CstF-64 and 3'-UTR cis-element determine Star-PAP specificity for target mRNA selection by excluding PAPα

Abstract

Almost all eukaryotic mRNAs have a poly (A) tail at the 3'-end. Canonical PAPs (PAPα/γ) polyadenylate nuclear pre-mRNAs. The recent identification of the non-canonical Star-PAP revealed specificity of nuclear PAPs for pre-mRNAs, yet the mechanism how Star-PAP selects mRNA targets is still elusive. Moreover, how Star-PAP target mRNAs having canonical AAUAAA signal are not regulated by PAPα is unclear. We investigate specificity mechanisms of Star-PAP that selects pre-mRNA targets for polyadenylation. Star-PAP assembles distinct 3'-end processing complex and controls pre-mRNAs independent of PAPα. We identified a Star-PAP recognition nucleotide motif and showed that suboptimal DSE on Star-PAP target pre-mRNA 3'-UTRs inhibit CstF-64 binding, thus preventing PAPα recruitment onto it. Altering 3'-UTR cis-elements on a Star-PAP target pre-mRNA can switch the regulatory PAP from Star-PAP to PAPα. Our results suggest a mechanism of poly (A) site selection that has potential implication on the regulation of alternative polyadenylation.

Figure 1.

Star-PAP controls distinct mRNA target independent of PAPα. (A) qRT-PCR analysis of mRNAs after knockdown of Star-PAP in HEK 293 cells expressed as fold-reductions relative to the control cells. (B) Measurement of uncleaved pre-mRNA of BIK, NQO1 and GCLC by qRT-PCR expressed relative to total mRNA level in presence and absence of Star-PAP knockdown. Total mRNA levels are shown in D. (C) qRT-PCR analysis of the expression of BIK and NQO1 mRNA level after Star-PAP knockdown followed by rescue with stable expressed FLAG-Star-PAP insensitive to the siRNA used for the knockdown or stable expressed FLAG-PAPα in HEK 293 cells. (D) Total mRNA levels of corresponding genes in B. (E–G) 3′-RACE assay of BIK, NQO1 and GAPDH under the similar conditions as in C. (H) Western blot showing knockdown of Star-PAP. (I) Western blot of NQO1, BIK and control β-Tubulin from HEK 293 cell lysates after knockdown of Star-PAP and rescue with FLAG-Star-PAP or -PAPα as described in C. Wherever (–) siRNA is indicated, we have used control scrambled siRNA. (J) RNA immunoprecipitation analysis of RNA Pol II, Star-PAP and PAPα on UTR RNAs as indicated.

Figure 2.

Star-PAP and PAPα compete with each other for CPSF-160 binding. (A) GST-pulldown using GST-Star-PAP or -PAPα to pulldown various CPSF subunits from HEK 293 cell lysates. (B) Quantifications of gels from C–E, and plot of relative bound CPSF-160 fraction on addition of increasing His-Star-PAP or -PAPα to compete the binding of CPSF with other PAP. (C) Control GST-pulldown experiment of CPSF-160 by GST-PAPα with no competitor PAP addition. (D) GST-pulldown of CPSF-160 and hFIP1 by GST-PAPα in the presence of increasing additions of (0, 12.5 nM, 25 nM, 50 nM, 100 nM, 200 nM) His-Star-PAP. (E) GST-pulldown of CPSF-160 and hFIP1 by GST-Star-PAP in the presence of increasing additions of (0, 12.5 nM, 25 nM, 50 nM, 100 nM, 200 nM) His-PAPα. (F–G) GST-pulldown experiment with control GST- from HEK 293 cell lysates in the presence of increasing His-Star-PAP and -PAPα additions.

Figure 3.

Star-PAP recognition of target mRNA is driven by a -AUA- core motif upstream of PAS. (A) BIK UTR RNA sequence showing Star-PAP binding region and mutations of AUA motif. (B) RNA EMSA experiment of Star-PAP with short RNA oligo having AUA in the sequence (C) with AUA to GGG mutation on the oligo (D) with NQO1 UTR RNA (E) with a mutation of AUA to GGG on NQO1 UTR. (F) Putative Star-PAP binding motif obtained by in silico analysis of Star-PAP target mRNAs at the Star-PAP binding region with core -AUA- motif. (G) RNA EMSA experiment of Star-PAP with GCLC UTR RNA (H) BIK UTR and (I) BIK UTR with AUA to GGG mutation in the Star-PAP binding region. F: free probe, B: Star-PAP-RNA binary complex. (J) Schematic of reporter mini gene construct of FLAG-NQO1 expressed from pCMV promoter and driven by NQO1 PAS (Star-PAP regulated distal poly(A) site that controls overall NQO1 expression, see Supplementary Figure S3) or control SV40 UTR. The sequence of the Star-PAP binding USE and the mutation of the AUA (U Mut) or introduction of U-rich DSE (D-Mut), or both (DU-Mut) is indicated. (K and L) Western blot analysis of FLAG-NQO1 HEK 293 cell lysates after transfection of the reporter constructs under the conditions as indicated. (M) qRT-PCR analysis of FLAG-NQO1 expression with a forward primer from FLAG and reverse primer from NQO1 CDS from HEK 293 cells after transfection of the reporter constructs.

Figure 4.

Suboptimal DSE at the 3′-UTR of Star-PAP target mRNAs prevent CstF-64 binding that excludes PAPα from the UTR. (A) BIK UTR sequence indicating the suboptimal DSE region and the insertion of U-rich DSE (UUUUUU) at the UTR. RNA EMSA experiment of CstF-64 with (B) GCLC (C) BIK UTR and (D) mutant BIK with U-rich DSE. F-free probe, B- CstF-64-RNA binary complex. (E) RIP analysis of Star-PAP, CstF-64, CPSF-160 and RNA Pol II on mRNAs as indicated. (F) qRT-PCR analysis of BIK, NQO1 and GCLC mRNA expressions under the conditions as indicated. (G) Uncleaved mRNA measurement by qRT-PCR expressed relative to the total mRNA levels under conditions as in F. (H) Western blot analysis of NQO1, BIK and control α-Actin from lysates of HEK 293 cells after knockdown of Star-PAP, CstF-64, CPSF-160, PAPα or control cells. (I) Reporter constructs as in Figure 3J, showing DSE region and the U-rich DSE insertions. (J) Reporter assay by western blot analysis using anti-FLAG antibody from HEK 293 cells after transfection of the reporter construct under the conditions as indicated. (K) qRT-PCR analysis under the similar conditions as in J.

Figure 5.

Model of Star-PAP mediated poly(A) site/UTR selection. Star-PAP recognition of AUA motif and suboptimal DSE mediated exclusion of PAPα is indicated.