A Single Transmembrane Site in the KCNE-encoded Proteins Controls the Specificity of KvLQT1 Channel Gating
2002; Elsevier BV; Volume: 277; Issue: 28 Linguagem: Inglês
10.1074/jbc.m200564200
ISSN1083-351X
AutoresYonathan F. Melman, Andrew Krumerman, Thomas V. McDonald,
Tópico(s)Neuroscience and Neuropharmacology Research
ResumoKCNEs are a family of genes encoding small integral membrane proteins whose role in governing voltage-gated potassium channel gating is emerging. Whether each member of this homologous family interacts with channel proteins in the same manner is unknown; however, it is clear that the functional effect of each KCNE on channel gating is different. The specificity of KCNE1 (minK) and KCNE3 control of activation of the potassium channel KvLQT1 maps to a triplet of amino acids within the KCNE transmembrane domain by chimera analysis. We now define the structural determinants of functional specificity within this triplet. The central amino acid of the triplet (Thr-58 of minK and Val-72 of KCNE3) is essential for the specific control of voltage-dependent channel activation characteristics of both minK and KCNE3. Using site-directed mutations that substitute minK and KCNE3 residues, we determined that a hydroxylated central amino acid is necessary for the slow sigmoidal activation produced by minK. The precise spacing of the hydroxyl group was required for minK-like activation. An aliphatic amino acid substituted at position 58 of minK is capable of reproducing KCNE3-like kinetics and voltage-independent constitutive current activation. The bulk of the central residue is another critical parameter, indicating precise positioning of this portion of the KCNE proteins within the channel complex. An intermediate phenotype produced by several smaller aliphatic-substituted mutants yields conditional voltage independence that is distinct from the voltage-dependent gating process, suggesting that KCNE3 traps the channel in a stable open state. From these results, we propose a model of KCNE-potassium channel interaction where the functional consequence depends on the precise contact at a single amino acid. KCNEs are a family of genes encoding small integral membrane proteins whose role in governing voltage-gated potassium channel gating is emerging. Whether each member of this homologous family interacts with channel proteins in the same manner is unknown; however, it is clear that the functional effect of each KCNE on channel gating is different. The specificity of KCNE1 (minK) and KCNE3 control of activation of the potassium channel KvLQT1 maps to a triplet of amino acids within the KCNE transmembrane domain by chimera analysis. We now define the structural determinants of functional specificity within this triplet. The central amino acid of the triplet (Thr-58 of minK and Val-72 of KCNE3) is essential for the specific control of voltage-dependent channel activation characteristics of both minK and KCNE3. Using site-directed mutations that substitute minK and KCNE3 residues, we determined that a hydroxylated central amino acid is necessary for the slow sigmoidal activation produced by minK. The precise spacing of the hydroxyl group was required for minK-like activation. An aliphatic amino acid substituted at position 58 of minK is capable of reproducing KCNE3-like kinetics and voltage-independent constitutive current activation. The bulk of the central residue is another critical parameter, indicating precise positioning of this portion of the KCNE proteins within the channel complex. An intermediate phenotype produced by several smaller aliphatic-substituted mutants yields conditional voltage independence that is distinct from the voltage-dependent gating process, suggesting that KCNE3 traps the channel in a stable open state. From these results, we propose a model of KCNE-potassium channel interaction where the functional consequence depends on the precise contact at a single amino acid. Chinese hamster ovary change in Gibbs' free energy voltage at half-maximal activation trapping of the open state The diversity of potassium ion (K+) currents observed in native tissues exceeds the number of K+ channel genes identified. The explanations for this functional diversity include alternative splicing (1McCormack K. McCormack T. Tanouye M. Rudy B. Stühmer W. FEBS Lett. 1995; 370: 32-36Crossref PubMed Scopus (94) Google Scholar), heteromultimeric assembly of channel subunits (2Wang W. Xia J. Kass R.S. J. Biol. Chem. 1998; 273: 34069-34074Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar), and association with accessory subunits that alter channel properties. Such a family of accessory subunits is represented by theKCNEs. The KCNE genes encode a family of small type I transmembrane proteins with one membrane-spanning segment whose role in potassium channel regulation is emerging (3Abbott G.W. Sesti F. Splawski I. Buck M.E. Lehmann M.H. Timothy K.W. Keating M.T. Goldstein S.A. Cell. 1999; 97: 175-187Abstract Full Text Full Text PDF PubMed Scopus (1180) Google Scholar, 4Sanguinetti M.C. Trends Pharmacol. Sci. 2000; 21: 199-201Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar). Although they themselves do not form functional ion channels, they associate with a variety of voltage-gated K+ channels and exert control over gating kinetics (5Tristani-Firouzi M. Sanguinetti M.C. J. Physiol. (Lond.). 1998; 510: 37-45Crossref Scopus (136) Google Scholar), voltage dependence, drug sensitivity (6Busch A.E. Busch G.L. Ford E. Suessbrich H. Lang H.J. Greger R. Kunzelmann K. Attali B. Stuhmer W. Br. J. Pharmacol. 1997; 122: 187-189Crossref PubMed Scopus (108) Google Scholar, 7Wang H.S. Brown B.S. McKinnon D. Cohen I.S. Mol. Pharmacol. 2000; 57: 1218-1223Crossref PubMed Scopus (21) Google Scholar, 8Sesti F. Abbott G.W. Wei J. Murray K.T. Saksena S. Schwartz P.J. Priori S.G. Roden D.M. George A.L., Jr. Goldstein S.A. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 10613-10618Crossref PubMed Scopus (445) Google Scholar), and conductance (9Sesti F. Goldstein S.A. J. Gen. Physiol. 1998; 112: 651-663Crossref PubMed Scopus (191) Google Scholar). The importance of the KCNEs is reflected by the mutations within these proteins that are linked to human genetic disease stemming from defective repolarization of cardiac (hereditary Long Q-T syndrome) (10Splawski I. Tristani-Firouzi M. Lehmann M.H. Sanguinetti M.C. Keating M.T. Nat. Genet. 1997; 17: 338-340Crossref PubMed Scopus (677) Google Scholar, 11Schulze-Bahr E. Wang Q. Wedekind H. Haverkamp W. Chen Q. Sun Y. Rubie C. Hordt M. Towbin J.A. Borggrefe M. Assmann G., Qu, X. Somberg J.C. Breithardt G. Oberti C. Funke H. Nat. Genet. 1997; 17: 267-268Crossref PubMed Scopus (372) Google Scholar, 12Tyson J. Tranebjaerg L. Bellman S. Wren C. Taylor J.F. Bathen J. Aslaksen B. Sorland S.J. Lund O. Malcolm S. Pembrey M. Bhattacharya S. Bitner-Glindzicz M. Hum. Mol. Genet. 1997; 6: 2179-2185Crossref PubMed Scopus (265) Google Scholar), sporadic and drug-induced Long Q-T syndrome (3Abbott G.W. Sesti F. Splawski I. Buck M.E. Lehmann M.H. Timothy K.W. Keating M.T. Goldstein S.A. Cell. 1999; 97: 175-187Abstract Full Text Full Text PDF PubMed Scopus (1180) Google Scholar), and skeletal (familial periodic paralysis) (13Abbott G.W. Butler M.H. Bendahhou S. Dalakas M.C. Ptacek L.J. Goldstein S.A. Cell. 2001; 104: 217-231Abstract Full Text Full Text PDF PubMed Scopus (265) Google Scholar) tissues. The first member of the family, KCNE1 (14Takumi T. Ohkubo H. Nakanishi S. FASEB J. 1989; 5: 331-337Google Scholar), encodes the minK protein and associates with another Long Q-T syndrome-linked protein, KvLQT1, to produce the slowly activating delayed rectifier K+ current, Iks (15Sanguinetti M.C. Curran M.E. Zou A. Shen J. Spector P.S. Atkinson D.L. Keating M.T. Nature. 1996; 384: 80-83Crossref PubMed Scopus (1524) Google Scholar, 16Barhanin J. Lesage F. Guillemare E. Fink M. Lazdunski M. Romey G. Nature. 1996; 384: 78-80Crossref PubMed Scopus (1401) Google Scholar).KCNE3 encodes a protein that interacts with either KvLQT1 (17Schroeder B.C. Waldegger S. Fehr S. Bleich M. Warth R. Greger R. Jentsch T.J. Nature. 2000; 403: 196-199Crossref PubMed Scopus (422) Google Scholar) or Kv3.4 (13Abbott G.W. Butler M.H. Bendahhou S. Dalakas M.C. Ptacek L.J. Goldstein S.A. Cell. 2001; 104: 217-231Abstract Full Text Full Text PDF PubMed Scopus (265) Google Scholar), a channel protein expressed in skeletal muscle. The KCNE proteins appear to have the ability to associate with a variety of voltage-gated channels, and correspondingly, several K+channels proteins are capable of associating with several different KCNE proteins (3Abbott G.W. Sesti F. Splawski I. Buck M.E. Lehmann M.H. Timothy K.W. Keating M.T. Goldstein S.A. Cell. 1999; 97: 175-187Abstract Full Text Full Text PDF PubMed Scopus (1180) Google Scholar, 13Abbott G.W. Butler M.H. Bendahhou S. Dalakas M.C. Ptacek L.J. Goldstein S.A. Cell. 2001; 104: 217-231Abstract Full Text Full Text PDF PubMed Scopus (265) Google Scholar, 17Schroeder B.C. Waldegger S. Fehr S. Bleich M. Warth R. Greger R. Jentsch T.J. Nature. 2000; 403: 196-199Crossref PubMed Scopus (422) Google Scholar, 18McDonald T.V., Yu, Z. Ming Z. Palma E. Meyers M.B. Goldstein S.A.N. Fishman G.I. Nature. 1997; 388: 289-292Crossref PubMed Scopus (309) Google Scholar, 19Zhang M. Jiang M. Tseng G.N. Circ. Res. 2001; 88: 1012-1019Crossref PubMed Scopus (168) Google Scholar). The relevance of these additional interactions to native channels is still uncertain. Despite the structural similarities between KCNE1 and KCNE3, the functional consequences of their association with KvLQT1 are distinctly different. The dramatically different effects of KCNE1 and KCNE3 on KvLQT1 provide a powerful tool to investigate how the KCNEs control K+ channel activity. Using KCNE1/KCNE3 chimeras, we have previously shown that a three-amino acid segment, the "activation triplet," within the transmembrane domain of minK is sufficient and necessary to confer the specificity of KCNE control of voltage-dependent channel activation (20Melman Y.F. Domenech A. de la Luna S. McDonald T.V. J. Biol. Chem. 2001; 276: 6439-6444Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). Our findings also showed that different regions of the KCNE proteins are responsible for different functions, namely the modulation of kinetics of activation and deactivation. In this paper, we refine our investigation of the role of individual amino acids within the KCNE activation triplet and show that the central residue is largely responsible for differences in activation kinetics between KCNE1 and KCNE3. By substituting homologous residues at this site, we identify structural requirements for minKversus KCNE3-like kinetics. This work provides more precise information on structural requirements and specificity of KCNE interaction and control of voltage-gated K+ channels. CHO1 cells were maintained in Ham's F-12 medium supplemented with 10% fetal calf serum and penicillin and/or streptomycin at 37 °C and 5% CO2. Gene transfer was performed using 15 μg of Qiagen Midiprep-purified plasmid DNA. Cells were electroporated at 225 V, 72 ohm, and 1800 microfarads with cytomix media. We performed electrophysiology studies 24–48 h after transfection. Molar ratios of 7:7:2 of kvlqt1:KCNE:green fluorescence protein plasmid were used to allow identification of transfected cells by fluorescence. We used the whole cell configuration of the patch clamp technique (21Hammill O.P. Marty A. Neher E. Sakmann B. Sigworth F.J. Pfluegers Arch. Eur. J. Physiol. 1981; 391: 85-100Crossref PubMed Scopus (15174) Google Scholar) to measure K+ currents carried by the heterologously expressed recombinant channels. The details of the techniques were identical to those described previously (20Melman Y.F. Domenech A. de la Luna S. McDonald T.V. J. Biol. Chem. 2001; 276: 6439-6444Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). All mutants were constructed using a variation of the QuikChange (Stratagene) PCR-based site directed mutagenesis system as previously described (20Melman Y.F. Domenech A. de la Luna S. McDonald T.V. J. Biol. Chem. 2001; 276: 6439-6444Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). The sequence of the wild-type KCNE1 region is ctgggattcttcggcttcttcactctgggcatcatgctgagctac, and the sequence of KCNE3 region is gtcatgttcctatttgctgtaactgtcggcagcctcatcctgggatac. Codons 57–59 of minK and 71–73 of KCNE3 are underlined. Primer pairs consisted of this sequence and its reverse complement with the desired mutations within the underlined stretches. We chose specific sequences based on standard human codon usage tables to introduce the desired mutations. As we previously showed (20Melman Y.F. Domenech A. de la Luna S. McDonald T.V. J. Biol. Chem. 2001; 276: 6439-6444Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar), the kinetics and voltage dependence of activation, current density, and the appearance of a constitutively activated component of current from KvLQT1 are dependent on the co-expression of minK (KCNE1) or KCNE3 (Fig.1, a–c). We also showed that a stretch of three amino acids within the transmembrane domain of KCNE1 and KCNE3 was both necessary and sufficient to confer the unique properties of current activation on KvLQT1. Specifically, the substitution of KCNE1 residues 57–59 into KCNE3 produced a current that is voltage-dependent and activated slowly whereas the converse transplantation of KCNE3 residues 71–73 into KCNE1 yielded a more rapidly activating current with an additional component of constitutively active voltage-independent current (Fig. 1 d). To examine the contributions of each amino acid within the triplet toward control of channel gating, we constructed all six possible single and double swaps of KCNE1 residues 57–59 into the corresponding region of KCNE3 (residues 71–73). The mutants were examined for their ability to produce currents resembling either minK·KvLQT1 or KCNE3·KvLQT1 complexes. Specifically, we asked whether the currents exhibited slow sigmoidal activation kinetics of KvLQT1-minK channels or rapid monoexponential activation plus a voltage-independent component characteristic of KvLQT1·KCNE3 channels. Fig. 1, e–i, shows currents from KCNE3 with minK-substituted amino acids within the activation triplet. All of the mutants containing the substitution V72T, either alone (Fig.1 e) or in combination with another residue within the triplet (Fig. 1, g and i), abolished the usual KCNE3-mediated constitutive current and exhibited slow sigmoidal-type activation kinetics that resemble minK. In contrast, the substitution of the other residues within the triplet did not eliminate the constitutively activated voltage-independent current (Fig. 1,d, f, and h), and monoexponential activation was seen. These results suggest that the specific control of activation by the KCNE proteins may be largely confined to just one amino acid, Thr-58, of minK (Val-72 of KCNE3). However, some contribution to activation gating by the two flanking residues is evident as seen in Fig. 1 h (KCNE3 T71F/G73L) where there is a slow but non-sigmoidal component of activation in addition to the constitutively active current. To test our hypothesis of the critical role of residue Val-72 of KCNE3, we constructed KCNE3 mutants A69T KCNE3 and L75T KCNE3. These two mutants contain a threonine located one turn up (N terminus) or one turn down the transmembrane α-helix of minK and, hence, would face the same direction as the threonine in V72T KCNE3. In contrast to the dramatic effects of the substitution of threonine at position 72 of KCNE3 (Fig. 1 b), the substitution at positions 69 and 75 had virtually no effect on KCNE3 gating (Fig. 3, a and b). To confirm this single amino acid hypothesis, we substituted KCNE3 amino acids into the activation triplet of minK (residues 57–59). The insertion of the first and/or third residues (threonine and glycine) of the triplet into KCNE1 did not reproduce the kinetics seen with KCNE3 (Fig. 2 b). When the central valine of KCNE3 was substituted into minK with either of the flanking residues of KCNE3, the slow sigmoidal activation of minK was replaced by rapid monoexponential activation (Fig. 2, a andc). When the central valine was inserted alone or with threonine in position one there appeared a distinct constitutively active voltage-independent component (Fig. 2, a andd). A further examination of T58V revealed that the voltage-independent component could be diminished or eliminated and the activation could be slowed by administering a preconditioning hyperpolarization pulse (−120 mV for 3 s) rather than the more physiological holding potential (− 80 mV) (Fig. 2 d). TheV h of the voltage-activated component of this mutant is not significantly left-shifted relative to minK (T58V, 27.7 ± 1.01 mV, n = 10, versus WT minK, 30.3 ± 3.1 mV, n = 10). Therefore a simple left-shifted activation process that leaves channels open at hyperpolarized potentials does not account for this conditional voltage-independent state but rather represents a distinct and stable open state that arises, which arises from our mutation. We also introduced threonines into T58V minK at positions 55 and 61, one turn N- or C-terminal to residue 58, respectively. Similar to our results with KCNE3, the mutation of Gly-55 failed to restore minK-like gating (Fig. 3 c). However, the insertion of a threonine at position 61 had a more pronounced effect and produced currents resembling minK (Fig. 3 d), which contained both less constitutive current and slower activation kinetics (Fig. 3, e and f). Thus, although the precise position of threonine is critical for its function in generating minK-like currents, this function may be possible from spatially proximal positions as well. We further investigated the role of the middle residue within the KCNE triplet by making both conservative and divergent amino acid substitutions. Specifically, we altered the bulkiness of the side chain using both larger (isoleucine) and smaller (alanine) aliphatic and polar residues (serine and threonine). In substituting the hydroxylated residue serine into KCNE3 (V72S), we observed slow voltage-dependent activation with sigmoidal kinetics and no voltage-independent constitutive current (Fig.4 a). This current was nearly identical to that produced by the V72T KCNE3 mutant, suggesting the need for a hydroxyl group at this position in producing slow sigmoidal voltage-dependent activation. The support for this hypothesis comes from V72C-KCNE3. Cysteine is isoelectric with serine, but currents are almost completely voltage-independent and resemble KCNE3 (Fig. 4 b). V72I-KCNE3, a mutant that introduces a larger aliphatic residue, retained voltage-independent current and rapid activation characteristic of KCNE3 (Fig. 4 d). The substitution of the smaller aliphatic amino acid, alanine at this position results in decreased voltage-independent constitutive current and slower activation kinetics (Fig. 4 c). This finding suggests that a minimal size of hydrophobic residue is necessary to keep the channel open during hyperpolarized potentials; conversely, hydroxylation eliminates this voltage-independent state entirely. That two sets of isosteric residues (threonine and valine, cysteine and serine) gave dramatically different currents solely with the introduction of a hydroxyl group suggests that this group has an important role in control KvLQT1 gating. We investigated this possibility further by substituting into the central position of KCNE3 tyrosine, a residue that is both bulky and hydroxylated and where the hydroxyl group is at a different spacing from the carbon backbone than serine or threonine. Although V72Y KCNE3 has both rapid and slow activation, there is a significant voltage-independent component and no sigmoidal minK-like kinetics (Fig. 4 e), suggesting that a hydrophobic group keeps the channel open in a voltage-independent manner and that precise spacing of the hydroxyl group is necessary to produce activation kinetics typical of Iks. We made comparable mutations in minK and observed similar trends. The substitution of neither serine nor alanine caused voltage-independent current (Fig. 4, f and g, respectively), in parallel to the findings with KCNE3. The activation kinetics of these two mutants (T58A minK and T58S minK) are accelerated relative to minK as estimated by rise times to half-maximum following an 80-mV depolarization (t 12 of minK = 706 ± 27 (n = 10), T58A minK = 314 ± 15 (n = 6), T58S minK = 358 ± 14 (n = 12)) (Fig. 3, f and g). However, T58A minK exhibits a non-sigmoidal activation of current (Fig.4 g), whereas T58S minK has a sigmoidal time course similar to wild type minK (Fig. 4 f), illustrating the importance of the hydroxyl group in reproducing minK-like activation kinetics. The mutations at minK position Thr-58 that introduced non-hydroxylated groups of various sizes exhibited prepulse-dependent constitutive opening (Fig. 5,a–d). The substitution of isoleucine produced a voltage-independent constitutive component that could be decreased by prolonged hyperpolarizations in a manner that is similar to the current seen with the T58V substitution. We also found that T58C minK behaves in a similar manner. Comparable with substitutions at the central position in KCNE3, T58S minK and T58C minK reveals that two isoelectric residues exhibit markedly different characteristics, further supporting the importance of an appropriately spaced hydroxylated side chain in producing Iks-like activation. In our previous work, we showed that the activation triplet of minK and KCNE3 did not control KvLQT1 deactivation kinetics (20Melman Y.F. Domenech A. de la Luna S. McDonald T.V. J. Biol. Chem. 2001; 276: 6439-6444Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). Specifically, the slowed deactivation kinetics of minK·KvLQT1 complexes are not observed when minK residues 57–59 are inserted into KCNE3, despite the slowing of activation kinetics. Therefore, we examined the deactivation properties of the site-directed mutants presented here. Wild type KvLQT1-KCNE1 and KvLQT1-KCNE3 channels close with a monoexponential time course after the removal of the depolarizing stimulus (Fig.5 f). Consistent with our previous work, most mutations within the activation triplet did not affect deactivation kinetics of the parent KCNE backbone, and deactivation kinetics are comparable with the wild type protein (Fig. 5 f). However, the deactivation kinetics of T58V minK, T58I minK and T58C minK, and T71F/G73L KCNE3 could not be fit to a monoexponential time course. Each of these mutants deactivates with a time course best fit with two time constants (Fig. 5, e and f). Each of these mutants was also previously observed to have a voltage-independent component that could be eliminated or reduced by prolonged hyperpolarizing conditioning pulses. The V h of the voltage-dependent component of these minK mutants was not left-shifted to an extent that would account for the constitutive current we observed (T58VminK V h = 27.7 ± 1.07 (n = 10), T58I minK V h = 26.7 ± 1.07 (n = 4), T58C minKV h = −8.5 mV ± 1.5 (n = 13), T71F/G73L KCNE3 −14.93 ± 1.4 mV (n = 7)versus KvLQT1 alone, V h = −15.2 ± 0.3 mV (n = 10), which shows no such prepulse dependence) and thus cannot be explained merely by a hyperpolarizing shift in activation gating. The bi-exponential deactivation kinetics reflects two open states and supports this hypothesis. The faster time constant represents closure of the voltage-dependent open state, and relaxation of an apparent "inducible" voltage dependence is reflected in the decay of current on a much slower time scale. Significantly, the recording of deactivation kinetics under conditions of prolonged hyperpolarizing prepulses converted the deactivation kinetics to a mono-exponential time course, as if hyperpolarization eliminated one of the open states (Fig. 5 e, right trace). In this work, we have refined our understanding of the structural determinants for functional specificity of the KCNE proteins within a previously mapped activation triplet of amino acids in the transmembrane domain. We show that the central residue of the triplet (T58 for minK or V72 for KCNE3) is the most influential amino acid in determining kinetics of voltage-dependent activation of KvLQT1. A central valine is necessary and sufficient to confer specificity of control of KvLQT1 channel activation by KCNE3. The introduction of mutations spaced one α-helical turn in the amino or carboxyl direction failed to override the phenotype of Val-72, suggesting that its precise positioning is critical for its function. Through detailed analysis, we show that the middle residue (Thr-58) is critical for minK as well, but that surrounding residues within the minK activation triplet functionally contribute to IKs-type activation kinetics. The amino acids in the central position of the activation triplet of minK and KCNE3 are isosteric with only the hydroxyl group of threonine in minK differing from the valine of KCNE3. In summary, we found that a precisely spaced hydroxyl group in the central amino acid produces mink-like kinetics, and larger aliphatic side chains produce KCNE3-like activation. In this study we have focused on the role of a small transmembrane region of KCNEs that specify the control of KvLQT1-gating kinetics. Other investigators have highlighted the functional importance of the cytoplasmic C-terminal tail that is the most conserved region among the KCNEs. It is clear that removal of the minK C terminus (22Tapper A.R. George A.L., Jr. J. Gen. Physiol. 2000; 116: 379-390Crossref PubMed Scopus (79) Google Scholar) or mutation of key sites within the C terminus (9Sesti F. Goldstein S.A. J. Gen. Physiol. 1998; 112: 651-663Crossref PubMed Scopus (191) Google Scholar, 23Takumi T. Moriyoshi K. Aramori I. Ishii T. Oiki S. Okada Y. Ohkubo H. Nakanishi S. J. Biol. Chem. 1991; 2266: 22192-22198Abstract Full Text PDF Google Scholar, 24Bianchi L. Shen Z. Dennis A.T. Priori S.G. Napolitano C. Ronchetti E. Bryskin R. Schwartz P.J. Brown A.M. Hum. Mol. Genet. 1999; 8: 1499-1507Crossref PubMed Scopus (155) Google Scholar) abolishes the functional effects that minK exerts over KvLQT1. These findings raise the interesting question, is the C-terminal portion of minK in fact the region that controls gating of the channel? If this were the case, we would expect that the channel interacts with the KCNE C terminus in two different ways to produce minK or KCNE3-like currents. Our data argue against such a mechanism, because the specificity of minK and KCNE3 actions resides entirely within the transmembrane segments centered on the key single residue (Thr-58 or Val-72). An exchange of the C termini of minK and KCNE3 had no functional consequence (20Melman Y.F. Domenech A. de la Luna S. McDonald T.V. J. Biol. Chem. 2001; 276: 6439-6444Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). The C terminus of minK and KCNE3 are the most highly conserved portions of the molecule, suggesting that its function in the channel complex is also conserved and common between the two. This site of interaction may serve as a crucial point of contact or anchoring that is required for precisely positioning the transmembrane domain to exert its specific effect. The transmembrane domain is postulated to lie in contact with or in close proximity to the KvLQT1 pore or pore-lining helices (22Tapper A.R. George A.L., Jr. J. Gen. Physiol. 2000; 116: 379-390Crossref PubMed Scopus (79) Google Scholar, 25Goldstein S.A.N. Miller C. Neuron. 1991; 7: 403-408Abstract Full Text PDF PubMed Scopus (110) Google Scholar, 26Wang K.W. Tai K.K. Goldstein S.A.N. Neuron. 1996; 16: 571-577Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar, 27Tapper A.R. George A.L., Jr. J. Biol. Chem. 2001; 276: 38249-38254Abstract Full Text Full Text PDF PubMed Google Scholar). Accordingly, the transmembrane segments of the KCNEs are in a better position to effect voltage-dependent gating. Our finding that some mutants with KCNE3-like behavior can be induced to conduct minK-like currents when preconditioned may be an indication of the mechanism by which KCNE3 converts a voltage-dependent channel into a voltage independent one. After a conditioning hyperpolarized holding potential (−120 mV), these mutant channels open with minK-like kinetics and then become trapped in an apparently constitutive open state until the membrane is hyperpolarized again to −120 mV. Once the conditional constitutively open state is entered, the reentry into a closed state occurs slowly as seen with the bi-exponential deactivation time course. This suggests the presence of two open conducting states, one resulting from voltage-dependent opening that is rapidly reversible and another that can occur through trapping of the channel into a very stable conducting state. KCNE3 may trap a component of the KvLQT1-gating structure in the open state, producing the constitutively active voltage-independent current. T58V minK, T58I minK, and T58C minK appear only to be able to trap partially the channel in this state and only so when the channel assumes the open state driven by the electric field across the membrane during depolarization. These preconditioning effects on gating are reminiscent of the Cole-Moore effect observed in K+ currents of squid axon (28Cole K.S. Moore J.W. Biophys. J. 1960; 1: 161-202Abstract Full Text PDF PubMed Scopus (234) Google Scholar) that has also been described for IKS in Xenopus oocytes (29Tzounopoulos T. Maylie J. Adelman J.P. Biophys. J. 1998; 74: 2299-2305Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar). The classical Cole-Moore effect is a shift in the onset of channel opening dependent on holding potential but with otherwise normal kinetics and is explained as moving the channel from one of several sequential closed states to another. Thus, more hyperpolarized holding potentials will force the channel into a closed state that requires more conformational changes to reach the open state and delay the onset of activation. Tzounopoulos et al. (29Tzounopoulos T. Maylie J. Adelman J.P. Biophys. J. 1998; 74: 2299-2305Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar) found that minK-IKS exhibited more complex activation changes with holding potential variations that could not be fully explained by a model of sequential, identical, and first-order transitions. Our results with the conditional KCNE3-minK mutants suggest that we are observing an additional discrete open state that is highly stable as well as the altered transition from closed to open states. The precedence for "trapped" states of ion channels exists for Na+ and K+ channels. Voltage-gated Na+ channels are altered by β-scorpion toxins by binding the voltage sensor of the channels while they are in the outward position, thus trapping them in the open state (30Cestele S., Qu, Y. Rogers J.C. Rochat H. Scheuer T. Catterall W.A. Neuron. 1998; 21: 919-931Abstract Full Text Full Text PDF PubMed Scopus (313) Google Scholar). Furthermore, prolonged hyperpolarizations could reverse this trapping in a fashion similar to that we have demonstrated with several of our mutations. Additional support comes from Isk produced by minK in Xenopus oocytes that could be altered by the cross-linking agent MTSSP to induce a constitutively active state similar to that seen with KCNE3-like mutants (31Varnum M.D. Maylie J. Busch A. Adelman J.P. Neuron. 1995; 14: 407-412Abstract Full Text PDF PubMed Scopus (17) Google Scholar). We previously used thermodynamics to help understand the control of kinetics of both opening and closure by minK and KCNE3 chimera mutants (20Melman Y.F. Domenech A. de la Luna S. McDonald T.V. J. Biol. Chem. 2001; 276: 6439-6444Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). That different regions exert control over opening and closing suggests the activation energy for the forward and reverse transitions states is different. This can be envisioned as an altered channel conformation resulting from the different electric field in these two conditions. In such a model, the structure of the transition state is different in the forward and reverse directions and hence could be controlled by different regions of minK. Our observation of bi-exponential kinetics of deactivation for channels with inducible voltage independence suggests that not all open channels are trapped, that non-trapped channels close faster, and the trapped channels close much slower requiring extreme hyperpolarizing forces. Thermodynamically, the trapping of the open state (Otrapped) is equivalent to a large negative change in free energy (ΔGtrapped) relative to the initial open state without necessarily changing the ΔG between closed state and open state (Fig. 6). Thus, the transition back to the closed state will involve a reversal of the trapped interaction and necessarily has a high transition energy requiring supraphysiological differences in the electric field across the membrane. Work from several groups has shed light on the nature of minK regulation of KvLQT1. Cysteines inserted in the transmembrane domain of minK are susceptible to reactivity with cadmium (32Tai K.K. Goldstein S.A. Nature. 1998; 391: 605-608Crossref PubMed Scopus (111) Google Scholar) and methanethiosulfonate (26Wang K.W. Tai K.K. Goldstein S.A.N. Neuron. 1996; 16: 571-577Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar) reagents from only one face of the membrane, leading the authors to conclude that the minK transmembrane segment lines the conduction pathway of the KvLQT1/KCNE1 channel. Extensive mutagenesis of minK (23Takumi T. Moriyoshi K. Aramori I. Ishii T. Oiki S. Okada Y. Ohkubo H. Nakanishi S. J. Biol. Chem. 1991; 2266: 22192-22198Abstract Full Text PDF Google Scholar) showed that the membrane proximal C terminus of minK is least tolerant to changes, suggesting that this segment is critical to the function of the protein. Support for this finding came from biochemical assays showing that the C terminus of minK may interact with the pore region of KvLQT1 (33Romey G. Attali B. Chouabe C. Abitbol I. Guillemare E. Barhanin J. Lazdunski M. J. Biol. Chem. 1997; 272: 16713-16716Abstract Full Text Full Text PDF PubMed Scopus (122) Google Scholar). Recent work has provided further support to the location of minK in close proximity to the S6 segment of KvLQT1 near cysteine 331 (27Tapper A.R. George A.L., Jr. J. Biol. Chem. 2001; 276: 38249-38254Abstract Full Text Full Text PDF PubMed Google Scholar). Because precise spacing of the hydroxyl group at position 58 of minK was critical, it is tempting to hypothesize that a specific hydrogen bond occurs at this residue with KvLQT1. By aligning putative transmembrane segments of minK and KvLQT1, it appears that two residues within the channel could provide possible interacting hydrogen bonds with Thr-58 of minK, Tyr-268 in S5, and Ser-339 in S6. As such, the transmembrane domain would be in a position to interact with much of the channel machinery thought to be involved in channel gating including the S5, S6 segments, and the KvLQT1 pore itself. Other possible interpretations are still possible such as Thr-58 facing a water-filled cavity or hydrogen bonding along the carbonyl backbone. Our work characterizes a small region of the KCNEs that is essential for specific modulation of activation kinetics of the KvLQT1 channel. Further investigations into the nature of the interaction of KvLQT1 and this region will allow a further understanding of the control of voltage-gated potassium channels by KCNEs and a better understanding of potassium channel-gating mechanisms. Our work, which localizes one specific role of KCNE control of channel gating to a highly localized region of the protein and investigates the role of various residues at this site, sheds new light on the potential mechanism of the gating of these channels. Furthermore, the identification and characterization of such a region that is so critical to controlling gating kinetics provide a valuable structural reference point for further study of K+ channel gating. We thank Mr. Michael W. Rajala and Catherine Newnham for technical advice concerning oligonucleotide purification as well as Dr. Nancy Carrasco for helpful suggestions.
Referência(s)