The voltage-gated potassium channel Kv1.3 plays a crucial role in the immune system response. In leukocytes, the channel is coexpressed with the dominant negative regulatory subunit KCNE4, which associates with Kv1.3 to trigger intracellular retention and accelerating C-type inactivation of the channel. Previous research has demonstrated that the main association between these proteins occurs through both C-termini. However, these data fail to fully elucidate the KCNE4-dependent modulation of channel kinetics. In the present study, we analyzed the contribution of each KCNE4 domain to the modulation of Kv1.3. Our results further confirmed that the C-terminus of KCNE4 is the main determinant involved in the association-triggered intracellular retention of the channel. Additionally, interactions throughout the transmembrane region were also observed. Both the C-terminus and, especially, the transmembrane domain of KCNE4 accentuated the C-type inactivation of Kv1.3. Our data provide, for the first time, the molecular effects that a KCNE peptide, such as KCNE4, exerts on a Shaker channel, such as Kv1.3. Our results pave the way for understanding the molecular mechanisms underlying potassium channel modulation and suggest that KCNE4 participates in the conformational rearrangement of the Kv1.3 architecture, altering the C-type inactivation of the channel.

Constructs

Rat (r)Kv1.3 in pRcCMV was obtained from T. C. Holmes (University of California–Irvine, Irvine, CA, USA) and introduced into pCDNA3.1, pEYFPC1 and pECFP-C1 (Clontech Laboratories). Kv1.3, tagged with hemagglutinin (HA) extracellularly between S3 and S4, was obtained from D. B. Arnold (University of Southern California, Los Angeles, CA, USA). Murine (m)KCNE4 in pSGEM was a gift from M. Sanguinetti (University of Utah, Salt Lake City, UT, USA). Human (h)KCNE2 in pHA was obtained from S. de la Luna (Centre for Genomic Regulation, Barcelona, Spain). mKCNE4 and hKCNE2 were subcloned and inserted into pECFP-N1 and pEYFP-N1. The plasma membrane marker pDsRed-tagged pleckstrin homology (PH) domain of Akt (Akt-PHpDsRed) was obtained from F. Viana (Instituto de Neurociencias de Alicante, Universidad Miguel Hernández-CSIC, San Juan de Alicante, Spain).

MluI sites were introduced into CFP-tagged mKCNE4 and hKCNE2 to isolate either their N-terminal (Nt), transmembrane (TMD) or C-terminal (Ct) domains. Next, these fragments were obtained by PCR and subcloned and inserted into reciprocal KCNE plasmids. Several KCNE2/4 chimeras (X-X-X) were generated. KCNE(X-X-X) identifies the Nt, TMD and Ct fragments from each peptide: KCNE(2-2-4)-CFP (Nt and TMD of KCNE2 with the Ct of KCNE4), KCNE(4-4-2)-CFP (Nt and TMD of KCNE4 with the Ct of KCNE2), KCNE(2-4-2)-CFP (Nt and Ct of KCNE2 with the TMD of KCNE4), KCNE(4-2-4)-CFP (Nt and Ct of KCNE4 with the TMD of KCNE2), KCNE(4-2-2)-CFP (TMD and Ct of KCNE2 with the Nt of KCNE4) and KCNE(2-4-4)-CFP (TMD and Ct of KCNE4 with the Nt of KCNE2). The full sequence of the KCNE2/4 chimeras can be found in Supplemental Fig. S1.

Mutations were generated with the QuikChange XL SiteDirected Mutagenesis Kit (Agilent Technologies) and transformed into XLGold^® ultracompetent cells. The mutants were subsequently sequenced using a BigDye Terminator v3.1 kit (Applied Biosystems) at CCiT University of Barcelona. The sequences were analyzed using Clone Manager Professional v7 (Sci Ed Software LLC).

Cell culture

HEK293 cells were seeded on 100-mm culture dishes for protein extraction and biochemistry, on 15x15 mm square poly-D-lysine-coated coverslips for immunocytochemistry, and on 35-mm dishes for electrophysiology experiments. The cells were cultured at 37 °C and 5% CO₂ in Dulbecco’s modified Eagle’s medium (DMEM; Lonza) supplemented with 10% fetal bovine serum (FBS; Gibco), 10,000 U/mL penicillin, 100 µg/mL streptomycin (Thermo Fisher Scientific) and 4 mM L-glutamine (Sigma‒Aldrich).

Cells were transfected 24 h before the experiment using Lipotransfectin^® (Attendbio Research). For coimmunoprecipitation, 4 µg of Kv1.3-HA and 4 µg of KCNE2/4 chimeras were transfected; for confocal microscopy, 500 ng of Kv1.3-YFP and 750 ng of the remaining constructs were transfected; and for patch clamp experiments, 150 ng of Kv1.3-pCDNA3.1 and 450 ng of the remaining constructs were transfected. No cell size changes were observed in any group after transfection being 15.52 ± 1.46, 15.52 ± 1.48, 15.41 ± 0.74 pF (n=5-6) for Kv1.3, KCNE4 and KCNE2 expressing cells, respectively,

Protein extraction, co-IP and Western blot

The cells were washed twice with cold PBS and lysed for 15 min at 4 °C with gentle mixing in lysis buffer (150 mM NaCl, 50 mM HEPES, 10% glycerol, 1% Triton X-100, pH 7.2) supplemented with protease inhibitors: 1 µg/mL leupeptin, 1 µg/mL pepstatin, 1 µg/mL aprotinin and 1 µM phenylmethylsulfonylfluoride (PMSF). The homogenates were centrifuged for 10 min at 16,200 × g at 4 °C, after which the supernatants were collected. The protein concentration was assessed using the Bio-Rad Protein Assay (Bio-Rad Laboratories).

For coimmunoprecipitation (coIP) experiments, 1 mg of protein per condition was precleared with 25 µL of protein G-Sepharose^® beads (GE Healthcare) for 1 h at 4 °C with gentle mixing. The beads were removed by centrifugation for 30 s at 1,000 × g. Next, the supernatants were incubated in a Micro Bio-Spin^® Chromatography column (Bio-Rad Laboratories) containing 50 µL of protein G-Sepharose^® beads and 2.5 µg of rabbit anti-GFP antibody (Genscript Biotech, Piscataway). The antibody had previously been crosslinked to Sepharose beads for 30 min at room temperature (RT, 20-25 °C) with 5.2 mg/mL dimethyl pimelimidate (DMP). The samples were incubated in the column for 2-3 h at RT with gentle mixing. Centrifugation for 30 s at 1,000 × g yielded the supernatants. The columns were washed five times with 500 µL of wash buffer (150 mM NaCl, 50 mM HEPES, 1% X-Triton, pH 7.4) supplemented with protease inhibitors and eluted with 100 µL of 0.2 M glycine (pH 2.5).

Eluants were added to 20 µL of Laemmli buffer with β-mercaptoethanol and 6 µL of Tris-HCl (1 M, pH 10), boiled for 10 min and loaded on SDS‒PAGE gels. Next, the proteins were transferred to PVDF membranes (Merck Millipore) and blocked for 1 h in 0.5% Tween^®-20 in PBS supplemented with 5% skimmed powder milk. The membranes were immunoblotted overnight with mouse antibodies against Kv1.3 (1/200, UC Davis/NIH Neuromab Facility) or against GFP (1/500, Hoffmann-La Roche). The filters were washed with PBS containing 0.5% Tween^®-20 and incubated for 2 h with horseradish peroxidase-conjugated secondary antibodies (Bio-Rad Laboratories).

Observation of colocalization by confocal microscopy

The cells were fixed for 10 min with 4% paraformaldehyde (PFA) at RT, washed twice with PBS without potassium and mounted on microscope slides with Mowiol + DABCO (Sigma Aldrich). The samples were left to dry at RT, and images were obtained on a Zeiss LSM 880 Airyscan confocal microscope from the Advanced Optical Microscopy Unit (CCiTUB). Images analyzed with ImageJ were processed using a previously described method (18) with the JACoP plugin to calculate colocalization coefficients (19).

Electrophysiology

Voltage-gated K⁺ currents were evoked in HEK293 cells using the whole-cell configuration of the patch clamp technique on a HEKA EPC10 USB amplifier using the acquisition software PATCHMASTER v2x91 (HEKA Elektronik GmbH).

Currents were recorded at RT at a frequency of 0.1 Hz and filtered and sampled at 2 kHz. Pipettes were made from borosilicate capillaries (GC120T-10; Warner Instruments) using a programmable puller (P-97; Sutter Instruments) and polished on a microforge (MF-830; Narishige). The pipette resistance was between 2 and 4 MΩ. The pipette solution contained (in mM) 80 K-aspartate, 42 KCl, 3 phosphocreatine, 10 KH₂PO₄, 3 MgATP, 5 HEPES-K, and 5 EGTA-K (adjusted to pH 7.25 with KOH). The bath solution contained (in mM) 145 NaCl, 4 KCl, 1.8 CaCl₂, 1 MgCl₂, 10 HEPES-Na, and 10 D-glucose (adjusted to pH 7.4 with NaOH). The data were analyzed using a FITMASTER v2x91 (HEKA Elektronik GmbH) and GraphPad Prism 5 (GraphPad Software).

The cells were held at -80 mV as the holding potential. To study the voltage dependency of activation, currents were elicited after a 250 ms pulse from -80 to +60 mV in 10 mV increments, followed by a 250 ms pulse at -40 mV to visualize tails. The tail current was measured, normalized, and adjusted to a Boltzmann equation of the form:

Equation 1.

where V is the independent variable (voltage), V₅₀ is the voltage at which half of the channels are open, and k is the slope of the curve.

To study the voltage dependency of inactivation, the cells were depolarized at +40 mV for 1 s, and then a 250-ms pulse was applied from -60 to +40 mV in 10-mV increments. The current of the second pulse was normalized and fitted to a Boltzmann equation.

To study use-dependent inactivation, a train of 15 depolarizing pulses of 200 ms at +60 mV was applied at a frequency of 2.5 Hz. The peak current of the obtained traces was normalized to that of the first pulse and expressed as a function of the number of pulses. To obtain the decay constant, we fitted the curve to an exponential equation of the form:

Equation 2.

where t is the independent variable (time), τ is the time constant and A is the amplitude. To further characterize the kinetics of inactivation, a long depolarizing pulse (5 s at +60 mV) was applied and fitted to an exponential decay.

To study the molecular rearrangements affected by KCNE4, we applied a very long depolarizing pulse (1 min at 0 mV) as previously described (20). The elicited traces were fitted to a biexponential equation of the form:

Equation 3.

  where τ_fast and τ_slow represent the time constants for the fast and slow components, respectively. Similarly, A_fast and A_slow represent the amplitudes of both kinetic components. To determine their contributions, the amplitude of the fast component was relative to that of A_fast+A_slow and is expressed as a percentage.

Recovery from inactivation was studied by applying a classic two-pulse protocol separated by a variable time interpulse at -90 mV. The first pulse was at +40 mV for 2 s, which allowed for full inactivation of the channels. The second test pulse was at +40 mV for 250 ms. Fractional recovery (FR) was calculated as:

Equation 4.

where I_peak is the peak current at the first pulse, I_plateau is the current at the end of the first pulse (steady state) and I_test is the peak current of the test pulse. FR as a function of interpulse duration was fitted to an exponential equation.

Free energy calculations and cycle analysis

From 5 s pulses to +60 mV, free energy calculations were also performed as previously described (21). Briefly, at long depolarizations, the number of channels in the closed state is negligible, so an equilibrium of the form 𝑂 ⇌ 𝐼 can be assumed. The free energies of inactivation, recovery, and their difference were calculated as:

Equation 5

where R is the universal gas constant, T is the temperature, τ is the time constant of inactivation and r is the ratio of the steady-state current at the end of the pulse divided by the peak current.

The free energy of inactivation (∆G_inactivation) was calculated for all KCNE2/4 chimeras. Thermodynamic cycle analysis was performed to analyze the coupling between the three different domains of KCNE4 as previously described (22). Thus, two cycles were constructed to assess the coupling (∆∆G_C) of the C-terminus and transmembrane regions in the absence (cycle I) or presence (cycle II) of the N-terminus of KCNE4. The differences between couplings (∆∆∆G_C) represented the influence of the N-terminus on C-terminus-transmembrane coupling.

Equation 6.              

Finally, the mean contribution of each domain was calculated assuming that their contributions were independent. For the contribution of the C-terminus, we averaged the differences in free energy in all the transitions that added the C-terminus of KCNE4 (KCNE2 → KCNE(2-2-4); KCNE(2-4-2) → KCNE(2-4-4); KCNE(4-2-2) → KCNE(4-2-4); KCNE(4-4-2) → KCNE4). The same was performed for the transmembrane domain (KCNE2 → KCNE(2-4-2); KCNE(2-2-4) → KCNE(2-4-4); KCNE(4-2-2) → KCNE(4-4-2); KCNE(4-2-4) → KCNE4) and the N-terminal domains (KCNE2 → KCNE(4-2-2); KCNE(2-4-2) → KCNE(4-4-2); KCNE(2-2-4) → KCNE(4-2-4); KCNE(2-4-4) → KCNE4).

Homology Kv1.3/KCNE4 docking model

We used our previously published docking model for Kv1.3/KCNE4 (12), which was validated on the cryo-EM structure of Kv1.3 (23, 24). Briefly, the largest part of Kv1.3 was modeled after Kv1.2 (PDB code 2R9R), except for the first N-terminal 49 amino acids (modeled after 1PXE; zinc-binding domain from neural zinc finger factor-1) and the C-terminus (3HGF; nucleotide-binding domain of the reticulocyte binding protein Py235). KCNE4 was modeled with a KCNE1 structure (PDB code 2K21) except for residues 99-170 in the C-terminus (HTC; a complex of recombinant hirudin and human α-thrombin). Homology modeling was performed on the SwissModel Protein Modeling on the Expert Protein Analysis System (ExPASy) Molecular Biology website (http://kr.expasy.org/). Energy minimization was performed with Yet Another Scientific Artificial Reality Application (YASARA) v23.5.19. Docking of Kv1.3 and KCNE4 was performed by first using a Kv7.1-KCNE1 structure (25) as a template and then docking regions of interaction (17, 23, 24).

UCSF Chimera (26) was used for visualization. Putatively interacting residues were identified with the Find Clashes/Contacts tool. The overlap between two atoms i and j was calculated as:

Equation 7.

where r_VDW is the van der Waals radius for each atom and d_ij is the distance between the two atoms. Only overlap values greater than -0.4 Ǻ were considered to indicate contacts.

Statistical analysis

The results are expressed as the means ± SE from at least three independent experiments. When only two conditions were analyzed, differences were assessed by Student’s t test. For multiple analysis, ANOVA was performed followed by either Tukey’s test (for colocalization analysis) or Dunnett’s test (for electrophysiology experiments). If the residuals were significantly non-Gaussian, as measured with the Kolmogorov‒Smirnov test, differences were assessed with a Kruskal‒Wallis test followed by Dunn’s post hoc test.