Quantitative phosphoproteomics reveals extensive cellular reprogramming during HIV-1 entry

Abstract

Receptor engagement by HIV-1 during host cell entry activates signaling pathways that can reprogram the cell for optimal viral replication. To obtain a global view of the signaling events induced during HIV-1 entry, we conducted a quantitative phosphoproteomics screen of primary human CD4(+) T cells after infection with an HIV-1 strain that engages the receptors CD4 and CXCR4. We quantified 1,757 phosphorylation sites with high stringency. The abundance of 239 phosphorylation sites from 175 genes, including several proteins in pathways known to be impacted by HIV-receptor binding, changed significantly within a minute after HIV-1 exposure. Several previously uncharacterized HIV-1 host factors were also identified and confirmed through RNAi depletion studies. Surprisingly, five serine/arginine-rich (SR) proteins involved in messenger RNA splicing, including the splicing factor SRm300 (SRRM2), were differentially phosophorylated. Mechanistic studies with SRRM2 suggest that HIV-1 modulates host cell alternative splicing machinery during entry in order to facilitate virus replication and release.

Figure 1. Phosphoproteomics workflow

(A) Phosphoproteomics workflow for T cell stimulations, processing, and fractionation. Virus and mock-treated samples were independently processed following mixing with a common reference lysate. HIV/mock phosphorylation site fold-changes were calculated by dividing the ratio of reference/mock by reference/HIV phosphorylation site fold-changes. Abbreviations: light (L), heavy (H), hydrophilicity interaction liquid chromatography (HILIC), immobilized metal affinity chromatography (IMAC), liquid chromatography tandem mass spectrometer (LC/MS/MS). (B) Primary CD4⁺T cells were incubated with purified HIV-1 or a mock vesicle preparation for various lengths of time at 37°C and probed for phosphorylated p38 (pT180/pY182) by immunoblot to determine optimal kinetics for large scale phosphoproteomics stimulations. (C) Distribution of phosphorylation site fold-changes. HIV/mock phosphorylation site fold-changes are shown according to the mixture model designation of HIV-responsive (red) or HIV-nonresponsive, i.e. null (blue). For clarity, only a subset of SRm300-responsive phosphorylation sites is indicated. See also Figure S1.

Figure 2. Immunoblot and bioinformatic validation of HIV-responsive phosphorylation site ratios

Primary CD4⁺ T cells were incubated with purified HIV-1 or a mock vesicle preparation and probed for phosphorylated p38 (pT180/pY182), ETS-1 (pS282/pS285), and ERK1/2 (pT202/pY204) by immunoblot. The chemokine CXCL12 (SDF-1) served as a positive control for pERK1/2. Odds ratio and p-values calculated with Fisher’s Exact with Benjamini and Hochberg corrections. (B) Putative kinase substrate motifs were searched for within up-, or down-regulated HIV-responsive phosphorylation. The amino acid position weighted matrix, centered on phosphorylated serine (position 0), is shown for each top-scoring motif. (C) Overrepresentation of the CaMKII and PKA motif within HIV up- and down-regulated phosphorylation sites was calculated with the Fisher’s Exact test with each p-value indicated above each bar. ϕ = L/V/I/M/Y/F, pS = phosphoserine. See also Table S4.

Figure 3. Bioinformatic analysis of genes from HIV-responsive phosphorylation sites

(A) Interaction network of HIV-responsive phosphoproteins. Functional or physical interactions (see methods) between genes from HIV-responsive phosphopsites were extracted from the STRING database (version 9.0). Underlined genes indicate SR-proteins. Genes indicated as ‘HIV-1 replication’ were taken from Table 1. (B) HIV-1 host factors are overrepresented within HIV-responsive phosphoproteins. Overlaps of genes from HIV-responsive phosphoproteins and genes from the HIV-1 human protein interaction database (HHPID), HIV-1 RNAi screens, and West Nile Virus (WNV) siRNA screen were calculated with the hypergeometric test using the human genome as background (n=20,402). See also Table S6.

Figure 4. RNAi-mediated knockdown of HIV-responsive phosphoproteins

(A) MAGI cells were transfected with individual siRNAs then infected with NL4.3 pseudovirions with an X4-tropic Env, R5-tropic Env, or VSV-G. Infection values (luciferase activity) were normalized to a non-targeting control siRNA. (B) Lysates from depleted MAGI cells were analyzed by immunoblot (top panel). SRm300 protein levels were monitored by FACS (bottom panel): grey area (non-targeting control), solid line (siRNA#1), dotted line (siRNA#2), short dashed line (siRNA#3), long dashed line (siRNA#4). (C) Jurkat cells were nucleofected with indicated pools of siRNA and analyzed by immunoblot (top panel). SRm300 protein levels were monitored by FACS (bottom panel): grey area (non-targeting control), solid line (SRm300 pool). (D) Nucleofected jurkat cells were infected as in (A). Bars=SEM, n=3. (E) Relative percent infection of SRm300-depleted MAGI cells with replication competent 89.6, NL4.3, and GFP-expressing VSV-G pseudovirions. Bars=SEM, n=4, *p<0.05 Mann-Whitney, ns: not significant (p>0.05).

Figure 5. SRRM2 regulates alternative splicing of HIV-1

(A) RT-PCR analysis of HIV-1 transcripts. Exon composition of sequence-confirmed transcripts are indicated to the right of a representative 89.6 DNA gel, positioned with respect to the HIV-1 genome above. D1-4: 5’ splice sites. A1-7: 3’splice sites. (B) Relative quantification of HIV-1 transcripts. Band intensities of corresponding RT-PCR products of tat and nef alternatively spliced products were divided by tat-1 and nef-2, respectively. Comparisons were made between each SRRM2 siRNA and the non-targeting control siRNA. (C) HIV-1 release following depletion of SRRM2 as measured by p24 collected on day 5 post infection. Comparisons were made between each SRRM2 siRNA and the non-targeting control siRNA. Bars=SEM, n=4, *p<0.05 Mann-Whitney, ns: not significant (p>0.05). See also Figure S2.

Figure 6. HIV-responsive SR-proteins regulate HIV-1 splicing and virion release

(A) Lysates from depleted MAGI cells were analyzed by immunoblot with each antibody. (B) Relative quantification of HIV-1 transcripts. Band intensities of tat and nef alternatively spliced products (top panel) were divided by tat-1 and nef-2, respectively for quantification (bottom panel). n=6 (C) HIV-1 release following depletion of SR proteins as measured by p24 collected on day 5 post infection. Comparisons were made between each SRRM2 siRNA and the non-targeting control siRNA. Bars=SEM, n=4, *p<0.05, **p<0.01 Mann-Whitney, ns: not significant (p>0.05). See also Figure S3.