Yra1 splicing auto-regulation

Ubiquitination of Yra1 mutants

Yra1 levels have to be tightly regulated since overexpression of Yra1 protein is detrimental for cell growth and causes mRNA export defects (Espinet et al., 1995;

Preker et al., 2002; Rodriguez-Navarro et al., 2002). Importantly Yra1 contains the second largest intron in the S. cerevisiae genome and is able to inhibit its own splicing leading to YRA1 pre-mRNA export and cytoplasmic degradation dependent on the de-capping activator Edc3 (Dong et al., 2010a; Dong et al., 2007a; Preker and Guthrie, 2006). One question is whether Yra1 ubiquitination may have a role in this splicing auto-regulation. For this reason, all the yra1 mutants analyzed in this study contain the YRA1 intron sequence except for the yra1Δintron mutant. Yra1 contains 22 lysines.

Notably, mutating all 8 N-terminal (HA-yra1KR1-8) or 14 C-terminal (HA-yra1KR9-15, HA-yra1KR16-22, HA-yra1KR9-22) lysines to arginines does not abrogate Yra1 ubiquitination indicating redundancy among the lysines and impossibility to define the minimal region important for Yra1 ubiquitination. Even the truncated mutant yra1(1-210) that lacks the 16 C-terminal amino acids (of which 4 lysines) is ubiquitinated.

Ubiquitination is fully abrogated only in the HA-yra1allKR mutant in which all the lysines are mutated to arginines (Figure 2.1 and data not shown).

Figure 2.1: Mutagenesis of all Yra1 lysines (K) to arginines (R) abrogates Yra1 ubiquitination

Ubiquitination assay of shuffled YRA1 strains transformed with a cupper inducible His-Ubiquitin expressing vector (Ubi vector, 2µ plasmid) or Empty vector. His-Ubiquitin was overexpressed by cupper induction over-night. His-Ubiquitinated proteins were affinity purified and the Ubiquitinated forms of Yra1 were detected by Western Blot analysis with an αHA antibody. Western Blot of input samples with αPgk1 was used as loading control and with αHA to assess total Yra1 levels. Shown is one representative experiment from at least 3 that were performed by Valentina Infantino and Ivona Bagdiul.

To identify which lysine may be preferentially ubiquitinated in vivo, a shuffled strain was constructed expressing a His-Yra1 fusion protein. The His-Yra1 fusion protein purified on Nickel beads was fractionated on SDS-PAGE and the bands corresponding to non-modified and modified forms of His-Yra1 subjected to Mass Spec (Figure 2.2). These analyses identified K201 as the most probable lysine modified by ubiquitination (Prof. Adibekian, Unige). The K201 was identified as ubiquitinated when purifying His-Yra1 from two independently shuffled yeast strains.

As shown in the cartoon, K201 lies within the C-terminal variable region of Yra1;

however, this does not exclude that, in the absence of K201, other lysines become

Ubiquitinated. In this regard, the HA-yra1KR9-15 and HA-yra1KR9-22 mutants, in which lysine 201 is replaced by arginine (KR15), are still Ubiquitinated (Figure 2.1).

To exclude that the observed ubiquitination was linked to the long overnight cupper induction of His-Ubiquitin from the 2µ plasmid, Yra1 ubiquitination was examined after a 3h cupper induction. In this condition as well, only the HA-yra1allKR mutant exhibited absence of ubiquitination (Figure 2.3). For the same purpose, we tested Yra1 ubiquitination expressing His-Ubiquitin from a Centromeric plasmid with only 2h of cupper induction (Figure 2.4). Although the overall Ubiquitin signal is weaker, only the HA-Yra1allKR mutant (lane 3) shows lack of ubiquitination comparable to the Δslx8Δtom1 mutant (lane 9) used as negative control. These observations indicate that Yra1 ubiquitination can occur on several lysines in a redundant manner.

Figure 2.2: Purification of His-Yra1 for Mass Spec analysis of ubiquitinated forms.

Extracts from two His-Yra1 shuffled strains (colonies 5 and 7) were affinity purified on Nickel beads.

Samples were eluted under stringent conditions and fractionated on polyacrylamide gels that were either subjected to Western blotting with anti-Yra1 (left) or stained by Coomassie (right). Mass spec analysis was performed on selected bands corresponding to Yra1 and the expected slower migrating His-Yra1-Ubi in the two His-Yra1 strains. These bands are absent in the eluted HA-Yra1 extract used as negative control. Experiment performed by Ivona Bagdiul.

Figure 2.3: HA-yra1allKR is the only mutant to lack ubiquitination when expressing His-Ubiquitin from a cupper inducible 2µ plasmid for 3h.

Ubiquitination assay of shuffled HA-YRA1 wild-type and mutant strains transformed with a cupper inducible His-Ubiquitin expressing 2µ plasmid or Empty vector. His-Ubiquitin was induced with cupper for 3h. Ubiquitination assay was performed by affinity purification of His-Ubiquitinated proteins. The Ubiquitinated forms of Yra1 were detected by Western Blot with an αHA antibody. Western Blot of input samples with αPgk1 was used as loading control; αHA antibody was used to assess input Yra1 levels. Experiment performed by Ivona Bagdiul.

Figure 2.4: HA-yra1allKR is the only mutant to lack ubiquitination when expressing His-Ubiquitin from a cupper-inducible Centromeric plasmid.

Ubiquitination assay of shuffled HA-YRA1 wild-type and mutant strains transformed with a cupper inducible His-Ubiquitin expressing Centromeric plasmid or Empty vector. His-Ubiquitin was induced with cupper for 2h. Ubiquitination assay was performed by affinity purification of His-Ubiquitinated proteins. The Ubiquitinated forms of Yra1 were detected by Western Blot with an αHA antibody.

Western Blot of input samples with αPgk1 was used as loading control; αHA antibody was used to assess Yra1 input levels. Experiment performed by Ivona Bagdiul.

Yra1 splicing analysis in yra1 mutants

Although we could not identify the Yra1 minimal region required for ubiquitination, we decided to analyse YRA1 pre-mRNA splicing in the HA-yra1KR9-15 and HA-yra1KR16-22 mutants, as well as in the HA-yra1(1-210) mutant lacking the last 16 amino acids. To facilitate the detection of the effects of Yra1 on its own pre-mRNA splicing, we worked in Δedc3 background, which leads to the cytoplasmic stabilization of YRA1 pre-mRNA that escaped splicing (Dong et al., 2010b). Northern blot analyses specific for YRA1 mRNA and pre-mRNA were performed in the HA-YRA1 WT, HA-yra1KR9-15, HA-yra1KR16-22, HA-yra1(1-210), and HA-yra1Δintron shuffled strains in presence or absence of Edc3 (Figure 2.5 A). The HA-yra1(1-210) mutant shows an increase in YRA1(1-210) mRNA and a decrease in pre-mRNA by Northern analysis with a corresponding increase in HA-yra1(1-210) protein levels as shown by Western blot (Figure 2.5 B). The HA-yra1Δintron was included as a positive control since the lack of intron abrogates the splicing regulation of Yra1 resulting in an increase of HA-Yra1 protein levels (Figure 2.5). Northern blot quantification shows the increase of the YRA1 mRNA and the decrease of the YRA1 pre-mRNA resulting in a decrease in the pre-mRNA/mRNA ratio in the case of the HA-yra1(1-210), HA-yra1Δintron and slightly in the HA-yra1KR16-22 mutants (Figure 2.6). These results indicate the importance of the Yra1 C-terminal region in the splicing auto-inhibitory loop, in agreement with studies performed with the yra1ΔC11 mutant (Preker and Guthrie, 2006).

Figure 2.5: The HA-yra1(1-210) mutant cannot inhibit YRA1 pre-mRNA splicing resulting in Yra1 overexpression.

A) One representative Northern blot out of three that was quantified in Figure 2.6 using PhosphoImager software. The probe (283bp) recognizes exon2 of YRA1 transcripts and was designed to detect both YRA1 pre-mRNA and mRNA. The probe for SCR1 was used as loading control. The analyzed shuffled strains are listed below.

B) Western Blot analysis of HA-Yra1 levels in HA-YRA1 WT, yra1KR9-15, yra1KR16-22, HA-yra1(1-210) and HA-yra1Δintron shuffled in WT and Δedc3 backgrounds was performed using αHA antibodies; αPgk1 was used to control for loading. One representative Western blot out of three is shown.

C) Quantification of the Western Blot shown in B using Lycor Software. Average and standard deviation of three independent experiments is shown. Significant changes with respect to HA-YRA1 WT are indicated. *=p value<0.05. Experiment set up by Valentina Infantino and performed by Ivona Bagdiul.

Figure 2.6: The HA-yra1(1-210) and HA-yra1KR16-22 mutants prevent inhibition of Yra1 pre-mRNA splicing to different extents.

Northern blot quantification of YRA1 mRNA (A), YRA1 pre-mRNA (B) levels in HA-YRA1 WT, HA-yra1KR9-15, HA-yra1KR16-22, HA-yra1(1-210) and HA-yra1Δintron shuffled in WT and Δedc3 backgrounds. Data were normalized to the ncRNA SCR1. The YRA1 pre-mRNA/YRA1 mRNA ratio is shown in C. *=p value<0.05 **=p value<0.01 in blue when referring to HA-YRA1 WT in WT background and in red when referring to HA-YRA1 WT in Δedc3 background. Average and standard deviation of three independent experiments are shown. Experiment set up by Valentina Infantino and performed by Ivona Bagdiul.

These data were confirmed by quantitative reverse transcriptase PCRs (qRT-PCR) analyses specific for YRA1 mRNA, YRA1 pre-mRNA and YRA1 total RNA (pre-mRNA and (pre-mRNA) (Figure 2.7) confirming the Northern blot quantifications. The HA-yra1KR16-22,Δedc3 mutant showed a significant decrease in the YRA1 pre-mRNA/YRA1 mRNA ratio in comparison to the HA-YRA1 WT,Δedc3 (Figure 2.7 D), indicating a possible impairment of HA-yra1KR16-22 in the splicing auto-inhibition.

This is also supported by the Northern quantification, which shows a slight but significant increase in the HA-yra1KR16-22 WT mRNA (Figure 2.6 A), a decrease in the pre-mRNA (Figure 2.6 B), and in the pre-mRNA/mRNA ratio (Figure 2.6 C).

Figure 2.7: The HA-Yra1(1-210) mutant prevents inhibition of YRA1 pre-mRNA splicing.

RT-qPCR analysis of YRA1 mRNA (A), YRA1 pre-mRNA (B), YRA1 total-RNA (C) (pre-mRNA and mRNA) levels in HA-YRA1 WT, HA-yra1KR9-15, HA-yra1KR16-22, HA-yra1(1-210), HA-yra1Δintron shuffled in WT and Δedc3 background. Data were normalized to the ncRNA SCR1 and relative values to HA-YRA1 WT in WT background are shown. The ratio between YRA1 pre-mRNA with YRA1 mRNA and YRA1 total-RNA is shown in D and E. *=p value<0.05 **=p value<0.01 in blue when referring to HA-YRA1 WT in WT background and in red when referring to HA-HA-YRA1 WT in Δedc3 background. Average and standard deviation of three independent experiments is shown. Experiment set up by Valentina Infantino and performed by Ivona Bagdiul.

The slight effect of the HA-Yra1KR16-22 mutant in preventing the splicing auto-inhibitory regulation does not result in increased protein expression. In contrast,

the HA-Yra1(1-210) and the HA-Yra1Δintron showed an increase in protein levels as expected (Figures 2.5 B and C).

These observations suggest that the Yra1 C-terminal region is required for splicing inhibition in order to regulate Yra1 expression levels confirming studies with the yra1ΔC11 mutant (Preker and Guthrie, 2006). The slight effect of the HA-yra1KR16-22 mutant on splicing inhibition could result from the mutations lying at the C-terminus rather than from impaired ubiquitination since this mutant still retains ubiquitination at more N-terminal lysines (Figures 2.1, 2.3, 2.4). However, one cannot fully exclude that ubiquitination of the very C-terminal lysines may contribute to Yra1 function in splicing auto-regulation.

Yra1 splicing in ΔE3 Ubiquitin ligase mutants

Given the limitations in using Yra1 KR mutants, we decided to investigate the potential effect of loss of E3 Ubiquitin ligases to assess the role of Yra1 ubiquitination in Yra1 splicing auto-regulation. For this purpose, we performed the same RT-qPCR analysis in HA-YRA1 WT shuffled in WT and Δedc3 background of Δslx5Δtom1, Δslx8Δtom1, Δtom1, Δslx5 and Δslx8 mutants (Figure 2.8). In none of the mutants did we observe a decrease in YRA1 pre-mRNA with a concomitant increase in YRA1 mRNA indicating that the variations observed are not strictly related to the impairment of YRA1 splicing auto-inhibition. Interestingly the Δslx8Δtom1Δedc3 triple mutant showed a significant decrease of the YRA1 pre-mRNA/YRA1 mRNA ratio compared to the WT Δedc3 single mutant. The same decrease is observed in the case of the YRA1 pre-mRNA/YRA1 total-RNA ratio (Figures 2.8 D, E). In contrast, the YRA1 pre-mRNA/YRA1 mRNA and YRA1 pre-mRNA/YRA1 total-RNA ratios increased highly significantly in Δslx8Δtom1 WT in comparison to the WT. These observations indicate that the effect of preventing splicing inhibition is specific for the Δslx8Δtom1 mutant only in the Δedc3 background. This phenotype is therefore unlikely to reflect a direct regulation by Yra1 ubiquitination. Furthermore in Δslx5Δtom1, the other mutant that abrogates Yra1 ubiquitination, we observed an increase in YRA1 pre-mRNA both in WT and Δedc3 background, rather suggesting that splicing inhibition is enhanced (Figures 2.8 D, E). Overall, we could not assess the role of Yra1 ubiquitination in the

splicing auto-regulation using E3 ligase mutants because of possible indirect effects of their targets on this process.

Figure 2.8: Yra1 pre-mRNA splicing analysis by qRT-PCR in ΔE3 ligase mutants.

RT-qPCR analysis of YRA1 mRNA (A), YRA1 pre-mRNA (B), YRA1 total-RNA (C) (pre-mRNA and mRNA) levels in HA-YRA1 WT shuffled in WT and Δedc3 background of, Δ^tom1, Δ^slx5, Δ^slx8, Δslx5Δtom1, Δslx8Δtom1 and Δslx5Δslx8 mutants. Data were normalized to the ncRNA SCR1 and relative values to HA-YRA1 WT in WT background are shown. The YRA1 pre-mRNA/YRA1 mRNA and YRA1 pre-mRNA/YRA1 total-RNA ratios are shown in D/E and F. *=p value<0.05 **=p value<0.01 in blue when referring to HA-YRA1 WT in WT background and in red when referring to HA-YRA1 WT in Δedc3 background. Average and standard deviations of two independent experiments are shown.

Experiment set up by Valentina Infantino and performed by Ivona Bagdiul.

The aim of my project was to elucidate the potential mRNA export independent role of Yra1 ubiquitination by Slx5-Slx8. Unfortunately, it was not possible to define the

minimal region of Yra1 required for ubiquitination because the lysine usage is redundant. The only yra1 mutant lacking ubiquitination is the HA-yra1allKR, which is not useful to define the function of Yra1 ubiquitination by Slx5-Slx8 since it has been shown by our laboratory (Iglesias et al., 2010) that its sickness can be rescued by Δmlp1 and Δmlp2, suggesting that the HA-yra1allKR may be impaired in mRNA export. In addition, the same study (Iglesias et al., 2010) has shown that the HA-yra1KR9-22, mutant that presents less lysine mutated than the HA-yra1allKR, has an a mRNA export defect of ~30%. Therefore, we decided to investigate the potential effect of Yra1 overexpression in the pathways affected by Slx5-Slx8 to elucidate a possible mRNA export-independent role of Yra1 that may be linked to Slx5-Sx8 regulation.