Voltage-gated Na+ channels play important functional roles in the generation of electrical excitability in most vertebrate and invertebrate species. These channels are members of a superfamily that includes voltage-gated K+, voltage-gated Ca2+ and cyclic-nucleotide-gated channels. There are nine genes encoding voltage-gated Na+ channels in mammals, with a tenth homologous gene that has not been shown to encode a functional channel. Other vertebrate and invertebrate species have a smaller number of Na+ channel genes. The mammalian genes can be classified into five branches in a phylogenetic tree, and they are localized on four chromosomes. Four of the branches representing the four chromosomal locations probably resulted from the chromosomal duplications that led to the four Hox gene clusters. These duplications occurred close to the emergence of the first vertebrates. The fifth branch probably evolved from a separate ancestral Na+ channel gene. There are two branches in the invertebrate tree, although members of only one of those branches have been demonstrated to encode functional voltage-gated Na+ channels. It is possible that the other branch may have diverged, so that its members do not represent true voltage-gated Na+ channels. Vertebrate and invertebrate Na+ channels appear to be derived from a single primordial channel that subsequently evolved independently in the two lineages.
Voltage-gated Na+ channels play a critical role in controlling the electrical excitability of animal cells, being primarily responsible for the rising phase of the action potential. The electrophysiological properties of all voltage-gated Na+ channels are not identical, however, and small differences in these properties can have significant effects on the electrical excitability of the cell (for a review, see Hille, 2001). Na+ channels with different functional or pharmacological properties have been observed in different tissues of the same species and in different species by electrophysiological recording (Mandel, 1992). In addition, a variety of different Na+ channel isoforms have been cloned, functionally expressed and characterized (Goldin, 1999).
The Na+ channel consists of a highly processed α subunit of approximately 260 kDa and contains four homologous domains termed I–IV. Within each domain, there are six transmembrane segments called S1–S6, and a hairpin-like P loop between S5 and S6 that comprises part of the channel pore (Fig. 1). The α subunit is associated with accessory subunits in the tissues of certain species, such as the β subunits in mammals (Catterall, 1993) and the tipE subunits in flies (Feng et al., 1995). The purpose of this review is to examine the phylogenetic relationships among the different voltage-gated Na+ channels that have been identified thus far. The structure and functional characteristics of the isoforms will not be discussed in detail as these topics have recently been reviewed (Catterall, 2000; Goldin, 2001).
For the purpose of this review, the voltage-gated Na+ channels can be considered in three distinct categories. Mammalian Na+ channels represent the largest group of genes that have been identified and studied (Table 1). These isoforms have all been classified as members of a single gene family with nine members. These will be referred to using the nomenclature that was proposed recently (Goldin et al., 2000), which consists of the prefix Nav to indicate the principal permeating ion and the principal physiological regulator, followed by a number that indicates the gene subfamily (currently Nav1 is the only subfamily). The number following the decimal point identifies the specific channel isoform (e.g. Nav1.1). There is a tenth isoform termed Nax that has approximately 50 % sequence identity to the other mammalian Na+ channels. Because none of the Nax channels has been functionally expressed, it is possible that this gene does not encode a voltage-gated Na+ channel, which is why it has not been assigned a number. Representatives of all nine isoforms and Nax have been identified and characterized from humans and rats, with representatives of some of the isoforms from other species. Only full-length sequences of the mammalian Na+ channels have been used in this analysis.
The other two groups of Na+ channel that will be discussed include the non-mammalian vertebrate Na+ channels and the invertebrate Na+ channels (Table 2). There is no systematic nomenclature for these channels, and they have been named in a variety of ways. A consistent nomenclature has been adopted for the purposes of this review. The name consists of the first letters of the genus and species, followed by Nav to indicate the principal permeating ion and the principal physiological regulator, and a number that was either part of the original name or is arbitrary.
There are many fewer representatives of non-mammalian Na+ channel sequences. Only five full-length sequences have been determined for non-mammalian vertebrate Na+ channels, and each of these is from a separate species. Lopreato et al. (2001) have determined partial sequences for six genes in one species, Sternopygus macrurus. This is the first attempt to identify all the Na+ channel genes in a single non-mammalian vertebrate species, so these data are particularly informative with respect to phylogeny. Therefore, these sequences have been included in the current analysis.
More invertebrate Na+ channel genes have been identified, with 13 full-length sequences representing 11 species. However, more than one full-length sequence has been determined for only two invertebrate species, Blattella germanica and Halocynthia roretzi. Two sequences have been determined from Drosophila melanogaster, with one full-length (DmNav1) and one that is almost complete (DmNav2). Blackshaw et al. (1999) have attempted to identify all the Na+ channel genes in an invertebrate species, Hirudo medicinalis. They determined partial sequences for four distinct genes, and these sequences have been included in the analysis.
Phylogeny of Na+ channels
Voltage-gated Na+ channels are members of a superfamily that includes voltage-gated K+ channels, voltage-gated Ca2+ channels and cyclic-nucleotide-gated channels (Jan and Jan, 1992; Hille, 2001). The Na+ and Ca2+ channel members all contain four homologous domains, whereas the K+ and cyclic-nucleotide-gated channels in the superfamily consist of tetramers of single-domain subunits. On the basis of structure alone, it might be expected that Na+ and Ca2+ channels would have evolved from the single-domain channels. This is in fact the most widely accepted scheme, for the reasons reviewed by Hille (1987, 1988). Bacteria, like eukaryotic cells, typically have high intracellular K+ and low intracellular Ca2+ concentrations, but many have no requirement for Na+ or Cl–. Consistent with this lack of necessity for Na+ or Cl–, bacteria do not generally express either Na+ or Cl– channels, whereas they clearly express K+ channels (Milkman, 1994). Therefore, it is likely that an ancestral primordial prokaryotic channel was more similar to the voltage-gated K+ channels (Ranganathan, 1994). In fact, it has been suggested that channels with a cytoplasmic gate, such as the cyclic-nucleotide-gated channels and mechanosensitive channels, may represent the primordial channel from which the voltage-gated channels evolved (Jan and Jan, 1994; Anderson and Greenberg, 2001).
It is likely that the Ca2+ channels evolved from the K+ channels during the evolution of the stem eukaryotes (Hille, 1989). Protozoans generally use Ca2+ as the inward charge carrier, and purely Na+-dependent action potentials are not common until the cnidarians (Hille, 1989; Anderson and Greenberg, 2001). Na+ channels have not been detected in protozoa, algae or higher plants (Hille, 1989). According to this scenario, the Na+ channels evolved early in the metazoan era from an ancestral channel resembling the T-type (Cav3 family) Ca2+ channels, before the separation of diploblasts and triploblasts (Spafford et al., 1999).
The four domains of the Na+ channel have more similarity to the four domains of the Ca2+ channels than to each other, further supporting the argument that Na+ channels evolved after the subunit duplications leading to the Ca2+ channels (Hille, 1989; Strong et al., 1993). On the basis of sequence similarities, Strong et al. (1993) suggested that the original duplication event resulted in a two-domain channel consisting of domains I/III and II/IV, each of which then duplicated to result in the first four-domain Ca2+ channel. In this context, it is interesting to note that no channels consisting of two homologous domains have been observed. In addition, no K+ channels with multiple homologous domains have been observed, suggesting either that there was strong selective pressure against this form of channel or that the selectivity change from K+ to Ca2+ occurred before the gene duplication event (Anderson and Greenberg, 2001). The primordial Na+ channel then evolved from the Ca2+ channels, and this primordial Na+ channel subsequently evolved independently in vertebrates and invertebrates (Fig. 2) (Strong et al., 1993).
Evolution of vertebrate Na+ channels
The mammalian vertebrate voltage-gated Na+ channels represent the best-characterized family in that all the different isoforms have been identified and cloned as full-length channels from two species, human and rat. The isoforms are in five distinct branches on the phylogenetic tree (Fig. 2). This division correlates well with the chromosomal localization of the genes encoding the channels, as described by Plummer et al. (1999). The genes encoding Nav1.1, Nav1.2, Nav1.3, Nav1.7 and Nax are clustered together on chromosome 2 in humans and mice (Table 1), and these channels form the largest branch of the tree. Nav1.1, Nav1.2, Nav1.3 and Nav1.7 share a number of common characteristics, including expression in the nervous system and block by nanomolar concentrations of tetrodotoxin (TTX). Nax is unusual in that this channel has sequence differences in functionally important regions, including fewer charges in the S4 voltage sensors and the absence of the IFM motif that is critical for fast inactivation (Fig. 1) (George et al., 1992a; Goldin, 2001). In addition, this is the only mammalian isoform that has not been functionally expressed in an exogenous system. Consistent with these differences, the gene for Nax has diverged more than any of the other channels in this branch of the tree. This divergence suggests that Nax may have evolved a unique functional role. In support of this hypothesis, mice in which the gene encoding Nax has been knocked out demonstrate functional deficits in salt intake behavior, suggesting that this channel has a physiological function other than as a voltage-gated Na+ channel (Watanabe et al., 2000).
The other mammalian Na+ channel genes that are clustered together are those encoding the Nav1.5, Nav1.8 and Nav1.9 isoforms. These genes are clustered on chromosome 9 in humans and the orthologous region of chromosome 3 in mice (Table 1). These isoforms also share functional characteristics, including the fact that they are all considered ‘TTX-resistant’ Na+ channels, requiring micromolar concentrations of TTX before they are blocked. Although the gene for Nav1.9 is clustered on the chromosome with the other two genes, this channel appears to represent a distinct branch of the tree, and it may have evolved from a distinct ancestral Na+ channel gene (Plummer and Meisler, 1999).
The final two isoforms, Nav1.4 and Nav1.6, each represent a separate branch of the tree, and the genes encoding these channels are located on separate chromosomes. The gene for Nav1.4 is located on chromosome 11 in humans and chromosome 17 in mice, and the gene for Nav1.6 is located on chromosome 15 in humans and chromosome 12 in mice (Table 1). The properties of these two channels are generally similar to those of the channels expressed in the nervous system, including block by nanomolar concentrations of TTX. Nav1.6 is the most closely related channel to the branch containing the nervous system Na+ channels, and this isoform is also expressed in the nervous system. Nav1.4 is the primary channel expressed in adult skeletal muscle, and is unique in this regard.
Plummer et al. (1999) observed that each of the chromosome segments containing the Na+ channel genes represents one of the Hox gene clusters. They suggested that the initial expansion of the Na+ channel genes was associated with the genomic duplications that occurred after the divergence of prevertebrates from invertebrate chordates and resulted in the four mammalian Hox gene clusters. Later tandem duplications of the Na+ channel genes on chromosomes 2 and 9 in humans (chromosomes 2 and 3 in mice) resulted in the two clusters of Na+ channel genes. Nav1.9 was more distantly related to Nav1.5 and Nav1.8 in the analysis of Plummer et al. (1999), and it appears on a separate branch in the current analysis (Fig. 2). Plummer et al. (1999) hypothesized that the ancestral chromosome might have had two Na+ channel genes, one that represented the ancestor of Nav1.9 and a second that represented the ancestor of the other mammalian Na+ channels. This model assumes that the ancestor of Nav1.9 was subsequently lost from three of the four chromosome segments.
The analysis of the non-mammalian vertebrate Na+ channels that have been characterized thus far is consistent with that of the mammalian channels and provides further information about the timing of the duplication events. The most informative data were obtained by Lopreato et al. (2001), who cloned partial sequences for six separate isoforms from the teleost fish Sternopygus macrurus. These sequences probably represent all the Na+ channels in that species. As pointed out by Lopreato et al. (2001), these isoforms align well with the branches of the mammalian channels (Fig. 2). SmNav3 and SmNav4 are in the branch with Nav1.1, SmNav5 aligns with Nav1.6, and SmNav1 and SmNav6 are in the branch with Nav1.4. SmNav2 forms a distinct branch of its own, although it is more closely related to Nav1.5 and Nav1.8 than to the other Na+ channel isoforms. SmNav2 was located on a branch with Nav1.5, Nav1.8 and Nav1.9 in the analysis of Lopreato et al. (2001). This branching pattern is not as robust as that for the other vertebrate channels though, because it is based on partial sequence data. Lopreato et al. (2001) suggested that the initial Na+ channel duplication into four genes occurred early in vertebrate history close to the emergence of the first vertebrates, so that a common ancestor of mammals and teleost fish already had four distinct genes. This hypothesis is consistent with the conclusions of Plummer et al. (1999). The duplications occurred as an initial event leading to two pairs of Hox genes, each of which then duplicated to result in the four clusters that are now present. These initial chromosomal duplications were then followed by tandem duplications, as suggested by Plummer et al. (1999), and these duplications were independent in teleosts and mammals. The relationships between the mammalian and teleost genes shown in Fig. 2 are supported by the corresponding tissue expression patterns of the mammalian and teleost channels encoded within each cluster (Lopreato et al., 2001).
An additional means of generating diversity in Na+ channels is through post-transcriptional processing, including alternative splicing and RNA editing. Alternative splicing has been demonstrated for five isoforms present in the mammalian nervous system, Nav1.1, Nav1.2, Nav1.3, Nav1.6 and Nav1.7 (Ahmed et al., 1990; Sarao et al., 1991; Schaller et al., 1992; Gustafson et al., 1993; Belcher et al., 1995; Plummer et al., 1997, 1998; Dietrich et al., 1998; Oh and Waxman, 1998). The proportion of differentially spliced transcripts depends on various factors, including age of development, the tissue of origin and the presence of modulatory agents such as dibutyryl cyclic AMP (Gustafson et al., 1993; Plummer et al., 1997; Dietrich et al., 1998; Oh and Waxman, 1998). However, electrophysiological differences resulting from alternative splicing have only been demonstrated for two alternatively spliced forms of Nav1.6 (Dietrich et al., 1998), so it is not clear how much functional diversity in mammalian Na+ channels results from alternative splicing. No alternative splicing has been demonstrated for non-mammalian vertebrate Na+ channels, and no RNA editing has been demonstrated for any vertebrate Na+ channels. This is in contrast to the situation with the invertebrate Na+ channels, which will be discussed below.
Evolution of invertebrate Na+ channels
The data concerning the invertebrate Na+ channels are much less complete than those for mammalian Na+ channels in that more than one full-length sequence has been determined from only two species, Blattella germanica and Halocynthia roretzi. A full-length sequence and a collection of sequences comprising most of the coding region have been cloned from Drosophila melanogaster, and it is likely that these represent the only Na+ channel genes in that species (Littleton and Ganetzky, 2000). Blackshaw et al. (1999) have made the most extensive effort to identify all the Na+ channel genes in an invertebrate species, having isolated four partial clones from Hirudo medicinalis. The data for these four species suggest that there are two branches of the Na+ channel tree for invertebrates (Fig. 2), although HrNav1 is somewhat divergent from the first branch. The existence of two invertebrate Na+ channel branches was previously suggested by Spafford et al. (1999). The two invertebrate branches are independent of the branching that occurred during vertebrate evolution (Fig. 2). This result is consistent with the conclusion of Plummer et al. (1999) that the vertebrate gene duplications occurred after the divergence of prevertebrates from invertebrate chordates. The four genes from Hirudo medicinalis align with two in each of the branches, suggesting that two additional gene duplication events occurred later during the evolution of this species.
It is difficult to evaluate the functional significance of the two branches of invertebrate Na+ channels. Functional expression has been reported for only three channels, DmNav1 (Feng et al., 1995), MdNav1 (Smith et al., 1997) and BgNav1 (Tan et al., 2001; Liu et al., 2001b), and all these channels are in the same branch of the phylogenetic tree. However, K. Dong and colleagues have preliminary evidence for functional expression of the BgNav2 channel (K. Dong, personal communication). They suggest that the properties of this channel are so unique that it is questionable whether it should be considered a voltage-gated Na+ channel. If BgNav2 does not represent a true voltage-gated Na+ channel, it would suggest that there is only one branch of voltage-gated Na+ channels in invertebrates and that the second branch has diverged to the point that these channels carry out unique functions. If this is the case, then the functional diversity of Na+ channels in many, if not most, invertebrate species would have to result from post-transcriptional regulatory events, as suggested by Liu et al. (2001a), unless additional Na+ channel genes remain to be discovered in these species.
The two major post-transcriptional processes that would increase Na+ channel diversity are alternative splicing and RNA editing. Alternative splicing of Na+ channels has been shown to occur in two invertebrate species, Drosophila melanogaster (Loughney et al., 1989; Thackeray and Ganetzky, 1995) and Blattella germanica (Liu et al., 2001a). Spafford et al. (1999) pointed out that 85 % of the intron splice junctions in a jellyfish Na+ channel (PpNav1) are also found in mammalian Na+ channels with similar locations, suggesting that the positions of the introns within the coding regions have been retained from a common ancestor. However, no alternatively spliced variants were detected in PpNav1, suggesting that alternative splicing was not a primordial means of generating diversity in Na+ channels. RNA editing has been demonstrated for some invertebrate Na+ channels, including DmNav1 from Drosophila melanogaster (Hanrahan et al., 2000; Reenan, 2001). Hanrahan et al. (2000) observed that two of the three characterized RNA editing sites in Drosophila melanogaster are also conserved in Drosophila virilis, suggesting that this process has been maintained throughout the 61–65 million years of divergence between these two species.
Considering the fact that invertebrate species appear to have only 2–4 Na+ channel genes, in contrast to the nine mammalian genes, it is likely that invertebrates and vertebrates use fundamentally different means of generating diversity in Na+ channel function. The major means of generating diversity in invertebrate Na+ channels may involve alternative splicing and RNA editing, whereas vertebrate Na+ channel diversity may result primarily from the presence of multiple genes.
Voltage-gated Na+ channels are encoded by a family of genes that have been highly conserved throughout evolution, which undoubtedly reflects the critical functional role of these proteins in regulating electrical excitability. Because the majority of Na+ channels have been cloned from mammalian species, there is most information about the diversity of mammalian Na+ channels. The recent identification of additional Na+ channels in non-mammalian vertebrates suggests that many of the duplication events leading to the different isoforms occurred early in vertebrate evolution. There is much less definitive information about the invertebrate Na+ channels. Although a significant number of genes have been identified and sequenced, the limited amount of functional information makes it difficult to determine whether those genes encode true voltage-gated Na+ channels. Future efforts to express the invertebrate channels should help to more clearly define the roles of Na+ channel genes in these species.
I thank Drs Miriam Meisler and Ke Dong for critical reading of the manuscript, Dr George Gutman for help with the phylogenetic analysis, and Drs Ke Dong, Bob Maue, Greg Lopreato and Harold Zakon for supplying sequence data prior to publication. Work in the author’s laboratory is supported by grants from the NIH and AHA.