Post-Transcriptional Gene Silencing of KChIP2 and Na_vβ1 in Neonatal Rat Cardiac Myocytes Reveals a Functional Association between Na and Ito currents

siRNAs

siRNA corresponding to rat KChIP2 was designed as recommended [8], with 5’ phosphate, 3’ hydroxyl, and two base overhangs on each strand; it was chemically synthesized and annealed by Qiagen (Valencia, CA). The following gene-specific sequences to all known rat KChIP2 splice variants (5’-AATGCCTTTGACACCAACCAC-3’) and the rat Navβ1 subunit (5’-AATTACGAGCACAACACCAGC-3’) were used successfully. A Blast search of these sequences against the rat genome was performed and did not reveal matches with any other gene. In order to identify transfected cells, the siRNAs were labelled on the 3’-end of the sense strand with Rhodamine.

Two different negative controls were used for both KChIP2 and Navβ1 silencing. A commercially available non-silencing control siRNA labelled with fluorescein (Qiagen) was used. In addition, we used single base pair mismatch siRNAs for both KChIP2 and the Navβ1 subunit: KChIP2 (5’-AATGCCTATGACACCAACCAC-3’) and rat Navβ1 subunit (5’-AATTACGGGCACAACACCAGC-3’). The mismatched base is underlined. None of these 2 sequences matched any genes in the rat genome.

Neonatal Rat ventricular Myocytes Isolation and Transfections

Neonatal rat ventricular myocytes (NRVMs) were isolated from 1- or 2-day old Sprague-Dawley rats and cultured as previously described [9]. Briefly, hearts were obtained following decapitation. The hearts were immersed in CBFHH (Calcium, Bicarbonate-Free Hank’s with HEPES) and minced with a straight razor. Cells were then digested with a Trypsin 250 solution (0.2g of trypsin in CBFHH) at 37°C and then put in FBS. Cells were pelleted by centrifugation at 1000 RPM for 5 minutes and were resuspended in warm DMEM containing 5% FBS, 1% Penicillin-Streptomycin and Vitamin B₁₂ (1:1000). Cells were then preplated for 30 minutes at 37°C to allow fibroblasts to adhere to the plate. Cells were pelleted again and resuspended in DMEM containing 5% FBS, 1% Penicillin-Streptomycin, Vitamin B₁₂ (1:1000) and bromodeoxyuridine (1:100, to inhibit fibroblast growth). NRVMs were then plated at a concentration of about 1 million cells/35mm plate.

Transfections of siRNA for endogenous gene targeting were carried out using TransMessenger transfection reagent according to the manufacturer’s instructions (Qiagen) 24 to 72 hours after initial plating. Briefly, 2 µg of siRNA was mixed with the transfection reagent and incubated at 37°C with the cells in serum free DMEM for a minimum of 3 hours. Complete DMEM (with FBS and Penicillin-Streptomycin) was then added to the cells without removing the siRNA mixture. The following siRNAs were transfected into NRVMs: KChIP2 siRNA, Na_vβ1 siRNA, control non-silencing siRNA, single base pair mismatch siRNA for KChIP2 (control) and single base pair mismatch siRNA for Navβ1 subunit (control). Cells were used 48 to 72 hours after transfections.

HEK 293 Cells Transfection

Transient transfections into human embryonic kidney (HEK293) cells were performed using the Polyfect transfection kit (Qiagen) according to the manufacturer’s protocol for 24 hours for electrophysiological recordings and 48 hours for co-immunoprecipitations. 1µg of each cDNA was used.

Real-time PCR, Western Blotting and Co-Immunoprecipitation

Fluorescence-based kinetic real-time PCR was performed using a Perkin-Elmer Applied Biosystems Model 7900 sequence detection system as previously described; in brief, the relative mRNA abundance expressed as arbitrary units was calculated using the expression levels of all transcripts normalized to 18S rRNA. This value was then normalized to mRNA levels measured from NRVMs transfected with either the non-silencing control siRNA or the control single base pair mismatch siRNA for the appropriate transcript [2]. For Western Blotting and co-immunoprecipitation, cell lysates were prepared as previously described [10]. Briefly, cells were washed with cold PBS and then harvested in lysis buffer. Cells were briefly sonicated and then lysed for an hour at 4°C. Cells were then centrifuged for 4 minutes at 4000 rpm at 4°C. Protein concentration of the supernatant was measured. The lysis buffer contained for 10 ml: 0.05M HEPES, 0.15 M KCl, 0.5 mM CaCl2, 1% IGEPAL, 1 tablet of the proteases inhibitor cocktail Complete Mini (Roche) and 100 µl of 40 mM AEBSF. All samples were run in duplicate on Tris-HCl precast gels (Bio-Rad, Hercules, CA) in 25mmol/L Tris, 192mmol/L glycine and 0.1% (w/v) SDS running buffer as previously described [11]. Primary antibody incubations were performed overnight at 4°C using antibodies to the following: KChIP2 (gift of Dr. James Trimmer), Kv4.2 (AB560, Chemicon), Kv4.3 (SC-11686, Santa Cruz), Kv1.4 (AB5180, Chemicon), Na_v1.5 (Upstate), Na_vβ1 (antibody raised to the following epitope: KRRSETTAETFTEWTFR), Cav1.2 (ACC-003, Alomone) and Kir2.1 (AB5374, Chemicon). Bands on the Western blots were considered specific if they were not present when the gel was stained with pre-immune serum and if the peptide epitope reduced the band intensity. Further, antibody specificity was assessed by isolation of proteins from cells transfected with cDNAs encoding the proteins of interest. Only with satisfaction of these criteria were antibodies then used in immunopreciptation experiments. Relative band densities were quantified using ImageQuant software (Molecular Dynamics). To normalize for protein loading, we used GAPDH (from RDI-TRK #5G4-6C5) for Kv4.2, Kv4.3, Kv1.4 and Na_v1.5 and calsequestrin (from RDI #CALSEQabR) for KChIP2 and Na_vβ1 because of their respective sizes.

Co-immunoprecipitation experiments were performed using lysates from NRVMs transfected with a control non-silencing siRNA for interaction between Na_vβ1 and Kv4.2 or Kv4.3, or from transfected HEK 293 cells for interaction experiments between KChIP2 and Nav1.5 or Kv4.3. The cell lysates were pre-absorbed with Protein A Dynabeads (Dynal, Norway) for 2 hours at 4°C. The unbound extracts were then incubated overnight at 4°C with Protein A Dynabeads crosslinked to either Kv4.2, Kv4.3 or KChIP2 polyclonal antibodies. The beads were then washed four times with lysis buffer. The bound complexes were eluted from the beads by adding 2X loading buffer to the beads and heating the samples at 70°C for 10 minutes. For Co-immunprecpitation between KChIP2 and Na_v1.5, the elution was done at 37°C for 20 minutes. Western blots were then performed and probed with the appropriate antibodies.

Electrophysiology

Macroscopic currents from transfected NRVMs were recorded using the whole-cell configuration of the patch-clamp technique [12]. Currents were recorded using an Axopatch 200B patch-clamp amplifier (Axon Instruments, Foster City, CA) interfaced to a personal computer. Voltage commands were issued and data collected with custom-written software. Low resistance electrodes (~2MΩ when filled with internal solution) were pulled from borosilicate, and a routine series resistance compensation of the Axopatch 200B amplifier was performed to values >80% to minimize voltage-clamp errors. To ensure the quality of the voltage clamp, we measured the time constant of membrane charge capacitance. The time constant of membrane charge capacitance was measured by fitting the decay of the uncompensated capacity transient. The cell capacitance (C_m) was calculated by integrating the area under the uncompensated capacity transient elicited by a 20 mV hyperpolarizing test pulse from a holding potential of −80 mV and averaged18 pF. The uncompensated R_series was therefore less than 2MΩ. The internal solution contained (in mM): potassium glutamate 130, KCL 9, NaCl 10, MgCl2 0.5 and MgATP 5 and HEPES 10 (pH 7.2 with KOH). Myocytes were superfused with a physiological salt solution containing (in mM): 138 NaCl, 4 KCl, 2 CaCl2, 1 MgCl2, 0.33 NaHPO4, 10 HEPES, 10 Glucose (pH 7.4 with NaOH). The pipette-to-bath liquid junction potential was −17 mV and was corrected off-line. Membrane currents were filtered at 5 kHz and digitized with 12-bit resolution. Experiments were performed at room temperature (22–23°C). Whole-cell currents were elicited by a family of depolarizing voltage steps from a holding potential of −80 mV. A 40 ms prepulse to −40 mV was used to inactivate I_Na. In order to record the current-voltage (I/V) relationships for I_to, Ca²⁺ currents were blocked with nifedipine after action potential and all other current recordings were made.

For sodium channel family recordings both in NRVMs and HEK 293 cells the following solutions were used: the internal solution contained (in mmol/L) NaCl 10, CsF 105, EGTA 10, and Cs-HEPES 10 adjusted to pH 7.4. The bath solution contained (in mmol/L) NaCl 140, KCl 5, MgCl₂ 1, CaCl₂ 2, HEPES 10, and glucose 10 adjusted to pH 7.4.For sodium currents recordings in control NRVM, a low external sodium concentration was used (25mM). Whole cell sodium current densities were measured by dividing the peak current obtained from stepping in 10 mV intervals from −80 to 50 mV from the holding potential of −120 mV (I/V protocol) by the cell capacitance. Time course of recovery from inactivation (τ_rec) was studied using a 2-pulse protocol with a 30 ms prepulse to −30 mV followed by varying rest intervals at −120 mV and a 30 ms test pulse to −30 mV. The recovery kinetics were determined by fitting the normalized peak current amplitudes to the following equation:

The voltage dependence of steady state inactivation was determined by 500 ms prepulses ranging from −140 to −30 mV. The voltage dependence of inactivation was determined by fitting the normalized peak currents to a Boltzmann distribution:

Statistical analysis

Data are expressed as mean ± SEM. When indicated, a Student’s t-test was performed using statistical software in Origin (Microcal ™ Software, Inc). Differences were considered to be significant at a P value <0.05.

Co-Immunoprecipitation

In heterologous expression studies, co-transfection of Na_vβ1 subunits with Kv4.3 produces a current that appeared to recapitulate cardiac I_to with the highest fidelity compared to other known K⁺ channel accessory subunits [3]. Thus we sought to determine if the subunits underlying I_to and I_Na interact in cardiac myocytes. Antibodies to either anti-Kv4.2 or Kv4.3 precipitated Na_vβ1 subunits from NRVM protein extracts (n=4) (Fig. 1A).

In order to further evaluate the interactions of Na and K channel subunits, channel proteins were expressed in a cell culture system. First, in order to validate our technique, we performed a co-immunoprecipitation between KChIP2 and Kv4.3, which are known to interact. As expected, in HEK 293 cells expressing both KChIP2 and Kv4.3, the KChIP2 antibody was able to pull-down the Kv4.3 protein (Fig. 1B). Co-immunoprecipitation experiments were then performed using an anti-KChIP2 antibody, which did not reveal an association between Na_v1.5 and KChIP2 (n=4) (Fig. 1B, lane 1). Figure 1B (lane 1) shows the IP lane of HEK 293 cells transfected with both Na_v1.5 and KChIP2 which did not exhibit a band for Na_v1.5. Na_v1.5 was detected in the supernatant, consistent with the absence of direct association between Na_v1.5 and KChIP2.

KChIP2 modulation of Na_v1.5 biophysical properties

We have previously shown that Na_vβ1 can modulate the biophysical properties of Kv4.3 [3] and this is consistent with the interaction of these channel subunits as assessed by co-immunoprecipitation (Fig. 1A). However, it appears that KChIP2, an accessory subunit of Kv4.3, does not reciprocally interact with Na_v1.5 (Fig. 1B). To further assess the potential interaction between KChIP2 and Na_v1.5, we co-expressed these channel subunits in HEK 293 cells and studied the biophysical properties of I_Na using the patch-clamp technique (Fig. 2). Co-expression of KChIP2 did not affect the gating properties of the Na current (Fig. 2). However, in this heterologous expression system, KChIP2 did increase the Na_v1.5 current density and this increase was greater when co-expressing both KChIP2 and Na_vβ1 compared with KChIP2 alone (At −30 mV, Na_v1.5: 495 pA/pF ± 67 n=8, Na_v1.5 + KChIP2: 648 pA/pF ± 52* n=14, Na_v1.5 + Na_vβ1: 682 pA/pF ± 71* n=12, Na_v1.5 + Na_vβ1 + KChIP2: 795 pA/pF ± 55* n=10) (* p<0.05 for voltages −40 to +10 mV compared to Na_v1.5) (Fig. 2A). Still, KChIP2 did not modulate steady-state inactivation (Fig. 2B) and recovery from inactivation (Fig. 2C). The V_1/2 values for steady-state inactivation are: Na_v1.5: −88.6 mV ± 2.5 n=8, Na_v1.5 + KChIP2: −93.4 mV ± 2.2 n=14, Na_v1.5 + Na_vβ1: −90.2 mV ± 1.8 n=12, Na_v1.5 + Na_vβ1 + KChIP2: −91.4 mV ± 2.5 n=10 (Fig. 2B). The time constant of recovery from inactivation are Na_v1.5: 8.2 ms ± 1.7 n=8, Na_v1.5 + KChIP2: 10.1 ± 1.1 n=14, Na_v1.5 + Na_vβ1: 8.9 ms ± 1.4 n=12, Na_v1.5 + Na_vβ1 + KChIP2: 10.9 ms ± 1.9 n=10 (Fig. 2C)

Gene silencing of KChIP2

In order to further explore the association between Na channel and I_to channel subunits and to examine the functional consequences of this interaction, we used RNA interference (RNAi) to silence the expression of KChIP2 and examine the electrophysiological consequences in NRVMs. Small interfering RNAs (siRNAs) are potent RNAi reagents for sequence-specific, post-transcriptional gene silencing [13] [14] [15] [16]. The use of siRNA has been remarkably robust in selective gene suppression in Caenorhabditis elegans, Drosophila melanogaster, Trypanosoma brucei and plants[17] [14] [15]. Direct introduction of siRNA into mammalian cell lines, [18] [19] [20] cultured mammalian neurons [21] and native cardiac myocytes [22] has been effective in suppressing endogenous and heterologously expressed genes.

An siRNA was designed to a common region of rat KChIP2 to silence the expression of all splice variants in NRVMs. Ventricular myocytes were successfully transfected (>75% efficiency) with either KChIP2 specific siRNA labelled with rhodamine or a commercially available non-silencing control siRNA labelled with fluorescein (Qiagen), or as another non-silencing control, a single base pair mismatch siRNA against KChIP2. KChIP2 siRNA nearly completely eliminated the expression of KChIP2 mRNA as measured using kinetic quantitative real-time PCR with primers designed to recognize all KChIP2 splice variants (Fig. 3A). RNA isolated from non-transfected myocytes, myocytes transfected with the non-silencing control siRNA and myocytes transfected with the control single base pair mismatch siRNA for KChIP2 were not statistically different and exhibited approximately 20-fold greater levels of KChIP2 mRNA.

KChIP2 immunoreactive protein was effectively suppressed by siRNA. Proteins isolated from ventricular myocytes transfected at the same time as those used for mRNA measurements, revealed robust expression of KChIP2 protein in non-transfected controls and cells transfected with non-silencing siRNA (Fig. 3B). The KChIP2-specific siRNA essentially eliminated KChIP2 protein expression (Figs. 3B and 3C). Thus the expression of KChIP2 can effectively be knocked down in NRVMs in primary culture using siRNA.

We next examined the effect of KChIP2 gene silencing on the expression of other potassium channel subunits expressed in the ventricle. We found no significant difference in the steady-state levels of Kv4.2 or Kv4.3 mRNAs, the α subunits that underlie cardiac I_to in the rat[1], in the KChIP2 specific siRNA transfected cells compared with the controls (Fig. 3A). Kv1.4 is the channel subunit that underlies ventricular I_to early in development. Kir2.1 is the α subunit that encodes for the inward rectifier K current (I_K1). KVLQT1 and erg are the α subunits that encode for IKs and IKr respectively. KChIP2 siRNA did not significantly alter the steady-state mRNA levels of Kv1.4, Kir2.1, KVLQT1 or erg when compared to non-silencing siRNA-transfected controls (Fig. 3A). To further assess the possible link between I_to and I_Na, we then investigated the effect of KChIP2 silencing on the sodium channel subunits. Surprisingly, the steady-state level of Na_vβ1 mRNA was as drastically reduced as KChIP2 mRNA in presence of KChIP2 siRNA compared to controls (Fig 3A). The level of Na_v1.5 mRNA was also significantly reduced in the presence of KChIP2 siRNA while the levels of Ca_v1.2 (L-type Ca²⁺ channel α1 subunit), Kir2.1, erg and KVLQT1 were unchanged (Fig. 3A).

In contrast to the mRNA levels, KChIP2 siRNA dramatically suppressed the levels of Kv4.2 and Kv4.3 immunoreactive proteins. Figure 3B shows Western blots employing polyclonal antibodies specific to Kv4.2, Kv4.3 and Kv1.4. The levels of protein expression normalized to GAPDH revealed at least a 10-fold reduction in the levels of Kv4.2 and Kv4.3 immunoreactive proteins in NRVMs transfected with KChIP2 siRNA when normalized to myocytes transfected with a non-silencing siRNA (Fig. 3C). There was also a significant reduction in the level of Kv1.4 protein (Fig. 3B) in cells transfected with KChIP2 siRNA when normalized to myocytes transfected with a non-silencing siRNA but this reduction was not as dramatic as that of Kv4.x proteins (Fig. 3C). Thus, silencing KChIP2 expression does not suppress Kv4.x subunit transcription but reduces the expression of the immunoreactive proteins. The possible mechanisms include suppression of translation or a failure of trafficking Kv subunits to the cell membrane with accelerated protein degradation. In contrast, both the mRNA and protein levels of Na_v1.5 and Na_vβ1 were significantly reduced following KChIP2 silencing (Fig. 3B, C), suggesting a functional link between KChIP2 and I_Na. This was not a generalized effect on ion channel subunit expression as Cav1.2 and Kir2.1 protein levels were unchanged (Fig. 3B, C).

Electrophysiological effects of KChIP2 silencing

In order to study the functional consequences of KChIP2 gene suppression, we performed patch clamp recordings on the isolated NRVMs. For currents recordings (Fig. 4A–C, middle panel), cells were held at −80 mV, stepped to −40 mV (to inactivate I_Na) followed by voltage steps from −70 to +70 mV in 10 mV increments for 500 msec. Only current records elicited by steps to −70, 0, 30, 50 and 70 mV are shown for clarity. I_Ca,L is observed with steps to 0 mV and I_to at 30, 50 and 70 mV. Not surprisingly, whole-cell current recordings revealed a complete absence of I_to in KChIP2 siRNA transfected cells, (Fig. 4C middle panel and 4E) consistent with the reduction seen in mRNA and protein levels (Fig. 3). In contrast, non-transfected NRVMs, and NRVMs transfected with control non-silencing siRNA consistently expressed I_to (density range: 8 – 11 pA/pF) (Fig 4A, B middle panel and Table 1). The absence of I_to would have predicted prolongation of the action potential duration; however, quite surprisingly we could not elicit action potentials in any myocytes transfected with KChIP2 siRNA (Fig. 4C left panel) despite the presence of a normal resting membrane potential (KChIP2 siRNA −79.5 ± 12.2 vs controls −84.5 ± 7.8, p=N.S.) (Table 1). In order to confirm that we were injecting sufficient current in these cells to generate an action potential, we determined the threshold of excitability of cells in all groups. Despite injecting current (100pA for 4ms) several times threshold for eliciting an action potential in control cells, action potentials were never elicited in the myocytes transfected with the KChIP2 siRNA. The voltage trace seen in Fig. 4C left panel, most likely corresponds to the discharge of the membrane capacity after the large current injection.

Remarkably, post-transcriptional gene silencing of KChIP2 led to complete suppression of voltage-dependent sodium current (I_Na) (Fig. 4C middle and right panel) thus explaining the absence of action potentials in these NRVMs. To characterize the I_Na in the different cell groups, families of sodium currents were recorded using the protocols shown in the inset of the right panels (Fig. 4A–C). Low external sodium (25 mM) was used for the non-transfected and control siRNA cells in order to maintain a controlled sodium current clamp. However, in presence of KChIP2 siRNA, even with physiological extracellular sodium concentration, little to no sodium current was elicited (Fig. 4C right panel, Fig. 4D). To ensure that the absence of I_Na was not the result of a shift in the voltage dependence of gating, we held the cells at a more negative potential (−120 mV) and delivered test pulses over a wider voltage range and were still unable to elicit significant sodium currents (Fig.4D). The absence of I_to and I_Na was not the result of an artefact or a non-specific toxic effect in the KChIP2 siRNA transfected cells since other ionic currents such as the L-type Ca²⁺ current and I_K1 were still present and their current densities were not different than control cells (Fig. 4C middle panel and Table 1). Further, cells transfected with control siRNA exhibited the same complement of ionic currents as the non-transfected controls and robust action potentials (Fig 4A–B and Table 1).

Gene silencing of Na_vβ1

To further examine the link between Na_vβ1 and I_to channel subunits, we used an siRNA specific for Na_vβ1. Transfection of NRVMs with Na_vβ1 siRNA significantly reduced both the mRNA and protein level of Na_vβ1 (Fig. 5A). In addition to Na_vβ1, the mRNA levels of Na_v1.5 and surprisingly KChIP2 were also significantly reduced (Fig. 5A) in presence of Na_vβ1 siRNA compared to non-transfected and control transfected cells. Silencing Na_vβ1 did not affect the steady-state levels of Kv4.x mRNA but significantly reduced the expression of the immunoreactive proteins (Fig. 5A and B). However it is unclear whether the reduction in Kv4 proteins is caused directly by Na_vβ1 silencing or is the result of KChIP2 mRNA reduction following Na_vβ1 silencing. Patch clamp recording of isolated ventricular myocytes transfected with Na_vβ1 siRNA revealed a marked reduction in I_Na and I_to in Na_vβ1 siRNA transfected cells (Fig. 5C). Current densities of I_to at +70 mV were 10.7 ± 2.6 pA/pF in control cells compared to 2.5 ± 1.2 pA/pF in Na_vβ1 siRNA transfected cells; however, the biophysical properties of I_to were not different compared to control cells (Fig. 4E, 5C and Table 1). Measurement of I_Na in control cells were made in presence of low extracellular sodium (25mM), whereas physiological extracellular sodium concentration was used in cells transfected with the Na_vβ1 siRNA. There was a dramatic reduction of I_Na in presence of the Navβ1 siRNA compared to control cells or NRVMs transfected with control non-silencing siRNA (Fig. 5C–D and Fig. 4D). Action potentials could not be elicited in the Na_vβ1 siRNA transfected cells, likely the result of the very small I_Na.