Gammaretroviral pol sequences act in cis to direct polysome loading and NXF1/NXT-dependent protein production by gag-encoded RNA

Abstract

Background: All retroviruses synthesize essential proteins via alternatively spliced mRNAs. Retrovirus genera, though, exploit different mechanisms to coordinate the synthesis of proteins from alternatively spliced mRNAs. The best studied of these retroviral, post-transcriptional effectors are the trans-acting Rev protein of lentiviruses and the cis-acting constitutive transport element (CTE) of the betaretrovirus Mason-Pfizer monkey virus (MPMV). How members of the gammaretrovirus genus translate protein from unspliced RNA has not been elucidated.

Results: The mechanism by which two gammaretroviruses, XMRV and MLV, synthesize the Gag polyprotein (Pr65Gag) from full-length, unspliced mRNA was investigated here. The yield of Pr65Gag from a gag-only expression plasmid was found to be at least 30-fold less than that from an otherwise isogenic gag-pol expression plasmid. A frameshift mutation disrupting the pol open reading frame within the gag-pol expression plasmid did not decrease Pr65Gag production and 398 silent nucleotide changes engineered into gag rendered Pr65Gag synthesis pol-independent. These results are consistent with pol-encoded RNA acting in cis to promote Pr65Gag translation. Two independently-acting pol fragments were identified by screening 17 pol deletion mutations. To determine the mechanism by which pol promoted Pr65Gag synthesis, gag RNA in total and cytoplasmic fractions was quantitated by northern blot and by RT-PCR. The pol sequences caused, maximally, three-fold increase in total or cytoplasmic gag mRNA. Instead, pol sequences increased gag mRNA association with polyribosomes ~100-fold, a magnitude sufficient to explain the increase in Pr65Gag translation efficiency. The MPMV CTE, an NXF1-binding element, substituted for pol in promoting Pr65Gag synthesis. A pol RNA stem-loop resembling the CTE promoted Pr65Gag synthesis. Over-expression of NXF1 and NXT, host factors that bind to the MPMV CTE, synergized with pol to promote gammaretroviral gag RNA loading onto polysomes and to increase Pr65Gag synthesis. Conversely, Gag polyprotein synthesis was decreased by NXF1 knockdown. Finally, overexpression of SRp20, a shuttling protein that binds to NXF1 and promotes NXF1 binding to RNA, also increased gag RNA loading onto polysomes and increased Pr65Gag synthesis.

Conclusion: These experiments demonstrate that gammaretroviral pol sequences act in cis to recruit NXF1 and SRp20 to promote polysome loading of gag RNA and, thereby license the synthesis of Pr65Gag from unspliced mRNA.

Figure 1

Gammaretroviral pol sequence is required for efficient Pr65 ^Gag protein production. (A) HEK293T cells were transfected with the indicated XMRV and MLV expression constructs. Cell lysate was harvested 48 hrs later and probed with anti-CA antibody (upper panel) or anti-β-actin antibody (lowel panel). (B) The magnitude difference in XMRV Pr65^Gag protein level in cells tranfected with the the gag-pol or gag-only expression plasmids was determined by comparing the gag-only signal with serial dilutions of lysate from cells transfected with gag-pol.

Figure 2

Protein synthesis by pol is not required to promote Pr65 ^Gag protein production. HEK293T cells were transfected with the indicated constructs (XMRV gag-pol, XMRV gag, XMRV gag-pol with a frameshift mutation just after the XMRV gag stop codon, codon optimized XMRV gag, or empty vector), and harvested 48 hrs later. (A) Schematic representation of XMRV constructs showing interruption of pol translation by introduction of a frameshift mutation just after the stop codon of XMRV gag. (B) HEK293T cell lysate was probed with anti-CA antibody (upper panel) and anti-β-actin antibody (lowel panel). (C) Virus-like particles (VLPs) pelleted from the supernatant by ultracentrifugation were collected and analyzed by immunoblotting with anti-CA antibody.

Figure 3

Either of two pol fragments promote Gag protein production. HEK293T cells were transfected with the indicated constructs and cell lysate was harvested 48 hrs later. (A) Schematic of the XMRV pol sequence fragments that were cloned out of frame and downstream of XMRV gag in which the natural UAG stop codon was replaced with UGA. (B) Cell lysate was probed with anti-CA antibody (upper panel) and anti-β-actin antibody (lowel panel).

Figure 4

Deletion analysis of the nucleotide 2232–3456 pol fragment . HEK293T cells were transfected with the indicated constructs derived from XMRV and harvested 48 hrs later for western blot. (A) Schematic showing truncation mutants generated from the XMRV pol fragment 2232–3456. (B) Cell lysate was probed with anti-CA antibody (upper panel) and anti-β-actin antibody (lowel panel).

Figure 5

Deletion mapping of pol fragment 4543–5199 . HEK293T cells were transfected with the indicated XMRV constructs, and harvested 48 hrs later. (A) Schematic of the truncation of XMRV pol fragment 4543–5199. Within this fragment is a known splice acceptor site (SA) and a potential splice donor (SD) site. Single point mutation within the AG sequence of the SA site and within the GU sequence of the SD site are shown. (B) Cell lysate was probed with anti-CA antibody (upper panel).

Figure 6

The XMRV pol sequence contributes a modest extent to steady-state level and nucleocytoplasmic export of gag mRNA. HEK293T cells were transfected with the indicated XMRV constructs. 48 hrs later, RNA from total (T) and cytoplasmic (C) fractions was collected. (A) XMRV mRNA was detected by northern blot with a radiolabeled gag mRNA probe. (B) The total cellular and cytoplasmic fractions used in northern blots were subjected to western blot with anti-HSP90 and anti-Histone H3 antibodies to monitor potential contamination of cytoplasmic preparations with nuclear contents. (C) Phosphorimager quantification of XMRV mRNA. Quantification was accomplished by detecting the bands and subtracting the background for each lane. This experiment - from transfection to northern blot quantitation - was repeated on three occasions with comparable results (see text). A representative experiment is shown.

Figure 7

Gammaretroviral pol promotes the association of gag mRNA with polyribosomes. HEK293T cells were transfected with expression plasmids for gag-pol, gag, gag with pol fragment 2232–3456, or gag with pol fragment 4543–5199. 48 h post-transfection cell lysate was harvested and loaded onto a 15% to 55% linear sucrose gradient. After acceleration for 3 hrs at 210,000 × g RNA content across the gradient was assessed by reading absorbance at 254 nM (upper panel). The position of migration of the various ribosomal components and polysomes is indicated. Fractions were collected from the gradient (lower panel) and gag mRNA in fractions 3 to 10 was quantified by qRT-PCR relative to GAPDH as a control (left Y axis) and displayed against the percent sucrose (right Y axis). The ratio of gag RNA signal expressed from gag-pol plasmid versus the gag-alone alone plasmid is shown for each fraction. This experiment was repeated on three separate occasions, using comparable constructs from MLV and XMRV, and a representative experiment with MLV is shown.

Figure 8

MPMV CTE and NXF1/NXT promote gammaretroviral Gag protein production. HEK293T cells were transfected with plasmids expressing gag-pol, gag, or gag with the 2232–3456 pol fragment. In addition, cells were transfected with a plasmid in which a single MPMV CTE was cloned downstream of gag (A), or cells were co-transfected with expression plasmids for NXF1 and NXT, as indicated in (B). Cell lysate was harvested 48 hrs later and probed with anti-CA antibody (upper panel) or anti-β-actin antibody (lowel panel). Cell lysate from (B) was probed additionally with anti-NXF1 and anti-NXT antibody, as indicated. Similar results were obtained with identical constructs from MLV and XMRV. A representative experiment with MLV derived constructs is shown.

Figure 9

Identification of a gammaretroviral CTE (γ-CTE) in pol . (A) Secondary structure model of the minimal pol RNA fragment that stimulates Pr65^Gag production. This structure includes a loop with the NXF1-binding motif (AAGACA) found in the MPMV/MLV CTE (identical for both viruses). Nucleotides that were mutated to generate γ-CTE^mut (ATCGCT) are highlighted in red. In (B) and (C), HEK293T cells were transfected with expression plasmids for gag, gag-pol, gag with 4 copies of either wild-type or mutant γ-CTE, NXF1, and NXT, as indicated. Cell lysate was harvested 48 hrs later and probed with anti-CA antibody and anti-β-actin antibody. Similar results were obtained with identical constructs from MLV and XMRV. A representative experiment with MLV derived constructs is shown.

Figure 10

The effect of NXF1 KD on Pr65 ^Gag protein production. (A) Diagram showing protocol for short-term selection of cells with NXF1 KD. HEK293 cells were transduced with lentiviral vectors bearing miR30 frameworks targeting either luciferase or NXF1. Then, cells were transfected with the indicated plasmids, and finally selected in pools with puromycin. (B) 55 hrs after addition of puromycin cell lysate was probed by western blot with anti-gammaretrovirus CA antibody, anti-HIV-1 CA antibody, anti-NXF1 antibody, or anti-β-actin antibody, as indicated. Similar results were obtained with identical constructs from MLV and XMRV. A representative experiment with MLV derived constructs is shown.

Figure 11

SRp20 promotes Pr65 ^Gag production in an NXF1-dependent manner. HEK293T cells were co-transfected with expression vectors for either gag-pol, gag, or gag with the 2232–3456 pol fragment. Each of these was co-transfected with either FLAG-tagged SRp20 (A) or FLAG-tagged SRp40 (B). In (C), cells were transfected with an expression plasmid for gag and either wild-type or mutant SRp20. One mutant, SRp20ΔRRM, lacks the RNA recognition motif. The other mutant, SRp20R3A, disrupts binding to NXF1. 48 hrs later cell lysate was probed with anti-CA antibody, anti-Flag antibody, and anti-β-actin antibody, as indicated. Similar results were obtained with identical constructs from MLV and XMRV. A representative experiment with MLV derived constructs is shown.

Figure 12

SRp20 promotes gag mRNA association with polysomes in an NXF1-dependent manner. HEK293T cells were transfected with expression plasmids for gag and either SRp20 wild-type, SRp20ΔRRM, or SRp20R3A. 48 h post-transfection cell lysate was loaded onto a 15% to 55% linear sucrose gradient. After acceleration for 3 hrs at 210,000 × g, RNA content across the gradient was assessed by reading absorbance at 254 nM (upper panel). The position of migration of the various ribosomal components and polysomes is indicated. Fractions were collected from the gradient (lower panel) and relative gag mRNA in fractions 3 to 10 was quantified by qRT-PCR using GAPDH as a control (left Y axis) and displayed against the percent sucrose (right Y axis). The ratio of gag RNA signal expressed in the presence of SRp20 wild-type, versus the gag RNA signal in the absence of SRp20 plasmid is shown for each fraction. This experiment was repeated on three separate occasions with the same result. Similar results were obtained with identical constructs from MLV and XMRV. A representative experiment with MLV-derived constructs is shown.

Figure 13

Model for the post-transcriptional regulation of gammaretroviral gag mRNA by NXF1 and SRp20. Translation of gammaretroviral Gag polyprotein requires recruitment of NXF1 to the gag mRNA. Recruitment of NXF1 might be direct, as it is with the MPMV CTE and appears to be with the γ-CTE located in pol, or indirect, via binding SRp20. (A) When NXF1 and SRp20 proteins are present at endogenous levels, direct binding of NXF1 to the γ-CTE in pol appears to be essential for Gag translation. (B) When NXF1 is overexpressed, some Gag protein is produced by gag alone, perhaps due to indirect NXF1 recruitment by SRp20. (C) When SRp20 is overexpressed, significant increase in Gag protein is observed with either gag alone or with gag-pol.