SLC35A2 modulates paramyxovirus fusion events during infection

Paramyxoviruses are significant human and animal pathogens that include mumps virus (MuV), Newcastle disease virus (NDV) and the murine parainfluenza virus Sendai (SeV). Despite their importance, few host factors implicated in paramyxovirus infection are known. Using a recombinant SeV expressing destabilized GFP (rSeVCdseGFP) in a loss-of-function CRISPR screen, we identified the CMP-sialic acid transporter (CST) gene SLC35A1 and the UDP-galactose transporter (UGT) gene SLC35A2 as essential for paramyxovirus infection. SLC35A1 knockout (KO) cells showed significantly reduced binding and infection of SeV, NDV and MuV due to the lack of cell surface sialic acids, which act as their receptors. However, SLC35A2 KO cells revealed unknown critical roles for this factor in virus-cell and cell-to-cell fusion events during infection with different paramyxoviruses. While the UGT was essential for virus-cell fusion during SeV entry to the cell, it was not required for NDV or MuV entry. Importantly, the UGT promoted the formation of larger syncytia during MuV infection, suggesting a role in cell-to-cell virus spread. Our findings demonstrate that paramyxoviruses can bind to or enter A549 cells in the absence of canonical galactose-bound sialic-acid decorations and show that the UGT facilitates paramyxovirus fusion processes involved in entry and spread.


Introduction
The Paramyxovirus family includes the major human and animal pathogens measles virus (MV), mumps virus (MuV), human parainfluenza virus (hPIV), Newcastle disease virus (NDV) and the highly pathogenic zoonotic Hendra (HeV) and Nipah (NiV) viruses imposing a significant burden on global public health, while also causing substantial economic losses [1,2].
Paramyxoviruses are single-stranded negative-sense RNA enveloped viruses containing a fusion (F) and an attachment glycoprotein on their surface.These glycoproteins are essential for virus entry and infection.Attachment proteins vary across the different genera of paramyxoviruses, and can be either the glycoprotein (G), the hemagglutinin (H), or the hemagglutinin-neuraminidase (HN).The Avulavirus (e.g., NDV), Rubulavirus (e.g., MuV), and Respirovirus (e.g., SeV) genera attach to the cell surface through the virus HN protein that binds sialic acid-containing cell surface molecules [3].Sialic acids also serve as attachment receptors for many other viruses, including influenza virus, reovirus, adenovirus, and rotavirus [4].After the virus attaches to the host cell, the F protein undergoes a conformational change that triggers the fusion of the host cell and viral membranes.Virus-cell membrane fusion leads to the release of the viral ribonucleoprotein complex into the cytosol, allowing for viral replication and transcription to occur [5].In some cases, the virus enters and fuses with the endosomal membrane [6,7].In addition, the F protein can facilitate cell-to-cell fusion and syncytia formation, for example during MuV infection [8][9][10].
Sialic acids are bound to carbohydrate chains on glycoproteins and glycolipids in the Golgi apparatus via different glycosidic linkages.The most common linkage types are α2,3-linkage to a galactose residue, α2,6-linkage to a galactose residue, α2,6-linkage to an N-acetylgalactosamine residue, and α2,8-linkage to another sialic acid moiety on a glycan [4].Sialic acids and galactose are transported into the Golgi by the CMP-sialic acid transporter (CST) and the UDP-galactose transporter (UGT) encoded by SLC35A1 and SLC35A2, respectively.CST facilitates the assembly of sialic acid onto glycoproteins and glycolipids [11].SeV and MuV are reported to only use α2,3-linked sialic acid to attach to cells [12][13][14][15], while NDV can bind to both α2,3-linked and α2,6-linked sialic acids [16].All reported paramyxovirus receptors involve sialic acids linked to a galactose, suggesting that this glycan motif may be essential for sialic aciddependent virus infection.
Targeting host factors essential to the viral lifecycle is one promising avenue for antiviral drug development [17,18].Unfortunately, the list of known cellular host factors and their importance in modulating the paramyxovirus lifecycle is relatively sparse when compared to other viruses such as influenza virus and coronavirus [19][20][21][22].Even less is known about the common and divergent host protein requirements among different paramyxoviruses.Given the significance of paramyxoviruses in disease and the lack of clear candidates for a host-directed antiviral drug design, unbiased and high-throughput screening for host factor dependencies remains a necessary research objective for this virus family.
In this work, we used the murine paramyxovirus Sendai virus (SeV) which causes respiratory infection in mice and is widely used as a model paramyxovirus [23][24][25][26][27][28], to perform CRISPR-Cas9-based screenings for essential pro-viral host factors.We leveraged a novel recombinant SeV strain expressing a destabilized eGFP (dseGFP) reporter that allowed for sensitive measurements of viral genome replication and transcription within the infected cell, permitting more accurate analysis of the CRISPR-Cas9 knockout (KO) library screening results.
Consistent with several published screens for other sialic-acid dependent RNA viruses, we found that the top essential pro-viral genes included SLC35A1 [29][30][31][32] and SLC35A2 [32,33].SLC35A1 serves as an essential gene for the expression of the virus attachment receptors.In contrast, we discovered that UGT, in addition to contributing to virus attachment, plays independent roles in paramyxovirus virus-cell and cell-cell fusion processes.

CRISPR knock-out screen identifies SLC35A1 and SLC35A2 as essential factors for Sendai virus infection
To identify host factors essential for paramyxovirus infection, we developed genome-wide CRISPR KO libraries in A549 cells to screen for infection with the model virus SeV (Fig 1A).Cas9stable A549 cells were generated by transduction with a lentivirus expressing Cas9.Several single cell clones of A549-Cas9 cells were selected based on the expression of Cas9 as determined by western blot (Fig S1A).The Cas9 activity of the clones was then confirmed by an eGFP knockout assay where higher Cas9 activity results in a lower percentage of GFP positive cells (Fig S1B) [34].The A549-Cas9 single cell clone 1 with the highest Cas9 efficiency (Fig S1C ) was selected and transduced at a low multiplicity of infection (MOI) of 0.3, with the Human CRISPR KO lentiviral single guide (sg) RNA Library Brunello [35] followed by puromycin selection.
For screening, we generated a recombinant SeV expressing a destabilized eGFP (rSeVC dseGFP ).This virus was generated by inserting a destabilized eGFP (dseGFP) between SeV NP and P genes (Fig S2A).The destabilization is due to a fused proline-glutamate-serinethreonine-rich (PEST) peptide to eGFP, which reduces the half-life of GFP from 20 hours to 2 hours and cause a 90% signal loss [36,37].As shown in Fig S2B, the rSeVC dseGFP did not show signs of attenuation in virus titer (10 8.28 vs 10 8.35 TCID 50 /ml) but exhibited lower eGFP intensity compared with rSeVC eGFP in infected A549 cells.To identify host factors regulating infection regardless of the antiviral response, we performed three screens using different immunostimulatory conditions.First, transduced cells were infected with either rSeVC dseGFP nonstandard viral genomes (nsVG)-negative stocks in the absence or presence of the JAK/STAT signaling inhibitor Ruxolitinib.Stocks without nsVGs lack strong immunostimulatory molecules [38], whereas drug treatment precludes interferon signaling.Second, another batch of transduced cells were infected with rSeVC dseGFP stock with a high content of immunostimulatory nsVGs (nsVG positive), which induce strong immune responses [39].Among the subset of sgRNAs that were enriched in the GFP-negative cell population in all three independent screenings relative to the control (S1-S3 Tables), we identified the genes SLC35A1 and SLC35A2 encoding the CMP-sialic acid transporter and the UDP-galactose transporter as significantly enriched (log fold change >2.5, p<0.01) (Fig 1C).

SLC35A1 and SLC35A2 are essential for SeV infection
The SLC35A1 gene encodes the CMP-sialic acid transporter (CST) necessary for the sialylation of proteins and lipids [11].The SLC35A2 gene encodes the UDP-galactose transporter (UGT) which is required not only for the galactosylation of N-and O-glycans on glycoproteins but also for the synthesis of galactosylceramide and galactosyl diglyceride [11].CST and UGT are found in the membrane of the Golgi apparatus and transport CMP-sialic acid and UDP-galactose from the cytosol into Golgi vesicles for the generation of glycans (Fig 2A, adapted from [11]).The terminal sugar chains of sialylated glycoproteins and gangliosides, such as GD1a and GQ1b, that act as SeV receptors are shown in Fig 2B .To validate the functional significance of SLC35A1 and SLC35A2 during SeV infection, these genes were disrupted in A549-Cas9 cells using sgRNAs.A control cell line was made by transducing a scramble sgRNA that did not target any specific host gene.Transduced cells were then selected with puromycin followed by singe cell cloning.We then tested for the presence of surface sialic acids and galactose in the KO cell lines by staining with the lectins Sambucus Nigra Agglutinin (SNA) and Erythrina Cristagalli Lectin (ECL) to detect cell-surface sialic acid and galactose, respectively as previously described [40,41]   We then used a SeV reporter virus expressing eGFP (rSeVC eGFP ) to directly assess the impact of SLC35A1 or SLC35A2 during infection.As a control, vesicular stomatitis virus (VSV) was used as it does not depend on sialic acid for entry.We looked for GFP expression at 24 hpi with an MOI of 1.5 for SeV and an MOI of 0.015 for VSV as a readout of infection and virus replication.Absence of SLC35A1 and SLC35A2 resulted in loss of infectivity in most cells, with only a few cells showing viral replication (Fig. 3A).We confirmed absence of SeV replication in SLC35A1 KO cells and drastically reduced replication in SLC35A2 KO cells by evaluating SeV NP mRNA expression by qPCR (Fig. 3B).In contrast, VSV infection proceeded normally in the absence of SLC35A1 or SLC35A2.To exclude the possibility of other defects that may result in the restriction of SeV infection in the KO cells, we complemented SLC35A1 KOs with DNA expressing SLC35A1-GFP and SLC35A2 KOs with cDNA expressing SLC35A2-GFP.At 24 hpi with 3 MOI of rSeVC miRF670 , we observed recovered SeV replication in complemented KOs (Fig 3C and 3D).Infection using high MOIs of rSeVC dseGFP or rSeVC eGFP showed more cells infected in SLC35A1 KO cells, and an even larger number in SLC35A2 KO cells at 24hpi, but in both cases the percentage of cells infected was significantly less than controls, suggesting that SeV can enter more cells when used at high MOIs independent of SLC35A1 and SLC35A2 but with limited spread (Fig S3).Overall, these data confirmed the critical, yet not completely overlapping, roles of SLC35A1 and SLC35A2 during SeV infection.

SLC35A1 and SLC35A2 differentially impact infection with several paramyxoviruses
To assess whether the observed non-overlapping functions of SLC35A1 and SLC35A2 were maintained during infection with other paramyxoviruses, we infected single and double KO A549 cells with the Respirovirus rSeVC eGFP , the Avulavirus rNDV eGFP , or the Rubulavirus MuV at an MOI of 1.5 and look for either GFP expression (SeV and NDV) or stained for MuV NP at 24 hpi.The double KO cell was made by transducing SLC35A1 KO cells with SLC35A2 sgRNA and it was confirmed by lectin staining (Fig 4A).In all cases, the absence of SLC35A1 or both SLC35A1 and SLC35A2 resulted in loss of infectivity in most cells, with only a few cells showing viral replication.However, for NDV and MuV infections, the absence of SLC35A2 alone resulted in an intermediate number of infected cells showing more infected cells than the double KO cells but significantly less than control cells (Fig. 4B and 4C).These data demonstrate that SLC35A2 has non-redundant functions with SLC35A1 during paramyxovirus entry and spread and suggest a differential impact of SLC35A2 during infection with different paramyxoviruses.

SLC35A2 differentially impacts SeV, NDV, and MuV infection and spread in A549 cells
To further investigate the impact of SCL35A2 on paramyxovirus infection and spread, we followed the infection in SCL35A2 KO cells through a 4-day infection period (Fig. 5).Interestingly, infections in SLC35A2 KO cells displayed varied phenotypes across these viruses.As shown before, SeV infection was drastically reduced to one or two cells per image (5X magnification) and the virus did not spread throughout the time course (Fig. 5A-B).NDV could infect a larger proportion of SLC35A2 KO cells before the infected cells died (Fig S4), but again, there was no evidence of virus spread, compared to NDV replication in control cells (Fig. 5C-D).In contrast, MuV infected and spread well in SLC35A2 KO cells as evidenced by staining for the virus NP (Figure 5E-F).Taken together, these data indicate that SLC35A2 plays differential roles in the infection and spread of different paramyxoviruses during infection of A549 cells.

SLC35A2 is essential for virus-cell fusion during SeV infection
We next focused on investigating where SLC35A2 played a critical role during the virus infection cycle.Based on the organization of the terminal sugar chains of sialylated glycoproteins and the apparent requirement for galactose for sialic acid proximal binding (Fig. 2B), we began by evaluating whether SLC35A2 impacts the attachment of SeV to the cell surface.As expected, SLC35A1 KO cells exhibited robust restriction of SeV binding evidenced by consistently negative cell surface HN staining in both infected and non-infected groups after co-incubation of virus and cells for 1 hours at 4C.However, incubation of SeV with SLC35A2 KO cells under the same conditions resulted in positive cell surface HN staining, albeit lower than control cells (Fig. 6A), suggesting that the impact of SLC35A2 on paramyxovirus infection extends beyond the virus binding step.
To directly test the impact of SLC35A2 in the SeV virus-cell fusion process, we tested for intracellular detection of the SeV internal M (matrix) protein as a marker for fusion [42] using a recombinant SeV rSeV-M-HA [43] in which the M protein is fused with an HA tag.In brief, SeV was incubated with cells at 37C for 3 hours to allow virus entry and fusion.The cells were then treated with Proteinase K followed by fixation, permeabilization, and staining of SeV M and envelope HN proteins.Proteinase K treatment was used to remove cell surface attached viral particles and the HN protein staining was used as a control for the presence of virus attached to the outside of the cells.As shown in Figure 6B, SeV HN was detected in both cell lines, similar to Figure 6A, but not after proteinase K treatment.However, the M protein was detected in the control cells but not in SLC35A2 cells, regardless of proteinase K treatment, suggesting that SLC35A2 is essential for efficient fusion of the virus and cell membrane during virus entry.
Lastly, to confirm that SLC35A2 did not directly affects SeV genome replication and transcription, we took advantage of the recombinant reporter SeV virus rSeVC eGFP∆FHN+GFtail .This virus was made by removing the original SeV F and HN and inserting a chimeric VSV glycoprotein G fused with the C-terminal tail of SeV F (GFtail) thus replicating as SeV but entering the cells as VSV (Fig. S5A).As expected, rSeVC eGFP∆FHN+GFtail can infect SLC35A1 KO cells using the VSV GFtail protein for entry (Fig S5B).Then we asked whether rSeVC eGFP∆FHN+GFtail can replicate without SLC35A2.Although both viruses infect control cells to a similar degree, rSeVC eGFP was unable to infect and replicate in SLC35A2 KO cells but rSeVC eGFP∆FHN+GFtail infected and spread normally in these cells (Fig. 6C).In addition, SeV transcription measured by SeV NP mRNA expression showed that SLC35A1 and SLC35A2 deletions have no effect on the viral polymerase activity (Fig. 6D).These data demonstrate that SLC35A2 is not essential for virus binding or viral genome replication and transcription but is critical for virus-cell fusion during SeV infection.KOs at MOIs of 1.5 or 6.Cellular RNA was collected 24 hours later and analyzed by qPCR.
Expression of mRNA was calculated relative to the housekeeping index.Data represent the mean of three independent experiments.Significance was calculated with a two-way ANOVA.ns: not significant.

SLC35A2 is implicated in syncytia formation during MuV infection
As MuV infected and transcribed in SLC35A2 KO cells (Fig. 4E-F), we next asked whether SLC35A2 impacted other steps of the MuV replication cycle.As shown in Fig. 7A and 7B, there was no significant difference in the virus production of infectious viral particles between control and SLC35A2 KO cells infected at MOI of 1.5 or 15, indicating that SLC35A2 does not impact MuV infectious particle production.Interestingly, we noticed less syncytia formation in SLC35A2 KO cells at both low and high MOI.To confirm this difference, we quantified the number of syncytia formed under low and high MOI conditions at 48 hpi.We defined an NP-positive large cell with more than three nuclei as syncytia.As shown in Fig. 7C-E, SLC35A2 KO cells formed less syncytia compared to control cells at both MOIs, suggesting that SLC35A2 is implicated in MuVinduced syncytia formation suggesting a role in cell-to-cell virus spread.

Discussion
This manuscript presents the results of a CRISPR-Cas9 knockout screen identifying the transporters genes SLC35A1 and SLC35A2 as important for paramyxovirus infection and the characterization of their roles during paramyxovirus infections.SLC35A1 is essential for the attachment of SeV, NDV, and MuV to the cell due to its role in exposing the viral receptor sialic acid on the cell surface, similar to what has been described for influenza virus and porcine delta coronavirus [29,31].Interestingly, while the role of SLC35A2 is assumed to be related to virus attachment as galactose is typically considered the sugar to which sialic acid is linked [4], we found that SLC35A2 is not essential for virus attachment to the cell surface.Instead, SLC35A2 is crucial for the fusion of SeV with the cells and for MuV induced cell-to-cell fusion and syncytia formation suggesting a specific role for this protein in fusion events during virus infection.
Virus-cell fusion is an important target for antiviral drug development.To date, research on the fusion process and anti-fusion strategies have mainly focused on the role the viral proteins F and HN play during viral infection [44,45].However, our understanding of the host factors involved in the processes of virus-cell and cell-to-cell fusion that occur during infection are limited to the described role of 25HC in interfering with NiV induced cell-to-cell fusion [46], and a the role of soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) protein USE1 in the glycosylation and expression of MuV fusion protein [47].To our knowledge, no host factors have been identified to affect paramyxovirus virus-to-cell fusion.
Specific N-glycans on the F protein of several paramyxoviruses are important for the fusion activity of the protein [48][49][50].However, we show successful infection of SLC35A2 KO and control cell lines with the same MuV virus stocks and production of similar levels of MuV infectious viral particles in SLC35A2 KO and control cells indicating normal activity of the F protein during virus entry in these conditions.These data suggest that either different glycosylations or factors beyond F glycosylation are impacted by SLC35A2 during MuV infection.In addition, SLC35A2 has been identified as an HIV X4 strain-specific restriction factor in primary target CD4+ T cells [51] and lack of SLC35A2 resulted in decreased influenza virus polymerase activity using a viral replicon system [52].However, our SeV and MuV genome replication data suggests SLC35A2 has no effect on viral replication and transcription during infection with these viruses.
Sialic acids are well studied in influenza virus infections since sialic acids on cell surface glycoproteins and glycolipids serve as receptors for the influenza virus.Loss of SLC35A1 causes reduced or abolished levels of sialylation on the cell surface, resulting in a severe impairment of influenza virus docking and entry [40,52].The role of SLC35A2, in contrast, is less clear.One study showed that absence of SLC35A2 abolishes influenza H1N1 replication [53] but another study showed no effect on influenza H7N9 virus binding and internalization [52].As we have confirmed that SLC35A2 affects paramyxovirus virus-cell and cell-to-cell fusion, there is a possibility that this protein also affects fusion processes in infections with influenza and other viruses.
Interestingly, SeV could bind to SLC35A2 KO cells and both NDV and MuV infected SLC35A2 KO cells better than SLC35A1 KO cells.These observations coupled with data from lectin staining showing that while SLC35A2 KO deleted all surface galactose it only decreased cell surface sialic acid expression levels, suggest that while sialic acid is essential for the binding of these viruses to cells, galactose is not, indicating these viruses may utilize alternative sialic acid receptors beyond the classic sialic acid linked to galactose.Articles describing sialic acids as a receptor for viruses focus on sialic acid linked to galactose [4,12,13,40,[54][55][56][57][58][59] but alternative linkages, such as α2,6-linkage to N-acetylgalactosamine are reported [4,60].Based on the data reported here, we hypothesize that either α2,6-linkage to N-acetylgalactosamine or other sialic acid linked molecules contribute to virus entry into A549 cells.Although further research is needed, our work provides new insights to the field.Syncytia formation occurs late in the MuV life cycle when the two envelope proteins expressed on infected cells mediate fusion with neighboring cells [61].Viral components can be transmitted between the fused infected cells through syncytia.Here, we observed that the MuV F mRNA level in SLC35A2 KO cells was significantly lower at low MOI.Additionally, the amount of infectious viral particles produced in both cell lines was the same, suggesting that the lower level of viral mRNA is likely related to the reduced syncytia formation in SLC35A2 KO cells.This finding suggests that SLC35A2 promotes MuV cell-to-cell transmission through MuV induced syncytia formation.
In this work, we used the reporter virus rSeVC dseGFP for screening.Compared to GFP, dseGFP, with a half-life of only two hours [36,37], has significant advantages in CRISPR screening as it better reflects real-time viral replication activity.The identification of SLC35A1 and SLC35A2 in the dsGFP negative cell population indicates that our screening method is very effective.Here, we only show host factors identified from dseGFP negative cell population.Future work will focus on dseGFP low and high cell populations, where we aim to identify polymeraserelated host factors, including those involved in non-standard viral genome generation.
In conclusion, our CRISPR-Cas9 KO screen identified SLC35A1 and SLC35A2 as critical host factors for paramyxovirus infection.SLC35A1 is essential for viral binding, while SLC35A2 is crucial for SeV-cell fusion and implicated in MuV-induced cell-to-cell fusion.Our findings show that even without SLC35A2, SeV can bind, NDV can enter and express genes, and MuV can complete its life cycle, indicating that galactose is not essential for viral attachment.This suggests the existence of alternative sialic acid receptors.Also, the reduced syncytia formation in SLC35A2 knockout cells at low MOI suggests that SLC35A2 deletion may inhibit MuV cell-to-cell transmission.These insights highlight the potential of targeting SLC35A2 for therapeutic interventions against paramyxovirus infections.

Genetically modified cell lines
Lentiviruses for transduction were generated by co-transfecting a plasmid expressing the gene of interest or sgRNA together with the lentivirus packaging plasmids psPAX2 (Addgene, #12260) and pMD2.G (Addgene, #12259) into Lenti-293T cells using a TransIT-Lenti Transfection Reagent (Mirus Bio, # MIR 6604).Supernatants were collected 48 hours post transfection and lentivirus were detected by a Lenti-X™ GoStix™ Plusx kit (TaKaRa, #631280).A549 cells were then transduced with 500ul supernatants containing the lentivirus and 8 ug/ml polybrene (1200rpm, 30℃, 2 hours).The cells were transferred to 6 well plates the second day followed by antibiotic selection (Blasticidin 10 ug/ml for 1 week, Puromycin 0.5 ug/ml for 1 week, Hygromycin 400ug/ml for 2 weeks).Surviving cells were single cell cloned and confirmed by western blot or lectin staining.

Recombinant reporter viruses rescue
The reporter viruses rSeVC eGFP and rSeVC dseGFP were rescued using the SeV Cantell strain reverse genetic system as described before [64].First, two full-length plasmids pSL1180-rSeV-C eGFP and pSL1180-rSeV-C dseGFP were made by replacing the miRF670 gene of pSL1180-rSeV-C miRF670 with an eGFP or a destabilized eGFP (dseGFP) gene [36].Additional nucleotides were inserted downstream of the dsGFP gene to ensure that the entire genome followed the "rule of six".The viruses were rescued by co-transfecting full-length plasmids and the three helper plasmids to BSR-T7/5 cells using Lipofectamine LTX with Plus Reagent (Invitrogen, #15338100).
The expression of GFP or dsGFP was monitored daily using fluorescence microscopy.At 4 days post-transfection, the cell cultures were harvested, and the supernatants were used to infect 10day-old specific-pathogen-free embryonated chicken eggs via the allantoic cavity after repeated freeze-thaw cycles.After incubation for 40 hours at 37℃, 40-70% humidity, the allantoic fluids were harvested and the TCID 50 was measured using LLC-MK2 cells.
The reporter virus rSeVC GFP-∆FHN+GFtail virus was generated and rescued by replacing SeV F and HN gene with a VSV G gene while retaining the SeV F protein tail.In brief, a VSV-GFtail plasmid was made by replacing G tail (G protein 490-511aa) with SeV F tail (F protein 524-565aa) using plasmid pMD2.G (Addgene, #12259), then the pSL1180-rSeVC GFP-∆FHN+GFtail full-length plasmid was made by cloning GFtail and deleting SeV F and HN gene from pSL1180-rSeVC eGFP through PCR and In-Fusion cloning (TaKaRA Bio, #638948).Virus was rescued as described above, and after five serial passages on A549-SLC35A1 KO cells, virus titer increased from 10 2 to above 10 7 TCID 50 /mL.Sanger sequencing confirmed the rescued virus sequence but revealed a D99G mutation within the M protein.

A549-Brunello CRISPR KO library
A549-Cas9 stable cells were transduced at a low MOI (~0.3) with Human CRISPR Knockout Pooled Library Brunello (Addgene, #73178) [35].Transduction conditions and antibiotic concentration were optimized for the A549-Cas9 stable cell line (cell seeding density: 8 x 10 4 per well (6-well plate); Puromycin concentraction: 0.5ug/ml.Lentivirus library was tittered to achieve a 30-50% infection rate and transductions were performed with 1.35x10 8 cells to achieve a representation of at least 500 cells per sgRNA per replicate.Puromycin was added 2 days post transduction and was maintained for 5 -7 days.The library cells containing sgRNA were used for CRISPR screening.Throughout the screen, the cells numbers were maintained at over 4x10 7 cells to ensure coverage of at least 500 cells per sgRNA.1).After demultiplexing according to the barcode sequences, the genes enrichment between cell populations was analyzed by MAGeCK (Version 0.5.9).The output tables were loaded and visualized with Prism 10.

Lectin staining
Cells were seeded at a confluency of 1 × 10 5 cells/well in a 12-well plate a day prior to staining.
The next day, the cells were washed twice with PBS and fixed using 2% PFA at RT for 15 minutes.

RNA extraction and RT-qPCR
Total RNA of infected cells and control samples were extracted using Kingfisher and a MagMAX™ mirVana™ Total RNA Isolation Kit (Thermo Fisher, #A27828) following manufacturer's guidelines.300-500ng of total RNA was used for cDNA synthesis with high-capacity RNA to cDNA kit (Thermo Fisher, #18080051).qPCR was performed using SYBR green (Thermo Fisher, # S7564) and 5 μM of reverse and forward primers (Table 2) for SeV NP, NDV NP, MuV F, and VSV NP genes on an Applied Biosystems QuantStudio 5 machine.Primers used for qPCR are listed in Table 2. Relative copy numbers were normalized to human GAPDH and human β -actin expression as described previously [69].

SLC35A1 and SLC35A2 complementation
To complement SLC35A1 or SlC35A1 to KO cells, KO cell lines were transduced with the SLC35A1-GFP or SLC35A2-GFP cDNA expression constructs.In brief, SLC35A1-GFP and SLC35A2-GFP were obtained from plasmids pEGFP.N3-SLC35A1-GFP and pEGFP.N3-SLC35A2-GFP (Addgene 186281 and Addgene 186284) [74].We performed codon optimization for the sgRNA binding sites of the target genes and switched to pLenti-Hygro plasmid backbone pLenti CMV Hygro DEST (Addgene, 17454) for lentivirus packaging.Then, SLC35A1-GFP and SLC35A2-GFP were separately introduced into their KO cell lines using the lentiviral transduction system.Following hygromycin selection, cell lines expressing the complemented genes were obtained, namely A549-SLC35A1 KO+A1-GFP and A549-SLC35A2 KO+A2-GFP.Subsequently, we infected these two cell lines with SeV reporter virus rSeVC miRF670 at a MOI of 3 and observed viral replication using fluorescence microscopy.

NDV induced cell death quantification
To quantitatively analyze NDV-induced cell death, cells infected with rNDV-eGFP were collected at 24 and 48 hpi and analyzed by Cytek flow cytometry.In brief, to collect all the cells including cell debris and dead floating cells, debris and dead floating cells in supernatant were collected by spin down.The attached cells were collected after trypsinization.Then cells from each condition were merged and stained with eBioscience™ Fixable Viability Dye eFluor™ 506 (ThermoFosher, #65-0866-14) in 1:400 dilution on ice for 10 minutes.Cells are fixed with 2% PFA for 10 minutes at room temperature followed by flow analysis.

MuV immunofluorescence
The infected cells were fixed using 2% PFA at RT for 15 minutes at specific time points postinfection followed by permeabilizing with 0.2% Triton X-100 (Sigma-Aldrich, # X100) for 10 minutes.Anti-MuV NP antibody (Thermo Fisher, #6008) was diluted in PBS at a 1:500 dilution and incubated at RT for 1 hour.Secondary antibodiy was diluted in PBS at a 1:500 dilution and incubated at RT for 30 minutes.The nuclei were stained with a 1:100,000 dilution of Hoechst 33342 (Invitrogen, # H-3570) along with the secondary antibody.

Sendai virus binding experiment
To determine the binding capability of the Sendai virus to KO cells, we performed a virus-cell binding assay.Briefly, KO cells or control cells were incubated with the virus at a MOI of 30 at 4℃ for one hour.Following this, the cells were fixed with 4% paraformaldehyde (PFA) (Fisher Scientific, #50-980-495) for 10 minutes, blocked with 3% BSA for 30 minutes, and then stained with a HN Monoclonal Antibody-Alexa Fluor™ 647 (Thermo, # 51-6494-82) for 30 minutes.Finally, the results were analyzed using Cytek flow cytometry.During the procedure, cells were washed in PBS supplemented with 2% BSA and 2 mM EDTA (Corning, # 46-034-CI) 3 times between each step.

Sendai virus Fusion experiment
To detect whether SeV underwent membrane fusion with the cell membrane, we referred to the Ebola membrane fusion experiment [42].In brief, control cells or KO cells were incubated with 400 MOI rSeV-M-HA at 37℃ for 3 hours, then treated with 0.5mg/ml proteinase K (NEB, #P8107S) at 37℃ for 70 minutes to remove virus bound to the cell surface but not fused with the cells.Next, the samples were fixed and permeabilized with eBioscience™ Foxp3 / Transcription Factor Staining Buffer Set (Invitrogen, #00552300).The intracellular M-HA protein was then detected by a HA-PE mAb (Biolegend, #901518).SeV HN was detected using a HN Monoclonal Antibody-Alexa Fluor™ 647 (Thermo, #51-6494-82) as a control for cell surface proteins.Finally, the results were analyzed by Cytek flow cytometry.The detection of M protein in the proteinase K-treated group indicates that the virus particles have fused with the cells.

MuV infectious particles measurement by TCID 50
96-well plates were prepared the day before the experiment by seeding 20,000 A549wt cells per well the day before titration.The collected samples were serially diluted 10-fold in infection media, ranging from 1:10 to 1:10 8 .The plated cells were washed once with PBS, and 100 µL of each dilution was added to the cells, with each sample tested in triplicate.After incubating in a 37℃ incubator for 4 days, the cells were stained using the immunofluorescence method described above.The TCID 50 /ml was calculated based on the fluorescence results.

MuV-induced syncytia quantification
After immunofluorescence staining, 3 images of each sample were captured using an inverted fluorescence microscope at both 20x and 5x magnifications.The 5x images were used for quantifying the MuV-induced syncytia.Fiji software was used to count the total number of nuclei in each 20x image, and the number of syncytia was manually marked and counted.Syncytia were defined as MuV-NP positive giant cells containing more than three nuclei.The number of syncytia per 1000 cells was then calculated.The results included three biological replicates.

Statistics
Statistics were calculated using GraphPad Prism Version 10 (GraphPad Software, San Diego, CA).

Fig 1 .
Fig 1. CRISPR screen workflow and sgRNA enrichment analysis.(A) Summary of the (Fig 2C).As shown in Fig 2D, SLC35A1 KO cells lack cell surface sialic acid while having more exposed galactose [41].SLC35A2 KO cells lack galactose in the cell surface and as expected since most of terminal sialic acid are linked to galactoses, have a significantly reduced level of sialic acid.

Fig 2 .
Fig 2. Roles of SLC35A1 and SLC35A2 and lectin staining analysis in A549 KO cells.(A)

Fig 4 .
Fig 4. Impact of SLC35A1 and SLC35A2 single KOs or double KOs in infection with different

A549-
Cas9 single cell clones were transduced with pXPR_011 lentivirus at an MOI of ~1.0 in 12 well plates.pXPR_011 plasmid was a gift from John Doench & David Root[34] (Addgene, #59702).Transduced cells were transferred to 6 well plates on day 3 post transduction and treated with puromycin.On day 9 post transduction, surviving cells from each single cell clone were collected and eGFP signal was detected by spectral flow cytometry.Active Cas9-expressing lines resulted in a reduction of eGFP when transduced with pXPR_011 as this vector delivers both eGFP and a sgRNA targeting eGFP.Because eGFP is linked to puromycin gene with a 2A site, abrogation of eGFP will have no impact on puromycin resistance.The lower eGFP percentage of a single cell clone indicates higher Cas9 activity.
A549-Brunello library was infected by SeV reporter virus rSeVC dseGFP nsVG negative stock at an MOI of 10 or rSeVC dseGFP nsVG positive stock at a MOI of 3 and cells were harvested followed by cell sorting to isolate the eGFP negative cell population.After screening, the sorted cells and two aliquots of the original CRISPR library (Mock) were pelleted and frozen at -80℃.Genomic DNA (gDNA) was isolated using the QIAamp DNA Blood Midi (Qiagen, #51183) or QIAamp DNA Bloop Mini (Qiagen, #51104) kit according to the manufacturer's instructions.The concentration of each gDNA was measured on a Qubit Fluorimeter with the Qubit dsDNA Quantification Assay Kit the manufacturer's instructions.The final pool was submitted to the DNA Sequencing Innovation Lab (Washington University School of Medicine) for sequencing on the Illumina NextSeq-Mid platform with a 15% spike-in of PhiX DNA, yielding a total of 109,525,309 reads (Table