Electrophysiological properties of mouse and epitope-tagged human cardiac sodium channel Nav1.5 expressed in HEK293 cells [version 1; peer review: 1 approved, 1 approved with reservations]

Background: The pore-forming subunit of the cardiac sodium channel, Nav1.5, has been previously found to be mutated in genetically determined arrhythmias. Nav1.5 associates with many proteins that regulate its function and cellular localisation. In order to identify more in situ Nav1.5 interacting proteins, genetically-modified mice with a high-affinity epitope in the sequence of Nav1.5 can be generated. Methods: In this short study, we (1) compared the biophysical properties of the sodium current (INa) generated by the mouse Nav1.5 (mNav1.5) and human Nav1.5 (hNav1.5) constructs that were expressed in HEK293 cells, and (2) investigated the possible alterations of the biophysical properties of the human Nav1.5 construct that was modified with specific epitopes. Results: The biophysical properties of mNav1.5 were similar to the human homolog. Addition of epitopes either up-stream of the Nterminus of hNav1.5 or in the extracellular loop between the S5 and S6 transmembrane segments of domain 1, significantly decreased the amount of INa and slightly altered its biophysical properties. Adding green fluorescent protein (GFP) to the N-terminus did not modify any of the measured biophysical properties of hNav1.5. Conclusions: These findings have to be taken into account when planning to generate genetically-modified mouse models that harbour specific epitopes in the gene encoding mNav1.5.


Introduction
The voltage-gated cardiac sodium channel Na v 1.5 is responsible for the initial phase of the cardiac action potential and plays a central role in cardiac impulse propagation 1 . Its role in human disorders has been underlined by the findings of several hundred mutations in its gene, SCN5A, that are linked to inherited cardiac electrical disorders such as congenital long QT syndrome and Brugada syndrome 2 . In recent years, it has been demonstrated that Na v 1.5 interacts with and is regulated by different proteins (recently reviewed by Shy et al. 3 ). Many of these interacting proteins were also found to be mutated in patients with genetically-determined cardiac arrhythmias 4 . The generation of genetically-modified mouse models, harbouring mutations in the Scn5a gene, has proven to be a very informative approach to investigate the various human phenotypes that are linked to the genetic variants of this gene 5 . Since Na v 1.5 interacts with many proteins during its life cycle in cardiac cells, it would be of great interest to generate a mouse model that permits the biochemical purification of Na v 1.5 with high efficiency, hence allowing the co-purification of interacting proteins. The identity of these co-purified proteins may then be determined by using mass spectrometry-based technologies. In order to do this, one needs to first generate a knock-in mouse model, where a high-affinity epitope would be added to the mouse Scn5a gene that codes for Na v 1.5.
The goals of this short study were (1) to compare the biophysical properties of the sodium current (I Na ) generated by mouse Na v 1.5 and human Na v 1.5 constructs expressed in HEK293 cells, and (2) to investigate the possible alterations of the biophysical properties of human Na v 1.5 constructs that were modified with specific epitopes. We used the common fluorescent GFP and YFP proteins as epitopes, which provide the advantage of being detectable without the use of antibodies. However, these tags can only be added to the N-and C-termini, which are both intracellular in Na v 1.5, and which are thus, not easily accessible. Therefore, we additionally chose the FLAG-epitope (Sigma-Aldrich), which consists of a short sequence that can be inserted into the extracellular loops of Na v 1.5. The results of these studies will have to be taken into account when planning the generation of a mouse line bearing an epitope-tagged Na v 1.5 channel.

Cellular electrophysiology
All experiments were performed in the whole-cell voltage-clamp mode. The extracellular solution contained (in mM): 50 NaCl, 80 NMDG-Cl, 5 CsCl, 2 CaCl 2 , 1.2 MgCl 2 , 10 HEPES, 5 Glucose, adjusted to pH 7.4 with CsOH, and with an osmolality of 280-290 mOsm. The internal solution consisted of (in mM): 70 CsAsp, 60 CsCl, 1 CaCl 2 , 1 MgCl 2 , 10 HEPES; 11 Cs 2 EGTA, 5 Na 2 ATP, adjusted to pH 7.2 with CsOH, and with an osmolality of 297 mOsm. Recordings were performed at room temperature (20-22°C) using a VE-2 amplifier (Alembic Instruments, Montreal, Canada). Data was acquired by Clampex 10.2 (Axon Instruments, Union City, Canada). Membrane resistance was ≥ 1 GΩ and access resistance ≤ 6.1 MΩ. Transfected cells were recognized by the addition of 1 µl/ml Dynabeads CD8 (Invitrogen) into the extracellular solution. Current-voltage (I/V) curves were assessed by depolarising cells from a holding potential of -100 mV to voltages of between -80 and 40 mV during 20 ms. Steady-state inactivation properties were measured by the following protocol: the cells were kept at a holding potential of -100 mV and then hyper-and depolarised during 500 ms to voltages of between -120 and 0 mV in steps of 5 mV, followed by 20 ms at the voltage that elicited the maximal response during the I/V-protocol. Voltage-dependent activation was read either from the I/V-or the steady-state inactivation-protocol. To characterise the recovery from inactivation, the cells were depolarised from a holding potential of -100 mV for 100 ms, repolarised to -100 mV at a recovery time of 0.25-3000 ms, and depolarised again for 25 ms. By varying the time of the first depolarisation step from 3 to 3000 ms followed by 25 ms of repolarisation, the onset of slow inactivation was determined (see insets of Figure 2 and Figure 4).

Data analyses and statistics
Peak values for all protocols were detected and measured by Clampfit 10.2 and I/V-relationships were fitted using KaleidaGraph 3.5 (Synergy Software, Reading, USA). Values were normalised to membrane capacitance. The following formula was used to fit I/V-curves and to calculate reversal potentials: I Na = (G max (V-V rev,Na ))/(1+e V-V0.5/K ) with I Na = sodium current in pA, G max = max. conductance = 60Ω -1 , V rev,Na = reversal potential = 40 mV, K = (-zδF)/FR = equilibrium constant = -5, V 0.5 = voltage for 50% of maximum current = -20 mV. Activation and inactivation curves were fitted with the Boltzmann equation f 0 = 1/(1+ e V-V0.5/K ) with f 0 = fraction of open channels/total available channels. Statistical analyses were performed using twotailed Student's t-tests. A p value <0.05 was considered statistically significant.

Results
Electrophysiological properties of human and mouse Na v 1.5 are comparable To compare the biophysical properties of the cardiac sodium channel Na v 1.5 from the human (hNa v 1.5) or the mouse sequence (mNa v 1.5), we measured the electrophysiological properties of hNa v 1.5 and mNa v 1.5, transiently expressed in HEK293 cells. Representative I Na recordings are shown in Figure 1. The responses to all applied protocols revealed similar characteristics for both channels, except for the reversal potential and the slope of steady-state inactivation ( Figure 2 and Table 1). The peak currents from the I/V-protocol were at -15 mV for both channels (Figure 2A). Furthermore, activation and inactivation of 50% of the channels occurred for both channels at ~-28 mV and ~-71 mV, respectively. In addition, the slopes of the activation curve were comparable for both channels (6.00 mV/e-fold in human and 6.24 mV/e-fold in mouse). Significant differences could be detected in the reversal potential V rev (51.0 mV and 56.6 mV, P<0.01) and in the slope of the inactivation curve (5.95 mV/e-fold and 6.67 mV/e-fold, P<0.01) ( Figure 2B). In addition, mNa v 1.5 had a tendency to recover faster from inactivation ( Figure 2C). The fraction of channels entering into a slow inactivation state was similar for both channel types ( Figure 2D). FLAG-tag inserted at the L299/V300 site alters voltagedependent activation of hNa v 1.5 The second set of experiments addressed the effects of adding epitopes to Na v 1.5 on its biophysical properties. To do this, we as-sessed the influence of these epitopes on I Na by expressing differently tagged hNa v 1.5 in HEK293 cells and performing whole-cell voltage-clamp experiments similar to those described above. YFPand GFP-tags were added to the N-terminus; the FLAG-tag was inserted into the extracellular loop linking S5 to S6 of domain I, between residues Leu-299 and Val-300. Representative I Na recordings for all transfected constructs are shown in Figure 3 and the data is summarised in Table 2. With the exception of the GFP-tagged construct, tagging of hNa v 1.5 led to a significant decrease in peak current I max ( Figure 4A) compared to the control WT hNa v 1.5 (FLAG: 57 pA/pF with P<0.01, YFP: 120 pA/pF with P<0.05, WT hNa v 1.5: 240 pA/pF). Adding GFP did not affect any of the biophysical prop-  erties of the human sodium channel, while a shallower activation slope (6.87 vs. 5.91 mV/e-fold, P<0.05, Figure 4B and Table 2) was observed for the YFP-tagged channel. The most pronounced effects were observed for the FLAG-tagged hNa v 1.5. The activation slope was significantly shallower (6.96 vs. 5.91 mV/e-fold, Figure 4B and Table 2), indicating that the activation of this channel is less sensitive to voltage changes. In addition, the V 1/2 of activation was shifted towards more positive voltages by about 5 mV, with -23.9 mV in FLAG-hNa v 1.5, compared to -28.9 mV in untagged hNa v 1.5. Finally, the reversal potential was decreased in the FLAG-hNa v 1.5 (FLAG 39.3 mV and untagged 51.8 mV, Figure 4B). Recovery from inactivation ( Figure 4C) and onset of slow inactivation ( Figure 4D) were comparable for all channels.
Raw data from sodium current recordings obtained from tagged human sodium channels (Na v 1.5) expressed in HEK293 cells

Discussion
The present study demonstrates (1) that the biophysical properties of mouse Na v 1.5 are essentially similar to the human homolog when expressed in HEK293 cells, and (2) that adding epitopes either upstream of the N-terminus of human Na v 1.5 or in one of the extracellular loops reduces the amount of I Na and alters some of its biophysical properties. Interestingly, GFP in the N-terminus was the only epitope that did not modify any of the measured biophysical properties of hNa v 1.5. The most pronounced effects could be observed by the insertion of the FLAG-tag in an extracellular loop. In this construct, not only was the amount of I Na drastically decreased, but also the activation properties of the sodium channel were altered. The smaller changes found in the properties of YFP-hNa v 1.5 might be partially linked to the different vector used for this epitope, especially since no alterations could be observed for GFP-hNa v 1.5.
These findings have to be taken into account when planning to generate genetically-modified mouse models that harbour specific epitopes in the mouse Na v 1.5 gene. Different combinations   Tagging with N-terminal YFP and FLAG (L299/V300) significantly decreases peak currents. Calculated reversal potentials are marked with square data points. (B) Voltage-dependence of activation and steady-state inactivation. The data was fitted with the Boltzmann formula. The activation slope of FLAG-and YFP-tagged channels is shallower compared to the untagged hNa v 1.5. V 1/2 is shifted by 5 mV for FLAG-hNa v 1.5. (C) Recovery from inactivation. The duration between the depolarising steps was varied from 0.25 to 3000 ms. No differences between the different channels could be detected. (D) Onset of slow inactivation. The duration of the first step was varied from 0.25 to 3000 ms. The relative number of channels entering slow inactivation is similar for all four channel types. (A-B): n(untagged) = 7, n(FLAG) = 11, n(YFP) = 11, n(GFP) = 8. (C): n(untagged) = 12, n(FLAG) = 5, n(YFP) = 11, n(GFP) = 8. (D): n(untagged) = 10, n(FLAG) = 8, n(YFP) = 10, n(GFP) = 8. **P<0.01 obtained by two-tailed Student's t-tests; error bars indicate standard errors. D of epitopes and insertion sites might reveal better candidates for in-vivo approaches. Furthermore, additional studies should be performed in HEK293 cells co-expressing other subunits and regulating proteins, and in native cardiomyocytes in order to assess the effects of added epitopes on the interactions with these proteins. expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.