Fig 2: In vitro adaptation of the early HIV-1 CTL escape variant RKLM selects for a CPSF6-resistant virus.(A) CEM cells were infected with replication-competent GFP reporter viruses for spreading infections. (B) HeLa cells were infected with VSV-G-pseudotyped GFP reporter viruses in the presence or absence of aphidicolin (Aph). Both double mutants were different from RKLM in their ability to infect aphidicolin-treated cells (p<0.02). (C) HeLa cells transfected with siRNA against CPSF6, CypA or a combination of the two were infected with VSV-G-pseudotyped GFP reporter viruses. RKLM, but not the two double mutants, was significantly rescued by either CPSF6 or CypA depletion (p<0.01). Results are one representative experiment of at least three independent experiments, except for the double knockdown experiment in C. Standard deviations of a single triplicate experiment are indicated with the error bars.

Fig 3: In vitro-derived CPSF6-resistant virus may alter CPSF6 binding.(A) HeLa cells stably transduced with the control vector LPCX or one overexpressing CPSF6-358 were infected with VSV-G-pseudotyped GFP reporter viruses. Results are one representative experiment of at least three independent experiments. (B) WT, RKLM and RKLM+T107I CA tubes were incubated with HeLa cell lysates as described in the Materials and Methods section. Complexes were pelleted and analyzed by quantitative immunoblotting for CPSF6 and CA. Input represents 10% of each reaction withdrawn prior to pelleting. The three individual panels shown for each protein are from one scan of a single immunoblot and were not manipulated. (C) Average of non-normalized data from three independent experiments that utilize HIV-1 CA tubes. One representative result is shown in B. Values represent the ratio of CPSF6 to CA in the pelleted CA complexes. The difference in CPSF6 binding between RKLM and RKLM+T107I was significant at lower concentrations of lysates. P values were determined by unpaired t-test. (D) HIV-1 CA-NC tubular complexes were incubated with cell lysates either from 293T cells transfected with HA-tagged rhesus TRIM5a (rhT5a) or HeLa cells for endogenous CPSF6. Ultracentrifugation through a 60% sucrose cushion was used to pellet HIV-1 CA-NC tubes and associated proteins. A small amount of the mixture was saved before ultracentrifugation and used as input. We have noted that HIV-1 CA-NC tubes carrying the RKLM mutations have reduced amounts of CA-NC in the pelleted fraction. At least two independent experiments were performed for CA binding to rhesus TRIM5a while five experiments were done for CA binding to CPSF6. One representative data for each experiment is shown here. (E) Summary of binding experiments using HIV-1 CA-NC tubes for three mutants. One representative result is shown in D. Binding of mutants carrying either RKLM+S41A or RKLM+T107T to endogenous CPSF6 is shown as relative to binding of CPSF6 to RKLM CA-NC. The level of CPSF6 in the pelleted fraction measured by LICOR was normalized by the level of CA-NC. The difference between RKLM and RKLM+S41A was not statistically significant (p = 0.71) whereas the difference between RKLM and RKLM+T107I was significant (p<0.01). The mean and standard error of the mean of five independent experiments are shown by horizontal bars.

Fig 4: CPSF6 and CypA coordinate to suppress early HIV-1 CTL escape mutants in HLA-B27+ subjects.(A) HeLa cells were infected with RT-normalized VSV-G-pseudotyped GFP reporter viruses in the presence or absence of aphidicolin (Aph). Viral infectivity by RKLM and RKLM+A105T was significantly lower in Aph-treated cells than in non-treated cells (p<0.05). (B) VSV-G-pseudotyped GFP reporter viruses were used to infect HeLa cells after siRNA knockdown of CPSF6. CPSF6 depletion significantly increased infectivity of RKLM (p = 0.002). RKLM was different from both double mutants in responding to CPSF6 depletion (p<0.002). (C) Activated primary CD4+ T cells were infected with RT-normalized VSV-G-pseudotyped GFP reporter viruses. Increase in viral infectivity of both double mutants relative to RKLM was statistically significant (p<0.01). (D) VSV-G-pseudotyped GFP reporter viruses were used to infect HeLa cells after siRNA knockdown of CPSF6. (E) RT-normalized VSV-G-pseudotyped GFP reporter viruses were used to infect HeLa cells in the presence or absence of CsA. Relative to CsA-untreated cells, CsA significantly increased infectivity of RKLM (p<0.01). Response of RKLM to CsA treatment differed from that of double mutants (p<0.001). (F) RT-normalized VSV-G-pseudotyped GFP reporter viruses were used to infect HeLa cells after siRNA knockdown of CypA. Infectivity of RKLM was higher in CypA-depleted cells than in control cells (p<0.01). RKLM differed from double mutants in its sensitivity to CypA depletion (p<0.001). (G) HeLa cells transfected with siRNA against CypA, CPSF6 or both simultaneously were infected with RT-normalized VSV-G-pseudotyped GFP reporter viruses. Results are one representative experiment of at least two experiments, except for A where the average of viral infectivity from four independent experiments was plotted with the standard error of the mean. Standard deviations of a single triplicate experiment are indicated with the error bars.

Fig 5: Endogenous CPSF6 restricts cell cycle-dependent HIV-1 CA mutants.(A, B) HeLa cells transfected with either non-targeting siRNA or siRNA targeting CPSF6 were infected with RT-normalized VSV-G-pseudotyped GFP reporter viruses. Increase in viral infectivity of T54A by CPSF6 depletion was statistically significant (B; p<0.002) and different from that of T54A+N74D or T54A+A105T (p<0.001). (C) HeLa cells transfected with either non-targeting siRNA or siRNA targeting CPSF6 were infected with VSV-G-pseudotyped GFP carrying CA mutants that lost the cell cycle independence so that 0.5%-1% cells were infected with the viruses two days after infection. (D) An experiment similar to that described in C was carried out, except that all these CA mutants still maintain the cell cycle independence. Results are one representative experiment of at least two experiments. Standard deviations of a single triplicate experiment are indicated with the error bars.