Skip to main content

Systematic characterization of the branch point binding protein, splicing factor 1, gene family in plant development and stress responses

Abstract

Background

Among eukaryotic organisms, alternative splicing is an important process that can generate multiple transcripts from one same precursor messenger RNA, which greatly increase transcriptome and proteome diversity. This process is carried out by a super-protein complex defined as the spliceosome. Specifically, splicing factor 1/branchpoint binding protein (SF1/BBP) is a single protein that can bind to the intronic branchpoint sequence (BPS), connecting the 5′ and 3′ splice site binding complexes during early spliceosome assembly. The molecular function of this protein has been extensively investigated in yeast, metazoa and mammals. However, its counterpart in plants has been seldomly reported.

Results

To this end, we conducted a systematic characterization of the SF1 gene family across plant lineages. In this work, a total of 92 sequences from 59 plant species were identified. Phylogenetic relationships of these sequences were constructed, and subsequent bioinformatic analysis suggested that this family likely originated from an ancient gene transposition duplication event. Most plant species were shown to maintain a single copy of this gene. Furthermore, an additional RNA binding motif (RRM) existed in most members of this gene family in comparison to their animal and yeast counterparts, indicating that their potential role was preserved in the plant lineage.

Conclusion

Our analysis presents general features of the gene and protein structure of this splicing factor family and will provide fundamental information for further functional studies in plants.

Background

In eukaryotes, canonical splicing removes noncoding intronic sequences and assembles the coding elements into mature mRNAs while alternative splicing (AS) generates different multiple transcripts that encode proteins with distinct structures and functions by differential usage of exons or splice site [58, 70]. The resulting transcripts of AS greatly contribute to post-transcriptional regulation, biological complexity and proteome diversity in eukaryotes [20, 50, 74]. Given that on average there are approximately 8 exons in each transcript in the human transcriptome and the degenerative nature of corresponding splice sites [20], pre-mRNA splicing is sophistically catalysed by the spliceosome. Spliceosome is a multi-megadalton protein complex, which consists of five (U1, U2, U4, U5 and U6) small nuclear ribonucleoprotein particles (snRNPs) and over 100 spliceosomal proteins [74]. Furthermore, the early assembly of spliceosome complex E or the commitment complex is an ATP-independent process and contains U1 snRNPs, SF1 and U2 snRNP auxiliary factors (U2AF large and U2AF small subunits) [48, 51]. Subsequently, the pre-spliceosome complex A is formed by replacing SF1 with SF3b155/SAP155 of U2 snRNPs [19, 67, 77]. Stepwise assembly of the following spliceosome during the splicing reaction has been reported as well [44, 63]; however, splice site recognition is a critical step during early assembly of the spliceosome. The current model describes the binding of U1 snRNP and U1 snRNA to a short stretch of 6 nucleotides at the 5′ splice site, of splicing factor 1 (SF1)/mammalian branch point binding protein (mBBP) at the branch point, and of U2 snRNP auxiliary factors at the 3′ splice site [46]. These three cis-elements are necessary but usually insufficient to define a specific exon–intron boundary. Thus, additional splicing enhancers or silencers located at exons and introns may allow the recognition of genuine splice sites during early spliceosome assembly [29].

Importantly, SF1 preferentially binds to the intron branch point sequence (BPS) which is adjacent to the binding site (polypyrimidine tract, Py) of U2AF large subunits (mammal U2AF65 and fission yeast U2AF59), bridging U1 and U2AF to form an intermediate lariat structure [58, 81]. In particular, SF1 is characterized by the presence of two types of RNA binding motifs at the N-terminus, a K homology/Quaking 2 (KH/QUA2) domain which originated from the human heterogeneous ribonucleoprotein (hnRNP) K protein [17, 66] and one or two zinc knuckle motif(s) (CX2CX4HX4C, X represents any amino acid). SF1 also contains a proline-rich region at C-terminus [2, 3]. Intriguingly, the yeast KH domain specifically binds to the BPS of pre-mRNAs with a Gly-Pro-Arg-Gly motif and the variable loop of the KH domain [39] and is necessary for spliceosome assembly [57]. The first but not the second zinc knuckle domain in yeast has been demonstrated to bind RNA with high affinity [16]. Moreover, the stability of the SF1–U2AF65–RNA complex is further affected by the phosphorylation status of several SF1 serine residues (Ser20, Ser80 and Ser82) in vitro [45]. The proline-rich region of SF1 interacts with U1 snRNP Prp40/FBP11 in yeast and human [2, 38]. In regards to its interaction partner, the U2AF large subunit, the N-terminal of SF1 interacts with its non-canonical RNA recognition motifs (RRM) or U2AF homology motif (UHM) [57, 62], whereas the other two RRMs of U2AF large subunit bind to the Py region [65].

A previous study in fission yeast (Schizosaccharomyces pombe) suggests that the initial co-recognition of the branch site and 3′ splice site is pivotal for correct splicing of target pre-mRNAs [60]. Because of the importance of splice site recognition for gene expression and protein diversity, SF1 has been demonstrated to play essential roles in a number of eukaryotic species including human (Homo sapiens), mice (Mus musculus), budding yeast (Saccharomyces cerevisiae), common fruit fly (Drosophila melanogaster) and roundworm (Caenorhabditis elegans) [2, 27, 47, 64, 68]. For example, in humans, missense mutation of splicing factors which are responsible for splice site recognition, such as SF1, has been linked to tumourigenesis [33]. Similarly, heterozygous SF1 (+/−) knockdown mice are susceptible to colon tumourigenesis induced by an organotrophic carcinogen, azoxymethane [64], and SF1 has been found to associate with beta-catenin/TCF4 complex, suggesting its role in carcinogenesis [49]. In contrast, knockdown of SF1 suppresses the development of germ cell tumours in mice [83], indicating its tissue dependency in cancer research. Furthermore, the molecular function of SF1 has been extensively studied in yeast. For instance, a sf1 mutant strain causes frequent exon skipping in fission yeast [52]. Additionally, SF1 has been proposed to recognize sub-optimal sequences in specific introns and lead to nuclear accumulation of pre-mRNA with aberrant splicing [73]. However, increasing evidence indicates that this protein is a regulator of splice site recognition and does not reduce general splicing, specifically during alternative splicing by targeting a subset of genes [46, 52, 68]. This hypothesis is supported by the fact that knockdown of SF1 in both yeast and human extracts only slightly affects the splicing outcome [22]. RNAi targeting of this gene has been demonstrated to not affect the splicing pattern of several splicing marker genes tested [68].

In comparison to studies in human and yeast, few reports have been published related to plant SF1 genes. Similar functions of the Arabidopsis SF1 gene were proposed in an early study in 2014 [30]. This plant SF1 homologue is reportedly responsible for the splicing of a group of transcripts. The loss-of-function mutant (atsf1–2) of this gene leads to abnormal development (early flowering and dwarfism) and ABA or heat stress sensitivity in Arabidopsis [30, 36]. Subsequently, the domain structure and its functional relationships have been substantially investigated [36], and the RRM domain is considered crucial to maintain its function in plants. Moreover, SF1 may have a different mechanism of 3′ splice site recognition in plant because the plant SF1 homologs contain a different RRM domain compared with fungal and metazoan counterparts [53, 78]. On the other hand, a study found that AtSF1 may be likely to play a functional role in the cytoplasm because it was found to shuttle between the nucleus and cytoplasm [54]. However, no related investigations have been conducted on the phylogenetic analysis of plant SF1 genes and their regulatory mechanisms. Although it is a highly conserved family and has conserved functions in eukaryotes, plant SF1 genes may have overlapping and distinct roles compared to the mammalian genes. Hence, studying the phylogenetic relationship and regulatory mechanism of plant SF1 genes may make us understand the evolutionary history, characteristics an expression profile of this gene family and predict specific functions in plants. This can lay the foundation for further functional studies in Viridiplantae. To this end, we systematically identified 92 SF1 sequences from 59 plant species, ranging from algae to higher plants. Meanwhile, the gene and protein structure, potential regulation at promoter regions and expression pattern of these genes were further investigated. In this study, we hypothesize that plant SF1 is structurally different from its counterparts in animals and yeast, but it is conserved among lower and higher plants, indicating its specific role in alternative splicing in branch point recognition.

Methods

Sequence acquisition and identification of plant SF1 genes

The Arabidopsis thaliana SF1 protein sequence (AT5G51300) was used to search similar sequences in all available plant species from the Phytozome v12.1 database (https://phytozome.jgi.doe.gov/pz/portal.html) [18] by running the BLASTp program with an e-value cutoff = 1e-10 (the other parameters were the default settings) [7]. Then, the retrieved protein sequences were examined and filtered using the HMMER score (default settings) [31], which contained PF16275 (Splicing factor 1 helix-hairpin domain, SF1-HH), PF00013 (K Homology domain, KH_1) and PF00076 (RNA recognition motif, RRM_1). Finally, 92 putative SF1 sequences from 59 plant species were identified. Detailed information including groups, plant species, common names and number of SF1 homologs reported for each plant species for subsequent analysis are listed in Table S1. Subcellular location prediction of identified SF1 proteins was carried out using WoLF PSORT (https://wolfpsort.hgc.jp/) [25].

Construction of molecular phylogenetic tree of plant SF1 genes

Protein sequences of the aforesaid plant SF1 genes were extracted from Phytozome v12.1 database for phylogenetic relationship analysis. The sequences with the longest coding sequences were chosen for genes with multiple different splicing isoforms. Then, multiple SF1 protein sequences were aligned with the Muscle v3.8 software with default settings [13]. The molecular phylogenetic tree of plant SF1 genes was then constructed using the maximum likelihood method (ML, JTT + G + I model) via PhyML v3.0 program with the following parameters: initial tree: BioNJ; discrete gamma model: yes; number of categories: 4; gamma shape parameter: 0.709; proportion of invariant: 0.021 subtree patterns aliasing: no [21]. FigTree v1.4.3 was used to visualize and edit the phylogenetic tree.

Gene structure, protein domain and multiple Em for motif elicitation (MEME) analysis

Required genomic, cDNA, and peptide sequences and all SF1 gene structures were downloaded from the Phytozome v12.1 database. Corresponding intron phases were generated using the online program Gene Structure Display Server 2.0 (GSDS2.0) (http://gsds.cbi.pku.edu.cn) [26]. Correlation analysis of SF1 exons were performed by using the piece2 webserver (http://www.bioinfogenome.net/piece/search.php? tdsourcetag=s_pctim_aiomsg) [76]. SF1 protein sequences were used to search for matching Pfam families using the HMMER website (https://www.ebi.ac.uk/Tools/hmmer/) [14]. Then, protein domain patterns were drawn by using TBtools software [8] according to the full Pfam resultant table. Conserved motifs of plant SF1 cDNA sequences and protein sequences were analysed on the MEME online program (http://meme-suite.org/tools/meme) [5] considering a maximum of the 10 most preserved motifs predicted for each sequence and leaving other settings on the default parameters.

Motif prediction in promoter regions of plant SF1 genes

The 1.5-kb 5′-flanking sequences of plant SF1 genes were extracted from genomic data available in Phytozome database. Prediction of plant putative cis-elements was performed with the online server PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) [37]. Motifs related to tissue-specific expression, internal hormones and external environmental stress response were selected for further analysis and discussion.

Expression analysis base on microarray datasets and gene expression experiments

Expression data of Arabidopsis, S. tuberosum, G. max, S. lycopersicum, P. trichocarpa and B. distachyon, including tissue specificity and stress responses, were extracted from the eFP browser series of the Bio-analytic Resource for plant biology (http://bar.utoronto.ca/) [34]. Expression values of selected plant SF1 genes were log transformed (lg) to generate visualize expression difference heatmaps by using BAR HeatMapper Tool program (http://bar.utoronto.ca/ntools/cgi-bin/ntools_heatmapper.cgi).

Gene expression experiments

Total RNA of samples from different plant tissues were extracted by RNeasy Mini kit (QIAGEN, USA) and subsequently reversed transcribed into cDNA by FastKing gDNA Dispelling RT SuperMix FastKing (TIANGEN, China) according to the manufacturer’s instruction. RT-PCR amplification were programmed as followings: 95 °C, 3 min; 95 °C, 30 s; 52 °C, 15 s; 72 °C, 45 s; 26/30 cycles; 72 °C 5 min. SYBR Premix Ex TaqTM (Accurate Biotechnology Co., Ltd. Hunan China) was used for quantitative real-time RT-PCR analysis which was conducted on the StepOne Plus real-time PCR system following optimized program: 95 °C, 30 s; 95 °C, 5 s; 60 °C, 30 s; 40 cycles. The data were normalized to the expression of internal reference genes (Table S6) and the transcript abundance was determined by the comparative CT value method [61].

Analysis of protein-protein interaction network and structural conservation

A protein-protein interaction network was generated by the STRING website (https://string-db.org) [12] with representative protein sequences from Arabidopsis. The following basic settings were employed: meaning of network edges, evidence (line colour indicates the type of interaction evidence); and active interaction sources, experiments.

There are three domains in the Arabidopsis SF1 protein. The phosphorylation and U2AF65 binding of the N-terminal domain of splicing factor 1 during 3′ splice site recognition of Homo sapiens (PDBID: 2M0G, identity: 36%, E-value: 7E-17) was similar to that of the K Homology domain. The structure for recognition of the intron branch site RNA by splicing factor 1 of Homo sapiens (PDBID: 1K1G, identity: 47%, E-value: 9E-27) can be used as the template for the splicing factor 1 helix-hairpin domain. Therefore, homology modelling was performed with modeller [43] based on two crystal structures. The amino acid conservation scores were calculated using the ConSurf Web server based on the ML method [4]. Input attributes were the 3D model and multiple sequence alignment (Figure S4). Related figures were created based on Pymol with default settings [79].

Analysis gene structure evolution with orthologue group of SF1 genes

Reconstruction of the evolutionary history of the structure of the plant SF1 family of orthologous genes was carried out by searching AT5G51300.1 in the PIECE 2 sever (http://www.bioinfogenome.net/piece/index.php). This provided an exon-intron display for orthologous genes from gene structure data sets linked to the phylogenetic tree.

Results

Sequence identification and phylogenetic analysis of the plant SF1 gene family

To identify SF1 gene family members in plants, we carried out a BLASTp search using the Arabidopsis AtSF1 (AT5G51300) amino acid sequence against the Phytozome database (v12.1). After filtering the sequence without SF1 signature or truncated sequences, a total of 92 sequences from 59 plant species were retrieved, which were roughly classified as 7 algae, 5 bryophyta, 1 basic angiosperm, 21 monocots, and 58 eudicots (Table S1). Specifically, the only species with four copies of plant SF1s was Eutrema salsugineum (salt cress) (Table S1). In particular, three copies of SF1 genes were observed in five species, including Panicum virgatum (Switchgrass), Triticum aestivum (common wheat), Daucus carota (carrot), Kalanchoe laxiflora (milky widow’s thrill) and Salix purpurea (purple osier willow). Additionally, 20 plant species contained two copies, and 33 species, including the model plant Arabidopsis, possessed only one copy of plant SF1s, respectively. The relatively larger number of SF1 genes and higher number of plant species in this work demonstrated the universality and complexity of the SF1 gene family. The retrieved sequences of 59 plant species provided us with more complete information to analyse the phylogenetic relationship of the SF1 gene family. Subsequently, a rooted phylogenetic tree was constructed based on the abovementioned 92 protein sequences by using the maximum likelihood method. The tree’s bootstrap (threshold: 0–1) was represented by a colour gradient (Fig. 1). In general, all SF1 protein sequences were clustered into four major clades including alga (in yellow), other land plants (in green), monocots (in pink) and eudicots (in blue), and one species (Amborella trichopoda) belonged to basic angiosperm (shown in colourless). The phylogenetic tree of SF1s (Figs. 1 and 2, left panel) with clear topology and overall high bootstrap values was similar to evolutionary trend from lower plants to higher plants reported in other studies. For example, the genes of algae in the yellow branch were representative members of the lineage that diverged before the evolution of land plants, which was the basal part of the phylogeny. In the blue branch, five sequences from Kalanchoe with higher BS values formed a subclade, showing their closer evolutionary relationships. Additionally, Cagra.3782 s0026.1.p from Capsella grandiflora and Carubv10025900m from C. rubella formed a subclade with the Arabidopsis sequences, because they all belong to Brassicaceae, which is consistent with the APG IV system (Fig. 1 and Table S1). Usually, some homologous SF1 sequences from the same species were clustered in the same small branch next to each other; these species included cashew, soybean, apple, woodland strawberry, quinoa, carrot, Colorado blue columbine, maize, common wheat, cereal grass, moss and bog moss (Fig. 1 and Table S1). In contrast, some other homologous SF1 members from the same species were clustered into the different subclades, such as purple osier willow, poplar, eastern cottonwood, salt cress, potato diploid kalanchoe, milky widow’s thrill, hall’s panicgrass, switchgrass, green algae and volvox (Fig. 1 and Table S1).

Fig. 1
figure1

Circular phylogenetic tree of the SF1 gene family available in plants. The phylogenetic tree of SF1 genes in plants was constructed based on maximum-likelihood with JTT + G model by using PhyML v3.037. A total of 92 protein sequences from 59 plant species were chosen to calculate the phylogenetic relationship for tree construction. Bootstrap values are labelled at each major branch. The corresponding information of each transcript such as species name, common name, number of identified transcripts and their transcript ID (nomenclature) are shown in Table S1 (taxonomies based on APG-IV system)

Fig. 2
figure2

Gene structure comparisons and conserved motif identification among plant SF1 genes. From left panel to right panel: vertical phylogenetic tree, genomic organization and identified cDNA conserved motifs by MEME analysis. Intron phase 0, 1 and 2 are shown on the gene structure. The conserved sequence of 10 identified motifs represented by different coloured boxes are listed below. Some long genes were reduced to one-half of their original length to fit this picture

Gene structure and conserved motif analysis

It is necessary to compare the exon-intron organization and conserved motifs of the plant SF1 gene family to clarify their evolutionary process and potential function. The gene structure models of SF1 genes were attached to the phylogenetic tree (Fig. 2), and the corresponding intron phase of each was also displayed (Fig. 2, Table S2). Figure 2 (middle, panel) shows that the gene length and structure of each member of the SF1 family exhibits significant differences. For example, the gene structure of 23 members of 92 SF1 family genes did not contain intron sequences; this subset accounts for 15.7% of the total number of members. Forty-eight sequences of SF1 genes had 2 exon-1 intron organizations, accounting for 52.2% of all genes. In particular, some genes from algae had multiple exons, including Vocar.0008 s0294.1.p (Volvox carteri) which contained the most exons (19 exons). Moreover, different gene structures were also observed at the same sub-branch. For instance, two sequences from Zea mays (maize) (Zm00008a037777_P01, 3 exons and Zm00008a007621_P01, 4 exons) were observed to have distinctive gene structures. Although the dissimilation of gene structure of each member of SF1s was substantial, we found that the length of CDSs did not significantly change (Fig. 2). Thus, whether it influences the differentiation of their gene function needs to be further investigated. Further investigation on conserved motifs by using Multiple Em for Motif Elicitation (MEME) search tool demonstrated that most SF1 genes (79 sequences) exhibited similar sequence signatures and the same order and all contained the 10 analysed motifs, except one sequence of Micromonas pusilla (50949) had a different position (Fig. 2, right panel). Although no obvious differences in identified conserved motifs were found among basal angiosperm, monocots and eudicots, sequences from the same species were found to have different motifs (Fig. 2). For example, Aqcoe5G406900.1.p and Aqcoe7G039300.1.p from the eudicot Aquilegia coerulea had 10 motifs and 9 motifs, respectively. The same situation was found in D. carota; DCAR_006843, DCAR_008506 and DCAR_004968 had 10 motifs, 9 motifs and 10 motifs, respectively. Intriguingly, the CDS length of DCAR_008506 was the longest. Notably, some sequences from algae and moss had fewer conserved motifs. For example, in bryophyta, the sequences of Physcomitrella patens (Pp3c7_10890V3.1.p and Pp3c11_24710V3.1.p), Sphagnum fallax (Sphfalx0015s0077.1.p and Sphfalx0010s0197.1.p) and Marchantia polymorpha (Mapoly0009s0189.1.p) had nine motifs. In algal plants, the sequences of 145,219 and 62,857 from Micromonas had only 7 motifs and 6 motifs, respectively. Moreover, although the sequences of Volvox carteri (Vocar.0007 s0345.1.p and Vocar.0008 s0294.1.p) and Chlamydomonas reinhardtii (Cre12.g553750.t1.1 and Cre09.g386731.t1.1) contained multiple exons, they had 9 motifs, indicating their sequence variation had little influence on function classes. Further correlation analysis of the SF1 exon regions were carried out to elucidate the gain/loss of introns. Correlations between transcripts of plant SF1s are shown in Fig. 3, providing additional information for phylogenetic analysis. For example, there is more similarity between PGSC0003DMT400081859 and Migut.D02531.2 because of multiple exact matches between the exons of the two transcripts.

Fig. 3
figure3

Analysis of gene structure evolution with orthologue group of SF1 genes. Exon-intron structure and intron phase (right panel) are linked to the plant species tree (left panel). Genes with red colour represent the members of the plant SF1 genes. Different coloured lines mean different exon comparison results between species

Analysis of protein domain and conserved motifs in peptides

The protein domains were analysed by using the above selected 92 peptide sequences from 59 plant species; the peptides’ annotations were splicing factor-related and conserved protein motifs were predicted according to the retrieved peptide sequences by MEME analysis (Fig. 4). Consequently, all SF1s were found having SF1_HH N-terminal domain on the N-terminal of the peptides followed by a KH domain and a C-terminal domain, namely, an RNA recognition motif (RRM) (Fig. 4, middle panel). Interestingly, in algae, 3 peptides from M. pusilla (145219), V. carteri (Vocar.0008 s0294.1.p) and C. reinhardtii (Cre09.g386731.t1.1) had two RRM domains. The amino acid lengths of SF1 proteins ranged from 499 aa to 1583 aa, and most of them possessed 700 to 800 amino acids (Table S3). Consistently, most of them are approximately 700 to 800 amino acids in length. Subcellular location prediction showed that the majority of SF1 proteins were had nuclear localization (86, 93.4%) (Table S3). Moreover, proteins of 30,147.m014250 (Ricinus communis) and Migut.F01191.1.p (Mimulus guttatus) were located in the vacuoles; proteins of Traes_2DL_6F03F05FA.4 (T. aestivum) and 145,219 (M. pusilla) were predicted to be cytoplasmic; proteins of GSMUA_Achr5P25100_001 (Musa acuminata) and Cre09.g386731.t1.1 (C. reinhardtii) were located in the chloroplast and endoplasmic reticulum, respectively.

Fig. 4
figure4

Comparisons of protein domains and conserved motif identification among plant SF1 genes. Protein domain (middle panel) and identified protein conserved motifs (right panel) identified by MEME analysis are shown against the vertical phylogenetic tree (left panel). The conserved sequence of 10 identified motifs represented by different coloured boxes are listed below

MEME analysis for SF1 peptide sequences was used to predict a total of 10 conserved motifs, which are presented as coloured boxes and cover most of the protein (Fig. 4, right panel). Further analysis showed that 77 peptides had all 10 motifs, accounting for approximately 83.7% of all SF1 protein sequences analysed in the study. Interestingly, all sequences from moss have 10 conserved motifs in the analysis, suggesting the conservation of SF1 proteins in bryophyta. Furthermore, almost all eudicots had 10 conserved motifs—except Anacardium occidentale (Anaoc.0018 s0425.1.p) and C. grandiflora (Cagra.3782 s0026.1.p) which lacked motif 2 and Malus domestica (MDP0000558834), Fragaria vesca (mrna21192.1-v1.0-hybrid) and Brassica rapa (Brara.C01481.1.p) which lacked motif 10—while most monocots had eight conserved motifs. In contrast, algal plants only possess approximately half of the predicted 10 motifs due to their peptides with integrant protein domains, implying the least degree of conservation and divergence of plant SF1 proteins in algae. T motifs that all algae shared were motif 3, motif 4, motif 5 and motif 9.

Analysis of promoter and tissue-specific expression of SF1 genes

To further analyse the regulation of plant SF1 genes at the transcriptional level, the 1.5-kb upstream sequences of plant SF1 genes were obtained from the Phytozome database, then the cis-elements of each promoter were identified by using the PlantCARE program (Table S4) [37]. Consequently, a total of 108 motifs were predicted. Generally, eight cis-elements related to tissue-specific expression among them were selected (Fig. 5 and Table S4), including HD-Zip1 for differentiation of the palisade mesophyll cells, the RY-element which regulates seed-specific expression, the AACA_motif and GCN4_motif involved in endosperm expression, and the CAT-box, CCGTCC-box, dOCT, and OCT for meristem expression. Further analysis showed that there were only 50 promoters of SF1 genes which had tissue-specific regulatory cis-elements. Particularly, the CAT-box and CCGTCC-box turned up at the highest frequency and greatest abundance in the promoters of SF1 genes. Both of them regulate meristem-specific expression and play key roles during development and growth of plants. Consistently, purple false brome (Brachypodium distachyon) of monocots not only had a CAT-box and CCGTCC-box, but was also highly expressed in young leaves, internode, adventitious roots and roots (Fig. 5 and Figure S2). However, no motifs were found to link the high expression of two SF1s of Glycine max (soybean) in SAM and root-tip (Figure S1). Additionally, the AACA_motif was only detected in Solanum tuberosum (PGSC0003DMP400032853) of potato, suggesting its specific role in regulating endosperm-specific negative expression. Further, HD-Zip 1 was present in Podel.03G113200.1.p of Populus deltoides (eastern cottonwood) and Spipo17G0046100 of Spirodela polyrhiza (greater duckweed). The RY-element was detected in the promoter of the dicot model plant Arabidopsis, and low expression was also reported in dry seed in Arabidopsis (Fig. 6), suggesting that the RY-element is involved in seed-specific negative expression of Arabidopsis. Moreover, expression levels in the same tissue type showed significant differences during different growth stages; for example, the expression level in stamen of flower stage 15 of Arabidopsis was obviously higher than that of the other flower development stages. However, the expression levels of different growth stages of Solanum lycopersicum were not only similar but lower, and no motifs were found in the promoter in tomato (Figs. 5 and S1). Furthermore, different expression patterns were detected in several SF1 genes with multiple copies (Figs. 6, S1 and S6). For instance, similar tissue expression profiles were detected in two SF1 homologues from the dicot Populus trichocarpa (Potri.001G126400.1 and Potri.003G107200.1) and the monocot Zea mays (Zm00008a007621_P01 and Zm00008a037777_P01) (Figure S1 and S5). In contrast, two SF1 genes of S. tuberosum showed differential expression patterns, similar to in G. max (Figs. 6 and S1).

Fig. 5
figure5

Analysis of motifs related to tissue specificity in the plant SF1 promoter regions. Eight cis-acting motifs are represented in different color triangles. Positions of these identified motifs are labeelled along the 1.5 kb 5′-flanking regions of each SF1 gene. The line solid and dotted represents regions with basic pairs and regions of no sequences or annexed base N respectively. Symbols on above the line represent the motifs at the plus strand, whereas symbols on below the line represent the motifs at the minus strand. Function of motifs: AACA-motif, involved in endosperm-specific negative expression; CAT-box, cis-acting regulatory element related to meristem expression; CCGTCC-box, cis-acting regulatory element related to meristem specific activation; dOCT, cis-acting regulatory element related to meristem specific activation; GCN4_motif, cis-regulatory element involved in endosperm expression; HD-Zip1, element involved in differentiation of the palisade mesophyll cells; RY-element, cis-acting regulatory element involved in seed-specific regulation. The black vertical lines represent break at that particular branch; OCT, cis-acting regulatory element related to meristem specific activation

Fig. 6
figure6

Expression patterns analysis of SF1 genes in several plants. a RT-PCR analysis of AtSF1 and PtSF1 associated with their isoforms expression in roots, shoots, leaves, and flowers. Gene models of each isoform (AtSF1 and PtSF1) are indicated (black, coding region; white, non-coding untranslated regions). RNA samples of mannitol treatment were prepared from 7-day-old seedlings at the exposure to 350 mM mannitol. The Arabidopsis actin2 and the poplar 18S rRNA gene were used as the internal expression control. b Real-time RT-PCR expression analysis of AtSF1 and PtSF1 associated with their isoforms. c Expression patterns of Arabidopsis and Solanum tuberosum (potato) SF1 genes. Expression data were obtained from the plant eFP browser microarray datasets, transformed by Lg conversion and presented as a heatmap. Red colour represents high levels of transcript abundance, and blue represents low transcript abundance

Analysis of promoter and internal and external hormones expression of SF1 genes

In long-term evolution and development, plants have gradually formed mechanisms of adaptation and resistance to adversity to maintain their life and sustain growth. To understand the regulatory mechanisms of internal and external stimuli on plant SF1s, cis-acting elements involved in hormone and stress were studied with the PlantCARE database (Fig. 7, Table S4). Finally, 19 hormone- and stress-related motifs were selected from 92 promoter sequences of plant SF1s. There are 12 hormone-related motifs including abscisic acid (ABRE), auxin (AuxRE, AuxRE-core, TGA-box, TGA-element), ethylene (ERE), gibberellin (GARE-motif, P-box, TATC-box), MeJA (CGTCA-motif, TGACG-motif), and salicylic acid (TCA-element) and five stress-related motifs including low-temperature (LTR), drought (MBS), wound (WUN-motif) and anoxic (ARE, GC-motif) motifs. Almost each SF1 sequence had a great diversity of cis-elements in its promoter regions except some sequences such as Araha.13031 s0002.1 and Traes_2AL_3D6729692.1 which did not contain a single motif due to the sequences contain ‘N’ or no promoter, suggesting that multiple hormones-mediated signalling pathways are closely related to SF1 plants resistance. Analysis showed that more than half of SF1 promoters contained ABRE, CGTCA-motif, TGACG-motif and ARE, respectively. Moreover, external hormone signals also affect the abundance of SF1 transcripts (Figure S3). For example, in Arabidopsis (AT5G51300.1), MJ (methyl jasmonate) inhibited its expression (Fig. 7), and treatment with other hormones like ACC (a precursor of ethylene), IAA (auxin), ABA and GA (gibberellin) regulates the expression of AT5G51300.1.

Fig. 7
figure7

Analysis of motif-related hormone and stresses in the plant SF1 promoter regions. Nineteen cis-acting elements are represented in different color symbols. Positions of these identified motifs are labeelled along the 1.5 kb 5′-flanking regions of each SF1 gene. The line solid and dotted represents regions with basic pairs and regions of no sequences or annexed base N respectively. Symbols on above the line represent the motifs at the plus strand, whereas symbols on below the line represent the motifs at the minus strand. Function of motifs: ABRE, cis-acting element involved in the abscisic acid responsiveness; ARE, cis-acting regulatory element essential for the anaerobic induction; AT-rich sequence, element for maximal elicitor-mediated activation (2copies); AuxRE, part of an auxin-responsive element; AuxRR-core, cis-acting regulatory element involved in auxin responsiveness; CGTCA-motif, cis-acting regulatory element involved in the MeJA-responsiveness; ERE, ethylene-responsive element; GARE-motif, gibberellin-responsive element; GC-motif, enhancer-like element involved in anoxic specific inducibility; LTR, cis-acting element involved in low-temperature responsiveness; TATC-box, cis-acting element involved in gibberellin-responsiveness; TCA-element, cis-acting element involved in salicylic acid responsiveness; MBS, MYB binding site involved in drought-inducibility; P-box, gibberellin-responsive element; TC-rich repeats, cis-acting element involved in defence and stress responsiveness; TGA-box, part of an auxin-responsive element; TGACG-motif, cis-acting regulatory element involved in the MeJA-responsiveness; TGA-element, auxin-responsive element; WUN-motif, wound-responsive element

Analysis of protein-protein interaction network and structural conservation

Protein-protein interaction (PPI) network analysis can systematically reveal the working principle of proteins in biological systems, the molecular mechanisms of biological signals and energy metabolism, and the functional relationships between proteins. In this study, we generated protein-protein interaction networks of the SF1 protein according to the representative protein sequence of Arabidopsis (AT5G51300) using the STRING database based on experiments (Fig. 8a). Finally, 10 predicted functional partners of the SF1 protein were obtained, including CDC5 (AT1G09770.1), AT1G10580 (AT1G10580.1), ATU2AF65A (AT4G36690.1), AT2G33440 (AT2G33440.1), AT2G33435 (AT2G33435.1), AT1G60900 (AT1G60900.1), AT1G60830 (AT1G60830.1), MAC3B (AT2G33340.1), MAC3A (AT1G04510.1), and AT1G31870 (AT1G31870.1) (Fig. 8a). CDC5, MAC3A and MAC3B are components of the MAC complex that probably regulate defence responses through transcriptional control and thereby are essential for plant innate immunity. All of them may be involved in pre-mRNA splicing and DNA repair. AT1G10580 is pre-mRNA-processing factor 17, and AT1G31870 is splicing factor CWC26. Both proteins participate in RNA splicing and pre-mRNA processing. AT2G33440, AT2G33435 and AT1G60830 are RNA recognition motif-containing proteins whose main molecular functions are involved in pre-mRNA splice site binding. ATU2AF65A and AT1G60900 are splicing factor U2af large subunit A and B, respectively, and they are necessary for the splicing of pre-mRNA. AT5G51300 (splicing factor-like protein 1) has already been demonstrated to be necessary for the splicing of pre-mRNA, development, and abscisic acid (ABA) responses. In general, SF1 protein and its functional partners are generally involved in RNA splicing and pre-mRNA processing, and some of them also possess functions in defence response to bacteria (Fig. 8a).

Fig. 8
figure8

Representative interaction network and conserved amino acid sequence analysis of plant SF1s. a Interaction network of Arabidopsis (AT5G51300) based on experimental data. Each network node represents all proteins produced by a single, protein-coding gene locus. Different coloured nodes represent query proteins and the first shell of interactors. Filled nodes represent that some 3D structure is known or predicted, while empty nodes represent proteins of unknown 3D structure. Edges represent protein-protein associations in which proteins jointly contribute to a shared function. b Conserved domains of plant SF1s. The 3D structure of plant SF1 were generated according to the Arabidopsis sequence (AT5G51300) and represented with their target RNA. The ribbon colored by the ConSur Grade (1-blue to 9-purple) represent the conservation grades of the identified peptides of SF1s

The A. thaliana SF1 protein includes three domains: splicing factor 1 helix-hairpin domain (residue: 126–237), KH domain (residue: 244–330) and RNA recognition motif (residue: 482–552). Multiple-sequence alignment revealed that the conservations of these domains are relatively high (Figure S4), suggesting similar functions of these genes. Furthermore, a 3D model of the splicing factor 1 helix-hairpin domain and KH domain were reconstructed according to two crystal structures by using a homology modelling approach (Fig. 8b). The first domain (helix-hairpin domain) forms a secondary, hydrophobic interface with U2AF65 (UHM) [80]. The second one (KH domain) is present in a wide variety of nucleic acid-binding proteins [15]. Therefore, we superimposed the crystal structure of U2AF65 (2M0G) and RNA (1K1G) on the structure from homology modelling to observe the interaction. The residues with higher ConSurf Grade are more conserved. The ConSurf Grade of 198 (74.4%) residues was over 7, and the ConSurf Grade of 111 (41.7%) residues was over 9. More importantly, the binding domain of RNA was highly conserved (Fig. 8b). All of the import residues had a ConSurf Grade higher than 7, except for Val288. The residues at position 288 have similar physiochemical properties, such as Val and Ile. Another domain was not as preserved as splicing factor 1 helix-hairpin domain with a loop interacting with U2AF65. However, the important residues have relatively high ConSurf Grade, and only two residues (Lys146 and Asp147) have ConSurf Grades less than 7. In the lower plants, these two residues are replaced by Ile, Gly, Tyr, Thr, Ala and Gly, Ser, or His. At the same time, they are lost in many species. Therefore, the functions of these domains are conserved. The RNA binding domain is much more conserved than the U2AF65 binding domain, especially in lower plants.

Discussion

It is well known that mature mRNA is formed by sequentially ligating exons to maintain a particular reading frame for protein translation [60]. In human, nearly all annotated protein-coding genes undergo alternative splicing [56, 75]. In plants, over 80% of intron-containing genes exhibit splicing isoforms [11, 82]. Furthermore, the process of splicing is tightly regulated by initial recognition of the splice site during early spliceosome assembly. Therefore, proteins which are responsible for this recognition are important to study and provide valuable targets for genetic control of splicing in eukaryotes [35, 71]. To this end, the branch point binding protein SF1, which connects both 5′ and 3′ splice site determination complexes, emerges as crucial component for splice site choice.

Comparison of structural and functional conservation among plant SF1 genes

In this study, we systematically characterized 92 plant SF1 genes from 59 different species. Although over 50% (34/59) of these species maintained one copy of SF1 gene, 26 plant species contained multiple SF1 members (Table S1), suggesting their functional redundancy. Intriguingly, most of the SF1 genes had one single exon encoding the target protein product except for several algal sequences (Fig. 2), indicating that an ancient gene transposition duplication event may have influenced the evolution of this gene family across the plant lineage [24]. However, further evidence is needed to confirm this hypothesis. At the molecular level, SF1 is an important component to mediate early spliceosome assembly and splice site recognition. Therefore, substantial investigations have been carried out to elucidate its molecular function in both animals and plants. For example, the primary amino acid sequence and domain architecture of SF1 proteins have been reported to be conserved among eukaryotic organisms such as yeast, human, metazoans and plants [2, 6, 30, 47]. SF1 proteins are normally characterized by three domains: KH/QUA2, zinc finger and RRM [36]. However, plant SF1 proteins have been documented to contain an additional RRM domain while lacking UHM-specific features [36]. A previous study demonstrated that a truncated plant SF1 protein without an RRM domain still has sufficient activity for pre-mRNA splicing in response to ABA treatment [36]. Thus, the potential function of this additional domain in planta needs to be further investigated. Furthermore, post-translational modification such as serine phosphorylation by KIS kinase has been reported to enhance the assembly of the SF1–U2AF65–RNA tri-complex [45, 80] or to recruit other splicing factors during splice site recognition [2, 28].

Functional diversification of plant SF1 genes revealed by their expression patterns

SF1 is considered a pivotal component connecting the 5′ and 3′ splice site definition complexes. Furthermore, substantial evidence has demonstrated that SF1 plays crucial roles during splice site recognition among a variety of eukaryotic organisms [46, 52, 68]. However, its role in cell viability remains disputed. Accumulating evidence suggests that SF1 may not be essential for viability and may only control subsets of genes in plants and animals [22, 83], indicating that an alternative mechanism may exist in addition to SF1-mediated splice site recognition [23, 46, 72]. Furthermore, the function of SF1 can be further affected by cell, tissue, or organ specificity. For example, mouse SF1 transcripts have been detected in the brain and heart, implying their tissue-specific regulation at the transcriptional level [83]. Additionally, SF1 is highly expressed in differentiated villous cells, but it is not observed in adenoma or undifferentiated intestinal crypt cells of the intestinal epithelium [49]. In plants, interestingly, SF1 has been found to be involved in a number of plant developmental processes and stress responses [30, 36]. In particular, SF1 has been observed to influence flowering time and leaf size in Arabidopsis and Populus, coincident with its relative high expression in flower parts and leaves (Fig. 6a). Importantly, the SF1 splicing isoforms also exhibit similar expression pattern as SF1 by our qRT-PCR and RT-PCR expression analysis, implicating a reciprocal regulation between SF1 expression and splicing differences during flower and leaf development (Fig. 6a and b). Meanwhile, the expression of SF1 associated with their isoforms were strongly induced by mannitol treatment, indicating a potential function involving the drought stress. Furthermore, transcripts of SF1 are unevenly distributed in several monocots and eudicots (Figs. 6c, S1 and S2), suggesting their potential role during plant development in these species.

In comparison to tissue specificity, more cis-elements involved in hormone and stress responses were observed within promoter regions of plant SF1 genes (Fig. 7 and Table S5), indicating their putative role in response to internal and external stimuli. The Arabidopsis SF1 has been demonstrated to participate in ABA signalling [30, 36], coinciding with the presence of an ABRE motif at its own 5′-flanking region (Fig. 7). Furthermore, Arabidopsis SF1 is induced by IAA at 1 h after treatment and repressed by MeJA (MJ). The AuxRR-core and CGTCA-motifs observed in its promoter region may be responsible for this regulation (Fig. 7). However, further intergrated investigation by using both bioinoformatic and experimental data is required to further strengthen this hypothesis in future functional investigations [9, 10].

Composition of splice site determination complex reveals diverged mechanism to define exon-intron boundary among eukaryotes

In general, eukaryotic SF1s have similar molecular functions to mediate early splice site recognition. Specifically, Arabidopsis SF1 has been proposed to have similar function to its yeast or metazoan counterparts [30, 36]. However, different eukaryotic organisms may evolve their own recognition mechanism during early spliceosome assembly through SF1. First, the target BPS of SF1 is distinct in yeast compared to the sequences in animals and plants. In particular, yeast intronic BPS is a conserved seven-nucleotide sequence (UACUAAC), whereas mammalian SF1 has been reported to bind more degenerate sequences (YNCURAY; N, any nucleotide; R, A or G; Y, C or U) [32]. No conserved BPS has been observed in nematodes and plants at this stage [40, 42]. This poses the question of how SF1 recognizes the BPS in these organisms and whether the additional RRM in plants contributes to this recognition [30]. Second, different coordinative mechanisms are present in a variety of organisms. For example, as the interaction partner of SF1 to coordinate 3′ splice site recognition, mammalian U2AF65 interacts with U2AF small subunit (U2AF35). A similar interaction complex has been found in fission yeast, S. pombe, except the small U2AF subunit is named U2AF23 [69]. In contrast, budding yeast lacks a U2AF35-like small U2AF factor, and the other two proteins (BBP/SF1 and Mud2p/U2AF65) are proposed to form a stable complex during splicing [55]. Furthermore, splicing reactions in animals requires the binding of U2AF65 to Py sequences downstream of BPS, while neither of these two components are necessary for yeast splicing [1, 59]. Intriguingly, plants show a distinct splicing pattern in comparison to animals. For example, a high proportion of intron-retention events has been observed in plants, whereas exon skipping is the dominant AS type in animals [55]. SF1 has been proposed to enhance splicing efficiency of introns containing weakly conserved 3′ splice sites in C. elegans [41]. Therefore, it is tempting to speculate that this difference may result from different SF1-centred splice site recognition between animals and plants.

Conclusion

In this work, we comprehensively identified 92 SF1 sequences from 59 plant species, ranging from algae to eudicots. Subsequent phylogenetic and expression analyses have been carried out to elucidate the conservation and functional regulation of this gene family. By considering the connecting role of SF1 during splice site recognition, we hypothesize that plant SF1s may overlap with but also have distinct function from their animal counterparts. Understanding the molecular mechanism of this protein family in plants provides intriguing possibility to manipulate crop traits through genetic control of plant splicing.

Availability of data and materials

The data are included within the article and its supporting files.

Abbreviations

SF1/BBP:

Splicing factor 1/branchpoint binding protein

BPS:

Branchpoint sequence

RRM:

RNA binding motif

AS:

Alternative splicing

snRNPs:

Small nuclear ribonucleoproteins

U2AF:

U2 snRNP auxilliary factor

KH/QUA2:

K homology/Quaking 2

hnRNP:

Heterogeneous ribonucleoprotein

UHM:

U2AF homology motif

ABA:

Abscisic acid

SF1-HH:

Splicing factor 1 helix-hairpin domain

KH_1:

Homology domain

RRM_1:

RNA recognition motif

GA:

Gibberellin

IAA:

Auxin

MJ:

Methyl jasmonate

PPI:

Protein-protein interaction

ML:

Maximum likelihood

MEME:

Multiple Em for Motif Elicitation

References

  1. 1.

    Abovich N, Liao XC, Rosbash M. The yeast MUD2 protein: an interaction with PRP11 defines a bridge between commitment complexes and U2 snRNP addition. Genes Dev. 1994;8:843–54..

    CAS  PubMed  Google Scholar 

  2. 2.

    Abovich N, Rosbash M. Cross-intron bridging interactions in the yeast commitment complex are conserved in mammals. Cell. 1997;89:403–12..

    CAS  PubMed  Google Scholar 

  3. 3.

    Arning S, Grüter P, Bilbe G, Krämer A. Mammalian splicing factor SF1 is encoded by variant cDNAs and binds to RNA. Rna-a Publication Rna Soc. 1996;2:794–810..

    CAS  Google Scholar 

  4. 4.

    Ashkenazy H, Abadi S, Martz E, Chay O, Mayrose I, Pupko T, Bental N. ConSurf 2016: an improved methodology to estimate and visualize evolutionary conservation in macromolecules. Nucleic Acids Res. 2016;44:W344–50..

    CAS  PubMed  PubMed Central  Google Scholar 

  5. 5.

    Bailey TL, Boden M, Buske FA, Frith M, Grant CE, Clementi L, Ren J, Li WW, Noble WS. MEME SUITE: tools for motif discovery and searching. Nucleic Acids Res. 2009;37:W202–8..

    CAS  PubMed  PubMed Central  Google Scholar 

  6. 6.

    Berglund JA, Chua K, Abovich N, Reed R, Rosbash M. The splicing factor BBP interacts specifically with the pre-mRNA Branchpoint sequence UACUAAC. Cell. 1997;89:781..

    CAS  PubMed  Google Scholar 

  7. 7.

    Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, Madden TL. BLAST plus: architecture and applications. BMC Bioinformatics. 2009;10:421..

    PubMed  PubMed Central  Google Scholar 

  8. 8.

    Chen C, Rui X, Hao C, He Y. TBtools, a toolkit for biologists integrating various HTS-data handling tools with a user-friendly interface; 2018..

    Google Scholar 

  9. 9.

    Chen MX, Sun C, Zhang KL, Song YC, Tian Y, Chen X, Liu YG, Ye NH, Zhang JH, Qu SC, Zhu FY*. SWATH-MSfacilitated proteomic profiling of fruit skin between Fuji apple and a red skin bud sport mutant. BMC Plant Bio. 2019;19:445–55..

    CAS  Google Scholar 

  10. 10.

    Chen MX, Zhu FY, Gao B, Ma KL, Ye NH, Zhang YJ, Fernie AR, Chen X, Hu QJ, Tian Y, Liu TY, Zhang JH, Liu YG. Recognition of the complex genomic organization of rice genes and their coding abilities by using single-molecule long-read sequencing based proteogenomics. Plant Physiol. 2020;182:1–17..

    Google Scholar 

  11. 11.

    Chen MX, Zhu FY, Wang FZ, Ye NH, Gao B, Chen X, Zhao SS, Fan T, Cao YY, Liu TY, Su ZZ, Xie LJ, Hu QJ, Wu HJ, Xiao S, Zhang J, Liu YG. Alternative splicing and translation play important roles in hypoxic germination in rice. J Exp Bot. 2019;70:817–33..

    CAS  PubMed  Google Scholar 

  12. 12.

    Damian S, Andrea F, Stefan W, Kristoffer F, Davide H, Jaime HC, Milan S, Alexander R, Alberto S, Tsafou KP. STRING v10: protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res. 2015;43:D447..

    Google Scholar 

  13. 13.

    Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004;32:1792–7..

    CAS  PubMed  PubMed Central  Google Scholar 

  14. 14.

    Finn RD, Jody C, William A, Miller BL, Wheeler TJ, Fabian S, Alex B, Eddy SR. HMMER web server: 2015 update. Nucleic Acids Res. 2015;43:30–8..

    Google Scholar 

  15. 15.

    García-Mayoral MF, Hollingworth D, Masino L, Díaz-Moreno I, Kelly G, Gherzi R, Chou CF, Chen CY, Ramos A. The structure of the C-terminal KH domains of KSRP reveals a noncanonical motif important for mRNA degradation. Structure. 2007;15:485–98..

    PubMed  Google Scholar 

  16. 16.

    Garrey SM, Rodger V, Andrew J, B. An extended RNA binding site for the yeast branch point-binding protein and the role of its zinc knuckle domains in RNA binding. J Biol Chem. 2006;281:27443–53..

    CAS  PubMed  Google Scholar 

  17. 17.

    Gibson TJ, Thompson JD, Heringa J. The KH domain occurs in a diverse set of RNA-binding proteins that include the antiterminator NusA and is probably involved in binding to nucleic acid. FEBS Lett. 1993;324:361–6..

    CAS  PubMed  Google Scholar 

  18. 18.

    Goodstein DM, Shengqiang S, Russell H, Rochak N, Hayes RD, Joni F, Therese M, William D, Uffe H, Nicholas P. Phytozome: a comparative platform for green plant genomics. Nucleic Acids Res. 2012;40:D1178–86..

    CAS  PubMed  Google Scholar 

  19. 19.

    Gozani O, Potashkin J, Reed R. A potential role for U2AF-SAP 155 interactions in recruiting U2 snRNP to the branch site. Mol Cell Biol. 1998;18:4752–60..

    CAS  PubMed  PubMed Central  Google Scholar 

  20. 20.

    Graveley BR. Alternative splicing: increasing diversity in the proteomic world. Trends Genet. 2001;17:100–7..

    CAS  PubMed  Google Scholar 

  21. 21.

    Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol. 2010;59:307–21..

    CAS  PubMed  PubMed Central  Google Scholar 

  22. 22.

    Guth S, Valcárcel J. Kinetic role for mammalian SF1/BBP in spliceosome assembly and function after polypyrimidine tract recognition by U2AF. J Biol Chem. 2000;275:38059–66..

    CAS  PubMed  Google Scholar 

  23. 23.

    Haihong S, Green MR. RS domains contact splicing signals and promote splicing by a common mechanism in yeast through humans. Genes Dev. 2006;20:1755–65..

    Google Scholar 

  24. 24.

    Hofberger JA, Nsibo DL, Govers F, Bouwmeester K, Schranz ME. A complex interplay of tandem- and whole-genome duplication drives expansion of the L-type lectin receptor kinase gene family in the brassicaceae. Genome Biol Evol. 2015;7:720–34..

    CAS  PubMed  PubMed Central  Google Scholar 

  25. 25.

    Horton P, Park K-J, Obayashi T, Fujita N, Harada H, Adams-Collier CJ, Nakai K. WoLF PSORT: protein localization predictor. Nucleic Acids Res. 2007;35(Web Server):W585–7..

    PubMed  PubMed Central  Google Scholar 

  26. 26.

    Hu B, Jin J, Guo AY, Zhang H, Luo J, Gao G. GSDS 2.0: an upgraded gene feature visualization server. Bioinformatics. 2014;31:1296..

    PubMed  PubMed Central  Google Scholar 

  27. 27.

    Hu ZS, Sun Y, Chen JJ, Zhao YR, Qiao H, Chen RH, Wen XH, Deng YQ, Wen JK. Deoxynivalenol globally affects the selection of 3′ splice sites in human cells by suppressing the splicing factors, U2AF1 and SF1. RNA Biol. 2020;17(4):584–95..

    CAS  PubMed  Google Scholar 

  28. 28.

    Ingham RJ, Karen C, Caley H, Sabine D, Lim CSH, Joanna Y, Kadija H, Judith R, Gerald G, Geraldine M. WW domains provide a platform for the assembly of multiprotein networks. Mol Cell Biol. 2005;25:7092–106..

    CAS  PubMed  PubMed Central  Google Scholar 

  29. 29.

    Jana K, Sophie HM, Angela KM, Igor V. Branch site haplotypes that control alternative splicing. Hum Mol Genet. 2004;13:3189–202..

    Google Scholar 

  30. 30.

    Jang YH, Park H-Y, Lee KC, Thu MP, Kim S-K, Suh MC, Kang H, Kim J-K. A homolog of splicing factor SF1 is essential for development and is involved in the alternative splicing of pre-mRNA in Arabidopsis thaliana. Plant J. 2014;78:591–603..

    CAS  PubMed  Google Scholar 

  31. 31.

    Johnson LS, Eddy SR, Portugaly E. Hidden Markov model speed heuristic and iterative HMM search procedure. Bmc Bioinformatics. 2010;11:431..

    PubMed  PubMed Central  Google Scholar 

  32. 32.

    Keller EB, Noon WA. Intron splicing: a conserved internal signal in introns of animal pre-mRNAs. Proc Natl Acad Sci U S A. 1984;81:7417–20..

    CAS  PubMed  PubMed Central  Google Scholar 

  33. 33.

    Kenichi Y, Masashi S, Yuichi S, Daniel N, Yasunobu N, Ryo Y, Yusuke S, Aiko SO, Ayana K, Masao N. Frequent pathway mutations of splicing machinery in myelodysplasia. Nature. 2011;478:64..

    Google Scholar 

  34. 34.

    Kiana T, Brady SM, Ryan A, Eugene L, Provart NJ. The botany Array resource: e-Northerns, expression angling, and promoter analyses. Plant J. 2005;43:153–63..

    Google Scholar 

  35. 35.

    Kotake Y, Sagane K, Owa T, Mimorikiyosue Y, Shimizu H, Uesugi M, Ishihama Y, Iwata M, Mizui Y. Splicing factor SF3b as a target of the antitumor natural product pladienolide. Nat Chem Biol. 2012;3:570..

    Google Scholar 

  36. 36.

    Lee KC, Yun HJ, Kim SK, Park HY, Thu MP, Lee JH, Kim JK. RRM domain of Arabidopsis splicing factor SF1 is important for pre-mRNA splicing of a specific set of genes. Plant Cell Rep. 2017;36:1–13..

    Google Scholar 

  37. 37.

    Lescot M, Dehais P, Thijs G, Marchal K, Moreau Y, Van De Peer Y, Rouze P, Rombauts S. PlantCARE, a database of plant cis-acting regulatory elements and a portal to tools for in silico analysis of promoter sequences. Nucleic Acids Res. 2002a;30:325–7...

    CAS  PubMed  PubMed Central  Google Scholar 

  38. 38.

    Lin K, Lu R, Tarn WY. The WW domain-containing proteins interact with the early spliceosome and participate in pre-mRNA splicing in vivo. Mol Cell Biol. 2004;24:9176–85..

    CAS  PubMed  PubMed Central  Google Scholar 

  39. 39.

    Liu, Z., Luyten, I., Bottomley, M.J., Messias, A.C., Houngninou-Molango, S., ., Sprangers, R., ., Zanier, K., Krämer, A., Sattler, M. (2001). Structural basis for recognition of the intron branch site RNA by splicing factor 1. Science 294, 1098–1102..

    CAS  PubMed  Google Scholar 

  40. 40.

    Long M, Horvitz HR. Mutations in the Caenorhabditis elegans U2AF large subunit UAF-1 alter the choice of a 3′ splice site in vivo. PLoS Genet. 2009;5:e1000708..

    Google Scholar 

  41. 41.

    Long M, Zhiping T, Yanling T, Sebastian H, Horvitz HR. In vivo effects on intron retention and exon skipping by the U2AF large subunit and SF1/BBP in the nematode Caenorhabditis elegans. Rna-a Publication Rna Soc. 2011;17:2201–11..

    Google Scholar 

  42. 42.

    Lorković ZJ, Wieczorek Kirk DA, Lambermon MH, Filipowicz W. Pre-mRNA splicing in higher plants. Trends Plant Sci. 2000;5:160–7..

    PubMed  Google Scholar 

  43. 43.

    Madhusudhan MS, Marti-Renom MA, Eswar N, John B, Pieper U, Karchin R, Shen MY, Sali A. Comparative protein structure modeling. Curr Protoc Bioinform. 2014;47:5.6.1 editoral board, Andreas D. Baxevanis ... [et al.]..

    Google Scholar 

  44. 44.

    Makarova OV, Makarov EM, Lührmann R. The 65 and 110 kDa SR-related proteins of the U4/U6.U5 tri-snRNP are essential for the assembly of mature spliceosomes. EMBO J. 2014;20:2553–63..

    Google Scholar 

  45. 45.

    Manceau V, Swenson M, Caer JL, Sobel A, Kielkopf C, Maucuer A. Major phosphorylation of SF1 on adjacent Ser-pro motifs enhances interaction with U2AF(65). FEBS J. 2010;273:577–87..

    Google Scholar 

  46. 46.

    Margherita C, Nicolas A, Goranka T, Mihaela Z, Angela KM. Analysis of in situ pre-mRNA targets of human splicing factor SF1 reveals a function in alternative splicing. Nucleic Acids Res. 2011;39:1868..

    Google Scholar 

  47. 47.

    Mazroui R, Puoti A, Krämer A. Splicing factor SF1 from Drosophila and Caenorhabditis: presence of an N-terminal RS domain and requirement for viability. Rna-a Publication Rna Soc. 1999;5:1615–31..

    CAS  Google Scholar 

  48. 48.

    Michaud S, Reed R. An ATP-independent complex commits pre-mRNA to the mammalian spliceosome assembly pathway. Genes Dev. 1991;5:2534..

    CAS  PubMed  Google Scholar 

  49. 49.

    Miki S, Yasuyoshi N, Masashi I, Kazufumi H, Masaya O, Setsuo H, Tesshi Y. Involvement of splicing factor-1 in beta-catenin/T-cell factor-4-mediated gene transactivation and pre-mRNA splicing. Gastroenterology. 2007;132:1039–54..

    Google Scholar 

  50. 50.

    Mo C, Manley JL. Mechanisms of alternative splicing regulation: insights from molecular and genomics approaches. Nat Rev Mol Cell Biol. 2009;10:741–54..

    Google Scholar 

  51. 51.

    Mount SM, Pettersson I, Hinterberger M, Karmas A, Steitz JA. The U1 small nuclear RNA-protein complex selectively binds a 5′ splice site in vitro. Cell. 1983;33:509–18..

    CAS  PubMed  Google Scholar 

  52. 52.

    Noriko H, Tomoko A, David F, Tokio T. Mutations in the SF1-U2AF59-U2AF23 complex cause exon skipping in Schizosaccharomyces pombe. J Biol Chem. 2007;282:2221–8..

    Google Scholar 

  53. 53.

    Park HY, Lee HT, Lee JH, Kim JK. Arabidopsis U2AF65 regulates flowering time and the growth of pollen tubes. Front Plant Sci. 2019;10:569..

    PubMed  PubMed Central  Google Scholar 

  54. 54.

    Park HY, Lee KC, Jang YH, Kim SK, Thu MP, Lee JH, Kim JK. The Arabidopsis splicing factors, AtU2AF65, AtU2AF35, and AtSF1 shuttle between nuclei and cytoplasms. Plant Cell Rep. 2017;36:1113–23..

    CAS  PubMed  Google Scholar 

  55. 55.

    Qiang W, Li Z, Bert L, Rymond BC. A BBP-Mud2p heterodimer mediates branchpoint recognition and influences splicing substrate abundance in budding yeast. Nucleic Acids Res. 2008;36:2787–98..

    Google Scholar 

  56. 56.

    Qun P, Ofer S, Lee LJ, Frey BJ, Blencowe BJ. Deep surveying of alternative splicing complexity in the human transcriptome by high-throughput sequencing. Nat Genet. 2008;40:1413–5..

    Google Scholar 

  57. 57.

    Rain JC, Rafi Z, Rhani Z, Legrain P, Krämer A. Conservation of functional domains involved in RNA binding and protein-protein interactions in human and Saccharomyces cerevisiae pre-mRNA splicing factor SF1. Rna-a Publication Rna Soc. 1998;4:551–65..

    CAS  Google Scholar 

  58. 58.

    Reed R. Mechanisms of fidelity in pre-mRNA splicing. Curr Opin Cell Biol. 2000;12:340–5..

    CAS  PubMed  Google Scholar 

  59. 59.

    Rutz B, Seraphin B. Transient interaction of BBP/ScSF1 and Mud2 with the splicing machinery affects the kinetics of spliceosome assembly. RNA. 1999;5:819–31..

    CAS  PubMed  PubMed Central  Google Scholar 

  60. 60.

    Sasaki-Haraguchi N, Ikuyama T, Yoshii S, Takeuchi-Andoh T, Frendewey D, Tani T. Cwf16p associating with the nineteen complex ensures ordered exon joining in constitutive pre-mRNA splicing in fission yeast. PLoS One. 2015;10:e0136336..

    PubMed  PubMed Central  Google Scholar 

  61. 61.

    Schmittgen TD, Livak KJ. Analyzing realtime PCR data by the comparative CT method. Nat Protoc. 2008;3:1101–8..

    CAS  PubMed  Google Scholar 

  62. 62.

    Selenko P, Gregorovic G, Sprangers R, Stier G, Rhani Z, Krämer A, Sattler M. Structural basis for the molecular recognition between human splicing factors U2AF65 and SF1/mBBP. Mol Cell. 2003;11:965–76..

    CAS  PubMed  Google Scholar 

  63. 63.

    Shih-Peng C, Der-I K, Wei-Yü T, Soo-Chen C. The Prp19p-associated complex in spliceosome activation. Science. 2003;302:282–1..

    Google Scholar 

  64. 64.

    Shitashige M, Satow R, Honda K, Ono M, Hirohashi S, Yamada T. Increased susceptibility of Sf1(+/−) mice to azoxymethane-induced colon tumorigenesis. Cancer Sci. 2010;99:1862–7..

    Google Scholar 

  65. 65.

    Sickmier EA, Frato KE, Shen H, Paranawithana SR, Green MR, Kielkopf CL. Structural basis for Polypyrimidine tract recognition by the essential pre-mRNA splicing factor U2AF65. Mol Cell. 2006;23:49–59..

    CAS  PubMed  PubMed Central  Google Scholar 

  66. 66.

    Siomi H, Matunis MJ, Michael WM, Dreyfuss G. The pre-mRNA binding K protein contains a novel evolutionarily conserved motif. Nucleic Acids Res. 1993;21:1193..

    CAS  PubMed  PubMed Central  Google Scholar 

  67. 67.

    Staley JP, Guthrie C. Mechanical devices of the spliceosome: motors, clocks, springs, and things. Cell. 1998;92:315–26..

    CAS  PubMed  Google Scholar 

  68. 68.

    Tanackovic G, Kramer A. Human splicing factor SF3a, but not SF1, is essential for pre-mRNA splicing in vivo. Mol Biol Cell. 2005;16:1366–77..

    CAS  PubMed  PubMed Central  Google Scholar 

  69. 69.

    Tao H, Josep V, Query CC. Pre-spliceosome formation in S.pombe requires a stable complex of SF1-U2AF(59)-U2AF(23). EMBO J. 2014;21:5516–26..

    Google Scholar 

  70. 70.

    Tom M, Bosiljka T. Alternative pre-mRNA splicing and proteome expansion in metazoans. Nature. 2002;418:236–43..

    Google Scholar 

  71. 71.

    Uehara T, Minoshima Y, Sagane K, Sugi NH, Mitsuhashi KO, Yamamoto N, Kamiyama H, Takahashi K, Kotake Y, Uesugi M. Selective degradation of splicing factor CAPERα by anticancer sulfonamides. Nat Chem Biol. 2017;13:675..

    CAS  PubMed  Google Scholar 

  72. 72.

    Valcárcel J, Gaur RK, Singh R, Green MR. Interaction of U2AF65 RS region with pre-mRNA branch point and promotion of base pairing with U2 snRNA [corrected]. Science. 1996;273:1706–9..

    PubMed  Google Scholar 

  73. 73.

    Vincent G, Olivier G, Micheline FR, Alper R, Alain J, Ulf N. Nuclear retention of unspliced mRNAs in yeast is mediated by perinuclear Mlp1. Cell. 2004;116:63–73..

    Google Scholar 

  74. 74.

    Wahl MC, Will CL, Reinhard L. The spliceosome: design principles of a dynamic RNP machine. Cell. 2009;136:701–18..

    CAS  PubMed  Google Scholar 

  75. 75.

    Wang ET, Rickard S, Shujun L, Irina K, Lu Z, Christine M, Kingsmore SF, Schroth GP, Burge CB. Alternative isoform regulation in human tissue transcriptomes. Nature. 2008;456:470–6..

    CAS  PubMed  PubMed Central  Google Scholar 

  76. 76.

    Wang Y, Xu L, Thilmony R, You FM, Gu YQ, Coleman-Derr D. PIECE 2.0: an update for the plant gene structure comparison and evolution database. Nucl Acids Res. 2016;45:1015..

    PubMed  Google Scholar 

  77. 77.

    Will, C.L., Schneider, C., ., Macmillan, A.M., Katopodis, N.F., Neubauer, G., ., Wilm, M., ., Lührmann, R., Query, C.C. (2014). A novel U2 and U11/U12 snRNP protein that associates with the pre-mRNA branch site. EMBO J 20, 4536–4546..

    Google Scholar 

  78. 78.

    Worden AZ, Lee JH, Mock T, Rouze P, Simmons MP, Aerts AL, Allen AE, Cuvelier ML, Derelle E, Everett MV, Foulon E, Grimwood J, Gundlach H, Henrissat B, Napoli C, McDonald SM, Parker MS, Rombauts S, Salamov A, Von DP, Badger JH, Coutinho PM, Demir E, Dubchak I, Gentemann C, Eikrem W, Gready JE, John U, Lanier W, Lindquist EA, Lucas S, Mayer. Green evolution and dynamic adaptations revealed by genomes of the marine picoeukaryotes micromonas. Science. 2009;324:268–72..

    CAS  PubMed  Google Scholar 

  79. 79.

    Yuan S, Chan HCS, Hu Z. Using PyMOL as a platform for computational drug design. Wiley Interdiscip Rev Comput Mol Sci. 2017;7:e1298..

    Google Scholar 

  80. 80.

    Yun Z, Tobias M, Ivona B, Thomas K, Hyun-Seo K, Peijian Z, Nina MU, Sieber SA, Angela KM, Michael S. Structure, phosphorylation and U2AF65 binding of the N-terminal domain of splicing factor 1 during 3′-splice site recognition. Nucleic Acids Res. 2013;41:1343–54..

    Google Scholar 

  81. 81.

    Zamore PD, Patton JG, Green MR. Cloning and domain structure of the mammalian splicing factor U2AF. Nature. 1992;355:609–14..

    CAS  PubMed  Google Scholar 

  82. 82.

    Zhu FY, Chen MX, Ye NH, Shi L, Ma KL, Yang JF, Cao YY, Zhang Y, Yoshida T, Fernie AR. Proteogenomic analysis reveals alternative splicing and translation as part of the abscisic acid response in Arabidopsis seedlings. Plant J. 2017;91:518–33..

    CAS  PubMed  Google Scholar 

  83. 83.

    Zhu R, Heaney J, Nadeau JH, Ali S, Matin A. Deficiency of splicing factor 1 suppresses the occurrence of testicular germ cell tumors. Cancer Res. 2010;70:7264–72..

Download references

Acknowledgements

Not applicable.

Funding

This work was supported by the National Natural Science Foundation of China (NSFC31701341), Natural Science Foundation of Jiangsu Province (SBK2020042924), NJFU project funding (GXL2018005), which are responsible for the design of the study and data collection. The Natural Science Foundation of Guangdong Province (2018A030313030) and the Natural Science Foundation of Hunan Province (2019JJ50263) play a role in data analysis and interpretation. Shenzhen Virtual University Park Support Scheme to CUHK Shenzhen Research Institute (YFJGJS1.0) and Hong Kong Research Grant Council (AoE/M-05/12, AoE/M-403/16, GRF14160516, 14177617, 12100318) are responsible for the manuscript revision.

Author information

Affiliations

Authors

Contributions

F.Y.Z., M.X.C., and C.W. designed the experiments. K.L.Z., Z.F., and J.F.Y. performed the experiments. K.L.Z., J.F.Y., Y.T., and M.X.C. analysed the data. K.L.Z. and M.X.C. wrote the manuscript. G.F.H., Y.M.F., and J.H.Z. critically commented on and revised the manuscript. All authors have read and approved the manuscript.

Corresponding author

Correspondence to Fu-Yuan Zhu.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors have no conflicts of interest to declare.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Additional file 1:

: Figure S1. Expression patterns of Glycine max (soybean), Solanum lycopersicum (Tomato) and Populus trichocarpa (Poplar) SF1s. Figure S2. Expression pattern of Brachypodium distachyon (Purple false brome) SF1. Figure S3. Expression of Arabidopsis SF1 gene is affected by multiple phytohormone treatments.

Additional file 2:

Figure S4. Multiple alignment of plant SF1 protein sequences.

Additional file 3:

Figure S5. Expression patterns of Zea mays (maize) and Kalanchoe fedtschenkoi (diploid Kalanchoe) SF1s.

Additional file 4:

Figure S6. The full uncropped gel photos of RT-PCR.

Additional file 5:

Table S1. SF1 genes identified from 59 plant species. Table S2. Characteristics of plant SF1 gene structures. Table S3. Predicted subcellular localization of plant SF1 proteins. Table S4. Information of cis-elements identified among plant SF1s.

Additional file 6:

Table S5. List of motifs identified in the 5′-flanking regions of plant SF1s.

Additional file 7:

Table S6. Primers used for RT-PCR and qPCR analysis.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zhang, KL., Feng, Z., Yang, JF. et al. Systematic characterization of the branch point binding protein, splicing factor 1, gene family in plant development and stress responses. BMC Plant Biol 20, 379 (2020). https://doi.org/10.1186/s12870-020-02570-6

Download citation

Keywords

  • Alternative splicing
  • Expression profile
  • Phylogenetics
  • Plants
  • Promoter
  • Splicing factor