Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Identification of Specific Pore Residues Mediating KCNQ1 Inactivation

Identification of Specific Pore Residues Mediating KCNQ1 Inactivation THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 276, No. 17, Issue of April 27, pp. 13600 –13605, 2001 © 2001 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Identification of Specific Pore Residues Mediating KCNQ1 Inactivation A NOVEL MECHANISM FOR LONG QT SYNDROME* Received for publication, September 13, 2000, and in revised form, January 11, 2001 Published, JBC Papers in Press, January 17, 2001, DOI 10.1074/jbc.M008373200 Guiscard Seebohm, Constanze R. Scherer, Andreas E. Busch, and Christian Lerche‡ From the Aventis Pharma Deutschland GmbH, DG Cardiovascular Diseases, Building H821, D-65926 Frankfurt am Main, Germany Inactivation plays an important physiological role such as KCNQ1 inactivation bears electrophysiological char- acteristics different from classical N- and C-type inacti- determining the sodium spike in neurons and myocytes. Ge- vation in Shaker-like potassium channels. However, the netic defects, resulting in impaired inactivation of sodium molecular site of KCNQ1 inactivation has not yet been channels, can cause myotonia and a form of long QT syndrome determined. KCNQ2 channels do not exert a fast inacti- (LQTS) (6, 7). vation in contrast to KCNQ1 channels. By expressing The KCNQ gene family represents a group of recently iden- functional chimeras between KCNQ1 and KCNQ2 in Xe- tified voltage-gated potassium channels, and disease-causing nopus oocytes, we mapped the region of this inactiva- mutations have been identified in 4 of 5 known KCNQ genes tion to transmembrane domain S5 and the pore loop H5 (8 –13). KCNQ1 (formerly called KvLQT1), the founding mem- and finally narrowed down the site to positions Gly ber of this family (8), coassembles with MinK (also called IsK or and Val in KCNQ1. Exchanging these two amino acids KCNE1) protein (14) generating slowly activating potassium individually with the analogous KCNQ2 residue abol- currents. These constitute the cardiac I conductance, one Ks ished inactivation. Furthermore, a KCNQ1-like inactiva- component of the delayed rectifier repolarizing current I (15– tion was introduced into KCNQ2 by mutagenesis in the 17). Mutations in either gene cause LQTS, whereby LQT1 corresponding region, confirming its relevance for the mutations occur in KCNQ1 and LQT5 mutations in MinK (8, inactivation process. As KCNQ1 inactivation involves 18 –21). the regions S5 and H5, it exhibits a geography distinct Homomeric KCNQ1 channels are characterized by fast acti- from N- or C-type inactivation. Native cardiac I chan- Ks vation and delayed inactivation. Delay of inactivation was ex- nels comprising KCNQ1 and accessory MinK subunits plained by two open states separating closed and inactivated do not inactivate because of the functional interaction states in a linear gating scheme. Inactivation of KCNQ1 is of KCNQ1 with MinK. Mutations in KCNQ1 can lead incomplete, supposedly resulting from a weak voltage sensitiv- to long QT1 syndrome, an inherited form of arrhyth- mia. The long QT1 mutant KCNQ1(L273F) displays a ity of the rates of both onset of and recovery from inactivation. pronounced KCNQ1 inactivation. Here we show that Alternatively, the partial inactivation can also be caused by when expressing mutant I channels formed from subconductance states. The lack of sensitivity to high extracel- Ks KCNQ1(L273F) and MinK, MinK association no longer lular TEA and potassium, together with delayed onset distin- eliminates KCNQ1 inactivation. This results in smaller guish KCNQ1 inactivation from classical C-type inactivation in repolarizing currents in the heart and therefore repre- Shaker-like channels. The weak voltage sensitivity of KCNQ1 sents a novel mechanism leading to long QT syndrome. inactivation and the lack of a Shaker-like ball structure within KCNQ1 differ from N-type inactivation (22, 23). In contrast to the existing biophysical data on KCNQ1 inac- Ion channels regulate the membrane potential of excitable tivation, its molecular determinants have only been discussed cells. These are proteins containing aqueous pores and undergo tentatively so far. Franqueza et al. (24) analyzed three different conformational changes leading to the defined gating states LQT1 mutations located within or close to the intracellular “open” and “closed,” where open channels are conducting and linker between the S4 and S5 transmembrane domains. After closed channels are nonconducting. In many channels there are expression in Xenopus oocytes, some mutants showed overall additional “inactivated” states, which are also nonconducting altered gating characteristics compared with KCNQ1-WT, but and which follow the open states because of an activating in none of them inactivation was abolished. The LQT1 muta- physiological stimulus. Among voltage-gated potassium chan- tion L273F (8) exhibited a pronounced macroscopic inactivation nels, there are two major types of inactivation. N-type inacti- when expressed in Xenopus oocytes (19). (The position number vation is mediated by an intracellular “ball,” located within the according to Shalaby et al. (19) is L272F.) The mutation is N terminus of either the pore-forming protein or a modulatory located in transmembrane segment S5, indicating a possible b-subunit, which plugs the ion channel pore (1– 4). In C-type involvement of the S5 region in channel inactivation. inactivation the outer vestibule of the pore itself undergoes Inactivation of KCNQ1 is abolished by coassembly with conformational changes (5). MinK (22, 23), suggesting that interacting domains of these two proteins might also influence the inactivation process it- self. Pusch et al. (25) have presented evidence for MinK inter- * The costs of publication of this article were defrayed in part by the action with the outer vestibule of KCNQ1. MinK also might payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ To whom correspondence should be addressed. Tel.: 33-673-846986; The abbreviations used are: LQTS, long QT syndrome; LQT, long Fax: 49-69-305-16393; E-mail: [email protected]. QT; TEA, tetraethylammonium; WT, wild type. 13600 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Geography of KCNQ1 Inactivation 13601 directly interact with the pore region of KCNQ1 as suggested mutated in neonatal epilepsy (9, 10, 31). As shown in Fig. 1, by both a yeast two-hybrid study and a cysteine-scanning study KCNQ2 did not inactivate like KCNQ1. In a gain of function in MinK (26, 27). approach, we transferred the ability to inactivate from KCNQ1 It was our aim to gain insights into the as yet unknown to KCNQ2. Substituting the S5/H5/S6 domain from KNCQ1 molecular mechanism underlying KCNQ1 inactivation and its into KCNQ2, caused a dramatic change in the biophysical possible pathophysiological relevance. characteristics of the resulting construct (Q2S5-S6Q1) com- pared with KCNQ2-WT (Fig. 1). Macroscopic inactivation is EXPERIMENTAL PROCEDURES conferred at positive voltages, suggesting that this part of the Molecular Biology—For the construction of chimeras, silent point protein is necessary for the inactivation process. mutations were introduced producing restriction endonuclease sites at By contrast it was possible to substitute large segments of corresponding positions in KCNQ1 and KCNQ2, namely a SacI site (at KCNQ2 into KCNQ1 without abolishing inactivation. A 261 262 KCNQ1 amino acids Glu -Leu ), an NsiI site (at KCNQ1 amino KCNQ1 construct including substituted KCNQ2 amino acids 301 302 303 acids Asp -Ala -Leu ), and a BamHI-site (at KCNQ1 amino acids 348 349 from the end of H5 to the end of S6 (Q1S6Q2, Fig. 1) main- Gly -Ser ). Other chimeric joining regions were created by recombi- tained inactivation. Interestingly the KCNQ2 S6 region some- nant polymerase chain reaction resulting in constructs described in Fig. 1. how enhances intrinsic KCNQ1 inactivation. Another KCNQ1 Site-directed mutagenesis was performed by polymerase chain reac- chimeric construct containing the KCNQ2 S5/H5 linker (Q1S5/ tion using cloned Pyrococcus furiosus DNA polymerase (Stratagene). All H5linkerQ2) revealed that exchanging this region does not constructs were reconfirmed by automated DNA sequencing. For Xeno- remove inactivation (Fig. 1). Thus the borders of the crucial pus oocyte expression, capped cRNA was synthesized using the SP6 and TM region for inactivation could be assigned to the S5 segment and T7 mMessage mMachine kits (Ambion). GenBank /EBI accession parts of the H5 pore loop. numbers for sequences used are as follows: hKCNQ1, AJ006345; hKCNQ2, NM004518; and MinK(KCNE1), M26685. Fig. 1A illustrates differences in the protein sequences of the Two Electrode Voltage Clamp Technique—Xenopus laevis oocytes related KCNQ channels in the described regions. With site- were obtained from tricaine-anesthetized animals. Ovaries were colla- directed mutagenesis, we exchanged the different amino acids genase-treated (1 mg/ml, Worthington, type II) in OR2 solution (NaCl in KCNQ1 to the corresponding residues in KCNQ2. Expres- 82.5 mM,KCl2mM, MgCl 1mM, HEPES 5 mM, pH 7.4) for 120 min and sion of these point mutants showed that two single amino acid subsequently stored in recording solution ND96 (NaCl 96 mM, KCl 2 substitutions, G272C and V307L, are capable of abolishing mM, CaCl 1.8 mM, MgCl 1mM, HEPES 5 mM, pH 7.4) with additional 2 2 sodium pyruvate (275 mg/liter), theophylline (90 mg/liter) and Genta- inactivation (Fig. 1). Mutation G272C, adjacent to the site of mycin (50 mg/liter) at 18 °C. Oocytes were individually injected with 10 clinical mutation L273F (as described above), caused addition- ng of cRNA encoding WT or mutant KCNQ subunits, or coinjected with ally slightly slowed activation kinetics. To exclude the possibil- 10 ng of cRNA of WT or mutant KCNQ1 and 5 ng of hMinK cRNA. ity that the introduced cysteine might be stabilizing the chan- Standard two electrode voltage clamp recordings were performed at nel via formation of novel intramolecular disulfide bonds, we 22 °C with a Turbo Tec 10CX (NPI) amplifier, an ITC-16 interface substituted glycine with threonine at this position. Now, acti- combined with Pulse software (Heka) and Origin version 5.0 (Microcal Software) for data acquisition. Macroscopic currents were recorded 3– 4 vation and deactivation kinetics of G272T were very similar to days after injection. The pipette solution contained 3 M KCl. All fitting KCNQ1-WT (Table I); however inactivation still was com- procedures were based on the simplex algorithm. Student’s t test was pletely abolished. The other key mutation V307L located in the used to test for statistical significance, which was assumed with p , pore loop showed slightly slowed activation and deactivation 0.05. kinetics compared with KCNQ1-WT (Table I), but again abol- To calculate the fraction of inactivated channels, a double exponen- ished inactivation. The slight changes in activation kinetics are tial fit to the tail currents was done according to the formula: I(t) 5 A 3 exp(2t/t ) 2 A 3 exp(2t/t ). The faster and the slower component not reflected by altered current-voltage relationships and thus s f f represent recovery from inactivation and deactivation respectively, shifted voltage dependence of activation cannot account for the whereby A and t are the amplitude and time constant for the slow s s lack of inactivation. A further observation was that in the component, and A and t are the amplitude and time constant for the f, f noninactivating mutants, the slow component of deactivation fast component. The fraction of inactivated channels is given by A /A , f s was decreased or abolished (Table I). Possibly the slow compo- where the amplitude A is related to the degree of activation, and the nent of deactivation in KCNQ1 is somehow connected to the amplitude A indicates the degree of inactivation. The method was previously described in detail (22, 23). inactivation process. To corroborate our results, we made additional point muta- RESULTS tions in the inactivating KCNQ2 chimera Q2S5-S6Q1. Analo- Inactivation in KCNQ1 channels becomes apparent in a hook gous substitution of glycine 272 in S5 (KCNQ1 numbering) and in the tail current when repolarizing the membrane potential valine 307 in H5 (KCNQ1 numbering) by cysteine and leucine, after a depolarizing pulse (Refs. 22, 23, and Fig. 1D). This trait respectively, again abolished inactivation in this chimeric con- is also described for HERG channels where it is even more struct. Effects on activation and deactivation kinetics were only pronounced (28, 29). The hook is attributed to rapid recovery of minor (Fig. 2, A and B). channels from inactivation at a rate much faster than deacti- Using a final approach, we attempted to introduce a KCNQ1- vation (Fig. 1D). Onset of KCNQ1 inactivation can be revealed like inactivation into KCNQ2 through single amino acid ex- using a double-pulse protocol (23, 29, 30). A conditioning 2-s changes. We constructed the KCNQ2 mutations C242G, pulse to 40 mV is applied to activate and inactivate channels L243F, and L272V. These positions correspond to the residues 273 273 307 followed by a 20-ms hyperpolarizing interpulse, which tran- Gly , Leu , and Val in KCNQ1, which we recognized as siently removes inactivation. During a final test pulse to 40 being important for KCNQ1 inactivation. Introducing a phen- mV, onset of inactivation becomes visible (Fig. 1C). Throughout ylalanine at position 243 in KCNQ2 resulted in inactivating this study we considered both the hook and the onset of inac- currents interestingly similar to the KCNQ1 mutant L273F tivation as a marker for the presence or absence of inactivation (Fig. 3A). The mutant KCNQ2(L272V) did not exert a KCNQ1- in mutant KCNQ1 channels. like inactivation as shown in Fig. 3B. In a construct containing To define the structural determinants of KCNQ1 inactiva- both mutations L243F and L272V, an inactivation was appar- tion, we constructed a series of functional chimeras between ent although the currents were small in amplitude (Fig. 3C). the closely related channels KCNQ1 and KCNQ2. KCNQ2 is a Substituting the cysteine at KCNQ2 position 242 by glycine component of the neuronal noninactivating M-current and is individually or together with the mutations described above 13602 Geography of KCNQ1 Inactivation FIG.1. Chimera and point mutants identifying regions of importance for KCNQ1 inactivation. A, protein sequence alignment of hKCNQ1, hKCNQ2, and rKCNQ3 showing the important regions for KCNQ1 inactivation. Postulated transmembrane segments S5 and S6 and the inner pore loop H5 are indicated above. Gray bars indicate the KCNQ2 spanning sequence in the Q1S6Q2 and Q1S5/H5linkerQ2 chimeras. These amino acids had a low impact on the inactivation process as shown below in B, C, and D. Differences in S5 and H5 of KCNQ1 and KCNQ2/KCNQ3 are marked by arrows. Amino acids that can be exchanged in KCNQ1 without affecting inactivation are marked by gray arrowheads and those that abolish inactivation by black arrowheads. Point mutants V308I and V310L and a double mutant T365A/L366W were still inactivating (data not shown). KCNQ3 forms heteromers with KCNQ2, which are not inactivating (13). Asterisk indicates the site of the LQT1 mutant L273F. B, model of important chimeras and mutants showing the typical backbone of a voltage-dependent potassium channel with the intracellular C and N terminus and six transmembrane domains. Black and gray lines indicate KCNQ1 and KCNQ2 sequences, respectively. C, representative current traces recorded in an oocyte injected with cRNA from the respective construct shown in B. A double-pulse protocol was used to show the onset or lack of inactivation. After stepping the membrane potential for 2 s from 2100 mV to 40 mV, a second 1-s pulse to 40 mV was applied following different 20-ms interpulses (to 2120 mV, 290 mV, and 260 mV) abrogating inactivation. The insets magnify the important regions. D, representative current traces at different voltages recorded by stepping the membrane potential for 3 s from 2100 mV to 50 mV in 10 Geography of KCNQ1 Inactivation 13603 TABLE I Activation and deactivation time constants of WT-KCNQ1 and mutant KCNQ1 channels Oocytes were injected with 10 ng of cRNA. Currents were activated by a 3-s pulse to 40 mV and followed by a hyperpolarizing pulse to 260 mV. Kinetics were analyzed by exponential fits to the rising phase for activation and to the tail current for deactivation. Values are mean 6 S.E., numbers of oocytes are shown in parenthesis. Constants WT-KCNQ1 KCNQ1(G272C) KCNQ1(G272T) KCNQ1(V307L) Activation, t 33 6 3 ms (9) 103 6 7 ms (9) 41 6 2 ms (8) 76 6 4ms(9) fast Activation, t 0.85 6 0.09 s (9) 0.68 6 0.05 s (9) 0.27 6 0.0 s (8) 0.67 6 0.06 s (9) slow Deactivation, t 222 6 2 ms (9) 267 6 18 ms (9) 238 6 4 ms (8) 420 6 30 ms (9) fast Deactivation, t 1.7 6 0.4 s (9) — 6.5 6 0.1 s (8) — slow FIG.2. Point mutants abolishing KCNQ1 inactivation in the inactivating chimera Q2S5-S6Q1. A and B, model of the chimera Q2S5-S6Q1 including introduced mutations G272C (A) and V307L (B). Black and gray lines indicate KCNQ1 and KCNQ2 sequences, respectively. Double-pulse protocol as shown in Fig. 1 demonstrate lack of inactivation in both point mutations shown. Right, current traces at different voltages of oocytes injected with cRNA of each construct; protocols as described in Fig. 1. Insets show magnified tail currents. The respective IV relationship is shown plotted from 10 oocytes for each construct. Error bars indicate S.E. All vertical bars represent 2 mA, horizontal bars, 0.5 s. resulted in a mutant that did not produce measurable currents activation of KCNQ1(L273F) as well as a reduced current am- under our conditions. plitude of KCNQ1(L273F)/MinK compared with KCNQ1-WT/ The only crystal structure of the pore-forming part of a MinK. We expressed homomeric mutant and heteromeric potassium channel presently available derives from the KcsA mutant/MinK channels and compared them to KCNQ1-WT and channel of Streptomyces lividans (32). In this structure, we I channels (Fig. 4). The greatly enhanced inactivation of Ks checked the locations of the residues equivalent to those that KCNQ1(L273F) is shown in Fig. 3B. Tail current analysis after we identified as being involved in KCNQ1 inactivation. Inter- a 3-s preconditioning pulse to 40 mV resulted in 86 6 3% 40 70 estingly, the amino acids in the KcsA channel, Leu and Val inactivated mutant channels in contrast to only 31 6 1% in corresponding to the clinically relevant residue KCNQ1(L273) KCNQ1-WT (Fig. 4, A and B; Table II). Around 30% inactivated and the residue KCNQ1(V307) respectively, are in close vicin- WT channels were also calculated previously using the same ity within one subunit. The closest distances of these amino analysis (22, 23). Time courses of inactivation and recovery acids are around 4 Å for all four subunits, thereby enabling were significantly affected compared with WT channels as well attractive interactions. Within the S5/H5/S6 region, KCNQ1 (Table II). After coinjecting the same amounts of KCNQ1-WT shares about 36% identity and 46% homology to KcsA. Accord- or KCNQ1(L273F) cRNA together with MinK cRNA, we ob- ingly the same residues in KCNQ1 can also be expected to be in served a decreased current amplitude for mutant I channels Ks close proximity. This raises the possibility that intramolecular consistent with previous data (19) (Fig. 3, C and D). The cur- interaction of transmembrane segment S5 and the helical part rent-voltage relationship of activation was not shifted in mu- of the pore loop H5 is involved in the inactivation process, tant I channels compared with WT channels (Fig. 4E). Most Ks putatively stabilizing the three-dimensional conformation of interestingly, using the double-pulse protocol, we demon- the inactivated channel. Exchange of glycine at KCNQ1 posi- strated that MinK is no longer able to completely abolish the tion 272 by a side chain-containing amino acid can be antici- pronounced inactivation in KCNQ1(L273F) (Fig. 4D). More- pated to change the conformation of the encompassing S5 re- over, this inactivation in mutant I channels was accelerated Ks gion and therefore possibly disrupt interaction of amino acids compared with homomeric KCNQ1(L273F), further indicating 273 307 Leu and Val . Introduction of a KCNQ1-like inactivation functional assembly of the mutant KCNQ1 subunit with MinK into KCNQ2 by exchanging leucine 243 with the more exten- (Table II). Tail currents of KCNQ1(L273F)/MinK channels dis- sive phenylalanine in S5 supports the hypothesis of interaction played a weak hook confirming the persistence of inactivation of domains S5 and H5 facilitating KCNQ1 inactivation. in these heteromers (Fig. 4D). By analyzing this hook, we We next examined the pathophysiological role of KCNQ1 calculated that more than 20% of mutant I channels were in Ks inactivation by studying kinetics of the LQT1 mutant L273F, the inactivated state after a 3-s activating pulse to 40 mV (Fig. which is located next to Gly and in the KcsA channel model 4F). Thus, the significant number of inactivated, i.e. noncon- in close proximity to Val . The mutant was analyzed previ- ducting, mutant I channels contributes to the decreased am- Ks ously (19); the authors reported a pronounced macroscopic in- plitudes of KCNQ1(L273F) and KCNQ1(L273F)/MinK com- mV increments in an oocyte injected with cRNA of the respective construct shown in B. Tail currents show a characteristic hook reflecting recovery from inactivation. The insets magnify the important regions. E, IV relationship of constructs shown in B recorded as described in D (n 5 10 –16). Amplitudes at the end of 3-s test pulses were plotted against voltage. Error bars indicate S.E. All vertical bars represent 2 mA; horizontal bars, 0.5 s. 13604 Geography of KCNQ1 Inactivation FIG.3. Introduction of fast inactivation into KCNQ2 by point mutation. A, B, and C, models of the KCNQ2 channel subunit including introduced mutations L243F (A), L272V (B), and L243F/L272V are drawn on the left side (C). Double-pulse protocols as performed in Fig. 1 were used demonstrating existence or lack of inactivation in the point mutants. Current traces of oocytes injected with cRNA of each construct were measured at different voltages with protocols as described in Fig. 1. The insets show magnified important regions. The respective IV relationships are shown plotted from 5–7 oocytes for each construct. Error bars indicate S.E. All vertical bars represent 1 mA, horizontal bars, 0.5 s. FIG.4. Characterization of the LQT1 mutant L273F in oocytes. A, representative current traces at different voltages recorded as described in Fig. 1 in an oocyte injected with KCNQ1-WT cRNA. Double-pulse protocol as described in Fig. 1 was used to show the onset of inactivation. The insets magnify the important regions. B, current traces of a L273F-coinjected oocyte recorded as described in A. C, current traces at different voltages of a KCNQ1-WT/MinK coinjected oocyte recorded as described in A. Double-pulse protocol of a KCNQ1-WT/MinK coinjected oocyte recorded as described in A. The insets magnify important regions. D, current traces of a L273F/MinK-coinjected oocyte recorded as described in A. E, IV relationship of mutant L273F/MinK (n 5 14) and of KCNQ1-WT/MinK (n 5 17) from oocytes recorded as in B and D. Amplitudes after 3-s test pulses were plotted against voltage. F, The fraction of inactivated channels of mutant L273F/MinK and of KCNQ1WT/MinK was determined by tail current analysis (double exponential fit) and calculated by the amplitudes of recovery from inactivation (A ) and deactivation (A ) (see fast slow “Experimental Procedures”). All vertical bars represent 2 mA, horizontal bars, 0.5 s. pared with those of KCNQ1-WT and KCNQ1/MinK. In vivo this In classical C-type inactivation, conformational changes in is expected to result in a decreased repolarizing I conduct- the outer vestibule of the pore are coupled to the inactivation Ks ance constituting a new mechanism leading to LQT1 syndrome. process. This is reflected in the dependence of the inactivation to extracellular potassium and TEA (33, 34). In KCNQ1, TEA DISCUSSION sensitivity of the inactivation rate is not determinable because Inactivation is best understood and intensively studied for of very low TEA affinity to the channel (22, 23, 35). However, classical N- and C-type mechanisms in ShakerB potassium KCNQ1 inactivation is independent of extracellular potassium channels. Fast N-type inactivation is realized by a ball plug-in concentration; a principal difference from C-type behavior. In mechanism, where a ball-forming domain of the channel binds Shaker-like potassium channels, certain mutations within the to its receptor thereby blocking the ion pathway from the in- P-region H5 and the sixth membrane-spanning region S6 have tracellular side (1–5). This occurs independently of the mem- been shown to affect the time course of inactivation (33, 36). brane potential. KCNQ1 inactivation exerts slower inactiva- Notably, also in our study KCNQ1 inactivation was enhanced tion gating and a weak voltage dependence. Further stressing by modification of this region (Q1S6Q2), but identified residues the differences from the N-type mechanism, the identified res- essential for KCNQ1 inactivation within the S5 region and the idues at KCNQ1 positions 272, 273, and 307 are not expected to helical part of the H5 segment have not been reported to be be accessible for a complex ball structure from the intracellular side and thus probably cannot account for a potential receptor substantially relevant for C-type inactivation. In conclusion, site. both C-type and KCNQ1 inactivation occur within the pore Geography of KCNQ1 Inactivation 13605 TABLE II REFERENCES Characteristics of inactivation in KCNQ1-WT and KCNQ1(L273F) in 1. Armstrong, C. M., and Bezanilla, F. (1977) J. Gen. Physiol. 70, 567–590 homomeric and heteromeric expression together with MinK in 2. Zagotta, W. N., Hoshi, T., and Aldrich, R. W. (1990) Science 250, 568 –571 Xenopus oocytes 3. Hoshi, T., Zagotta, W. N., and Aldrich, R. W. (1990) Science 250, 533–538 4. Rettig, J., Heinemann, S. H., Wunder, F., Lorra, C., Parcej, D. N., Dolly, J. O., Oocytes were injected with 10 ng of KCNQ1 cRNA. For heteromeric and Pongs, O. (1994) Nature 369, 289 –294 expression 5 ng of MinK cRNA were added. The double-pulse protocol 5. Liu, Y., Jurman, M. E., and Yellen, G. (1996) Neuron 16, 859 – 867 with two 40-mV pulses separated by a 20-ms pulse to 2120 mV was 6. Lerche, H., Heine, R., Pika, U., George, A. L., Jr., Mitrovic, N., Browatzki, M., used to analyze the onset of inactivation. The time constants of onset Weiss, T., Rivet-Bastide, M., Franke, C., Lomonaco, M., Ricker, K., and were calculated by exponential fits to the decaying phase of the currents Lehmann-Horn, F. (1993) J. Physiol. 470, 13–22 at the beginning of the second pulse. To determine recovery from 7. Bennett, P. B., Yazawa, K., Makita, N., and George, A. L., Jr. (1995) Nature inactivation and the fraction of inactivated channels, currents were 376, 683– 685 activated by a 3-s pulse to 40 mV followed by a hyperpolarizing pulse to 8. Wang, Q., Curran, M. E., Splawski, I., Burn, T. C., Millholland, J. M., Van Raay, T. J., Shen, J., Timothy, K. W., Vincent, G. M., De Jager, T., Schwarz, 260 mV. Tail currents were analyzed by double exponential fits. The P. J., Towbin, J. A., Moss, A. J., Atkinson, D. L., Landes, G. M., Connors, faster and the slower component constitute recovery from inactivation T. D., and Keating, M. T. (1996) Nat. Genet. 12, 17–23 and deactivation respectively. Values are mean 6 S.E., numbers of 9. Charlier, C., Singh, N. A., Ryan, S. G., Lewis, T. B., Reus, B. E., Leach, R. J., oocytes are shown in parenthesis. and Leppert, M. (1998) Nat. Genet. 18, 53–55 10. Singh, N. A., Charlier, C., Stauffer, D., DuPont, B. R., Leach, R. J., Melis, R., Inactivated t t onset recovery Ronen, G. M., Bjerre, I., Quattlebaum, T., Murphy, J. V., McHarg, M. L., channels Gagnon, D., Rosales, T. O., Peiffer, A., Anderson, V. E., and Leppert, M. ms % (1998) Nat. Genet. 18, 25–29 11. Biervert, C., Schroeder, B. C., Kubisch, C., Berkovic, S. F., Propping, P., WT-KCNQ1 15 6 1(n 5 12) 45 6 3(n 5 13) 31 6 1(n 5 12) a Jentsch, T. J., and Steinlein, O. K. (1998) Science 279, 403– 406 KCNQ1(L273F) 60 6 6(n 5 6) 98 6 5(n 5 6) 86 6 3(n 5 6) 12. Kubisch, C., Schro ¨ der, B. C., Friedrich, T., Lu ¨ tjohann, B., El Amraoui, A., WT-KCNQ1/ —— 0 Marlin, S., Petit, C., and Jentsch, T. J. (1999) Cell 96, 437– 446 MinK 13. Lerche, C., Scherer, C. R., Seebohm, G., Derst, C., Wei, A. D., Busch, A. E., and KCNQ1(L273F)/ 26 6 2(n 5 8) 111 6 3(n 5 8) 24 6 5(n 5 12) Steinmeyer, K. (2000) J. Biol. Chem. 29, 22395–22400 MinK 14. Takumi, T., Ohkubo, H., and Nakanishi, S. (1988) Science 242, 1042–1045 15. Sanguinetti, M. C., Curran, M. E., Zou, A., Shen, J., Spector, P. S., Atkinson, Significance of homomeric KCNQ1-WT versus KCNQ1(L273F). D. L., and Keating, M. T. (1996) Nature 384, 80–83 Significance of homomeric KCNQ1(L273F) versus heteromeric 16. Barhanin, J., Lesage, F., Guillemare, E., Fink, M., Lazdunski, M., and Romey, KCNQ1(L273F)/MinK. G. (1997) Nature 384, 78–80 17. Attali, B. (1996) Nature 384, 24 –25 region spanning S5 to S6, but the defined critical structures 18. Chouabe, C., Neyroud, N., Guicheney, P., Lazdunski, M., Romey, G., and Barhanin, J. (1997) EMBO J. 16, 5472–5479 appear to be different. 19. Shalaby, F. Y., Levesque, P. C., Yang, W. P., Little, W. A., Conder, M. L., In summary our results define geographical peculiarities for Jenkins-West, T., and Blanar, M. A. (1997) Circulation 96, 1733–1736 20. Wollnik, B., Schroeder, B. C., Kubisch, C., Esperer, H. D., Wieacker, P., and KCNQ1 inactivation, which were not yet reported for classical Jentsch, T. J. (1997) Hum. Mol. Gen. 6, 1943–1949 N- and C-type inactivation in Shaker-like potassium channels. 21. Schulze-Bahr, E., Wang, Q., Wedekind, H., Haverkamp, W., Chen, Q., and KCNQ1 inactivation is characterized by an involvement of the Sun, Y. (1997) Nat. Genet. 17, 267–268 22. Pusch, M., Magrassi, R., Wollnik, B., and Conti, F. (1998) Biophys. J. 75, pore region H5 and the S5 transmembrane region, putatively 785–792 interplaying with each other. It is believed that KCNQ1 inac- 23. Tristani-Firouzi, M., and Sanguinetti, M. C. (1998) J. Physiol. (Lond.) 510, 37– 45 tivation is abolished by interaction with MinK in native het- 24. Franqueza, L., Lin, M., Shen, J., Splawski, I., Keating, M. T., and Sanguinetti, eromeric I channels (22, 23). As suggested previously, inter- Ks M. C. (1999) J. Biol. Chem. 274, 21063–21070 action between KCNQ1 and MinK might involve the pore 25. Pusch, M., Bertorello, L., and Conti, F. (2000) Biophys. J. 78, 211–226 26. Romey, G., Attali, B., Chouabe, C., Abitbol, I., Guillemare, E., Barhanin, J., region of KCNQ1 (26, 27). Our results demonstrate an essential and Lazdunski, M. (1997) J. Biol. Chem. 272, 16713–16716 function of the pore region for KCNQ1 inactivation, therefore 27. Tai, K. K., and Goldstein, S. A. (1998) Nature 391, 605– 608 allowing the interesting possibility that MinK interaction may 28. Sanguinetti, M. C., Jiang, C., Curran, M.-E., and Keating, M. T. (1995) Cell 81, 299 –307 occur in the same region that is important for inactivation. 29. Smith, P. L., Baukrowitz, T., and Yellen, G. (1996) Nature 379, 833– 836 MinK possibly abolishes inactivation by weakening the inter- 30. Schonherr, R., and Heinemann, S. H. (1996) J. Physiol. (Lond.) 493, 635– 642 31. Wang, H. S., Pan, Z., Shi, W., Brown, B. S., Wymore, R. S., Cohen, I. S., Dixon, play between H5 and S5. In the LQT1 mutant L273F this J. E., and McKinnon, D. (1998) Science 282, 1890 –1893 interplay might be stronger, thereby stabilizing inactivation 32. Doyle, D. A., Morais, C.abral, J., Pfuetzner, R. A., Kuo, A., Gulbis, J. M., such that MinK no longer can abolish inactivation. This in turn Cohen, S. L., Chait, B. T., and MacKinnon, R. (1998) Science 280, 69 –77 33. LopezBarneo, J., Hoshi, T., Heinemann, S. H., and Aldrich, R. W. (1993) may contribute to the deleterious phenotype of the LQT1 Recept. Channel 1, 61–71 mutant. 34. Choi, K. L., Aldrich, R. W and Yellen, G. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 5092–5095 Acknowledgments—We thank K. Steinmeyer, H. Lerche, B. Attali, P. 35. Hadley, J. K., Noda, M., Selyanko, A. A., Wood, I. C., Abogadie, F. C., and Ruppersberg, A. Wei, I. Gutcher, and E. Kostenis for fruitful discus- Brown, D. A. (2000) Br. J. Pharmacol. 129, 413– 415 sions and careful proofreading. 36. Hoshi, T., Zagotta, W. N., and Aldrich, R. W. (1991) Neuron 7, 547–556 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry Unpaywall

Identification of Specific Pore Residues Mediating KCNQ1 Inactivation

Seebohm, Guiscard; Scherer, Constanze R.; Busch, Andreas E.; Lerche, Christian
Journal of Biological Chemistry – Apr 1, 2001
Download PDF
6 pages

Loading next page...
 
/lp/unpaywall/identification-of-specific-pore-residues-mediating-kcnq1-inactivation-t94fGECvdy

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
0021-9258
DOI
10.1074/jbc.m008373200
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 276, No. 17, Issue of April 27, pp. 13600 –13605, 2001 © 2001 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Identification of Specific Pore Residues Mediating KCNQ1 Inactivation A NOVEL MECHANISM FOR LONG QT SYNDROME* Received for publication, September 13, 2000, and in revised form, January 11, 2001 Published, JBC Papers in Press, January 17, 2001, DOI 10.1074/jbc.M008373200 Guiscard Seebohm, Constanze R. Scherer, Andreas E. Busch, and Christian Lerche‡ From the Aventis Pharma Deutschland GmbH, DG Cardiovascular Diseases, Building H821, D-65926 Frankfurt am Main, Germany Inactivation plays an important physiological role such as KCNQ1 inactivation bears electrophysiological char- acteristics different from classical N- and C-type inacti- determining the sodium spike in neurons and myocytes. Ge- vation in Shaker-like potassium channels. However, the netic defects, resulting in impaired inactivation of sodium molecular site of KCNQ1 inactivation has not yet been channels, can cause myotonia and a form of long QT syndrome determined. KCNQ2 channels do not exert a fast inacti- (LQTS) (6, 7). vation in contrast to KCNQ1 channels. By expressing The KCNQ gene family represents a group of recently iden- functional chimeras between KCNQ1 and KCNQ2 in Xe- tified voltage-gated potassium channels, and disease-causing nopus oocytes, we mapped the region of this inactiva- mutations have been identified in 4 of 5 known KCNQ genes tion to transmembrane domain S5 and the pore loop H5 (8 –13). KCNQ1 (formerly called KvLQT1), the founding mem- and finally narrowed down the site to positions Gly ber of this family (8), coassembles with MinK (also called IsK or and Val in KCNQ1. Exchanging these two amino acids KCNE1) protein (14) generating slowly activating potassium individually with the analogous KCNQ2 residue abol- currents. These constitute the cardiac I conductance, one Ks ished inactivation. Furthermore, a KCNQ1-like inactiva- component of the delayed rectifier repolarizing current I (15– tion was introduced into KCNQ2 by mutagenesis in the 17). Mutations in either gene cause LQTS, whereby LQT1 corresponding region, confirming its relevance for the mutations occur in KCNQ1 and LQT5 mutations in MinK (8, inactivation process. As KCNQ1 inactivation involves 18 –21). the regions S5 and H5, it exhibits a geography distinct Homomeric KCNQ1 channels are characterized by fast acti- from N- or C-type inactivation. Native cardiac I chan- Ks vation and delayed inactivation. Delay of inactivation was ex- nels comprising KCNQ1 and accessory MinK subunits plained by two open states separating closed and inactivated do not inactivate because of the functional interaction states in a linear gating scheme. Inactivation of KCNQ1 is of KCNQ1 with MinK. Mutations in KCNQ1 can lead incomplete, supposedly resulting from a weak voltage sensitiv- to long QT1 syndrome, an inherited form of arrhyth- mia. The long QT1 mutant KCNQ1(L273F) displays a ity of the rates of both onset of and recovery from inactivation. pronounced KCNQ1 inactivation. Here we show that Alternatively, the partial inactivation can also be caused by when expressing mutant I channels formed from subconductance states. The lack of sensitivity to high extracel- Ks KCNQ1(L273F) and MinK, MinK association no longer lular TEA and potassium, together with delayed onset distin- eliminates KCNQ1 inactivation. This results in smaller guish KCNQ1 inactivation from classical C-type inactivation in repolarizing currents in the heart and therefore repre- Shaker-like channels. The weak voltage sensitivity of KCNQ1 sents a novel mechanism leading to long QT syndrome. inactivation and the lack of a Shaker-like ball structure within KCNQ1 differ from N-type inactivation (22, 23). In contrast to the existing biophysical data on KCNQ1 inac- Ion channels regulate the membrane potential of excitable tivation, its molecular determinants have only been discussed cells. These are proteins containing aqueous pores and undergo tentatively so far. Franqueza et al. (24) analyzed three different conformational changes leading to the defined gating states LQT1 mutations located within or close to the intracellular “open” and “closed,” where open channels are conducting and linker between the S4 and S5 transmembrane domains. After closed channels are nonconducting. In many channels there are expression in Xenopus oocytes, some mutants showed overall additional “inactivated” states, which are also nonconducting altered gating characteristics compared with KCNQ1-WT, but and which follow the open states because of an activating in none of them inactivation was abolished. The LQT1 muta- physiological stimulus. Among voltage-gated potassium chan- tion L273F (8) exhibited a pronounced macroscopic inactivation nels, there are two major types of inactivation. N-type inacti- when expressed in Xenopus oocytes (19). (The position number vation is mediated by an intracellular “ball,” located within the according to Shalaby et al. (19) is L272F.) The mutation is N terminus of either the pore-forming protein or a modulatory located in transmembrane segment S5, indicating a possible b-subunit, which plugs the ion channel pore (1– 4). In C-type involvement of the S5 region in channel inactivation. inactivation the outer vestibule of the pore itself undergoes Inactivation of KCNQ1 is abolished by coassembly with conformational changes (5). MinK (22, 23), suggesting that interacting domains of these two proteins might also influence the inactivation process it- self. Pusch et al. (25) have presented evidence for MinK inter- * The costs of publication of this article were defrayed in part by the action with the outer vestibule of KCNQ1. MinK also might payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ To whom correspondence should be addressed. Tel.: 33-673-846986; The abbreviations used are: LQTS, long QT syndrome; LQT, long Fax: 49-69-305-16393; E-mail: [email protected]. QT; TEA, tetraethylammonium; WT, wild type. 13600 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Geography of KCNQ1 Inactivation 13601 directly interact with the pore region of KCNQ1 as suggested mutated in neonatal epilepsy (9, 10, 31). As shown in Fig. 1, by both a yeast two-hybrid study and a cysteine-scanning study KCNQ2 did not inactivate like KCNQ1. In a gain of function in MinK (26, 27). approach, we transferred the ability to inactivate from KCNQ1 It was our aim to gain insights into the as yet unknown to KCNQ2. Substituting the S5/H5/S6 domain from KNCQ1 molecular mechanism underlying KCNQ1 inactivation and its into KCNQ2, caused a dramatic change in the biophysical possible pathophysiological relevance. characteristics of the resulting construct (Q2S5-S6Q1) com- pared with KCNQ2-WT (Fig. 1). Macroscopic inactivation is EXPERIMENTAL PROCEDURES conferred at positive voltages, suggesting that this part of the Molecular Biology—For the construction of chimeras, silent point protein is necessary for the inactivation process. mutations were introduced producing restriction endonuclease sites at By contrast it was possible to substitute large segments of corresponding positions in KCNQ1 and KCNQ2, namely a SacI site (at KCNQ2 into KCNQ1 without abolishing inactivation. A 261 262 KCNQ1 amino acids Glu -Leu ), an NsiI site (at KCNQ1 amino KCNQ1 construct including substituted KCNQ2 amino acids 301 302 303 acids Asp -Ala -Leu ), and a BamHI-site (at KCNQ1 amino acids 348 349 from the end of H5 to the end of S6 (Q1S6Q2, Fig. 1) main- Gly -Ser ). Other chimeric joining regions were created by recombi- tained inactivation. Interestingly the KCNQ2 S6 region some- nant polymerase chain reaction resulting in constructs described in Fig. 1. how enhances intrinsic KCNQ1 inactivation. Another KCNQ1 Site-directed mutagenesis was performed by polymerase chain reac- chimeric construct containing the KCNQ2 S5/H5 linker (Q1S5/ tion using cloned Pyrococcus furiosus DNA polymerase (Stratagene). All H5linkerQ2) revealed that exchanging this region does not constructs were reconfirmed by automated DNA sequencing. For Xeno- remove inactivation (Fig. 1). Thus the borders of the crucial pus oocyte expression, capped cRNA was synthesized using the SP6 and TM region for inactivation could be assigned to the S5 segment and T7 mMessage mMachine kits (Ambion). GenBank /EBI accession parts of the H5 pore loop. numbers for sequences used are as follows: hKCNQ1, AJ006345; hKCNQ2, NM004518; and MinK(KCNE1), M26685. Fig. 1A illustrates differences in the protein sequences of the Two Electrode Voltage Clamp Technique—Xenopus laevis oocytes related KCNQ channels in the described regions. With site- were obtained from tricaine-anesthetized animals. Ovaries were colla- directed mutagenesis, we exchanged the different amino acids genase-treated (1 mg/ml, Worthington, type II) in OR2 solution (NaCl in KCNQ1 to the corresponding residues in KCNQ2. Expres- 82.5 mM,KCl2mM, MgCl 1mM, HEPES 5 mM, pH 7.4) for 120 min and sion of these point mutants showed that two single amino acid subsequently stored in recording solution ND96 (NaCl 96 mM, KCl 2 substitutions, G272C and V307L, are capable of abolishing mM, CaCl 1.8 mM, MgCl 1mM, HEPES 5 mM, pH 7.4) with additional 2 2 sodium pyruvate (275 mg/liter), theophylline (90 mg/liter) and Genta- inactivation (Fig. 1). Mutation G272C, adjacent to the site of mycin (50 mg/liter) at 18 °C. Oocytes were individually injected with 10 clinical mutation L273F (as described above), caused addition- ng of cRNA encoding WT or mutant KCNQ subunits, or coinjected with ally slightly slowed activation kinetics. To exclude the possibil- 10 ng of cRNA of WT or mutant KCNQ1 and 5 ng of hMinK cRNA. ity that the introduced cysteine might be stabilizing the chan- Standard two electrode voltage clamp recordings were performed at nel via formation of novel intramolecular disulfide bonds, we 22 °C with a Turbo Tec 10CX (NPI) amplifier, an ITC-16 interface substituted glycine with threonine at this position. Now, acti- combined with Pulse software (Heka) and Origin version 5.0 (Microcal Software) for data acquisition. Macroscopic currents were recorded 3– 4 vation and deactivation kinetics of G272T were very similar to days after injection. The pipette solution contained 3 M KCl. All fitting KCNQ1-WT (Table I); however inactivation still was com- procedures were based on the simplex algorithm. Student’s t test was pletely abolished. The other key mutation V307L located in the used to test for statistical significance, which was assumed with p , pore loop showed slightly slowed activation and deactivation 0.05. kinetics compared with KCNQ1-WT (Table I), but again abol- To calculate the fraction of inactivated channels, a double exponen- ished inactivation. The slight changes in activation kinetics are tial fit to the tail currents was done according to the formula: I(t) 5 A 3 exp(2t/t ) 2 A 3 exp(2t/t ). The faster and the slower component not reflected by altered current-voltage relationships and thus s f f represent recovery from inactivation and deactivation respectively, shifted voltage dependence of activation cannot account for the whereby A and t are the amplitude and time constant for the slow s s lack of inactivation. A further observation was that in the component, and A and t are the amplitude and time constant for the f, f noninactivating mutants, the slow component of deactivation fast component. The fraction of inactivated channels is given by A /A , f s was decreased or abolished (Table I). Possibly the slow compo- where the amplitude A is related to the degree of activation, and the nent of deactivation in KCNQ1 is somehow connected to the amplitude A indicates the degree of inactivation. The method was previously described in detail (22, 23). inactivation process. To corroborate our results, we made additional point muta- RESULTS tions in the inactivating KCNQ2 chimera Q2S5-S6Q1. Analo- Inactivation in KCNQ1 channels becomes apparent in a hook gous substitution of glycine 272 in S5 (KCNQ1 numbering) and in the tail current when repolarizing the membrane potential valine 307 in H5 (KCNQ1 numbering) by cysteine and leucine, after a depolarizing pulse (Refs. 22, 23, and Fig. 1D). This trait respectively, again abolished inactivation in this chimeric con- is also described for HERG channels where it is even more struct. Effects on activation and deactivation kinetics were only pronounced (28, 29). The hook is attributed to rapid recovery of minor (Fig. 2, A and B). channels from inactivation at a rate much faster than deacti- Using a final approach, we attempted to introduce a KCNQ1- vation (Fig. 1D). Onset of KCNQ1 inactivation can be revealed like inactivation into KCNQ2 through single amino acid ex- using a double-pulse protocol (23, 29, 30). A conditioning 2-s changes. We constructed the KCNQ2 mutations C242G, pulse to 40 mV is applied to activate and inactivate channels L243F, and L272V. These positions correspond to the residues 273 273 307 followed by a 20-ms hyperpolarizing interpulse, which tran- Gly , Leu , and Val in KCNQ1, which we recognized as siently removes inactivation. During a final test pulse to 40 being important for KCNQ1 inactivation. Introducing a phen- mV, onset of inactivation becomes visible (Fig. 1C). Throughout ylalanine at position 243 in KCNQ2 resulted in inactivating this study we considered both the hook and the onset of inac- currents interestingly similar to the KCNQ1 mutant L273F tivation as a marker for the presence or absence of inactivation (Fig. 3A). The mutant KCNQ2(L272V) did not exert a KCNQ1- in mutant KCNQ1 channels. like inactivation as shown in Fig. 3B. In a construct containing To define the structural determinants of KCNQ1 inactiva- both mutations L243F and L272V, an inactivation was appar- tion, we constructed a series of functional chimeras between ent although the currents were small in amplitude (Fig. 3C). the closely related channels KCNQ1 and KCNQ2. KCNQ2 is a Substituting the cysteine at KCNQ2 position 242 by glycine component of the neuronal noninactivating M-current and is individually or together with the mutations described above 13602 Geography of KCNQ1 Inactivation FIG.1. Chimera and point mutants identifying regions of importance for KCNQ1 inactivation. A, protein sequence alignment of hKCNQ1, hKCNQ2, and rKCNQ3 showing the important regions for KCNQ1 inactivation. Postulated transmembrane segments S5 and S6 and the inner pore loop H5 are indicated above. Gray bars indicate the KCNQ2 spanning sequence in the Q1S6Q2 and Q1S5/H5linkerQ2 chimeras. These amino acids had a low impact on the inactivation process as shown below in B, C, and D. Differences in S5 and H5 of KCNQ1 and KCNQ2/KCNQ3 are marked by arrows. Amino acids that can be exchanged in KCNQ1 without affecting inactivation are marked by gray arrowheads and those that abolish inactivation by black arrowheads. Point mutants V308I and V310L and a double mutant T365A/L366W were still inactivating (data not shown). KCNQ3 forms heteromers with KCNQ2, which are not inactivating (13). Asterisk indicates the site of the LQT1 mutant L273F. B, model of important chimeras and mutants showing the typical backbone of a voltage-dependent potassium channel with the intracellular C and N terminus and six transmembrane domains. Black and gray lines indicate KCNQ1 and KCNQ2 sequences, respectively. C, representative current traces recorded in an oocyte injected with cRNA from the respective construct shown in B. A double-pulse protocol was used to show the onset or lack of inactivation. After stepping the membrane potential for 2 s from 2100 mV to 40 mV, a second 1-s pulse to 40 mV was applied following different 20-ms interpulses (to 2120 mV, 290 mV, and 260 mV) abrogating inactivation. The insets magnify the important regions. D, representative current traces at different voltages recorded by stepping the membrane potential for 3 s from 2100 mV to 50 mV in 10 Geography of KCNQ1 Inactivation 13603 TABLE I Activation and deactivation time constants of WT-KCNQ1 and mutant KCNQ1 channels Oocytes were injected with 10 ng of cRNA. Currents were activated by a 3-s pulse to 40 mV and followed by a hyperpolarizing pulse to 260 mV. Kinetics were analyzed by exponential fits to the rising phase for activation and to the tail current for deactivation. Values are mean 6 S.E., numbers of oocytes are shown in parenthesis. Constants WT-KCNQ1 KCNQ1(G272C) KCNQ1(G272T) KCNQ1(V307L) Activation, t 33 6 3 ms (9) 103 6 7 ms (9) 41 6 2 ms (8) 76 6 4ms(9) fast Activation, t 0.85 6 0.09 s (9) 0.68 6 0.05 s (9) 0.27 6 0.0 s (8) 0.67 6 0.06 s (9) slow Deactivation, t 222 6 2 ms (9) 267 6 18 ms (9) 238 6 4 ms (8) 420 6 30 ms (9) fast Deactivation, t 1.7 6 0.4 s (9) — 6.5 6 0.1 s (8) — slow FIG.2. Point mutants abolishing KCNQ1 inactivation in the inactivating chimera Q2S5-S6Q1. A and B, model of the chimera Q2S5-S6Q1 including introduced mutations G272C (A) and V307L (B). Black and gray lines indicate KCNQ1 and KCNQ2 sequences, respectively. Double-pulse protocol as shown in Fig. 1 demonstrate lack of inactivation in both point mutations shown. Right, current traces at different voltages of oocytes injected with cRNA of each construct; protocols as described in Fig. 1. Insets show magnified tail currents. The respective IV relationship is shown plotted from 10 oocytes for each construct. Error bars indicate S.E. All vertical bars represent 2 mA, horizontal bars, 0.5 s. resulted in a mutant that did not produce measurable currents activation of KCNQ1(L273F) as well as a reduced current am- under our conditions. plitude of KCNQ1(L273F)/MinK compared with KCNQ1-WT/ The only crystal structure of the pore-forming part of a MinK. We expressed homomeric mutant and heteromeric potassium channel presently available derives from the KcsA mutant/MinK channels and compared them to KCNQ1-WT and channel of Streptomyces lividans (32). In this structure, we I channels (Fig. 4). The greatly enhanced inactivation of Ks checked the locations of the residues equivalent to those that KCNQ1(L273F) is shown in Fig. 3B. Tail current analysis after we identified as being involved in KCNQ1 inactivation. Inter- a 3-s preconditioning pulse to 40 mV resulted in 86 6 3% 40 70 estingly, the amino acids in the KcsA channel, Leu and Val inactivated mutant channels in contrast to only 31 6 1% in corresponding to the clinically relevant residue KCNQ1(L273) KCNQ1-WT (Fig. 4, A and B; Table II). Around 30% inactivated and the residue KCNQ1(V307) respectively, are in close vicin- WT channels were also calculated previously using the same ity within one subunit. The closest distances of these amino analysis (22, 23). Time courses of inactivation and recovery acids are around 4 Å for all four subunits, thereby enabling were significantly affected compared with WT channels as well attractive interactions. Within the S5/H5/S6 region, KCNQ1 (Table II). After coinjecting the same amounts of KCNQ1-WT shares about 36% identity and 46% homology to KcsA. Accord- or KCNQ1(L273F) cRNA together with MinK cRNA, we ob- ingly the same residues in KCNQ1 can also be expected to be in served a decreased current amplitude for mutant I channels Ks close proximity. This raises the possibility that intramolecular consistent with previous data (19) (Fig. 3, C and D). The cur- interaction of transmembrane segment S5 and the helical part rent-voltage relationship of activation was not shifted in mu- of the pore loop H5 is involved in the inactivation process, tant I channels compared with WT channels (Fig. 4E). Most Ks putatively stabilizing the three-dimensional conformation of interestingly, using the double-pulse protocol, we demon- the inactivated channel. Exchange of glycine at KCNQ1 posi- strated that MinK is no longer able to completely abolish the tion 272 by a side chain-containing amino acid can be antici- pronounced inactivation in KCNQ1(L273F) (Fig. 4D). More- pated to change the conformation of the encompassing S5 re- over, this inactivation in mutant I channels was accelerated Ks gion and therefore possibly disrupt interaction of amino acids compared with homomeric KCNQ1(L273F), further indicating 273 307 Leu and Val . Introduction of a KCNQ1-like inactivation functional assembly of the mutant KCNQ1 subunit with MinK into KCNQ2 by exchanging leucine 243 with the more exten- (Table II). Tail currents of KCNQ1(L273F)/MinK channels dis- sive phenylalanine in S5 supports the hypothesis of interaction played a weak hook confirming the persistence of inactivation of domains S5 and H5 facilitating KCNQ1 inactivation. in these heteromers (Fig. 4D). By analyzing this hook, we We next examined the pathophysiological role of KCNQ1 calculated that more than 20% of mutant I channels were in Ks inactivation by studying kinetics of the LQT1 mutant L273F, the inactivated state after a 3-s activating pulse to 40 mV (Fig. which is located next to Gly and in the KcsA channel model 4F). Thus, the significant number of inactivated, i.e. noncon- in close proximity to Val . The mutant was analyzed previ- ducting, mutant I channels contributes to the decreased am- Ks ously (19); the authors reported a pronounced macroscopic in- plitudes of KCNQ1(L273F) and KCNQ1(L273F)/MinK com- mV increments in an oocyte injected with cRNA of the respective construct shown in B. Tail currents show a characteristic hook reflecting recovery from inactivation. The insets magnify the important regions. E, IV relationship of constructs shown in B recorded as described in D (n 5 10 –16). Amplitudes at the end of 3-s test pulses were plotted against voltage. Error bars indicate S.E. All vertical bars represent 2 mA; horizontal bars, 0.5 s. 13604 Geography of KCNQ1 Inactivation FIG.3. Introduction of fast inactivation into KCNQ2 by point mutation. A, B, and C, models of the KCNQ2 channel subunit including introduced mutations L243F (A), L272V (B), and L243F/L272V are drawn on the left side (C). Double-pulse protocols as performed in Fig. 1 were used demonstrating existence or lack of inactivation in the point mutants. Current traces of oocytes injected with cRNA of each construct were measured at different voltages with protocols as described in Fig. 1. The insets show magnified important regions. The respective IV relationships are shown plotted from 5–7 oocytes for each construct. Error bars indicate S.E. All vertical bars represent 1 mA, horizontal bars, 0.5 s. FIG.4. Characterization of the LQT1 mutant L273F in oocytes. A, representative current traces at different voltages recorded as described in Fig. 1 in an oocyte injected with KCNQ1-WT cRNA. Double-pulse protocol as described in Fig. 1 was used to show the onset of inactivation. The insets magnify the important regions. B, current traces of a L273F-coinjected oocyte recorded as described in A. C, current traces at different voltages of a KCNQ1-WT/MinK coinjected oocyte recorded as described in A. Double-pulse protocol of a KCNQ1-WT/MinK coinjected oocyte recorded as described in A. The insets magnify important regions. D, current traces of a L273F/MinK-coinjected oocyte recorded as described in A. E, IV relationship of mutant L273F/MinK (n 5 14) and of KCNQ1-WT/MinK (n 5 17) from oocytes recorded as in B and D. Amplitudes after 3-s test pulses were plotted against voltage. F, The fraction of inactivated channels of mutant L273F/MinK and of KCNQ1WT/MinK was determined by tail current analysis (double exponential fit) and calculated by the amplitudes of recovery from inactivation (A ) and deactivation (A ) (see fast slow “Experimental Procedures”). All vertical bars represent 2 mA, horizontal bars, 0.5 s. pared with those of KCNQ1-WT and KCNQ1/MinK. In vivo this In classical C-type inactivation, conformational changes in is expected to result in a decreased repolarizing I conduct- the outer vestibule of the pore are coupled to the inactivation Ks ance constituting a new mechanism leading to LQT1 syndrome. process. This is reflected in the dependence of the inactivation to extracellular potassium and TEA (33, 34). In KCNQ1, TEA DISCUSSION sensitivity of the inactivation rate is not determinable because Inactivation is best understood and intensively studied for of very low TEA affinity to the channel (22, 23, 35). However, classical N- and C-type mechanisms in ShakerB potassium KCNQ1 inactivation is independent of extracellular potassium channels. Fast N-type inactivation is realized by a ball plug-in concentration; a principal difference from C-type behavior. In mechanism, where a ball-forming domain of the channel binds Shaker-like potassium channels, certain mutations within the to its receptor thereby blocking the ion pathway from the in- P-region H5 and the sixth membrane-spanning region S6 have tracellular side (1–5). This occurs independently of the mem- been shown to affect the time course of inactivation (33, 36). brane potential. KCNQ1 inactivation exerts slower inactiva- Notably, also in our study KCNQ1 inactivation was enhanced tion gating and a weak voltage dependence. Further stressing by modification of this region (Q1S6Q2), but identified residues the differences from the N-type mechanism, the identified res- essential for KCNQ1 inactivation within the S5 region and the idues at KCNQ1 positions 272, 273, and 307 are not expected to helical part of the H5 segment have not been reported to be be accessible for a complex ball structure from the intracellular side and thus probably cannot account for a potential receptor substantially relevant for C-type inactivation. In conclusion, site. both C-type and KCNQ1 inactivation occur within the pore Geography of KCNQ1 Inactivation 13605 TABLE II REFERENCES Characteristics of inactivation in KCNQ1-WT and KCNQ1(L273F) in 1. Armstrong, C. M., and Bezanilla, F. (1977) J. Gen. Physiol. 70, 567–590 homomeric and heteromeric expression together with MinK in 2. Zagotta, W. N., Hoshi, T., and Aldrich, R. W. (1990) Science 250, 568 –571 Xenopus oocytes 3. Hoshi, T., Zagotta, W. N., and Aldrich, R. W. (1990) Science 250, 533–538 4. Rettig, J., Heinemann, S. H., Wunder, F., Lorra, C., Parcej, D. N., Dolly, J. O., Oocytes were injected with 10 ng of KCNQ1 cRNA. For heteromeric and Pongs, O. (1994) Nature 369, 289 –294 expression 5 ng of MinK cRNA were added. The double-pulse protocol 5. Liu, Y., Jurman, M. E., and Yellen, G. (1996) Neuron 16, 859 – 867 with two 40-mV pulses separated by a 20-ms pulse to 2120 mV was 6. Lerche, H., Heine, R., Pika, U., George, A. L., Jr., Mitrovic, N., Browatzki, M., used to analyze the onset of inactivation. The time constants of onset Weiss, T., Rivet-Bastide, M., Franke, C., Lomonaco, M., Ricker, K., and were calculated by exponential fits to the decaying phase of the currents Lehmann-Horn, F. (1993) J. Physiol. 470, 13–22 at the beginning of the second pulse. To determine recovery from 7. Bennett, P. B., Yazawa, K., Makita, N., and George, A. L., Jr. (1995) Nature inactivation and the fraction of inactivated channels, currents were 376, 683– 685 activated by a 3-s pulse to 40 mV followed by a hyperpolarizing pulse to 8. Wang, Q., Curran, M. E., Splawski, I., Burn, T. C., Millholland, J. M., Van Raay, T. J., Shen, J., Timothy, K. W., Vincent, G. M., De Jager, T., Schwarz, 260 mV. Tail currents were analyzed by double exponential fits. The P. J., Towbin, J. A., Moss, A. J., Atkinson, D. L., Landes, G. M., Connors, faster and the slower component constitute recovery from inactivation T. D., and Keating, M. T. (1996) Nat. Genet. 12, 17–23 and deactivation respectively. Values are mean 6 S.E., numbers of 9. Charlier, C., Singh, N. A., Ryan, S. G., Lewis, T. B., Reus, B. E., Leach, R. J., oocytes are shown in parenthesis. and Leppert, M. (1998) Nat. Genet. 18, 53–55 10. Singh, N. A., Charlier, C., Stauffer, D., DuPont, B. R., Leach, R. J., Melis, R., Inactivated t t onset recovery Ronen, G. M., Bjerre, I., Quattlebaum, T., Murphy, J. V., McHarg, M. L., channels Gagnon, D., Rosales, T. O., Peiffer, A., Anderson, V. E., and Leppert, M. ms % (1998) Nat. Genet. 18, 25–29 11. Biervert, C., Schroeder, B. C., Kubisch, C., Berkovic, S. F., Propping, P., WT-KCNQ1 15 6 1(n 5 12) 45 6 3(n 5 13) 31 6 1(n 5 12) a Jentsch, T. J., and Steinlein, O. K. (1998) Science 279, 403– 406 KCNQ1(L273F) 60 6 6(n 5 6) 98 6 5(n 5 6) 86 6 3(n 5 6) 12. Kubisch, C., Schro ¨ der, B. C., Friedrich, T., Lu ¨ tjohann, B., El Amraoui, A., WT-KCNQ1/ —— 0 Marlin, S., Petit, C., and Jentsch, T. J. (1999) Cell 96, 437– 446 MinK 13. Lerche, C., Scherer, C. R., Seebohm, G., Derst, C., Wei, A. D., Busch, A. E., and KCNQ1(L273F)/ 26 6 2(n 5 8) 111 6 3(n 5 8) 24 6 5(n 5 12) Steinmeyer, K. (2000) J. Biol. Chem. 29, 22395–22400 MinK 14. Takumi, T., Ohkubo, H., and Nakanishi, S. (1988) Science 242, 1042–1045 15. Sanguinetti, M. C., Curran, M. E., Zou, A., Shen, J., Spector, P. S., Atkinson, Significance of homomeric KCNQ1-WT versus KCNQ1(L273F). D. L., and Keating, M. T. (1996) Nature 384, 80–83 Significance of homomeric KCNQ1(L273F) versus heteromeric 16. Barhanin, J., Lesage, F., Guillemare, E., Fink, M., Lazdunski, M., and Romey, KCNQ1(L273F)/MinK. G. (1997) Nature 384, 78–80 17. Attali, B. (1996) Nature 384, 24 –25 region spanning S5 to S6, but the defined critical structures 18. Chouabe, C., Neyroud, N., Guicheney, P., Lazdunski, M., Romey, G., and Barhanin, J. (1997) EMBO J. 16, 5472–5479 appear to be different. 19. Shalaby, F. Y., Levesque, P. C., Yang, W. P., Little, W. A., Conder, M. L., In summary our results define geographical peculiarities for Jenkins-West, T., and Blanar, M. A. (1997) Circulation 96, 1733–1736 20. Wollnik, B., Schroeder, B. C., Kubisch, C., Esperer, H. D., Wieacker, P., and KCNQ1 inactivation, which were not yet reported for classical Jentsch, T. J. (1997) Hum. Mol. Gen. 6, 1943–1949 N- and C-type inactivation in Shaker-like potassium channels. 21. Schulze-Bahr, E., Wang, Q., Wedekind, H., Haverkamp, W., Chen, Q., and KCNQ1 inactivation is characterized by an involvement of the Sun, Y. (1997) Nat. Genet. 17, 267–268 22. Pusch, M., Magrassi, R., Wollnik, B., and Conti, F. (1998) Biophys. J. 75, pore region H5 and the S5 transmembrane region, putatively 785–792 interplaying with each other. It is believed that KCNQ1 inac- 23. Tristani-Firouzi, M., and Sanguinetti, M. C. (1998) J. Physiol. (Lond.) 510, 37– 45 tivation is abolished by interaction with MinK in native het- 24. Franqueza, L., Lin, M., Shen, J., Splawski, I., Keating, M. T., and Sanguinetti, eromeric I channels (22, 23). As suggested previously, inter- Ks M. C. (1999) J. Biol. Chem. 274, 21063–21070 action between KCNQ1 and MinK might involve the pore 25. Pusch, M., Bertorello, L., and Conti, F. (2000) Biophys. J. 78, 211–226 26. Romey, G., Attali, B., Chouabe, C., Abitbol, I., Guillemare, E., Barhanin, J., region of KCNQ1 (26, 27). Our results demonstrate an essential and Lazdunski, M. (1997) J. Biol. Chem. 272, 16713–16716 function of the pore region for KCNQ1 inactivation, therefore 27. Tai, K. K., and Goldstein, S. A. (1998) Nature 391, 605– 608 allowing the interesting possibility that MinK interaction may 28. Sanguinetti, M. C., Jiang, C., Curran, M.-E., and Keating, M. T. (1995) Cell 81, 299 –307 occur in the same region that is important for inactivation. 29. Smith, P. L., Baukrowitz, T., and Yellen, G. (1996) Nature 379, 833– 836 MinK possibly abolishes inactivation by weakening the inter- 30. Schonherr, R., and Heinemann, S. H. (1996) J. Physiol. (Lond.) 493, 635– 642 31. Wang, H. S., Pan, Z., Shi, W., Brown, B. S., Wymore, R. S., Cohen, I. S., Dixon, play between H5 and S5. In the LQT1 mutant L273F this J. E., and McKinnon, D. (1998) Science 282, 1890 –1893 interplay might be stronger, thereby stabilizing inactivation 32. Doyle, D. A., Morais, C.abral, J., Pfuetzner, R. A., Kuo, A., Gulbis, J. M., such that MinK no longer can abolish inactivation. This in turn Cohen, S. L., Chait, B. T., and MacKinnon, R. (1998) Science 280, 69 –77 33. LopezBarneo, J., Hoshi, T., Heinemann, S. H., and Aldrich, R. W. (1993) may contribute to the deleterious phenotype of the LQT1 Recept. Channel 1, 61–71 mutant. 34. Choi, K. L., Aldrich, R. W and Yellen, G. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 5092–5095 Acknowledgments—We thank K. Steinmeyer, H. Lerche, B. Attali, P. 35. Hadley, J. K., Noda, M., Selyanko, A. A., Wood, I. C., Abogadie, F. C., and Ruppersberg, A. Wei, I. Gutcher, and E. Kostenis for fruitful discus- Brown, D. A. (2000) Br. J. Pharmacol. 129, 413– 415 sions and careful proofreading. 36. Hoshi, T., Zagotta, W. N., and Aldrich, R. W. (1991) Neuron 7, 547–556

Journal

Journal of Biological Chemistry – Unpaywall

Published: Apr 1, 2001

There are no references for this article.