Modulation of exon skipping by high-affinity hnRNP A1-binding sites and by intron elements that repress splice site utilization

Abstract

The RNA-binding protein hnRNP A1 is a splicing regulator produced by exclusion of alternative exon 7B from the A1 pre-mRNA. Each intron flanking exon 7B contains a high-affinity A1-binding site. The A1-binding elements promote exon skipping in vivo, activate distal 5Ј splice site selection in vitro and improve the responsiveness of pre-mRNAs to increases in the concentration of A1. Whereas the glycine-rich C-terminal domain of A1 is not required for binding, it is essential to activate the distal 5Ј splice site. Because A1 complexes can interact simultaneously with two A1-binding sites, we propose that an interaction between bound A1 proteins facilitates the pairing of distant splice sites. Based on the distribution of putative A1-binding sites in various pre-mRNAs, an A1-mediated change in pre-mRNA conformation may help define the borders of mammalian introns. We also identify an intron element which represses the 3Ј splice site of exon 7B. The activity of this element is mediated by a factor distinct from A1. Our results suggest that exon 7B skipping results from the concerted action of several intron elements that modulate splice site recognition and pairing.

High-affinity hnRNP A1 binding sites and duplex-forming inverted repeats have similar effects on 5??? splice site selection in support of a common looping out and repression mechanism

“…The backbone of the pre-mRNA substrates used in the following experiments contains sequences from the hnRNP A1 pre-mRNA corresponding to parts of exon 7 and alternative exon 7B, their respective 59 splice sites, and a common 39 splice site derived from the adenovirus major late exon L2 (Fig+ 1A)+ This basic premRNA is spliced predominantly to the proximal 59 splice site of exon 7B (RNA 45; Fig+ 1B, lane 1)+ However, when high-affinity binding sites for hnRNP A1 (A1BS) are inserted downstream of each 59 splice site, splicing occurs predominantly at the distal 59 splice site (RNA 36: Fig+ 1B,lane 4), consistent with previous reports (Chabot et al+, 1997;Blanchette & Chabot, 1999)+ Thus, the relative efficiency of distal 59 splice site utilization is approximately 50 times superior for a pre-mRNA containing two high-affinity A1 binding sites than for a premRNA lacking these sites+ Intermediate effects are observed when only one A1BS is included downstream of either the proximal or the distal 59 splice site (Fig+ 1B, lanes 2 and 3; Chabot et al+, 1997)+ This result suggests that the mechanism underlying A1 action is unlikely to involve direct repression of the 59 splice site located immediately upstream of the A1BS+ A simple model invoking changes in 59 splice site recognition is also difficult to envision because the presence of A1BS does not strongly affect the assembly of U1-dependent complex on either 59 splice sites (Chabot et al+, 1997)+ To explain the strong shift in 59 splice site selection in the absence of important effects on U1 snRNP binding, we have proposed that the binding of A1 is followed by an interaction between bound A1 molecules that loops out the proximal 59 splice site region (Blanchette & Chabot, 1999)+ The intermediate shift observed with substrates containing only one A1BS (RNA 39 and RNA 42) can be attributed to less frequent A1 binding events occurring in the other half of the pre-mRNA, a situation that would provide a partner A1 molecule required for the proposed interaction+ In the absence of high-affinity A1 binding sites, distal 59 splice site use could still occur, albeit inefficiently, possibly due to weak A1 binding+ With such a pre-mRNA, depletion of A1 from a nuclear extract completely abrogates distal 59 splice site use (M+ Blanchette, S+ Hutchison, unpubl+ results), while supplementing a nuclear extract with large quantities of recombinant A1 stimulates the use of the distal donor site (S+ Hutchison, unpubl+ results)+ A duplex structure mimics the effect of A1 binding sites According to the A1/A1 interaction model, a similar effect on 59 splice site selection should be obtained if, instead of interactions between bound A1 molecules, the change in conformation was mediated by basepairing interactions between complementary sequences (Fig+ 1D)+ It is already known that placing an exon in the loop of a hairpin can promote exon skipping (Solnick & Lee, 1987) 1. High-affinity A1-binding sites (A1BS) and inverted repeats affect 59 splice site selection in a similar manner+ A: Schematic diagram representing substrates with or without A1BS, inverted repeats, or both+ The inverted repeats are 20-nt-long stretches of a perfect complementary sequence, whereas the A1BS corresponds to the CE1a sequence that contains the UAGAGU hexanucleotide sequence (Chabot et al+, 1997)+ B: Uniformly labeled pre-mRNA substrates were incubated in a HeLa nuclear extract in the absence (lanes …”

“…High-affinity A1-binding sites (A1BS) and inverted repeats affect 59 splice site selection in a similar manner+ A: Schematic diagram representing substrates with or without A1BS, inverted repeats, or both+ The inverted repeats are 20-nt-long stretches of a perfect complementary sequence, whereas the A1BS corresponds to the CE1a sequence that contains the UAGAGU hexanucleotide sequence (Chabot et al+, 1997)+ B: Uniformly labeled pre-mRNA substrates were incubated in a HeLa nuclear extract in the absence (lanes 1-4, 5, 9, 13, 17-20) or in the presence of increasing amounts (2, 4, 8 pmol) of oligonucleotide TS10 (lanes 6-8, 10-12, and 14-16)+ The position of the distal and the proximal lariat molecules is indicated+ The arrows in lanes 17-20 indicate the position of the lariat intron product+ C: RNA 59 with long stretches (115 nt) of inverted repeats and a control RNA were incubated in a HeLa nuclear extract as described above in the absence (lanes 1 and 2) or in the presence of increasing amounts of oligo TS10 (lanes 3-5)+ Splicing products in this panel were resolved on a 7% denaturing acrylamide gel+ Substrate RNAs are schematically represented in the above panel+ D: Mechanistic parallels between the A1BS-mediated and the inverted repeats-mediated switch in splice site selection+ The A1 molecules bound to the high-affinity sequences interact to generate a loop that contains the proximal 59 splice site+ An interaction between inverted repeats would essentially produce the same structure, albeit in an A1-independent manner+ increasing amounts of an oligonucleotide (TS10) carrying the DNA version of A1 binding sites shifts selection toward the proximal 59 splice site (RNA 36;Fig+ 1B, lanes 10-12;Blanchette & Chabot, 1999)+ This effect is abrogated by the addition of an excess of recombinant hnRNP A1 proteins (Blanchette & Chabot, 1999)+ The addition of TS10 to RNA 153 promoted a reduction in distal 59 splice site usage (Fig+ 1B, lanes 14-16), although the overall effect was much less pronounced than what is observed on RNA 36 (e+g+, Fig+ 1B, compare lane 14 with lane 10)+ This result suggests that in the absence of high-affinity sites, A1 can make a small yet detectable contribution to the increase in distal 59 splice site utilization promoted by inverted repeats+ Because this is not the case when the complementarity between inverted repeats is more extensive (Fig+ 1C, lanes 1-4), one likely explanation is that the binding of A1 to weaker sites stimulates duplex formation consistent with the reported activity of A1 in RNA annealing (Pontius & Berg, 1990;Munroe & Dong, 1992;Mayeda et al+, 1994)+ Importantly, the amplitude of switch obtained with inverted repeats is as dramatic as the effect obtained with A1 binding sites despite the fact that the proximal 59 splice site remains available to U1 snRNP+ If inverted repeats had been less efficient at promoting this switch in our pre-mRNA, an A1-mediated change in pre-mRNA conformation would have been unlikely+ RNA duplex formation can therefore recapitulate the effect of A1 binding elements on 59 splice site selection, confirming that remodeling pre-mRNA structure can have a strong effect on distal 59 splice site selection+…”

“…1999)+ In contrast to the effect of SR proteins (Eperon et al+, 1993), we have been unable to correlate strong A1-mediated shifts towards distal 59 splice site utilization with equivalent changes in the assembly of U1 snRNP-dependent complexes on competing 59 splice sites (Chabot et al+, 1997)+ These results suggest that the mechanism by which A1 binding sites control 59 splice site selection may not directly affect 59 splice site recognition+ Another observation that argues against the notion that the A1BS-bound hnRNP A1 directly interferes with the binding of U1 snRNP to the proximal 59 splice site is that positioning a single high-affinity A1 binding site downstream of either the proximal or the distal 59 splice site switches splicing toward the distal site+ Based on these observations, we proposed that RNA-bound A1 molecules interact with one another to change the spatial position of the competing 59 splice sites; the proximal 59 splice site would now be part of the looped out portion of the pre-mRNA, and the distal 59 splice site could become more advantageously positioned for commitment with the 39 splice site (Chabot et al+, 1997;Blanchette & Chabot, 1999)+ The fact that A1 molecules can self-interact (Cartegni et al+, 1996;Blanchette & Chabot, 1999) is consistent with this model+ Moreover, we have shown that A1 molecules can interact while simultaneously binding to two high-affinity sites (Blanchette & Chabot, 1999)+ The current study provides additional evidence in support of the A1/A1 interaction model+ First, replacing high-affinity A1 binding sites with sequences that can form a stable RNA duplex strongly decreases proximal 59 splice site utilization and activates distal 59 splice site use to a similar extent+ Second, the insertion of spacer elements that increase the distance between A1 binding sites and 59 splice sites does not compromise distal 59 splice site utilization+ If A1 binding sites were sterically interfering with the recognition of a nearby splice site or the activity of a nearby enhancer, changing their relative positions in the pre-mRNA should have had a more important impact+ If the presence of an A1BS downstream of each 59 splice site changes the conformation of the pre-mRNA and promotes loop formation, the effect of inserting spacer elements in the pre-mRNA carrying inverted repeats should be similar to the effect obtained by inserting spacer elements in the pre-mRNA containing A1BS+ The fact that this is the case brings more credence to the view that a similar mechanism is used by high-affinity A1 binding sites and inverted repeats to promote distal 59 splice site utilization+ Thus, although our results do not constitute a proof that loop formation occurs in the presence of A1BS, they support the notion that a change in the conformation of the pre-mRNA is responsible for shifting splice site selection+ HnRNP A1-mediated loop formation therefore represents the simplest model that can easily accommodate all the results that we have obtained so far+ Our model is also consistent with the crystal structure of a shortened version of A1 lacking the glycinerich domain (UP1) bound as a dimer to single-stranded telomeric DNA (Ding et al+, 1999)+ Each molecule in this dimer interacts with binding sites on two different oligonucleotides+ Thus, A1 molecules bound to highaffinity sites on the pre-mRNA may initially interact through their glycine-rich domains, an event that may be further stabilized by "cross-strand" interactions that wou...…”

“…The hnRNP A1 protein functions in mammalian splice site selection+ HnRNP A1 antagonizes the activity of SR proteins by shifting splicing toward the most distal 59 splice site (Mayeda & Krainer, 1992;Mayeda et al+, 1993;Cáceres et al+, 1994;Yang et al+, 1994)+ HnRNP A1 has also been involved in the control of 39 splice site selection (Bai et al+, 1999;Caputi et al+, 1999;Del GattoKonczak et al+, 1999;Bilodeau et al+, 2001;Tange et al+, 2001;Zhu et al+, 2001)+ We have previously characterized the impact of high-affinity A1 binding intron elements (CE1a and CE4) in the control of alternative splicing of the A1 pre-mRNA (Chabot et al+, 1997;Blanchette & Chabot, 1999;Hutchison et al+, 2000)+ Notably, although A1 binding elements activate splicing of the most upstream 59 splice site in a model premRNA in vitro, this strong shift in 59 splice site selection is not accompanied by significant changes in the assembly of U1 snRNP-dependent complexes on the competing donor site (Chabot et al+, 1997)+ This observation prompted us to propose a model that posits that modulation of 59 splice site selection by A1 is caused by an interaction between A1 molecules bound to sequences on each side of the proximal 59 splice site (Blanchette & Chabot, 1999; Fig+ 1D)+ The resulting change in premRNA conformation may hinder splicing between the closest pair of splice sites+ Moreover, by reducing the spatial distance separating the distal 59 splice site from the 39 splice site, the interaction between A1 molecules could increase the competitive advantage of the distal donor site+ In support of this model, we have shown that A1 can interact simultaneously with two highaffinity binding sites (Blanchette & Chabot, 1999)+ Here, we present additional evidence in support of the A1/A1 interaction model of splicing control+ We show that substituting A1 binding elements for sequences that can form an RNA duplex also promotes distal 59 splice site selection+ Moreover, the effects of introducing spacer elements at various positions in model pre-mRNAs are consistent with the A1/A1 interaction model+ Finally, placing a 59 splice site in between high-affinity A1 binding sites or inside a loop formed by inverted repeats represses splicing+ Our analysis supports the notion that an interaction between bound A1 molecules represses proximal 59 splice site utilization by placing this site in a loop+…”

Protein kinase clk/STY is differentially regulated during erythroleukemia cell differentiation: a bias toward the skipped splice variant characterizes postcommitment stages

“…We speculate that either NMD proceeds in a slow fashion or there is another mechanism involved in the processing of mRNA. Autoregulation via alternative splicing and NMD has been described for other members of the splicing machinery such as PTB [47], SC35 [48], TIA and TIAR [49], Srp20 [50], ADAR2 [51] and hnRNPA1 [52]. These observations led Wollerton and co-workers to propose a common mechanism for gene expression control [47].…”