Person:

Singh, Parmit

Allosteric integrase (IN) inhibitors (ALLINIs), which are promising preclinical compounds that engage the lens epithelium-derived growth factor (LEDGF)/p75 binding site on IN, can inhibit different aspects of human immunodeficiency virus 1 (HIV-1) replication. During the late phase of replication, ALLINIs induce aberrant IN hyper-multimerization, the consequences of which disrupt IN binding to genomic RNA and virus particle morphogenesis. During the early phase of infection, ALLINIs can suppress HIV-1 integration into host genes, which is also observed in LEDGF/p75- depelted cells. Despite this similarity, the roles of LEDGF/p75 and its paralog hepatoma-derived growth factor like 2 (HDGFL2) in ALLINI-mediated integration retargeting are untested. Herein, we mapped integration sites in cells knocked out for LEDGF/p75, HDGFL2, or both factors, which revealed that these two proteins in large part account for ALLINI-mediated integration retargeting during the early phase of infection. We also determined that ALLINI-treated viruses are defective during the subsequent round of infection for integration into genes associated with speckle-associated domains, which are naturally highly targeted for HIV-1 integration. Class II IN mutant viruses with alterations distal from the LEDGF/p75 binding site moreover shared this integration retargeting phenotype. Altogether, our findings help to inform the molecular bases and consequences of ALLINI action.

Integration of HIV-1 into genes depends on multiple factors including transcription, splicing, LEDGF/p75, and CPSF6. But, the precise role of mRNA splicing in HIV-1 integration targeting is unclear. Transcriptional intron content can show high rate of splicing due to high transcription rate, whereas genes with higher number of introns can show high rate of splicing due to intron number, despite comparatively low transcription rate. To differentiate the roles of splicing and transcription in HIV-1 integration, we compared gene populations targeted by HIV-1 to random gene populations generated in silico (RIC, for random integration control). Analyzing gene groups with same numbers of introns, we determined that genes containing >10 introns were preferentially enriched for HIV-1 integration (ratio >1; R2 =0.9). We then parsed the human transcriptome into genes containing 0-10 introns (71% of all genes) and genes with minimally 11 introns (29% of all genes). Remarkably, we found that genes with >10 introns harbored 57% of all HIV-1 integration sites compare to 35% of sites in genes with <11 introns. This preference was independent of gene length: length-normalized genes with <11 introns were targeted 1.3 times greater than random whereas genes with >10 introns were targeted 2.1 times over the RIC (P < e-10). To assess the roles of virus-host interactions, we analyzed integration sites from LEDGF knockout (LKO) and CPSF6 knockout (CKO) cell lines. For genes with >10 introns, integration was similarly reduced by 20-21% in both cell types. Integration into genes with <11 introns was by contrast reduced by 2% and 7.6% in LKO and in CKO cells, respectively, indicating that LEDGF/p75 preferentially directs integration to genes with >10 introns. Consistent with this interpretation, analysis of a recently published LEDGF/p75 ChIP-Seq dataset revealed a strong correlation between number of genic LEDGF/p75 interaction sites and intron number (R2 =0.7; P < e-5), with 14 as the average number of introns for genes with minimally one LEDGF/p75 binding site. In LKO cells, integration into genes with LEDGF/p75 interaction sites was reduced by 14% (12% in genes with >10 introns), while integration into genes that lacked LEDGF/p75 interaction sites was reduced by 6% and 21% in LKO and CKO cells, respectively. Moreover, integration into genes with <11 introns that lacked LEDGF/p75 interaction sites was reduced in CKO but not in LKO cells. Our results reveal a previously unrecognized role for LEDGF/p75 in mediating HIV-1 integration into highly spliced genes that are expressed at comparatively low levels.

Selection of a suitable chromatin environment during retroviral integration is a tightly regulated and multilayered process that involves interplay between viral and host factors. However, whether intrinsic chromatin dynamics during mitosis modulate retroviral genome invasion is currently poorly described. Direct interaction between the spumaretrovirus prototype foamy virus (PFV) Gag protein and cellular chromatin has been described as a major determinant for integration site selection. A previous Gag chromatin-binding site (CBS)–nucleosome co-crystal structure revealed an interaction with the histone H2A-H2B acidic patch via a highly conserved arginine anchor residue. Yet, the molecular mechanisms regulating Gag-chromatin capture during PFV infection remain obscure. Here, we investigated the kinetics of Gag-chromatin interactions during mitosis and proviral integration of PFV-infected synchronized cells. Using Gag CBS variant viruses, we showed that alteration of Gag affinity for nucleosome binding induced untimely chromatin tethering during mitosis, decreased infectivity and redistributed viral integration sites to markers associated with late replication timing of host chromosomes. Mutant Gag proteins were moreover defective in their ability to displace the histone H4 tail from the nucleosome acidic patch of highly condensed mitotic chromatin. These data indicate that the mitotic chromatin landscape during Gag–nucleosome interactions hosts PFV integration site selection determinants and that spumaretroviruses evolved high- affinity chromatin binding to overcome early mitosis chromatin condensation for optimal viral DNA tethering, integration and infection.

Pre-mRNA splicing regulates promoter-proximal Pol II pausing and alternative polyadenylation (APA). Splicing inhibitors increase pausing and use of proximal polyadenylation sites (PAS) for comparatively long (>4 introns) transcripts by impairing P-TEFb recruitment, which is a core component of the super elongation complex (SEC). Similar to splicing, the CFIm complex composed of CPSF6 and CPSF5 promotes the use of distal PAS. CPSF6 binds capsid (CA) to target integration into highly spliced genes and gene-dense (GD) regions. We previously reported that HIV-1 preferentially integrates into genes harboring >10 introns and Pol II-paused genes. Based on the connection between intron number, pausing, and APA, we hypothesized that CPSF6 targets HIV-1 integration into paused genes regulated by U2 snRNP and P-TEFb. U2 snRNP-regulated APA genes, which represent 5% of the transcribed genome, contained 24% of integration sites (3x compared to RIC or random integration control; p<1E-5). In contrast, nonregulated genes were targeted similarly to all genes (p< 0.2). Moreover, we found that paused genes regulated by PAF-1, which is important for APA, were preferentially targeted (3.5x RIC; p<1E-5), whereas the reciprocal gene set was preferentially avoided (p<1E-5). To test the role of splicing, we mapped integration sites in Jurkat T cells in the presence of the U2 snRNP inhibitor Pladenolide B (Plad B) or SEC inhibitor KL-2. Plad B significantly reduced genic integration in PAF-1 paused genes but not in unpaused genes. We defined chromosomes with reduced genic integrations as Plad B sensitive chromosomes (PBSC) and the remaining chromosomes as Plad B insensitive chromosomes (PBIC). KL-2 reduced genic integration significantly for PBSC but not for PBIC, suggesting that splicing targets HIV-1 into genes regulated by P-TEFb. To test the roles of integration targeting cofactors, we mapped sites for CPSF6- defective CA mutant viruses or WT HIV-1 in LEDGF/p75 knockout (LKO) cells. Despite predictably similar levels of splicing, PBSC (avg. intron content 9.6 and GD 11.8/Mb) were targeted 2.6x to RIC, whereas PBIC (avg. intron 9.5 and 6.6 genes/Mb) were but 1.2x. PBSC supported significantly less genic integration for CA mutants and for WT virus in LKO cells (p<1E-7). However, while PBIC were significantly less targeted by WT virus in LKO cells, these genes were significantly more targeted by CA mutants (p<1E-7 for both comparisons). Thus, CPSF6-CA interactions preferentially target HIV-1 integration into paused genes regulated by the SEC and U2 snRNP. Acknowledgment: HUCFAR 5P30AI060354-18.

HIV-1 integration targeting is mediated by interactions of viral integrase and capsid with LEDGF/p75 and CPSF6, respectively. HIV-1 also prefers to integrate into highly spliced genes, but the role of pre-mRNA splicing in HIV-1 integration is unclear. Splicing inhibitors reduce the association of exon junction complex (EJC) with genes containing introns but not with intronless genes. The EJC causes promoter-proximal Pol II pausing, which is released by super elongation complex (SEC). To assess the role of Pol II pausing in integration, genes were stratified as paused versus unpaused, which revealed 50%/40% respective HIV-1 integration targeting preferences in HEK293T cells; random paused and unpaused values were 15%/34%. In LEDGF/p75 knockout (LKO), CPSF6 knockout (CKO), and double knockout (DKO) cells, unpaused gene targeting dropped marginally from 40% to 37%, 38%, and 35%, respectively. However, in paused genes, 50% targeting reduced to 31%, 24%, and 20% in respective LKO, CKO and DKO cells. Thus, HIV-1 prefers integration into Pol II-paused genes. To test the role of splicing, we mapped integration sites in the presence of splicing inhibitor Pladienolide B (PladB) in WT and LKO Jurkat T cells. Genic integration in WT cells was reduced in a dose-dependent manner, with maximum reduction of 2.7% (P <2E-31) at 9 nM. In LKO cells, genic integration was reduced by 3.6% (P <4E-13), but only at 9 nM inhibitor, suggesting both LEDGF/p75-dependent and independent roles for pre- mRNA splicing in integration targeting. Similar to HEK293T cells, integration in paused and unpaused genes was 55% and 38% in WT Jurkat cells, while in LKO cells, these respective values were 25% and 35%. PladB significantly reduced integration into paused genes in both WT and LKO, but only at 9 nM in LKO cells. By contrast, whereas PladB significantly increased integration into unpaused genes (1.2% at 9 nM; P <2E-16) in WT cells, it had no effect in LKO cells. To rule out the role of pausing itself, we assessed the SEC inhibitor KL-2 in HEK293T cells, which revealed significant reductions in the amount of integration in genes, paused genes, and speckle-associated domains. Yet, integration into unpaused genes was unaffected. Thus, our results reveal an interplay between pre-mRNA splicing with Pol II pausing and transcriptional elongation in HIV-1 integration targeting. Acknowledgment: Harvard University Center for AIDS Research (HU CFAR NIH/NIAID fund 5P30AI060354-17).

HIV-1 integrase and capsid proteins interact with host proteins to direct preintegration complexes to active transcription units within gene-dense regions of chromosomes for viral DNA integration. Analyses of spatially-derived genomic DNA coordinates, such as nuclear speckle-associated domains, lamina-associated domains, super enhancers, and Spatial Position Inference of the Nuclear (SPIN) genome states, have further informed the mechanisms of HIV-1 integration targeting. Critically, however, these different types of genomic coordinates have not been systematically analyzed to synthesize a concise description of the regions of chromatin that HIV-1 prefers for integration. To address this informational gap, we have extensively correlated genomic DNA coordinates of HIV-1 integration targeting preferences. We demonstrate that nuclear speckle-associated and speckle-proximal chromatin are highly predictive markers of integration and that these regions account for known HIV biases for gene-dense regions, highly transcribed genes, as well as the mid-regions of gene bodies. In contrast to a prior report that intronless genes were poorly targeted for integration, we find that intronless genes in proximity to nuclear speckles are more highly targeted than are spatially-matched intron-containing genes. Our results additionally highlight the contributions of capsid and integrase interactions with respective CPSF6 and LEDGF/p75 host factors in these HIV-1 integration targeting preferences.

Selection of a suitable chromatin environment during retroviral integration is a tightly regulated process. Most retroviruses, including spumaretroviruses, require mitosis for nuclear entry. However, whether intrinsic chromatin dynamics during mitosis modulates retroviral genome invasion is unknown. Previous work uncovered critical interactions of prototype foamy virus (PFV) Gag with nucleosomes via a highly conserved arginine anchor residue. Yet, the regulation of Gag-chromatin interaction and its functional consequences for spumaretrovirus biology remain obscure. Here, we investigated the kinetics of chromatin binding by Gag during mitosis and proviral integration in synchronized cells. We showed that alteration of Gag affinity for nucleosome binding induced untimely chromatin tethering during mitosis, decreased infectivity, and redistributed viral integration sites to markers associated with late replication timing of chromosomes. Mutant Gag proteins were, moreover, defective in their ability to displace the histone H4 tail from the nucleosome acidic patch of highly condensed chromatin. These data indicate that the chromatin landscape during Gag-nucleosome interactions is important for PFV integration site selection and that spumaretroviruses evolved high-affinity chromatin binding to overcome early mitosis chromatin condensation.

This presentation examines the correlation between R-loops and HIV-1 integration, highlighting experimental results generated in our study. The work was presented at the Behavior of HIV in Viral Environments (B-HIVE) Center Face-to-Face Meeting.