HTLV-1 Tax plugs and freezes UPF1 helicase leading to nonsense-mediated mRNA decay inhibition

Abstract

Up-Frameshift Suppressor 1 Homolog (UPF1) is a key factor for nonsense-mediated mRNA decay (NMD), a cellular process that can actively degrade mRNAs. Here, we study NMD inhibition during infection by human T-cell lymphotropic virus type I (HTLV-1) and characterise the influence of the retroviral Tax factor on UPF1 activity. Tax interacts with the central helicase core domain of UPF1 and might plug the RNA channel of UPF1, reducing its affinity for nucleic acids. Furthermore, using a single-molecule approach, we show that the sequential interaction of Tax with a RNA-bound UPF1 freezes UPF1: this latter is less sensitive to the presence of ATP and shows translocation defects, highlighting the importance of this feature for NMD. These mechanistic insights reveal how HTLV-1 hijacks the central component of NMD to ensure expression of its own genome.

Fig. 1

HTLV-1 Tax protects reporter and endogenous mRNAs from NMD in mammalian cells. a RNA decay assays were carried out in HeLa cells. The stability of Gl-PTC and Gl-WT mRNA was analysed after RNA quantification by qRT-PCR. mRNA half-lives (t_1/2 = ln(2)/λ with λ the time constant of the decay curves) are indicated in front of their respective conditions. b Same as a except that HeLa cells were co-transfected with HTLV-1 molecular clone. c RNA decay assays of Gl-PTC mRNA in the presence of WT HTLV-1 molecular clone (continuous line, diamonds), truncated ΔpX HTLV-1 molecular clone complemented with Tax protein (continuous line, triangle), truncated ΔpX HTLV-1 molecular clone alone (dashed line, square) and empty vector (dashed line, circle). d RNA decay assays examining the stability of endogenous GADD45α mRNA in the presence of an empty vector (dashed line, square), Tax (continuous line, cross) or a WT HTLV-1 molecular clone (continuous line, diamond; e), f Same as d with SMG5 and MAP3K14 endogenous mRNA. The values represented in each graph correspond to the mean of at least three biological replicates, and the error bars correspond to the SD. Half-lives were calculated for each replicate, and P values were calculated by performing a Student’s t-test (unpaired, two-tailed) ns: P > 0.05; *P < 0.05; **P < 0.01

Fig. 2

HTLV-1 Tax interacts directly with UPF1-HD and inhibits its ATPase activity. a Schematic diagram showing the UPF1 and Tax domains and sites of mutant derivatives used for this study. Structural domains are represented by rectangles and the protein truncations used by black lines. b Pulldown experiment using GST-Tax (lanes 1–6) or GST-His tag (lanes 7–12) as bait. After incubation, protein mixtures before (input 20% of total) or after precipitation (precipitate) were separated on a 10% SDS-PAGE gel and visualised by coomassie staining. c A SDS-PAGE gel illustrating the co-purification of human UPF1-HD-His and HTLV-1 GST-Tax proteins by two sequential affinity purifications (Nickel and Glutathione columns). Samples from the Escherichia coli lysate (lane 1: L), flow through (lanes 2 and 4: FTHIS and FTGST) and eluate (lanes 3 and 5: EHIS and EGST) fractions were loaded on the gel. The purity of the fractions was evaluated by coomassie staining (upper panel). The proteins were visualised by western blot analysis using anti-His and anti-Tax antibodies targeting UPF1 HD-His and GST-Tax, respectively (lower panels). d Graph showing the percentage of [α-³²P]ATP hydrolysed as a function of time by UPF1-HD (continuous line, diamond), UPF1-HD/Tax_CACA (continuous line, square), UPF1-HD/Tax (dotted line, circle) and Tax_CACA (dashed lane, cross) under conditions of steady-state ATPase turnover. Aliquots from the reaction mixture were quenched 0, 1, 5, 10 and 15 min before TLC chromatography (see Methods). The values are the mean of at least three biological replicates, and the error bars correspond to the SD. e UPF1-HD, UPF1-HD/Tax_CACA complex and BSA proteins were spotted onto a nitrocellulose membrane and exposed to [α-³²P]ATP (top panel). The levels of the proteins used were also analysed by slot-blot on a different membrane. The membrane was incubated with anti-His and anti-Tax antibodies (middle and lower). Nonsignificant lanes were removed as indicated by the vertical white line. Uncropped scans related to Fig. 2 are available in Supplementary Fig. 6

Fig. 3

Tax binding near the entry site of RNA decreases UPF1 affinity for its substrate. a The upper panel shows the levels of endogenous UPF1 immunoprecipitation (IP) from mock (lane 1), Tax (lane 2), pCMV- HTLV-1 ΔpX (lane 3) and pCMV- HTLV-1 WT (lane 4)-transfected HeLa cells. The lower panel represents the relative quantification of Gl-PTC RNA recovery upon UPF1 RIP. The ΔΔCt method of mRNA quantification was applied to each experimental condition. The control sample (Ctrl) derived from IP using immunoglobulin G (IgG) antibody. The values represented in each graph correspond to the mean of at least three biological replicates, and the error bars correspond to the SD. ***P < 0.005 with Student’s t-test (two-tailed, unpaired). b Representative native 8% polyacrylamide gel illustrating the interaction of UPF1-HD protein with a 30mer-ssRNA (grey line) labelled with ³²P (black star) with or without Tax_CACA factor. The RNA substrate (1 nM) was incubated with increasing concentrations of UPF1-HD (0, 1, 3, 5 and 10 nM) alone or with 300 nM of Tax_CACA. The absence of an interaction between Tax_CACA and the substrate was also verified (lane 6). c Recapitulative tables of Bio-Layer Interferometry experiments listing K_D, k_on and k_off in real-time measurements of UPF1-HD and UPF1-HD/Tax complex binding to 5′ biotinylated 30mer-ssDNA. These data were obtained using the BLItz^® System instrument and BLItz ^®Pro 1.2 software (ForteBio). d Crystal structure of human UPF1-HD complexed with RNA substrate showing the position of the R843 residue. The magnification shows the structural model of the R843C mutant produced using I-Tasser software. e Protein co-precipitation with 3′ end-biotinylated 30-mer ssRNA (Bt-RNA). Combinations of UPF1-HD (lanes 1 and 2), UPF1-HD_DE633AA (lanes 3 and 4) or UPF1-HD_R843C (lanes 5 and 6) were mixed with Tax_CACA (lanes 2, 4 and 6) and incubated in a buffer containing 200 mM NaCl before co-precipitation. Tax_CACA was incubated with Bt-RNA alone as a control for aspecific binding (lane 7). Input (20% of total) and pull-down fractions were analysed by 12% SDS-PAGE followed by coomassie blue staining. f Recapitulative tables of BLItz listing K_D, k_on and k_off in real-time measurements for UPF1-HD_WT, UPF1-HD_DE633AA, UPF1-HD_R843C and UPF1-HD_R843S binding to Tax. Uncropped scans related to Fig. 3 are available in Supplementary Fig. 7

Fig. 4

The UPF1-HD/Tax complex shows residual RNA-binding ability. a Quantification of precipitated Gl-PTC from Tax immunoprecipitation (black bars) vs. IgG immunoprecipitation control (white bars) using the globin 3′ oligos. The histogram represents the quantification of relative RNA recovery upon normalisation to input RNA under condition 1 that was set to 1. The ΔΔCt method of mRNA quantification was applied to each experimental condition. The construct encoding Tax (conditions 1, 3 and 4) or empty vector (lane 2) were co-transfected with reporter plasmids expressing Gl-PTC (lanes 1–3) or Gl-WT (lane 4). The cells were treated with non-targeting siRNA (siCtr; lanes 1, 2 and 4) or UPF1 siRNA (siUPF1; lane 3). The values represented in each graph correspond to the mean of at least three biological replicates, and the error bars correspond to the SD. ***P < 0.005 with Student’s t-test (two-tailed, unpaired). Immunoprecipitated Tax was analysed by western blotting using anti-Tax antibody (upper left panel). b Schematic representation of the double-RIP workflow. c Western blot analysis monitoring the presence of both UPF1 and Tax in each fraction. d Agarose gel showing the products of RT-PCR using globin 3′ oligonucleotides with input and immunoprecipitated RNA samples. e Real-time sensorgram of the Bio-Layer interferometry experiment showing UPF1-HD binding to 5′-biotinylated 30mer-DNA (red dashed line), and the interaction of UPF1-HD with Tax_CACA when Tax_CACA was added at 330 s (blue line). f Association and dissociation curves of UPF1-HD (10 nM) and UPF1-HD/Tax (70 nM) preformed complex to the DNA-coated sensor with or without ATP or ADPNP. Given that the k_on of the UPF1-HD/Tax_CACA complex is one order of magnitude lower than that UPF1-HD, the concentration of the protein complex was adjusted to have almost the same association kinetics with DNA than the enzyme alone. Biosensor DNA tips were loaded with: UPF1-HD (brown), UPF1-HD supplemented with 2 mM of ADPNP (red) or 2 mM of ATP (orange), UPF1-HD/Tax (dark blue), UPF1-HD/Tax supplemented with 2 mM of ADPNP (middle blue) or 2 mM of ATP (light blue). Uncropped scans related to Fig. 4 are available in Supplementary Fig. 8

Fig. 5

Tax blocks UPF1 translocation. a Schematic representation of the DNA substrate used for the magnetic tweezers' set-up. b Experimental magnetic tweezer traces showing the activity of UPF1-HD under a saturating concentration of ATP (blue line). The trace showing the force monitoring (red line) is shown above the recorded track. The number of unwound bases is deduced from the molecular extension Z(t) obtained at F = 10 piconewton (pN). From 1000 to 2500 s, the helicase unwound the ~1200 bp DNA hairpin. From 2500 to 3200 s, the DNA hairpin refolded while the UPF1-HD translocated on the ss-DNA. c The upper and lower panels show two types of enzymatic activity detected for UPF1-HD in the presence of Tax. On the upper panel, UPF1-HD is unwinding the DNA hairpin between 3000 and 3625 s before dissociation induced by Tax binding. On the lower panel, UPF1-HD is unwinding the whole DNA hairpin between 500 and 1750 s and translocating between 1750 and 1920 s on the ss-DNA. Tax blocks UPF1-HD translocation twice between 1920 and 2370 s and 2414 and 3660 s. A sliding event and the final dissociation of UPF1-HD can be observed at 2372 and 3660 s, respectively

Fig. 6

Model proposed for Tax-mediated inhibition of UPF1 activity. The fate of UPF1 depends on the time frame of interaction with Tax that occurs during HTLV-1 infection. Tax can prevent the recruitment of UPF1 to the RNA substrate. Alternately, Tax can affect the activity of previously bound RNA and NMD-engaged UPF1: the translocation activities of UPF1 can be blocked by Tax binding. The UPF1/Tax complex then dissociates from the RNA substrate or remains stuck in P-bodies