REVIEW EVOLUTIONARY LINK BETWEEN PROKARYOTIC AND EUKARYOTIC K + CHANNELS

Transcription

1 The Journal of Experimental Biology 201, (1998) Printed in Great Britain The Company of Biologists Limited 1998 JEB REVIEW EVOLUTIONARY LINK BETWEEN PROKARYOTIC AND EUKARYOTIC K + CHANNELS CHRISTIAN DERST 1 AND ANDREAS KARSCHIN 2, * 1 Institute for Normal and Pathological Physiology, University of Marburg, Marburg, Germany and 2 Max- Planck-Institute for Biophysical Chemistry, Molecular Neurobiology of Signal Transduction, Am Fassberg 11, Göttingen, Germany *Author for correspondence ( akarsch@gwdg.de) Accepted 30 July; published on WWW 22 September 1998 Considering the importance of K + channels in controlling the crucial K + gradient across the plasma membranes of all living cells, it comes as no surprise that, besides being present in every eukaryotic cell, these integral membrane proteins have recently also been identified in prokaryotes. Today, approximately a dozen successfully completed and many more ongoing sequencing projects permit a search for genes related to K + channels in the genomes of both eubacteria and archaea. The coding regions of homologues show a remarkable variety in primary structure. They predict membrane proteins with one, two, three and six hydrophobic segments surrounding a putative K + -selective pore (H5) and the presence or absence of a cytosolic putative NAD + -binding domain Summary (PNBD) that probably senses the reducing power of the cell. The analysis of kinships on the basis of phylogenetic algorithms identifies sequences closely related to eukaryotic voltage-dependent Kv channels, but also defines members of a primordial class of prokaryotic K + channel (containing the 2TMS/PNBD motif). Considering the unique mechanisms that may account for the assembly of modern proteins from different ancestral genes, and with more primary sequence data soon to appear, a scheme for the evolutionary origin of K + channels comes within reach. Key words: K + channel, membrane protein, evolution, NAD-binding domain, K + transport system, archaea, bacteria. Introduction Eukaryotic K + channels have been isolated from vertebrates, arthropods, molluscs, nematodes, protozoa, fungi, yeast and higher plants. They are involved in many cellular processes such as osmoregulation, K + homeostasis, secretory processes, signal transduction and membrane excitability. Channel proteins contain four (or two dimeric) subunits that differ significantly in structure. In voltage-gated (Kv) channels, they are composed of six hydrophobic, probably membranespanning, α-helices (Baumann et al. 1987; Kamb et al. 1987; Papazian et al. 1987). In inwardly rectifying (Kir) channels, they consist of only two transmembrane segments (Ho et al. 1993; Kubo et al. 1993), and in the recently discovered tandem-repeat or duplicate-pore channels they contain either four (2+2; Fink et al. 1996; Goldstein et al. 1996; Lesage et al. 1996a) or eight segments, as in the yeast channel TOK (6+2; Ketchum et al. 1995; Zhou et al. 1995; Lesage et al. 1996b). Common to all K + channels, however, is the presence of a conserved stretch of amino acid residues around a T-X(4)- G-[Y/F/L]-G motif that is thought to line the K + -selective ionconduction pathway H5 (P region; Doyle et al. 1998). The fourfold arrangement of the Kv channel theme (six transmembrane segments) is present in each subunit of voltagedependent Na + and Ca 2+ channels. It is plausible, therefore, that they may have evolved secondary to K + channels by two events of tandem gene duplication (Ranganathan, 1994; Hille, 1996). K + transport across membranes and the maintenance of ionic homeostasis play an essential role not only in eukaryotes but also in prokaryotes, i.e. eubacteria and archaea. Various transport systems have been described genetically and molecularly. Complexes consisting of integral membrane proteins and variably cytosolic components, such as Trk (Schlösser et al. 1993, 1995), Kdp (Hesse et al. 1994), Kup (Dorsch et al. 1991; Schleyer and Bakker, 1993) or Kef (Munro et al. 1991), mediate net uptake or efflux of K + and maintain the ion gradient between the cytosol and the surrounding environment (Jan and Jan, 1997). Some complexes contain motifs in their hydrophobic segments with

2 2792 C. DERST AND A. KARSCHIN limited resemblance to the core region (S5 H5 S6) or to the segment that confers voltage-sensitivity (S4) in Kv channels (Jan and Jan, 1994). Yet, from their moderate homology to eukaryotic K + channels, it is more than just speculation that they function as true ion channels, defined by the rate at which ions pass through the water-filled pore. Two other proteins with a stretch of amino acid residues homologous to the H5 region have been reported, one from Escherichia coli (Kch; Milkman, 1994), the other from the gram-positive soil bacterium Streptomyces lividans (KcsA; Schrempf et al. 1995). KcsA (here termed slikcha) exhibits a channel architecture with two hydrophobic regions surrounding a pore that is indeed similar to eukaryotic K + channels (MacKinnon et al. 1998). The 0.32 nm (3.2 Å) X-ray resolution structure of slikcha, described recently by Doyle et al. (1998), now provides new insight into the structure of K + channels. Kch from E. coli (here termed ecokcha) displays a primary structure different from that of other known K + channels. It is composed of an N- terminal part resembling that of Kv channels (i.e. six membrane-spanning segments surrounding an H5 region; Johansson and von Heijne, 1996) and a putatively cytoplasmic C-terminal part homologous to two repeats in the K + transport system subunit TrkA and the cytosolic domains of KefB and KefC (see Fig. 1; Milkman, 1994; Jan and Jan, 1997). This region may be involved in the glutathione-dependent gating of KefC (Miller et al. 1997) and in TrkA has NAD + -binding capacity, possibly conveying channel regulation by the cellular redox state (Schlösser et al. 1993). Owing to limited homology to the NAD + -binding site in bacterial dehydrogenases (Rossman fold), we have used the term putative NAD + - binding domain (PNBD) for this structure in the remainder of this article. Little further information is available on the presence and role of prokaryotic pore proteins with high ion throughput rates equivalent to eukaryotic K + channels. The plethora of prokaryotic sequence information that accumulates from multiple genome sequencing projects allows a search to be made of these databases for more K + channel homologues. Together with the functional characterization of the gene products, this may help in deciphering structural themes inherent to prokaryotes and possibly in revealing hints about the origin of eukaryotic K + channels. Prokaryotic K + channel structures lacking a PNBD A current homology search with ecokcha and eukaryotic K + channel sequences in GenBank and several databases of unfinished (and unpublished) sequences of prokaryotic genomes (T.I.G.R., The Institute for Genomic Research, personal communication) identifies a host of novel prokaryotic genes related to K + channels (for details, see Table 1). These encode integral membrane proteins with a tertiary structure consisting of one, two, three or six transmembrane segments (TMSs) adjacent to or surrounding a pore region and an optional PNBD (Fig. 1). It is interesting to note that complete open reading frames of K + channel Membrane topology of prokaryotic K + channels: 6TMS/PNBD 2TMS/PNBD 1TMS 2TMS 3TMS 6TMS mjakchab operon in Methanococcus jannaschii: homologues are found in several representatives of eubacteria and archaea (for further information, see but are completely absent in other species, e.g. Haemophilus influenzae (Fleischmann et al. 1995), Mycoplasma genitalium (Fraser et al. 1995), Mycoplasma pneumoniae (Himmelreich et al. 1996) and Borrelia burgdorferi (Fraser et al. 1997). Although other K + -transport systems (Kef and Trk) may substitute functionally, e.g. in H. influenzae, no known K + -transport systems are present in the small genomes of Mycoplasma spp. and B. burgdorferi. The four structures with six TMSs (lacking a C-terminal PNBD) identified in Methanococcus jannaschii (mjakcha; Bult et al. 1996), Methanobacterium thermoautotrophicum (mthkcha; Smith et al. 1997), Aquifex aeolicus (aqakcha; Deckert et al. 1998) and Deinococcus radiodurans (drakchb; T.I.G.R.) topologically resemble the eukaryotic motif of Kv * mjakcha mjakchb K + channel TrkA TMS Pore/H5 PNBD Fig. 1. Putative membrane topology of prokaryotic K + channels (A) and illustration of the mjakchab operon in the archaeon Methanococcus jannaschii (B). Hydrophobicity blots were made using PEPTIDESTRUCTURE according to the algorithm of Kyte and Doolittle (1982). Database searches were performed with the GCG program in the Heidelberg Unix Sequence Analysis Resources (HUSAR) using BLASTN or TBLASTN algorithms (Altschul et al. 1990) on GenBank (release 105) or The Institute of Genomic Research databases ( For details, see text. TMS, transmembrane segment; PNBD, putative NAD + -binding domain. A B

3 Evolutionary link between prokaryotic and eukaryotic K + channels 2793 channels. Indeed, a TBLASTN homology search for the topscoring sequences in other organisms reveals that their sequence is more similar to eukaryotic than to other prokaryotic sequences. A graphic profile of the level of identity of mjakcha illustrates the similarity to both eukaryotic Kv8.1 channels (black line) and ecokcha channels (red line) (Fig. 2A, alignment available on the internet). It is evident that similarity to ecokcha is only greater within the first half of the pore, but not in the rest of the protein. However, the first four hydrophobic segments and the positioning of a basic Similarity value mjakcha Kv8.1 mjakcha ecokcha Residues in mjakcha sequence 4 3 A B bsukcha mjakchb bsukcha Kv2.2 residue at every third position in the voltage-sensor S4 of mjakcha suffice to demonstrate strong homology to Kv channels and suggest that both families share a closely related common ancestor. The primary structures of two other sequences from Streptomyces lividans (slikcha; Schrempf et al. 1995) and Bacillus subtilis (bsukcha; Kunst et al. 1997), although significantly different in length, resemble eukaryotic Kir channels: they are composed of two TMSs and lack a PNBD. A TBLASTN search with bsukcha and an equivalent alignment shows that it is indeed similar to eukaryotic channels, but of the Kv type. Of all known sequences, bsukcha demonstrates the highest homology throughout the coding region to Kv channels (e.g. Kv2.2; Fig. 2B). In the C- terminal region, however, it is also similar to the C-terminal PNBD of 2TMS/PNBD structures, such as mjakchb (see below). Despite this similarity, it remains speculative whether the C terminus of bsukcha represents an evolutionary intermediate between the C-terminal structures of eukaryotic Kv channels and prokaryotic PNBD motifs. A membrane topology unknown from eukaryotic K + channels is predicted from several other novel sequences that translate into integral proteins with uneven numbers of TMSs. For example, three putative TMSs arise from the highly identical sequences bcakcha from Bacillus caldotenax and bstkcha from Bacillus stearothermophilus (Barstow et al. 1987), and a unique structure with only one N-terminal hydrophobic region, a pore region and a short hydrophilic N- terminal segment, is found in the archaeon Archaeoglobus fulgidus (afukchb; Klenk et al. 1997). They may simply represent truncated pseudogenes, but it is possible that they assemble with other structures to form heteromeric macromolecules Residues in bsukcha sequence Fig. 2. Comparison of homologies between mjakcha and ecokcha and Kv8.1 (A) and between bsukcha and mjakchb and Kv2.2 (see Table 1 for channel abbreviations) (B). The diagram displays the output from comparisons of two sequences using the GAP algorithm (Needlemen and Wunsch, 1970). Homology scores between sequences inside an 11 amino acid window were rated as 4 (identity), 3 (strong), 2 (moderate), 1 (low) or 0 (no homology or gaps); homology is scored according to a normalized Dayhoff matrix (Gribskov and Burgess, 1986). Note that the putative 110 residue N- terminal domain of bsukcha (yugo) was omitted, since it represents an additional open reading frame accidentally merged with bsukcha by a sequencing error. Two transmembrane segments plus a PNBD is there an ancient structural motif in prokaryotes? All other sequences retrieved from the database search with ecokcha contain a core region built from two transmembrane segments and linked to a C-terminal PNBD (2TMS/PNBD). These genes are currently found in the genomes of Helicobacter pylori (Tomb et al. 1997), Synechocystis sp. (Kaneko et al. 1996), Mycobacterium tuberculosis (Cole et al. 1998), Deinococcus radiodurans (T.I.G.R.), Thermatoga maritima (T.I.G.R.), Methanococcus jannaschii (Bult et al. 1996), Methanobacterium thermoautotrophicum (Smith et al. 1997) and Archaeoglobus fulgidus (Klenk et al. 1997). Their coding regions vary in length from 314 to 378 amino acid residues owing to significant variations in the less-conserved terminal domains. There are some indications to suggest that they are purely prokaryotic developments and possibly close to the origin of K + channels. First, applying algorithms for sequence comparison, they appear in an independent phylogenetic cluster (Fig. 3A) with no equivalent structure in eukaryotes. Second, sequence variations occasionally occur both in the cytosolic PNBD and in the highly conserved

4 2794 C. DERST AND A. KARSCHIN hydrophobic pore region. A rare tyrosine-to-phenylalanine exchange in the unique pore signature GYG, for example, is present in syskcha of the cyanobacterium Synechocystis sp. and in hpykcha of the gram-negative ulcer-causing bacterium Helicobacter pylori. Also, a glycine is altered to a serine (GYS) in tmakcha of the extreme thermophilic eubacterium Thermatoga maritima, which may be of crucial functional consequence. A mutation of the first residue in this signature A slo afukchc ecokcha Kir3.2 Kir3.4 Kir3.3 Kir3.1 Kir6.1 Kir6.2 Kir2.1 Kir2.2 Kir2.3 Kir5.1 Kir1.1 Kir1.2 Kir family (2TMS) eag herg elk TREK_2 afukchb eag family mjakchb mjakchc bcakcha drakcha syskchb afukcha syskcha tmakcha mthkchb hpykcha 2TMS/PNBD Kch family Fig. 3. (A) Phylogenetic tree of prokaryotic and eukaryotic K + channel genes. Prokaryotic sequences are shown in red, eukaryotic sequences in blue. For each sequence, a 101 amino acid window around the conserved T- X-G-X-G signature in the H5 region was isolated, ensuring that only equivalent regions within the S5 H5 S6 region were used for computation. Following a multiple CLUSTAL alignment (Thompson et al. 1994), the standard CLUSTREE algorithm (Saitou and Nei, 1987) was used for calculation. The Gapc parameter was set to 20 to avoid extensive gap formation. Red dots indicate misplaced sequences; blue dots represent prokaryotic sequences without structural homologues. For channel abbreviations, see Table 1. (B) Similarities in the sequence window (see text) between the prokaryotic 6TMS and 2TMS/PNBD families and other members of the different K + channel families. TMS, transmembrane segment; PNBD, putative NAD + -binding domain; Kir, inwardly rectifying K + channel; eag, ether-à-gogo; Kch, K + channel; Kv, voltage-sensitive K + channel. B 6TMS family 50 56% 35 39% 36 48% 44 49% 2TMS/PNBD family Kv family Kir family Two-pore channels 34 44% 30 39% 36 52% 40 51% Kv3.1 Kv3.2 Kv3.3 shaw Kv1.1 Kv1.2 Kv1.3 Kv1.6 Kv1.4 Kv1.5 shaker Kv4.2 Kv4.3 Kv4.1 shal Kv5.1 Kv2.1 Kv2.2 shab Kv6.1 Kv8.1 Kv7.1 bsukcha mthkcha drakchb slikcha vcokcha KvLQT1 mjakcha Kv family (6TMS) ork_1 TREK_1 CeK_C24A3_1 CeK_F21C3_1 TWIK_1 TOK_2 TWIK_2 CeK_F21C3_2 ork_2 CeK_C24A3_2 TOK_1 mtukcha 6TMS family 2TMS/PNBD family 40 51% 38 52% 6TMS/2TMS Kch family Two-pore K + channels (2+2/6+2 TMS)

5 Evolutionary link between prokaryotic and eukaryotic K + channels 2795 (SYG) in inwardly rectifying Kir3.2 (GIRK2) subunits from mutant ataxic weaver mice causes loss of K + selectivity (Patil et al. 1995). Pores that lose the ability to discriminate between K + and Na + have also been engineered in other K + channels by mutations in the same motif (Heginbotham et al. 1994). It may be hypothesized that these and other uncommon exchanges of residues in the pore of 2TMS/PNBD channels (e.g. the absence of a proline at the fourth position after GYG) date back to the most ancestral K + (cation)-selective channels. Third, there is an intriguing functional resemblance between the C-terminal cytosolic PNBD/TrkA and auxiliary (β-) subunits of Kv and Kir channels in that both are controlled by the cellular redox state and thus represent a conserved machinery for channel regulation (McCormack and McCormack, 1994; Jan and Jan, 1997). However, overall sequence homology between these structures is absent, and PNBDs themselves may therefore be regarded as prokaryotic developments. Sequence variations nevertheless occur even in the conserved G-X-G-X(2)-G motif of the Rossman fold. None of the glycine residues is maintained in hpykcha, for example, although the rest of the sequence is clearly homologous to a PNBD. These sequence variations may be relevant for the binding of PNBDs to different substrates, for example NAD +, glutathione or nucleic acids (Schlösser et al. 1993; Nagy and Rigby, 1995; Miller et al. 1997). Are K + channels assembled from different genes? Tetrameric K + channel proteins in eukaryotes are often composed of different pore-forming subunits and variably associate with β-subunits or ATP-binding cassette (ABC) proteins. It should, however, also be considered that the modern K + channel subunit itself may be a functional mosaic that arose from different ancestral genes (Gilbert, 1985). These may encode channel-like or non-related proteins and may be located close together in the genome or on different chromosomes. The genes of several prokaryotic channels are indeed genomically coupled to neighbouring known and unknown open reading frames with sometimes overlapping sequences. Thus, as is typical of prokaryotic genes, coexpression of polycistronic transcripts may occur. The stop codon of the gene for syskchb, for example, overlaps with the start codon of the gene (amt1) for an ammonium transporter. Similarly, the gene for afukchb is located in close vicinity to napa-2, which encodes a Na + /H + antiporter with homology to KefC (Reizer et al. 1992). Following concerted transcription or transcriptional regulation, transport processes across the membrane may be functionally coupled. Alternatively, in the A. fulgidus genome, the close vicinity of the gene for afukchc to the N5,N10-methylenetetrahydromethanopterin dehydrogenase gene (mtd) may implicate coupling to metabolic enzymes. In the genome of the methanogenic and strictly anaerobic Methanococcus janaschii, the first completely sequenced genome from an archaeon, 2TMS/PNBD members are found. One of them, mjakchb, is genomically coupled in one operon to mjakcha, another K + channel homologue with six hydrophobic segments and close similarity to eukaryotic Kv channels (Fig. 1). The 209 amino acid residues of mjakcha are directly followed by the first of the 333 amino acid residues of mjakchb. Although there is no experimental evidence at present, a physical association between the two gene products, if stoichiometrically transcribed, is not unlikely. Thus, the ancestral genome of M. janaschii may provide insight into the evolutionary origin of certain K + channel subunits. Phylogenetic clusters It remains a challenge to infer the ancestry of K + channels from the few modern examples that have survived several hundred million years of evolution. So far, comparison of primary sequences offers the most meaningful method for revealing the kinships between K + channels from different domains of life. Hence, we have constructed a phylogenetic tree from the majority of known sequences by means of the CLUSTREE algorithm using a variable-sequence window (Fig. 3). By selecting a region (e.g. of 101 amino acid residues) centred on the conserved G-[Y/F]-G signature in the pore, only comparable core regions including the channel s pore and neighbouring hydrophobic segments were employed for computation. We point out that the analysis of partial sequences emphasizes the relationships between genes and phylogenetic clusters, rather than reflecting the true evolutionary distances to the nearest ancestor. Irrespective of the window size, this analysis correctly groups most eukaryotic Kir, Kv, ether-à-gogo (eag)-like and tandem-pore channels in the phylogenetic cluster that matches classification according to functional channel properties. This indicates that, after divergence from the prokaryotes 1.4 billion years ago, these Na + /Ca 2+ channels Kv? 6TMS Loss of PNBD 6TMS/PNBD Kir? Loss/gain of 4TMS Loss/gain of 4TMS Tandem-pore K + channels?? prokaryotic K + transporter KefC/TrkH TOK (yeast) 6+2TMS Eukaryotic Gene fusion 2TMS mjakchab operon Loss of 6TMS + 2TMS/PNBD PNBD 2TMS/PNBD Prokaryotic Fig. 4. Scheme of the possible evolutionary relationships between eukaryotic and prokaryotic K + channels. The diagram permits evolutionary progress along the red or green arrows. A gene fusion event originating from the mjakchab operon is illustrated as an alternative origin of yeast TOK channels. TMS, transmembrane segment; PNBD, putative NAD + -binding domain; Kv, voltagesensitive K + channel; Kir, inwardly rectifying K + channel.

6 2796 C. DERST AND A. KARSCHIN Table. 1. Summary of prokaryotic K + channels Genome Number of Organism (genome size) Name ORF/Acc No. position amino acids TMS PNBD Pore motif Archaeoglobus fulgidus afukcha AF1673/AE ITTTGYGEVKP (2.18 Mb) afukchb AF1246/AE TTTVGYGDVTP afukchc AF0715+AF0716/ AE * 2 + ITTVGYGDIYF ( ) Aquifex aeolicus aqakcha aq-1863/ae ATTVGYGDITP (1.50 Mb) Bacillus caldotenax bcakcha lctb/x LLSVGYGDVTP Bacillus stearothermophilus bstkcha lctb/x LLSVGYGDVTP Bacillus subtilis bsukcha yugo-b /Z VSTVGYGDYVP (4.20 Mb) Borrelia burgdorferi (1.44 Mb) Deinococcus radiodurans drakcha LTTVGFGEVHP (3.00 Mb) drakchb VTTVGYGDISP Escherichia coli ecokcha ECOKCH/AE MSTVGYGDIVP (4.60 Mb) Haemophilus influenzae (1.83 Mb) Helicobacter pylori hpykcha HP0490/AE MTATGFGALNE (1.66Mb) Methanococcus jannaschii mjakcha MJ0139/U ITTVGYGDITP (1.66 Mb) mjakchb MJ0138.1/U ISTVGYGDYTP mjakchc MJ1357/U ITTTGYGDFTP Methanobacterium mthkcha MTH505AE ITTVGYGDIVP thermoautotrophicum mthkchb MTH1520/AE IATVGYGDYSP (1.75 Mb) Mycobacterium tuberculosis mtukcha Rv3200c/AL LSTTGYGDITP (4.40 Mb) Mycoplasma genitalium (0.58 Mb) Mycoplasma pneumoniae (0.81 Mb) Streptomyces lividans slikcha Skc1/Z ATTVGYGDLYP Synechocystis sp. syskcha sll0993/d VFGVGYGEVRP (3.57 Mb) syskchb sll0536/d LATVGFGETHP Thermotoga maritima tmakcha > VSTVGYSIPEN (1.80 Mb) Vibrio cholerae vcokcha >167 2? ITTVGYGDMYP (2.50 Mb) Organism (genome size in parentheses), reference name of channel, open reading frame (ORF) title, database accession number (Acc no.), genome position, number of amino acids in the ORF, number of transmembrane segments (TMS), presence or absence (+/ ) of a putative NAD + -binding domain (PNBD) and the pore motif are listed for the channels discussed in the text. Note that four species do not contain any K + channel homologue. *Two ORFs have been merged; a possible frameshift in the original sequence. The original sequence contains a frameshift that accidentally merged two open reading frames; only yugo-b codes for a K + channel homologue, the putative N-terminal domain (yugo-a) has been omitted. Unfinished sequences (The Institute for Genomic Research).

7 Evolutionary link between prokaryotic and eukaryotic K + channels 2797 genes have deviated substantially from those in other families and thus lack close relatives within the prokaryotic domain. Most importantly, the prokaryotic 2TMS/PNBD motifs (devoid of closely related eukaryotic sequences) cluster independently and thus are represented as a separate K + channel subfamily. Also, as outlined above, prokaryotic channels with 6TMS/2TMS topology that lack a PNBD appear in one phylogenetic group with eukaryotic Kv channels. Although their exact location in the phylogenetic tree may change with window size, and some members, e.g. mtukcha, appear to be misplaced, they always cluster with Kv channels. This is possibly due to the high degree of overall diversity and low degree of conservance in prokaryotic channels. Remarkably, ecokcha is not closely related to any of the main groups within the K + channel superfamily or to other cation channels (Milkman, 1994). Comparison of the first four transmembrane segments shows no apparent homology to Kv channels, but sequence similarity is restricted to the core region (S5 H5 S6) (Fig. 2A). A voltage-sensing S4 element and other functional gate structures are missing and may have been acquired following continued selection pressure during evolution (Milkman, 1994; Ranganathan, 1994). Evolutionary scheme Considering these data, the relationships between prokaryotic K + channels and their eukaryotic derivatives may follow the scheme proposed in Fig. 4. Little sequence identity can be found between 2TMS/PNBD channels and eukaryotic K + channels. The way in which evolution proceeded cannot be deduced unequivocally from the limited prokaryotic sequence data available. It seems clear, however, that these genes were not imported from eukaryotes through horizontal gene transfer: the typical codon usage, the sequence variations in different bacterial strains, and the existence of members common to both eubacteria and archaea instead date back to a primordial K + channel precursor. Whether 2TMS/PNBD channels have lost four transmembrane domains or whether 6TMS/PNBD structures (ecokcha) have gained four helices is unknown. The high level of homology between prokaryotic 2TMS/6TMS channels and eukaryotic Kv channels favours the view that exchange of four transmembrane domains has occurred more than once during evolution (e.g. also in the evolution of Kir channels) and that the smaller proteins do not necessarily reflect the simpler design of the ancestral channel. Channels lacking a PNBD nevertheless appear to have evolved closer to the transition to eukaryotic Kv channels than channels with a PNBD, arguing for the loss of this domain during prokaryotic evolution. The evolutionary origin of eukaryotic tandem-pore channels is still elusive (Salkoff and Jegla, 1995). It is only of limited use to trace the evolutionary history of K + channels by deducing relationships simply from the tertiary structure of proteins. On the basis of the hallmark of prokaryotes to host gene clusters organized in operons, it is tempting to speculate that an operon similar to mjakchab is ancestral to the gene for the yeast tandem-pore channel TOK. Replacement of the first stop codon by a coding triplet would give rise to the 6+2 topology known from TOK channels. Such an evolutionary process cannot be deduced directly from the modest sequence similarity between these proteins (e.g. comparing the first TOK pore with mjakcha and the second TOK pore with mjakchb). However, yeast TOK channels may have successfully survived a long evolutionary process that masks such a homology. The presence of several K + channel homologues with different topology in the genome of an archetype prokaryote at least reminds us of the option that Kv, Kir and TOK channels have originated from more than one prokaryotic K + channel structure. Heterologous expression has only been achieved for one prokaryotic K + channel (Schrempf et al. 1995; Heginbotham et al. 1997), and further functional data are needed to classify these structures. It is not known whether these translate into classic K + channels or settle in the grey zone where the distinctions between ion channels, membrane pumps, transporters and enzymes break down and they can no longer be treated as separate groups of proteins (Jan and Jan, 1992). With a passage rate of 1 million K + per second, some transport systems, such as KefC, sharing topological features with prokaryotic K + channels, approach the rates of true ion channels (Munro et al. 1991; Jan and Jan, 1997). It is beyond the scope of this review to speculate further about the origin of prokaryotic K + channels and ion channels in general. More prokaryotic sequences and genomes will have to be unravelled to explain how this large protein family has expanded from its precursors. A relationship between K + channels with PNBD and the multifunctional Trk/Kdp/Kef systems remains vague at best (Jan and Jan, 1997). With more data appearing daily, this situation may soon change. We thank The Institute for Genomic Research for supplying sequence data prior to publication, Dr E. Wischmeyer for experimental data and discussions, and Professor J. Daut for his generous support. References ALTSCHUL, S. F., GISH, W., MILLER, W., MYERS, E. W. AND LIPMAN, D. J. (1990). Basic local alignment search tool. J. molec. Biol. 215, BARSTOW, D., MURPHY, J., SHARMAN, A., ATKINSON, T. AND MINTON, N. (1987). Sequence of the Bacillus caldotenax and Bacillus stearothermophilus lctb genes. Nucleic Acids Res. 15, BAUMANN, A., KRAH-JENTGENS, I., MÜLLER, R., MÜLLER-HOLTKAMP, F., SEIDEL, R., KECSKEMETHY, N., CASAL, J., FERRUS, A. AND PONGS, O. (1987). Molecular organization of the maternal effect region of the Shaker complex of Drosophila: characterization of an IA transcript with homology to vertebrate Na + channel. EMBO J. 6, BULT, C. J., WHITE, O., OLSEN, G. J., ZHOU, L., FLEISCHMANN, R. D., SUTTON, G. G., BLAKE, J. A., FITZGERALD, L. M., CLAYTON, R. A. AND GOCAYNE, J. D. (1996). Complete genome sequence of the methanogenic archaeon, Methanococcus jannaschii. Science 273,

8 2798 C. DERST AND A. KARSCHIN COLE, S. T., BROSCH, R., PARKHILL, J., GARNIER, T., CHURCHER, C., HARRIS, D., GORDON, S. V., EIGLMEIER, K., GAS, S. AND BARRY III, C. E. (1998). Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, DECKERT, G., WARREN, P. V., GAASTERLAND, T., YOUNG, W. G., LENOX, A. L., GRAHAM, D. E., OVERBEEK, R., SNEAD, M. A., KELLER, M., AUJAY, M., HUBER, R., FELDMAN, R. A., SHORT, J. M., OLSEN, G. J. AND SWANSON, R. V. (1998). The complete genome of the hyperthermophilic bacterium Aquifex aeolicus. Nature 392, DORSCH, D. C., HELMER, G. L., SUTTON, S. H., SLVACION, F. F. AND EPSTEIN, W. (1991). Genetic analysis of K + transport loci in Escherichia coli: evidence for three constitutive systems mediating uptake of potassium. J. Bacteriol. 173, DOYLE, D. A., CABRAL, J. M., PFUETZNER, R. A., KUO, A., GULBIS, J. M., COHEN, S. L., CHAIT, B. T. AND MACKINNON, R. (1998). The structure of the potassium channel: molecular basis of K + conduction and selectivity. Science 280, FINK, M., DUPRAT, F., LESAGE, F., REYES, R., ROMEY, G., HEURTEAUX, C. AND LAZDUNSKI, M. (1996). Cloning, functional expression and brain localization of a novel unconventional outward rectifier K + channel. EMBO J. 15, FLEISCHMANN, R. D., ADAMS, M. D., WHITE, O., CLAYTON, R. A., KIRKNESS, E. F., KERLAVAGE, A. R., BULT, C. J., TOMB, J.-F., DOUGHERTY, B. A. AND MERRICK, J. M. (1995). Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science 269, FRASER, C. M., GOCAYNE, J. D., WHITE, O., ADAMS, M. D., CLAYTON, R. A., FLEISCHMANN, R. D., BULT, C. J., KERLAVAGE, A. R., SUTTON, G. AND KELLEY, J. M. (1995). The minimal gene complement of Mycoplasma genitalium. Science 270, FRASER, C. M., CASJENS, S., HUANG, W. M., SUTTON, G. G., CLAYTON, R., LATHIGRA, R., WHITE, O., KETCHUM, K. A., DODSON, R. AND HICKEY, E. K. (1997). Genomic sequence of a Lyme disease spirochaete, Borrelia burgdorferi. Nature 390, GILBERT, W. (1985). Genes in pieces revisited. Science 228, GOLDSTEIN, S. A. N., PRICE, L. A., ROSENTHAL, D. N. AND PAUSCH, M. A. (1996). ORK1, a potassium-selective leak channel with two pore domains cloned from Drosophila melanogaster by expression in Saccharomyces cerevisiae. Proc. natn. Acad. Sci. U.S.A. 93, GRIBSKOV, M. AND BURGESS, R. R. (1986). Sigma factors from E. coli, B. subtilis, phage SP01 and phage T4 are homologous proteins. Nucleic Acids Res 14, HEGINBOTHAM, L., LU, Z., ABRAMSON, T. AND MACKINNON, R. (1994). Mutations in the K + channel signature sequence. Biophys. J. 66, HEGINBOTHAM, L., ODESSEY, E. AND MILLER, C. (1997). Tetrameric stoichiometry of a prokaryotic K + channel. Biochemistry 36, HESSE, J. E., WIECZORTEK, L., ALTENDORF, K., REICIN, A. S., DORUS, E. AND EPSTEIN, W. (1984). Sequence homology between two membrane transport ATPases, the Kdp-ATPase of Escherichia coli and the Ca-ATPase of sarcoplasmic reticulum. Proc. natn. Acad. Sci. U.S.A. 81, HILLE, B. (1996). Origin and evolution of channels and signaling mechanisms. In Brain and Evolution (ed. N. Elsner and H. U. Schnitzler), pp Stuttgart, New York: Thieme. HIMMELREICH, R., HILBERT, H., PLAGENS, H., PIRKL, E., LI, B. C. AND HERRMANN, R. (1996). Complete sequence analysis of the genome of the bacterium Mycoplasma pneumoniae. Nucleic Acids Res. 24, HO, K., NICHOLS, C. G., LEDERER, W. J., LYTTON, J., VASSILEV, P. M., KANAZIRSKA, M. V. AND HEBERT, S. C. (1993). Cloning and expression of an inwardly rectifying ATP-regulated potassium channel. Nature 362, JAN, L. Y. AND JAN, Y. N. (1992). Tracing the roots of ion channels. Cell 69, JAN, L. Y. AND JAN, Y. N. (1994). Potassium channels and their evolving gates. Nature 371, JAN, L. Y. AND JAN, Y. N. (1997). Cloned potassium channels from eukaryotes and prokaryotes. A. Rev. Neurosci. 20, JOHANSSON, M. AND VON HEIJNE, G. (1996). Membrane topology of Kch, a putative K + channel from Escherichia coli. J. biol. Chem. 271, KAMB, A., IVERSON, L. E. AND TANOUYE, M. A. (1987). Molecular characterization of Shaker, a Drosophila channel that encodes a potassium channel. Cell 50, KANEKO, T., SATO, S., KOTANI, H., TANAKA, A., ASAMIZU, E., NAKAMURA, Y., MIYAJIMA, N., HIROSAWA, M., SUGIURA, M. AND SASAMOTO, S. (1996). Sequence analysis of the genome of the unicellular cyanobacterium Synechocystis sp. strain PCC6803. II. Sequence determination of the entire genome and assignment of potential protein-coding regions. DNA Res. 3, KETCHUM, K. A., JOINER, W. J., SELLERS, A. J., KACZMAREK, L. K. AND GOLDSTEIN, S. A. N. (1995). A new family of outwardly rectifying potassium channel proteins with two pore domains in tandem. Nature 376, KLENK, H. P. et al. (1997). The complete genome sequence of the hyperthermophilic, sulphate-reducing archaeon Archaeoglobus fulgidus. Nature 390, KUBO, Y., BALDWIN, T. J., JAN, Y. N. AND JAN, L. Y. (1993). Primary structure and functional expression of a mouse inward rectifier potassium channel. Nature 362, KUNST, F. et al. (1997). The complete genome sequence of the grampositive bacterium Bacillus subtilis. Nature 390, KYTE, J. AND DOOLITTLE, R. F. (1982). A simple method for displaying the hydropathic character of a protein. J. molec. Biol. 157, LESAGE, F., GUILLEMARE, E., FINK, M., DUPRAT, F. AND LAZDUNSKI, M. (1996a). TWIK-1, a ubiquitous human weakly inward rectifying K + channel with a novel structure. EMBO J. 15, LESAGE, F., GUILLEMARE, E., FINK, M., DUPRAT, F., LAZDUNSKI, M., ROMEY, G. AND BARHANIN, J. (1996b). A ph-sensitive yeast outward rectifier K + channel with two pore domains and novel gating properties. J. biol. Chem. 271, MACKINNON, R., COHEN, S. L., KUO, A., LEE, A. AND CHAIT, B. T. (1998). Structural conservation in prokaryotic and eukaryotic potassium channels. Science 280, MCCORMACK, T. AND MCCORMACK, K. (1994). Shaker K + channel β- subunits belong to an NAD(P)H-dependent oxidoreductase superfamily. Cell 79, MILKMAN, R. (1994). An Escherichia coli homologue of eukaryotic potassium channel proteins. Proc. natn. Acad. Sci. U.S.A. 91, MILLER, S., DOUGLAS, R. M., CARTER, P. AND BOOTH, I. R. (1997). Mutations in the glutathione-gated KefC K + efflux system of Escherichia coli that cause constitutive activation. J. biol. Chem. 272, MUNRO, A. W., RITCHIE, G. Y., LAMB, A. J., DOUGLAS, R. M. AND BOOTH, I. R. (1991). Cloning and DNA sequence of the gene for

9 Evolutionary link between prokaryotic and eukaryotic K + channels 2799 the glutathione-regulated potassium efflux system KefC of Escherichia coli. Molec. Microbiol. 5, NAGY, E. AND RIGBY, W. F. C. (1995). Glyceraldehyde-3-phosphate dehydrogenase selectively binds AU-rich RNA in the NAD + - binding region (Rossmann fold). J. biol. Chem. 270, NEEDLEMAN, S. B. AND WUNSCH, C. D. (1970). A general method applicable to the search for similarities in the amino acid sequence of two proteins. J. molec. Biol. 48, PAPAZIAN, D. M., SCHWARZ, T. L., TEMPEL, B. L., TIMPE, L. C. AND JAN, L. Y. (1987). Cloning of genomic and complementary cdna from Shaker, a putative potassium channel gene from Drosophila. Science 237, PATIL, N., COX, D. R., BHAT, D., FAHAM, M., MYERS, R. M. AND PETERSON AS, (1995). A potassium channel mutation in weaver mice implicates membrane excitability in granule cell differentiation. Nature Genetics 11, RANGANATHAN, R. (1994). Evolutionary origins of ion channels. Proc. natn. Acad. Sci. U.S.A. 91, REIZER, J., REIZER, A. AND SAIER, M. H. (1992). The putative Na + /H + antiporter (NapA) of Enterococcus hirae is homologous to the putative K + /H + antiporter (KefC) of Escherichia coli. Microbiol. Lett. 94, SAITOU, N. AND NEI, M. (1987). The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molec. Biol. Evol. 4, SALKOFF, L. AND JEGLA, T. (1995). Surfing the DNA databases for K + channels nets yet more diversity. Neuron 15, SCHLEYER, M. AND BAKKER, E. P. (1993). Nucleotide sequence and 3 -end deletion studies indicate that the K + -uptake protein Kup from Escherichia coli is composed of a hydrophobic core linked to a large and partially essential hydrophilic C terminus. J. Bacteriol. 175, SCHLÖSSER, A., HAMANN, A., BOSSEMEYER, D., SCHNEIDER, E. AND BAKKER, E. P. (1993). NAD + binding to the Escherichia coli K + - uptake protein TrkA and sequence similarity between TrkA and domains of a family of dehydrogenases suggest a role for NAD + in bacterial transport. Molec. Microbiol. 9, SCHLÖSSER, A., MELDORF, M., STUMPE, S., BAKKER, E. P. AND EPSTEIN, W. (1995). TrkH and its homolog, TrkG, determine the specificity and kinetics of cation transport by the Trk system of Escherichia coli. J. Bacteriol. 177, SCHREMPF, H., SCHMIDT, O., KÜMMERLEN, R., HINNAH, S., MÜLLER, D., BETZLER, M., STEINKAMP, T. AND WAGNER, R. (1995). A prokaryotic potassium ion channel with two predicted transmembrane segments from Streptomyces lividans. EMBO J. 14, SMITH, D. R., DOUCETTE-STAMM, L. A., DELOUGHERY, C., LEE, H., DUBOIS, J., ALDREDGE, T., BASHIRZADEH, R., BLAKELY, D., COOK, R. AND GILBERT, K. (1997). Complete genome sequence of Methanobacterium thermoautotrophicum deltah: functional analysis and comparative genomics. J. Bacteriol. 179, THOMPSON, J. D., HIGGINS, D. G. AND GIBSON, T. J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, TOMB, J. F., WHITE, O., KERLAVAGE, A. R., CLAYTON, R. A., SUTTON, G. G., FLEISCHMANN, R. D., KETCHUM, K. A., KLENK, H. P., GILL, S. AND DOUGHERTY, B. A. (1997). The complete genome sequence of the gastric pathogen Helicobacter pylori. Nature 388, ZHOU, X. L., VAILLANT, B., LOUKIN, S. H., KUNG, C. AND SAIMI, Y. (1995). YKC1 encodes the depolarization-activated K + channel in the plasma membrane of yeast. FEBS Letts. 373,