. 5). In a previous study, defects in PAPSS2 caused the Pakistani type of spondyloepimetaphyseal dysplasia62,64. However, no information is available on why the ubiquitously expressed PAPSS1 cannot compensate for the loss of PAPSS2 activity. Differences in subcellular compartmentalization30, catalytic efficiency, and expression patterns during cartilage and bone development62,63 might explain the differential effects of PAPSS1 and PAPSS2.

In summary, our study findings confirmed the significance of HepS in HSV-1 infection and identified PAPSS1 as the key factor for HepS biosynthesis. The CRISPR-mediated inactivation of PAPSS1 abolished the sulfated HepS expression on the cell surface, which resulted in the reduced binding of various pathogenic viruses in HAP1 cells. Furthermore, in some other cells that express PAPSS2, such as RPE-1 and A549 cells, PAPSS2 compensated for PAPSS1 in HepS biosynthesis. Therefore, our study provides further insights into the host factors required for HSV-1 infection and HepS biosynthesis regulation.

Methods

Cells and virus stocks

HAP1 cells were purchased from Horizon (Cambridge, UK) and cultured in Iscove’s Modified Dulbecco’s Medium (IMDM) (Nakalai Tesque, Kyoto, Japan) supplemented with 10% fetal bovine serum (FBS), 1% GlutaMAX (Thermo, Waltham, MA, USA), and 1% penicillin-streptomycin (Pen/St) (Thermo). HEK293T (CRL-3216; ATCC), A549 (CCL-185; ATCC), and Huh7 (RCB 1942; RIKEN BRC Cell Bank) cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% FBS, 1% GlutaMAX, and 1% Pen/St. RPE-1 cells (CRL-4000; ATCC) were cultured in DMEM/F-12 medium (Nakarai Tesque) supplemented with 10% FBS, 1% GlutaMAX, and 1% Pen/St. Vero cells (CCL-81; ATCC) were cultured in Eagle’s minimal essential medium supplemented with 10% calf serum (CS), 1% L-glutamine (Thermo), and 1% Pen/St. AGS cells (CRL-1739; ATCC) were cultured in RPMI-1640 medium (Nakalai Tesque) supplemented with 10% FBS, 1% GlutaMAX, and 1% Pen/St. HepAD38 cells (a gift from Christoph Seeger) were cultured as described previously65. All cells were maintained at 37 °C and 5% CO2 unless otherwise indicated.

HSV-1 strain F and HSV-2 strain 186 were propagated in Vero cells. HCoV strain, OC43, was obtained from ATCC (VR-1558) and propagated in Huh7 cells at 35 °C. HBV genotype D (subtype ayw) was prepared from HepAD38 cells, as described previously65.

Antibodies, plasmids, and reagents

Anti-HepS (10E4 epitope; H1890, 1:50) mouse monoclonal antibody was purchased from US Biological Life Sciences (Salem, MA, USA). Anti-XYLT2 (G-1; sc-374134, 1:500), anti-EXT2 (A-2; sc-514092, 1:500), anti-PAPSS1 (A-2; sc-376244, 1:500), anti-PAPSS2 (SQ-19; sc-100801, 1:500) and normal mouse IgM (sc-3881, 1:50) antibodies were obtained from Santa Cruz Biotechnology (Dallas, TX, USA). Anti-IRF2BPL (NBP2–56241, 1:500) antibody was purchased from Novus Biologicals (Littleton, CO, USA). Anti-VANGL2 (clone 2G4; MABN750, 1:500) antibody was obtained from Merck (Darmstadt, Germany). Anti-GAPDH (D16H11; #5174, 1:2000), horseradish peroxidase-conjugated anti-mouse or anti-rabbit secondary antibodies (#7074 and #7076, 1:1000 and 1:1000) were purchased from Cell Signaling Technology (Danvers, MA, USA). A horseradish peroxidase-conjugated anti-rat (SA00001-15, 1:1000) secondary antibody was obtained from Proteintech Group (Rosemont, IL, USA). Alexa Fluor 488-conjugated anti-mouse IgM secondary antibody (A-10680, 1:50) was obtained from Thermo.

The full-length cDNAs of EXT2, NECTIN1, PAPSS1, and XYLT2 were obtained by reverse transcriptase-PCR (RT-PCR) and cloned into the CSII-CMV-MCS-IRES2-Bsd vector (RIKEN BioResource Center, Wako, Japan) using the In-Fusion cloning system (Takara Bio, Kusatsu, Japan). The primers used for RT-PCR are listed in Supplementary Table 3. To KO the gene of interest, sgRNA was designed using CHOPCHOP (https://chopchop.cbu.uib.no)66 and cloned into px459 (a gift from Feng Zhang; #48139, Addgene, Watertown, MA, USA) or lentiviral expression constructs containing Cas9 and blasticidin- or neomycin-resistant genes, generated by VectorBuilder® (Shenandoah, TX, USA). The oligos of the used sgRNAs are shown in Supplementary Table 4. Inserted DNA sequences of each vector were confirmed through direct DNA sequencing.

Surfen hydrate (S6951) and heparinase II (#P0736) were purchased from Sigma-Aldrich and New England Biolabs (Ipswich, MA, USA), respectively. miRNA inhibitors (SMI-001-MIMAT0019709 and SMC-2101) were purchased from Bioneer Corporation (Daejeon, Republic of Korea). Lipofectamine 2000 was obtained from Thermo.

Genome-wide CRISPR screen

HAP1/Cas9 cells were generated through lentiviral transduction with lentiCas9-blast (a gift from Feng Zhang; #52962, Addgene), followed by antibiotic selection with blasticidin for seven days. To establish CRISPR KO libraries, 120 million HAP1/Cas9 cells were transduced with each lentiviral human GeCKO v2 library A and B (SureGuide GeCKO v2.0 Human Exome CRISPR Library; Agilent Technologies, West Cedar Creek, TX, USA) at an MOI of 0.3. Cells were subsequently selected using puromycin and cultured for 7 days. This culturing equates to >500-fold coverage of the library after selection. Twenty million selected cells (estimated coverage: 150 cells/sgRNA) were subjected to the first HSV challenge at an MOI of 0.1 with 5 mg/mL of pooled human IgG (Equitech-Bio, Kerrville, TX, USA). Survivor cells were cultured for 2 weeks, followed by a second HSV challenge at an MOI of 0.1 without IgG. After an additional 2 weeks, 20 million viable cells were harvested, followed by gDNA extraction with the QIAmp DNA Mini Kit (QIAGEN, Hilden, Germany). To ensure that HSV-1 attacked all the cells during the selection period, the same number of HAP1/Cas9 cells were infected with HSV-1 and cultured likewise, and the survival of none of the cells was confirmed. The control group was mock-infected with PBS and cultured for the same period as the HSV-challenged group. To evaluate sgRNA enrichment, regions containing sgRNA cassette were amplified using the PrimeSTAR GXL DNA Polymerase (Takara Bio) and primers complementary to a common sequence of the lentivirus. For the first round of PCR, 35 reactions containing 1 μg gDNA were set up and amplified for 24 cycles. First, the PCR products were pooled, mixed, and subsequently subjected to second PCR. Then 5 mL of the mixed first PCR product were amplified for six cycles with index primers during the second round. PCR products were purified using the Agencourt AMPure XP beads (Beckman Coulter, Pasadena, CA, USA) and sequenced on the Illumina HiSeq2500 (Illumina, San Diego, CA, USA) according to the manufacturers’ instructions. Primer sequences used in the study are listed in Supplementary Table 5.

Generation of clonal knockout and complemented cell lines

HAP1, AGS, or A549 cells were transfected with three different px459-sgRNA targeting the gene of interest (triple-target CRISPR) as described previously67. Next, the cells were selected using puromycin for 5 days. RPE-1 cells were transduced with a lentiviral vector generated by VectorBuilder®, followed by 7 days of antibiotic selection using blasticidin or neomycin. Clonal lines were established by limited dilution after selection. KO was confirmed through Sanger sequencing or Western blotting.

Lentiviral transduction was performed to generate stable cell lines complementarily expressing a selected gene of interest under a CMV promoter. Lentivirus was produced by transfecting HEK293T cells with CSII plasmids, pCMVR8.74 (a gift from Dider Trono and Yasuo Ariumi; #22036, Addgene) and pCMV-VSV-G (a gift from Bob Weinberg; #8454, Addgene). KO cells were transduced with the lentivirus encoding the gene of interest. Transduced cell populations were selected with blasticidin and the transgene expression was confirmed by immunoblotting.

Heparinase treatment

Cells were washed twice with PBS and incubated with 2 U/mL heparinase II diluted in PBS for 1 h at 37 °C. The cells were washed twice again with PBS and then subjected to subsequent experiments.

HepS staining

Cells were harvested using the TrypLE Select Enzyme (Sigma-Aldrich), washed with PBS, and subsequently incubated with monoclonal anti-HepS antibody diluted in PBS + 2% FBS. After 1 h of incubation at 4 °C, the cells were washed with PBS + 2% FBS and incubated with Alexa 488-conjugated secondary antibody. The cells were again washed twice with PBS + 2% FBS. Fluorescence was measured using the FACS Canto2 (BD Biosciences, Freemont, CA, USA) and analyzed using FlowJo software (FlowJo Inc, Ashland, OR, USA).

Cell viability assay

Cell viability was measured using the Cell Titer 96 AQueois One Solution reagent (Promega, Madison, WI, USA). In brief, wells containing 100 μL of the medium in 96-well plates were supplemented with 20 μL of 3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS) reagent, followed by 2 h of incubation at 37 °C. Then, the absorbance was recorded at 490 nm using a 96-well plate reader (Sunrise Basic; Tecan Japan, Kawasaki, Japan).

Progeny virus titer measurement

Cells were infected with HSV-1 at an MOI of 1. After 24 h of incubation at 37 °C, the cells were collected with the supernatant and stored immediately at −80 °C. The samples were frozen and thawed thrice before titration. On the next day, 100 μL of the samples serially diluted in PBS were inoculated on pre-seeded Vero cells, then incubated at 37 °C for 1 h adsorption. After adsorption, 1 mL fresh medium containing 5% CS was added, and the cells were incubated for 48 h. These cells were fixed with 10% formaldehyde and stained with 0.5% crystal violet. Plaques were observed and counted.

Virus binding assay

For the virus binding assay, 2 × 105 cells were seeded into 24-well plates. On the next day, the medium was replaced with fresh medium containing 5 μg/mL of surfen hydrate or DMSO, followed by incubation for 1 h at 37 °C. The cells were inoculated with viruses in 100 μL ice-cold PBS. After 1 h of adsorption on an ice bath, the cells were washed with ice-cold PBS three times and collected for DNA extraction using the DNeasy Blood and Tissue Kit (QIAGEN) or RNA extraction with the NucleoSpin RNA Plus (Takara Bio). The HSV-1 and 2 viral DNA levels were measured via qPCR as previously described68. The HBV DNA and cDNA of HCoV OC43 were measured and normalized by GAPDH using the following primers: For HBV, 5′-GTG TCT GCG GCG TTT TAT CA-3′ and 5′-GAC AAA CGG GCA ACA TAC CTT-3′; for HCoV OC43, 5′-CGA TGA GGC TAT TCC GAC TAG GT-3′ and 5′-CCT TCC TGA GCC TTC AAT ATA GTA ACC-3′; for GAPDH, 5′-TGC ACC ACC AAC TGC TAG C-3′ and 5′-GGC ATG GAC TGT GGT CAT GAG-3′.

Cell-to-cell spreading assay

WT RPE-1 cell monolayers were exposed to HSV-1 at an MOI of 50 for 1 h at 37 °C to enable virus entry. Then, the cells were washed twice with PBS, followed by treatment with 40 mM of citrate buffer (pH 3.0) for 1 min to deactivate the extracellular residual viruses. The cells were again washed twice with PBS and collected to inoculate onto 90% confluent monolayers of uninfected RPE-1 clones. Approximately 30 cells/well were inoculated with infection media (IMDM + 1% FBS) containing 5 mg/mL of pooled human IgG, which was previously reported to be a sufficient concentration to neutralize nearly all virions per 1 × 106 PFU69. At 72 hpi, the cells were fixed for 10 min in 10% formaldehyde and stained with 0.5% crystal violet. The number of plaques was counted manually by direct observation, and the area of plaques was analyzed using ImageJ Fiji software70.

Quantification of HepS chains

Glycosaminoglycans (GAGs) from cultured cells were prepared as described previously71 with slight modifications. Briefly, the cells were homogenized in ice-cold acetone at least three times, and air-dried. The dried materials (correspond to acetone powders) were exhaustively digested with heat-activated actinase E (10% [w/w] of dried materials) in 0.1 M borate buffer, pH 8.0, containing 10 mM CaCl2, at 55 °C for 48 h. The digest was treated with 5% trichloroacetic acid, and the resultant acid-soluble fraction was adjusted to contain 80% ethanol. The resultant precipitate was dissolved in water and subjected to gel filtration on a PD-10 column (Cytiva), using water as an eluent. The flow-through fractions were collected, evaporated to dryness, and dissolved in water (crude GAG fractions).

An aliquot of the crude GAG sample was digested with a mixture of heparinase (0.5 mIU, Seikagaku, Tokyo, Japan) and heparitinase (0.5 mIU, Seikagaku) in 20 mM sodium acetate, pH 7.0, containing 2 mM calcium acetate at 37 °C for 4 h. The digests were derivatized with fluorophore 2-aminobenzamide and analyzed by anion-exchange high-performance liquid chromatography on a PA-G column (YMC, Kyoto, Japan), as described previously72. The identification and quantification of the resulting disaccharides were achieved by comparison with authentic unsaturated HepS disaccharides (Seikagaku).

The HepS sulfation degree was calculated as the mol% of the total sulfated HepS disaccharides multiplied by their respective sulfate group numbers (1, 2, or 3).

Statistics and reproducibility

Data were processed using Microsoft Excel. All analyses were performed using two-tailed Student’s t-test. P < 0.05 was considered statistically significant. Results are presented as means ± standard error of measurement (SEM) of at least three independent experiments. The sample sizes and number of replicates were described in the figure legends. The experiments were not randomized and the investigators were not blinded.

Reporting summary

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