A virus-packageable CRISPR screen identifies host factors mediating interferon inhibition of HIV

Interferon (IFN) inhibits HIV replication by inducing antiviral effectors. To comprehensively identify IFN-induced HIV restriction factors, we assembled a CRISPR sgRNA library of Interferon Stimulated Genes (ISGs) into a modified lentiviral vector that allows for packaging of sgRNA-encoding genomes in trans into budding HIV-1 particles. We observed that knockout of Zinc Antiviral Protein (ZAP) improved the performance of the screen due to ZAP-mediated inhibition of the vector. A small panel of IFN-induced HIV restriction factors, including MxB, IFITM1, Tetherin/BST2 and TRIM5alpha together explain the inhibitory effects of IFN on the CXCR4-tropic HIV-1 strain, HIV-1LAI, in THP-1 cells. A second screen with a CCR5-tropic primary strain, HIV-1Q23.BG505, described an overlapping, but non-identical, panel of restriction factors. Further, this screen also identifies HIV dependency factors. The ability of IFN-induced restriction factors to inhibit HIV strains to replicate in human cells suggests that these human restriction factors are incompletely antagonized. Editorial note: This article has been through an editorial process in which the authors decide how to respond to the issues raised during peer review. The Reviewing Editor's assessment is that all the issues have been addressed (see decision letter).

. We identify the type I IFN pathway genes, STAT1, IFNAR1, STAT2 142 and IRF9 as the highest-scoring hits (magenta in Figure 1E). Therefore, the PIKAHIV screen functions as 143 designed: cells in which IFN signaling is compromised exhibit increased viral production and, therefore, 144 enriched HIV-CRISPR representation of sgRNAs in the secreted HIV virions. After the IFN pathway 145 genes, the Zinc Antiviral Protein (ZAP) and its modifier TRIM25 were the next to highest scoring hits.

146
ZAP is an antiviral effector that has potent activity against alphaviruses as well as moderate activity   Figure S2A and S2B). Therefore, we reasoned 160 that the true IFN-induced restriction factors that potently inhibit HIV in THP-1 cells were not identified in 161 this initial screen. Analysis of the CG dinucleotide content across the HIV-CRISPR genome shows high 162 levels of CG dinucleotides, particularly in the Cas9 and Puromycin resistance ORFs, that are potential 163 targets for ZAP-mediated RNA degradation (Figure 2A). Given its role in degradation of RNA with high 164 CG content, we hypothesized that ZAP could inhibit the full-length HIV-CRISPR genomic RNA that is 165 packaged into budding virions rather than the wt HIV genome. Thus, we determined whether or not ZAP 166 KO allows for increased packaging of the HIV-CRISPR vector in viral particles released from cells by 167 8 measuring both wild type HIV-1LAI genomes (HIV-Pol; black in Figure 2B) and HIV-CRISPR genomes 168 (cPPT-U6; gray in Figure 2B) with a ddPCR assay. Indeed, we find enhanced packaging of HIV-CRISPR 169 genomes relative to wild type HIV-1LAI genomes in the viral supernatant in cell populations with reduced 170 ZAP expression ( Figure 2B -10.5% in wt THP-1 cells; 24.8% and 31.6% in the ZAP-KO clonal lines).

171
Therefore, to circumvent the inhibitory effects of ZAP on the HIV-CRISPR vector, we repeated the 172 PIKAHIV screen in two ZAP-KO THP-1 clonal cell lines. As expected for a screen in ZAP-KO cells, ZAP is 173 no longer a significantly-scoring hit in the screen (rank # 1647/3812 in combined ZAP-KO screen data;

174
Supplemental Table S4). In addition, there is also no enrichment of N4BP1 or TRIM25 in the ZAP-KO 175 screens (rank # 3789/3812 and 3090/3812 in combined ZAP-KO screen data; Supplemental

178
To ask if ZAP knockout improves performance of the HIV-CRISPR screen, we analyzed read 179 counts across duplicates in the two independent ZAP-KO THP-1 clonal lines and compared the results to

187
By multiplying gene scores from both ZAP-KO screens ( Figure 2E; MAGeCK score on x-axis) we 188 identify a list of candidate hits. To ask which genes are most likely to contribute specifically to the IFN-189 mediated inhibition of HIV-1, we calculated the level of IFN induction of each of the top hits from an 190 existing THP-1 microarray dataset ( Figure 2E; IFN log2FC on y-axis and Supplemental Table S4). No hit 191 scored as highly as the type I IFN pathway genes (magenta in Figure 2E). Therefore, multiple genes, 192 rather than a single ISG, are responsible for the IFN-mediated inhibition of HIV infection in THP-1 cells.

234
We were surprised to find TRIM5 and Tetherin in this screen as HIV-1 is thought to be highly-

242
IFITM1 also contribute to IFNα inhibition ( Figure 4B). We find no effect of Tetherin KO on early steps of 243 HIV replication as expected given its role as a late-acting HIV restriction factor (Tetherin_1 = 6-fold, 244 Tetherin_2 = 6.4-fold in Figure 4B

248
(compare controls to MxB-KOs in Figure 4D), there is still a 4.8-fold inhibition of virus released from MxB-

249
KO clonal lines (magenta in Figure 4D). Since Tetherin is a well-characterized late-acting restriction 250 factor and was also a hit in our PIKAHIV screen, we asked if Tetherin is responsible for the late ISG block

261
The HIV-CRISPR screen also identifies HIV dependency factors

262
Although we designed our screen specifically to find IFN-induced factors restricting HIV-1 in THP-         Figure S2). Similar to the effect of ZAP, N4BP1 (Nedd4-binding protein 1) also has a 388 modest effect on HIV replication both after IFN pretreatment and when constitutively-expressed 389 (Supplemental Figure S2). In our screen, the anti-lentiviral function of N4BP1 appears to be genetically 390 linked to ZAP activity as N4BP1 is no longer a hit in the ZAP-KO screen (Supplemental Table S4).

391
Therefore, N4BP1 may modify or enhance ZAP-mediated antiviral activity similar to the modification of

425
In summary, we developed a novel screen that is highly sensitive to detect restriction factors for

574
sgRNA sequences present in the genomic DNA and viral supernatants were amplified by PCR and RT-575 PCR, respectively, using primers specific for the HIV-CRISPR construct (Supplemental Table S5) 576 (Toledo et al., 2015). Libraries were then barcoded/prepared for Illumina sequencing by a second round 577 of PCR (Supplemental Table S5). Each amplicon was then cleaned up through double-sided SPRI