Transcriptome analysis of alternative splicing eventsregulated by SRSF10 reveals position-dependentsplicing modulationXuexia Zhou1y Wenwu Wu1y Huang Li1 Yuanming Cheng1 Ning Wei1 Jie Zong2

Xiaoyan Feng1 Zhiqin Xie1 Dai Chen2 James L Manley3 Hui Wang14 and

Ying Feng14

1Key Laboratory of Food Safety Research Institute for Nutritional Sciences Shanghai Institutes forBiological Sciences Chinese Academy of Sciences Shanghai 200031 China 2Novel Bioinformatics Co LtdShanghai China 3Department of Biological Sciences Columbia University New York NY 10027 USA and4Key Laboratory of Food Safety Risk Assessment Ministry of Health Beijing 100021 China

Received August 5 2013 Revised December 17 2013 Accepted December 18 2013

ABSTRACT

Splicing factor SRSF10 is known to function as asequence-specific splicing activator Here we usedRNA-seq coupled with bioinformatics analysis toidentify the extensive splicing network regulated bySRSF10 in chicken cells We found that SRSF10promoted both exon inclusion and exclusion Motifanalysis revealed that SRSF10 binding to cassetteexons was associated with exon inclusion whereasthe binding of SRSF10 within downstream constitu-tive exons was associated with exon exclusion Thispositional effect was further demonstrated by themutagenesis of potential SRSF10 binding motifs intwo minigene constructs Functionally many ofSRSF10-verified alternative exons are linked topathways of stress and apoptosis Consistent withthis observation cells depleted of SRSF10 expres-sion were far more susceptible to endoplasmic re-ticulum stress-induced apoptosis than control cellsImportantly reconstituted SRSF10 in knockout cellsrecovered wild-type splicing patterns and consider-ably rescued the stress-related defects Togetherour results provide mechanistic insight intoSRSF10-regulated alternative splicing events in vivoand demonstrate that SRSF10 plays a crucial role incell survival under stress conditions

INTRODUCTION

Alternative splicing (AS) is a process by which differentcombinations of exons can be joined together to produce a

diverse array of messenger RNA (mRNA) isoforms froma single primary transcript This process is a commonmechanism for controlling gene expression and generatingprotein diversity in higher eukaryotic cells In humansgt95 of multiexonic protein-coding genes undergo AS(1) however a precise understanding of how theseevents are regulated and coordinated in vivo is stilllacking Given that AS plays a key role in the regulationof gene expression and that aberrant splicing has beenimplicated in a wide range of human diseases (23) it isimportant to fully understand the regulatory mechanismsby which AS is regulated in vivoThe inclusion of an alternative exon in mature mRNA

is largely dependent on the recognition and usage of theflanking splice sites by the splicing machinery Thisinvolves many cis-acting RNA elements known asexonic or intronic splicing enhancers and silencers whichare bound by different regulatory proteins Serinearginine-rich (SR) and heterogeneous nuclearribonucleoprotein (hnRNP) proteins are two well-characterized RNA-binding proteins that interact withthe enhancers or silencers and subsequently stimulate orrepress the inclusion of an alternative exon respectively(4ndash6) However recent genome-wide analyses revealedthat these regulatory proteins can modulate both exonactivation and repression in vivo which is likely dependenton their binding location within the pre-mRNA (7ndash9)Therefore it seems clear that RNA-binding proteinswidely use this strategy to regulate splicing in the cellHowever splicing-sensitive arrays and CLIP-seq analysisof SRSF1- and SRSF2-regulated splicing revealed thatspecific splicing outcomes were not directly linked to thepositional effect of SR protein binding in vivo Ratherthese outcomes were mainly determined by the balance

To whom correspondence should be addressed Tel +86 21 54920965 Fax +86 21 54920965 Email fengyingsibsaccn

yThese authors contributed equally to the paper as first authors

Published online 17 January 2014 Nucleic Acids Research 2014 Vol 42 No 6 4019ndash4030doi101093nargkt1387

The Author(s) 2014 Published by Oxford University PressThis is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (httpcreativecommonsorglicensesby-nc30) which permits non-commercial re-use distribution and reproduction in any medium provided the original work is properly cited For commercialre-use please contact journalspermissionsoupcom

of cooperation and competition between the SR proteinsand other regulatory proteins in vivo (10)SRSF10 (formally known as SRp38) is an atypical SR

protein that functions differently from the standard SRproteins Although its domain organization is similar tothat of other SR proteins SRSF10 was found to be unableto activate the splicing of a model substrate b-globin pre-mRNA in an in vitro assay (11) SRSF10 was initiallyidentified as a general splicing repressor that is activatedby dephosphorylation (1112) Subsequent studies showedthat phosphorylated SRSF10 acts as a sequence-specificactivator (13) Consistent with this role SRSF10 wasfurther demonstrated to be a key regulator of ASinfluencing the selection of mutually exclusive exons inthe GluR-B pre-mRNA in a sequence-specific manner(1314)In this study we took advantage of the previously

generated SRSF10-deficient (KO) chicken DT40 cells(12) and globally analyzed the AS events regulated bySRSF10 using an RNA-seq approach We found thatSRSF10 activates both exon inclusion and exclusionand it is involved in all of the common modes of ASin vivo Motif analysis revealed that SRSF10-activatedand -repressed exons could be distinguished by the ar-rangement of SRSF10-enriched sequence motifsSRSF10 binding to the alternative exon resulted in exoninclusion whereas SRSF10 binding in the region of thedownstream constitutive exon led to exon exclusion Thispositional effect was further demonstrated by mutatingthe SRSF10 binding motifs of two minigene reportersImportantly the insertion of SRSF10 consensus sequenceswithin an SRSF10-repressed exon was sufficient to induceexon inclusion Our results thus provide substantialinsight into mechanisms of SRSF10-regulated AS eventsin vivoFunctionally a large body of literature revealed that

many of SRSF10-regulated alternative exons areassociated with pathways of stress and apoptosisConsistent with this finding SRSF10 KO cells displayedelevated levels of apoptosis in response to endoplasmicreticulum (ER) stress than wild-type (WT) cellsStrikingly induction of SRSF10 in KO cells recoveredthe WT splicing patterns for the majority of SRSF10-verified splicing events and considerably rescued the ERstress-related defects These findings thus demonstratethat SRSF10 is required for cell survival under stress con-ditions probably reflecting abilities of SRSF10 inregulating a large subset of stressapoptosis-relatedsplicing events

MATERIALS AND METHODS

RNA sequencing

The complementary DNA (cDNA) libraries for paired-end sequencing were prepared using mRNA-Seq SamplePrep Kit (Illumina) according to the manufacturerrsquosinstructions After selecting 200-bp fragments byagarose gel electrophoresis Illumina paired-endsequencing adapters were ligated to the DNA fragmentsfor polymerase chain reaction (PCR) amplification and

sequencing with an Illumina HiSeqTM 2000 system Theraw sequence data and processed data have beensubmitted to Gene Expression Omnibus with accessionnumber GSE53354

Mapping of paired-end reads

Before read mapping clean high-quality reads wereobtained from the raw reads by removing the adaptor se-quences reads with gt5 ambiguous bases (noted as N)and low-quality reads containing more than half of thebases with qualities of lt20 The clean reads were thenaligned to the reference chicken genome (releasegalGal4) and the corresponding recently annotated genes(galGal40) using the TopHat program (v206) In thealignment preliminary experiments were performed tooptimize the alignment parameters (ndashread-mismatches 3ndashread-edit-dist 4 ndashread-realign-edit-dist 0 ndashmate-inner-dist 20 ndashmin-anchor 6 ndashcoverage-search ndashmicroexon-search ndashsolexa13-quals ndashb2-very-sensitive) to provide themost information on the AS events

Transcriptome reconstruction

The recent transcript annotation of the chicken genomesequence (galGal40) was obtained from ftpftpncbinlmnihgovgenomesGallus_gallusGFF KO and WTtranscriptomes were reconstructed respectively bycufflinks using the parameters (ndashGTF-guidegalGal40gff ndashfrag-bias-correct galGal4fa ndashmulti-read-correct ndashupper-quartile-norm) Then these two transcrip-tomes were merged with the original version by cuffmergeusing the parameters (ndashref-gtf galGal40gff ndashref-sequencegalGal4fa KO_WT_transcriptomestxt) to yield compre-hensive reannotated transcripts for subsequent ASanalysis

AS analysis

After mapping an lsquoaccepted_hitsbamrsquo file was generatedwhich includes information regarding the chromosomeposition for exonic reads and exonndashexon junction readsAll junction reads must meet two criteria (i) at least 6 ntof a read must match perfectly to each of the two flankingregions of a potential junction site and (ii) a junction sitethat has gt3 non-redundant reads in both the WT and KOsamples must be filtered Afterwards a series of Javaprograms implemented into the software named ASD(AS detector) were developed to fulfill the followingtasks (i) to reconstruct exon-clusters based on the afore-mentioned reannotated chicken transcriptome for identi-fication of five common modes of AS events for eachexon-cluster (ii) to count the number of junction readsthat align either to the inclusion or exclusion isoforms inboth the WT and KO samples respectively and calculatea P-value using junction read-counts between KO and WTsamples by Fisher exact test (iii) to calculate readcoverage for the alternative exon and its correspondinggene in both WT and KO samples respectively and cal-culate a second P-value by Fisher exact test based on thealternative exon read coverage relative to its gene readcoverage between WT and KO samples (iv) to combinethe two P-values to get an adjusted P-value using a

4020 Nucleic Acids Research 2014 Vol 42 No 6

weighted arithmetic equation (615) (see SupplementaryMethod) for assessing the statistical difference of ASbetween the two samples ASD is available at httpwwwnovelbiocomasdASDhtml

Detection of GA-rich region and motif analysis ofSRSF10-regulated exons

Based on SELEX-identified (11) and RNAcompete-predicted consensus sequences (16) for SRSF10 and alsotaking into account the degenerate nature of RNA-binding sequences we arbitrarily used GA-rich 6-mersthat include at least one G and two As with GA contentgt80 as potential SRSF10 binding sites Significantlyenrichments of GA-rich regions were identified bycounting the occurrences of each 6-mer (46 total occur-rences=4096 combinations) within five segments ofeach transcript with a SRSF10-regulated cassette exon(including the cassette exon and the two flanking consti-tutive exons and introns) The cassette exons which werenot affected by SRSF10 were treated as the controlgroup A hypergeometric test (Fisherrsquos right-tailed exacttest) was performed to detect the enrichment of eachhexamer sequence within the sequences undergoingSRSF10-regulated inclusion (or exclusion) events againstthe control sequences Among the 4096 combinations aBonferroni-adjusted P-value (or false discovery rateFDR) was calculated by multiplying the actual totalnumber of tests Subsequently the sums of the minuslog2-transformed P-values were calculated for GA-rich6-mers as a metric to determine the peaks of GA-rich6-mers within the five regions around the regulatedcassette exons Among the identified GA-rich peaks theoverrepresented GA-rich 6-mers (Plt 00001) along withflanking 5 nt on each side (a total of 16 nt) were furthercollected and submitted to Multiple EM for MotifElicitation (MEME) suite (17) for possible GA-richmotif analysis The parameter setting for MEME was anumber of single motif per sequence of 0 or 1 a minimummotif width of 5 a maximum motif width of 8 amaximum number of different motifs of 3 and a searchof only the given strand Finally MEME searching motifenrichments were manually edited to have a length of 6-mer

RNA isolation and RT-PCR validation

RNA extraction reverse-transcription and PCR amplifi-cation were all carried out as previously described (14)Briefly PCR primers were designed to amplify two ormultiple isoforms with different sizes For complexsplicing events isoform-specific restriction digestion andor DNA sequencing was used to validate PCR productidentity Glyceraldehyde-3-phosphate dehydrogenase(GAPDH) was used as an internal control for normaliza-tion Sequences for primer pairs used in this study arelisted in the Supplementary Table S3

Plasmids construction cell culture and transfection

Minigenes were constructed by amplifying genomicsequences spanning exon 12ndash14 of KDM6A and exon8ndash10 of EIF4ENIF1 which were then inserted into

CMV-Tag2b between the BamH1 and Xho1 sites respect-ively Creating SRSF10-binding motifs on the minigeneconstructs was made as previously described (7)Specifically three copies of the consensus SRSF10binding sequences or random sequences were insertedinto the middle of exon 13 of KDM6A Similar strategywas used to construct other plasmids used in the studyFor deletion mutation constructs sequences for primerpairs were available upon request DT40 cells of differentbackground including WT SRSF-deficient (KO) andSRSF10-deficient containing exogenously expressed hem-agglutinin-tagged SRSF10 (KO-re) were maintained es-sentially as described previously (12) Lipofectamine2000(Invitrogen) was used for plasmid transient transfectionaccording to manufacturerrsquos instructions

Cells treated with ER stress inducer apoptosis assay andwestern blotting

Cells were treated with the ER stress inducer (eitherthapsigargin 05 uM or tunicamycin (TM) 2 ugml) for28 h Cell apoptosis was determined by propidium iodidestaining and flow cytometry For recovery after 16 h treat-ment with ER stress inducer cells were collected andwashed with phosphate-buffered saline once then main-tained in fresh medium Numbers of surviving cells weredetermined by trypan blue exclusion Actin monoclonalantibodies were purchased from Santa Cruz humaneIF2 [pSer52] antibody was from Biosource Westernblotting was performed as previously described (18)

Statistic analysis

The data represent meanplusmnSD format from at least threeindependent experiments except where indicatedStatistical analysis was performed by Studentrsquos t-test at asignificant level of Plt 005

RESULTS

RNA-seq and transcriptome analysis of WT andSRSF10-deficient DT40 cells

We have previously shown that SRSF10 functions as asequence-specific splicing activator that is capable ofregulating AS both in vitro and in vivo (1314) To under-stand the global profile of the AS networks regulated bySRSF10 we used an RNA-seq approach to analyze theeffects of SRSF10 depletion on splicing regulation inchicken DT40 cells (12) Briefly WT cells and SRSF10-KO cells were cultured in standard medium to 90 con-fluence Total RNA was isolated from these cells andcDNA libraries were prepared individually by performinga series of procedures including poly(A) enrichmentRNA fragmentation random hexamer-primed cDNAsynthesis linker ligation size selection and PCR amplifi-cation (Supplementary Figure S1A) The libraries werethen sequenced in a 2 100-bp format using an IlluminaHiSeq instrumentAfter quality filtering we generated a total of 53 001 728

and 53 138 030 clean reads for the WT and KO cells re-spectively (Table 1) Then we used TopHat (v206) to

Nucleic Acids Research 2014 Vol 42 No 6 4021

align these reads against the chicken genome sequence(galGal4) and its corresponding annotated genes(galGal40gff3) (Supplementary Figure S1B) Thepaired-end reads had an approximate insert size of192plusmn16bp (Supplementary Figure S2) which is consist-ent with the size of the cDNA selected during the librarypreparation Given that accurate mapping of reads thatspan splice junctions is a critical component for the iden-tification of AS events a series of experiments were per-formed to optimize the alignment parameters that wouldensure the number of reads mapping to splice junctions(see the lsquoMaterials and Methodsrsquo Section)After proper optimization a total of 46 152 593 (87)

and 45 354 730 (85) mapped reads were generated forthe WT and KO cells respectively Almost 99 of boththe WT and KO reads aligned to the reference genome in aunique manner whereas lt1 filtered as multi-mappedreads Among the mapped reads 15 314 481 WT(332) and 15 505 045 KO (342) reads were alignedagainst splice junctions (Table 1) A comprehensive anno-tation of the potentially novel transcripts relative to theknown transcripts is indispensable for dissecting potentialAS events occurring in the chicken transcriptome To thisend Cufflink packages were used to reconstruct thechicken transcriptome based on the read distribution ofthe genomic sequences and then merge it with the currentreference database The final merged chicken transcrip-tome had a significantly expanded number of annotatedtranscripts and was sufficient for the following AS analysis(Supplementary Figure S1B)

Identification of SRSF10-affected splicing events

There are five major modes of AS events described inmetazoan organisms including cassette exons alternative50 splice sites alternative 30 splice sites mutually exclusiveexons and retained introns (Figure 1A) To identify thechanges in these five AS events on the loss of SRSF10 wedeveloped a Java program that can search and locate setsof alternative exons and junctions from the genomic co-ordinates of all annotated chicken transcripts Exonicreads and junction reads that align either to the inclusionor exclusion isoforms in both the WT and KO sampleswere counted respectivelyComparison of the junction read counts between the

WT and KO samples can reveal whether a change in thesplicing pattern has occurred due to the depletion ofSRSF10 this would be further adjusted by comparison

of exon read coverage between the WT and KO cells(see lsquoMaterials and Methodsrsquo Section) With an adjustedP-value cutoff of lt005 we were able to identify 173 ASevents from a total number of 10 652 splicing events thatchanged significantly in the KO samples Specifically themajority of the affected splicing events (130) belonged tothe cassette exon category 37 of the remaining events fellinto other splicing modes including 16 for alternative 50

splice sites 17 for alternative 30 splice sites 3 for mutuallyexclusive exons and 1 for intron retention In additionanother six events (including coordinated two cassetteexons and complicated splicing events) were placed intothe lsquoundefinedrsquo category as they could not be classified asany AS mode (Figure 1B and Supplementary Table S1)In summary SRSF10 is involved in all of the commonmodes of AS and cassette exons are the most frequenttargets for SRSF10 regulation

Validation of SRSF10-affected splicing events

To verify whether these predicted AS events are realtargets regulated by SRSF10 we designed primer pairsthat were able to detect alternative exons in 45 potentialtarget transcripts which were chosen based on bothP-values and read coverage signals These primers wereused to analyze the RNA isolated from the WT and KOcells by reverse transcriptase-polymerase chain reaction(RT-PCR) Although there were no or only minorchanges in the splicing patterns in the 13 transcriptsexamined (t-test Pgt 005 date not shown) 32 were sig-nificantly different between the WT and KO samples

Cassee exonA

Alternave 5rsquo splice site

Alternave 3rsquosplice site

Mutually exclusive exon

Retained intron

Type of AS of different AS events of affectedB

Cassette exon 3211 130

Alternative 5rsquo splice site 1606 16

Alternative 3rsquo splice site 1738 17

Mutually exclusive exon 251 3

Retained intron 1089 1

Undefined 2757 6

Total 10652 173

Figure 1 Common AS events and the SRSF10-affected splicing eventsobserved in DT40 cells (A) Diagrams of the exclusion and inclusionisoforms for the five modes of AS events that were examined Whiteboxes flanking constitutive exons black boxes alternative splicedexonsregions solid lines splice junctions supporting the inclusionisoform dotted lines splice junctions supporting the exclusionisoform (B) Summary of the different AS events identified in DT40cells and the splicing events affected by SRSF10 depletion as revealedby analysis of RNA-seq data

Table 1 Summary of RNA-seq data mapping results

Sample WT KO

Clean reads 53 001 728 (1000) 53 138 030 (1000)Mapped reads 46 152 593 (871) 45 354 730 (854)Unique mapped readsa 45 512 948 (986) 44 766 437 (987)Multi-mapped readsb 639 645 (14) 588 293 (13)Junction readsc 15 314 481 (332) 15 505 045 (342)Unmapped reads 6 847 282 (129) 7 782 321 (146)

aThat include paired-end reads with an unalignable matebWith multiple alignmentscAll junction reads have 3 offset alignments

4022 Nucleic Acids Research 2014 Vol 42 No 6

(t-test Plt 005 Supplementary Table S2) indicating ahigh consistency (73) between the RNA-seq predictionsand RT-PCR results in our study Unexpectedly 24 (72)events contained unannotated splice junctions The highpercentage of unannotated transcripts indicated that thecurrently available chicken transcript annotations are in-complete Table 2 shows the top 12 splicing events thatwere dramatically affected by SRSF10 (t-test Plt 0001)Figure 2 shows representative examples of these 12 eventswith RNA-seq read coverage RT-PCR results and quan-tification of their RNA products measured as the inclu-sionexclusion (InEx) ratio Specifically the WT cellspredominantly expressed cassette exon-containingmRNA isoforms such as for transcripts encodingEIF4ENIF1 and CAST however in the KO cellsSRSF10 depletion significantly increased exonskipping to nearly 90 (ETF4ENIF1) and 60 (CAST)(Figure 2A)

These results are in agreement with previous findingsthat SRSF10 functions as a splicing activator to promoteexon inclusion in vivo However the increased inclusion ofexon 31 of the EPRS pre-mRNA and exon 13 of theKDM6A pre-mRNA was observed in the SRSF10-defi-cient cells (Figure 2B) suggesting that SRSF10 can alsoactivate exon exclusion in vivo In addition the selectionof alternative 30 and 50 splice sites was similarly affected bySRSF10 This was illustrated by two cases in which theinclusion of the BAP1 alternative 50 exon and the CDK13alternative 30 exon were significantly decreased on SRSF10depletion (Figure 2C and D) Taken together these resultsprovide strong evidence that SRSF10 can promote bothexon inclusion and exclusion in vivo

Restoration of WT-splicing patterns by exogenouslyexpressed SRSF10 in KO cells

We next tested whether exogenously expressed SRSF10could rescue the splicing variation observed in the KOcells To this end we first examined the splicing patternsof 32 validated transcripts in KO-re cells which highlyexpress flu-tagged SRSF10 (12) along with thoseobtained from WT and KO cells for comparison

Strikingly the induction of SRSF10 restored the WTsplicing patterns of all of the examined transcripts withone exception the inclusion of EPRS exon 11 was notdecreased to the WT level on SRSF10 re-expression(data not shown) Representative RT-PCR results of theexamined transcripts and quantification of their RNAproducts are shown in Figure 3The levels of SRSF10-activated exons were all increased

to nearly the WT levels for the EIF4ENIF1 CASTand PHF20L1 transcripts on SRSF10 re-expression(Figure 3A compare lane 3 with lanes 1 and 2 for eachtarget) In contrast a decreased inclusion ratio wasobserved for the SRSF10-repressed exons such as forthe transcripts encoding CEP44 L3MBTL2 andKDM6A (Figure 3B compare lane 3 with lanes 1 and 2for each target) Similar restoration was reproduced forother types of alternative exons such as the alternative 50

exons of SLAM and BAP1 (Figure 3C compare lane 3with lanes 1 and 2 for each) and the alternative 30 exon ofCDK13 (Figure 3D compare lane 3 with lanes 1 and 2)These results strongly suggest that SRSF10 is directlyinvolved in the AS regulation of these significant tran-scripts in vivoWe then determined whether the minor changes that

were observed in several transcripts on SRSF10 deficiencycould be rescued by the re-expression of SRSF10Surprisingly we found that these minor changes revertedback to the normal WT splicing levels in KO-re cells aswas observed for the MEF2A mutually exclusive exon(Figure 3E compare lane 3 with lanes 1 and 2) Thisresult provides evidence that SRSF10 also regulates theAS of pre-mRNAs such as MEF2A in vivo althoughthe output effect can be negligible Together thesefindings demonstrate that SRSF10 exerts differential acti-vation or repression effects on the various splicing eventsin vivo

Position-dependent activity revealed by SRSF10 bindingmotif analysis

Next we investigated whether the distribution of SRSF10binding motifs differs between SRSF10-activated and

Table 2 List of top 12 splicing events affected by SRSF10

Gene symbol Gene description AS type Annotation of twospliced variantsa

EIF4ENIF1 Eukaryotic translation initiation factor 4E nuclear import factor 1 Exon inclusion A+UCAST Calpastatin Exon inclusion A+UCASP1 Caspase 1 apoptosis-related cysteine peptidase Exon inclusion A+UPHF20L1 PHD finger protein 20-like 1 Exon inclusion A+UHSF2 Heat shock transcription factor 2 Exon inclusion A+AEPRS Glutamyl-prolyl-tRNA synthetase Exon exclusion A+AKDM6A Lysine (K)-specific demethylase 6A Exon exclusion A+UMDM4 Mdm4 p53 binding protein homolog (mouse) Exon exclusion A+UCDK13 Cyclin-dependent kinase 13 Alt 30 SS A+USLTM SAFB-like transcription modulator Alt 50 SS A+UBAP1 BRCA1-associated protein-1 Alt 50 SS A+URBBP5 Retinoblastoma binding protein 5 Alt 50 SS A+U

Alt alternative SS splice siteaBased on chicken NCBI and Ensemble databases lsquoUrsquo stands for unannotated splice transcripts whereas A stands for annotated transcripts

Nucleic Acids Research 2014 Vol 42 No 6 4023

-repressed cassette exons To this end we simply definedGA-rich hexamers as potential SRSF10 binding sitesbased on the SELEX-identified (11) and RNAcompete-predicted (16) sequences for SRSF10 (see lsquoMaterials andMethodsrsquo Section) We then searched five transcriptregions including the cassette exon the two flanking con-stitutive exons and the two flanking introns for GA-richhexamer enrichments among 78 transcripts with SRSF10-activated exons and 52 transcripts with SRSF10-repressedcassette exons Another 80 SRSF10-unaffected eventsserved as controls As shown in Figure 4A no GA-richhexamers were found in the upstream exon or in the twoflanking introns for both groups However someoverrepresented GA-rich hexamers (eg GAAAGT GAAGAA and AGAAAG) were only found within theSRSF10-activated exons whereas other overrepresentedGA-rich hexamers (eg GAAAAG AGAAA and GAAAAT) were found in the flanking downstream constitutive

exons for both groups (Figure 4B) These overrepresentedGA-rich 6-mers was further analyzed by MEME forpossible SRSF10 binding motifs (Figure 4C seelsquoMaterials and Methodsrsquo Section)

In summary SRSF10-activated exons showed a pre-dominant enrichment of GA-rich hexamers withinthe exons themselves over the flanking downstream con-stitutive exon (Figure 4A) However SRSF10-repressedexons were associated with the enrichment of GA-richhexamers only in the downstream constitutive exon(Figure 4A) Thus the binding of SRSF10 to the alterna-tive exon was associated with exon inclusion whereas thebinding of SRSF10 within the downstream constitutiveexon was associated with exon exclusion This positionaleffect as to inducing exon inclusion on the cassette exonand skipping on the downstream exon was only previouslydemonstrated with typical SR protein SRSF1 using MS2tethering approach (19)

EPRS6 0

B

2 4EIF4ENIF1

A

79nt

00

20

40

60

RNA-seq reads

31

InE

x

00

06

12

18

24 180nt

RNA-seq reads

9

InE

x

WT

KOWT KO

WT

KOWT KO

0306091215

135nt

KDM6A

RNA-seq reads

13

InE

x

05

10

15

20

25 CAST84nt

RNA-seq reads

13

InE

x

0

WT KO

WT

KO

00

WT KO

WT

KO

15

2 BAP1

Ex

54nt

C

06

09

12

180nt

CDK13

Ex

D

0

05

1

WT KO

WT

KO

RNA-seq reads

In

q

0

03

WT

KO

RNA-seq reads

In

WT KO WT KO

Figure 2 SRSF10 can promote both exon inclusion and exclusion in vivo SRSF10-activated cassette exons (A) SRSF10-repressed cassette exons(B) SRSF10-activated alternative 50 exons (C) and SRSF10-activated alternative 30 exons (D) were identified by RNA-seq analysis and validated byRT-PCR Each case is schematically diagrammed with mapped RNA-seq reads that cover the corresponding exons (left panels) Black boxesalternative exons black lines introns white boxes flanking constitutive exons The primer pairs used for RT-PCR are indicated by arrowsSRSF10-regulated AS changes were analyzed by RT-PCR (right bottom panels) Quantification of the RNA products was measured as the inclu-sionexclusion (InEx) ratio as shown in the top right histogram Values shown are the meanplusmnSD n=3 Significance for each detected change wasevaluated by Studentrsquos t-test

4024 Nucleic Acids Research 2014 Vol 42 No 6

Mechanistic insights into SRSF10-regulated exoninclusion and exclusion

For SRSF10-activated exons potential SRSF10 bindingmotifs were observed predominantly within cassette exonsover downstream exons We thus reasoned that thebinding of SRSF10 to cassette exons is associated withexon inclusion To test this we analyzed individual casesthat displayed SRSF10-dependent exon inclusion such asEIF4ENIF1 exon 9 We constructed a minigene plasmidin which a genomic DNA fragment spanning EIF4ENIF1exons 8ndash10 was preceded by the CMV promoter andfollowed by an SV(40) polyA site (Figure 5A top)Splicing was assayed following transient expression in

DT40 cells Total RNA was extracted from the transfectedcells and exon inclusionexclusion was analyzed byRT-PCR (Figure 5B) Consistent with its endogenoussplicing pattern the EIF4ENIF1 exon 9 was almost fullyincluded in the WT cells (lane 1) whereas SRSF10 deple-tion significantly decreased the inclusion of exon 9 tonearly 30ndash40 (lane 2) indicating that inclusion ofEIF4ENIF1 exon 9 was SRSF10 dependent We identifiedfour GA-rich clusters GA1 and GA2 within exon 9(Figure 5A bottom) whereas GA3 and GA4 withinexon 10 (Supplementary Figure S3A)We next examined in more detail the role of the indi-

vidual GA elements in exon inclusion To this end weintroduced a single GA deletion into the minigene

A

0

1

2

3 EIF4ENIF1 WT KO KO-re

WT KO KO-re1 2 3

0

1

2

3CAST

WT KO KO-re

WT KO KO-re1 2 3

0

2

4

6PHF20L1 WT KO KO-re

WT KO KO-re1 2 3

C

SLTM

0

2

4

6 WT KO KO-re

WT KO KO-re1 2 3

BAP1

0

1

2

3

WT KO KO-re

WT KO KO-re1 2 3

B

0

05

1

15CEP44

WT KO KO-re

WT KO KO-re1 2 3

0

03

06

09L3MBTL2

WT KO KO-re

WT KO KO-re1 2 3

0

05

1

15 KDM6A WT KO KO-re

WT KO KO-re1 2 3

0

15

3

45D

CDK13 WT KO KO-re

WT KO KO-re1 2 3

E

MEF2A

0

2

4

6

WT KO KO-re

WT KO KO-re1 2 3

Figure 3 Exogenous expression of SRSF10 in KO cells restores WT splicing patterns of SRSF10-targeted transcripts SRSF10-activated cassetteexons (A) SRSF10-repressed cassette exons (B) alternative 50 exons (C) alternative 30 exons (D) and mutually exclusive exons (E) were analyzed forsplicing changes in WT KO and KO-re cells by RT-PCR (left panels) The primers used for RT-PCR are the same as in Figure 3 Bar graphs showthe ratios of inclusion versus exclusion (AndashD) or exon a versus exon b (E) (right panels) Values shown are the meanplusmnSD n=3 The significance ofeach detected change was evaluated by Studentrsquos t-test

Nucleic Acids Research 2014 Vol 42 No 6 4025

construct and analyzed the effects of each mutant on theexon inclusion in vivo As shown in Figure 5C deletion ofGA1 led to nearly 80 exon exclusion which is similar tothe effects caused by SRSF10 deficiency (compare lanes 3and 4 with lanes 1 and 2) In contrast the other threemutants displayed no effect on exon exclusion and stillremained responsive to SRSF10 deficiency (comparelanes 5ndash6 and Supplementary Figure S3B) Takentogether these results provide strong evidence that theGA1 element has SRSF10-dependent enhancer activityTo provide direct evidence that SRSF10-dependent

enhancer activity within the GA1 fragment can lead toexon inclusion we made two other constructs in whichthree copies of the SRSF10 consensus motif or random

sequences were inserted in the middle exon of the GA1plasmid (Figure 5B) When SRSF10 binding motifs werepresent in the GA1 plasmid exon inclusion was restorednearly to the WT level compared with the controlsequence (Figure 5C compare lane 9 lane 7 and lane 1)However the differential effects on exon inclusionbetween the two inserted sequences were completely di-minished on SRSF10 depletion (Figure 5C comparelane 10 and lane 8)

For SRSF10-repressed exons a key insight from themotif analysis was the enrichment of SRSF10 bindingmotifs within the downstream constitutive exon To inves-tigate the possible role of downstream SRSF10 bindingmotifs in the regulation of exon exclusion we selected

A

B

C

Figure 4 Hexamer and MEME analysis of SRSF10-regulated exon inclusion and exclusion (A) The sum of the minus log2-transformed P-values of GA-rich hexamers within the five regions around regulated cassette exons was compared with control cassette exons that are notaffected by SRSF10 The red line represents SRSF10-mediated exon inclusion compared with the control (left) and the blue line representsSRSF10-mediated exon exclusion compared with the control (right) There is no obvious peak in the upstream exon or the two flanking intronsaround the cassette exons (B) A representative list of top GA-rich hexamers was shown for SRSF10-activated cassette exons -activateddownstream constitutive exons and -repressed downstream constitutive exons respectively FC is fold change P-value stands for statisticallysignificance of GA-rich hexamers between SRSF10-regulated group and controls and FDR indicates false discovery rate Shown in brackets arethe number of overrepresented GA-rich 6-mers and the percentage of exons containing at least one overrepresented 6-mer (C) Potential GA-richmotifs derived from the aforementioned overrepresented GA-rich 6-mers using MEME (Plt 00001)

4026 Nucleic Acids Research 2014 Vol 42 No 6

an exon from the KDM6A gene which displayed exonexclusion in the WT cells but showed increased inclusionon SRSF10 depletion (Figure 2B) This effect wasreproduced in the transient transfection of the WT orKO cells with a KDM6A minigene construct (Figure 5Flanes 1 and 2) Although potential SRSF10 binding motifswere not observed within cassette exon 13 two GA-richelements (named GA5 and GA6) were identified withindownstream constitutive exon 14 (Figure 5D bottom)Although the deletion of GA5 had little effect on exoninclusion the deletion of GA6 increased exon inclusionfrom 20 to nearly 40 (Figure 5F compare lanes 5and 6 with lanes 3 and 4) These results suggest that theGA6 motif partially contributes to exon 13 exclusionImportantly insertion of SRSF10 consensus motifs inthe exon 14 restored the WT splicing pattern of GA6substrate (Figure 5E and F compare lanes 7ndash10 withlanes 1ndash2) further demonstrating that SRSF10-dependent

binding motifs within the constitutive exon resulted inexon skipping

Induction of exon inclusion by the manipulation ofSRSF10 binding motifs

We next considered whether the introduction of SRSF10binding sequences in a naturally excluded exon would besufficient to induce inclusion To this end mutations weremade in the KDM6A minigene construct to allow for theinsertion of three copies of the SRSF10 consensus motif orrandom sequences just downstream of the 30 splice site inthe cassette exon (Figure 5E) As shown in Figure 5F90 of the mutated exon with the inserted SRSF10binding motifs was included compared with the 20exon inclusion observed with the control sequence(Figure 5F compare lane 13 and 11) Moreover theeffect on exon inclusion critically depended on themotifs that were directly bound by SRSF10 as it was

GTCTAGAGCAAGCCATTCTCTCCCCTGGCCAGAACTCTGGAAACTACTTTGCTCCCATTCCATTGGAAGACCACTCTGAAAACAAAGTGGACATCCTAGAAATGCTACAGAAAGCCAAAGTGGACTTAAAACCTCTTCTCTCAAGTCTTTCAGCCAACAAGGAAAAGCTTAGAGAGAGCA

(GA1)

(GA2)

exon9

exon9

EIF4ENIF1 minigeneA

B GA1-Random

GA1-SRSF10

KDM6A minigene

exon13

AATACTTCAGATGTTTGGAGCAGTGGCCATACAGTGTCACATCCTCCTGTACAACAACAAATTCATTCTTGGTGCTTGACACCACAGAAATTACAG

exon14(GA5)

(GA6)

exon14

D

KDM6A-SRSF10 KDM6A-Random

E 14

GA6-SRSF10 GA6-Random

C

WT KO WT KO WT KO WT KO WT KO

0

02

04

06

08

1WTKO

Rela

ve

exon

incl

usio

n

WT KO WT KO WT KO WT KO WT KO WT KO WT KO

1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14

F

0

02

04

06

08

1WTKO

Rela

ve

exon

incl

usio

n

DD D

D

Figure 5 Characteristics of SRSF10-regulated exon inclusion and exclusion (A) Schematic representation of the EIF4ENIF1 minigene construct(top) The intact genomic region of EIF4ENIF1 from exon 8ndash10 was cloned into the CMV-Tag2b vector The exon 9 sequences are shown at thebottom with two GA-rich motifs (GA1 and GA2) marked with a different color The PCR primers used to detect the alternatively spliced productsare indicated by arrows (B) Three copies of the SRSF10 consensus binding motif or random sequences were inserted within the middle exon of theGA1-deletion construct (C) RT-PCR analysis of the EIF4ENIF1 minigene or each of the indicated mutants transfected in WT or KO cellsrespectively The percentage of exon inclusion is shown at the bottom of the gel Values shown are the meanplusmnSD n=3 (D) Schematic repre-sentation of the KDM6A minigene construct (top) The intact genomic region of KDM6A from exon 12ndash14 was cloned into the CMV-Tag2b vectorThe exon 14 sequences are shown at the bottom with two potential GA-rich motifs (GA5 and GA6) The PCR primers are indicated by arrows (E)Three copies of the SRSF10 consensus binding motif or random sequences were either inserted within exon 14 using the GA6-deletion construct (top)or within exon 13 of the KDM6A minigene (bottom) (F) RT-PCR analysis of the KDM6A minigene or each of indicated mutant transfected in WTor KO cells respectively The histogram below shows mean percentage of exon inclusionplusmnSD n=3

Nucleic Acids Research 2014 Vol 42 No 6 4027

diminished in response to SRSF10 depletion (Figure 5Flane 14) Taken together these results suggest that theactivation activity of SRSF10 binding sequences withinthe cassette exon is dominant over the repression activityof the downstream exon

SRSF10 protects DT40 cells from ER stress-inducedapoptosis

Through an extensive literature review we learned that alarge subset of genes have functional annotations relatedto stress and apoptosis among SRSF10-regulated ASevents such as genes of HSF2 (20) CASP1 (21) CAST(22) SLTM (23) EIF4ENIF1 (2425) FKBP14 (26)BAP1 (27) SEC16A (28) or MDM4 (29) BecauseSRSF10 is not required for cell survival (12) we ques-tioned whether SRSF10 favors cell growth under stressconditions by regulating these splicing events To thisend we treated WT KO and KO-re cells with differentstress inducers and observed the responses of these celllines to each treatment Although there were no obviousdifferences observed among these cell lines in response tomost stresses ER stress induced by either thapsigargin orTM exerted greater deleterious effects on the KO cells

compared with the WT and KO-re cells As shown inFigure 6A significant apoptosis was observed in the KOcells after 28 h of treatment with TM as measured bypropidium iodide staining and flow cytometry

We further measured the ability of these cell lines torecover after ER stress Whereas both the WT and KO-re cells resumed rapid growth after an 1-day lag the KOcells showed considerably delayed recovery resuminggrowth only after at least a 3-day lag (Figure 6B) Theseresults reflect the loss of SRSF10 rather than an unrelatedproperty of the cells because (i) another independentSRSF10-KO cell line behaved identically to the first(data not shown) and (ii) stable expression of exogenousSRSF10 (KO-re cells) considerably rescued the ER stress-related defects In addition in contrast to heat shock-induced rapid dephosphorylation (12) SRSF10 remainedphosphorylated and dephosphorylated SRSF10 was notdetected regardless of short-term or long-term treatmentof ER stress (Figure 6C) Consistent with this findingnone of SRSF10-dependent exon inclusion or skippingwas reversed under ER stress conditions (SupplementaryFigure S4) Therefore we concluded that it is phosphory-lated SRSF10 that plays a critical role in cell survival after

A

50WT

TM 2μμgml 0hr 28hr

30

40

50 WTKOKO-re

)

WT

10

20

30

Apop

tosi

s (

KO

0

10

0hr 28hr

KO-re

C

TM(1μ

WT KO6

7

lion

)

WT

KO

B

phospho-eIF2α

SRSF10

gml) 0 0 5 2 6 12 24 0 0 5 2 6 12 24 hrs

2

3

4

5

ell c

ou

nts

(m

ill KO-re

α-tubulin

total-eIF2α

0

1

2

day0 day1 day2 day3

Ce

Figure 6 SRp38 protects DT40 cells from ER stress-induced apoptosis (A) WT KO and KO-re cells were treated with the ER stress-inducertunicamycin (TM 2 ugml) for 28 h Apoptosis was measured by propidium iodide staining and flow cytometry (right panel) (B) Cells were treatedwith TM (2 ugml) for 16 h and allowed to recover in medium for 1ndash3 days The number of surviving cells was determined by trypan blue exclusionThe error bars represent the SD (C) WT and KO cells were subjected to TM-induced ER stress for the indicated times Immunoblotting wasperformed with whole cell extracts using antibodies against SRSF10 phospho-eIF2A and b-actin

4028 Nucleic Acids Research 2014 Vol 42 No 6

ER stress perhaps reflecting its ability to regulate manystress- and apoptosis-related AS events

DISCUSSION

In this study we globally analyzed SRSF10-regulated ASevents based on deep sequencing of the chicken transcrip-tome We found that SRSF10 is involved in all of thecommon modes of AS with cassette exons being themost frequent targets for SRSF10 regulation The moststriking finding of our analysis is that SRSF10 showsposition-dependent activity Later we discuss thepossible mechanisms by which SRSF10 regulates exon in-clusion and exclusion compared with the previouslyreported mechanisms of splicing regulation by otherRNA regulatory proteins

We previously demonstrated that SRSF10 functions asa sequence-specific splicing activator by directly binding toexonic splicing enhancers The proposed mechanism forSRSF10 activation involves the stabilization of thebinding of U1 and U2 small nuclear ribonucleoproteins(snRNPs) to the 50 splice site and branch site respectively(13) Consistent with this positive regulation in this studywe identified that SRSF10 activates a large subset of exoninclusions in chicken cell lines including cassette exonsalternative 50 exons and alternative 30 exons Motifanalysis and functional studies conducted on a modelminigene provided strong evidence that when bindingwithin an alternative exon SRSF10 acts as an activatorto promote exon inclusion This was further corroboratedby previous findings in which SRSF10 was shown toenhance the inclusion of a triadin alternative 30 exon bydirectly binding to the exon (14)

However we also observed that SRSF10 can repressexon inclusion in vivo and thus sought to determine themechanism underlying this SRSF10-dependent exon ex-clusion It is possible that when binding occurs withinflanking constitutive exons SRSF10 may stabilize theinteractions of U1 andor U2 snRNPs with the nearbyconstitutive splice sites thereby decreasing the competi-tiveness of the alternative sites Kinetic analysis of splicesite competition has demonstrated that a minor modula-tion of splice site recognition might be converted into amajor functional consequence (30) This suggests that re-pression by SRSF10 results from its usual splicing activa-tor activity with partial activation of the flankingconstitutive sites preventing the cassette exon from inclu-sion A similar principle has only been previouslydemonstrated for the SR protein SRSF1 (19) It is import-ant to note that the manipulation of SRSF10 bindingsequences could significantly switch an excluded exoninto an included exon indicating that SRSF10-dependentactivation activity predominates over its repressionactivity

Recent genome-wide studies of splicing regulatorsincluding the SR proteins hnRNPs and other RNA regu-latory proteins have revealed that they can all functionboth positively and negatively in regulating exon inclu-sion regardless of their predominant splicing activatoror repressor activity which has been previously

demonstrated using model genes (7ndash1031) The oppositesplicing activity could be attributed to their differentbinding locations in the pre-mRNA although the pos-itional effect exerted by individual splicing regulatorsappears to be fundamentally distinct However in somecases the final splicing outcomes are not directly linked tothe positional effect of individual protein binding in vivothus emphasizing the combinatory control of AS in vivoMost AS events are likely subjected to regulation bymultiple different splicing regulators which may act syn-ergistically or antagonistically In this study we observedthe differential effects exerted by SRSF10 in modulatingexon inclusion Whether this reflects differences in theaffinity of SRSF10 for target exons or complex mechan-isms involving SRSF10 and other regulatory proteinsremains to be determinedOur data also revealed that 70 of the validated

splicing variants caused by the loss of SRSF10 have notbeen annotated which indicates that the current chickentranscriptome is incomplete Importantly these newlyidentified variants do not result from aberrant splicingevents as they are also observed to be normally expressedin WT cells albeit at much lower levels than their corres-ponding dominant isoforms As WT cells are more resist-ant to ER stress-induced apoptosis than KO cells thisraises the possibility that SRSF10 protects cells fromstress-induced apoptosis by maintaining the normallevels of these major isoforms in cells By contrastSRSF10 deletion results in isoform switching in theseevents and renders cells susceptible to specific apoptosistriggers This possibility was further demonstrated by theobservation that KO cells with reconstituted SRSF10 re-covered the WT splicing patterns for the majority ofSRSF10-verified splicing events and considerably rescuedthe ER stress-related defects How these novel splicevariants resulting from the loss of SRSF10 contribute toER stress-induced apoptosis is currently unclear and willbe an important focus of future studiesTogether our findings provide mechanistic insights into

SRSF10-regulated exon inclusion and exclusion whichdepend on its binding location with respect to theregulated exons within the pre-mRNA By regulatingmany stress- and apoptosis-related splicing events ourfindings demonstrate that SRSF10 is required for cell via-bility under conditions of stress

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online

FUNDING

Ministry of Science and Technology of China[2012CB524900 and 2012BAK01B00] National NaturalScience Foundation [31170753 and 31070704] and theOne Hundred Talents Program of the Chinese Academyof Sciences China Postdoctoral Science Foundation[2013M541564 to WW] Funding for open accesscharge National Natural Science Foundation [31170753]

Conflict of interest statement None declared

Nucleic Acids Research 2014 Vol 42 No 6 4029

REFERENCES

1 GoodellMA BroseK ParadisG ConnerAS andMulliganRC (1996) Isolation and functional properties ofmurine hematopoietic stem cells that are replicating in vivoJ Exp Med 183 1797ndash1806

2 DavidCJ ChenM AssanahM CanollP and ManleyJL(2010) HnRNP proteins controlled by c-Myc deregulate pyruvatekinase mRNA splicing in cancer Nature 463 364ndash368

3 SinghRK and CooperTA (2012) Pre-mRNA splicing in diseaseand therapeutics Trends Mol Med 18 472ndash482

4 ChenM and ManleyJL (2009) Mechanisms of alternativesplicing regulation insights from molecular and genomicsapproaches Nat Rev Mol Cell Biol 10 741ndash754

5 NilsenTW and GraveleyBR (2010) Expansion of theeukaryotic proteome by alternative splicing Nature 463 457ndash463

6 ChenJJ LinKK HuqueM and AraniRB (2000) Weightedp-value adjustments for animal carcinogenicity trend testBiometrics 56 586ndash592

7 FengY ChenM and ManleyJL (2008) Phosphorylationswitches the general splicing repressor SRp38 to a sequence-specific activator Nat Struct Mol Biol 15 1040ndash1048

8 VethanthamV RaoN and ManleyJL (2008) Sumoylationregulates multiple aspects of mammalian poly(A) polymerasefunction Genes Dev 22 499ndash511

9 LlorianM SchwartzS ClarkTA HollanderD TanLYSpellmanR GordonA SchweitzerAC de la GrangeP AstGet al (2010) Position-dependent alternative splicing activityrevealed by global profiling of alternative splicing events regulatedby PTB Nat Struct Mol Biol 17 1114ndash1123

10 FinckBN GroplerMC ChenZ LeoneTC CroceMAHarrisTE LawrenceJC Jr and KellyDP (2006) Lipin 1 is aninducible amplifier of the hepatic PGC-1alphaPPARalpharegulatory pathway Cell Metab 4 199ndash210

11 BiondoPD BrindleyDN SawyerMB and FieldCJ (2008)The potential for treatment with dietary long-chainpolyunsaturated n-3 fatty acids during chemotherapy J NutrBiochem 19 787ndash796

12 ShinC FengY and ManleyJL (2004) DephosphorylatedSRp38 acts as a splicing repressor in response to heat shockNature 427 553ndash558

13 NunezE KwonYS HuttKR HuQ CardamoneMDOhgiKA Garcia-BassetsI RoseDW GlassCKRosenfeldMG et al (2008) Nuclear receptor-enhancedtranscription requires motor- and LSD1-dependent genenetworking in interchromatin granules Cell 132 996ndash1010

14 FengY ValleyMT LazarJ YangAL BronsonRTFiresteinS CoetzeeWA and ManleyJL (2009) SRp38regulates alternative splicing and is required for Ca(2+) handlingin the embryonic heart Dev Cell 16 528ndash538

15 AlvesG and YuYK (2011) Combining independent weightedP-values achieving computational stability by a systematicexpansion with controllable accuracy PLoS One 6 e22647

16 RayD KazanH CookKB WeirauchMT NajafabadiHSLiX GueroussovS AlbuM ZhengH YangA et al (2013)

A compendium of RNA-binding motifs for decoding generegulation Nature 499 172ndash177

17 BaileyTL BodenM BuskeFA FrithM GrantCEClementiL RenJ LiWW and NobleWS (2009) MEMESUITE tools for motif discovery and searching Nucleic AcidsRes 37 W202ndashW208

18 LiH WangZ ZhouX ChengY XieZ ManleyJL andFengY (2013) Far upstream element-binding protein 1 and RNAsecondary structure both mediate second-step splicing repressionProc Natl Acad Sci USA 110 E2687ndashE2695

19 HanJ DingJH ByeonCW KimJH HertelKJ JeongSand FuXD (2011) SR proteins induce alternative exon skippingthrough their activities on the flanking constitutive exons MolCell Biol 31 793ndash802

20 AkerfeltM TrouilletD MezgerV and SistonenL (2007) Heatshock factors at a crossroad between stress and developmentAnn N Y Acad Sci 1113 15ndash27

21 MaloyKJ and PowrieF (2011) Intestinal homeostasis andits breakdown in inflammatory bowel disease Nature 474298ndash306

22 Porn-AresMI SamaliA and OrreniusS (1998) Cleavage of thecalpain inhibitor calpastatin during apoptosis Cell Death Differ5 1028ndash1033

23 ChanCW LeeYB UneyJ FlynnA TobiasJH andNormanM (2007) A novel member of the SAF (scaffoldattachment factor)-box protein family inhibits gene expression andinduces apoptosis Biochem J 407 355ndash362

24 LadomeryM (2013) Aberrant Alternative Splicing Is AnotherHallmark of Cancer Int J Cell Biol 2013 463786

25 FerraiuoloMA BasakS DostieJ MurrayELSchoenbergDR and SonenbergN (2005) A role for the eIF4E-binding protein 4E-T in P-body formation and mRNA decayJ Cell Biol 170 913ndash924

26 PenkertRR and KalejtaRF (2012) Tale of a tegumenttransactivator the past present and future of human CMV pp71Future Virol 7 855ndash869

27 TemmeN and TudzynskiP (2009) Does botrytis cinerea IgnoreH(2)O(2)-induced oxidative stress during infectionCharacterization of botrytis activator protein 1 Mol PlantMicrobe Interact 22 987ndash998

28 FarhanH WeissM TaniK KaufmanRJ and HauriHP(2008) Adaptation of endoplasmic reticulum exit sites toacute and chronic increases in cargo load EMBO J 272043ndash2054

29 PerryME (2010) The regulation of the p53-mediated stressresponse by MDM2 and MDM4 Cold Spring Harb PerspectBiol 2 a000968

30 YuY MaroneyPA DenkerJA ZhangXH DybkovOLuhrmannR JankowskyE ChasinLA and NilsenTW (2008)Dynamic regulation of alternative splicing by silencers thatmodulate 50 splice site competition Cell 135 1224ndash1236

31 LicatalosiDD MeleA FakJJ UleJ KayikciM ChiSWClarkTA SchweitzerAC BlumeJE WangX et al (2008)HITS-CLIP yields genome-wide insights into brain alternativeRNA processing Nature 456 464ndash469

4030 Nucleic Acids Research 2014 Vol 42 No 6