Genes affecting viral entry were selectively identified with enhanced resolution with the CRE/floxed reporter over the lentiviral reporter.
We predicted that the known envelope receptors would be discovered by the screen with signal strength specificity of the selection. In other words, as the specificity of the molecular events is increased toward viral entry and fewer genes associated with other viral activities affect the selection, the frequency of sgRNAs targeting genes that influence entry will be enriched as there will be fewer competing gene sgRNA sequencing amplicons in the low-infection population. Thus, we expected that that the normalized Log fold change, a normalized frequency count and the primary statistical output of MaGeCK robust rank aggregation (RRA) score will increase as the pathway specificity increases, resulting in fewer genes affected (Li et al., 2014). Indeed, ACE2 was identified for Spike-mNG entry but not for VSVG-mNG or VSVG-CRE (Figure 3). Even though ACE2 expression in the ACE2-293T was driven from a virally integrated cDNA, the sgRNAs in the Brunello library were able to cut and disable the gene thereby preventing entry of Spike-mNG as expected. 293T cells without the ACE2 transgene were highly resistant to the reporter virus prior to expression of ACE2 (unpublished data, (Ou et al., 2020). We attempted to conduct the comparative screen using Spike-CRE, however we were unable to generate sufficient Spike-CRE reporter virus. In the case of VSVG-CRE, our screen identified the low-density lipoprotein receptor (LDLR, Figure 3C) which is a known receptor for VSV-G (Finkelshtein et al., 2013; Nikolic et al., 2018). The VSVG-mNG screen did not identify the LDLR (Figure 3B) with the LDLR producing a modes LFC (0.83) and weakly significant RRA score (0.0023) and rank of 205, which is trending toward importance for viral entry but much weaker than in the VSVG-CRE which had a LFC (1.32) and a highly significant RRA score (9.23*10-6) and is ranked 13th. We interpret this improvement in the detection of the LDLR within the VSVG-CRE reporter screen over the VSV-G-mNG screen due to the reduction in overall noise arising from the additional genes influencing retro transcription and viral integration present in the VSVG-mNG screen, but not in the VSVG-CRE screen.
To compare the effect of the reporter with the same envelope, we plotted the negative Log (RRA) of VSVG-mNG against negative Log (RRA) of VSVG-CRE for the top 15 ranked genes for loss of viral reporter expression in each screen (Figure 3D). Similarly, we plotted the Log (RRA) for VSVG-mNG against Spike-mNG to allow for comparison of the same reporter with different envelopes (Figure 3E). In these plots, genes strongly affecting both screens will appear in the upper righthand corner. In both plots, SLC35B2 appeared in the upper left indicating a role for facilitating viral entry (Figure 3D and 3E). This Golgi associated protein is a transporter for 3’-phosphoadenosine 5’-phosphosulfate and is required for sulfation of heparan sulfate proteoglycans and has been shown to contribute to entry for SARS-CoV-2, seasonal coronavirus, and a wide range of other viruses, likely by heparan sulfate mediated recruitment of viral particles to the cell surface (Clausen et al., 2020; Park et al., 2017; Schneider et al., 2021; Tandon et al., 2021; Wang et al., 2021; Q. Zhang et al., 2020).
The ER chaperone Gp96 (HSP90B1) was identified in VSVG-mNG and VSVG-CRE screens, but not in the Spike-mNG screen indicating that it is specific to VSV-G mediated entry (Figure 3). Prior work has shown that Gp96 facilitates the correct folding of proteins and is essential for the presence of functional LDLR at the cell surface and entry of VSV (Hastie et al., 2013). Its absence in our screen against Spike-mNG is consistent with it not being identified in other SARS-CoV-2 screens (Baggen et al., 2021; Daniloski et al., 2021; Schneider et al., 2021; Wang et al., 2021; Wei et al., 2021) and is consistent with HSP90B1 being a host factor associated with the cell surface expression of the LDLR and VSV-G mediated entry.