Smu1 and RED play an important role for the activation of human spliceosomes
by Sandra Maria Keiper
Date of Examination:2018-09-27
Date of issue:2019-07-08
Advisor:Prof. Dr. Reinhard Lührmann
Referee:Prof. Dr. Reinhard Lührmann
Referee:Prof. Dr. Heike Krebber
Files in this item
Name:PhD thesis_Sandra Keiper_SUB.pdf
Size:7.50Mb
Format:PDF
Abstract
English
In eukaryotes, pre-mRNA splicing is catalysed by the spliceosome, a highly complex and dynamic molecular machine, which assembles stepwise by the sequential recruitment of five small nuclear ribonucleoproteins (snRNPs) and numerous non-snRNP proteins. In humans, formation of the pre-catalytic, spliceosomal B complex with a stably associated tri-snRNP requires the action of the helicase Prp28, which displaces the U1 snRNP from the 5' splice site to be replaced by the U6 snRNA. At the same time, the so-called, B-specific proteins are recruited. Transformation of the B complex into an activated (Bact) spliceosome is initially triggered by the action of the helicase Brr2 that unwinds the U4/U6 RNA duplex. This results in the crucial release of the U4 snRNP, which allows the U6 snRNA to interact with the U2 snRNA forming essential components of the catalytic centre. Concomitantly, the B-specific proteins dissociate in the course of the activation, while the Bact-specific proteins, the Prp19/CDC5L complex and the intron binding complex (IBC) are stably integrated. Smu1 and RED are two B-specific proteins that are conserved among higher eukaryotes but absent in S. cerevisiae. Currently, little is known about the precise function of these proteins. Their transient association with the B complex suggests that they may be involved in the B-to-Bact complex transition, but previous studies proposed that these proteins form a functional module that is involved in the regulation of alternative splicing. However, it is not clear how Smu1 and RED might contribute to this process. Furthermore, it has not been investigated whether these proteins participate in constitutive splicing and thus it remains possible that Smu1 and RED – like several other B-specific proteins – are in general important for intron excision in higher eukaryotes. To provide clarity on this subject, I analysed the function of Smu1 and RED in pre-mRNA splicing both in vivo and in vitro. To address whether constitutive splicing is dependent on Smu1 and RED, an RNAseq analysis was performed with HeLa cells siRNA-depleted of Smu1 or RED. Knock-down of these two proteins resulted in profound changes in alternative splicing patterns and also led to the retention of constitutively spliced introns, suggesting that Smu1 and RED are important for splicing in general, and thus not only involved in the regulation of alternative splicing. A role for Smu1 and RED in constitutive splicing was also demonstrated in vitro, using HeLa nuclear extract that was immunodepleted of Smu1 and RED. By using a well-functioning Smu1-specific antibody, Smu1 was nearly quantitatively removed from the extract along with more than 90% of RED, suggesting that the majority of human Smu1 and RED exist as a dimer in HeLa cell extract. Splicing of MINX-120 was less efficient in the absence of Smu1/RED, with an apparent slow-down in the rate of mRNA production. Investigation of spliceosome assembly revealed a transient accumulation of spliceosomal B complexes in the Smu1/RED-depleted extract, while the subsequently formed Bact and C complexes were still formed but at a slower rate. These results indicate that defects in splicing triggered by the absence of Smu1 and RED were caused by impaired spliceosome activation. To investigate whether Smu1 and RED need to interact to fulfil their function, spliceosome assembly in the presence of the individually expressed proteins was investigated. As the addition of single proteins to the Smu1/RED-depleted extract did not restore the B-to-Bact transition and splicing product formation, Smu1 and RED do not appear to function on their own. This was attributable to poor or less stable binding to the spliceosome of the individual proteins compared to the Smu1/RED dimer. Removal of the WD40 domain of Smu1 abolished binding of the dimer to the spliceosome, and spliceosome assembly remained blocked at B complex level, demonstrating that the WD40 domain is essential for proper interaction of the dimer with the spliceosome. Truncation of RED’s N- and C-terminal regions, which contact U2 or tri-snRNP proteins, respectively, within the B complex, restored splicing and spliceosome activation partially or nearly fully, suggesting that the contacts RED establishes with U2 or U5 individually are not essential for the function of Smu1/RED. To determine whether Smu1 and RED function as a binding platform for other spliceosomal factors, I purified B complexes that accumulate in their absence and investigated their composition. As no additional proteins were missing, Smu1 and RED appear to play a direct role in splicing, as opposed to aiding the binding of other factors required for spliceosome activation. Thus, these results indicate that Smu1 and RED themselves are important for efficient conversion of the B complex into Bact. To provide evidence that Smu1 and RED are also involved in spliceosome activation in vivo, I knocked-down these proteins in HeLa cells and investigated endogenous spliceosome assembly by immunoblotting, using antibodies that recognise phosphopeptides specifically associated with assembled B or Bact complexes. Knock-down of Smu1 or RED led to an increased B complex signal and a decreased Bact complex signal, indicating that the activation of the spliceosome is also impaired in vivo in the absence of Smu1 and RED. While knock-down of Smu1 and RED affected introns of all sizes, the vast majority were shorter than 100 nt. Short introns constitute only a small fraction of introns in the human genome, and thus this result indicates that the splicing of very short introns in vivo is highly dependent on the presence of Smu1 and RED. In vitro splicing studies using truncated versions of the MINX-120 pre-mRNA and Smu1/RED-depleted extract, also demonstrated that splicing was more dependent on the presence of Smu1 and RED when intron length was shorter. Truncation of the intron to either 90 or 80 nt reduced the overall efficiency of splicing compared to MINX-120, but it also enhanced the inhibitory effect of Smu1/RED-depletion. While MINX-90 was spliced somewhat less efficiently than MINX-120, splicing of MINX-80 was nearly abolished in the absence of Smu1 and RED, and led to an apparent block at the B complex stage with little or no formation of Bact or catalytically-active C complexes. Thus, Smu1 and RED play a crucial role in the splicing of extremely short introns both in vitro and in vivo. To investigate whether the intron length per se or the distance between the 5’SS and the BS or between the BS and the 3’SS determines whether splicing is dependent on Smu1/RED or not, I compared spliceosome assembly on PM5 pre-mRNAs with a shortened 5’SS-BS distance or a shortened polypyrimidine (PY) tract (BS-3’SS distance). While spliceosome activation was not affected by the truncation of the PY tract in the absence of Smu1 and RED, shortening of the 5’SS-BS distance to ~55nt (as found in MINX-80) blocked the assembly at the B complex level, indicating that the distance between the 5’SS and the BS is the decisive factor for a strong Smu1/RED-dependence. The dependence of spliceosome activation on a minimal 5’SS-BS distance is probably due to physical limitations exerted by the intron. In this case, resolving this steric hindrance should allow spliceosome activation even without the support of Smu1 and RED. Indeed, when the MINX-80 pre-mRNA was cleaved into two RNAs, spliceosome activation was restored in the absence of Smu1/RED. This supports the idea that a short 5’SS-BS distance exerts a structural constraint, which can be relieved by physically separating the 5' and 3' regions of an intron. Based on my data and structural information obtained from the cryo-EM structure of the human B complex, a model for how 5’SS-BS distance might lead to the dependency of spliceosome activation on Smu1/RED was generated. In the human B complex, Smu1 and RED form a molecular bridge between the U2 snRNP protein SF3B3 and the RNA helicase Brr2. This interaction appears to be important for spliceosome activation, potentially either by tethering Brr2 in a position required to unwind the U4/U6 interaction during activation or by directly aiding in triggering Brr2 unwinding activity. In the B complex, the intron is base paired with the U6 ACAGAG box at the 5’SS and with the U2 snRNA at the BS region. In most cases the distance between the 5'SS and the BS of the intron is sufficiently long to grant the U2 domain enough flexibility to move towards Brr2 and to form a bridge directly, even when Smu1 and RED are absent, albeit it at a slower rate. However, when this distance is short, i.e. ~56 nt or less, the intron adopts a fully extended conformation that results in a structural constraint. This could potentially result in repositioning of the U2 domain away from Brr2 and/or hinder its ability to move towards Brr2, and in turn inhibit spliceosome activation. As Smu1 and RED extend the U2/Brr2 bridge, the negative effect of a short 5'SS to BS distance would be greatly enhanced in their absence. Taken together, my studies provide novel insights into the function of the B-specific proteins Smu1 and RED in splicing. They additionally elucidate how intron architecture impacts spliceosome assembly and splicing, and how spliceosomal proteins potentially help the splicing machinery to overcome the challenges created by short introns.
Keywords: Pre-mRNA splicing; RNA Biologie; spliceosome assmebly; RNAseq; Protein biochemistry