The voltage-gated potassium channel Kv3.3 is the causative gene of SCA13 (spinocerebellar ataxia type 13), an autosomal dominant neurological disorder. The four dominant mutations identified to date cause Kv3.3 channels to be non-functional or have altered gating properties in Xenopus oocytes. In the present paper, we report that SCA13 mutations affect functional as well as protein expression of Kv3.3 channels in a mammalian cell line. The reduced protein level of SCA13 mutants is caused by a shorter protein half-life, and blocking the ubiquitin–proteasome pathway increases the total protein of SCA13 mutants more than wild-type. SCA13 mutated amino acids are highly conserved, and the side chains of these residues play a critical role in the stable expression of Kv3.3 proteins. In addition, we show that mutant Kv3.3 protein levels could be partially rescued by treatment with the chemical chaperone TMAO (trimethylamine N-oxide) and to a lesser extent with co-expression of Kv3.1b. Thus our results suggest that amino acid side chains of SCA13 positions affect the protein half-life and/or function of Kv3.3, and the adverse effect on protein expression cannot be fully rescued.
- potassium channel
- protein expression
- protein stability
- side chain
- spinocerebellar ataxia type 13 (SCA13)
Voltage-gated potassium (Kv) channels are a large family of ion channels that play a major role in setting the resting membrane potential and the threshold and firing pattern of action potentials . The Kv3 subfamily, consisting of four genes (encoding Kv3.1, Kv3.2, Kv3.3 and Kv3.4), have some distinct functional properties. They activate at high thresholds (−10 mV) and activate and deactivate more rapidly than other Kv channels; these properties enable neurons to fire at high frequencies . The predicted membrane topology of a Kv3 subunit shows six transmembrane domains (S1–S6), a voltage sensor in S4, and a pore loop between S5 and S6 (Figure 1A). Kv3.3 is prominently expressed throughout the CNS (central nervous system), particularly in Purkinje cells [3–5]. Mutations in Kv3.3 cause the neurological disorder SCA13 (spinocerebellar ataxia type 13).
SCA13 belongs to a large group of autosomal dominant neurological disorders, SCAs, caused by degeneration of the cerebellum and the spinal cord . SCAs have a wide range of phenotypes, including cerebellar ataxia, dysarthria, extrapyramidal symptoms, oculomotor disturbance, cognitive impairment, epilepsy and so on . SCAs are highly heterogeneous, and 35 SCA loci have been mapped to a chromosome region so far; in 22 of these disorders the causative genes have been identified, including the recently identified causative gene of SCA36 . Many of them are caused by microsatellite repeat expansions, whereas others are caused by non-repeat mutations. Moreover, little is known about the normal functions of most SCA-affected proteins. SCA13 is caused by point mutations in the coding region of the Kv3.3 gene on chromosome 19q13, and four mutations have been identified in different populations [8–10]. The R366H mutation is located in the S2 domain; R420H and R423H are in voltage sensor S4, and F448L is located at the cytoplasmic end of S5 (Figure 1A). R366H was found in one individual with adult-onset ataxia, but it did not co-segregate with the disease phenotype in that family and did not meet the stringent criteria of causative mutations . R420H was associated with adult-onset progressive ataxia [9–11]. Patients with R423H or F448L mutation had early onset, but F448L also showed mental retardation and seizures [8,11].
Functional analyses of these SCA13 mutations were first conducted in Xenopus oocytes. R366H, R420H and R423H were non-functional when expressed alone, and they suppressed the current amplitude when co-expressed with wild-type Kv3.3; the F448L mutation did not change current amplitude, but altered gating properties [10–12]. However, the biochemical properties of these mutants and the underlying cellular mechanisms are largely unknown.
The aims of the present study were to investigate the functional and protein expression of SCA13 mutants in mammalian CHO (Chinese-hamster ovary) cells to determine whether the mutations affect protein stability, trafficking and cell-surface expression, and the effects of a chemical chaperone and heterotetramerization on protein expression. The effect of conservative amino acid replacements in SCA13 was also examined. This work deepens our understanding of mechanisms underlying the SCA13 disease.
Cell culture and plasmids
CHO pro-5 cells were purchased from the A.T.C.C. (Manassas, VA, U.S.A.). Cells were grown in minimum essential medium α (Cellgro) supplemented with 2 mM glutamine, 100 IU/ml penicillin, 100 μg/ml streptomycin and 10% (v/v) FBS at 37°C under 5% CO2. Human wild-type Kv3.3 and Kv3.1b cDNAs were purchased from Origene and were engineered into the eukaryotic expression vector pcDNA3 without 5′ or 3′ untranslated regions (Invitrogen). SCA13 mutations were introduced into Kv3.3 constructs by site-directed replication mutagenesis following the manufacturer's instructions (Stratagene). Complementary pairs of mutagenic primers were used to introduce mutations in Kv3.3 using a thermocycler. The integrity of constructs and the accuracy of mutations were confirmed by DNA sequencing (Cornell University).
CHO cells grown on 35-mm-diameter dishes were transiently transfected with 0.5 μg of plasmid using Lipofectamine™ Plus as recommended by the manufacturer (Invitrogen). In co-transfection experiments, 2 ng of Kv3.1b plasmid was co-transfected with 0.5 μg of either wild-type Kv3.3 or SCA13 plasmids at a 4:1 ratio so that most Kv3.3 subunits were in tetrameric complexes with Kv3.1b subunits.
At 20 h after transfection, cells were washed three times with ice-cold PBS containing 0.7 mM CaCl2 and 0.4 mM MgCl2, and incubated with 10 mM sodium periodate at 4°C in the dark for 20 min to oxidize cell-surface carbohydrates. The cells were washed three times with ice-cold PBS containing 0.7 mM CaCl2 and 0.4 mM MgCl2, and washed once with 100 mM sodium acetate (pH 5.5) containing 0.7 mM CaCl2 and 0.4 mM MgCl2. After that, cells were incubated with 2 mM biotin-LC-hydrazide (Thermo Fisher Scientific) in 100 mM sodium acetate (pH 5.5) containing 0.7 mM CaCl2 and 0.4 mM MgCl2 at 4°C for 15 min. This was followed by a wash with ice-cold PBS containing 10 mM Tris for three times to remove excessive biotins. The cells were lysed in 1 ml of cell lysis buffer (150 mM NaCl, 50 mM Tris/HCl, pH 7.4, 1% Nonidet P40, 0.5% deoxycholate, 0.01 mM leupeptin, 0.01 mM pepstatin A and 0.5 mM PMSF), and were collected into 1.5 ml tubes and rocked at 4°C for 40 min. An aliquot of cell lysates was taken out for total protein level analysis; the remaining lysates were centrifuged, and the supernatant was transferred to a new tube and incubated with 50 μl of streptavidin–agarose resins (Thermo Fisher Scientific) at 4°C. The biotin–streptavidin–agarose complexes were harvested by centrifugation, and were washed and resuspended in 15 μl of Laemmli sample buffer. This was used for surface protein level analysis by electrophoresis and immunoblotting.
SDS/PAGE and immunoblotting
Cell lysate samples were heated to 70°C for 5 min before separating by SDS/PAGE (9% gels). Proteins were then electrotransferred on to PVDF membranes (GE Healthcare) and incubated overnight with primary antibodies (rabbit anti-Kv3.3 from Sigma; mouse anti-Kv3.1b and anti-Kv1.1 from the UC Davis/NIH NeuroMab Facility, Department of Neurobiology, Physiology and Behaviour, University of California, Davis, CA, U.S.A.). The PVDF membranes were washed and incubated with horseradish peroxidase-linked anti-rabbit or anti-mouse secondary antibodies (1:10000 dilution, GE Healthcare), and target proteins were detected using Western lightning-ECL reagents (PerkinElmer) and Kodak Biomax light films. File images were scanned with a Microtek 8700 scanmaker and analysed with the Scion Image software. The PVDF membranes containing total proteins were stripped and re-blotted with anti-actin antibodies (Sigma) to control for protein loading amount; transfection efficiencies were routinely checked by co-transfection of 0.1 μg of GFP cDNA and immunoblotting with anti-GFP antibodies (results not shown).
CHO Pro-5 cells were co-transfected with 0.5 μg of wild-type Kv3.3 or SCA13 cDNAs and 0.3 μg of GFP cDNAs to visualize transfected cells. Whole-cell voltage clamp was performed at room temperature (23°C) using an Axopatch 200B amplifier (Axon Instruments) and patch pipettes were fabricated from Corning 8161 glass (Warner Instruments) and fire-polished (Micro Forge MF-90) to give a tip resistance of 1.2–2.0 MΩ when filled with an intracellular solution (1 mM MgCl2, 5 mM NaCl, 70 mM KCl, 65 mM KF, 10 mM EGTA, 5 mM glucose and 10 mM Hepes, pH 7.3). The bath solution contained 1 mM MgCl2, 2 mM CaCl2, 5 mM KCl, 150 mM NaCl, 5 mM glucose, and 10 mM Hepes, pH 7.3. Currents were obtained from cells that were held at −80 mV and depolarized to a maximum activating voltage of 120 mV. Leakage and capacitance currents were subtracted by standard P/n procedure, and series resistance was compensated to 80%. The peak current of command pulses was used to calculate maximum peak conductance (Gm) following Ohm's law [Gm=I/(Vp−EK)], where Vp represents the test voltage value and EK denotes the Nernst k equilibrium potential, which is −83 mV); maximum capacitance (Cm) was computed to normalize the difference in the cell surface area.
CHO cells co-transfected with either Kv3.3 and Kv3.1b or Kv3.3 and Kv1.1 together were collected with 1 ml of lysis buffer; cell lysates were centrifuged, and the supernatant was transferred into a new tube and pre-cleared with Protein A– or Protein G–Sepharose (Sigma–Aldrich). A total of 4 μl of anti-Kv3.3 antibody or anti-Kv3.1b antibody was added to the pre-cleared cell lysates and gently rocked at 4°C overnight. The antigen–antibody complexes were isolated with Protein A– or Protein G–Sepharose and eluted from the Sepharose beads in SDS sample buffer before being separated by SDS/PAGE.
Cycloheximide protein half-life experiment
Transfected CHO cells were treated with 20 μg/ml cycloheximide (Sigma–Aldrich) at approximately 20 h post-transfection. Cells at zero time point were harvested immediately as described above, otherwise cells were incubated with cycloheximide at 37°C for 1, 2 or 4 h. Cell lysates were generated and total proteins were separated by SDS/PAGE and immunoblotted. The films were scanned and band intensity was evaluated by densitometry. The protein levels at time 0 were set to 100, and values of the other time points were normalized to the value of time 0. Data were plotted and fitted using one-phase exponential decay to obtain the protein half-life.
Treatment by proteasome inhibitor MG132
MG132 (10 μM; Cayman Chemical) was added to transfected CHO cells at 15 h post-transfection and incubated for 2 h at 37°C. DMSO was added to the control group as the vehicle control. Cells were lysed and total protein amount was analysed as described above.
Treatments by chemical chaperone TMAO (trimethylamine N-oxide)
TMAO (100 mM; Sigma–Aldrich) was added to CHO cells after transfection. Cells were incubated at 37°C under 5% CO2 for 20 h before being processed, and surface and total proteins were analysed as described above for cell-surface biotinylation and immunoblotting.
Values in the text and Figures are presented as means±S.E.M. Statistical significance was determined by Student's t test or one-way ANOVA with Tukey's test. P<0.05 was considered to be significant.
SCA13 mutations affected functional and protein expression of Kv3.3
To examine the effect of SCA13 mutations on Kv3.3 channel activity, currents from transfected CHO cells were recorded by whole-cell voltage clamp. R336H and R420H mutants had no currents detected above background, and R423H showed substantially reduced current amplitude, whereas F448L exhibited similar current amplitude compared with wild-type Kv3.3 (Figure 1B). The maximum calculated conductance density (Gm/Cm) suggested that Kv3.3 was 17.85±1.11 nS/pF; R366H and R420H were 0.31±0.15 nS/pF and 0.73±0.15 nS/pF respectively, which were similar to the background level; R423H was 2.88±0.81 nS/pF, which was 16% of wild-type Kv3.3; and F448L was 21.62±0.85 nS/pF, which was statistically similar to the wild-type (Figure 1C). To rule out the possibility that the reduction in current level was due to failure of Kv3.3 proteins to tetramerize, we solubilized transfected cells under non-denaturing conditions and ran samples on non-denaturing gels followed by immunoblotting. The signals on immunoblots indicated that Kv3.3 and the four SCA13 mutants were all tetramerized with little or no evidence of monomers, dimers or trimers (results not shown). These results indicate that SCA13 mutations did not appear to alter tetramerization.
The effect of SCA13 mutations on protein expression was estimated by measuring surface channel protein levels by cell-surface biotinylation and immunoblotting, and total channel protein levels by immunoblotting. The three arginine-to-histidine mutants (R366H, R420H and R423H) exhibited significantly reduced surface and total protein levels, approximately 25–30% of wild-type Kv3.3 for surface protein and 28–38% of the wild-type Kv3.3 for total protein (Figures 1D and 1E). The F448L mutant showed a similar surface expression level as wild-type Kv3.3, although its total protein level was reduced to 70% of wild-type Kv3.3 (Figures 1D and 1E).
Relative surface expression (surface/total protein) can be indicative of different subcellular retention patterns of proteins compared with control. However the surface/total ratio averages of the three arginine-to-histidine mutants were approximately 0.8 compared with wild-type Kv3.3, but these values were not significantly different than control (Figure 1F). These results suggest that the arginine-to-histidine mutants had similar relative intracellular retention compared with control. On the other hand, the F448L mutant had a surface/total ratio of approximately 1.6, mirroring the fact that it had lower total protein expression, but similar surface protein expression compared with wild-type Kv3.3, whose surface/total ratio was set to 1.0 (Figure 1F). These results suggest that F448L was trafficked to the cell surface somewhat more efficiently than wild-type Kv3.3.
The results above suggest that cell-surface protein level and conductance density level of R366H, R420H and R423H were significantly decreased compared with Kv3.3, presumably because of the decreased total protein level and not due to intracellular retention, whereas the surface expression of F448L was not affected. We hypothesize that these three SCA13 arginine-to-histidine mutations affect protein stability.
SCA13 arginine-to-histidine proteins were unstable, degraded at faster rates through proteasomes and could be stabilized by a chemical chaperone
We focused next on only the three SCA13 arginine-to-histidine mutants because they exhibited a significant decrease in protein levels compared with the wild-type. The protein half-lives of Kv3.3 and the three SCA13 arginine-to-histidine mutants were estimated using cycloheximide to inhibit protein synthesis followed by immunoblotting; protein half-lives of wild-type Kv3.3, R366H, R420H and R423H were 2.2, 1.0, 1.2 and 1.2 h respectively. No reduction was found in the level of actin used as a control (Figure 2A). Not surprisingly, all three SCA13 arginine-to-histidine mutants, which had less total protein levels, showed shorter protein half-lives compared with control. The SCA13 mutants might be misfolded and unstable, and therefore be degraded faster through proteasomes initiated by quality control mechanisms in the ER (endoplasmic reticulum). This degradation pathway was tested using MG132, a proteasome inhibitor. Transfected CHO cells were incubated with MG132 or DMSO for 2 h before they were lysed, and cell lysates were used to analyse total proteins. Total protein levels of all mutants treated with MG132 were normalized to MG132-treated wild-type Kv3.3, and total protein levels of mutants treated with DMSO only were normalized to DMSO-treated wild-type Kv3.3. We then compared the protein expression of each SCA13 mutant treated with MG132 or DMSO only. Total protein expression of the SCA13 arginine-to-histidine mutants treated with MG132 was higher than their comparative controls (Figure 2B).
Chemical chaperones appear to interact with and stabilize proteins and may be effective in rescuing some mutants from the ER quality control mechanism [13,14]. We tested the effect of the chemical chaperone TMAO on increasing protein expression. Wild-type Kv3.3 and SCA13 arginine-to-histidine mutant-transfected cells were treated with 100 mM TMAO, and the signals were normalized to treated wild-type Kv3.3. TMAO partially increased/rescued total protein levels of all three SCA13 arginine-to-histidine mutants, among which R366H and R420H had ~50% and ~70% of wild-type Kv3.3 respectively, and R423H exhibited a total protein level of ~90% of Kv3.3 (Figure 2E). At the surface protein level, TMAO had little effect on R366H, but increased surface protein levels of R420H and R423H, in particular R423H (Figure 2C).
We then determined the effect of TMAO on the functional expression of R423H. After TMAO treatment, R423H did show an increase in the peak current (Figure 2D, top panel) and an increase in Gm/Cm compared with the untreated R423H (dashed line) (Figure 2D, bottom panel). The peak current of wild-type in TMAO treatment (Figure 2D, top panel) was lower compared with no TMAO. The explanation for this finding is unknown. In addition, R366H, R420H and R423H now had a relative surface expression ratio of approximately 0.5, 0.5 and 0.7 respectively, compared with Kv3.3 (Figure 2F). After TMAO treatment, all three SCA13 arginine-to-histidine mutants showed reduced surface/total ratios compared with no TMAO due to increased intracellular retention.
The results described above suggest that the SCA13 arginine-to-histidine mutants are unstable proteins and are degraded through proteasomes at faster rates, and that TMAO appears to stabilize the mutant proteins and helps protein folding.
SCA13 arginine-to-histidine mutant surface protein levels were not rescued by co-expression with Kv3.1b
Kv3 channels can form heterotetramers with different Kv3 members, and Kv3.3 should form functional heteromers with Kv3.1 [2,15,16]. It has been suggested that subunit folding and channel tetramerization steps alternate during channel biogenesis, and assembly of native subunits could help some mutants to fold properly . To study the consequence Kv3.1b has on SCA13 arginine-to-histidine mutants, co-transfection experiments were conducted. First, the formation of Kv3.1b–Kv3.3 heteromers was examined by immunoprecipitation and immunoblotting. Kv3.1b proteins were precipitated with the anti-Kv3.3 antibody, and Kv3.3 proteins were also precipitated by the anti-Kv3.1b antibody (Figures 3A and 3B). As a control, Kv1.1 co-transfected with Kv3.3 was not precipitated by the anti-Kv3.3 antibody (Figure 3C). These results suggest that Kv3.3 formed tetramers with Kv3.1b, but not with Kv1.1, in CHO cells. We then tested the effect of Kv3.1b on protein expression of Kv3.3. Kv3.3 and Kv3.1b were co-transfected into CHO cells at a 1:4 ratio so that most of Kv3.3 subunits were in Kv3.3–Kv3.1b complexes. The three SCA13 arginine-to-histidine mutant surface protein levels were not rescued with co-expression with Kv3.1b, and there was little effect on the total protein levels and the surface/total ratios (Figures 3D–3F). The above results suggest that Kv3.1b did not rescue surface protein levels of the SCA13 arginine-to-histidine mutants.
Lysine substitution at the three SCA13 arginine-to-histidine positions partially rescued protein expression, whereas alanine substitution was less effective
The three SCA13 arginine-to-histidine positions are conserved in Kv channels (Figure 4A) and these side chains could play a substantial role in maintaining the proper structure and/or function of Kv3.3 proteins. We further explored the importance of these SCA13 arginine-to-histidine positions by making conservative amino acid substitutions (R366K, R420K and R423K) to test their effects on protein expression. Lysine has a similar side chain as arginine and should be less disruptive than histidine. The surface protein, the total protein and the surface/total ratio were higher with arginine-to-lysine mutations compared with the SCA13 arginine-to-histidine mutations, although there was not a complete rescue to wild-type surface and total protein levels (Figures 4B–4D).
In addition, we also used alanine substitution because it is the most abundant amino acid in proteins, is found in a variety of secondary structures and is frequently used in alanine-scanning mutagenesis . The surface protein, total protein and surface/total ratio for the arginine-to-alanine mutations were more similar to the SCA13 arginine-to-histidine mutations (Figures 4E–4G).
The above results suggest that the SCA13 arginine-to-lysine substitutions partially rescued surface protein, total protein and the surface/total ratio compared with the arginine-to-histidine mutations, whereas the arginine-to-alanine substitutions were more similar to the mutants.
In the present study we have shown evidence that, although SCA13 R366H, R420H and R423H mutant Kv3.3 proteins are unstable and degraded faster through proteasomes, they are still trafficked to the CHO cell surface. A chemical chaperone (TMAO) also partially rescued the total protein levels of these SCA13 mutants. TMAO appears to form hydrogen bonds with the backbone nitrogen of solvent-exposed hydrophobic peptide regions, affects the free energy and destabilizes the unfolded state, and drives the free energy equilibrium towards the native state with higher folding rates [19,20]. Whatever the mechanism, these stabilized mutants do not appear to be trafficked to the cell surface as efficiently as wild-type Kv3.3.
SCA13 R366H and R420H mutants had little, if any, detectable voltage-gated currents, and this was similar to previous studies using the Xenopus oocyte expression system [10,11]. We also showed that the R423H mutant exhibited a lower current amplitude of 17% of wild-type Kv3.3 in CHO cells, which was inconsistent with another group's finding that R423H was non-functional in Xenopus oocytes . This difference may be due to the two different expression systems. Xenopus oocytes could be different from mammalian cell lines such as CHO cells in trafficking and intracellular processing pathways as suggested by several studies [21–23]. The SCA13 F448L mutant exhibited altered gating  and we found that this mutation had little effect on surface and total protein levels compared with wild-type.
The Kv3.3 arginine residues at positions 366, 420 and 423 are almost invariable among Kv channels, suggesting that they have fundamental roles in protein folding and/or function. Arg366 is an important basic residue in the S2 sensor that contributes to voltage-dependent properties, and the crystal structure of a voltage-gated potassium channel also suggests that it is important for electrostatic interactions . Indeed, mutating the analogous arginine residue (Arg297 in Shaker notation) to the neutral amino acid glutamine in the Shaker channel produced non-functional channels, but the conservative R297K mutation shifted the activation and inactivation curves without affecting other functional properties . Arg420 and Arg423 are critical to the voltage sensor not only because of their positive charges, but also their structural necessity for the voltage sensor or channel conformation. The crystal structure of the Kv1.2 channel suggests that the analogous arginine residues Arg420 and Arg423 face S1 and S2 helices, and make salt bridge interactions with acidic amino acids . Although there is no crystal structure for the Kv3.3 channel, these arginine residues are speculated to have a similar conformation given the fact that these residues are highly conserved and that their position in KvAP is consistent with Kv1.2 [24,26]. Comparing the protein expression results of arginine, histidine (the SCA-13 mutations), alanine and lysine at residues 366, 420 and 423, we find that histidine is the most disruptive mutation at all three positions since it gives the least total protein expression; alanine is in the middle of the three non-arginine mutations; and lysine is least disruptive, presumably because it has a side chain that resembles arginine.
Kv3.1b association slightly stabilized SCA13 mutant proteins, yet it had little effect on their surface expression. SCA13 patients are expected to have a 1:1 expression ratio of wild-type and mutant Kv3.3, and both could form tetramers with Kv3.1. Thus, in some brain regions such as DCN (deep cerebellar nuclei), SCA13 patients are expected to have a mixed combination of wild-type Kv3.3, SCA13 Kv3.3 and wild-type Kv3.1b. Our study focused on two cases: (i) homotetramers of wild-type Kv3.3 or SCA13 mutants; and (ii) heterotetramers from four Kv3.1b subunits and one type of either wild-type or SCA13 Kv3.3. Although not an exact reflection of the situation in patients, the present study sheds some light on the mechanisms underlying the SCA13 disease. Kv3.1b seems to have little effect on rescuing SCA13 mutants to the cell surface. It is, however, not clear that to what extent Kv3.1b could affect the function of SCA13 mutants, or vice versa. Nevertheless, it is expected that the dominant-negative SCA13 mutants could disrupt all Kv3 currents involving Kv3.3 subunits by forming tetramers. The functional importance of Kv3.3 and Kv3.1 has been underscored from several knockout mice and rescue studies [27–32]. Kcnc3-null mice had comparable gait with wild-type mice, but showed increased lateral deviation during ambulation and increased slips when crossing a narrow beam originated from application force [29–31]. Kcnc1/Kcnc3 double knockout mice exhibited tremor and all the signs of cerebellar gait ataxia [28,29].
Overall our studies extend our knowledge of the three SCA13 arginine-to-histidine mutants concerning the defects in both the functional and biochemical expression of these proteins.
Jian Zhao performed most of the experimental work. Jing Zhu performed the experiments in Figure 2(D). Jian Zhao, Jing Zhu and William Thornhill designed the experiments and wrote the paper.
This work was partially supported by the National Institutes of Health [grant number R15 NS48906 (to W.B.T.)].
Abbreviations: CHO, Chinese-hamster ovary; ER, endoplasmic reticulum; SCA, spinocerebellar ataxia; TMAO, trimethylamine N-oxide
- © The Authors Journal compilation © 2013 Biochemical Society