Structure of tandem RNA recognition motifs from polypyrimidine tract binding protein reveals novel features of the RRM fold

Abstract

Polypyrimidine tract binding protein (PTB), an RNA binding protein containing four RNA recognition motifs (RRMs), is involved in both pre-mRNA splicing and translation initiation directed by picornaviral internal ribosome entry sites. Sequence comparisons previously indicated that PTB is a non-canonical RRM protein. The solution structure of a PTB fragment containing RRMs 3 and 4 shows that the protein consists of two domains connected by a long, flexible linker. The two domains tumble independently in solution, having no fixed relative orientation. In addition to the betaalphabetabetaalphabeta topology, which is characteristic of RRM domains, the C-terminal extension of PTB RRM-3 incorporates an unanticipated fifth beta-strand, which extends the RNA binding surface. The long, disordered polypeptide connecting beta4 and beta5 in RRM-3 is poised above the RNA binding surface and is likely to contribute to RNA recognition. Mutational analyses show that both RRM-3 and RRM-4 contribute to RNA binding specificity and that, despite its unusual sequence, PTB binds RNA in a manner akin to that of other RRM proteins.

Mutations in RRM4 uncouple the splicing repression and RNA-binding activities of polypyrimidine tract binding protein

“…The sensitivity of RRM4 function to residues S552 and K553 near the C terminus of the domain is striking+ Upon deletion of these residues, splicing repression activity as well as RNA binding and footprinting downstream of the branch site are lost, whereas the presence of S552 and K553 rescues these activities+ Interestingly, S552 and K553 lie just outside the fourth b strand of the babbab motif and these residues are invariant in all PTB, nPTB, and regulator of differentiation (ROD1; Yamamoto et al+, 1999) isoforms identified (Fig+ 8)+ The results shown here are consistent with the NMR study of PTB RRMs 3 and 4, which shows that the five amino acids extending just beyond the fourth b strand of RRM4 lie close to the RNA-binding surface (Conte et al+, 2000)+ The results of this and the previous study are consistent with the idea that S552 and K553 contribute directly to RNA recognition by RRM4+…”

“…putative splicing corepressors+ Such an uncoupling phenotype for an alternative splicing regulator is a novel feature of the present study+ RRM4 appears to play a particularly important role in the splicing repression function of PTB, as an RRM4 deletion is not rescued by the remaining RRMs (1-3)+ Moreover, an RRM1 or RRM2 deletion is functionally neutral with respect to splicing repression activity+ Twelve RRM4 residues critical for the splicing repression function of PTB are identified (Fig+ 4, black bars)+ To assay the effects of PTB mutations, we developed a tissue-specific splicing extract that supports the neuralspecific pathway (exon inclusion), without a biochemical depletion step to remove PTB from the extract+ The neural extract contains nPTB, but levels of PTB isoforms are quite low+ Loss-of-function mutations are recognized in this assay as PTB variants that, in contrast to the wild-type protein, are unable to switch splicing to the nonneural pathway (exon skipping)+ The assay is also informative about regions of the protein that are nonessential for splicing repression+ Although the results of the present study show that PTB variants lacking the amino terminal domains, domain A, RRM 1, or RRM2 are fully active in the splicing switch assay, these regions are important, in vivo, for other functions+ Previous studies have shown that the nuclear localization of PTB requires the first 55 residues (domain A in the present study; Huang et al+, 1997; Perez et al+, 1997b), and protein dimerization requires RRMs 1 and 2 (Perez et al+, 1997b; Oh et al+, 1998)+ RRM4 is functionally distinct from the central region of the protein domain C, which contains an alanine-rich region consistent with its role as a flexible linker between RRMs 2 and 3+ Whereas the loss of RNA binding resulting from the deletion of domain C is closely linked to the loss of splicing repression activity, RRM4 mutations largely uncouple these two functions+ One interpretation of the RNA-binding behavior of the five domain C variants tested here is that RRM2, RRM3, or both actually extend beyond their predicted boundaries into domain C+ This interpretation would account for the global defects in RNA binding observed upon deletion of domain C+ It would be of interest to test this idea using structural approaches; however, information is lacking on this point+ A recent NMR study of PTB reports the structural characteristics of the region containing RRMs 3 and 4; however, the central linker region is absent from the studied proteins (Conte et al+, 2000)+ In a previous study, the human PTB isoforms PTB1, PTB2, and PTB4, which differ in the length of the central linker region, are reported to have differential effects on a muscle-specific exon of a-tropomyosin, although these isoforms have indistinguishable effects on the inclusion of a smooth muscle-specific exon found in a-actinin pre-mRNA (Wollerton et al+, 2001)+ In comparison, we have not observed any significant difference in the function of the corresponding rat PTB isoforms in the present study (rat PTB A1 corresponds to human PTB4; rat PTB A2 corresponds to human PTB1)+ The possibility that other neural pre-mRNAs are differentially affected cannot be ruled out, however+ The physiological importance of PTB isoforms is implicated from experiments documenting alterations in the ratio of human PTB isoforms in metastatic, compared to nonmetastatic rat prostate epithelial cancer cell lines (Wagner et al+, 1999)+ It has been speculated that these changes might induce irregularities in alternative splicing+…”

The KH domains of Xenopus Vg1RBP mediate RNA binding and self-association

“…Vg1RBP can self-associate in vivo+ Mixed-stage Xenopus oocyte extract was crosslinked with buffer alone (Ϫ) or 0+3-1+5 mg/mL DMS+ Samples were resolved using SDS-PAGE alongside mock treated (Ϫ) or crosslinked (ϩ) samples of recombinant R-K12-K34 and subjected to western blot analysis using a-Vg1RBP antiserum+ The migration of Vg1RBP monomers, dimers, and the large complex is indicated on the right+ 1328 A. Git and N. Standart noncanonical RRMs, such as those in PTB (Conte et al+, 2000)+ At very high protein concentrations, the RRMs displayed a striking preference for poly(U) relative to poly(C), (A) and (G) (data not shown); though the physiological implications of this observation are not clear+ That the RRMs are dispensable for the RNA-binding activity of Vg1RBP agrees with their complete absence from the Drosophila homolog (Fig+ 1; Nielsen et al+, 2000), suggesting that they are required by the vertebrate homologs to support either binding to as yet unknown RNAs or additional functions, such as proteinprotein interactions+ Indeed, the RRMs appear to enhance K12 self-association (Fig+ 6)+ Finally, we also show that Vg1RBP self-associates, via the K34 didomain, in a manner that is greatly stabilized by RNA, both as recombinant protein, and in oocyte lysates (Figs+ 6 and 7)+ The notion that KH domains mediate oligomerization was first indirectly implied by the finding that KH3 from human Nova-1 antigen forms an intermolecular tetramer when crystallized (Lewis et al+, 1999)+ A recent NMR study of the Nova KH3 domain independently supports our findings that KH domains, although able to interact in the absence of RNA, are greatly stabilized in their protein-protein interactions by RNA+ Here, homodimerization involves a specific protein-protein interface and results in a general stiffening of the protein such that the protein binds to RNA by a rigidification of the protein/RNA-binding surface mediated by protein dimerization (Ramos et al+, 2002)+ However, the primary sequence of the Nova KH3 interacting surface is not conserved in Vg1RBP+…”

Picornavirus IRES elements: RNA structure and host protein interactions

“…2B) (Jang and Wimmer, 1990;Luz and Beck, 1991). This protein harbors four RNA recognition motifs (RRM) that recognize U/C-rich sequences (Conte et al, 2000). Several picornavirus IRES elements have two polypyrimidine tracts located at each end of the IRES region.…”