A Chimeric NaV1.8 Channel Expression System Based on HEK293T Cell Line

Among the nine voltage-gated sodium channel (NaV) subtypes, NaV1.8 is an attractive therapeutic target for pain. The heterologous expression of recombinant NaV1.8 currents is of particular importance for its electrophysiological and pharmacological studies. However, NaV1.8 expresses no or low-level functional currents when transiently transfected into non-neuronal cell lines. The present study aims to explore the molecular determinants limiting its functional expression and accordingly establish a functional NaV1.8 expression system. We conducted screening analysis of the NaV1.8 intracellular loops by constructing NaV chimeric channels and confirmed that the NaV1.8 C-terminus was the only limiting factor. Replacing this sequence with that of NaV1.4, NaV1.5, or NaV1.7 constructed functional channels (NaV1.8/1.4L5, NaV1.8/1.5L5, and NaV1.8/1.7L5, respectively), which expressed high-level NaV1.8-like currents in HEK293T cells. The chimeric channel NaV1.8/1.7L5 displayed much faster inactivation of its macroscopic currents than NaV1.8/1.4L5 and NaV1.8/1.5L5, and it was the most similar to wild-type NaV1.8 expressed in ND7/23 cells. Its currents were very stable during repetitive depolarizations, while its repriming kinetic was different from wild-type NaV1.8. Most importantly, NaV1.8/1.7L5 pharmacologically resembled wild-type NaV1.8 as revealed by testing their susceptibility to two NaV1.8 selective antagonists, APETx-2 and MrVIB. NaV chimeras study showed that at least the domain 2 and domain 4 of NaV1.8 were involved in binding with APETx-2. Our study provided new insights into the function of NaV1.8 intracellular loops, as well as a reliable and convenient expression system which could be useful in NaV1.8 studies.


INTRODUCTION
Mammalian NaV is composed of four homologous domains, each domain contains six transmembrane segments, in which the first four segments (S1 -4) construct the VSM and the last two (S5 -6) constitute the PM (Catterall, 2012(Catterall, , 2014b. PMs from all four domains of the channel construct the central ion conducting pathway, which is surrounded by VSMs. When the membrane is depolarized, the S4 segment, which carries the gating charges, moves outward to allosterically open the central pore. Recent landmark progresses regarding the high resolution crystal structures of bacterial sodium channels and cryo-EM structures of cockroach Na V PaS and eel NaV1.4 have provided important insights to understand the structure-function relationship of mammalian NaVs (Payandeh et al., 2011(Payandeh et al., , 2012Zhang et al., 2012;Shen et al., 2017;Yan et al., 2017). In mammals, nine NaV subtypes have been characterized so far, and their roles in pathological conditions such as epilepsy, pain, ataxia and so on, have been verified by numerous studies (Catterall, 2014a;Emery et al., 2016;Schwarz et al., 2016). Among them, NaV1.8 channel is a major contributor to action potential (AP) generation and propagation in peripheral primary sensory neurons, and it was deemed as an ideal target for pain-treating drugs (Blair and Bean, 2003). The pivotal role of NaV1.8 in the pain transduction pathway was supported by gene knocking out studies, in which knocking out the NaV1.8 gene or genes of several NaV1.8 regulatory proteins in mouse attenuated the pain behavior (Akopian et al., 1999;Foulkes et al., 2006;Zhang and Verkman, 2010). Actually, several NaV1.8 selective antagonists characterized in recent years showed analgesic effect in animal pain models (Jarvis et al., 2007;Bagal et al., 2015;Payne et al., 2015).
The functional expression of NaV1.8 channel in mammalian cell lines is an important approach to study its electrophysiological and pharmacological properties. For this goal, the cell lines ND7/23 and SH-SY5Y were commonly used for functional NaV1.8 currents expression by using the transient transfection method (John et al., 2004;Dekker et al., 2005). And another choice is Xenopus oocytes, but it needs in vitro transcription and microinjection of NaV1.8 mRNA (Rabert et al., 1998). In recent years, commercial cell lines stably expressing NaV1.8 channel were available and were amenable to the drug high-throughput screening (HTS) method, such as the recombinant HEK293T cell line co-expressing the human NaV1.8 and β1 accessory subunit (Merck Millipore, Billerica, MA, United States). In this cell line, the mean NaV1.8 current was −900 ± 17 pA. Compared with the successful heterologous expression of functional NaV1.3 -1.5 and NaV1.7 currents in non-neuronal cell lines such as CHO-K1, COS-7 and HEK293T, transient transfection of NaV1.8 channel into them yielded no or low-level functional currents, even though it was co-expressed with the β subunits (John et al., 2004;Zhang et al., 2008). We speculated that some parts of the NaV1.8 protein sequence might be responsible for this. The ND7/23 cell line supported the functional expression of NaV1.8 channel, possibly because it has a background of dorsal root ganglion neuron and contains accessory proteins interacting with these sequences.
Actually, some signals in the NaV1.8 protein sequence preventing its surface trafficking were identified in previous studies (Zhang et al., 2008;Li et al., 2010). And proteins such as β3 subunit, annexin II light chain p11, PDZD2, contactin and calnexin were shown to regulate the functional expression of NaV1.8 (Okuse et al., 2002;Rush et al., 2005;Zhang et al., 2008;Shao et al., 2009;Li et al., 2010). Nonetheless, destroying or masking the identified signals, or co-expressing several of these partner proteins with NaV1.8 channel had not achieved highlevel functional currents expression in non-neuronal cell lines. Choi et al. (2004) and Vijayaragavan et al. (2004) had locked an unidentified retention/retrieval signal in the C-terminus of NaV1.8 by conducting chimeric channel studies. In ND7/23 cells, the chimeric channel by replacing NaV1.8 C-terminus with that of NaV1.4 (NaV1.8/1.4C) showed approximately 2folds increase of the current density, while the reverse chimera NaV1.4/1.8C showed approximate 7-folds decrease of its current when compared with the parental NaV1.4 channel (Choi et al., 2004). And replacing the NaV1.8 C-terminus with that of NaV1.7 constructed chimeric channel functionally expressed in tSA201 cells, the peak current was approximately 1 nA (Vijayaragavan et al., 2004). These studies suggested that the NaV1.8 C-terminus is an important factor regulating its functional expression. However, the roles of the other four intracellular loops in NaV1.8 were not systematically analyzed yet.

Toxins and Hazard Waste Treatment
MrVIB was a kindly gift from professor Paul Alewood (Institute for Molecular Bioscience, The University of Queensland, Australia). APETx-2 was purchased from Abcam (Abcam PLC, Cambridge, United Kingdom). These neurotoxins were used in laboratory only and the hazardous wastes were collected and sent for centralized treatment by Hunan Normal University.
University School of Medicine, United States). hNaV1.5 and rNaV1.8 cDNAs were subcloned into a pCMV-Blank vector (Beyotime, Shanghai, China) between sites Hind III and XbaI, with the Kozak sequence (GCCACC) added in front of the NaV coding sequence. The resulting constructs were named as pCMV-NaV1.5 and pCMV-NaV1.8. The rNaV1.4 and hNaV1.7 channels were cloned in the pCDNA3.1 plasmid, and the boundaries of the transmembrane domains were as presented in NCBI (NaV1.4 1 ; NaV1.7 2 ). We constructed the DCNaV1.8 channel by step-by-step substituting the transmembrane domains (D1 -4) of NaV1.5 with those of NaV1.8. To take domain I (D1) substitution as an example, briefly, the pCMV-NaV1.5 plasmid was linearized by a pair of oppositely directed primers to delete the NaV1.5 D1, the NaV1.8 D1 was amplified by PCR by a pair of primer containing the upstream and downstream joint sequences. The linearized plasmid and the amplified domain segment were purified and ligated by using the recombinant cloning kit following the manufacturer's instructions (CloneEZ R PCR Cloning kit, Genscript, Nanjing, China). The primers for linearizing pCMV-NaV1.5 plasmid and for amplifying NaV1.8 transmembrane domains were as listed in our previous studies (Tang et al., 2015. We used the enzymatic digestion and ligation method to replace the loops L1, L2, L3, or L5 of NaV1.8 with that of NaV1.5, as well as the NaV1.8 L5 with that of NaV1.4 or NaV1.7. For construction of these chimeras, the sites Hind III, BglII and XbaI were engineered to flank these loops by site-directed mutations (deletion and insertion). In the chimera NaV1.5/1.8L5, the NaV1.5 C-terminus was substituted with that of NaV1.8 by using the engineered EcoR V site (between 5313 and 5314 bp in NaV1.5 mRNA numbering) and the Xbal I site. The introduced restriction sites in the coding sequence of NaVs were finally deleted by site-directed mutations. All constructs were sequenced to ensure that the correct chimeras were made.

Electrophysiology
Whole-cell currents of cells transfected with wild-type or chimeric NaV channels were recorded in an EPC-10 USB patchclamp platform (HEKA Elektronik, Ludwigshafen, Germany). The recording pipets were pulled from glass capillaries (thickness = 0.225 mm) in a PC-10 puller (NARISHIGE, Tokyo, Japan). The pipet resistance was controlled to be 1.5 -2.0 M and only the tip was filled with pipet solution to minimize the fast capacitance. The standard pipet solution contains (in mM): 140 CsCl, 10 NaCl, 1 EGTA, 2 Mg-ATP, and 20 HEPES (pH = 7.4). Bath solution contains (in mM): 140 NaCl, 2 CaCl 2 , 1 MgCl 2, 5 KCl, 20 HEPES (pH = 7.4), and 10 Glucose. 1 https://www.ncbi.nlm.nih.gov/protein/NP_037310.1 2 https://www.ncbi.nlm.nih.gov/protein/NP_002968.1 300 nM TTX was added into the bath solution to block the endogenous NaVs when recording currents in ND7/23 cells. All chemicals were products of Sigma-Aldrich (Sigma-Aldrich, St. Louis, MO, United States) and dissolved in Milli-Q water. All experiments were conducted at room temperature (20 -25 • C). Data were collected by PatchMaster software and analyzed by Igo Pro 6.10A, OriginPro 8 and GraphPad Prism 5. The pipet capacitance and the cell capacitance were canceled by sequential fast and slow capacitance compensation by using the automatic computer-controlled circuit of the amplifier. To minimize the voltage error, the serial resistance (Rs) was controlled to be less than 10 M and 80% Rs compensation was used, the speed value for Rs compensation was set to be 10 µs. The LockIn extension was used to measure the cell capacitance in case of measuring the current density. The dose-response curves were fitted by a Hill equation to estimate the potency (IC 50 ) of the toxin. The NaVs repriming curves were fitted by a double exponential rising equation: y = a(1-e −x/Tau1 ) + c(1-e −x/Tau2 ), where Tau1 and Tau2 represented the fast and the slow time constant, respectively. The G-V and SSI curves were fitted by a Boltzmann equation: y = y steady + (y (0) -y steady )/(1 + exp[(V-V 1/2 )/K]), where V 1/2 , V and K represented the midpoint voltage of kinetics, the test voltage and the slope factor, respectively.

Data Analysis
Data were presented as MEAN ± SEM, n was presented as the number of separate experimental cells. Statistical significance was assessed by ONE-WAY ANOVA, multiple comparison between the groups was performed using Turkey method, statistical significance was accepted when p < 0.05.

The Intracellular Loops Regulated NaV1.8 Functional Expression
Voltage-gated sodium channels are tolerant of large module/modules substitution between different subtypes (Shaya et al., 2011;Tang et al., 2015). Our previous study showed that replacing D2, D3, or D4 of NaV1.5 with that of NaV1.8 constructed channels (NaV1.5/1.8D2, NaV1.5/1.8D3, and NaV1.5/1.8D4, respectively) functionally expressed in HEK293T cells . This inspired us to construct a chimeric channel which contains the intracellular loops of NaV1.5 and the transmembrane domains of NaV1.8, and it might express high-level functional currents in HEK293T cells. In Figure 1A, we determined the boundaries of the transmembrane domains (D1-4) and the intracellular loops connecting them (L1 -5), with the number bellow each domain indicating the first and the last amino acid, in NaV1.5 and NaV1.8 (NaV1.5 3 ; NaV1.8 4 ). The chimeric channel was constructed as described in the molecular cloning section and was named as DCNaV1.8. It expressed large NaV currents as the parental NaV1.5 channel when transfected into HEK293T cells ( Figure 1B). And more than 70% (34/47) FIGURE 1 | Strategy for NaV chimeras construction. (A) The topological structure of NaV1.5 and NaV1.8, the numbers indicated the border of the transmembrane domains, the representative current traces showed the NaV1.5 and NaV1.8 expressions in HEK293T cells. (B) Replacing the transmembrane domains of NaV1.5 with those of NaV1.8 constructed a chimeric channel (DCNaV1.8) functionally expressed in HEK293T cells. Currents were elicited by depolarizing to +10 mV from holding potential of -100 mV.
of the DCNaV1.8 transfected cells showed peak currents of 1 -4 nA, the mean peak current density was −97.5 ± 10.0 pA/pF (n = 47). These data suggested that the NaV1.8 intracellular loops were responsible for its poor functional expression in HEK293T cells.
the C-terminus substitution of NaV1.8 did not change its pharmacology.

DISCUSSION
The ion channels and receptors represented two primary drug targets in the cell membrane, and their high-level functional expression in mammalian cell lines was of particular importance for their pharmacological and biophysical researches. Lots of methods were used to overexpress ion channels and receptors. As in the case of M 3 muscarinic acetylcholine receptor (M3R), its functional expression in mammalian cells were substantially increased by codon optimization, fusing tags with the protein, and using virus as the transfection vector (Romero-Fernandez et al., 2011). Our previous study also showed that the NaV1.9 functional expression in ND7/23 cells could be enhanced by fusing an EGFP tag to its C-terminus . Another strategy was to use pharmacological chaperones, as shown in previous reports that atropine and lidocaine increased the functional expression of M3R and NaV1.8, respectively (Ward et al., 2002;Zhao et al., 2007). The right choice of the host cell line was also important. As for NaV1.8, it expressed functional currents in ND7/23 and SH-SY5Y cells but not in non-neuronal cell lines. However, the ND7/23 cells expressed abundant endogenous NaVs (Rogers et al., 2016), TTX must be used in the bath solution to separate the recombinant NaV1.8 currents. Finally, Co-expressing the accessory proteins and even the expression temperature could affect the expression level of ion channels, as shown in the case of NaV1.9 channel expressed in the HEK293T cells (Lin et al., 2016). Future works could attempt to express NaV1.8 and NaV1.9 channels by virus transfection or codon optimization.
The in vivo expressions of ion channels in neurons were subtly tuned and the dysregulations were closely related to diseases, as those of NaVs in neuropathic pain and ASIC1a in ischemia (Lai et al., 2003;Chai et al., 2010). It was shown that some motifs in ion channels were important in determining their subcellular locations (Standley et al., 2000;Xia et al., 2001;Garrido et al., 2003). In this study, by conducting a screening analysis, we confirmed that the NaV1.8 C-terminus was the limiting factor for its poor functional expression in HEK293T cells. Interestingly, a recent study showed that substituting the NaV1.9 C-terminus with that of NaV1.4 also constructed chimeric channel functionally expressed in various host cells (Goral et al., 2015). It remains to be elucidated if the NaV1.8 C-terminus functioned by ER retention, or on the other hand, by fast internalization mechanism. The second means that the NaV1.8 proteins were fast retrieved from membrane by internalization signals, as that of NaV1.2 channel in dendrites in hippocampal neurons (Garrido et al., 2001). It is of particular interest to investigate the in vivo role of NaV1.8 C-terminus in regulating channel expression.
The NaV1.8/1.7L5 channel, constructed by replacing NaV1.8 C-terminus with that of NaV1.7, expressed high-level functional currents in HEK293T cells by transient transfection, and this channel pharmacologically resembled wild-type NaV1.8. The screening analysis of the NaV1.8 intracellular loops provided us with more solid evidence that the role of the NaV1.8 C-terminus is unique. This expression system could be used in the electrophysiological and pharmacological studies of NaV1.8 channel, especially in analyzing the gating kinetics of diseaseassociated NaV1.8 mutants and in investigating the molecular mechanism of drug-NaV1.8 interaction by mutating channel. The peak current density of NaV1.8/1.7L5 in this study was about 2-folds of that reported by Vijayaragavan et al. (2004), possibly because we used the human NaV1.7 rather than the rat NaV1.7 as the parental channel, as sequence alignment revealed approximately 9% variation between their C-termini. In addition, sequence alignment showed the C-termini of human Nav1.7 and rat NaV1.8 had an identity of 54%, while that for rat NaV1.7 and rat NaV1.8 is 51.9%. When compared with those of NaV1.8/1.4L5 and NaV1.8/1.5L5 channels (Figures 2A,E), the macroscopic currents of NaV1.8/1.7L5 channel showed apparently accelerated inactivation. This is not surprising as the NaV C-terminus was proposed to interact with the inactivation particle (the D3 -4 linker) by using the calcium sensor calmodulin as a bridge (Sarhan et al., 2012) or its conformation change might directly affect the D4 transmembrane domain, which is a key determinant of NaV fast inactivation (Wang et al., 2014). Lots of electrophysiological studies have shown that swapping the C-terminus between NaV subtypes or mutation in the C-terminus region changed channels' inactivation kinetics (Deschenes et al., 2001;Mantegazza et al., 2001;Kass, 2006;Lee and Goldin, 2008;Nguyen and Goldin, 2010). We speculated that the sequence variation between different NaV C-termini is responsible for the observed differences in the inactivation kinetics of these three chimeric channels. However, the exact mechanism is currently unknown and requires further research.
We used MrVIB and APETx-2 to investigate the pharmacology of NaV1.8/1.7L5 and compared it with that of wild-type NaV1.8. Although we did not test more toxins or chemicals, we speculated that drugs targeting the extracellular/transmembrane parts of NaV1.8 should act on the NaV1.8/1.7L5 channel in the same way, as the structure of NaV is largely stabilized by its transmembrane domains. The dose-response curve for MrVIB blocking DRG TTX-R NaVs was slightly shifted when compared with those of NaV1.8 and NaV1.8/1.7L5. It might be explained by endogenous and heterologous expressed channels showing different susceptibility to the toxin. In fact, previous study showed that the MrIVB activity on NaV1.8 could be modulated by beta subunits (Wilson et al., 2011). In addition, by using the panel of NaV1.5/NaV1.8 chimeras, we showed that APETx-2 is not a classical NaV site 4 toxin (Figure 5), and the detail of the toxin binding site in NaV1.8 is still to be investigated.

AUTHOR CONTRIBUTIONS
CT and ZL designed the experiments and wrote the manuscript. CT, YZ, DT, and XZ performed the experiments and the data analysis. SL and PC helped to perform the data analysis.

FUNDING
This work was supported by National Natural Science Foundation of China (Grant Nos. 31600669, 31570782, and 31370817). This work was also funded by Hunan Provincial Innovation Foundation for Postgraduate (Grant No. CX2015B133).