A family of putative Kir potassium channels in prokaryotes
© Durell and Guy; licensee BioMed Central Ltd. 2001
Received: 20 June 2001
Accepted: 20 December 2001
Published: 20 December 2001
Prior to this report, members of the inward rectifier family, or Kir, have been found only in eukaryotes. Like most K+ channels, the pore-forming part of the protein is formed by four identical, or closely related, subunits. Each subunit contains a transmembrane M1-P-M2 motif that is followed by a relatively large C-terminus region unique to Kir's.
In searching unfinished microbial genomes for K+ channels, we identified five sequences in the prokaryote Burkholderia pseudomallei, Burkholderia cepacia, Burkholderia fungorum LB400, Magentospirillum magnetotacticum, and Nostoc Punctiforme genomes that code for proteins whose closest relatives in current sequence databases are eukaryote Kir's. The sequence similarity includes the C-terminus portion of Kir's, for which there are no other close homologs in current prokaryote sequences. Sequences of the pore-forming P and M2 segments of these proteins, which we call KirBac, is intermediate between those of eukaryotic Kir's and several other K+ channel families.
Although KirBac's are more closely related to Kir's than to other families of K+ channels, the intermediate nature of their pore-forming P and M2 segments suggests that they resemble an ancestral precursor to the eukaryotic Kir's. The similarity of KirBac to the bacterial KcsA channel, whose transmembrane structure has been solved, helps align Kir's with KcsA. KirBac's may assist in solving the three-dimensional structure of a member of the Kir family since bacterial membrane proteins are more easily expressed in the quantities necessary for crystallography.
Distance matrix for Kir's and KirBac's Below diagonal: Uncorrected distances Above diagonal: Jukes-Cantor distances
Z scores for M1 and M2 sequence profiles for alignments A and B.
We favor Alignment A for M1 for several additional reasons. 1) This alignment does not require any indels for alignment of KirBac's with most members of the Kv and KcsA-like families for the entire M1-P-M2 segments. Indel penalties were not included in our profile calculations. 2) Second site suppressor experiments on Kir2.1 strongly suggest that a serine residue in M1 forms a H bond with a glutamine residue in M2 . This can occur when Alignment A is used to develop homology models based on the KcsA structure. Futhermore, when a homology model (to be presented elsewhere) of KirBac1.1 was developed, the analogous glutamate residue in M2, which is absolutely conserved among KirBac's, can form H-Bonds to the two adjacent asparagines residues in M1; the first of these is analogous to the Kir2.1 serine mentioned above and the second is absolutely conserved among KirBac's. This interaction between the most polar conserved residue on M1 (the glutamate) and the most polar conserved residue on M2 (the asparagine) cannot occur with most other alignments of M1 when the strongly predicted Alignment A is used for M2. 3) When Alignment A is used to develop homology models, most residues that are highly conserved both within and between the different families interact with residues of other transmembrane segments. This point is illustrated in the helical wheel representations shown in Fig. 4 of the M1-P-M2 segments for the different families. Note that both M1 and M2 display patterns that we call unilateral conservation  in which residues that are exposed to lipid on the outer surface are poorly conserved and very hydrophobic, whereas those that interact with other protein residues tend to be more highly conserved. These patterns would not be expected to be the same in the Kv and Kbac6tm1 families because they have four additional transmembrane segments per subunit that should surround their core S5-P-S6 region. In these models, residues that are conserved among the different families tend to cluster together, either near the center of the pore, where they determine the K+ selectivity, or at the region where M1, P, and M2 interact within the subunit. In the latter case, most of the very highly conserved residues are small (glycine, alanine, serine, threonine, or cysteine). Small residues are common where axes of adjacent transmembrane α helices come close together . Although there is little sequence similarity for the M1 and M2 sequences when the Kir sequences are compared to those of the other families, the patterns of sequence conservation of the Kir and KirBac families are remarkably similar to those of the other families when Alignment A is used. Also note that many of the residues that are highly conserved within each family are identical or very similar to residues that are conserved within the KirBac family, as indicated by the red and orange dashed lines that encircle some of the side chains.
Three groups [2–4] have used Alignment A for M2 of KcsA and the eukaryotic Kir's. However, based on the results from yeast mutant screens that identify second site suppressor mutations in M1 and M2 segments in Kir2.1, Minor et al.  proposed Alignment B in which the Kir sequences of Fig. 3 would be shifted three positions to the left relative to the other sequences for M1, while those for M2 would be shifted three positions to the right. They proposed a model to explain their data in which the Kir2.1 has a structure different from that of KcsA in which M1 interacts with M1 helices of adjacent subunits throughout the entire transmembrane region. We are skeptical about the validity of this model because our three-dimensional modeling efforts indicate that the Minor et al. model requires exclusion of the P segment from the transmembrane region. It is highly unlikely that the only portion of the protein with substantial sequence identity between Kir's and KcsA and that determines the selectivity of the channels for potassium would have entirely different structures and/or exist in different locations in these two proteins. Our calculations indicate that Alignment B is clearly inferior to Alignment A for both M2 and M1; in fact, the Z values in Table 2 of Alignment B average zero, as expected for an incorrect alignment. Futhermore, Alignment A requires no indels for the M1-P-M2 regions for most sequences, which were not included in the calculations, whereas Alignment B requires two. Finally, in our hands homology models based on KcsA developed with Alignment A satisfy the mutagenesis data on which Alignment B is based as do models using alignment B, and models using Alignment A are more consistent with mutagenesis studies of other groups [3, 13]. No single model in which the P segment has the structure of KcsA can satisfy all of the second site suppressor data of Minor et al. . However, their data are from an open conformation and the KcsA structure is probably a closed conformation. Most of their data can be satisfied by a combination of two conformations by dramatically altering the position of M1 and the inner portion of M2 for the open conformation (personal observation). It is also conceivable that some of the second site suppressor data are due to an essential intermediate stage of protein folding that differs from the final structure.
Gene transfer between organisms often complicates the interpretation of their evolution. There are now three families of eukaryotic ion channel genes for which only a few homologs have been identified in prokaryotes: the first family is a glutamate-activated K+ channel, GluR0  from Synechocystis PCC 6803, that is homologous to eukaryotic ionotrophic glutamate receptors; the second family is a Na+ channel, NaChBac, from Bacillus halodurans,  and homologous sequences from Thermobifida fusca and Magnetococcus sp. MC1 (personal observation, http://www.jgi.doe.gov/) with only one 6TM motif per subunit that is homologous to the CatSper Ca2+ channel of sperm cells  and to each of four homologous 6TM motifs of the pore-forming subunit of eukaryotic Ca2+ channels; and KirBac is the third family. In each case, the nature of the prokaryotic sequence supports the hypothesis that the gene evolved first in the prokaryotes rather than being transferred to the prokaryote from a eukaryote. For example, mutagenesis experiments [17, 18] suggests that the ion selective region of eukaryotic glutamate receptor pores have a gross structure similar to that of K+ channels; however, only GluR0 has a K+ channel signature sequence and forms a K+ selective channel. This is consistent with the hypothesis that glutamate receptors evolved first in bacteria from K+ channels and then lost their selectivity in eukaryotes. Similarly, the fact that NaChBac is about equidistant from consensus sequences of all four 6TM motifs of eukaryotic Ca2+ channels  is consistent with the hypothesis that Ca2+ channels initially evolved first in prokaryotes as homotetramers from 6TM Kv-like channels and then underwent two consecutive gene duplication events to evolve into the eukaryotic Ca2+ channels that have only one pore-forming subunit that contains four consecutive 6TM motifs. Likewise, the hypothesis that KirBac evolved after transfer of a eukaryotic Kir to a bacteria is inconsistent with several findings: 1) all eukaryotic Kir's are more closely related to each other than to any KirBac, 2) the P-M2 region of KirBac's is more similar to that of some other bacterial K+ channels than it is to that of eukaryotic Kir's, and 3) the M1-P-M2 region of eukaryotic Kir's have numerous features that occur in no other K+ channels, including KirBac. Thus, the Kir's probably evolved first in prokaryotes as proteins similar to KirBac. This finding suggests KirBac has diverged less from the common ancestor to KirBac and eukaryotic Kir's than have the eukaryotic Kir's and supports the hypothesis that Kir's evolved first in bacteria.
The utility of bacterial channels in structural studies has been made abundantly clear by the KcsA structure. Currently little is known about the three-dimensional structure of Kir's. Thus, determination of the structure of KirBac would be a major breakthrough in understanding the structure and functional mechanisms of this important family of K+ channels. Also, it would be interesting to compare the functional properties of KirBac's to those of other Kir's. Chimeric experiments in which proteins are generated that combine part of a KirBac with part of a Kir could be useful in identifying the role of features that are conserved among eukaryotic Kir's but that are not present in KirBac's.
Sequence searches were performed with the web-based programs Blast and PsiBlast at http://www3.ncbi.nlm.nih.gov/BLAST/ for sequences that were deposited in data bases and by tblastn at http://www.ncbi.nlm.nih.gov/MicrobJ3last/unfinishedgenome.html for unpublished microbial sequences. The default matrix (BLOSUM62) and gap cost were used but a filter was not used in these searches. The Wisconsin Package Version 10.2, Genetics Computer Group (GCG), Madison, Wisc. USA. was used to align and edit multiple sequences and to calculated the distance matrices. Clustal W1.74  was also used to make multisequence alignments of the M1-P-M2 region for members of each Kir subfamily plus the other sequences illustrated in Fig. 3.
Quantification of the similarity of the transmembrane segments of the different channel families was accomplished by first transforming the multiple sequence alignments into log-odds residue profile matrices. This was done by the method of Henikoff & Henikoff , as previously described . In summary, the first step was to weight each sequence in a multiple sequence alignment block according to its degree of similarity to the other sequences, which has the effect of minimizing the influence of highly redundant sequences in the final profile. These weights were calculated according to the method of Henikoff & Henikoff , which is based on the residue diversity at each position of the alignment. Next, the sequence-weighted counts were used to calculate the observed occurrence frequency of amino acid residues at each column of the alignment block. To these real residue counts, pseudo-counts were added to better approximate the full set of related sequences in nature (of which only an incomplete, non-random sample is known). Calculation of the pseudo-counts was based on the degree of diversity and statistical substitution probabilities for the specific residues occurring in each of the alignment columns. The recommended value of 5.0 times the residue diversity was used for the total number of pseudo-counts, and the amino acid substitution probabilities were taken from the BLOSUM 62 matrix . A substitution matrix based on transmembrane helices  was also used in some cases, however, the results were not altered substantially. Finally, the log-odds of occurrence of a specific residue is obtained from the logarithm of the sum of real and pseudo-counts divided by the background frequency that would occur in a random sequence by chance. The latter was calculated from the relative occurrence of all amino acids in the SWISS-PROT protein sequence database . The final profiles were then constructed as matrices of dimension 20 by the number of positions in the multiple sequence alignment, where the column vectors provide the log-odds of occurrence of the 20 different amino acids at each position.
Having numerically represented the distribution of residues in the multiple sequence alignments, the similarity of two profiles was calculated according to the method of Pietrokovski . Specifically, the standard Pearson correlation coefficient was calculated for each aligned pair of column vectors and summed over the length of the alignment to provide a raw score. This was then converted to a Z-score, which is the number of standard deviations the raw score is from the mean of the normal distribution of scores that would occur by chance. This distribution was estimated from the scores obtained by randomly permuting the columns of one of the two profiles over 40 thousand times. In contrast to our previous method of calculating the chance distribution from the Blocks database , using the profiles corrects for the specific composition of amino acids in the segments. The Z-score provides a measure of the statistical significance that can be compared among pairs of aligned profiles. More positive scores are less likely to occur by chance, and thus indicate a greater probability that the two protein segments are homologous.
Inward rectifying K+ channel
Inward rectifying K+ channel homolog from bacteria
Voltage-gated K+ channel
a family of prokaryotic channels whose closest relatives are Kv channels.
- 2TM channel:
A channel with only two transmembrane segments per subunit
- 6TM channel:
A channel with six transmembrane segments per subunit.
- Reimann F, Ashcroft FM: Inwardly rectifying potassium channels. Curr Opin Cell Biol. 1999, 11: 503-508. 10.1016/S0955-0674(99)80073-8.View ArticlePubMedGoogle Scholar
- Doyle DA, Morais Cabral J, Pfuetzner RA, Kuo A, Gulbis JM, Cohen SL, Chait BT, MacKinnon R: The structure of the potassium channel: molecular basis of K+ conduction and selectivity. Science. 1998, 280: 69-77. 10.1126/science.280.5360.69.View ArticlePubMedGoogle Scholar
- Loussouarn G, Makhina EN, Rose T, Nichols CG: Structure and dynamics of the pore of inwardly rectifying K(ATP) channels. J Biol Chem. 2000, 275: 1137-1144. 10.1074/jbc.275.2.1137.View ArticlePubMedGoogle Scholar
- Capener CE, Shrivastava IH, Ranatunga KM, Forrest LR, Smith GR, Sansom MS: Homology modeling and molecular dynamics simulation studies of an inward rectifier potassium channel. Biophys J. 2000, 78: 2929-2942.PubMed CentralView ArticlePubMedGoogle Scholar
- Minor DL, Masseling SJ, Jan YN, Jan LY: Transmembrane Structure of an Inwardly Rectifying Potassium Channel. Cell. 1999, 96: 879-891.View ArticlePubMedGoogle Scholar
- Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ: Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997, 25: 3389-3402. 10.1093/nar/25.17.3389.PubMed CentralView ArticlePubMedGoogle Scholar
- Thompson JD, Higgins DG, Gibson TJ TJ: CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994, 22: 4673-4680.PubMed CentralView ArticlePubMedGoogle Scholar
- Wisconsin Package Version 10.2,. Genetics Computer Group (GCG), Madison, Wisc. USA.
- Lu Z, MacKinnon R: Electrostatic tuning of Mg2+ affinity in an inward-rectifier K+ channel. Nature. 1994, 371: 243-246. 10.1038/371243a0.View ArticlePubMedGoogle Scholar
- Guy HR: Models of voltage- and transmitter-activated channels based on their amino acid sequences. In Pasternak, C. A. (Ed.): Monovalent Cations in Biological Systems. Boca Raton, CRC,. 1990, 31-58.Google Scholar
- Eilers M, Shekar SC, Shieh T, Smith SO, Fleming PJ: Internal packing of helical membrane proteins. Proc Natl Acad Sci USA. 2000, 97: 5796-5801. 10.1073/pnas.97.11.5796.PubMed CentralView ArticlePubMedGoogle Scholar
- Choe S, Stevens CF, Sullivan JM: Three Distinct Structural Environments of a Transmembrane Domain in the Inwardly Rectifying Potassium Channel ROMK1 Defined by Perturbation. Proc Natl Acad Sci U S A. 92: 12046-12049.
- Loussouarn G, Phillips LR, Masia R, Rose T, Nichols CG: Flexibility of the Kir6.2 inward rectifier K+ channel pore. Proc Natl Acad Sci U S A. 2001, 98: 4227-4232. 10.1073/pnas.061452698.PubMed CentralView ArticlePubMedGoogle Scholar
- Chen GQ, Cui C, Mayer ML, Gouaux E: Functional characterization of a potassium-selective prokaryotic glutamate receptor. Nature. 1999, 402: 817-821. 10.1038/45568.View ArticlePubMedGoogle Scholar
- Durell SR, Guy HR: A putative prokaryote voltage-gated Ca2+ channel with only one 6TM motif per subunit. Biochem Biophys Res Commun. 2001, 281: 741-746. 10.1006/bbrc.2001.4408.View ArticlePubMedGoogle Scholar
- Ren D, Navarro B, Perez G, Jackson AC, Hsu S, Shi Q, Tilly JL, Clapham DE: A sperm ion channel required for sperm motility and male fertility. Nature. 2001, 413: 603-609. 10.1002/(SICI)1096-9861(19991101)413:4<603::AID-CNE9>3.0.CO;2-K.View ArticlePubMedGoogle Scholar
- Kuner T, Wollmuth LP, Karlin A, Seeburg PH, Sakmann B: Structure of the NMDA receptor channel M2 segment inferred from the accessibility of substituted cysteines. Neuron. 1996, 17: 343-352.View ArticlePubMedGoogle Scholar
- Panchenko VA, Glasser CR, Mayer ML: Structural Similarities between Glutamate Receptor Channels and K+ Channels Examined by Scanning Mutagenesis. J Gen Physiol. 2001, 117: 345-360. 10.1085/jgp.117.4.345.PubMed CentralView ArticlePubMedGoogle Scholar
- Henikoff JG, Henikoff S: Using substitution probabilities to improve position-specific scoring matrices. Comput. Appl. Biosci. 1996, 12: 135-143.PubMedGoogle Scholar
- Durell SR, Hao Y, Nakamura T, Bakker EP, Guy HR: Evolutionary relationship between K+ channels and symporters. Biophys. J. 1999, 77: 775-788.PubMed CentralView ArticlePubMedGoogle Scholar
- Henikoff S S, Henikoff JG: Position-based sequence weights. J. Mol. Biol. 1994, 243: 574-578.View ArticlePubMedGoogle Scholar
- Henikoff JG, Henikoff S: Amino acid substitution matrices from protein blocks. Proc Natl Acad Sci U S A. 1992, 89: 10915-10919.PubMed CentralView ArticlePubMedGoogle Scholar
- Ng PC, Henikoff JG, Henikoff S: PHAT: a transmembrane-specific substitution matrix. Bioinformatics. 2000, 16: 760-766. 10.1093/bioinformatics/16.9.760.View ArticlePubMedGoogle Scholar
- Bairoch A, Apweiler R: The SWISS-PROT protein sequence data bank and its supplement TrEMBL in 1998. Nucleic Acids Res. 1998, 26: 38-42. 10.1093/nar/26.1.38.PubMed CentralView ArticlePubMedGoogle Scholar
- Pietrokovski S: Searching databases of conserved sequence regions by aligning protein multiple-alignments. Nucleic Acids Res. 1996, 24: 3836-3845. 10.1093/nar/24.19.3836.PubMed CentralView ArticlePubMedGoogle Scholar
- Ren D, Navarro B, Yue L, Shi Q, Clapham DE: A prokaryotic Voltage-gated sodium channel. Science. 2001, 294: 2372-2375. 10.1126/science.1065635.View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article: verbatim copying and redistribution of this article are permitted in all media for any purpose, provided this notice is preserved along with the article's original URL.