Molecular determinants and evolutionary dynamics of wobble splicing.

Alternative splicing at tandem splice sites (wobble splicing) is widespread in many species, but the mechanisms specifying the tandem sites remain poorly understood. Here, we used synaptotagmin I as a model to analyze the phylogeny of wobble splicing spanning more than 300 My of insect evolution. Phylogenetic analysis indicated that the occurrence of species-specific wobble splicing was related to synonymous variation at tandem splice sites. Further mutagenesis experiments demonstrated that wobble splicing could be lost by artificially induced synonymous point mutations due to destruction of splice acceptor sites. In contrast, wobble splicing could not be correctly restored through mimicking an ancestral tandem acceptor by artificial synonymous mutation in in vivo splicing assays, which suggests that artificial tandem splice sites might be incompatible with normal wobble splicing. Moreover, combining comparative genomics with hybrid minigene analysis revealed that alternative splicing has evolved from the 3' tandem donor to the 5' tandem acceptor in Culex pipiens, as a result of an evolutionary shift of cis element sequences from 3' to 5' splice sites. These data collectively suggest that the selection of tandem splice sites might not simply be an accident of history but rather in large part the result of coevolution between splice site and cis element sequences as a basis for wobble splicing. An evolutionary model of wobble splicing is proposed.

Introduction

Alternative splicing plays a major role in the generation of proteomic and functional diversity in metazoan organisms (Black 2003; Blencowe 2006). A subtle variation of transcripts may come from alternative splicing at the tandem splice sites in close proximity (Hiller et al. 2004; Lai et al. 2006; Tsai and Lin 2006; Hiller and Platzer 2008). This often leads to the insertion/deletion (InDel) of a single or a few amino acids while maintaining the open reading frame (Hiller et al. 2004; Lai et al. 2006; Tsai and Lin 2006; Hiller and Platzer 2008). Such wobble splicing is predicted to exist in 30% of human genes and is active in at least 5% of genes (Hiller et al. 2004). Conservation of such events points to important functions (Nakhost et al. 2003; Hiller et al. 2008). In addition, experimental studies have also revealed functional differences for tandem sites that lack deep evolutionary conservation (Condorelli et al. 1994); thus, species-specific alternative splice events at tandem sites may have functional consequences (Hiller et al. 2008). Although wobble splicing only subtly changes the protein sequence and structure (Hiller et al. 2004), it might increase the functional diversity of proteins (Condorelli et al. 1994; Englert et al. 1995; Vogan et al. 1996; Hu et al. 1999; Maugeri et al. 1999; Merediz et al. 2000; Yan et al. 2000; Joyce-Brady et al. 2001; Burgar et al. 2002; Tadokoro et al. 2005; Unoki et al. 2006; Tsai et al. 2007, 2008).

Tandem splice sites are common in human genes, and a considerable fraction of genes have been confirmed to generate wobble splicing isoforms (Englert et al. 1995; Hiller et al. 2004; Tadokoro et al. 2005; Tsai and Lin 2006). Recent reports have indicated that several features of wobble splicing differ from those of constitutive splicing, such as high conservation of the intron sequence upstream of the tandem splice site (Akerman and Mandel-Gutfreund 2006; Koren et al. 2007; Tsai et al. 2007; Hiller et al. 2008), but the mechanisms and basic determinants of tandem splicing are still poorly understood. The clue to understanding this feature requires a promising candidate with conservation over longer evolutionary distances for further experimental studies (Hiller et al. 2008). Synaptotagmin I (SytI), a Ca2+ sensor for synchronous neurotransmitter release, generated two spliced isoforms (SytIVQ and SytI) in Aplysia californica, whereas expression of SytIVQ, but not of SytI, blocked 5-HT-mediated reversal of depression (Nakhost et al. 2003). Here, we envisage sytI as a model to carry out detailed evolutionary analysis of functional wobble splicing from different insects spanning more than 300 My of evolution: Drosophila melanogaster, Culex pipiens (Diptera), Bombyx mori (Lepidoptera), Tribolium castaneum (Coleoptera), Apis mellifera (Hymenoptera), and Pediculus humanus (Phthiraptera). Comparative analysis indicated that the occurrence of species-specific wobble splicing was related to tandem splice sites. Mutagenesis experiments demonstrated that tandem splice sites are necessary but not sufficient for wobble splicing. Moreover, combining comparative genomics with mutational analysis revealed that alternative splicing has evolved from the 3′ to 5′ tandem splice site in C. pipiens, as the result of an evolutionary shift of cis element sequences from 3′ to 5′ splice sites. An evolutionary model of wobble splicing is proposed.

Materials and Methods

Materials

Fruit flies (D. melanogaster), mosquitoes (C. pipiens), silkworms (B. mori), red flour beetles (T. castaneum), honeybees (A. mellifera), and human head lice (P. humanus) were obtained as previously reported (Yang et al. 2008). Water fleas (Daphnia magna) were donated by Dr. Li Shaonan of the Institute of Pesticide and Environmental Toxicology, Zhejiang University. Xenopus laevis were kindly provided by the Institute of Cell and Genetics, Zhejiang University. Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer's protocol. Genomic DNA was isolated using the Universal Genomic DNA Extraction Kit (TaKaRa, Dalian,China).

Gene Structure Analysis and Verification

Genomic DNA sequences and corresponding protein sequences of other insect sytI homologs were obtained through Blast searches, using the fruit fly (D. melanogaster) homolog sequence as bait (Perin et al. 1991). Contigs within which each homolog resides are described below: D. melanogaster: AABU01002692.1; C. pipiens: DS231887.1; B. mori: AADK01011502.1, AADK01020847.1; T. castaneum: NW_001092820.1; A. mellifera: NW_001253173.1, NW_001253172.1; P. humanus: DS235852.1; and Dma: D. magna (FJ550353). Other invertebrate and vertebrate sytI orthologs in Homo sapiens (human) and Lymnaea stagnalis (snail), among other species, were identified by Blast searches using the sequence of the most closely related organism and confirmed by available genome annotation and phylogenetic analysis (Geppert et al. 1994; Zhang et al. 1994; Craxton 2001; Nakhost et al. 2003; Craxton 2004) (supplementary fig. S1, Supplementary Material online). Potential alternative splicing sites were either verified with expressed sequence tag (EST) evidence or reverse transcriptase polymerase chain reaction (RT-PCR). Full-length cDNA clones were obtained using RT-PCR or the 5′/3′ rapid amplification of cDNA ends cDNA synthesis kit (Invitrogen, Carlsbad, CA).

RT-PCR and Verification of Wobble Splicing

Total RNA was reverse transcribed using SuperScript III RTase (Invitrogen) with oligo(dT)15 primer, and the resulting single-stranded cDNA product was treated with DNase I at 37 °C for 30 min. The specific primers flanking the wobble splicing site were designed according to the sytI sequence (table 1). RT-PCR products were gel purified and subjected to direct sequencing with corresponding specific primers. In addition, the products of RT-PCR were purified and cloned into the pGEM-T Easy vector (Promega, Madison, WI) and transformed into JM109 competent cells. Recombinant clones were identified by restriction enzyme digestion and PCR. Sequencing of individual clones was done using an automatic DNA sequencer.

Table 1

Primers Used for the RT-PCR and PCR Analysis

Primer	Sequence	Primer	Sequence
D/ENK verification and quantification
DmeEKVE5′	AGAGTTTGCCCTACGCCGATG	DmeEKVEq5′	GGAGTGGCGCGACCTGGTCA
DmeEKVE3′	AGACAATCCGCCCACGTCCATC	DmeEKVEq3′	CGCAGCGAGAAGCAGATATCT
MsD/ENKVE5′	MGCRATGAACAARACSCTRGTG	MsD/ENKVEq5′	CGAGAACTGCAGAGCGTCGA
MsD/ENKVE3′	TGAYAAWCCDCCRACRTCCATC	MsD/ENKVEq3′	CGTAACGCAGCGAGAAGCAG
AmeDNKVE5′	AGCGTACCTTATGCGGATGCG	AmeDNKVEq5′	AAGCGTCGAGGGTGAAGGTGGT
AmeDNKVE3′	GGATCTGATAAACCACCGACG	AmeDNKVEq3′	CGAAGCGAGAAGCAAATGTCAC
TcaENKVE5′	AGAACATTCCGTACGCCGATGC	TcaENKVEq5′	AATGGCGCGAGCTGCAGAGCGT
TcaENKVE3′	AATCCTCCAACGTCCATTTTC	TcaENKVEq3′	CGTAATGAGAAGCAAATGTCTC
BmoDNKVE5′	TTCAAGAACGTGCCATACGC	BmoDNKVEq5′	GTGAGCTTCAGAGTGTTGA
BmoDNKVE3′	GAGACCACCGACGTCCATTTTC	BmoDNKVEq3′	ACGTAACGCAATGAGAAAC
PhuENKVE5′	AATGTTCCGTATGCTGACGC	PhuENKVEq5′	GAGATTTACAAAGCGTGGA
PhuENKVE3′	TGACAAACCTCCTACATCCAT	PhuENKVEq3′	ACATAACGTAGCGAAAAGCA
DmaENKVE5′	AGTGCCACTCTGCATGGTCGA	SiGAPDH5′-1	CTACTGTTCATGCCACAACTGCT
DmaENKVE3′	AGGATGGCCTTACCAATGGGATC	SiGAPDG3′-1	TGTACTTGATGAGATCAATGACT
DroQEK5′	GGAGTGGAGGGACCTCCA	GgaQEK5′	GTGACTGAGGAGTGGAGGGA
DroQEK3′	GCAGTTGGCACATACCGCA	GgaQEK3′	CGGCAGTAGGCACGTAAC
VQ verification and quantification
DmeVQVE5′	GATCAAAAAGATCGAGCATG	DmeVQVEq5′	AGTTGTTGGGATCGGCGTAC
DmeVQVE3′	TTCTGCTCGCTCTGCTTGTCC	DmeVQVEq3′	TCCTCGGCATTTTCGGTGAG
CpiVQVE5′	AGGGCGAGGCGGAAAAGATC	CpiVQVEq5′	GGCATGAAAGGAGTCGACCT
CpiVQVE3′	GGACTTTGCCCAGCTTCTGC	CpiVQVEq3′	CGGTCAGTTCTTCCATGTC
AmeVQVE5′	CTGTAAGCGTAGTAGTTCTC	AmeVQVEq5′	AACTATTGGGCAGCACGTAC
AmeVQVE3′	ACCTCGCTCTGCTTACTTTC	AmeVQVEq3′	CTTCGGCATTGTCGGTAAG
BmoVQVE5′	GTGTCGGTGTGGTGATTCTG	BmoVQVEq5′	AGCTGTTGGGTTCAGCATAC
BmoVQVE3′	GTTTTCCGAGCTTTTGTTCC	BmoVQVEq3′	TCCTCAGCGTTTTCCGTGAG
PhuVQVE5′	ATGCGGATTCTGTATTCGAC	PhuVQVEq5′	ACAGTTGCTCGGTTCGGCA
PhuVQVE3′	ATCACTCGTTCCTCCCATAT	PhuVQVEq3′	TCTGCATTTTCCGTTAATTC
DpuVQVE5′	GGACTCAAAGGAGCCGTCGACCT	TcaVQVEq5′	GTCCAGTTGCTTGGCTCCTC
DpuVQVE3′	CTTGGCCTCGAGGATGACGACAGT	TcaVQVEq3′	GCATTTTCCGTCAGTTCGTC
Mutagenesis
AmeDNKWS5′	CAGGATAACAAGTTGGGTGAC	AmeDNKSDM5′	AATTACAGGACAACAAGTTGGGTGACATTTG
AmeDNKWS3′	TGTCACCCAACTTGTTATCCTG	AmeDNKSDM3′	CCAACTTGTTGTCCTGTAATTAATTTACACC
BmoVQSDM5′	GACTTGCATGTAGGTTCAACC	TcaENKA3′	GTCACCCAACTTGTTATCCTGTAATATC
BmoVQSDM3′	TATCAGGTTGAACCTACATG	AmeENKC3′	GGGATATCACCTGACCACCTTCACCCT
PhuVQSDM5′	TAGGTTCAGCCGGATATGGAAG	TcaENKin5′	CGACCCCTCCGTTTTTCGAC
PhuVQSDM3′	CTTCCATATCCGGCTGAACCTA	DmeVQin5′	TTTCTAGACGCGCCATTAAATCTTAGT
Variant-specific primers
AmeDNKSS3′	GTCACCCAACTGACCAC	CpiDmeVQ3′-3	TCAGGCTGAACCTTTTCT
CpiDmeVQ3′-1	CCATATCAGGCTTTTCTTTG	CpiDmeVQ3′-4	TCAGGCTGAACCGATTTTAG
CpiDmeVQ3′-2	CCATATCAGGCGATTTTAGGT

The primers were synthesized to perform site-directed mutagenesis. (The mutated nucleotides were shown in italics.)

Quantification of mRNA Isoforms

The quantification of the splice form ratio was performed using RT-PCR gel electrophoresis as described by Tadokoro et al. (2005). Total RNA was isolated from different development stages and tissue samples. RT-PCR was carried out under the following cycling conditions with an initial denaturation of 2 min at 94 °C followed by 30–35 cycles of denaturation at 94 °C for 30 s, annealing at 60–65 °C for 1 min, and extension at 72 °C for 15 s, with a final extension at 72 °C for 10 min. As sytI expression levels in tissues varied extensively, the amount of transcribed RNA in RT-PCR was adjusted to give a similar level of PCR products to facilitate the detection of two isoforms with a 6- or 9-nt difference. Amplified PCR products were separated by electrophoresis through a 12% polyacrylamide gel and then detected by silver staining. Images were captured through a Charge Coupled Device camera, and the quantification of mRNA isoforms was done by comparison of the integrated optical density of detected bands measured by the GIS 1D Gel Image System ver. 3.73 (Tanon, Shanghai, China). The primer sequences are listed in table 1.

Minigene Construction, Site-Directed Mutagenesis, and Transfection

Genomic DNA isolated from T. castaneum, A. mellifera, and B. mori was taken as template, and PCR was performed to attain the corresponding DNA segments encompassing the alternative spliced exon (but not for A. mellifera), the upstream intron, and the upward constitutive exon. Wild-type (WT) minigene DNA was cloned into the pGEM-T Easy vector (Promega). Site-directed mutagenesis was performed on the above gene constructs, having the “AA” changed into “AG” or “AG” changed into “AA,” by successive PCR (Aiyar et al. 1996) with an additional pair of mutating primers (table 1). WT and mutant constructs were further cloned into pBluescript-SK+ under the Baculovirus IE1 promoter (a gift from Chenliang Gong, Shuzhou University, China). All constructs were confirmed by sequencing. The linearized DNA was transfected into silkworm cells (BmN) with Lipofectin (Invitrogen) according to the manufacturer's protocol. Total RNAs were harvested 72 h after transfection using Trizol (Invitrogen). RT-PCR and electrophoresis were carried out to detect the splicing status.

The mosquito and fruit fly hybrid minigene containing mosquito exon 3, 5′ conserved part of intron 3 and fruit fly 3′ conserved part of intron 3 and exon 4 was constructed by PCR with primers (table 1). Based on this construct, hybrid minigenes containing a smaller 5′ conserved part of mosquito intron 3 and a smaller 3′ conserved part of fruit fly intron 3 were constructed by PCR with primers (table 1). All constructs were confirmed by sequencing and were transfected into silkworm cells. RT-PCR and electrophoresis were carried out to detect the splicing status. In addition, the spliced variant-specific primers for PCR were designed in a hybrid minigene experiment (table 1). Each splice product was amplified separately from cDNA using a shared forward primer and a reverse variant-specific primer.

In Vitro Transcription

WT or mutant minigenes were placed downstream of the T7 RNA polymerase promoter. Each transcription vector was linearized with BstXI and had ends trimmed with T4 DNA polymerase prior to in vitro transcription. Transcription was carried out at 37 °C for 1 h in the following system: linearized DNA 500 ng, 10XT7 RNA polymerase buffer 2 μl, 50 mM dithiothreitol 2 μl, 2.5 mM nucleoside triphosphate mix 4 μl, RNase inhibitor (40 U/μl) 0.5 μl, T7 RNA polymerase (Takara) 10 U, and diethylpyrocarbonate-treated water to 20 μl. The plasmid template DNA was then degraded with 2 U DNase I (Ambion, Austin, TX) after transcription. Synthesized RNAs were collected with Trizol (Invitrogen) and quantified.

Xenopus Embryo Microinjection

Microinjection was performed with a HARVARD microinjector according to the operating instructions. Each embryo was injected with 10 nl, and each group consisted of about 30 embryos. Injected embryos were incubated at 18 °C for 1 h before total RNA isolation. RT-PCR and electrophoresis were then carried out to evaluate splicing status.

Results

Species-Specific Wobble Splicing is Related to Synonymous Variation in Splice Sites

Two spliced isoforms (SytIVQ and SytI) generated by wobble splicing in sytI were previously reported in A. californica and were conserved in Drosophila and mammalian sytI (Nakhost et al. 2003). We subsequently analyzed the nature of alternative splicing in sytI orthologs from different insect species. The organisms analyzed were two dipteran species, including D. melanogaster (Adams et al. 2000) and one mosquito (C. pipiens) (http://flybase.org/blast/), the lepidopteran B. mori (Xia et al. 2004; Mita et al. 2004), the coleopteran T. castaneum (http://flybase.org/blast/), the hymenopteran A. mellifera (Honeybee Genome Sequencing Consortium 2006), and the phthirapteran P. humanus (http://flybase.org/blast/). The sequence data from these species allowed us to analyze the evolution of wobble splicing spanning at least 300 My.

Multiple sequence alignment revealed that an AG dinucleotide exists at the intron-distal splice site of the exon only in D. melanogaster, B. mori, and A. mellifera, whereas an AA dinucleotide is present within this site in T. castaneum, C. pipiens, and P. humanus (fig. 1). To validate correlations of this variation with the occurrence of alternative splicing, we designed primers flanking this site in each ortholog. Electrophoresis and sequencing analysis of RT-PCR products were performed to identify alternatively spliced isoforms. As a result, the SytIVQ alternatively spliced form was detected across the SytI orthologs of D. melanogaster, B. mori, and A. mellifera, occurring in a species-dependent manner (fig. 1; supplementary fig. S2, Supplementary Material online). The A. mellifera SytIVQ splicing form was much less pronounced than the corresponding form of D. melanogaster and B. mori (fig. 1; supplementary fig. S2, Supplementary Material online). The wobble splicing occurred in a developmental stage- and tissue-specific manner (supplementary fig. S2 and S2, Supplementary Material online). The smaller splicing form was found at a very low level in the larvae but increased during the stages from pupa to adult (supplementary fig. S2, Supplementary Material online). Tribolium castaneum was genomically incapable of splicing at this location due to loss of the intron (fig. 1). We failed to detect alternative splicing at the 3′ acceptor site in C. pipiens and P. humanus, where the AA instead of the AG sequence is located at orthologous splice junctions (fig. 1). Alternative splicing at this tandem site is generally conserved in distantly related species (i.e., D. magna; Loligo pealei; fig. 2), suggesting that SytIVQ should represent the ancestral state in insects. SytIVQ and SytI deviated during the course of insect evolution due to the loss of an intron-distal splice site at the exon 3–4 junction (fig. 1). At the diversion of hymenopteran and phthirapteran orders, an AG dinucleotide was converted into an AA in phthirapterans, whereas an AG remained in the hymenopterans. Similarly, at the diversion of coleopteran and hymenopteran orders, an AG dinucleotide was converted into an AA in the coleopterans, whereas SytIVQ remained in hymenopterans.

F

Conservation and divergence analysis of wobble splicing in insect sytI genes. (A). Schematic illustration of insect sytI gene structure. Exons are represented by black boxes and the exon–intron architecture is magnified within open boxes. Introns are represented by black lines. “Ins” or “Del” denotes the splicing pattern leading to amino acid insertion or deletion, respectively. Amino acid insertion or deletion is either indicated with red letters or inverted triangles. (B). Phylogeny of wobble splicing in insect sytI genes from Drosophila melanogaster (Dme), Culex pipiens (Cpi), Bombyx mori (Bmo), Tribolium castaneum (Tca), Apis mellifera (Ame), and Pediculus humanus (Phu). Exonic and intronic sequences are shown in capital and lowercase letters, respectively. The tandem splice sites [AG(N)3AG] diverged during the course of evolution, whereas the distal splice sites were lost in some species. Tribolium castaneum was genomically incapable of splicing at this location due to the intron loss. “WSed” denotes “wobble-spliced.”. (C) Verification of alternative splicing by direct sequencing of RT-PCR products at the former site (VQ) and the latter site [D/E(N)K]. Overlapping nucleotide sequence signals result from wobble splicing.

F

Phylogenetic analysis of wobble splicing in metazoan sytI genes. Abbreviations: Lpe: Loligo pealei (CAA51079.1, BAA09866); Aca: Aplysia californica (AAQ91785, P41823); Lst: Lymnaea stagnalis (AAO83847); Pdu: Platynereis dumerilii (ABR68850.1); Dja: Dugesia japonica (BP191518, BP188154); Sme: Schmidtea mediterranea (EG350483.1); Cca: Capitella capitata (EY537522.1); Dme: Drosophila melanogaster (AAF51205); Cpi: Culex pipiens (FJ550348); Bmo: Bombyx mori (FJ550351); Tca: Tribolium castaneum (FJ550350); Ame: Apis mellifera (FJ550352); Phu: Pediculus humanus (FJ550349); Dma: Daphnia magna (FJ550353); Spu: Strongylocentrotus purpuratus (AAB67801.3); Lva: Lytechinus variegatus (ABA29124.1); Hsa: Homo sapiens (BX488459.1, DA206355.1); Ptr: Pan troglodytes (DC525743.1, DC521107.1); Mfa: Macaca Fascicularis (CJ462565.1, CJ465758.1): Bta: Bos taurus (DN522068.1, DV892458.1); Rno: Rattus norvegicus (AA924659, CB731032.1); Mmu: Mus musculus (NP_033332.1, Q5R4J5.1); Gga: Gallus gallus (NP_990502.1, BU141735.1). Xtr: Xenopus tropicalis (CX838804, CX839099.1); Dre: Danio rerio (CK400350.1, EB976334.1). Phylogenetic tree data were adapted and modified from Dunn et al. (2008). Amino acid residues showing complete conservation are shaded in black. Residues shaded in green were confirmed to be alternatively deleted according to cDNA and EST evidence. Residues in red represent those lacking a tandem splice site. Residues in yellow are those with tandem splice sites without cDNA and EST evidence.

A similar situation also occurs in the seventh exon in sytI. We discovered that the B. mori sytI sequence had a Δ9 tandem acceptor resulting in an InDel of three amino acids (DNK) prior to the C2B domain (fig. 1). Cross-species alignment of this exon revealed that an AG dinucleotide existed at the intron-distal splice site in all of these species except A. mellifera, where an AA dinucleotide was present at this site (fig. 1). RT-PCR analysis followed by sequencing confirmed that this InDel of three amino acids (DNK) was also conserved across all sytI orthologs except for A. mellifera, which differed in the ratio of InDels (supplementary fig. S2, Supplementary Material online). Similar to the VQ site, wobble splicing at this site occurred in a developmental stage–specific and tissue-specific manner (supplementary fig. S2 and , Supplementary Material online). Moreover, D. melanogaster has a tandem acceptor separated by only 6 nt, resulting in the InDel of a dipeptide EK (fig. 1). We detected only one form of spliced transcript in A. mellifera embryos, larvae, pupae, and adults (supplementary fig. S2, Supplementary Material online). Similar to previous reports (Tadokoro et al. 2005; Hiller et al. 2006; Unoki et al. 2006; Tsai et al. 2007), these phylogenetic correlations between AG/AA dinucleotide variations and alternative splicing occurrence indicates that the tandem splice site is necessary for this type of alternative splicing.

Evolution of Wobble Splicing

SytI proteins are evolutionarily ancient proteins that are conserved throughout the animal kingdom. We analyzed the amino acid sequences of some SytI proteins in Platyhelminthes, Annelida, Mollusca, Arthropoda, Echinodermata, and Chordata. The sequence data from these species allowed us to trace the origin and divergence of wobble splicing. Similar to A. californica and D. melanogaster, two spliced isoforms (SytIVQ and SytI) generated by wobble splicing were found in squid (fig. 2; supplementary fig. S3, Supplementary Material online). In contrast, only one spliced isoform was found in another molluscan, L. stagnalis, where the AA instead of the AG sequence is located at orthologous sites (fig. 2; supplementary fig. S3, Supplementary Material online). Interestingly, alternative splicing at tandem acceptors has also been found in this exact region of the juxtamembrane linker of almost all vertebrate SytI proteins, from low-level fish to high-level mammals, where the amino acids ALK are inserted instead of VQ (fig. 2). Notably, two wobble spliced isoforms (SytIIQ and SytI) have been found in the linker region of SytI in sea urchin (Strongylocentrotus purpuratus), which seems to be intermediate between VQ and ALK, consistent with the evolutionary position of the Echinodermata between invertebrates and vertebrates. Intriguingly, a wobble splicing event at this site is generally conserved in Mollusca, Arthropoda, Echinodermata, and Chordata (fig. 2; supplementary fig. S3, Supplementary Material online), indicating that this case is among the few alternative splice events that might have a very deep evolutionary history spanning ∼550 My. During evolution, wobble splicing was lost in some species, such as T. castaneum and P. humanus. In addition, the InDel peptides have diverged into VQ (IQ) and ALK after vertebrate–invertebrate separation.

In contrast to the VQ site, we seldom detected the alternative QEK deletion in the equivalent region of insect D/E(N)K InDel in all vertebrate SytI proteins by Blast-searching against EST databases. As a result, most of these vertebrate sytI genes lack the tandem splice site in this region, except for Gallus gallus and Danio rerio (fig. 2; supplementary fig. S3, Supplementary Material online). The alternative splicing at these tandem splice sites of D. rerio and G. gallus was confirmed by electrophoresis and sequencing of RT-PCR products (data not shown). In addition, Schmidtea mediterranea, Dugesia japonica, and Capitella capitata (http://genome.jgi-psf.org/euk_cur1.html), which lack the tandem splice site at this site, are thought to generate only one larger spliced isoform, according to EST-based database analyses. However, alternative splicing at this tandem splice site is generally conserved in insect phyla (fig. 2), where an AG dinucleotide represents the ancestral state of this site, at least before the radiation of the coleopteran and hymenopteran orders (spanning ∼300 My). At the diversion of coleopteran and hymenopteran orders, an AG dinucleotide was converted into an AA in hymenopterans, whereas an AG remained in place in the coleopterans. Analysis of these orthologs from 12 Drosophila species (D. melanogaster, Drosophila simulans, Drosophila sechellia, Drosophila yakuba, Drosophila erecta, Drosophila ananassae, Drosophila pseudoobscura, Drosophila persimilis, Drosophila willistoni, Drosophila mojavensis, Drosophila virilis, and Drosophila grimshawi) (), three mosquitoes (Anopheles gambiae, C. pipiens, and Aedes aegypti) (Holt et al. 2002, http://flybase.org/blast/), and the lepidopteran B. mori (Xia et al. 2004; Mita et al. 2004) indicates an AG in these orthologs, whereas AA is also located in Nasonia vitripennis, another hymenopteran species (Honeybee Genome Sequencing Consortium 2006; http://flybase.org/blast/). Compared with the former VQ site, wobble splicing at this site is confined to a smaller phyletic branch (fig. 2).

Wobble Splicing Might be Destroyed by Single Synonymous Variation

Comparative analysis indicates that wobble splicing depends upon whether the 3′ splice junction sequence is AG or AA in insect phyla (fig. 1). It is noteworthy that this AG/AA variation is synonymous but occurs at tandem splice sites. In this way, single synonymous variation could alter wobble splicing. To mimic the natural evolution at tandem acceptors and to investigate their effect on wobble splicing, we introduced a single synonymous change into exon 4 of B. mori sytI. Site-directed mutagenesis at the equivalent position to P. humanus in the B. mori sytI exon 7 was performed, converting the AG into an AA dinucleotide at the 3′ splice junctions (fig. 3). The resulting minigene was evaluated for splicing efficiency and accuracy. As a result, evident wobble splicing of the WT minigene occurred not only in the silkworm cell transfection system but also in the Xenopus embryo injection system (fig. 3), indicating that this construct is clearly sufficient to direct wobble splicing in heterologous splicing systems. However, we only detected one spliced form in the mutated minigene construct in both in vivo splicing assays (fig. 3). Thus, wobble splicing could be abolished through artificial point synonymous substitution in the sytI exon. This means that this synonymous variation in sytI cannot evolve neutrally but is significantly constrained by splicing requirements.

F

Artificial synonymous mutation affects wobble splicing. (A) Schematic diagrams of WT and mutant (Mu) minigene constructs involved in in vivo splicing assays. T7 and IE1 promoters were used in the Xenopus embryo microinjection system and silkworm cell transfection, respectively. Bmo: Bombyx mori; Ame: Apis mellifera. (B) Electrophoresis analysis of splicing profiles. The minigene constructs were introduced into silkworm cells, and splicing profiles were analyzed by electrophoresis in parallel with RT-PCR products. In addition, RNA transcribed with the WT and mutant minigenes by T7 RNA polymerase was subjected to the Xenopus embryo microinjection system followed by RT-PCR and electrophoresis analysis. The small spliced forms are marked with arrows. The splicing patterns in B. mori and A. mellifera embryos were taken as controls (“first lane”). (C) Single synonymous variation can destroy wobble splicing. The minigene constructs were introduced into silkworm cells, and the splicing profiles were analyzed by direct sequencing of RT-PCR products. In addition, wobble splicing of WT and mutant minigenes was analyzed in the Xenopus embryo microinjection system. The mutated nucleotides are marked with asterisks. (D) Mimicking an ancestral tandem acceptor by single synonymous mutation fails to restore A. mellifera wobble splicing in in vivo splicing assays.

Mimicking an Ancestral Tandem Acceptor by Single Synonymous Mutation Fails to Restore Wobble Splicing

To test whether a wobble splicing event can be restored by mimicking an ancestral tandem acceptor, AA-to-AG site-directed mutagenesis was performed at the equivalent position in the A. mellifera sytI (fig. 3), which may create a novel 3′ tandem acceptor site like that in T. castaneum. WT and mutated minigene constructs encompassing exon 6, the downstream intron, and downstream exon 7, respectively, were prepared based on A. mellifera genomic DNA. In transfection experiments on silkworm cells, the mutated minigene could be correctly spliced, but we failed to detect the alternatively spliced form (fig. 3B and D). We found a similar trend in in vivo splicing assays using the Xenopus embryo injection system (fig. 3). In addition, we also failed to detect the alternatively spliced form with a variant-specific primer (data not shown). These results indicate that mimicking an ancestral tandem acceptor fails to restore wobble splicing. Mimicking a tandem acceptor by single synonymous mutation in P. humanus also showed similar result. Recognition of the splice sites involves an interaction between the cis elements and trans-acting factors (Liang et al. 2003). Failure to produce wobble splicing should not be due to a variation of trans-acting factors, because wobble splicing of the minigene could occur in both in vivo splicing assays (fig. 3). A recent study indicated that mutations of the region between the branch site and the NAGNAG 3′ splice site could affect the ratio of the distal–proximal AG selection (Tsai et al. 2007). In addition, a further report indicates that several features of wobble splicing differ from those of constitutive splicing, such as high conservation of the intron sequence upstream of the tandem splice site (Akerman and Mandel-Gutfreund 2006; Koren et al. 2007; Hiller et al. 2008). Moreover, single nucleotide polymorphisms (SNPs) around the NAGNAG motif could affect the splice site choice, which may lead to a change in wobble splicing pattern (Hiller et al. 2006; Tsai et al. 2007). Finally, a further experiment indicated that the mutation of AA to AG in the A. mellifera minigene caused wobble splicing after the A. mellifera sequence upstream of the tandem spice site was replaced by that of T. castaneum. These data suggest that failure to produce wobble splicing may contribute to a lack of cis elements, which are involved in precise recognition of the tandem splice sites. Overall, our evidence from mutagenesis experiments directly demonstrates that the tandem splice sites are essential, but not alone sufficient, for wobble splicing.

Alternative Splicing Evolved from 3′ to 5′ Tandem Sites as a Result of an Evolutionary Shift of cis Element Sequences

We failed to detect 3′ alternative splicing in exon 4 of C. pipiens, whereas AA, instead of AG, is located at orthologous distal 3′ splice junctions. However, we detected alternative splicing at the 5′ tandem donor site in exon 3, resulting in deletion of 11 amino acids (fig. 4). Interestingly, the first two InDel amino acids are VQ, which are identical to the deleted sequence resulting from 3′ wobble splicing in exon 4 in other species (fig. 4). Thus, the two isoforms, VQ InDel, are caused by alternative splicing at the 3′ tandem acceptor in D. melanogaster, B. mori, and A. mellifera, whereas VQ-like InDel isoforms are generated by alternative splicing at the 5′ tandem donor in C. pipiens (fig. 4). Wobble splicing, producing two spliced isoforms with or without VQ, can regulate 5-HT-mediated reversal of depression in Aplysia SytI (Nakhost et al. 2003). Considering the similarity of InDel peptides by 5′ and 3′ wobble splicing, it is tantalizing to think that loss of alternative splicing at the 3′ tandem acceptor may be compensated by alternative splicing at the 5′ tandem donor in C. pipiens.

F

Comparison of alternative splicing at the 5′ tandem donor in Culex pipiens and at the 3′ tandem acceptor in Drosophila melanogaster. (A) A schematic diagram of wobble splicing in C. pipiens (Cpi) and D. melanogaster (Dme). Exon sequence is represented by a black box and the alternative sequence (extension) by an empty box. Introns are shown by black lines; alternative splice events are shown by dashed lines; (B) Sequence alignment of sytI genes from Drosophila species. Abbreviations: Dse, Drosophila sechellia; Dsi, Drosophila simulans; Dya, Drosophila yakuba; Dps, Drosophila pseudoobscura; Der, Drosophila erecta; Dan, Drosophila ananassae; Dpe, Drosophila persimilis; Dwi, Drosophila willistoni; Dgr, Drosophila grimshawi; Dvi, Drosophila virilis; Dmo, Drosophila mojavensis; and Dvi, Drosophila virili. Blue and red indicate invariant exon and intron positions, respectively. Alternative and constitutive splice sites are indicated by an empty and black arrow, respectively. (C) Sequence alignment of sytI genes from Anopheles species. Aga: Anopheles gambiae; Cpi: C. pipiens; and Aae: Aedes aegypti. The distal splice site sequences “TCGGTACA” in mosquito, which are identical to that of D. melanogaster, are boxed. (D) Splicing assay using a hybrid minigene. Mosquito and fruit fly hybrid minigenes contained mosquito exon 3, the 5′ conserved part of intron 3, and fruit fly 3′ conserved part of intron 3 and exon 4. Transient transfection of silkworm cells was performed using a hybrid minigene. The hybrid minigene could give rise to four splice isoforms (a–d) by alternative splicing at the 5′ tandem donor and 3′ tandem acceptor. Variant-specific primers (VSP1-4) and shared primers (P1) were designed (table 1) (E) Identification of four splice variants by clone sequencing (a–d). (F) Identification of four splice variants by variant-specific primers. Each splice variant could be distinguished separately from cDNA using a shared forward primer and a reverse variant-specific primer. The products were confirmed to have expected sequences by sequencing. (G) The relative occurrence of the four isoforms from the hybrid minigene (Cpi–Dme) was calculated according to sequencing of cDNA clones, and cDNA clones from C. pipiens (Cpi) and D. melanogaster (Dme) were compared. Sixty to eighty clones were sequenced for every sample. “*” represents the data based on supplementary figure S2, Supplementary Material online.

Interestingly, D. melanogaster also contains a 5′ tandem donor sequence, where TCGGTACA is identical to the distal splice site in mosquito (fig. 4) but fails to undergo alternative splicing. To gauge cis elements that may contribute to species-specific alternative splicing, we compared genomic sequences flanking alternative splice sites from 12 Drosophila species spanning a ∼40-My divergence time (). Sequence alignment revealed invariant intronic and exonic regions in the vicinity of alternative 3′ splice sites (fig. 4), which may contain putative cis elements for 3′ tandem acceptor site selection. The remaining intron and exon sequences at the canonical 5′ splice site were relatively divergent. Similarly, genomic sequence alignment from three mosquito species (Holt et al. 2002; http://flybase.org/blast/) also revealed invariant intronic and exonic regions at the 5′ intron–exon boundary (fig. 4), whereas the remaining intronic and exonic sequences at canonical 3′ splice sites were highly divergent. Thus, alternative splicing diverged from 3′ tandem donor to 5′ tandem acceptor in C. pipiens after the Drosophilidae–Culicidae separation as a result of an evolutionary shift of cis element sequences from 3′ to 5′ splice sites.

To elucidate these conserved elements involved in wobble splicing, we constructed mosquito and fruit fly hybrid minigenes containing C. pipiens exon 3, 5′ conserved part of intron 3 and D. melanogaster 3′ conserved part of intron 3 and exon 4 (fig. 4). As a result, sequencing indicated that the 5′ wobble splicing site of the hybrid minigenes was identical to that of mosquito endogenous sytI transcripts, whereas the 3′ wobble splicing site was identical to that of fruit fly endogenous transcripts (fig. 4). In addition, each splice variant could be distinguished separately from cDNA using a shared forward primer and a reverse variant-specific primer (fig. 4). This indicates that the conserved cis elements of the hybrid minigene were sufficient to direct wobble splicing at the 5′ and 3′ tandem splice sites. Notably, analysis of expression ratio indicated that this minigene was spliced at a much lower relative splicing ratio at the 5′ intron–proximal tandem donor than the mosquito (fig. 4), suggesting that the fruit fly fragment could affect 5′ tandem acceptor selection and led to predominant usage of the intron-distal 5′ splice site. Conversely, the mosquito fragment also affected 3′ splice site selection of this minigene but led to predominant usage of the intron-proximal 3′ splice site.

Discussion

Synonymous Mutation-Mediated Wobble Splicing

Splice acceptors with the genomic NAGNAG motif may cause NAG InDels in transcripts that occur in 30% of human genes and are functional in at least 5% of human genes (Hiller et al. 2004). The intron phase determines the outcome of a NAG InDel and includes single amino acid InDels, exchange of single amino acids, an unrelated dipeptide, or the creation or destruction of a stop codon (Hiller et al. 2004). NAG-to-NAA conversion in phase 0 (synonymous mutation) cannot change the ancestral protein isoform sequence, whereas NAG-to-NAA conversion in phase 1 or 2 will possibly result in alteration of its amino acid sequence. Thus, NAG-to-NAA conversions in phase 1 or 2 are also constrained by the protein coding sequence, in addition to splicing. This is evidenced by the SNP data in tandem splice sites, where 77% (49 of 64) of SNPs in plausible NAGNAGs are translationally silent but introduce a novel variability at the protein level by changing the acceptor (Hiller et al. 2006). Therefore, wobble splicing in phase 0 is thought to be regulated more compatibly than in phases 1 and 2. Synonymous mutation can convert a tandem acceptor [AG(N)3AG] that allows alternative splicing to a nontandem acceptor, AG(N)3 AA, that allows only the expression of one transcript. Conversely, conversion of a nontandem acceptor to a tandem acceptor cannot simply mean the occurrence of alternative splicing, which still requires evolutionary coordination of a tandem site and cis element sequences.

Functional Implications

Wobble splicing provides a mechanism to create subtle changes, which may be of functional relevance by changing local hydrophobicity and charge (Hiller et al. 2006). Wobble splicing was detected at two conserved sites in sytI. The former has good evidential support for regulating 5-HT-mediated reversal of depression in Aplysia (Nakhost et al. 2003), whereas the conservation of the latter splicing event throughout insect evolution points to an important role for this domain. By analyzing the splice site within the context of the 3D structure of the protein domain in which it resides, we are able to propose functional consequences for the InDel of the DNK tripeptide sequence. Homology modeling evidence showed that the three amino acid–DNK tripeptide sequence is just in front of the C2B domain (supplementary fig. S4, Supplementary Material online). Intriguingly, a comparison between 3D structures of the protein domains of these sequences revealed an extra alpha-helix structure within the C2B domain linker region (supplementary fig. S4, Supplementary Material online). This subtle structural difference between the two splice isoforms might play an important role for Ca2+ binding and conformational change (Garcia et al. 2004). Moreover, regional electron capacity analysis revealed that the linker of the insertion splice isoform showed more hydrophilicity than the deletion isoform linker (supplementary fig. S4, Supplementary Material online). Because the D/E(N)K tripeptide sequence is located within the linker region between C2A and C2B domains, it seems likely that wobble splicing might regulate the precise interactions between C2A and C2B domains, or their binding activity with SNARE, by changing hydrophobicity and charge of the linker.

Interestingly, sytI is also a target of A-to-I RNA editing in some insects and squid (Reenan 2005; Yang et al. 2008). In addition, sytI also undergoes exon duplication and alternative splicing or other functional wobble splicing in C. elegans and A. californica (Nakhost et al. 2004; Mathews et al. 2007). These events provide another means to create subtle changes at posttranscriptional stages. Thus, more SytI isoforms are possible through RNA editing and alternative splicing, suggesting that there is a general adaptive benefit to managing the regulation of neurotransmitter release in this way.

An Evolutionary Model of Wobble Splicing

Although these tandem splice sites (i.e., NAGNAG) are common in human genes, only a small subset of sites with this motif is confirmed to be involved in alternative splicing (Liang et al. 2003; Hiller et al. 2004). A recent report indicated that several features of wobble splicing differ from those of constitutive splicing, such as high conservation of the intron sequence upstream of the tandem splice site (Akerman and Mandel-Gutfreund 2006; Koren et al. 2007; Tsai et al. 2007; Hiller et al. 2008). Moreover, evolutionary dynamics of cis element sequences also support this view, as evidenced by the shift of cis element sequences underlying tandem splice site selection (fig. 4). Our evidence from mutagenesis experiments directly demonstrates that the tandem splice sites are necessary but not sufficient alone for wobble splicing. These data suggest that the selection of tandem splice sites might not simply be an accident of history but rather in large part the result of evolutionary coadaptation of splice site and cis element sequences as a basis for wobble splicing, driven by positive selection. Based on these findings, a plausible evolutionary model for wobble splicing is proposed (fig. 5). The first phase involves the emergence of a novel splice site around an ancestral constitutive exon sequence in the course of evolution, for example, by synonymous mutation from A to G that can result in tandem acceptors. During phase II, the additional cis elements produced are compatible with a distal splice site through nucleotide mutation, which allows alternative splicing. The compatibility between splice site and cis elements was initially considered to lack biological functionality, but that such functionality was acquired during evolution. The newly created tandem acceptor site might be located within the regulatory range of a cis-acting element, and thus alternative splicing occurs immediately after the mutation event, as evidenced by SNP and EST data (Hiller et al. 2006). Alternatively, existing donor or acceptor motifs can function as splice sites if the strength of an adjacent constitutive splice site is reduced by a mutation (Hiller and Platzer 2008). Furthermore, the functional distal splice site might be destroyed during subsequent evolution, thus resulting in the loss of wobble splicing (i.e., A. mellifera SytIDNK). In addition to neutral mutation, an alternative option to remove a deleterious wobble splicing event might be to destroy the tandem splice site. Correspondingly, the additional cis elements are also degenerated. In such a case, wobble splicing fails to occur precisely even if the distal splice site is restored.

F

An evolutionary model of wobble splicing. The exon sequence is represented by a black box and the alternative sequence (extension) by a red box. Introns are represented by black lines. The ancestral constitutive splice events are shown in black dashed lines, whereas the alternative splice event is shown as a red dashed line. The cis elements for wobble splicing are denoted by stars and triangles. The first phase involves the emergence of a novel splice site around an ancestral constitutive exon sequence, for example, by synonymous mutation from A to G that can result in a NAGNAG acceptor. During phase II, the additional cis elements were made to be compatible with an intron-distal splice site through nucleotide mutation, which allows alternative splicing. Finally, the functional distal splice site might be destroyed during subsequent evolution, thus resulting in the loss of wobble splicing. Correspondingly, the additional cis elements would be degenerated.

Supplementary Material

Supplementary figures S1–S4 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/). Sequence data from this article have been submitted to GenBank (GenBank accession numbers—C. pipiens: FJ550348; B. mori: FJ550351; T. castaneum: FJ550350; A. mellifera: FJ550352; P. humanus: FJ550349; and D. magna: FJ550353).