Article Information

• PDF (287 kb)
• Full Text with Large Figures

PubMed

• Articles by Thomas Zimmer
• Articles by Klaus Benndorf

Related Articles

• D1 Receptors Physically Interact with N-Type Calcium Channel...

  D1 Receptors Physically Interact with N-Type Calcium Channels to Regulate Channel Distribution and Dendritic Calcium Entry Neuron, Volume 58, Issue 4, 22 May 2008, Pages 557-570 Alexandra E. Kisilevsky, Sean J. Mulligan, Christophe Altier, Mircea C. Iftinca, Diego Varela, Chao Tai, Lina Chen, Shahid Hameed, Jawed Hamid, Brian A. MacVicar and Gerald W. Zamponi Summary Dopamine signaling through D1 receptors in the prefrontal cortex (PFC) plays a critical role in the maintenance of higher cognitive functions, such as working memory. At the cellular level, these functions are predicated to involve alterations in neuronal calcium levels. The dendrites of PFC neurons express D1 receptors and N-type calcium channels, yet little information exists regarding their coupling. Here, we show that D1 receptors potently inhibit N-type channels in dendrites of rat PFC neurons. Using coimmunoprecipitation, we demonstrate the existence of a D1 receptor-N-type channel signaling complex in this region, and we provide evidence for a direct receptor-channel interaction. Finally, we demonstrate the importance of this complex to receptor-channel colocalization in heterologous systems and in PFC neurons. Our data indicate that the N-type calcium channel is an important physiological target of D1 receptors and reveal a mechanism for D1 receptor-mediated regulation of cognitive function in the PFC. Summary | Full Text | PDF (1580 kb)

• Caveolin-3 Is Adjacent to a Group of Extradyadic Ryanodine R...

  Caveolin-3 Is Adjacent to a Group of Extradyadic Ryanodine Receptors Biophysical Journal, Volume 89, Issue 3, 1 September 2005, Pages 1893-1901 David R.L. Scriven, Agnieszka Klimek, Parisa Asghari, Karl Bellve and Edwin D.W. Moore Abstract Caveolae are present in almost all cells and concentrate a wide variety of signaling molecules, receptors, transporters, and ion pumps. We have investigated the distribution of the ryanodine receptor, the Na/Ca exchanger, the predominant Na channel isoform rH1, and the L-type calcium channel, Ca1.2, relative to the muscle-specific caveolin isoform, caveolin-3, in adult rat ventricular myocytes. Three-dimensional immunofluorescence images were deconvolved and analyzed. Caveolin-3 colocalizes with all of these molecules at the surface of the cell, but there is no significant colocalization between caveolin-3 and either the Na/Ca exchanger or the Na channel in the cell interior. The distribution of the surface colocalization indicates that the caveolae that colocalize with each molecule form distinct populations. This organization indicates that there are multiple populations of caveolae separable by location and occupants. In the interior of the cell, caveolin-3 shows a marked colocalization with a population of ryanodine receptors that are separate from those within the dyad. Because of their location, the signaling molecules contained within these caveolae may have preferred access to the neighboring nondyadic ryanodine receptors. Abstract | Full Text | PDF (916 kb)

• The Tumor Suppressor eIF3e Mediates Calcium-Dependent Intern...

  The Tumor Suppressor eIF3e Mediates Calcium-Dependent Internalization of the L-Type Calcium Channel CaV1.2 Neuron, Volume 55, Issue 4, 16 August 2007, Pages 615-632 Eric M. Green, Curtis F. Barrett, Geert Bultynck, Steven M. Shamah and Ricardo E. Dolmetsch Summary Voltage-gated calcium channels (VGCCs) convert electrical activity into calcium (Ca) signals that regulate cellular excitability, differentiation, and connectivity. The magnitude and kinetics of Ca signals depend on the number of VGCCs at the plasma membrane, but little is known about the regulation of VGCC surface expression. We report that electrical activity causes internalization of the L-type Ca channel (LTC) Ca1.2 and that this is mediated by binding to the tumor suppressor eIF3e/Int6 (eukaryotic initiation factor 3 subunit e). Using total internal reflection microscopy, we identify a population of Ca1.2 containing endosomes whose rapid trafficking is strongly regulated by Ca. We define a domain in the II-III loop of Ca1.2 that binds eIF3e and is essential for the activity dependence of both channel internalization and endosomal trafficking. These findings provide a mechanism for activity-dependent internalization and trafficking of Ca1.2 and provide a tantalizing link between Ca homeostasis and a mammalian oncogene. Summary | Full Text | PDF (2216 kb)

• …more

Copyright © 2007 The Biophysical Society. All rights reserved.
Biophysical Journal, Volume 92, Issue 11, 3885-3892, 1 June 2007

doi:10.1529/biophysj.106.098889

Channels, Receptors, and Electrical Signaling

The Intracellular Domain of the β2 Subunit Modulates the Gating of Cardiac Na_v1.5 Channels

Thomas Zimmer^,   and Klaus Benndorf

Institute of Physiology II, Friedrich Schiller University, Jena, Germany

Address reprint requests to Thomas Zimmer, PhD, Institute of Physiology II, Friedrich Schiller University Kollegiengasse 9, 07743 Jena, Germany. Tel.: 49-3641-934372; Fax: 49-3641-933202.

The α-subunits of mammalian voltage-gated Na⁺ channels constitute a family of 10 different members 1 that are responsible for the initiation and propagation of action potentials in electrically excitable cells 2. These subunits interact with one or two small accessory β-subunits, thereby forming heteromultimeric protein complexes in the plasma membrane. As demonstrated by heterologous expression experiments, the α-subunit determines the main electrophysiological and pharmacological properties of a given Na⁺ channel 2. Functional analysis on the known β-subunits revealed their multifunctional nature: first, β1 and β3 subunits shift steady-state gating, accelerate Na⁺ current decay kinetics and recovery from inactivation, and increase the current amplitude of several α-subunits 3,4,5,6,7,8,9,10,11,12,13,14,15,16,17. The β1-induced gating effects on α-subunits strongly depend on the degree of sialylation of the extracellular β1 domain 18. Second, β1 subunits are capable of interfering with the action of fatty acids 19, pharmaceuticals 20, or natural toxins 21 on α-subunits. Third, β-subunits contain extracellular immunoglobulin (Ig) domains that are structurally homologous to the V-set of the Ig superfamily 22. They can act as cell adhesion molecules (CAM) 23 and recruit via intracellular interactions cytoskeletal proteins, such as ankyrin, to points of cell-cell contact 24. Thus, β-subunits serve as important communication links between the extracellular and intracellular environment of neurons. The conclusions from these in vitro studies were strongly supported by recent data on β1^(−/−) and β2^(−/−) knock-out mice 25,26,27. These experiments demonstrated not only the significance of both β-subunits for regulation of neuronal Na⁺ channel density and localization, but also the fundamental function of β1 for nodal architecture and neurite outgrowth, thus pointing to the in vivo action of accessory β-subunits as CAM in the central and peripheral nerve system.

In the mammalian heart, the Na⁺ channel isoform Na_v1.5 dominates the I_Na in atrial and ventricular cardiomyocytes 1,28. Studies on the subunit composition of cardiac Na⁺ channels suggested an in vivo association of Na_v1.5 with β1 and β2 subunits 29. Immunocytochemical studies further demonstrated a remarkable cardiac expression of both β-subunits by Western blotting 29,30, and suggested a preferential colocalization of Na_v1.5 channels with β2 instead of β1 subunits at intercalated disks of ventricular cardiomyocytes 31. However, electrophysiological measurements in mammalian cell lines or Xenopus oocytes showed that only the β1 but not the β2 subunit exerts modulatory effects on Na_v1.5 gating 15,29. Thus, the lack of any electrophysiological effect of β2 on Na_v1.5 channels conflicts with the immunocytochemical data. A possible explanation could be that, in contrast to the situation in native cardiomyocytes, Na_v1.5 does not associate with β2 upon heterologous expression, as concluded from a missing colocalization of respectively labeled subunit constructs 32.

We observed previously that the β1 transmembrane segment was required for a current increase and for an accelerated recovery from inactivation of Na_v1.5 channels 15. To exert β1-like effects on Na_v1.5, either the extracellular or the intracellular domain of the β1 subunit was necessary in addition to the short membrane spanning region. For example, deletion of the β1 intracellular domain (β11Δ) resulted in a faster recovery from inactivation and produced larger whole-cell currents, indicating that this shortened β1 variant still efficiently interacts with Na_v1.5. In contrast, modulatory effects of the β2 subunit on Na_v1.5 are unknown. To create a functional channel complex in which the β2 intracellular domain is in close contact with Na_v1.5, we attached this β2 sequence to β11Δ resulting in β112. We expected that possible functional effects of the β2 intracellular domain may become visible when the spatial requirements for an α/β2 interaction are fulfilled. In this study, we used such a β-subunit construct and observed severe modulation of Na_v1.5 function during recovery from inactivation that has not been observed in voltage-gated Na⁺ channels before.

Plasmids pSP64T-hH1, pNa200, and pSPNaβ coding for human Na_v1.5 (hH1, accession No. M77235), for Na_v1.2 (rat brain IIA Na⁺ channel, accession No. X61149), and for the rat β1 subunit (accession No. M91808) were kindly provided by Dr. A. L. George (Vanderbilt University), Dr. A. L. Goldin (University of California), and Dr. W. Stühmer (Max Planck Institute, Göttingen), respectively. The β2 subunit (accession No. U37026) was isolated by RT-PCR from the human brain astrocytoma cell line 1321N1, as previously described 32. The isolation of sequences for mouse Na_v1.4 (mH2; accession No. AJ278787) and mouse Na_v1.5 (mH1; accession No. AJ271477) has been previously described 33.

We subcloned the cDNAs of the β1 and β2 subunit, and of deletion variant β11Δ into the in vitro transcription vector pGEMHEnew, resulting in plasmids pGEM-β1, pGEM-β2, and pGEM-β11Δ, as previously described 15. This vector contains the T7 promoter, a 5′-untranslated region (UTR) of the Xenopus β-globin gene and a multicloning site. To create the construct β112, the desired β1 and β2 subunit regions were first separately amplified by PCR and then linked by a recombinant PCR step using the internal primer pair 5′-ACTTGACCACCATCTCCGCCACGAGCCATA-3′ and 5′-GGCGGAGATGGTGGTCAAGTGTGTGAGGAG-3′. The recombinant fragment was subcloned into the BamHI/HindIII sites of pGEMHEnew resulting in pGEM-β112. PfuTurbo DNA polymerase (Stratagene, La Jolla, CA) was used for all PCR reactions to minimize PCR-mediated nucleotide exchanges. The plasmid was expected to encode chimera β112 consisting of amino acids M1–M178 of β1 plus amino acids V179–K215 of β2 (amino acid numbering according to the premature full-length proteins). Mouse Na_v1.4 was subcloned into pSP64T by releasing the hH1 coding region in pSP64T-hH1 using BglII/SmaI digestion and inserting a BamHI/XhoI fragment of mouse Na_v1.4 (mH2) released from the respective pTSV40Gnew vector 33, resulting in plasmid pSP64T-mNa_v1.4. To allow for a respective blunt-end ligation, the XhoI site was treated with Klenow enzyme. Mouse Na_v1.5 was subcloned into a pSP64-PolyA derivative (Promega, Mannheim, Germany) by releasing the β1 subunit region in pSPNaβ using HindIII/HincII digestion and inserting a HindIII/NotI fragment of mouse Na_v1.5 (mH1) released from pTSV40Gnew-mH1 33 after treatment of the NotI site with Klenow enzyme, resulting in vector pSP-mNa_v1.5. The correctness of the DNA constructs was checked by the dideoxy DNA sequencing method.

Capped cRNAs were prepared by digestion of respective plasmids with NotI (pNa200 and pSP64T-mNa_v1.4), EcoRI (pSP-mNa_v1.5), SpeI (pSP64T-hH1), XbaI (pGEM-β1 and pGEM-β11Δ), and HindIII (pGEM-β112 and pGEM-β2), followed by in vitro transcription reaction with SP6 (pSP64 derivatives) and T7 (pNa200 and all pGEM derivatives) RNA polymerase (Roche Diagnostics GmbH, Mannheim, Germany). Thus, the cRNAs of all Na⁺ channel isoforms and of the β-subunit variants were composed of the β-globin 5′-UTR and the respective α- or β-subunit sequence.

Oocytes from Xenopus laevis were obtained as previously described 32. Glass micropipettes were used to inject a volume of capped cRNA per oocyte of ∼40–60 nl. The final cRNA concentrations used to inject oocytes were 0.02μg/μl (Na_v1.5) and 0.01μg/μl (Na_v1.2, Na_v1.4). The different cRNAs encoding the β-subunit variants were at a concentration of ∼0.1μg/μl so that the final molar ratio of α- to β-subunit was ∼1:50 at the cRNA level. Injected oocytes were incubated for 3 days at 18°C in Barth medium. Under those conditions the amplitudes of I_Na, measured 3 days after injection at the test potential of -25mV (Na_v1.5) and −10mV (Na_v1.2, Na_v1.4), were between 0.5 and 5.0μA. The recovery from inactivation was determined from Na⁺ currents with an amplitude between 1 and 4μA.

Whole-cell Na⁺ currents were recorded with the two-microelectrode voltage clamp technique using a commercial amplifier (OC725C, Warner Instruments, Hamden, CT). The glass microelectrodes were filled with 3M KCl. The microelectrode resistance was between 0.2 and 0.5 MΩ. The bath solution contained (in mM): 96 NaCl, 2 KCl, 1.8 CaCl₂, 10 Hepes/KOH, pH 7.2. The currents were elicited by test potentials from −80 to 40mV (holding potential of −120mV). The pulsing frequency was 0.1Hz. We used only cells that produced a peak current amplitude <5μA. Steady-state activation (m_∞) was evaluated by fitting the Boltzmann equation m_∞={1+exp[−(V−V_m)/s]}⁻¹ to the normalized conductance as function of voltage. V is the test potential, V_m the midactivation potential, and s the slope factor in mV. Recovery from inactivation was determined with a standard protocol (Fig. 1) at a frequency of 0.1Hz. Recording and analysis of the data were performed on a PC with the ISO2 software (MFK, Niedernhausen, Germany). The sampling rate was generally 20kHz.

Display large version of this figure

Figure 1

Atypical recovery from inactivation in Na_v1.5 channels produced by the intracellular domain of the β2 subunit in β112. The figure illustrates the domain structures of β1, β11Δ, and β112, and the corresponding Na_v1.5 current traces at the indicated recovery intervals. The respective terminal amino acids of the β1 (black) and β2 (gray) subunit regions are indicated. We observed an accelerated recovery process and a current increase when coexpressing wild-type β1 subunit or deletion variant β11Δ. Recovery time constants τ_rec determined with monoexponential fits were (in ms): 5.87±0.18 for Na_v1.5, 3.20±0.13 for Na_v1.5/β1, and 3.95±0.29 for Na_v1.5/β11Δ (n=8–15). Peak-current amplitudes relative to Na_v1.5 increased significantly by factor 3.47±0.32 (β1) and 1.33±0.15 (β11Δ). Attachment of the β2 intracellular domain to the β1 transmembrane segment produced a paradoxical recovery behavior of wild-type Na_v1.5 channels. The pulse protocol is illustrated (pulse frequency 0.1Hz). Calibration bars: 20ms and 0.5μA.

To test for a possible effect of the intracellular domain of the β2 subunit on Na_v1.5, we attached this domain to the transmembrane plus extracellular domain of β1, resulting in chimera β112. We coexpressed the chimera with Na_v1.5 channels in Xenopus oocytes and investigated whole-cell currents by the two-microelectrode voltage-clamp technique. Coexpression of chimera β112 produced an Na_v1.5 recovery behavior that was not seen before in any other voltage-gated Na⁺ channel (Fig. 1, bottom): Short recovery intervals up to ∼20ms allowed Na_v1.5 channels to recover efficiently from inactivation resulting in increased current amplitudes. This process occurred almost as fast as in the case of Na_v1.5/β1 channels (Figure 2A, right) and should be therefore mediated by β1 regions in chimera β112. Longer recovery intervals, however, led to a paradoxical decrease of the current (Fig. 1, bottom). After ∼1s, a steady-state current level with an amplitude of 43±2% of the value at 20ms was reached (Figure 2A, left). We observed this atypical recovery in both mouse and human Na_v1.5 channels coexpressed with β112, but neither in Na_v1.2 nor in Na_v1.4 channels: Coexpression of β112 with the neuronal or skeletal muscle isoform produced typical β1-like effects, as accelerated recovery from inactivation (Figure 2BC), increased current densities, and faster inactivation (data not shown). In conclusion, a close contact of cardiac-specific Na_v1.5 with the β2 intracellular domain led to a distinct interaction of both proteins, producing an atypical recovery from inactivation.

Display large version of this figure

Figure 2

Recovery from inactivation in mouse Na_v1.5 (A), rat Na_v1.2 (B), and mouse Na_v1.4 (C) channels upon coexpression of β112. The figure shows that Na_v1.5/β112 currents decreased at recovery intervals >20ms and reached a steady-state level after 1–2s with current amplitudes of <50% of the value at 20ms. In case of human Na_v1.5 very similar data were consistently obtained (data not shown). In Na_v1.2 and Na_v1.4 channels, an atypical recovery was not observed. Instead, the typical β1-like modulatory effect on recovery was obtained when coexpressing β112. Number of measurements: n=9 for Na_v1.5, n=15 for Na_v1.5/β1, n=15 for Na_v1.5/β112, n=7 for Na_v1.2, n=6 for Na_v1.2/β1, n=6 for Na_v1.2/β112, n=9 for Na_v1.4, n=5 for Na_v1.4/β1, and n=6 for Na_v1.4/β112. For pulse protocol see Fig. 1 (test pulse to −30mV for Na_v1.5, and to −20mV for Na_v1.2 and Na_v1.4). Error bars represent mean±SE.

To exclude that an irreversible “run-down” caused the time-dependent current decrease at hyperpolarized potentials, we successively prolonged the recovery intervals Δt from 1ms to 7s, and shortened the time periods again from 7s to 1ms using the same oocyte. We found that the atypical recovery was reversible (Fig. 3).

Display large version of this figure

Figure 3

The atypical recovery in Na_v1.5/β112 channels was reversible. The recovery interval Δt was successively prolonged until 7s (■) and then again shortened to 1ms (○). The figure illustrates fractional recovery at intervals up to 7s (left) and up to 300ms (right). For pulse protocol see Fig. 1 (n=6). Error bars represent mean±SE.

We next investigated whether Na_v1.5 inactivation kinetics were affected by β112 (Fig. 4). After a recovery interval of 15ms, inactivation time constants τ_h were similar for Na_v1.5, Na_v1.5/β1, and Na_v1.5/β112 channels (e.g., τ_h in ms at −25mV: 2.05±0.14 for Na_v1.5, 2.04±0.08 for Na_v1.5/β1, and 2.17±0.17 for Na_v1.5/β112; n=5). At the same time, current amplitudes were significantly larger when coexpressing either β1 or β112, compared to cells expressing Na_v1.5 alone (Figure 5A). This current increase is mediated by the extracellular domain and the transmembrane segment of β1, because it was also found when coexpressing β11Δ 15. Upon longer recovery intervals, the β2 intracellular domain successively decreased the current amplitude and concomitantly decelerated the inactivation kinetics (Figure 4B and Figure 5B). Interestingly, a minor deceleration of inactivation upon longer recovery intervals occurred also in Na_v1.5 and Na_v1.5/β1 channels (Fig. 4).

Display large version of this figure

Figure 4

Effect of β112 on inactivation kinetics of Na_v1.5 channels. Current traces from Na_v1.5, Na_v1.5/β1, and Na_v1.5/β112 channels at −25mV, and corresponding plots of the inactivation time constants τ_h as function of voltage after recovery intervals of 15ms (■) and 4000ms (○). Please note that a significant deceleration of inactivation upon longer recovery intervals occurred also in Na_v1.5 and Na_v1.5/β1 channels (p<0.05 for τ_h values in ms at −10mV: 1.29±0.05 at 15ms versus 1.70±0.08 at 4000ms for Na_v1.5, and 1.18±0.07 at 15ms versus 1.72±0.04 at 4000ms for Na_v1.5/β1). Calibration bar is: 5ms, and 0.9μA for Na_v1.5, 1.9μA for Na_v1.5/β1, 1.0μA for Na_v1.5/β112 at 15ms, and 0.3μA for Na_v1.5/β112 at 4000ms. Number of experiments is: n=6 for Na_v1.5, n=5 for Na_v1.5/β1, and n=5 for Na_v1.5/β112. Error bars represent mean±SE.

Display large version of this figure

Figure 5

Effect of the β2 intracellular domain on current amplitude and inactivation time constants in Na_v1.5 channels. (A) Current amplitudes at recovery intervals Δt of 15 and 1000ms (n=5 each). At Δt=15ms, the modulatory effect of the β1 region on Na_v1.5 current amplitude can be observed also when coexpressing β112 (* indicates p<0.05 versus Na_v1.5 at 15ms). However, at longer recovery intervals currents through Na_v1.5/β112 channels decreased successively (compare also Fig. 1 to Fig. 3). (B) Current amplitude as a function of the inactivation time constant τ_h in Na_v1.5/β112. The decreased current amplitudes upon longer recovery intervals are accompanied by decelerated inactivation. Data points were obtained from five independent measurements (illustrated by the five different symbols). The increase of τ_h was caused by successively prolonging the recovery interval (15, 50, 100, 300, 500, and 1000ms; recovery potential, −120mV).

These data suggested that, upon interaction with the β2 intracellular domain, Na_v1.5 channels enter an additional closed state at hyperpolarized potentials in a time-dependent manner. As shown in Fig. 6, Na_v1.5/β112 channels captured in this closed state (−120mV for 10s) could not escape when stepping the voltage of the prepulse up to −80mV: The currents at the test pulse were constant at nearly 30% of the maximal current. Further prepulse depolarizations allowed for a transition to the open and/or inactivated state during this prepulse. During the recovery pulse of 15ms at −120mV, channels recovered from inactivation. This interval was too short for a full transition to the additional closed state. Consequently, channel activation at the test pulse was facilitated and currents increased. Such an atypical recovery behavior was not observed in Na_v1.5/β1 channels: The 15-ms recovery interval was too short for a complete recovery of channels that were inactivated during the prepulse. Consequently, the current at the test pulse decreased (Fig. 6).

Display large version of this figure

Figure 6

Effect of the prepulse potential on recovery of Na_v1.5/β112 channels at a short recovery interval (15ms). The figure illustrates the relative current amplitude (normalized to −1) at the test pulse (−30mV) as a function of the prepulse potential (from −120 to +20mV). Number of measurements, n=7 each. Error bars represent mean±SE.

Reduced whole-cell currents and slowed inactivation in Na_v1.5/β112 channels at longer recovery intervals suggested that the current decrease during the recovery from inactivation is at least to some extent the result of a shift of steady-state activation toward depolarized potentials. To test this hypothesis, we compared steady-state activation curves at the recovery intervals of 15 and 4000ms. At 15ms, we obtained similar midactivation potentials (V_m) for Na_v1.5, Na_v1.5/β1, and Na_v1.5/β112 channels (Table 1). At 4000ms, V_m values did not change in case of Na_v1.5 and Na_v1.5/β1 channels (p>0.05). In contrast, chimera β112 significantly shifted Na_v1.5 steady-state activation by ∼9mV to depolarized potentials (15 vs. 4000ms; Table 1, Fig. 7). This shift can fully explain the observed whole-cell current reduction during recovery from inactivation. When applying a test pulse around the midactivation potential (−30mV), the respective depolarized shift of the steady-state activation curve decreased the channel number responding to the voltage stimulus. Consequently, the phenomenon of reduced peak currents upon longer recovery intervals was not seen when applying a test pulse to 0mV (Figure 7C). However, the altered steady-state activation alone cannot explain the significantly decelerated inactivation kinetics in Na_v1.5/β112 channels. Current decay time constant remained markedly larger also at test potentials more positive than 0mV (Fig. 4).

Display large version of this figure

Figure 7

Longer recovery intervals shift steady-state activation in Na_v1.5/β112 channels. (A) Currents of Na_v1.5/β112 channels measured after the indicated recovery time points. (B) Representative steady-state activation curves in Na_v1.5/β112 channels for recovery intervals of 15ms (■) and 4000ms (○). For midactivation potentials (V_m) and statistics, see Table 1. (C) Recovery from inactivation in Na_v1.5/β112 channels at test potentials of 0 vs. −30mV. Peak current reduction at −30mV depended on the shift of the steady-state activation curve. At 0mV, all available channels are activated and consequently, a current reduction was not obtained. Number of measurements, n=5. All values were normalized with respect to the maximum current obtained at −30mV.

Finally, we investigated the voltage dependence of the atypical recovery and compared the effect of two distant recovery potentials (−80 and −140mV). As shown in Figure 8A (left), we observed a much faster initial recovery at −140mV compared to −80mV. In control experiments with Na_v1.5/β1 channels, recovery was much faster at the more hyperpolarized potential (Figure 8A, right) and both recovery potentials produced the maximal steady-state current level. In contrast to this, availability of Na_v1.5/β112 channels passed a maximum and reached finally a steady-state value at ∼30% of the maximal current. Interestingly, this steady state was independent of the recovery potential: We observed a similar steady-state level at both −140 and −80mV. Consequently, when investigating the effect of the β2 intracellular domain in β112 on both Na_v1.5 inactivation time course and steady-state activation, we obtained similar data for both potentials at long recovery intervals (Figure 8BC). Therefore, we conclude that the channel transition to the additional closed state at potentials negative to −80mV is voltage independent.

Display large version of this figure

Figure 8

Atypical recovery from inactivation in Na_v1.5/β112 channels at recovery potentials of −140 and −80mV. (A) Recovery from inactivation in Na_v1.5/β112 (left) and Na_v1.5/β1 (right) channels. Shorter intervals led to a faster recovery at the more hyperpolarized potential in both Na_v1.5/β112 and Na_v1.5/β1 channels. In contrast to Na_v1.5/β1, atypical recovery occurred in Na_v1.5/β112 at both −140 and −80mV, resulting in similar current levels at intervals >0.3s independently of the recovery potential. (B) Effect of the β2 intracellular domain on Na_v1.5 steady-state activation after long recovery intervals. A clear shift of steady-state activation occurred at −140mV after longer recovery intervals (* p<0.05; midactivation potentials V_m, −31.1±1.8mV at 15ms versus −23.9±1.9mV at 4000ms), similarly as observed at −120mV (see Table 1). At 4000ms, steady-state activation was similar at −80mV (V_m, −22.8±1.7mV; n=16) compared to −140mV. Neither the recovery interval nor the recovery potential had an effect on Na_v1.5/β1 channels (at −140mV, V_m=−30.2±1.6 at 15ms and −30.3±2.0 at 4000ms; at −80mV, V_m=−28.9±1.5mV at 4000ms; n=11). (C) Inactivation time constant τ_h as function of voltage at a recovery interval of 4000ms. Deceleration of inactivation occurred only in Na_v1.5/β112 and did not depend on the recovery potential (n=16 for Na_v1.5/β112, n=11 for Na_v1.5/β1). Error bars represent mean±SE.

In this study we report the first functional effects of a specific β2 sequence on cardiac Na_v1.5 channels: The β2 intracellular domain is capable of generating an atypical recovery behavior of cardiac Na_v1.5 channels. At longer recovery intervals the β2 intracellular domain shifted steady-state activation toward depolarized potentials and decelerated channel inactivation (Figure 4 and Figure 7). Thus, the paradoxical current decrease upon longer recovery intervals observed at test potentials negative to −10mV (e.g., −30mV in Fig. 1) is not simply caused by an overshooting recovery from inactivation, as observed previously in voltage-gated K⁺ channels 34. Our data suggest the existence of a novel closed state in Na_v1.5 channels accessible at hyperpolarized potentials that does neither occur in neuronal Na_v1.2 nor in skeletal muscle Na_v1.4 channels.

How can we interpret our results in terms of an Na_v1.5 gating model? We suggest that at hyperpolarized voltages, channels switch rapidly from the inactivated to the normal closed state in a voltage-dependent manner. This process is largely complete within 15–30ms. After these short recovery intervals, channel activation produced current amplitudes and inactivation kinetics that were similar to those seen for Na_v1.5/β1 channels (see increased current amplitudes upon coexpression of either β1 or β112 in Figure 5A). Prolongation of the recovery interval, however, favors a transition to an additional closed state, until a steady state is reached (after ∼1s). This transition is voltage independent at potentials negative to −80mV (see Fig. 8). Channel activation from this additional closed state requires more positive voltages. At the same time, inactivation kinetics are decelerated. In conclusion, our data indicate that Na_v1.5 channels can be inactivated also by hyperpolarization. Inactivation kinetics have been demonstrated more than 20 years ago to be caused by the first latency of the channels 35. Single-channel analysis would be required to define the contribution of a possibly altered first latency to the biophysical phenomenon observed with Na_v1.5/β112 channels. Significantly larger inactivation time constants at test pulse potentials positive to 0mV suggest that also other single-channel properties are altered by β112: A test pulse to 0mV increased whole-cell currents and decelerated inactivation kinetics upon longer recovery intervals (compare Figure 4C and Figure 7C). It is interesting to note that a small Na_v1.5 channel fraction obviously enters the additional closed state also when coexpressing β1 or even in the absence of any other subunit (see decelerated inactivation at recovery interval of 4000ms; left and middle panel in Fig. 4).

Although we measured currents through wild-type Na⁺ channels, we used an artificial construct to provoke the unusual Na_v1.5 gating phenomenon. Consequently, our data raise the question why such an atypical recovery has not yet been observed when coexpressing wild-type β2? Previously, we found that Na_v1.5 and β2 are not colocalized upon heterologous expression 32, in contrast to immunocytochemical data obtained in cardiomyocytes 31. However, missing colocalization is incompatible with an efficient intracellular assembly and trafficking of the channel complex to the plasma membrane, and thus also with a modulatory action of, respectively, associated β-subunits 15,32. When anchoring the intracellular domain of the β2 subunit to a β1 subunit deletion variant that interacts with Na_v1.5 (β11Δ; 15), we reconstituted a channel complex in which the β2 sequence should interact with Na_v1.5. In respective control experiments, we investigated the subcellular localization of Na_v1.5/β11Δ and Na_v1.5/β112 channels using different variants of the green fluorescent protein, and we indeed observed a strict α/β colocalization that was not seen when using wild-type β2 (data not shown).

However, we have to consider that the β2 intracellular domain may not interact at the correct β2 binding site in Na_v1.5 when using chimera β112. It is imaginable that β1 and β2 normally bind distinct sites on α-subunits. Thus, it is possible that chimera β112 binds specifically at the β1 site, thereby directing the β2 intracellular domain to Na_v1.5 regions normally interacting with β1. Moreover, protein folding of the intracellular domain of β2 may be affected in the chimera, despite an obviously normal trafficking and correct localization in the plasma membrane of heterologous host cells. Consequently, the paradoxical recovery effect could be rather specific for the chimera and may not be produced in vivo by wild-type β2. The fact that the recovery behavior reported in this study has not been seen in native cardiac Na⁺ channels before is consistent with this idea 36.

In conclusion, the results of our study provide more insight into the biophysical properties of wild-type Na_v1.5 channels by suggesting the existence of an additional closed state at hyperpolarized potentials that has not been detected so far. This state does neither exist in neuronal nor skeletal muscle Na⁺ channels, providing further evidence for the unique feature of Na_v1.5 to interact with β-subunits also at intracellular regions. Further research is necessary to provide more insight into the nature of this additional closed state by identifying the interacting amino acid residues in Na_v1.5. At the same time we have to notice that conclusions on the physiological significance of our results for an in vivo Na_v1.5/β2 interaction in the heart are still strictly limited.