PFV-1 GAG interacts with AGO2 independently of miRNAs and P-bodies. (A) 293T cells were transfected with pc13 and myc-AGO2 vectors. IPs anti-PFV-1 proteins or anti-myc were performed 48 hpt. IP (pellet) and cell lysates (input) were analyzed by WB as indicated. RNase treatment was carried out before IP. (B) 293T cells were transfected with PFV-1 EGFP-GAG and myc-AGO2, myc-PAZ9 or pcDNA3 (mock) vectors. Anti-myc or anti-HA (used as a negative control), IPs were performed 48 hpt. WBs were also performed as indicated. The WB signals were further quantified and the enrichment of co-immunoprecipitated PFV-1 GAG was calculated in each case using the following formula: [(GAG signal in pellet /GAG signal in input)/ AGO2/PAZ9 signal in pellet] × 100. (C) PFV-1 GAG mutated in the GRI domain, which is involved in nucleic acid interaction, was constructed in the pcgp1 background (GAGΔGRI). 293T cells were transfected with the mutated PFV-1 GAG and the myc-AGO2 expression vectors. Anti-myc IPs were performed 48 hpt. Anti-PFV-1 WBs were performed on immunoprecipitated fractions (pellet) and cell extracts (input). An anti-myc WB was also performed to control IP efficiency (bottom). (D) 293T cells were transfected with PFV-1 EGFP-GAG. IPs directed against EGFP-GFP were performed 48 hpt. Immunoprecipitated fractions (pellet) and cell lysates (input) were analyzed by WB as indicated.

The retrovirus core GAG protein recruits AGO2 on viral RNAs without eliciting RNA silencing. (A) 293T cells were transfected with the empty psiCHECK2 vector or a psiCHECK2 containing the encapsidation sequences of PFV-1 (PFV-1 ψ), as well as myc-AGO2 and PFV-1 GAG expressing vectors as indicated. IPs directed against the myc tag were performed 48 hpt. RNAs extracted from IP fractions were assayed for the presence of the Renilla RNA by RT–qPCR. Fold enrichment was calculated using the following formula: 2^(Ct input-Ct IP) – (Ct input obtained in the no Ab condition – Ct IP obtained in the no Ab condition). As control, a reporter containing the HIV-1 encapsidation sequences (psiCHECK2 HIV-1 ψ) was also tested. The RNAs extracted from myc-AGO2 IPs in the presence or absence of PFV-1 GAG were also assayed for the presence of the GAPDH mRNA by RT–qPCR. Surprisingly, although several miRNAs are predicted to target the GAPDH mRNA (according to miRBase, http://www.mirbase.org/), no RNA enrichment was observed in the myc-AGO2 IP fraction. Anti-myc WBs were also performed to control IP efficiency (bottom, one representative example is shown). (B) 293T cells were transfected with the psiCHECK2 vector or PFV-1 ψ and increasing amount of PFV-1 GAG. Dual-luciferase assays were performed 48 hpt. The results were normalized with the values obtained with each construct in the absence of GAG (Renilla/Firefly ratio or Renilla/Firefly ratio obtained in the absence of GAG). (C) 293T cells were transfected as in (B) and RT–qPCRs directed against Renilla and Firefly luciferases were performed 48 hpt. The ratio Renilla/Firefly was calculated and the values obtained in the absence of GAG were used as reference. (D) 293T cells (devoid of HIV-1 GAG) and 293T* cells (stably expressing HIV-1 GAG-POL) were transfected with the empty psiCHECK2 vector or a psiCHECK2 containing the encapsidation sequences of HIV-1 (HIV-1 ψ) as well as myc-AGO2 as indicated. As control, the reporter described in (A) was also tested. IPs directed against the myc tag were realized 48 hpt. RNAs extracted from immunoprecipitated fractions were assayed for the presence of the renilla RNA by RT–qPCR. Anti-myc WBs were also performed to control IP efficiency (bottom, one representative example is shown). (E) 293T and 293T* cells were transfected with the psiCHECK2 vector or the HIV-1 ψ and Dual-luciferase assays were performed 48 hpt. The results were normalized with the values obtained with each construct in the absence of GAG (Renilla/Firefly ratio or Renilla/Firefly ratio obtained in the absence of GAG). (F) 293T and 293T* cells were transfected with the psiCHECK2 vector or the HIV-1 ψ and RT–qPCRs were performed 48 hpt as in (C).

AGO2 is required for HIV-1 and PFV-1 replication. (A) Jurkat cells were transfected with the HIV-1 provirus and indicated siRNAs. Viral production was measured 48 hpt using anti-p24 ELISA assays. Results are the mean of at least three independent experiments. *P < 0.01 (Student's test). (B) HIV-1 VLP production from 293T* cells was measured 48 hpt of the indicated siRNAs using anti-p24 ELISA assays. Results are the mean of three independent experiments. *P < 0.005 (Student's test). (C) Electron microscopy of HIV-1 VLPs produced in 293T* transfected with siRNA directed against AGO2 or p54/RCK, as a control. The size and the number of VLPs per cell are indicated in each condition. Arrows show VLPs. Right panels are enlargements of squares indicated in left panels. (D) 293T cells were transfected with the PFV-1 provirus and indicated siRNAs (T1). Separate cells were transfected with Firefly luciferase driven by the PFV-1 Long Terminal Repeat (LTR), activated by the transactivator Tas (T2). Supernatants of T1 were used to infect cells from T2. Luciferase expression was quantified in T2 cells. *P < 0.005 (Student's test).

Retroviral mRNAs interact with the RNAi machinery independently of miRNAs and P-bodies. (A) 293T cells were transfected with the PFV-1 pc13 provirus together with vectors encoding myc-AGO2/myc-PAZ9 (left) or DCP1-flag (right). IPs directed against the myc tag (left) or the flag tag (right) were performed 48 hpt. IP (pellet) and cell-extract (input) RNAs were further analyzed by RT–PCR. IP efficiency was controlled by western blots (WBs). (B) IP RNAs described in (A) were analyzed by RT–PCR using primers discriminating spliced (S) and unspliced (U) RNAs. RNase treatment was performed before IP to assess potential DNA contamination (due to DNA transfection). (C) 293T cells were transfected with the HIV-1 pNL4.3 provirus and the myc-AGO2 or myc-PAZ9 expression vectors. IPs directed against the myc tag were performed 48 hpt. RNAs were extracted from immunoprecipitated fractions and assayed for the presence of spliced (black) and unspliced (gray) viral RNAs by RT–qPCR. The log(fold enrichment, FE) = {log[2^{∧(Ct noAb – Ct IP)}]} was calculated. The log representation was chosen in order to underscore the FE obtained with spliced RNAs bound to PAZ9. The details of our calculation are shown.

HIV-1 GAG interacts with AGO2 independently of miRNAs and P-bodies. (A) Jurkat cells were transfected with myc-AGO2 or DCP1-flag vectors and further infected with HIV-1. Anti-myc or anti-flag IPs were performed 48 hpt. IP fractions (pellet) or cell lysates (input) were analyzed by WBs. (B) Similar experiments were conducted in 293T cells transfected with pNL4.3 and myc-AGO2 or myc-PAZ9 vectors. RNase treatment was performed as described in Figure 2A. The WB signals were further quantified and the enrichment of co-immunoprecipitated HIV-1 GAG was calculated in each case using the following formula: [(GAG signal in pellet/GAG signal in input)/ AGO2/PAZ9 signal in pellet] × 100. (C) 293T* cells were transfected with Flag/HA-tagged versions of AGO2 wild-type (FH-AGO-WT) or mutated at position 529 (Y529A, Y529E and Y529F). Anti-flag IPs were performed 48 hpt. WBs were executed as indicated on IP fractions (pellet) or cell lysates (input). (D) IPs directed against endogenous AGO1 and AGO2 were performed in 293T* cells. WBs were executed as indicated on IP fractions (pellet) or cell lysates (input).

AGO2 is encapsidated in retroviral particles. (A) 293T cells were transfected with the PFV-1 provirus and hA3G-V5, myc-AGO2 or myc-PAZ9 vector. hA3G-V5 was used as a positive control (38,39). PFV-1 virions were purified on a 20% sucrose cushion 48 hpt and assayed for the presence of myc-AGO2/PAZ9 and hA3G-V5 by WBs. (B) The experiments described in (A) were repeated except that 293T cells were transfected with the PFV-1 provirus and a GW182-EGFP vector. (C) 293T* and 293T cells were transfected with the myc-PAZ9 expression vector. The culture supernatants were purified on a sucrose cushion 48 hpt and assayed for the presence of myc-PAZ9 by WB. (D) The experiments described in (C) were repeated using a GW182-EGFP vector.