Genome-wide analysis of alternative pre-mRNA splicing in Arabidopsis t
http://www.100md.com
《核酸研究医学期刊》
1 Plant Mutation Exploration Team, Plant Functional Genomics Research Group and 2 Genomic Knowledge Base Research Team, Bioinformatics Group, RIKEN Genomic Sciences Center (GSC), RIKEN Yokohama Institute, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama 230-0045, Japan and 3 Laboratory of Plant Molecular Biology, RIKEN Tsukuba Institute, 3-1-1 Koyadai, Tsukuba 305-0074, Japan
* To whom correspondence should be addressed. Tel: +81 29 836 4359; Fax: +81 29 836 9060; Email: sinozaki@rtc.riken.jp
Present address: Kei Iida, Faculty of Bio-Science, Nagahama Institute of Bio-Science and Technology, 1266 Tamura-cho, Nagahama 526-0829, Japan
ABSTRACT
We mapped RIKEN Arabidopsis full-length (RAFL) cDNAs to the Arabidopsis thaliana genome to search for alternative splicing events. We used 278 734 full-length and 3'/5' terminal reads of the sequences of 220 214 RAFL cDNA clones for the analysis. Eighty-nine percent of the cDNA sequences could be mapped to the genome and were clustered in 17 130 transcription units (TUs). Alternative splicing events were found in 1764 out of 15 214 TUs (11.6%) with multiple sequences. We collected full-length cDNA clones from plants grown under various environmental conditions or from various organs. We then analyzed the correlation between alternative splicing events and environmental stress conditions. Alternative splicing profiles changed according to environmental stress conditions and the various developmental stages of plant organs. In particular, cold-stress conditions affected alternative splicing profiles. The change in alternative splicing profiles under cold stress may be mediated by alternative splicing and transcriptional regulation of splicing factors.
INTRODUCTION
Arabidopsis thaliana Heynh. is a model organism used to study various molecular systems in the development, environmental responses and metabolism of higher plants. Its complete genomic sequence has been determined (1), and extensive large-scale, full-length cDNA collections have been made (2–4). From work on the human genome sequence, alternative splicing is now thought to be important to the complexity of gene function (5). Alternative splicing events produce additional transcripts from genes to mediate the complicated functions of the human body. Alternative splicing events are also important in higher plants. Large-scale alternative splicing in A.thaliana was first analyzed by Haas et al. (3,6). They used 180 000 expressed sequence tags (ESTs), including 80 000 RIKEN Arabidopsis full-length (RAFL) cDNAs, to detect 1188 genes containing alternative splicing variations. In addition, Zhu et al. (7) have reported 327 alternative splicing events in an analysis using 180 000 ESTs. In our project, we had collected more than 270 000 sequences by the end of 2003 (Figure 1a). We expected that a large number of EST sequences would allow us to detect many alternative splicing events. It was also of help that our EST sequences were obtained from full-length cDNA, because all EST sequences have information on the 5' and/or 3' terminal sites.
Figure 1. Comparison of sequencing resources and detected alternative splicing events between this work and previous work. (a) A.thaliana cDNA sequence resources used in our analysis (in December 2003), (b) Venne diagram of the genes with alternative splicing events detected in this work and in previous work by Hass et al. (6).
Our RAFL cDNA collection has an additional advantage for the analysis of alternative splicing events. We have constructed 18 cDNA libraries of expressed genes from Arabidopsis plants grown under various environmental conditions or from plant organs at various developmental stages. Therefore, each RAFL cDNA clone has associated information on the conditions or organs in which it is expressed. To use this information, we analyzed the relationship between the expression of alternatively spliced transcripts and plant growth conditions. Previous studies suggest that alternative splicing events occur in response to environmental changes or at particular developmental stages (8–10). However, there have been few reports on changes in alternative splicing profiles according to expressional conditions at the whole transcriptome level. We discuss the molecular mechanism of cold-inducible changes in alternative splicing profiles.
MATERIALS AND METHODS
Data set
We used 278 734 sequences from RAFL cDNA clones. They included 92 654 RAFL 5' terminal read sequences, 172 653 RAFL 3' terminal read sequences and 13 427 RAFL full-length read sequences (Figure 2). We analyzed 248 514 mapped cDNA clones. About 190 000 unpublished sequences were also used for the analysis. These sequences can be downloaded from RARGE (http://rarge.gsc.riken.jp/) and have been deposited in the DNA database of Japan (DDBJ).
Figure 2. Data flow of clustering for the analysis of alternative splicing events in RAFL sequences.
Mapping the RAFL cDNA clone sequences to the Arabidopsis genome
We mapped the RAFL cDNA sequences to the Arabidopsis genome using BLAST (11). We clustered the results in two steps. In the first step, to detect long and identical exons, we chose sequences with 95% identity and a length of 50 bp as exons. In the second step, to detect micro-exons (3) or other small exons, we chose sequences with 85% identity and a length of 15 bp where each HSP (high-scoring segment pair) was consistent with exons detected in the first step. Although a micro-exon is defined as an exon with a length of 3–25 bp (3), we did not treat HSPs with a length of <15 bp as exons. It is difficult to detect such micro-exons using BLAST. In some cases, this problem causes the incorrect detection of exon skip-type (ES-type) alternative splicing events. In addition, 10 bp sequences on exon–intron boundaries usually belong to both of the two neighboring exons. To avoid the incorrect detection of exon–intron structure as a result, we used 15 bp sequences as a spacer to check the consistency of the exon–intron structure.
After mapping the RAFL cDNA sequences to the genome, we clustered mapped sequences into transcription units (TUs) according to the method of Okazaki et al. (12). Sequences with the same direction and overlapping nucleotides were clustered into single TUs.
Detection of alternative splicing
Before searching for alternative splicing events, we surveyed the genomic exon–intron structure of each TU. We aligned sequences clustered into single TUs on the genome sequence and considered contiguous nucleotides as genomic exons if each nucleotide was on an exon in at least one RAFL cDNA sequence. Following detection of the genomic exon–intron structure, we searched for alternative splicing events. For ES-type alternative splicing, we simply searched for sequences without genomic exons. If an exon loss occurred on the 3' terminal site of a TU, the alternative splicing event was categorized as an alternative terminal exon type (AT-type) one. The genomic exon reflected the longest exon. For alternative donor type (AD-type) and alternative acceptor type (AA-type) alternative splicing, we searched for exons shorter than genomic exons. We did not check the 5' sites of initial exons or 3' sites of terminal exons because these are not splice sites. We used an additional rule in the search for AD/AA-type alternative splicing events to avoid misdetection. The BLAST algorithm was sometimes not good at finding identical sequences at an exon–intron boundary if the similarities of the boundary sequences were not high because of sequence read errors. To detect AD/AA-type events, we used the lengths of sequences that could not be aligned on the genome as spacers. In considering spacers, we considered a sequence as an alternatively spliced clone of the AD/AA type if Ld – Ls > Lt (AD/AA) . For a retained intron type (RI-type) alternative splicing event, we searched for introns on genomic exons. As in the detection of other types of alternative splicing, we used nucleotides that could not be aligned to the genome as spacers. We considered a sequence to be an RI-type alternatively spliced clone if Li – Ls > Lt (RI) (where Li is the length of an intron found on a genomic exon, Ls is the spacer length and Lt (RI) is a threshold length for the detection of RI-type alternative splicing; we used Lt (RI) = 20).
The sequences used contained both 5' and 3' terminal read sequences of the RAFL clones. We deselected uninformative sites of each sequence to check alternative splicing. (For example, the 3' terminal sites of the 5' terminal read sequences were not used to check AT-type alternative splicing.)
We used the BLAST algorithm for mapping the RAFL cDNA clones to the genome. The BLAST algorithm is good for searching for highly similar sequences, but sometimes it will not detect sequences with low similarity, especially exon–intron boundary sequences. Other programs are better tuned for mapping sequences to a genome, e.g. GeneSeqer (13) and Sim4 (14). We tried these programs, but they gave worse results than BLAST because our sequence set included some poor-quality terminal read sequences. As we did not use sequences with <95% identity, good sequences remained that could be mapped to the genome. Considering these issues, we chose BLAST.
Analysis of alternative splicing profiles in each library
We analyzed the relationships between alternative splicing events and environmental conditions or the various developmental stages of plant organs. In this analysis, we compared the measured and expected numbers of clones that have specific alternative splicing profiles in each library. For this comparison, we counted clones without alternative exons in ES-type alternative splicing events, those with short terminal sites in AT-type alternative splicing, those with short exons in AD or AA sites and those with unspliced introns in RI-type alternative splicing. Expected numbers were calculated as the product of the number of library members and the probability of each alternative splicing event. We used the chi-square test to check whether the differences between measured and expected values were statistically significant. We used the statistics software ‘R’ (http://cran.r-project.org/) to compute P-values in the chi-square test.
RESULTS
Mapping the RAFL cDNA clone sequences to the Arabidopsis genome
We mapped 248 514 (89%) of 278 734 RAFL cDNA clone sequences to the A.thaliana genome (1) using the BLAST search (11). We used a mapping rule in which each exon has 95% identity to the genome in a 50 bp region. Haas et al. (3) reported micro-exons in some Arabidopsis genes. To detect micro-exons or other small exons, we used an additional rule in which exons 15 bp are considered to be micro-exons only if they occur between mapped exons. cDNA clones with mapping coverage of <90% of the corresponding full-length exons were not used. After mapping these sequences to the genome, we constructed TUs (12). Sequences that are encoded on the same strands of the same chromosome and overlap by at least 1 nt were clustered into single TUs. Using this rule, we analyzed the whole Arabidopsis genome to identify 17 130 TUs out of 248 514 sequences (Figure 2). Each TU was estimated to contain 14.5 sequences on average. The TU with the most sequences contained 1335 sequences, encoding the dnak-type molecular chaperone hsc 70.1. TUs with 30 sequences accounted for almost 90% of all TUs.
Detection of alternative splicing
We searched for alternative splicing events in 15 214 TUs with two or more sequences. We divided events into five types: ES-type, AT-type, AD-type, alternative AA-type and RI-type events, following Zhou et al. (15) and Haas et al. (3) (Figure 3). Although AT is a subtype of ES, we treat AT-type events as an independent category. These two alternative splicing events may belong to different types, because selection of an alternative terminal exon can occur as a result of alternative polyadenylation.
Figure 3. Types of alternative splicing: (a) exon skip, (b) alternative terminal exon (subtype of exon skip), (c) alternative donor site, (d) alternative acceptor site, (e) retained intron. [based on Zhou et al. (15) and Haas et al. (3).
We detected 1764 TUs (11.6% of multisequence TUs) with splicing variants (Table 1; http://rarge.gsc.riken.jp/a_splicing/) including 1315 TUs that were newly detected as alternatively spliced TUs (Figure 1b). Among the 15 214 TUs, we identified 273 (1.8%) ES-type and 458 (3.0%) AT-type events. In both ES-type and AT-type alternative splicing events, there was at least one alternative exon. If the alternative exon was on the terminal site of the TU, this alternative splicing event was classified as an AT-type one. These results indicate that the alternative exon tended to be on the terminal sites of TUs rather than on their initial or middle sites. We identified 250 (1.6%) AD-type and 321 (2.1%) AA-type events. When TUs contained a shorter exon than the genomic exon, they were considered to be AD-type or AA-type events (details in Materials and Methods). We could not find much difference among occurrences of these two alternative splicing events. Of the 15 214 TUs, 790 (5.2%) had RI-type alternative splicing events. RI-type TUs were slightly more common than the other types. Some RI-type alternative splicing events may be due to immature mRNAs. In our data, there were 267 TUs (34%) in which RI-type alternative splicing events were supported by more than two transcripts. When using other EST data registered with the NCBI, the proportion of TUs including more than two transcripts with an unspliced intron was 65% (data not shown).
Table 1. Numbers of TUs including alternatively spliced transcripts
Change of alternative splicing profiles
We analyzed whether alternative splicing profiles change with environmental conditions or the developmental stages of plants. We collected RAFL cDNAs from 18 cDNA libraries prepared from Arabidopsis plants grown under various abiotic stress conditions (2). Thus, each RAFL cDNA clone was associated with a library (Table 2). Using this information, we could analyze the relationship between alternative splicing profiles and growth conditions. Significant changes of alternative splicing profiles were observed in several libraries.
Table 2. Alternative splicing events in each library
From the analysis of ES-type alternative splicing (Figure 3a), we compared the measured and expected numbers of clones without alternatively spliced exons in each library. Libraries RAFL07 (cold-treated plants), RAFL15 (siliques and flowers), RAFL17 (rehydration) and RAFL21 (various stresses) had statistically significant differences (P < 1.0 x 10–7) between measured and expected values (Table 2 and Figure 4a). In those except RAFL21, the numbers of alternatively spliced clones without alternative exons were fewer than expected.
Figure 4. Comparison of the measured values and expected values of alternatively spliced transcripts in each library: (a) exon skip, (b) alternative terminal exon, (c) alternative donor site, (d) alternative acceptor site, (e) retained intron. Arrows indicate the libraries with statistically significant differences (P < 1.0 x 10–7) between measured values and expected values.
For AT-type alternative splicing, we compared the measured and expected numbers of clones with short termination events. Libraries RAFL07, RAFL21, RAFL06 (various developmental stages) and RAFL19 (siliques and flowers) had statistically significant differences between measured and expected values (Figure 4b). In RAFL06 and RAFL07, the observed numbers of clones with short termination sites were greater than expected. In libraries RAFL19 and RAFL21, the numbers were smaller than expected.
For AD-type and AA-type alternative splicing, we analyzed the number of clones with short exons. These two types of alternative splicing showed similar results. RAFL07 and RAFL21 had statistically significant differences (Figure 4c and d). In RAFL07 (cold-treated plants), the number of clones with short exons as a result of AD/AA-type alternative splicing was lower than expected. In contrast, in RAFL21 (various stresses), the number of clones with short exons was greater than expected.
For RI-type alternative splicing, we compared the measured and expected number of clones with at least one unspliced intron. RAFL04 (cold-treated plants), RAFL08 (dehydration-treated plants), RAFL09 (various developmental stages) and RAFL21 had statistically significant differences between the two values (Figure 4e). In all those except RAFL21, the numbers of alternatively spliced clones with unspliced introns were greater than expected.
These results suggest that alternative splicing profiles are significantly affected by environmental stress conditions. In particular, RAFL07 (cold-treated plants) and RAFL21 (various stresses) had very different alternative splicing profiles from those of other libraries (Figure 4).
DISCUSSION
This study presents a large-scale analysis of alternative splicing profiles in the A.thaliana transcriptome. We used 278 734 cDNA sequences obtained from our collection of Arabidopsis full-length cDNAs . We detected 1764 TUs with alternatively splicing transcripts out of 15 214 TUs with two or more sequences (Figure 2). We were able to detect many more TUs with alternative splicing events than reported previously because of the number of sequences we analyzed. Moreover, all the sequences used in this study had associated information on at least one terminal site, which means that cDNA sequences have more information than EST sequences. For example, we could determine both the 5' initiation and 3' termination sites of 13 580 out of 17 130 TUs mapped on the genome. The percentage of TUs with alternative splicing events was 11.6%. The percentages of alternatively spliced genes out of the detected transcripts are 42% in human (16), 41% in mouse (12) and 10% in rice (17). Johnson et al. (18) reported that the percentage of alternative splicing events was 74% in human transcripts on the basis of a search of alternative splicing events using microarray analysis. The occurrence of alternative splicing events in Arabidopsis is similar to that in rice. The large difference in alternative splicing frequencies between higher animals and higher plants may cause differences in variations of transcripts.
Alternative splicing profiles were affected by environmental stress conditions at the transcriptome level (Figure 4 and Table 2). In particular, libraries RAFL07 (cold-treated plants) and RAFL21 (plants treated with various stresses) had statistically significant differences between the measured and expected numbers of alternatively spliced clones in many types of alternative splicing. Because RAFL21 included cDNA clones isolated from plants grown under various stress and hormone-treatment conditions, the relationship between alternative splicing and the exact environmental stress was not clear. However, on the basis of the data on alternative splicing events in RAFL07, we can discuss the effect of cold stress on alternative splicing.
We analyzed the alternative splicing variants of transcripts for splicing factors because alternative splicing profiles could be affected by splicing factors. In the Arabidopsis genome, there are 33 genes annotated as splicing factors (ftp://ftpmips.gsf.de/cress/). We obtained 26 TUs corresponding to these genes annotated as splicing factors; 13 contained splicing variation (Table 3). Among these, 11 were identified from RAFL libraries prepared from cold-treated plants (RAFL04, 07, 12, 18). Transcripts for splicing factors expressed under cold stress showed different splicing from those identified under other growth conditions. The transcript for splicing factor SR1 has splicing variants under cold stress. The SR1 transcript is alternatively spliced at an acceptor site of exon 11 (Figure 5). Splicing variants of SR1 expressed under cold stress (RAFL04-19-E10) have a long exon 11 as a result of the selection of the acceptor site (Figure 5). Lazar and Goodman (19) reported that the SR1 transcript has splicing variants, and that its alternative splicing is regulated by temperature increase. They suggested that alternative splicing profiles may be changed at the genomic level as a result of alternative splicing of SR1 at high temperature. SR1 is a homolog of human general/alternative splicing factor SF2/ASF (20), which can affect alternative splicing of several genes (21,22). These results suggest that the change in the splicing profile of SR1 by temperature change affects the splicing profiles of various transcripts.
Table 3. Alternative splicing events in various splicing factor transcripts
Figure 5. Splicing variations of the splicing factor SR1 transcripts. Full-length read sequences are colored in blue, 5' terminal read sequences in red and 3' terminal read sequences in green. AD-type alternative splicing events on the junction site between intron 10 and exon 11 are indicated by red arrows.
In addition to SR1, transcripts of other splicing factors showed splicing variations under cold conditions (Table 3). Our microarray analyses revealed that some genes for splicing factors were induced (>2.5 times) under cold stress (gray cells in Table 3). Fourteen splicing factor genes on the 7k full-length cDNA microarray can be divided into two groups: a group of eight genes that have alternative splicing events, and a group of six genes that have no alternative splicing events. Out of the eight genes with splicing variations, five were induced >2.5 times under cold stress. In contrast, out of the six genes without splicing variations, only one was induced under cold stress. These observations suggest a correlation between alternative splicing events and the cold-inducibility of splicing factor genes, and that both transcriptional and splicing regulation of splicing factor genes affects genome-wide alternative splicing profiles under cold stress. According to us, splicing factors change alternative splicing profiles at the transcriptome level by binding to exon–intron junctions for splicing. Further study of the relationship between alternative splicing and expressional and alternative splicing regulation of splicing factor genes is necessary to explain how splicing factors affect alternative splicing.
We showed a correlation between alternative splicing events and cold-stress conditions. This correlation may be due to the cold-inducible expression of splicing factor genes. This means that alternative splicing events in certain splicing factors affect alternative splicing profiles at the whole transcriptome level.
ACKNOWLEDGEMENTS
We thank Ms Asako Kamiya, Maiko Nakajima, Junko Ishida, Akiko Enjyu and Mari Narusaka for their excellent technical assistance. We also thank Dr Yoshihide Hayashizaki, Prof. Joseph R. Ecker, Athanasios Theologis and Ronald W. Davis for their collaboration, and Prof. Michiko Go for critical reading of the manuscript. This work is supported by the Genome Research Program of RIKEN.
REFERENCES
Arabidopsis Genome Initiative. ( (2000) ) Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature, , 408, , 796–815.
Seki,M., Narusaka,M., Kamiya,A., Ishida,J., Satou,M., Sakurai,T., Nakajima,M., Enju,A., Akiyama,K., Oono,Y. et al. ( (2002) ) Functional annotation of a full-length Arabidopsis cDNA collection. Science, , 296, , 141–145.
Haas,B.J., Volfovsky,N., Town,C.D., Troukhan,M., Alexandrov,N., Feldmann,K.A., Flavell,R.B., White,O. and Salzberg,S.L. ( (2002) ) Full-length messenger RNA sequences greatly improve genome annotation. Genome Biol., , 3, , RESEARCH0029.
Yamada,K., Lim,J., Dale,J.M., Chen,H., Shinn,P., Palm,C.J., Southwick,A.M., Wu,H.C., Kim,C., Nguyen,M. et al. ( (2003) ) Empirical analysis of transcriptional activity in the Arabidopsis genome. Science, , 302, , 842–846.
Lander,E.S., Linton,L.M., Birren,B., Nusbaum,C., Zody,M.C., Baldwin,J., Devon,K., Dewar,K., Doyle,M., FitzHugh,W. et al. ( (2001) ) Initial sequencing and analysis of the human genome. Nature, , 409, , 860–921.
Haas,B.J., Delcher,A.L., Mount,S.M., Wortman,J.R., Smith,R.K.,Jr, Hannick,L.I., Maiti,R., Ronning,C.M., Rusch,D.B., Town,C.D. et al. ( (2003) ) Improving the Arabidopsis genome annotation using maximal transcript alignment assemblies. Nucleic Acids Res., , 31, , 5654–5666.
Zhu,W., Schlueter,S.D. and Brendel,V. ( (2003) ) Refined annotation of the Arabidopsis genome by complete expressed sequence tag mapping. Plant Physiol., , 132, , 469–484.
Shi,H., Xiong,L., Stevenson,B., Lu,T. and Zhu,J.K. ( (2002) ) The Arabidopsis salt overly sensitive 4 mutants uncover a critical role for vitamin B6 in plant salt tolerance. Plant Cell, , 14, , 575–588.
Macknight,R., Duroux,M., Laurie,R., Dijkwel,P., Simpson,G. and Dean,C. ( (2002) ) Functional significance of the alternative transcript processing of the Arabidopsis floral promoter FCA. Plant Cell, , 14, , 877–888.
Yoshimura,K., Yabuta,Y., Ishikawa,T. and Shigeoka,S. ( (2002) ) Identification of a cis element for tissue-specific alternative splicing of chloroplast ascorbate peroxidase pre-mRNA in higher plants. J. Biol. Chem., , 277, , 40623–40632.
Altschul,S.F., Madden,T.L., Schaffer,A.A., Zhang,J., Zhang,Z., Miller,W. and Lipman,D.J. ( (1997) ) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res., , 25, , 3389–3402.
Okazaki,Y., Furuno,M., Kasukawa,T., Adachi,J., Bono,H., Kondo,S., Nikaido,I., Osato,N., Saito,R., Suzuki,H. et al; FANTOM Consortium; RIKEN Genome Exploration Research Group Phase I & II Team. ( (2002) ) Analysis of the mouse transcriptome based on functional annotation of 60,770 full-length cDNAs. Nature, , 420, , 563–573.
Usuka,J., Zhu,W. and Brendel,V. ( (2000) ) Optimal spliced alignment of homologous cDNA to a genomic DNA template. Bioinformatics, , 16, , 203–211.
Florea,L., Hartzell,G., Zhang,Z., Rubin,G.M. and Miller,W. ( (1998) ) A computer program for aligning a cDNA sequence with a genomic DNA sequence. Genome Res., , 8, , 967–974.
Zhou,Y., Zhou,C., Ye,L., Dong,J., Xu,H., Cai,L., Zhang,L. and Wei,L. ( (2003) ) Database and analyses of known alternatively spliced genes in plants. Genomics, , 82, , 584–595.
Modrek,B., Resch,A., Grasso,C. and Lee,C. ( (2001) ) Genome-wide detection of alternative splicing in expressed sequences of human genes. Nucleic Acids Res., , 29, , 2850–2859.
Kikuchi,S., Satoh,K., Nagata,T., Kawagashira,N., Doi,K., Kishimoto,N., Yazaki,J., Ishikawa,M., Yamada,H., Ooka,H. et al; Rice Full-Length cDNA Consortium; National Institute of Agrobiological Sciences Rice Full-Length cDNA Project Team; Foundation of Advancement of International Science Genome Sequencing & Analysis Group; RIKEN. ( (2003) ) Collection, mapping, and annotation of over 28 000 cDNA clones from japonica rice. Science, , 301, , 376–379.
Johnson,J.M., Castle,J., Garrett-Engele,P., Kan,Z., Loerch,P.M., Armour,C.D., Santos,R., Schadt,E.E., Stoughton,R. and Shoemaker,D.D. ( (2003) ) Genome-wide survey of human alternative pre-mRNA splicing with exon junction microarrays. Science, , 302, , 2141–2144.
Lazar,G. and Goodman,H.M. ( (2000) ) The Arabidopsis splicing factor SR1 is regulated by alternative splicing. Plant Mol. Biol., , 42, , 571–581.
Lazar,G., Schaal,T., Maniatis,T. and Goodman,H.M. ( (1995) ) Identification of a plant serine-arginine-rich protein similar to the mammalian splicing factor SF2/ASF. Proc. Natl Acad. Sci. USA, , 92, , 7672–7676.
Caceres,J.F., Stamm,S., Helfman,D.M. and Krainer,A.R. ( (1994) ) Regulation of alternative splicing in vivo by overexpression of antagonistic splicing factors. Science, , 265, , 1706–1709.
Molin,M. and Akusjarvi,G. ( (2000) ) Overexpression of essential splicing factor ASF/SF2 blocks the temporal shift in adenovirus pre-mRNA splicing and reduces virus progeny formation. J. Virol., , 74, , 9002–9009.
Seki,M., Narusaka,M., Ishida,J., Nanjo,T., Fujita,M., Oono,Y., Kamiya,A., Nakajima,M., Enju,A., Sakurai,T. et al. ( (2002) ) Monitoring the expression profiles of 7000 Arabidopsis genes under drought, cold and high-salinity stresses using a full-length cDNA microarray. Plant J., , 31, , 279–292.(Kei Iida1,2, Motoaki Seki1,3, Tetsuya Sa)
* To whom correspondence should be addressed. Tel: +81 29 836 4359; Fax: +81 29 836 9060; Email: sinozaki@rtc.riken.jp
Present address: Kei Iida, Faculty of Bio-Science, Nagahama Institute of Bio-Science and Technology, 1266 Tamura-cho, Nagahama 526-0829, Japan
ABSTRACT
We mapped RIKEN Arabidopsis full-length (RAFL) cDNAs to the Arabidopsis thaliana genome to search for alternative splicing events. We used 278 734 full-length and 3'/5' terminal reads of the sequences of 220 214 RAFL cDNA clones for the analysis. Eighty-nine percent of the cDNA sequences could be mapped to the genome and were clustered in 17 130 transcription units (TUs). Alternative splicing events were found in 1764 out of 15 214 TUs (11.6%) with multiple sequences. We collected full-length cDNA clones from plants grown under various environmental conditions or from various organs. We then analyzed the correlation between alternative splicing events and environmental stress conditions. Alternative splicing profiles changed according to environmental stress conditions and the various developmental stages of plant organs. In particular, cold-stress conditions affected alternative splicing profiles. The change in alternative splicing profiles under cold stress may be mediated by alternative splicing and transcriptional regulation of splicing factors.
INTRODUCTION
Arabidopsis thaliana Heynh. is a model organism used to study various molecular systems in the development, environmental responses and metabolism of higher plants. Its complete genomic sequence has been determined (1), and extensive large-scale, full-length cDNA collections have been made (2–4). From work on the human genome sequence, alternative splicing is now thought to be important to the complexity of gene function (5). Alternative splicing events produce additional transcripts from genes to mediate the complicated functions of the human body. Alternative splicing events are also important in higher plants. Large-scale alternative splicing in A.thaliana was first analyzed by Haas et al. (3,6). They used 180 000 expressed sequence tags (ESTs), including 80 000 RIKEN Arabidopsis full-length (RAFL) cDNAs, to detect 1188 genes containing alternative splicing variations. In addition, Zhu et al. (7) have reported 327 alternative splicing events in an analysis using 180 000 ESTs. In our project, we had collected more than 270 000 sequences by the end of 2003 (Figure 1a). We expected that a large number of EST sequences would allow us to detect many alternative splicing events. It was also of help that our EST sequences were obtained from full-length cDNA, because all EST sequences have information on the 5' and/or 3' terminal sites.
Figure 1. Comparison of sequencing resources and detected alternative splicing events between this work and previous work. (a) A.thaliana cDNA sequence resources used in our analysis (in December 2003), (b) Venne diagram of the genes with alternative splicing events detected in this work and in previous work by Hass et al. (6).
Our RAFL cDNA collection has an additional advantage for the analysis of alternative splicing events. We have constructed 18 cDNA libraries of expressed genes from Arabidopsis plants grown under various environmental conditions or from plant organs at various developmental stages. Therefore, each RAFL cDNA clone has associated information on the conditions or organs in which it is expressed. To use this information, we analyzed the relationship between the expression of alternatively spliced transcripts and plant growth conditions. Previous studies suggest that alternative splicing events occur in response to environmental changes or at particular developmental stages (8–10). However, there have been few reports on changes in alternative splicing profiles according to expressional conditions at the whole transcriptome level. We discuss the molecular mechanism of cold-inducible changes in alternative splicing profiles.
MATERIALS AND METHODS
Data set
We used 278 734 sequences from RAFL cDNA clones. They included 92 654 RAFL 5' terminal read sequences, 172 653 RAFL 3' terminal read sequences and 13 427 RAFL full-length read sequences (Figure 2). We analyzed 248 514 mapped cDNA clones. About 190 000 unpublished sequences were also used for the analysis. These sequences can be downloaded from RARGE (http://rarge.gsc.riken.jp/) and have been deposited in the DNA database of Japan (DDBJ).
Figure 2. Data flow of clustering for the analysis of alternative splicing events in RAFL sequences.
Mapping the RAFL cDNA clone sequences to the Arabidopsis genome
We mapped the RAFL cDNA sequences to the Arabidopsis genome using BLAST (11). We clustered the results in two steps. In the first step, to detect long and identical exons, we chose sequences with 95% identity and a length of 50 bp as exons. In the second step, to detect micro-exons (3) or other small exons, we chose sequences with 85% identity and a length of 15 bp where each HSP (high-scoring segment pair) was consistent with exons detected in the first step. Although a micro-exon is defined as an exon with a length of 3–25 bp (3), we did not treat HSPs with a length of <15 bp as exons. It is difficult to detect such micro-exons using BLAST. In some cases, this problem causes the incorrect detection of exon skip-type (ES-type) alternative splicing events. In addition, 10 bp sequences on exon–intron boundaries usually belong to both of the two neighboring exons. To avoid the incorrect detection of exon–intron structure as a result, we used 15 bp sequences as a spacer to check the consistency of the exon–intron structure.
After mapping the RAFL cDNA sequences to the genome, we clustered mapped sequences into transcription units (TUs) according to the method of Okazaki et al. (12). Sequences with the same direction and overlapping nucleotides were clustered into single TUs.
Detection of alternative splicing
Before searching for alternative splicing events, we surveyed the genomic exon–intron structure of each TU. We aligned sequences clustered into single TUs on the genome sequence and considered contiguous nucleotides as genomic exons if each nucleotide was on an exon in at least one RAFL cDNA sequence. Following detection of the genomic exon–intron structure, we searched for alternative splicing events. For ES-type alternative splicing, we simply searched for sequences without genomic exons. If an exon loss occurred on the 3' terminal site of a TU, the alternative splicing event was categorized as an alternative terminal exon type (AT-type) one. The genomic exon reflected the longest exon. For alternative donor type (AD-type) and alternative acceptor type (AA-type) alternative splicing, we searched for exons shorter than genomic exons. We did not check the 5' sites of initial exons or 3' sites of terminal exons because these are not splice sites. We used an additional rule in the search for AD/AA-type alternative splicing events to avoid misdetection. The BLAST algorithm was sometimes not good at finding identical sequences at an exon–intron boundary if the similarities of the boundary sequences were not high because of sequence read errors. To detect AD/AA-type events, we used the lengths of sequences that could not be aligned on the genome as spacers. In considering spacers, we considered a sequence as an alternatively spliced clone of the AD/AA type if Ld – Ls > Lt (AD/AA) . For a retained intron type (RI-type) alternative splicing event, we searched for introns on genomic exons. As in the detection of other types of alternative splicing, we used nucleotides that could not be aligned to the genome as spacers. We considered a sequence to be an RI-type alternatively spliced clone if Li – Ls > Lt (RI) (where Li is the length of an intron found on a genomic exon, Ls is the spacer length and Lt (RI) is a threshold length for the detection of RI-type alternative splicing; we used Lt (RI) = 20).
The sequences used contained both 5' and 3' terminal read sequences of the RAFL clones. We deselected uninformative sites of each sequence to check alternative splicing. (For example, the 3' terminal sites of the 5' terminal read sequences were not used to check AT-type alternative splicing.)
We used the BLAST algorithm for mapping the RAFL cDNA clones to the genome. The BLAST algorithm is good for searching for highly similar sequences, but sometimes it will not detect sequences with low similarity, especially exon–intron boundary sequences. Other programs are better tuned for mapping sequences to a genome, e.g. GeneSeqer (13) and Sim4 (14). We tried these programs, but they gave worse results than BLAST because our sequence set included some poor-quality terminal read sequences. As we did not use sequences with <95% identity, good sequences remained that could be mapped to the genome. Considering these issues, we chose BLAST.
Analysis of alternative splicing profiles in each library
We analyzed the relationships between alternative splicing events and environmental conditions or the various developmental stages of plant organs. In this analysis, we compared the measured and expected numbers of clones that have specific alternative splicing profiles in each library. For this comparison, we counted clones without alternative exons in ES-type alternative splicing events, those with short terminal sites in AT-type alternative splicing, those with short exons in AD or AA sites and those with unspliced introns in RI-type alternative splicing. Expected numbers were calculated as the product of the number of library members and the probability of each alternative splicing event. We used the chi-square test to check whether the differences between measured and expected values were statistically significant. We used the statistics software ‘R’ (http://cran.r-project.org/) to compute P-values in the chi-square test.
RESULTS
Mapping the RAFL cDNA clone sequences to the Arabidopsis genome
We mapped 248 514 (89%) of 278 734 RAFL cDNA clone sequences to the A.thaliana genome (1) using the BLAST search (11). We used a mapping rule in which each exon has 95% identity to the genome in a 50 bp region. Haas et al. (3) reported micro-exons in some Arabidopsis genes. To detect micro-exons or other small exons, we used an additional rule in which exons 15 bp are considered to be micro-exons only if they occur between mapped exons. cDNA clones with mapping coverage of <90% of the corresponding full-length exons were not used. After mapping these sequences to the genome, we constructed TUs (12). Sequences that are encoded on the same strands of the same chromosome and overlap by at least 1 nt were clustered into single TUs. Using this rule, we analyzed the whole Arabidopsis genome to identify 17 130 TUs out of 248 514 sequences (Figure 2). Each TU was estimated to contain 14.5 sequences on average. The TU with the most sequences contained 1335 sequences, encoding the dnak-type molecular chaperone hsc 70.1. TUs with 30 sequences accounted for almost 90% of all TUs.
Detection of alternative splicing
We searched for alternative splicing events in 15 214 TUs with two or more sequences. We divided events into five types: ES-type, AT-type, AD-type, alternative AA-type and RI-type events, following Zhou et al. (15) and Haas et al. (3) (Figure 3). Although AT is a subtype of ES, we treat AT-type events as an independent category. These two alternative splicing events may belong to different types, because selection of an alternative terminal exon can occur as a result of alternative polyadenylation.
Figure 3. Types of alternative splicing: (a) exon skip, (b) alternative terminal exon (subtype of exon skip), (c) alternative donor site, (d) alternative acceptor site, (e) retained intron. [based on Zhou et al. (15) and Haas et al. (3).
We detected 1764 TUs (11.6% of multisequence TUs) with splicing variants (Table 1; http://rarge.gsc.riken.jp/a_splicing/) including 1315 TUs that were newly detected as alternatively spliced TUs (Figure 1b). Among the 15 214 TUs, we identified 273 (1.8%) ES-type and 458 (3.0%) AT-type events. In both ES-type and AT-type alternative splicing events, there was at least one alternative exon. If the alternative exon was on the terminal site of the TU, this alternative splicing event was classified as an AT-type one. These results indicate that the alternative exon tended to be on the terminal sites of TUs rather than on their initial or middle sites. We identified 250 (1.6%) AD-type and 321 (2.1%) AA-type events. When TUs contained a shorter exon than the genomic exon, they were considered to be AD-type or AA-type events (details in Materials and Methods). We could not find much difference among occurrences of these two alternative splicing events. Of the 15 214 TUs, 790 (5.2%) had RI-type alternative splicing events. RI-type TUs were slightly more common than the other types. Some RI-type alternative splicing events may be due to immature mRNAs. In our data, there were 267 TUs (34%) in which RI-type alternative splicing events were supported by more than two transcripts. When using other EST data registered with the NCBI, the proportion of TUs including more than two transcripts with an unspliced intron was 65% (data not shown).
Table 1. Numbers of TUs including alternatively spliced transcripts
Change of alternative splicing profiles
We analyzed whether alternative splicing profiles change with environmental conditions or the developmental stages of plants. We collected RAFL cDNAs from 18 cDNA libraries prepared from Arabidopsis plants grown under various abiotic stress conditions (2). Thus, each RAFL cDNA clone was associated with a library (Table 2). Using this information, we could analyze the relationship between alternative splicing profiles and growth conditions. Significant changes of alternative splicing profiles were observed in several libraries.
Table 2. Alternative splicing events in each library
From the analysis of ES-type alternative splicing (Figure 3a), we compared the measured and expected numbers of clones without alternatively spliced exons in each library. Libraries RAFL07 (cold-treated plants), RAFL15 (siliques and flowers), RAFL17 (rehydration) and RAFL21 (various stresses) had statistically significant differences (P < 1.0 x 10–7) between measured and expected values (Table 2 and Figure 4a). In those except RAFL21, the numbers of alternatively spliced clones without alternative exons were fewer than expected.
Figure 4. Comparison of the measured values and expected values of alternatively spliced transcripts in each library: (a) exon skip, (b) alternative terminal exon, (c) alternative donor site, (d) alternative acceptor site, (e) retained intron. Arrows indicate the libraries with statistically significant differences (P < 1.0 x 10–7) between measured values and expected values.
For AT-type alternative splicing, we compared the measured and expected numbers of clones with short termination events. Libraries RAFL07, RAFL21, RAFL06 (various developmental stages) and RAFL19 (siliques and flowers) had statistically significant differences between measured and expected values (Figure 4b). In RAFL06 and RAFL07, the observed numbers of clones with short termination sites were greater than expected. In libraries RAFL19 and RAFL21, the numbers were smaller than expected.
For AD-type and AA-type alternative splicing, we analyzed the number of clones with short exons. These two types of alternative splicing showed similar results. RAFL07 and RAFL21 had statistically significant differences (Figure 4c and d). In RAFL07 (cold-treated plants), the number of clones with short exons as a result of AD/AA-type alternative splicing was lower than expected. In contrast, in RAFL21 (various stresses), the number of clones with short exons was greater than expected.
For RI-type alternative splicing, we compared the measured and expected number of clones with at least one unspliced intron. RAFL04 (cold-treated plants), RAFL08 (dehydration-treated plants), RAFL09 (various developmental stages) and RAFL21 had statistically significant differences between the two values (Figure 4e). In all those except RAFL21, the numbers of alternatively spliced clones with unspliced introns were greater than expected.
These results suggest that alternative splicing profiles are significantly affected by environmental stress conditions. In particular, RAFL07 (cold-treated plants) and RAFL21 (various stresses) had very different alternative splicing profiles from those of other libraries (Figure 4).
DISCUSSION
This study presents a large-scale analysis of alternative splicing profiles in the A.thaliana transcriptome. We used 278 734 cDNA sequences obtained from our collection of Arabidopsis full-length cDNAs . We detected 1764 TUs with alternatively splicing transcripts out of 15 214 TUs with two or more sequences (Figure 2). We were able to detect many more TUs with alternative splicing events than reported previously because of the number of sequences we analyzed. Moreover, all the sequences used in this study had associated information on at least one terminal site, which means that cDNA sequences have more information than EST sequences. For example, we could determine both the 5' initiation and 3' termination sites of 13 580 out of 17 130 TUs mapped on the genome. The percentage of TUs with alternative splicing events was 11.6%. The percentages of alternatively spliced genes out of the detected transcripts are 42% in human (16), 41% in mouse (12) and 10% in rice (17). Johnson et al. (18) reported that the percentage of alternative splicing events was 74% in human transcripts on the basis of a search of alternative splicing events using microarray analysis. The occurrence of alternative splicing events in Arabidopsis is similar to that in rice. The large difference in alternative splicing frequencies between higher animals and higher plants may cause differences in variations of transcripts.
Alternative splicing profiles were affected by environmental stress conditions at the transcriptome level (Figure 4 and Table 2). In particular, libraries RAFL07 (cold-treated plants) and RAFL21 (plants treated with various stresses) had statistically significant differences between the measured and expected numbers of alternatively spliced clones in many types of alternative splicing. Because RAFL21 included cDNA clones isolated from plants grown under various stress and hormone-treatment conditions, the relationship between alternative splicing and the exact environmental stress was not clear. However, on the basis of the data on alternative splicing events in RAFL07, we can discuss the effect of cold stress on alternative splicing.
We analyzed the alternative splicing variants of transcripts for splicing factors because alternative splicing profiles could be affected by splicing factors. In the Arabidopsis genome, there are 33 genes annotated as splicing factors (ftp://ftpmips.gsf.de/cress/). We obtained 26 TUs corresponding to these genes annotated as splicing factors; 13 contained splicing variation (Table 3). Among these, 11 were identified from RAFL libraries prepared from cold-treated plants (RAFL04, 07, 12, 18). Transcripts for splicing factors expressed under cold stress showed different splicing from those identified under other growth conditions. The transcript for splicing factor SR1 has splicing variants under cold stress. The SR1 transcript is alternatively spliced at an acceptor site of exon 11 (Figure 5). Splicing variants of SR1 expressed under cold stress (RAFL04-19-E10) have a long exon 11 as a result of the selection of the acceptor site (Figure 5). Lazar and Goodman (19) reported that the SR1 transcript has splicing variants, and that its alternative splicing is regulated by temperature increase. They suggested that alternative splicing profiles may be changed at the genomic level as a result of alternative splicing of SR1 at high temperature. SR1 is a homolog of human general/alternative splicing factor SF2/ASF (20), which can affect alternative splicing of several genes (21,22). These results suggest that the change in the splicing profile of SR1 by temperature change affects the splicing profiles of various transcripts.
Table 3. Alternative splicing events in various splicing factor transcripts
Figure 5. Splicing variations of the splicing factor SR1 transcripts. Full-length read sequences are colored in blue, 5' terminal read sequences in red and 3' terminal read sequences in green. AD-type alternative splicing events on the junction site between intron 10 and exon 11 are indicated by red arrows.
In addition to SR1, transcripts of other splicing factors showed splicing variations under cold conditions (Table 3). Our microarray analyses revealed that some genes for splicing factors were induced (>2.5 times) under cold stress (gray cells in Table 3). Fourteen splicing factor genes on the 7k full-length cDNA microarray can be divided into two groups: a group of eight genes that have alternative splicing events, and a group of six genes that have no alternative splicing events. Out of the eight genes with splicing variations, five were induced >2.5 times under cold stress. In contrast, out of the six genes without splicing variations, only one was induced under cold stress. These observations suggest a correlation between alternative splicing events and the cold-inducibility of splicing factor genes, and that both transcriptional and splicing regulation of splicing factor genes affects genome-wide alternative splicing profiles under cold stress. According to us, splicing factors change alternative splicing profiles at the transcriptome level by binding to exon–intron junctions for splicing. Further study of the relationship between alternative splicing and expressional and alternative splicing regulation of splicing factor genes is necessary to explain how splicing factors affect alternative splicing.
We showed a correlation between alternative splicing events and cold-stress conditions. This correlation may be due to the cold-inducible expression of splicing factor genes. This means that alternative splicing events in certain splicing factors affect alternative splicing profiles at the whole transcriptome level.
ACKNOWLEDGEMENTS
We thank Ms Asako Kamiya, Maiko Nakajima, Junko Ishida, Akiko Enjyu and Mari Narusaka for their excellent technical assistance. We also thank Dr Yoshihide Hayashizaki, Prof. Joseph R. Ecker, Athanasios Theologis and Ronald W. Davis for their collaboration, and Prof. Michiko Go for critical reading of the manuscript. This work is supported by the Genome Research Program of RIKEN.
REFERENCES
Arabidopsis Genome Initiative. ( (2000) ) Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature, , 408, , 796–815.
Seki,M., Narusaka,M., Kamiya,A., Ishida,J., Satou,M., Sakurai,T., Nakajima,M., Enju,A., Akiyama,K., Oono,Y. et al. ( (2002) ) Functional annotation of a full-length Arabidopsis cDNA collection. Science, , 296, , 141–145.
Haas,B.J., Volfovsky,N., Town,C.D., Troukhan,M., Alexandrov,N., Feldmann,K.A., Flavell,R.B., White,O. and Salzberg,S.L. ( (2002) ) Full-length messenger RNA sequences greatly improve genome annotation. Genome Biol., , 3, , RESEARCH0029.
Yamada,K., Lim,J., Dale,J.M., Chen,H., Shinn,P., Palm,C.J., Southwick,A.M., Wu,H.C., Kim,C., Nguyen,M. et al. ( (2003) ) Empirical analysis of transcriptional activity in the Arabidopsis genome. Science, , 302, , 842–846.
Lander,E.S., Linton,L.M., Birren,B., Nusbaum,C., Zody,M.C., Baldwin,J., Devon,K., Dewar,K., Doyle,M., FitzHugh,W. et al. ( (2001) ) Initial sequencing and analysis of the human genome. Nature, , 409, , 860–921.
Haas,B.J., Delcher,A.L., Mount,S.M., Wortman,J.R., Smith,R.K.,Jr, Hannick,L.I., Maiti,R., Ronning,C.M., Rusch,D.B., Town,C.D. et al. ( (2003) ) Improving the Arabidopsis genome annotation using maximal transcript alignment assemblies. Nucleic Acids Res., , 31, , 5654–5666.
Zhu,W., Schlueter,S.D. and Brendel,V. ( (2003) ) Refined annotation of the Arabidopsis genome by complete expressed sequence tag mapping. Plant Physiol., , 132, , 469–484.
Shi,H., Xiong,L., Stevenson,B., Lu,T. and Zhu,J.K. ( (2002) ) The Arabidopsis salt overly sensitive 4 mutants uncover a critical role for vitamin B6 in plant salt tolerance. Plant Cell, , 14, , 575–588.
Macknight,R., Duroux,M., Laurie,R., Dijkwel,P., Simpson,G. and Dean,C. ( (2002) ) Functional significance of the alternative transcript processing of the Arabidopsis floral promoter FCA. Plant Cell, , 14, , 877–888.
Yoshimura,K., Yabuta,Y., Ishikawa,T. and Shigeoka,S. ( (2002) ) Identification of a cis element for tissue-specific alternative splicing of chloroplast ascorbate peroxidase pre-mRNA in higher plants. J. Biol. Chem., , 277, , 40623–40632.
Altschul,S.F., Madden,T.L., Schaffer,A.A., Zhang,J., Zhang,Z., Miller,W. and Lipman,D.J. ( (1997) ) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res., , 25, , 3389–3402.
Okazaki,Y., Furuno,M., Kasukawa,T., Adachi,J., Bono,H., Kondo,S., Nikaido,I., Osato,N., Saito,R., Suzuki,H. et al; FANTOM Consortium; RIKEN Genome Exploration Research Group Phase I & II Team. ( (2002) ) Analysis of the mouse transcriptome based on functional annotation of 60,770 full-length cDNAs. Nature, , 420, , 563–573.
Usuka,J., Zhu,W. and Brendel,V. ( (2000) ) Optimal spliced alignment of homologous cDNA to a genomic DNA template. Bioinformatics, , 16, , 203–211.
Florea,L., Hartzell,G., Zhang,Z., Rubin,G.M. and Miller,W. ( (1998) ) A computer program for aligning a cDNA sequence with a genomic DNA sequence. Genome Res., , 8, , 967–974.
Zhou,Y., Zhou,C., Ye,L., Dong,J., Xu,H., Cai,L., Zhang,L. and Wei,L. ( (2003) ) Database and analyses of known alternatively spliced genes in plants. Genomics, , 82, , 584–595.
Modrek,B., Resch,A., Grasso,C. and Lee,C. ( (2001) ) Genome-wide detection of alternative splicing in expressed sequences of human genes. Nucleic Acids Res., , 29, , 2850–2859.
Kikuchi,S., Satoh,K., Nagata,T., Kawagashira,N., Doi,K., Kishimoto,N., Yazaki,J., Ishikawa,M., Yamada,H., Ooka,H. et al; Rice Full-Length cDNA Consortium; National Institute of Agrobiological Sciences Rice Full-Length cDNA Project Team; Foundation of Advancement of International Science Genome Sequencing & Analysis Group; RIKEN. ( (2003) ) Collection, mapping, and annotation of over 28 000 cDNA clones from japonica rice. Science, , 301, , 376–379.
Johnson,J.M., Castle,J., Garrett-Engele,P., Kan,Z., Loerch,P.M., Armour,C.D., Santos,R., Schadt,E.E., Stoughton,R. and Shoemaker,D.D. ( (2003) ) Genome-wide survey of human alternative pre-mRNA splicing with exon junction microarrays. Science, , 302, , 2141–2144.
Lazar,G. and Goodman,H.M. ( (2000) ) The Arabidopsis splicing factor SR1 is regulated by alternative splicing. Plant Mol. Biol., , 42, , 571–581.
Lazar,G., Schaal,T., Maniatis,T. and Goodman,H.M. ( (1995) ) Identification of a plant serine-arginine-rich protein similar to the mammalian splicing factor SF2/ASF. Proc. Natl Acad. Sci. USA, , 92, , 7672–7676.
Caceres,J.F., Stamm,S., Helfman,D.M. and Krainer,A.R. ( (1994) ) Regulation of alternative splicing in vivo by overexpression of antagonistic splicing factors. Science, , 265, , 1706–1709.
Molin,M. and Akusjarvi,G. ( (2000) ) Overexpression of essential splicing factor ASF/SF2 blocks the temporal shift in adenovirus pre-mRNA splicing and reduces virus progeny formation. J. Virol., , 74, , 9002–9009.
Seki,M., Narusaka,M., Ishida,J., Nanjo,T., Fujita,M., Oono,Y., Kamiya,A., Nakajima,M., Enju,A., Sakurai,T. et al. ( (2002) ) Monitoring the expression profiles of 7000 Arabidopsis genes under drought, cold and high-salinity stresses using a full-length cDNA microarray. Plant J., , 31, , 279–292.(Kei Iida1,2, Motoaki Seki1,3, Tetsuya Sa)