Vigilin binding selectively inhibits cleavage of the vitellogenin mRNA 3'-untranslated region by the mRNA endonuclease polysomal ribonuclease 1

Abstract

In Xenopus, estrogen induces the stabilization of vitellogenin mRNA and the destabilization of albumin mRNA. These processes correlate with increased polysomal activity of a sequence-selective mRNA endonuclease, PMR-1, and a hnRNP K homology-domain RNA-binding protein, vigilin. Vigilin binds to a region of the vitellogenin mRNA 3'-untranslated region (3'-UTR) implicated in estrogen-mediated stabilization. The vigilin-binding site in the vitellogenin B1 mRNA 3'-UTR contains two consensus PMR-1 cleavage sites. The availability of purified PMR-1 and recombinant vigilin made it possible to test the hypothesis that RNA-binding proteins interact with cis-acting elements to stabilize target mRNAs by blocking cleavage by site-specific mRNA endonucleases. Vigilin binds to the vitellogenin mRNA 3'-UTR site with at least 30-fold higher affinity than it exhibits for the albumin mRNA segment containing the mapped PMR-1 cleavage sites. This differential binding affinity correlates with differential in vitro susceptibility of the protein-RNA complexes to cleavage by PMR-1. Whereas recombinant vigilin has no detectable protective effect on PMR-1 cleavage of albumin mRNA, it retards in vitro cleavage of the vitellogenin mRNA 3'-UTR by purified PMR-1. The PMR-1 sites in the vitellogenin mRNA 3'-UTR are functional because they are readily cleaved in vitro by purified PMR-1. These results provide direct evidence for differential susceptibility to endonuclease-mediated mRNA decay resulting from the differential affinity of a RNA-binding protein for cis-acting stability determinants.

Figure 1

RNA sequence elements identified as in vitro cleavage substrates for PMR-1. (Upper) The solved secondary structure of a region in the 5′ end of albumin mRNA bearing mapped PMR-1 cleavage sites (16). APyrUGA consensus cleavage sites are boxed. (Lower) The sequence of the vitellogenin B1 3′-UTR bound by vigilin containing the only two APyrUGA elements present in the 3′-UTR of vitellogenin mRNA (18). The sites for cleavage of albumin mRNA within the APyrUGA elements are indicated (Upper) with open arrows, and the PMR-1 cleavage sites in the vitellogenin mRNA 3′-UTR determined in Fig. 2 are indicated with filled arrows, with the relative strength of each site represented by the size of each arrow.

Figure 2

Identification of PMR-1 cleavage sites in the vitellogenin mRNA 3′-UTR. Vitellogenin mRNA 3′-UTR labeled on the 5′ end with ³²P was digested with 30 units of PMR-1 and the resulting products were separated on a 6% polyacrylamide-urea gel. The sites of RNA cleavage were determined by comparison to a DNA sequencing ladder prepared from the cloned vitellogenin cDNA by using a primer beginning at the 5′ end of the encoded transcript. Cleavage sites identified in this manner are indicated on the vitellogenin sequence in Fig. 1.

Figure 3

Characterization of vigilin binding to vitellogenin mRNA 3′-UTR and albumin mRNAs. (A) EMSA analysis was performed on 500 fmol of a 160-nt uniformly labeled albumin RNA containing the mapped PMR-1 cleavage sites (16) (Left) and vitellogenin 3′-UTR (Right). Lanes marked “No protein” contain input RNA. RNAs were incubated with either 5 μg of protein extract from liver polysomes of control animals (− E extract), 5 μg of extract from polysomes of estrogen-treated animals (+ E extract), or 750 ng of recombinant vigilin (Vigilin). The vigilin–RNA complex is indicated by the open arrow. The probes used had equal specific activity, and the albumin RNA EMSA gel was exposed for 20 vs. 4 h for vitellogenin RNA gel. (B) Complexes assembled on ice as in A were crosslinked for 5 min at 0.12 joules/cm² at a distance of 5 cm. After digestion with RNase A to remove unbound probe, the protein samples were separated by SDS/PAGE on a 10% polyacrylamide gel and cross-linked peptides were visualized by PhosphorImager. The mobility of a 158-kDa maltose-binding protein-β-galactosidase fusion protein size marker is indicated. (C) Samples (10 μg) of extract from liver polysomes of control (− E) and estrogen-treated (+ E) frogs were analyzed by Western blot by using a polyclonal antibody to epitope-tagged hnRNP K homology domains 3–5 of human vigilin, which was expressed in Escherichia coli and then purified.

Figure 4

Vigilin-binding affinity for vitellogenin mRNA 3′-UTR and albumin mRNA. (A) A protein excess-binding curve EMSA was performed by using ³²P-labeled albumin mRNA or vitellogenin 3′-UTR and increasing amounts of polysome extract from estrogen-treated frogs. Binding activity was quantified by using the PhosphorImager to compare the relative amount of radiolabeled RNA shifted into slower-migrating complexes versus unbound RNA at the bottom of the gel. (B) The ability of the albumin transcript bearing the PMR-1 cleavage sites to displace vigilin-bound vitellogenin mRNA 3′-UTR was determined by a competition EMSA. Polysome extract (5 μg) from estrogen-treated frogs was incubated with 500 fmol of ³²P-labeled vitellogenin mRNA 3′-UTR plus the indicated amounts of unlabeled albumin mRNA competitor. The position of the retarded complex is indicated by the open arrow.

Figure 5

Vigilin binding selectively inhibits PMR-1 cleavage of vitellogenin mRNA 3′-UTR; 500 fmol of uniformly labeled albumin mRNA (Alb) or vitellogenin mRNA 3′-UTR (Vit) were incubated with buffer alone (No protein), 40 units of PMR-1 (PMR-1), 500 ng of recombinant vigilin (Vigilin), or PMR-1 plus vigilin (PMR-1 and Vigilin). Samples recovered from each reaction were electrophoresed on a 6% polyacrylamide-urea gel and visualized by PhosphorImager. Lane 1 contains a size marker of HinfI-cut φX174 DNA.