Expression of Alternatively Spliced Sodium Channel (cid:1) -Subunit Genes UNIQUE SPLICING PATTERNS ARE OBSERVED IN DORSAL ROOT GANGLIA*

Molecular medicine requires the precise definition of drug targets, and tools are now in place to provide ge-nome-wide information on the expression and alternative splicing patterns of any known gene. DNA microarrays were used to monitor transcript levels of the nine well-characterized (cid:1) -subunit sodium channel genes across a broad range of tissues from cynomolgus monkey, a non-human primate model. Alternative splicing of human transcripts for a subset of the genes that are expressed in dorsal root ganglia, SCN8A (Na v 1.6), SCN9A (Na v 1.7), and SCN11A (Na v 1.9) was characterized in detail. Genomic sequence analysis among gene family paralogs and between cross-species orthologs suggested specific alternative splicing events within transcripts of these genes, all of which were experimentally confirmed in human tissues. Quantitative PCR revealed that certain alternative splice events are uniquely expressed in dorsal root ganglia. In addition to characterization of human transcripts, alternatively spliced sodium channel transcripts were monitored in a rat model for neuropathic pain. Consistent down-regulation of all transcripts was observed, as well as significant changes in the splicing patterns of SCN8A and SCN9A. Alternative splicing of primary gene transcripts provides a mechanism to generate functionally distinct protein isoforms from a single gene. For the development of safe and efficacious therapeutic compounds, it is necessary to identify the reper-toire of proteins that can arise from a gene targeted for therapeutic

Alternative splicing of primary gene transcripts provides a mechanism to generate functionally distinct protein isoforms from a single gene. For the development of safe and efficacious therapeutic compounds, it is necessary to identify the repertoire of proteins that can arise from a gene targeted for therapeutic intervention and determine their tissue distribution within the body. The completion of several mammalian genome sequences, coupled with rich resources provided by extensive expressed sequence tag (EST) and cDNA sequencing, present opportunities for computational prediction of alternative splicing (1)(2)(3)(4). The UCSC genome browser (genome.ucsc.edu), which displays overlapping tracks of mRNAs, ESTs, and comparative genomic conservation, can also facilitate the identification of potential alternative splice events. DNA microarrays that monitor exon-exon junctions directly across a broad range of transcripts provide an additional resource to detect alterna-tive splicing on a genome-wide scale (5,6). These combined computational and experimental approaches, coupled with traditional laboratory validation, provide a wealth of information about alternative splicing and tissue-specific expression, which are essential to define a drug target at the molecular level.
Sodium channels are multisubunit protein complexes that play a pivotal role in the propagation of action potentials along neurons (7,8). The ␣-subunit genes encode the primary channel-forming pores within the cell membrane that allow ionspecific translocation. The ␣-subunit gene family contains nine paralogs (and one additional sodium channel-like gene, Na x ) that are highly conserved across vertebrate species (9). The channel protein structure includes four highly similar clusters of transmembrane helices that are connected by intracellular loops. A similar structure is found in calcium and potassium channels, indicating that this ion channel superfamily arose from a single primordial ion channel gene (8,10). Voltage-gated sodium channels perform a broad spectrum of functions within vertebrate cells, as is evident from the large number of paralogous genes and their tissue-selective expression patterns. For a particular sodium channel gene, subtle differences in channel properties can be attributed to alternative splicing, post-translational modification, changes in the expression of ancillary ␤-subunits, and mutation (7,8). Importantly, alternative splicing of transcripts derived from a common gene has been shown to generate biochemically and pharmacologically distinct sodium channel isoforms (11,12).
We chose to focus our attention on sodium channels expressed in dorsal root ganglia (DRG), 1 peripheral nervous system (PNS) structures found just outside the spinal column that play key roles in sensory transmission from the periphery to the brain. Channels expressed in DRG are known to play key roles in nociception (7,8,13). Modulation of DRG sodium channel activity may provide relief from neuropathic pain, a medical condition that is not well addressed by current medicinal therapies (13). Our goal was to catalog and quantify alternative splicing events that occur in SCN8A (encoding Nav1.6, PN4), SCN9A (encoding Nav 1.7, PN1), and SCN11A (encoding Nav 1.9, PN5). As an example, previous research had shown alternative splicing of SCN8A transcripts (14,15). In rat DRG, alternative splicing extends the reading frame of exon 11, resulting in a channel that has altered kinetics of inactivation and reactivation relative to the non-extended isoform (12). Developmentally regulated alternative splicing of SCN8A coding exon 18 in mouse and human results in a transcript that encodes a truncated, nonfunctional channel that appears in * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  fetal tissue but vanishes just after birth (14). Finally, a splicing event involving a mutually exclusive alternative exon 6 was inferred from human genomic sequence (15). A recent report of DRG-selective expression of an alternatively spliced, functionally distinct variant of the voltage-gated calcium channel CACNA1B (Cav2.2) suggests that DRG-specific alternative splicing may generate a unique constellation of ion channels within this physiologically important PNS subregion (16). In this study, we were able to detect both known and novel sodium channel transcripts in DRG that arise by alternative splicing. Quantitative PCR was used to assess the expression patterns of these alternatively spliced isoforms, and we found that alternative splicing of sodium channel mRNAs was most pronounced in DRG. These data indicate that effective treatment of neuropathic pain via antagonism of sodium channels must account for multiple channel isoforms.

MATERIALS AND METHODS
Sodium Channel Expression Compendium (Body Atlas)-Gene expression profiling of cynomolgus monkey mRNA was performed on ink-jet synthesized oligonucleotide microarrays designed to monitor ϳ47,000 human transcripts. This two-array set (Hu50K) is an updated version of human microarrays described previously (17). Probes for the nine voltage-gated sodium channel transcripts reported here were present on these arrays. Additional results from these experiments will be described in future publications. Microarray probe sequences were designed to hybridize near the transcript 3Ј-end. Messenger RNA amplification and hybridization conditions were performed as described previously (17). Individual samples were labeled with either fluorescent Cy3 or Cy5 dye and hybridized to a human microarray in replicate against a mass-balanced control pool of 220 individual tissue RNA samples. Each experiment was repeated with the Cy3 and Cy5 dyes reversed (a dye swap). Microarrays were purchased from Agilent Technologies (Palo Alto, CA). The housing, necropsy, and extraction of RNA from organs and tissues of all animals used in the monkey expression experiments was performed by MPI Research Inc. (Mattawan, MI), with the sole exception of RNA extraction from bone samples. Four cynomolgus (Macaca fascicularis) monkeys, 2 male and 2 female, were the source of the organs and tissues. All animals were matched for age, weight, and diet (Lab Diet® Certified Primate Diet 5048, PMI Nutrition International, Inc.) available ad libitum. Monkeys were fasted for 16 h pre-euthanasia and sedated with ketamine, followed by overdose of sodium pentobarbital solution and exsanguination. All samples were harvested, trimmed, and snap frozen in liquid nitrogen within 30 min with the majority of samples being frozen within 15 min.
Each sample was hybridized to 8 microarrays: 4 individuals and 2 microarrays for the dye-swap. For each gene, a log 10 ratio of individual sample to pool was generated by combining dye-swap microarray pairs. These 4 values, one for each individual, were then averaged in an error-weighted fashion (18) to produce a log 10 error-weighted average (tissue-to-pool) ratio for each gene. Error bars estimate the one standard deviation of this average combine the modeled errors calculated for each sample and the replicate error (18) Finally, for each gene, we transformed the log ratios to ratios and normalized linearly by scaling the largest tissue ratio to 1.

RT-PCR and Quantitative Real Time PCR (TaqMan®)-
Reverse transcription-polymerase chain reaction (RT-PCR) amplification from tissue-specific mRNA or total RNA was performed as described previously (5). The oligonucleotides used in this study (Table I) were obtained from Qiagen (Valencia, CA). Amplicons were subcloned into pCR2.1 using a TOPO-TA cloning kit (Invitrogen). Sequencing was performed by a commercial vendor (Lark Technologies Inc., Houston, TX). The sequences of all isoforms described in this study were deposited into GenBank TM (Table II).
TaqMan® is a registered trademark of Roche Applied Science. Taq-Man® primer probe reagents were obtained through the Applied Biosystems Assays-by-Design custom assay service (Foster City, CA). The primer-probe sets used in this study are shown in Table III. Probe sequences were designed to straddle the unique splice junctions characteristic of each alternative splice form. TaqMan® assays were performed on an ABI 7900 real time PCR instrument in 10-l assays that were run in triplicate in a 384-well format optical PCR plate. The assays were calibrated with isoform-specific RT-PCR clones using the standard curve method. 2 Standard curves generated from plasmid clones were linear across at least six orders of magnitude, and all reported values derived for total tissue RNA fell within the range of these standard curves.
Total RNA from human tissue was obtained from Clontech. Total rat dorsal root ganglia RNA from control and treated animals from a spinal nerve ligation neuropathic pain model was obtained as a gift from Dr. Hao Wang and colleagues (Merck Research Labs, West Point, PA). All of the handling of the animals and testing was performed in accordance with the policies and recommendations of the International Association for the Study of Pain (19) and received approval from the Institutional Animal Care and Use Committee of MRL, West Point, PA. The experimental treatment of the animals was exactly as described in Ref. 20. RNA was converted to cDNA for TaqMan® measurements using a commercially available kit from Applied Biosystems. All assays were normalized on a tissue-to-tissue basis by adding a constant amount of FIG. 1. Sodium channel body atlas from cynomolgus macaque monkey. Microarray measurements were made of sodium channel ␣-subunit transcripts across multiple tissue-derived RNA samples. For each gene, fluorescence intensity values were measured for labeled cRNA hybridized to 3Ј-positioned probe sequences. The tissue sample that generated the highest intensity value was assigned a value of 1, and all remaining samples were normalized on a linear scale. The tissues were sorted by SCN8A expression level, first by PNS, then CNS, and then all remaining tissues. Error bars estimate 1 S.D. of average measurements and are described in more detail under "Materials and Methods." input total RNA into the RT reaction. We chose this normalization method because we were unable to identify a single housekeeping gene that yielded satisfactory normalization data. Isoform levels within tissue RNAs were measured in triplicate on separate occasions, and the results were highly reproducible. A representative data set is shown in each case.
Quantitative PCR values were calculated by assuming that the sum of splicing events at a given site was equal to unity (e.g. SCN8A [exon 6N] ϩ [exon 6A] ϭ 1). This was applied to the most abundant measurement in each data set (SCN8A-adult brain; SCN9A exon 5-DRG; SCN9A exon 11-fetal brain) where the sum of the isoform measurements was adjusted to a value of 100%. All other measurements in the data set were normalized to these maximum values. Error bars are 1 S.D. of the average of triplicate measurements.

RESULTS
Body Atlas Expression Patterns of Sodium Channels-DNA microarrays afford the opportunity for genome-wide monitoring of transcription within any RNA sample. When RNA samples from diverse tissues throughout the body are hybridized, it becomes possible to assemble a transcriptional compendium or Body Atlas (5,21,22). This resource is essentially a semiquantitative, whole-genome Northern blot for all transcripts with corresponding probes on the array. The Body Atlas of the nine voltage-gated sodium channel paralogs across a broad range of tissues was obtained for cynomolgus macaque monkey transcripts profiled on human microarrays (Fig. 1). While entirely consistent with previous studies (7)(8)(9), these data are unique in that they provide a comprehensive overview of the entire gene family in a wide range of tissues. We present the data for monkey because our data for human were incomplete. Moreover, we believe the cynomolgus macaque data are a more accurate reflection of normal human biology because tissue RNA samples were obtained from healthy individuals under carefully controlled conditions. Microarrays using human sequences have routinely and successfully been used to profile non-human primate samples to study gene expression, including the use of samples from cynomolgus monkey (23). There is a high level of conservation of orthologous sodium channel sequences and expression patterns among mammals that is likely even higher among primates. On average, one would expect zero to only a few mismatched bases between monkey and human per 60-mer probe sequence. Of the 10 sodium channel probes that could be mapped to both human and chimpanzee genomic sequences, seven were perfect matches in both species. Finally, although sequence differences between human and cynomolgus monkey transcripts may affect individual probe intensities, they are unlikely to influence intensity ratios (between sample and pool), which are used here. Examination of the Body Atlas data reveals that SCN4A (Na v 1.4) and SCN5A (Na v 1.5) exhibit strikingly selective expression in striated muscle and heart, respectively. SCN3A (Na v 1.3) appears to be transcribed in numerous tissues. Channels SCN1A (Na v 1.1), SCN2A (Na v 1.2) and SCN8A (Na v 1.6) appear to be abundantly expressed in both PNS and CNS tissues. In contrast, SCN9A (Na v 1.7), SCN10A (Na v 1.8) and SCN11A (Na v 1.9) expression is strikingly selective to DRG, with only minor expression levels detected elsewhere in the body.
Detection of Alternative Splicing in SCN8A, SCN9A, and SCN11A-Within the vertebrate sodium channel paralog family, parsimonious clustering by protein sequence indicates that SCN8A and SCN9A occupy one branch of the sodium channel family tree that also includes SCN1A, SCN2A, and SCN3A (8,9). As mentioned, within SCN8A, two alternative splicing events with the potential to produce functional sodium channels have been described. The first involves the potential use of mutually exclusive, alternative exon 6 (coding exon 5) sequences that encode parts of transmembrane segments S3 and S4 within domain I (Fig. 2 and Refs. 14 and 15). The 92 nucleotide (nt) alternative exons, which are found in human, mouse, and rat genomic sequence and known to be used in other sodium channel family members, code for nearly identical amino acid sequences that differ at only two positions. The second, described in mouse and rat, involves the use of alternative 5Ј-splice donor sites in exon 12 (coding exon 11), which encodes a portion of the cytoplasmic loop between domains I and II (Fig. 2 and Ref. 12). The resulting channels differ by 11 amino acid residues, and these isoforms exhibit distinct electrophysiological properties (Table I and Ref. 12). Given conservation of the respective genomic sequences, we hypothesized that these alternative splice events might be expressed in human tissues. Gene-specific amplification primers for SCN8A were used to generate RT-PCR products from human DRG, and sequencing confirmed usage of both mutually exclusive exon 6 cassettes and alternative splice donor sites in exon 12 (Fig. 2). Inspection of genomic sequence of human, mouse and rat SCN9A suggested it has a gene structure similar to SCN8A. Specifically, the genome sequences from all three species encode potentially mutually exclusive exon 5 sequences, and evidence for alternative splicing of the rabbit paralog was deduced from cDNA sequences (24). Similarly, conservation of exon 11 alternative splice donor sites was found in human, mouse and rat genomic sequences, and evidence for alternative splicing was suggested by comparison of human, rat, and rabbit cDNA sequences (12). Confirmation that both of alternative splice events occur in human SCN9A was obtained by sequencing of RT-PCR products amplified from human DRG (Fig. 2).
Cross-species comparative studies using a combined paralog/ ortholog approach also revealed a novel, alternatively spliced isoform of SCN11A. Sodium channel genes SCN5A, SCN10A, and SCN11A share a similar gene structure (25). Murine SCN5A is alternatively spliced in heart tissues, with one transcript that is missing the RefSeq (NM_021544.1) exon 17 (26). The same splicing event is observed in rat transcripts of SCN5A (GenBank TM AF353637). This 159-nt exon codes for 53 amino acids situated in the cytoplasmic loop region between domains II and III. Electrophysiology measurements suggest that the Na v 1.5 channels encoded by these isoforms are functionally similar (26). The human SCN11A gene encodes an analogous 114 nt, 38 amino acid exon in the same cytoplasmic loop-encoding region of the transcript. Moreover, this loop region has only 48% protein sequence identity between human and mouse, whereas the overall channel identity is 72%. Therefore, by analogy to murine SCN5A, this splice variant of SCN11A may encode a functional sodium channel. RT-PCR across this region of human SCN11A using DRG total RNA gave predominantly the expected exon 16 ϩ amplicon, however a clone containing a smaller amplicon was isolated, and sequencing revealed it encoded a splice variant lacking exon 16.
Quantitation of Alternatively Spliced Isoforms-Real-time PCR, wherein exon-specific primers flank a splice junctionspecific, fluorescently labeled probe (commonly referred to as a TaqMan® assay) was used to quantify alternative splicing events. Using custom assays, we monitored the ratios of all of the alternative splicing events shown in Fig. 2 across two PNS (DRG and spinal cord) and two CNS (fetal and adult whole brain) tissue RNA samples (Fig. 3).
Several interesting conclusions can be drawn from these data. First, the tissue-specific differences in the overall expression levels of each channel transcript closely mirror the Body Atlas array measurements shown in Fig. 1. Second, almost all alternative splicing events are well represented, especially in DRG. This is particularly true of SCN9A, where we observed roughly equal abundance of both exon 5 and exon 11 alternative splice variants. The expression levels and splicing patterns observed in spinal cord were distinct from DRG, especially for SCN9A. Third, despite the relatively high abundance of certain splicing isoforms (e.g. SCN8A transcripts that include exon 6A), many of these alternatively spliced transcripts were not found in GenBank TM . Finally, we were unable to reliably detect the SCN11A ⌬exon 16 splice variant, suggesting it is expressed at very low levels in human DRG (data not shown).
Alternative Splicing Events Appear to Be Unlinked-In DRG, SCN8A, and SCN9A transcripts undergo frequent alternative splicing at two sites (exon 6N versus 6A and exon 12RS versus Junction-specific TaqMan® assays were calibrated against control plasmids by the standard curve method, and validated with control reactions against alternative junction sequences. The data were normalized to the highest measurements, which were assigned an arbitrary value of 100. Error bars are the standard deviations of triplicate measurements. Abbreviations are the same as Fig. 2. A, expression of alternative splice forms of SCN8A. B, expression of alternative splice forms of SCN9A. C, expression of SCN11A. The exon 16 skip variant was below the limits of reliable detection. exon 12EXT for SCN8A; exon 5N versus exon 5A and exon 11RS versus exon 11EXT for SCN9A). It is possible that these alternative splicing events are linked, e.g. SCN9A transcripts with exon 5N generally possess exon 11EXT. Conversely, alternative splicing at one site may not influence splicing at a distal site, in which case there should be a stochastic distribution of splicing-generated isoforms. To test this, we amplified cDNA prepared form DRG with PCR primers that span both alternative splice sites for SCN8A and SCN9A (Fig. 4). Individual amplicon clones were then screened at both alternative splice positions to determine the splicing pattern of each clone. We found all possible combinations of splice events, and the distribution of splicing events within the overall set of clones suggested that alternative splicing events were regulated independently with respect to one another. The implication of this finding is that DRG is populated with at least four distinct alternatively spliced isoforms of both SCN8A and SCN9A.
Sodium Channel Expression and Alternative Splicing in a Rat Neuropathic Pain Model-Nerve damage in the periphery can result in chronic neuropathic pain. Effective treatments for this condition may result from a more complete understanding of the biological changes that accompany nerve injury. We monitored isoform-specific changes in expression levels that occur in response to spinal nerve ligation in rat, a commonly used model for neuropathic pain (27). Total RNA from ipsilateral DRG of control-treated and nerve-damaged animals was harvested during the period in which maximal allodynia was observed in the injured animals (2 weeks post-injury). In control animals, the splicing pattern of rat SCN8A appeared similar to human SCN8A (Fig. 5A). In contrast, alternative splicing of rat SCN9A transcripts in control DRG was quantitatively different from human (Fig. 5B). In rat, 80% of the SCN9A transcripts include the exon 11 extension variant, whereas in human, this variant appeared to make up 45% of the overall DRG SCN9A transcript. The significance of this species-specific difference is unclear.
As reported previously, down-regulation of the SCN8A, SCN9A, and SCN11A sodium channel transcripts was observed in ipsilateral DRG in response to neuropathic injury (20, 28 -30). SCN8A expression is reduced to 17% of control (Fig.  5A), SCN9A to 27% (Fig. 5B), and SCN11A to 2.5% (data not shown). We also observed a significant change in the expression pattern of alternatively spliced isoforms. In Fig. 5C, we show the retention of alternative splice forms, which is simply the post-injury value divided by the control value for each individual splicing event. For SCN8A, selective retention of exon 6N was observed. For SCN9A, the exon 11RS-containing transcripts were enriched in relative abundance in response to treatment. FIG. 4. Alternative splicing events in SCN8A and SCN9A transcripts appear to be uncoupled. PCR primers (arrows) were used to generate independent clones of SCN8A (287 clones) or SCN9A (92 clones) from DRG cDNA that span both alternative splice sites. Taq-Man® assays were used to determine the alternative splicing events within each clone. The frequency of each clone is represented as a percentage.
FIG. 5. Quantitative PCR measurements of SCN8A and SCN9A alternative splicing in a rat spinal nerve ligation model. Assays designed against specific rat sequences were used to measure alternative splice variants in DRG total RNA isolated from control and nerveligated animals by the standard curve method. Measurements of exon 5A ϩ exon 5N or exon 11 RS ϩ exon 11 EXT for rat SCN8A or rat SCN9A in untreated DRG were assigned a value of 100. The corresponding measurements made for treated samples were normalized to these values. A, SCN8A expression and alternative splicing in control and Chung-treated rat DRG. The splicing pattern of human SCN8A in DRG is shown in the inset for comparison. The numbering of the rat SCN8A exons is different from human SCN8A owing to the absence of a 5Ј-UTR exon in rat RefSeq sequence. B, SCN9A expression and alternative splicing in the rat neuropathic pain model. The human SCN9A splicing pattern from DRG is shown in the inset. C, retention of alternative splice events in response to neuropathic injury. Values were calculated by dividing the treated values by control values for each individual splicing event.

DISCUSSION
Alternative splicing provides a mechanism to generate functionally diverse protein isoforms from a single genetic locus. As shown here, the SCN8A transcript undergoes both development-specific and tissue-specific splicing in humans. In development, exon 6N is highly expressed in fetal brain tissue, and splicing shifts to almost exclusive use of exon 6A in the adult brain. The extension of SCN8A exon 12 is observed almost exclusively in human DRG, with minor relative expression in spinal cord. The functional significance of these alternative splicing events is not yet clear. Interestingly, SCN9A shares similar gene architecture with SCN8A, with a duplicated exon 5 and alternative splice donor sites in exon 11. However, we find that the expression patterns of this transcript are entirely distinct. The SCN9A exon 5N is preferentially expressed in the PNS and CNS of adult tissues and significant usage of exon 5A was found only in DRG. Hence similar gene structure does not imply conserved patterns of splicing regulation.
The duplication of the exons encoding part of transmembrane helix S3 and all of helix S4 in domain I (exon 6 in SCN8A and exon 5 in SCN9A) is only partially conserved across the voltage-gated sodium channel gene family (Fig. 6). By examining 1) the degree of sequence conservation between human, mouse, and rat (genome.ucsc.edu), 2) human genomic sequence, 3) expressed mRNA and ESTs, and 4) published reports in the literature, this duplication appears in SCN2A (31), SCN3A (32), SCN5A (GenBank TM mRNAs), SCN8A (15), and SCN9A (Belcher et al., 1995 and this report). In human, SCN1A also appears to have a duplicated fifth coding exon, however the syntenic region in mouse and rat is interrupted by a single-base, frameshift mutation. It was not possible to find evidence of this exon duplication in the SCN4A, SCN10, or SCN11A genes. The functional significance of the duplicated exon is unclear. Characterization of alternatively spliced, exon 6N or exon 6A rat SCN2A channels failed to reveal detectable differences (33). On the other hand, conservation of this feature across family members and across species, coupled with clear examples of development-specific and/or tissue specific regulation suggest these alternative exons play an important role that has yet to be identified.
The alternative splice donor sites in coding exon 11 that give rise to an 11 amino acid extension in the cytoplasmic loop between domains I and II are less prevalent in the human sodium channel gene family (Fig. 6). In addition to SCN8A and SCN9A, there is clear evidence from cross-species conservation for this alternative splice event in SCN1A, and alternative splicing of this exon in SCN1A has been observed in rat (12,34). The GenBank TM collection of human mRNAs and ESTs indicate that SCN1A and SCN3A also undergo alternative splicing that deletes amino acids from this intracellular domain (Fig. 6). Characterization of the channels encoded by SCN8A alternatively spliced transcripts has demonstrated that the difference of 11 amino acids in the domain I-domain II cytoplasmic loop influences the inactivation and reactivation properties of the channel (12). It is intriguing that the extension isoforms are most highly expressed in DRG, suggesting a specialized role in the transmission of sensory signals.
Alternative splicing of the heart-specific SCN5A transcript in mouse generates variability in the cytoplasmic loop that connects domains II and III (26). While there is no obvious functional consequence of this 53 amino acid deletion that results from the loss of coding exon 17, this variant is highly expressed in mouse heart and conserved between mouse and rat (GenBank TM accession AF353637). Alternative splicing has not been observed in human SCN5A transcripts. Here we report a comparable exon-drop isoform of SCN11A isolated from FIG. 6. Predicted amino acid sequences of human alternatively spliced ␣-subunit sodium channels. A, gene symbols and synonymous protein channel names are shown in the left hand columns. Human peptide sequences encoded by mutually exclusive coding exon 5 segments are shown as a Clustal W alignment in the right hand column. The peptide sequence encoded by the alternative exon 5 proximal to exon 4 (termed neonatal, N by convention, Ref. 10) is shown on the top line while the sequence encoded by the exon 5 distal to exon 4 (adult -A) is shown on the bottom. Four amino acid residues from the upstream exon 4 and downstream exon 6 are underlined. For SCN4A, SCN10A, and SCN11A, searches using TBLASTN, examination of genomic sequence for open reading frames and cross-species sequence conservation, and evaluation of conserved splice site sequences failed to reveal evidence of an alternative exon. The single exon 5 sequences for these genes are shown. B, human peptide sequences encoded by alternative splice forms of coding exon 11. Cross-species (human, rat, mouse) searches for conserved open reading frames, conserved splice sites, and overall sequence conservation revealed evidence for a potential exon extension in SCN1A, SCN8A, and SCN9A. Alternative splice forms found in GenBank TM for SCN1A and SCN3A are also shown. Five residues encoded by coding exon 12 are underlined. No evidence for an exon 11 extension was found for the remaining genes. DRG, although this isoform appears to be rare. Further studies will determine if this transcript encodes a functional sodium channel.
Dorsal root ganglia are clearly unique with respect to sodium channel expression and alternative splicing. Specific sodium channel genes are expressed in DRG and almost nowhere else in the body. Moreover, alternative splicing within these transcripts generates a potentially diverse set of sodium channel isoforms. Similar DRG-selective alternative splicing of calcium channel CACNA1B transcripts has been reported recently (16). DRG is composed of heterogeneous cell types that were shown to differ in their expression patterns of CACNA1B transcripts. Similar observations were made with SCN9A transcripts in rat DRG (35). In this study, we have treated DRG as a homogeneous tissue and detected a diverse spectrum of alternative splicing. It will be of interest to determine whether unique transcripts are constrained to specific cell subtypes, and more importantly, which sodium channel isoforms contribute most to neuropathic pain. Such channels would be preferred targets of a future class of channel-specific antagonists. Our investigation of a rat neuropathic pain model revealed selective enrichment of SCN8A mRNAs encoding exon 5N and SCN9A transcripts that included exon 11RS, suggesting these isoforms may selectively contribute to neuropathic pain.
Examination of DRG-specific, voltage-gated sodium channels highlight an important theme with respect to basic biology and pharmaceutical compound development. We found a surprising diversity of alternative splice forms in the highest expressing tissue, DRG, which is also a region of therapeutic focus with respect to neuropathic pain. Our search was by no means exhaustive. We did not examine SCN10A transcripts for alternative splicing and are not certain we have identified all of the highly expressed splicing events that occur in SCN8A, SCN9A, or SCN11A. While our current knowledge of the human transcriptome is a powerful resource, we believe that much remains to be discovered about alternative splicing and that a thorough knowledge of these post-transcriptional events will be critical to the development of more effective and specific therapies in the treatment of disease and the maintenance of health.