C Terminus-mediated Control of Voltage and cAMP Gating of Hyperpolarization-activated Cyclic Nucleotide-gated Channels
2001; Elsevier BV; Volume: 276; Issue: 32 Linguagem: Inglês
10.1074/jbc.m103971200
ISSN1083-351X
AutoresCarlo Viscomi, Claudia Altomare, Annalisa Bucchi, Eva Camatini, Mirko Baruscotti, Anna Moroni, Dario DiFrancesco,
Tópico(s)Neuroscience and Neuropharmacology Research
ResumoThe hyperpolarization-activated cyclic nucleotide-gated (HCN) family of "pacemaker" channels includes 4 isoforms, the kinetics and cAMP-induced modulation of which differ quantitatively. Because HCN isoforms are highly homologous in the central region, but diverge more substantially in the N and C termini, we asked whether these latter regions could contribute to the determination of channel properties. To this aim, we analyzed activation/deactivation kinetics and the response to cAMP of heterologously expressed isoforms mHCN1 and rbHCN4 and verified that mHCN1 has much faster kinetics and lower cAMP sensitivity than rbHCN4. We then constructed rbHCN4 chimeras by replacing either the N or the C terminus, or both, with the analogous domains from mHCN1. We found that: 1) replacement of the N terminus (chimera N1–4) did not substantially modify either the kinetics or cAMP dependence of wild-type channels; 2) replacement of the C terminus, on the contrary, resulted in a chimeric channel (4-C1), the kinetics of which were strongly accelerated compared with rbHCN4, and that was fully insensitive to cAMP; 3) replacement of both N and C termini led to the same results as replacement of the C terminus alone. These results indicate that the C terminus of rbHCN4 contributes to the regulation of voltage- and cAMP-dependent channel gating, possibly through interaction with other intracellular regions not belonging to the N terminus. The hyperpolarization-activated cyclic nucleotide-gated (HCN) family of "pacemaker" channels includes 4 isoforms, the kinetics and cAMP-induced modulation of which differ quantitatively. Because HCN isoforms are highly homologous in the central region, but diverge more substantially in the N and C termini, we asked whether these latter regions could contribute to the determination of channel properties. To this aim, we analyzed activation/deactivation kinetics and the response to cAMP of heterologously expressed isoforms mHCN1 and rbHCN4 and verified that mHCN1 has much faster kinetics and lower cAMP sensitivity than rbHCN4. We then constructed rbHCN4 chimeras by replacing either the N or the C terminus, or both, with the analogous domains from mHCN1. We found that: 1) replacement of the N terminus (chimera N1–4) did not substantially modify either the kinetics or cAMP dependence of wild-type channels; 2) replacement of the C terminus, on the contrary, resulted in a chimeric channel (4-C1), the kinetics of which were strongly accelerated compared with rbHCN4, and that was fully insensitive to cAMP; 3) replacement of both N and C termini led to the same results as replacement of the C terminus alone. These results indicate that the C terminus of rbHCN4 contributes to the regulation of voltage- and cAMP-dependent channel gating, possibly through interaction with other intracellular regions not belonging to the N terminus. hyperpolarization-activated cyclic nucleotide-gated cyclic nucleotide binding amino acid polymerase chain reaction Together with voltage-dependent K+channels and cyclic nucleotide-gated channels, the recently cloned hyperpolarization-activated cyclic nucleotide-gated (HCN)1 channels belong to the superfamily of 6 transmembrane domain channels (1Clapham D.E. Neuron. 1998; 21: 1-5Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar). HCN channel subunits are the molecular building blocks of native hyperpolarization-activated "pacemaker" (f/h) channels of cardiac and neuronal cells (2DiFrancesco D. Annu. Rev. Physiol. 1993; 55: 455-472Crossref PubMed Scopus (683) Google Scholar, 3Pape H.C. Annu. Rev. Physiol. 1996; 58: 299-327Crossref PubMed Scopus (991) Google Scholar). A peculiar property of HCN channels is their dual modulation by voltage hyperpolarization and by intracellular cAMP (4DiFrancesco D. J. Physiol. (Lond.). 1999; 512: 367-376Crossref Scopus (91) Google Scholar); as in voltage-dependent K+ channels, voltage sensing is localized in the S4 transmembrane domain (5Vaca L. Stieber J. Zong X. Ludwig A. Hofmann F. Biel M. FEBS Lett. 2000; 4791–2: 35-40Crossref Scopus (42) Google Scholar, 6Chen J. Mitcheson J.S. Lin M. Sanguinetti M.C. J. Biol. Chem. 2000; 275: 36465-36471Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar), and as in cyclic nucleotide-gated channels, gating by cAMP requires direct binding of cAMP molecules to the cyclic nucleotide binding (CNB) domain located intracellularly at the C terminus (7DiFrancesco D. Tortora P. Nature. 1991; 351: 145-147Crossref PubMed Scopus (665) Google Scholar). Thus far, four HCN isoforms (1Clapham D.E. Neuron. 1998; 21: 1-5Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar, 2DiFrancesco D. Annu. Rev. Physiol. 1993; 55: 455-472Crossref PubMed Scopus (683) Google Scholar, 3Pape H.C. Annu. Rev. Physiol. 1996; 58: 299-327Crossref PubMed Scopus (991) Google Scholar, 4DiFrancesco D. J. Physiol. (Lond.). 1999; 512: 367-376Crossref Scopus (91) Google Scholar) have been cloned. When expressed heterologously, HCN channels display kinetic and modulatory properties similar to those of native If/Ih channels (8Santoro B. Liu D.T. Yao H. Bartsch D. Kandel E.R. Siegelbaum S.A. Tibbs G.R. Cell. 1998; 935: 717-729Abstract Full Text Full Text PDF Scopus (591) Google Scholar, 9Gauβ R. Seifert R. Kaupp B.U. Nature. 1998; 393: 583-587Crossref PubMed Scopus (381) Google Scholar, 10Ludwig A. Zong X. Jeglitsch M. Hofmann F. Biel M. Nature. 1998; 393: 587-591Crossref PubMed Scopus (797) Google Scholar, 11Ludwig A. Zong X. Stieber J. Hullin R. Hofmann F. Biel M. EMBO J. 1999; 189: 2323-2329Crossref Scopus (316) Google Scholar, 12Vaccari T. Moroni A. Rocchi M. Gorza L. Bianchi M.E. Beltrame M. DiFrancesco D. Biochim. Biophys. Acta. 1999; 1446: 419-425Crossref PubMed Scopus (67) Google Scholar, 13Ishii T.M. Takano M. Xie L.H. Noma A. Ohmori H. J. Biol. Chem. 1999; 274: 12835-12839Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar, 14Moroni A. Barbuti A. Altomare C. Viscomi C. Morgan J. Baruscotti M. DiFrancesco D. Pflügers Arch. Eur. J. Physiol. 2000; 439: 618-626Crossref PubMed Scopus (73) Google Scholar). However, different isoforms have qualitatively similar, but quantitatively different properties. Activation and deactivation kinetics for example are much faster for HCN1 than HCN2 or HCN4, whereas the HCN1 sensitivity to cAMP is much reduced; further, the activation threshold of HCN2 is more negative than that of either HCN1 or HCN4 (15Altomare C. Bucchi A. Camatini E. Baruscotti M. Viscomi C. Moroni A. DiFrancesco D. J. Gen. Physiol. 2001; 117: 519-532Crossref PubMed Scopus (128) Google Scholar, 16Moosmang S. Stieber J. Zong X. Biel M. Hofmann F. Ludwig A. Eur. J. Biochem. 2001; 268: 1646-1652Crossref PubMed Scopus (362) Google Scholar). These differences are important in that they can provide a molecular basis for the diverse properties of native channels in various cell types; it is known for example that the Ih current of hippocampal CA1 neurons has faster kinetics and a reduced cAMP sensitivity relative to the same current recorded in other regions (such as the thalamic relay neurons) or in cardiac myocytes (2DiFrancesco D. Annu. Rev. Physiol. 1993; 55: 455-472Crossref PubMed Scopus (683) Google Scholar,3Pape H.C. Annu. Rev. Physiol. 1996; 58: 299-327Crossref PubMed Scopus (991) Google Scholar). Furthermore, the If activation threshold in cardiac conduction tissue (Purkinje fiber) is substantially more negative than that in pacemaker cells of the sinoatrial node (17DiFrancesco D. Peracchia C. Handbook of Membrane Channels: Molecular and Cellular Physiology. Academic Press, New York1994: 335-343Crossref Google Scholar). Thus, differential expression of HCN isoforms may represent a basic mechanism underlying the variable kinetic and modulatory features of native f/h channels in different tissues. The observation that different isoforms have different features prompts the question of how kinetic and modulatory properties of HCN channels are determined. Because HCN isoforms are highly homologous in the "core" region (i.e. the TM domains S1 to S6) and in the CNB domain but diverge more substantially at both the N and C termini (18Ludwig A. Zong X. Hofmann F. Biel M. Cell Physiol. Biochem. 1999; 9: 179-186Crossref PubMed Scopus (75) Google Scholar), we asked whether the most diverging regions could have a role in affecting channel properties. In this work we have investigated the contribution of the N- and C-terminal regions to the channel kinetics and sensitivity to cAMP. To this aim, we used chimeras obtained by the replacement of regions of rbHCN4 with equivalent regions of mHCN1 channels. The results indicate that the C terminus, but not the N terminus, of rbHCN4 plays an important role in determining both the rate of channel activation/deactivation and the cAMP-dependent gating, presumably by an interaction with other intracellular channel regions that do not belong to the N terminus. Chimeric channels were obtained by standard molecular biology techniques (19Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd Ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1989Google Scholar) from mHCN1 (BCNG1, 20, kindly provided by Dr. B. Santoro, Columbia University) and rbHCN4 (HAC4, 13, kindly provided by Dr. H. Ohmori, University of Kyoto). The chimera N1–4 was obtained by replacing the N terminus of rbHCN4 (aa 1–228) with the N terminus of mHCN1 (aa 1–96), by means of a commonHincII restriction site, found in a region of high similarity among the two clones, upstream of the first transmembrane domain (S1). Chimera 4-C1 was made by replacing the C terminus of rbHCN4 (aa 545–1175) with the C terminus of mHCN1 (aa 413–910). To obtain this construct we introduced a KpnI restriction site in the C terminus of mHCN1 by PCR (Expand Long Template PCR System,Roche Molecular Biochemicals) according to the following procedure: 2 min at 94 °C, 30 times for 45 s at 94 °C, 45 s at 60 °C, and 2 min at 68 °C with the oligonucleotides CHCN1-for (5′-AGCACCGGTACCAAGGCAAG-3′, corresponding to aa 425–431) and CHCN1-rev (5′-CCCCGAATCATAAATTCGAAGCA-3′, corresponding to aa 907–910). The PCR product was cloned (Topo cloning kit, Invitrogen) and confirmed by sequencing (MWG Biotech). The CHCN1-for oligonucleotide included a point mutation that introduced into the mHCN1 sequence a KpnI restriction site, also present in rbHCN4, downstream of the last transmembrane domain S6. ThisKpnI site was used to assemble the chimeric construct 4-C1. The N1–4-C1 chimera was obtained in a similar way, by replacing the C terminus of N1–4 with the same PCR product. For functional expression, wild-type and chimeric channels were inserted into a mammalian expression vector (pCDNA3.1+, Invitrogen) and cotransfected with a green fluorescent protein-containing vector in modified HEK 293 (Phoenix (21Kinsella T.M. Nolan G.P. Human Gene Ther. 1996; 7: 1405-1413Crossref PubMed Scopus (672) Google Scholar)) cells with a standard calcium phosphate protocol, as described previously (14Moroni A. Barbuti A. Altomare C. Viscomi C. Morgan J. Baruscotti M. DiFrancesco D. Pflügers Arch. Eur. J. Physiol. 2000; 439: 618-626Crossref PubMed Scopus (73) Google Scholar). Currents were measured 48–96 h after transfection. Experiments were performed on cells incubated after transfection at 37 °C in 5% CO2 for 1–5 days. On the day of the experiment cells were dispersed by trypsin, plated at a low density on 35-mm plastic Petri dishes, and allowed to settle for 2–4 h; dishes were then placed on the stage of an inverted microscope and perfused with a Tyrode solution containing (mm): NaCl, 110; KCl, 30; CaCl2, 1.8; MgCl2, 0.5; Hepes-NaOH, 5; BaCl2, 1; MnCl2, 1; nifedipine, 0.02; NiCl, 0.1; pH 7.4. Fluorescent single cells were voltage-clamped as described previously (14Moroni A. Barbuti A. Altomare C. Viscomi C. Morgan J. Baruscotti M. DiFrancesco D. Pflügers Arch. Eur. J. Physiol. 2000; 439: 618-626Crossref PubMed Scopus (73) Google Scholar). In whole-cell experiments, the pipette solution contained (mm): KCl, 130; NaCl, 10; MgCl2, 0.5; EGTA-KOH, 1; Hepes-KOH, 5; ATP sodium salt, 2; creatine phosphate, 5; GTP, 0.1; pH 7.2. In inside-out patch experiments, large tipped pipettes were used for macro-patch measurements (resistance of about 0.8 megohm). Following seal formation and current recording in the cell-attached configuration, the patch was extruded and superfused with a solution containing (mm): potassium aspartate, 130; NaCl, 10; CaCl2, 2; EGTA, 5; HEPES-KOH, 5; pH 7.2. The pipette solution contained (mm): NaCl, 70; KCl, 70; CaCl2, 1.8; MgCl2, 1; BaCl2, 1; MnCl2, 2; HEPES-KOH, 5; pH 7.4. All experiments were run at room temperature (24–26 °C). Measurement of activation curves for HCN currents was performed by the use of standard voltage-clamp protocols (see for example Ref. 22DiFrancesco D. Ferroni A. Mazzanti M. Tromba C. J. Physiol. (Lond.). 1986; 377: 61-88Crossref Scopus (366) Google Scholar). Cells were held at a fixed holding potential of −35 mV, and steps of variable duration were applied to voltages in the activation range to reach steady-state current activation. The duration of steps varied according to the HCN channel isoform investigated because different isoforms have different activation/deactivation rates. For example at −105 mV the step duration was 2.5 s and 9 s for mHCN1 and rbHCN4, respectively. Activation curve data were then measured as normalized deactivation tail amplitudes at a test voltage where the current was fully deactivated (+5 mV). Activation curves were fitted by a standard Boltzmann distribution equationy(V) = 1/(1 + exp((V −V 12)/s)) where V 12is the half-activation voltage and s the inverse slope factor. Shifts of the current activation curve in macro-patch inside-out experiments were measured by a "trace superimposition" method developed previously (23Accili E.A. DiFrancesco D. Pflügers Arch. Eur. J. Physiol. 1996; 431: 757-762Crossref PubMed Scopus (35) Google Scholar). Briefly, the current was measured during steps applied at a fixed frequency (1/6 Hz) to near the midactivation voltage range, and the control trace was frozen on the digital oscilloscope; the current increase caused by superfusion with cAMP was then compensated by manual change of the holding potential, until superimposition of the traces before and after admission of cAMP was achieved. After a correction accounting for the change in applied voltage during cAMP perfusion, the displacement of the holding potential was used to calculate the shift of the activation curve. Sets of data were compared using independent Student's t test, and the significance level was set to p = 0.05. All data are plotted as mean ± S.E. values. The rationale behind our experimental approach is shown in Fig.1. A comparison among sequence topologies of the various isoforms of HCN channels (Fig. 1 a) indicates that the isoforms are highly homologous in the core region,i.e. the transmembrane (S) and CNB domains, but differ more substantially in the N terminus and the C terminus. To visualize the difference in homology between terminal regions and S and CNB regions, we constructed a multiple alignment sequence (CLUSTAL X (24Jeanmougin F. Thompson J.D. Gouy M. Higgins D.G. Gibson T.J. Trends Biochem. Sci. 1998; 23: 403-405Abstract Full Text Full Text PDF PubMed Scopus (2406) Google Scholar)) using known sequences for mHCN1 (20Santoro B. Grant S.G. Bartsch D. Kandel E.R. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 14815-14820Crossref PubMed Scopus (229) Google Scholar), hHCN2 (12Vaccari T. Moroni A. Rocchi M. Gorza L. Bianchi M.E. Beltrame M. DiFrancesco D. Biochim. Biophys. Acta. 1999; 1446: 419-425Crossref PubMed Scopus (67) Google Scholar), mHCN3 (10Ludwig A. Zong X. Jeglitsch M. Hofmann F. Biel M. Nature. 1998; 393: 587-591Crossref PubMed Scopus (797) Google Scholar), and rbHCN4 (13Ishii T.M. Takano M. Xie L.H. Noma A. Ohmori H. J. Biol. Chem. 1999; 274: 12835-12839Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar) and plotted the homology function against multiple alignment amino acid position in Fig. 1 b. The plot clearly indicates a high degree of homology in the region encompassing the 6 S and CNB domains, in which most residues are identical for all 4 isoforms, in contrast with the N- and C-terminal regions in which most substitutions are nonconservative. The protein identity between sequence pairs varies in the range 80–90% in the core region (S1 through the CNBD) and in the ranges 21–37% and 11–21% in the N and C termini, respectively (see Fig. 1 legend). Heterologous expression has indicated that although similar qualitatively, the kinetic and modulatory properties of individual isoforms differ quantitatively (8Santoro B. Liu D.T. Yao H. Bartsch D. Kandel E.R. Siegelbaum S.A. Tibbs G.R. Cell. 1998; 935: 717-729Abstract Full Text Full Text PDF Scopus (591) Google Scholar, 10Ludwig A. Zong X. Jeglitsch M. Hofmann F. Biel M. Nature. 1998; 393: 587-591Crossref PubMed Scopus (797) Google Scholar, 11Ludwig A. Zong X. Stieber J. Hullin R. Hofmann F. Biel M. EMBO J. 1999; 189: 2323-2329Crossref Scopus (316) Google Scholar, 13Ishii T.M. Takano M. Xie L.H. Noma A. Ohmori H. J. Biol. Chem. 1999; 274: 12835-12839Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar, 14Moroni A. Barbuti A. Altomare C. Viscomi C. Morgan J. Baruscotti M. DiFrancesco D. Pflügers Arch. Eur. J. Physiol. 2000; 439: 618-626Crossref PubMed Scopus (73) Google Scholar, 15Altomare C. Bucchi A. Camatini E. Baruscotti M. Viscomi C. Moroni A. DiFrancesco D. J. Gen. Physiol. 2001; 117: 519-532Crossref PubMed Scopus (128) Google Scholar, 16Moosmang S. Stieber J. Zong X. Biel M. Hofmann F. Ludwig A. Eur. J. Biochem. 2001; 268: 1646-1652Crossref PubMed Scopus (362) Google Scholar, 25Santoro B. Tibbs G.R. Ann. N. Y. Acad. Sci. 1999; 868: 741-764Crossref PubMed Scopus (310) Google Scholar). Some basic differences in kinetic properties and cAMP dependence between isoforms mHCN1 and rbHCN4 are shown in Figs. 2 and3.Figure 3Action of cAMP. Currents were measured during steps to the voltages indicated (holding potential was −35 mV) from macro-patches in the inside-out configuration, and cAMP (10 µm) was perfused on the intracellular side of the patch (see "Experimental Procedures"). Shifts of the activation curve obtained by a "trace superimposition" protocol (detailed under "Experimental Procedures") were 7.6 and 17.4 mV for isoforms 1 and 4, respectively. Mean shifts of the activation curve were 5.6 ± 0.5 (n = 10) for mHCN1 and 16.0 ± 0.9 mV (n = 15) for rbHCN4 currents.View Large Image Figure ViewerDownload Hi-res image Download (PPT) In Fig. 2 b, activation records from cells expressing mHCN1 (left) and rbHCN4 (right) show that, in agreement with previous data, the two isoforms have very different rates of activation. For example, in the records in Fig. 2 b the activation time constant at −95 mV, measured by fitting records after an initial delay (15Altomare C. Bucchi A. Camatini E. Baruscotti M. Viscomi C. Moroni A. DiFrancesco D. J. Gen. Physiol. 2001; 117: 519-532Crossref PubMed Scopus (128) Google Scholar), was 85.9 ms for mHCN1 and 2772.8 ms for rbHCN4. The mean time constant at the same voltage was 105.6.0 ± 9.8 ms (n = 8) and 3023.1 ± 348.2 ms (n= 17) for isoforms 1 and 4, respectively. We measured activation curves of mHCN1 and rbHCN4 from several cells (see "Experimental Procedures") and fitted the mean activation curves in Fig.2 c with a Boltzmann distribution equation. This yielded half-activation voltages, V 12 = −71.7 and −81.2 mV, and inverse slope coefficients, s = 8.6 and 10.0 mV for mHCN1 (n = 8) and rbHCN4 (n = 7), respectively. These results confirm previous work indicating that the range of activation of isoforms 1 and 4 are not too dissimilar (15Altomare C. Bucchi A. Camatini E. Baruscotti M. Viscomi C. Moroni A. DiFrancesco D. J. Gen. Physiol. 2001; 117: 519-532Crossref PubMed Scopus (128) Google Scholar,26Kaupp U.B. Seifert R. Annu. Rev. Physiol. 2001; 63: 235-257Crossref PubMed Scopus (314) Google Scholar). To further quantify the difference in voltage-dependent kinetics between the two isoforms, we fitted activation and deactivation records according to a single-exponential time course following an initial delay (15Altomare C. Bucchi A. Camatini E. Baruscotti M. Viscomi C. Moroni A. DiFrancesco D. J. Gen. Physiol. 2001; 117: 519-532Crossref PubMed Scopus (128) Google Scholar) and plotted in Fig. 2 the mean time constant of activation (d) and deactivation (e) for mHCN1 (τon, n = 9; τoff, n = 7 cells) and rbHCN4 (τon, n = 17; τoff,n = 11 cells). Clearly, both activation and deactivation rates for isoform 4 are severalfold slower than those for isoform 1. We then proceeded to verify whether the isoforms investigated have different degrees of sensitivity to internal cAMP. As shown in the sample records of Fig. 3, macro-patch currents were recorded in the inside-out configuration during voltage-clamp hyperpolarizations to the voltages indicated before and after perfusion of the intracellular side of the patch with 10 µm cAMP. In the patches shown in Fig. 3, cAMP shifted the activation curve of mHCN1 and rbHCN4 currents by 7.6 and 17.4 mV, respectively. The mean shifts induced by 10 µm cAMP on the current activation curve were 5.6 ± 0.5 mV (n = 10) and 16.0 ± 0.9 mV (n = 15) for mHCN1 and rbHCN4, respectively. These data confirm previous evidence indicating a different sensitivity to cAMP of isoforms 1 and 4 (11Ludwig A. Zong X. Stieber J. Hullin R. Hofmann F. Biel M. EMBO J. 1999; 189: 2323-2329Crossref Scopus (316) Google Scholar, 13Ishii T.M. Takano M. Xie L.H. Noma A. Ohmori H. J. Biol. Chem. 1999; 274: 12835-12839Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar, 25Santoro B. Tibbs G.R. Ann. N. Y. Acad. Sci. 1999; 868: 741-764Crossref PubMed Scopus (310) Google Scholar). In a first approach to investigating the role of N and C termini in channel gating, we constructed chimeric rbHCN4 channels where either the N or the C termini of rbHCN4 were substituted with those from mHCN1. Data obtained with two such chimeras are shown in Fig.4. On the left, we replaced the N terminus of rbHCN4 (aa 1–228) with that of mHCN1 (aa 1–96). The resulting chimera (N1–4) had activation and deactivation time constant curves similar to those of rbHCN4 (Fig. 4,c and d, left). For example, at −105 mV the activation time constant (τon) was 84.1 ± 9.1 ms and 2519.5 ± 386.0 ms for mHCN1 and rbHCN4, respectively, according to the data in Fig. 2 d. The τon of N1–4 current was 1904.1 ± 108.5 ms, not significantly different from the τon of rbHCN4 (p = 0.33). Similarly at −85 mV the deactivation time constant τoffof N1–4 current was 1634.4 ± 326.4 ms, not significantly different from τoff of rbHCN4 (1746.8 ± 305.2 ms,p = 0.81). In the whole voltage range investigated, differences between activation/deactivation time constants for N1–4 and wild-type rbHCN4 channels did not reach the significance level (p = 0.05). In Fig. 4, right, on the other hand, substitution of the C terminus of rbHCN4 (aa 545–1175) with that of mHCN1 (aa 413–910) led to a chimeric channel (4-C1), the activation and deactivation time constant curves of which clearly diverged from those of the wild-type rbHCN4 channel (c and d, right; note that the wild-type channel curves plotted both in left andright panels are the same as in Fig. 2, d ande). At −95 mV, for example, τon was 1423.7 ± 218.2 ms (n = 8) (compare with 3023.1 ms for rbHCN4), and at −85 mV, τoff was 465.0 ± 97.9 ms (n = 5) (compare with 1746.8 ms for rbHCN4). In the whole voltage range investigated, activation and deactivation time constants were intermediate between those of wild type channels. Interestingly, at the most negative voltages, deactivation of 4-C1 was almost as fast as that of mHCN1 (Fig. 4 d,right). The activation curves for both the N and C replacement chimeras fell close to, and in fact nearly overlapped with, the activation curve for the wild-type rbHCN4 channel (Fig. 4 b). Fitting activation curves with Boltzmann equations yielded half-activation voltagesV 12 = −83.5 and −80.6 mV and inverse slope factors s = 12.0 and 12.6 mV for the N1–4 and 4-C1 channels, respectively. These values can be compared with those for rbHCN4 (V 12 = −81.2 mV, s = 10.0 mV, see Fig. 2). To quantify the similarity between the activation curves of N1–4 and 4-C1 chimeras and wild-type rbHCN4 channels, we ran significance tests at voltages close to half-activation. AtV = −95, −85, and −75 mV, values of the activation curve of N1–4 and 4-C1 were not significantly different from those of rbHCN4 (p > 0.28). In the light of the time constant data (Fig. 4, c andd), the similarity between 4-C1 and rbHCN4 activation curves may seem surprising. However, a recent description of HCN channel voltage gating by an allosteric model has shown that changes in the position of the activation curve are attributable to changes in the voltage dependence of voltage sensors, located in the S4 region (15Altomare C. Bucchi A. Camatini E. Baruscotti M. Viscomi C. Moroni A. DiFrancesco D. J. Gen. Physiol. 2001; 117: 519-532Crossref PubMed Scopus (128) Google Scholar). Thus, the results of Fig. 4 b agree with the view that substitution of the C terminus does not impair the voltage-dependent distribution of "willing"versus "reluctant" voltage sensors. The results above suggest that the C terminus of rbHCN4 has a role in controlling voltage-dependent gating. Because the chimera carrying the C terminus of mHCN1 (4-C1) does accelerate both activation and deactivation kinetics of rbHCN4 but not to the extent of reproducing fully the kinetics of mHCN1, the role of the C terminus cannot be unique, and other channel regions are bound to contribute to determining the kinetic properties of rbHCN4. The N terminus of rbHCN4, on the other hand, could be replaced by the N terminus of mHCN1 without substantial modification of activation/deactivation kinetics, suggesting that the N terminus does not have a similarly crucial role in determining channel kinetics. As a way to investigate if the modifications induced by substitution of the C terminus could depend upon the interaction between C and N termini, we proceeded to construct a chimeric channel in which both rbHCN4 termini were simultaneously replaced with those from mHCN1 (Fig.5, N1–4-C1). This chimera had the same kinetic features of 4-C1. A Boltzmann fit of the mean activation curve for N1–4-C1 yielded values of half-activation voltageV 12 of −83.5 ± 0.4 mV and inverse slope factor s = 11.5 ± 0.4 mV (n = 4), which are comparable with the values obtained for 4-C1. Also, the time constants of activation and deactivation (τon, τoff) of N1–4-C1 were not significantly different from those of 4-C1 in the whole voltage range investigated. Finally, we addressed the issue of the cAMP dependence of the channel open probability by analyzing in inside-out patches the response to cAMP of the three chimeras investigated and comparing their response to that of wild-type channels. As shown in the sample records of Fig.6, the response of N1–4 to cAMP was only moderately reduced with respect to rbHCN4 (mean shift of the activation curve with 10 µm cAMP was 11.6 ± 0.7 mV,n = 4), whereas 4-C1 was totally insensitive to cAMP (the shift was 0 mV in n = 13 patches). The lack of a cAMP-induced shift of the 4-C1 activation curve contrasts with the presence of a reduced but measurable shift of 5.6 ± 0.5 mV obtained for the wild-type mHCN1 channel under similar conditions (Fig.3). The lack of cAMP dependence of the 4-C1 chimera is especially interesting because it indicates that the sensitivity to cAMP is not entirely determined by the cyclic nucleotide consensus region and suggests the presence of specific interactions between the C terminus and other regions of the channel able to affect the mechanism of cAMP-dependent gating. To verify the possible involvement of the N terminus in this interaction, we tested the responsiveness to cAMP of the double N and C termini replacement chimera N1–4-C1. Like the C replacement mutant, this double mutant was also irresponsive to cAMP (a shift of 0 mV was observed in all of the n = 6 patches analyzed). This favors the view that the cAMP sensitivity involves a coupling mechanism between the C terminus and the core channel region, presumably located in one or more intracellular S-linkers. Our results indicate that: 1) replacement of the N terminus of rbHCN4 with that from mHCN1 does not alter channel activation and deactivation kinetics (Fig. 4, left) nor does it modify the channel response to cAMP with respect to the wild-type rbHCN4 channel (Fig. 6, left); 2) replacement of the C terminus, on the other hand, leads to a chimeric channel with faster activation and deactivation kinetics (Fig. 4, right) and with total lack of response to cAMP (Fig. 6, middle). These data suggest that the C terminus of rbHCN4 channels participates in controlling voltage-dependent gating; the gating properties are however not determined by the C terminus alone because the 4-C1 chimera did not acquire the kinetics of mHCN1 fully (Fig. 4,right). This indicates that the C terminus could interact with other adjacent regions of the channel to exert its controlling action. Because replacement of the N terminus of 4-C1 with that of mHCN1 did not have any appreciable effects on channel kinetics (Fig.5), channel regions interacting with the C terminus are unlikely to be located in the N terminus and more likely reside within the intracellular S-linker regions and/or the innermost regions of S domains. An interaction of the C terminus with intracellular channel regions is also required to interpret the different response to cAMP of the mutant channels investigated. If binding of cAMP to the CNB domain were the only process responsible for the cAMP-mediated gating, replacement of the rbHCN4 C terminus with the one from mHCN1 should yield a cAMP dependence analogous to that of the wild-type mHCN1 channel. However, neither the single (4-C1) nor the double rbHCN4 mutant (N1–4-C1) carrying the C terminus of mHCN1 had a measurable sensitivity to cAMP (Fig. 6). Thus, after binding cAMP the C terminus must interact with other channel regions to mediate the action of cAMP. Here again, because replacement of the N terminus did not affect the loss of cAMP dependence of the C-terminal mutant 4-C1, channel sites interacting with the C terminus are unlikely to reside in the N terminus. Our data do not allow us to discriminate whether the C terminus-mediated process affecting voltage gating and that affecting the cAMP action share common intracellular mechanisms. However, until a more specific investigation of the residues involved is performed, this remains an intriguing possibility. Although replacement of the rbHCN4 C terminus by that of mHCN1 caused strong acceleration of activation and deactivation rates, it did not modify the activation curve (Fig. 4 b). This apparent contradiction can in fact be reconciled with a recent allosteric model of voltage gating of HCN channels (15Altomare C. Bucchi A. Camatini E. Baruscotti M. Viscomi C. Moroni A. DiFrancesco D. J. Gen. Physiol. 2001; 117: 519-532Crossref PubMed Scopus (128) Google Scholar). According to this model, the position of the activation curve reflects two mechanisms: one is the voltage dependence of closed/open transitions, and the other is the voltage dependence of voltage sensor displacement. Because the voltage dependence of closed/open transitions appears to be nearly equivalent for different HCN isoforms, changes in the position of the activation curve reflect mostly changes in the voltage dependence of the transitions of the sensors. Therefore according to this interpretation, lack of modification of the activation curve by substitution of the C terminus would simply reflect a lack of interaction between the C terminus and the distribution of S4 voltage sensors between the reluctant and willing states. Previous investigation of native f-channels in sinoatrial node myocytes has shown that perfusion with Pronase of the intracellular side of the membrane leads to a strong activation of channels by a >50-mV depolarizing shift of the activation curve (27Barbuti A. Baruscotti M. Altomare C. Moroni A. DiFrancesco D. J. Physiol. (Lond.). 1999; 520–3: 737-744Crossref Scopus (27) Google Scholar). This was interpreted to indicate the presence of an inhibitory mechanism operated by intracellular channel regions that were cleaved by Pronase. This effect was accompanied by abolishment of cAMP sensitivity, implying that the intracellular region responsible for the inhibitory action could reside on the C terminus. The data presented here agree with the above interpretation by indicating the possibility that the C terminus interacts with intracellular regions to modulate voltage- and cAMP-dependent gating. Disruption of intracellular regions leads to strong channel activation, as expected if the C terminus exerted an inhibitory effect on channel gating. An inhibitory interaction with intracellular channel regions might also explain the action of cAMP if cAMP binding occurs more favorably in the open state of the channel (4DiFrancesco D. J. Physiol. (Lond.). 1999; 512: 367-376Crossref Scopus (91) Google Scholar), when the inhibitory action is offset, in agreement with the view that cAMP actually operates by partial removal of the inhibitory action of the C terminus. We thank Drs. B. Santoro and S. A. Siegelbaum (Columbia University, New York) for kindly providing mHCN1 and Dr. H. Ohmori (University of Kyoto) for kindly providing rbHCN4.
Referência(s)