Kv10.1 Regulates Microtubule Dynamics during Mitosis

Kv10.1 (potassium voltage-gated channel subfamily H member 1, known as EAG1 or Ether-à-go-go 1), is a voltage-gated potassium channel, prevailingly expressed in the central nervous system. The aberrant expression of Kv10.1 is detected in over 70% of all human tumor tissues and correlates with poorer prognosis. In peripheral tissues, Kv10.1 is expressed almost exclusively during the G2/M phase of the cell cycle and regulates its progression—downregulation of Kv10.1 extends the duration of the G2/M phase both in cancer and healthy cells. Here, using biochemical and imaging techniques, such as live-cell measurements of microtubule growth and of cytosolic calcium, we elucidate the mechanisms of Kv10.1-mediated regulation at the G2/M phase. We show that Kv10.1 has a dual effect on mitotic microtubule dynamics. Through the functional interaction with ORAI1 (calcium release-activated calcium channel protein 1), it modulates cytosolic calcium oscillations, thereby changing microtubule behavior. The inhibition of either Kv10.1 or ORAI1 stabilizes the microtubules. In contrast, the knockdown of Kv10.1 increases the dynamicity of mitotic microtubules, resulting in a stronger spindle assembly checkpoint, greater mitotic spindle angle, and a decrease in lagging chromosomes. Understanding of Kv10.1-mediated modulation of the microtubule architecture will help to comprehend how cancer tissue benefits from the presence of Kv10.1, and thereby increase the efficacy and safety of Kv10.1-directed therapeutic strategies.

We have previously shown that, in non-neural cells, Kv10.1 localizes to the centrosomes and primary cilia [21], and that its expression occurs only during the G2/M phase of the cell cycle [22]. Downregulation of Kv10.1 induces accumulation of cells in the G2/M phase, suggesting that cells depleted of Kv10.1 require a longer time to complete G2/M, which implies the participation of the channel in the regulation of this phase of the cell cycle [22]. phosphorylated (p-) ATR normalized for actin was doubled in KD cells (0.41 ± 0.03 a.u. vs. 0.20 ± 0.01 a.u. in control; Figure 1A). p-CHK1 was increased by 60% upon Kv10.1 knockdown (0.14 ± 0.03 a.u. vs. 0.09 ± 0.02 a.u. in control; Figure 1B). The total level of CHK2 was increased by approximately 70% in Kv10.1-deficient HeLa cells (0.24 ± 0.01 a.u. in the KD vs. 0.14 ± 0.01 a.u. in controls; Figure  1C). However, we observed no significant differences in the expression of phosphorylated BRCA1 (Ser1524-A phosphorylation site for ATR [46]), total CHK1, p21, and p-histone H2A.X (Ser139; Figure S1). with either Scr-or KCNH1-siRNA, synchronized with double thymidine block at the G1/S border, and subsequently released from the block. Samples were collected at the indicated time-points after release from the block. Representative immunoblots and the results of densitometry analysis are plotted. The time interval corresponding to Kv10.1 expression is indicated in pink (n = 3 independent experiments; mean ± SEM; two-way ANOVA, repeated measures, Bonferroni post hoc test; the exact p-values indicate the significance of the global difference between the groups, and the p-value between single time points is indicated by asterisks: * p < 0.05, ** p < 0.01, *** p < 0.001). (D-H) HeLa and hTERT RPE1 cells were transiently transfected with either Scr-or KCNH1-siRNA, allowed to reach 70% confluency, harvested, and immediately used for single-cell gel electrophoresis (comet assay). (D) Representative images of the DNA comets stained with SYBR-Gold are given (max intensity projection, z-step, 0.  with either Scror KCNH1-siRNA, synchronized with double thymidine block at the G1/S border, and subsequently released from the block. Samples were collected at the indicated time-points after release from the block. Representative immunoblots and the results of densitometry analysis are plotted. The time interval corresponding to Kv10.1 expression is indicated in pink (n = 3 independent experiments; mean ± SEM; two-way ANOVA, repeated measures, Bonferroni post hoc test; the exact p-values indicate the significance of the global difference between the groups, and the p-value between single time points is indicated by asterisks: * p < 0.05, ** p < 0.01, *** p < 0.001). (D-H) HeLa and hTERT RPE1 cells were transiently transfected with either Scror KCNH1-siRNA, allowed to reach 70% confluency, harvested, and immediately used for single-cell gel electrophoresis (comet assay). (D) Representative images of the DNA comets stained with SYBR-Gold are given (max intensity projection, z-step, 0.2 µm, scale bar, 200 µm). (E-H) The comets were analyzed with the CASPLab software and Head DNA % and Olive Tail Moment are plotted in (E,F) for HeLa cells, and in (G,H) for hTERT RPE1 cells (n = 5 independent experiments with >750 comets per condition; Kruskal-Wallis test with Dunn's post hoc test, min to max range with the median indicated as a line are shown, with p-values indicated for each pair). (Kv10.1-potassium voltage-gated channel subfamily H member 1; KCNH1-potassium voltage-gated channel subfamily H member 1 gene; Scr-scrambled; p-phosphorylated; ATR-ATM and Rad3 related kinase; CHK-checkpoint kinase; HeLa-Henrietta Lacks; hTERT RPE1-human telomerase reverse transcriptase immortalized retinal pigmented epithelial).
To elucidate whether the activation of components of the checkpoint actually responded to increased DNA damage, we studied the presence and extent of DNA lesions using single-cell gel Cancers 2020, 12, 2409 4 of 26 electrophoresis, also known as the comet assay [47]. The analysis of DNA comets in asynchronously growing HeLa and hTERT RPE1 cells did not reveal any significant changes in the magnitude of DNA damage between the control and the Kv10.1 knockdown group ( Figure 1D-H). The observed differences in the DNA damage-induced checkpoint components upon Kv10.1 depletion are hence independent of DNA damage.

Kv10.1 Downregulation Leads to Activation of Spindle Assembly Checkpoint and Decreased Occurrence of Lagging Chromosomes in HeLa Cells
Although the initiation of DDR occurs in the nucleus, some of the DNA damage-induced checkpoint proteins are also localized at centrosomes and are linked to SAC during mitosis, independently of DNA damage [48][49][50][51][52][53][54][55][56][57][58][59]. We, therefore, set out experiments to investigate a possible activation of SAC upon Kv10.1 knockdown. HeLa cells were transfected with Scror KCNH1-siRNA and synchronized at the G1/S border using a double thymidine block. Samples of synchronized cells were collected at the indicated time-points and analyzed by immunoblotting (Figure 2A-G). Densitometry analysis of the immunoblots showed a 20% increase in CDC20 in Kv10.1 KD cells (0.28 ± 0.05 a.u.) compared to the control group (0.23 ± 0.04 a.u.; Figure 2A). Similarly, MAD2 was 27% higher in Kv10.1 deficient HeLa cells (0.26 ± 0.02 a.u.) compared to the control cells (0.19 ± 0.02 a.u.; Figure 2B). To elucidate whether the activation of components of the checkpoint actually responded to increased DNA damage, we studied the presence and extent of DNA lesions using single-cell gel electrophoresis, also known as the comet assay [47]. The analysis of DNA comets in asynchronously growing HeLa and hTERT RPE1 cells did not reveal any significant changes in the magnitude of DNA damage between the control and the Kv10.1 knockdown group ( Figure 1D-H). The observed differences in the DNA damage-induced checkpoint components upon Kv10.1 depletion are hence independent of DNA damage.

Kv10.1 Downregulation Leads to Activation of Spindle Assembly Checkpoint and Decreased Occurrence of Lagging Chromosomes in HeLa Cells
Although the initiation of DDR occurs in the nucleus, some of the DNA damage-induced checkpoint proteins are also localized at centrosomes and are linked to SAC during mitosis, independently of DNA damage [48][49][50][51][52][53][54][55][56][57][58][59]. We, therefore, set out experiments to investigate a possible activation of SAC upon Kv10.1 knockdown. HeLa cells were transfected with Scr-or KCNH1-siRNA and synchronized at the G1/S border using a double thymidine block. Samples of synchronized cells were collected at the indicated time-points and analyzed by immunoblotting (Figure 2A-G). Densitometry analysis of the immunoblots showed a 20% increase in CDC20 in Kv10.1 KD cells (0.28 ± 0.05 a.u.) compared to the control group (0.23 ± 0.04 a.u.; Figure 2A). Similarly, MAD2 was 27% higher in Kv10.1 deficient HeLa cells (0.26 ± 0.02 a.u.) compared to the control cells (0.19 ± 0.02 a.u.; Figure 2B). with double thymidine block at the G1/S border (A-I) or with consecutive thymidine and nocodazole treatments at metaphase and anaphase (J-M). (A-G) Samples were collected at the indicated time-points after cell cycle re-initiation and immunoblotted. Representative immunoblots and densitometry analysis results are given. The time interval corresponding to Kv10.1 expression is highlighted in pink (n = 3 independent synchronization experiments; mean ± SEM, two-way ANOVA, repeated measures, Bonferroni post hoc test, the given p-value shows the significance of the global difference between the two groups. (H,I) PLK1 enzymatic activity was measured in whole-cell lysates from HeLa (left grey panel) and hTERT RPE1 (right blue panel) cells synchronized at G2/M (H) and G1/S (I) (n = 4 independent synchronization experiments; two-way ANOVA, repeated measures, mean ± SEM is shown). (J-M) The mitotic spindles in synchronized HeLa cells were visualized with α-tubulin (green), and the chromosomes-with CENP-B (kinetochores, red) and Hoechst 33342 (DNA, blue) (scale bars represent, 10 µm and apply for all panels). We also tested the status of the APC/C complex, a direct target of the MCC. The APC/C protein CDC27 (cell division cycle protein 27) was downregulated by 20% in Kv10.1 KD (0.24 ± 0.02 a.u.) with respect to the control cells (0.30 ± 0.03 a.u.; Figure 2C). Notably, the most significant difference between the groups occurred six hours after the release (a 60% decrease: 0.45 ± 0.09 a.u. in controls vs. 0.28 ± 0.05 a.u. in KD cells; Figure 2C).
The APC/C activity requires CDC27 [60], and the absence of CDC27 would originate a compensatory upregulation of other factors. The kinase WEE1 is a plausible candidate for such an upregulation. WEE1 inactivates the cyclin B1/CDK1 complex both at the entry into mitosis [61] and at the metaphase-to-anaphase transition [62]. WEE1 expression was increased by 30% in Kv10.1-deficient cells (0.16 ± 0.01 a.u.) compared to the control group (0.12 ± 0.01 a.u.; Figure 2D). Previously, we have shown that Kv10.1 downregulation in HeLa delays the degradation of cyclin A2 during G2/M phase [22]. Interestingly, cyclin A2 degradation has been suggested as a reporter for SAC activity-the higher the rate of MCC formation, the later cyclin A2 is degraded [63,64]. Additionally, in line with our previous observations [22], slightly broader peaks of CDC20 and WEE1 at the G2/M phase hint to extended and stronger SAC activity and the prevention of premature activation of APC/C in the absence of Kv10.1.
Besides MCC, other factors can reinforce the SAC signal. For instance, pVHL (von Hippel-Lindau protein) enhances SAC response by increasing MAD2 levels, which in turn reduces chromosome instability [64,65]. Immunoblot analysis showed a moderate increase in the total levels of pVHL in Kv10.1 KD cells (0.34 ± 0.03 a.u. vs. 0.27 ± 0.02 a.u. in control; Figure 2E). While MCC detects unattached kinetochores [25], the accuracy of already-formed kinetochore-MT attachments is modulated by the AURKB (Aurora kinase B)-PLK1 (Polo-like kinase 1)-MCAK (mitotic centromere-associated kinesin) axis [66]. We assessed the involvement of this pathway by monitoring changes in its expression during the cell cycle in synchronized HeLa cells transiently transfected with either Scror KCNH1-siRNA ( Figure 2F,G). Interestingly, the total amount of AURKB was downregulated by 35% in Kv10.1-deficient HeLa cells (0.29 ± 0.03 a.u.) compared to control cells (0.40 ± 0.06 a.u.; Figure 2F). The localization of AURKB at the kinetochores is a crucial step in recruiting the SAC components [67] led by MAD2 [68].
Moreover, CHK1 and CHK2 help to correctly localize and activate AURKB, BUBR1, MAD1 at kinetochores [69][70][71][72]. Hence, although we observe a decrease in total AURKB in Kv10.1-deficient cells ( Figure 2F), we can speculate that the amount of "effective" AURKB at the kinetochores may increase due to upregulation of both phosphorylated CHK1 ( Figure 1B) and MAD2 ( Figure 2B). Even a marginal loss of PLK1 activity impairs chromosome congression [66,73]. Like AURKB, PLK1 was 30% lower in Kv10.1 KD cells (0.11 ± 0.01 a.u. vs. 0.16 ± 0.02 a.u. in control cells; Figure 2G). However, the decreased total amount of PLK1 did not correlate with the enzymatic activity measured during the G2/M phase. A moderately higher PLK1 activity was detected both in Kv10.1-deficient HeLa ( Figure 2H, left grey panel) and hTERT RPE1 cells ( Figure 2H, right blue panel). As expected, cells synchronized at G1/S showed low overall PLK1 activity ( Figure 2I). Thus, increased PLK1 enzymatic activity upon Kv10.1 loss in mitotic cells ( Figure 2H) can contribute to the strength of SAC by modulating the kinetochore-MT attachments and facilitating the recruitment of MCC to kinetochores.
The upregulation of major components of SAC in synchronized HeLa cells depleted of Kv10.1 (summarized in Table 1) raises two questions. (i) Are SAC components upregulated due to chromosome misalignment during metaphase? And (ii) Does the more robust and prolonged SAC in Kv10.1 KD cells decrease the occurrence of lagging chromosomes? We approached these questions using HeLa cells transfected with either Scror KCNH1-siRNA again, and synchronized at the metaphase, to study chromosome alignment, and at anaphase-for the occurrence of lagging chromosomes. Bipolar mitotic spindle, kinetochores, and DNA were visualized in fixed synchronized cells using a confocal fluorescence microscope ( Figure 2J,L). Kv10.1 knockdown had almost no effect on the number of defective metaphases, which are characterized by the presence of misaligned chromosomes or damaged DNA ( Figure 2J,K). In contrast, KCNH1-siRNA-treated HeLa cells had approximately 20% less lagging chromosomes (29.32 ± 0.45%, than the control cells (35.78 ± 0.45%; Figure 2L,M), suggesting a more efficient error-correction of kinetochore-MT attachments in KD cells.

Kv10.1 Knockdown Leads to Mitotic Spindle Tilting and Increased Microtubule Plus-End Growth
One of the causes of SAC activation can be an altered positioning of the mitotic spindle. HeLa cells with a larger tilting of the mitotic spindle require more time for biorientation of sister kinetochores due to prolonged activation of the spindle checkpoint [74]. Moreover, both PLK1 and pVHL, altered in Kv10.1-depleted cells, modulate the positioning of the mitotic spindle [75][76][77][78][79][80][81]. We measured the mitotic spindle angle in HeLa cells synchronized at metaphase ( Figure 3A). The mitotic spindle angle was determined as the angle between the line connecting the two poles and the growing substrate ( Figure 3B). As predicted, the mitotic spindle angle almost doubled in the Kv10.1 KD HeLa cells (17.20 • ) compared to the controls (9.68 • ; Figure 3C). substrate ( Figure 3B). As predicted, the mitotic spindle angle almost doubled in the Kv10.1 KD HeLa cells (17.20°) compared to the controls (9.68°; Figure 3C). To understand the exact mechanism underlying the effect of Kv10.1 expression on SAC activity and mitotic spindle positioning, we studied the behavior of MTs. It is well-established that both SAC and spindle orientation on one side and MT dynamics on the other are highly interdependent [26,82,83]. We studied the effect of Kv10.1 KD on MT dynamics in cells treated with DME (dimethylenastron). DME is an Eg5 kinesin inhibitor that induces the formation of monopolar prometaphase spindles without affecting the assembly rates of MT plus-tips [84]. Yet, cells with altered MT dynamics and mitotic spindle angle form asymmetric monopolar spindles [85,86]. HeLa cells were treated with 2.5 µM DME, and the mitotic spindles were inspected under a fluorescence To understand the exact mechanism underlying the effect of Kv10.1 expression on SAC activity and mitotic spindle positioning, we studied the behavior of MTs. It is well-established that both SAC and spindle orientation on one side and MT dynamics on the other are highly interdependent [26,82,83]. We studied the effect of Kv10.1 KD on MT dynamics in cells treated with DME (dimethylenastron). DME is an Eg5 kinesin inhibitor that induces the formation of monopolar prometaphase spindles without affecting the assembly rates of MT plus-tips [84]. Yet, cells with altered MT dynamics and mitotic spindle angle form asymmetric monopolar spindles [85,86]. HeLa cells were treated with 2.5 µM DME, and the mitotic spindles were inspected under a fluorescence microscope for the presence of asymmetric monopolar spindles ( Figure 3D). Kv10.1-deficient HeLa cells displayed more asymmetric monopolar spindles (33%) than the control group (24.20%; Figure 3E). The DME-assay has also been suggested as a screen-test for an altered mitotic spindle angle [86]. The DME-assay confirms the results from the direct measurement of the mitotic angle in metaphase-arrested cells: Kv10.1 downregulation is associated with an increased mitotic spindle angle.
To further validate the initial observation of increased MT dynamics in Kv10.1 KD cells, we measured the MT plus-end growth rates directly in living cells with transient Kv10.1 downregulation (upper blue panel, Figure 3F-I; Videos S1 and S2) or upregulation (lower green panel, Figure 3F-I; Videos S3 and S4). The growing MTs were detected by expressing EB3-tdTomato (end-binding protein 3 tandem dimer tomato fluorescent protein) and monitored in cells synchronized in prometaphase. The MT plus-end growth speed, growth length, and dynamicity were then computed with the U-Track software. It should be noted that we excluded the visibly asymmetric monopolar spindles from the analysis to ensure an unbiased assessment of MT dynamics. Due to that, however, cells with the most substantial effects are discarded, and our analysis display only minor changes in MT dynamics under the described conditions. Nevertheless, the growth speed was increased by 8.5% in Kv10.1 KD HeLa cells (20.77 µm·min −1 vs. 19.12 µm·min −1 in control cells; upper blue panel, Figure 3G). Similar to the growth speed, the frequency rate of switching from growth to shrinkage and vice-versa, known as dynamicity, was also increased in knockdown cells (top blue panel, Figure 3H). The displacement of the growing tip (growth length), was increased only by 3.6% upon Kv10.1 downregulation (upper blue panel, Figure 3I). Equivalent results were obtained in hTERT RPE1 cells with transient downregulation of Kv10.1 ( Figure S2).
Opposite to Kv10.1 downregulation, overexpression of the channel globally reduced the MT dynamics (lower green panel, Figure 3G-I; Videos S3 and S4). The growth speed, dynamicity, and growth length were also lower in pKCNH1-WT-expressing cells compared to the cells transfected with the empty vector. This result strongly suggests that the alteration of MT dynamics is directly dependent on Kv10.1.
Despite the negative correlation between the MT dynamics and Kv10.1 expression, the causal relationship between Kv10.1 presence and the observed phenotypes remains to be defined. Whether the changes in the expression levels of regulatory proteins appear first, altering the MT behavior, or if the Kv10.1-dependent changes in MT dynamicity lead to mitotic spindle tilting and subsequently to SAC activation remains to be elucidated. We, nevertheless, favor the possibility that Kv10.1 participates in G2/M regulation by modulating the MT behavior, since spindle orientation is not affected by SAC inhibition [74].

Block of Kv10.1 or ORAI1 Currents Reduces Microtubule Dynamics
The absence of Kv10.1 abolishes both its canonical function (flow of K + ) and any possible non-canonical function, such as its participation in supramolecular complexes. K + currents hyperpolarize the membrane and, therefore, increase the driving force for Ca 2+ [87], which controls MT dynamicity both in vitro and in vivo [37,38,[88][89][90]. Ca 2+ ions also regulate many MT-associated proteins present on the growing tips [40,91]. Besides, there is evidence that Kv10.1 interacts functionally with the Ca 2+ channel ORAI1 and participates in store-independent Ca 2+ entry (SICE) [41,92]. Thus, we tested the possible implication of such a functional tandem in MT growth during the G2/M phase.
We first studied the colocalization of Kv10.1 and ORAI1 in mitotic cells using proximity ligation assay (PLA). As shown in Figure 4A, using Kv10.1 and ORAI1-directed antibodies, the PLA gave rise to a signal, indicating the proximity of the two channels also in mitotic cells (zoomed-in inset, Figure 4A).  Further, we studied the effect of each of the channel conductance on MT dynamics in live-cell imaging experiments, as before. To block Kv10.1, we used the only specific blocker available, the monoclonal antibody mAb56 [15]. For ORAI1 inhibition, we used AnCoA4, a potent drug directed against the C-terminus of ORAI1 [93]. As controls, we used either the vehicle for AnCoA4 DMSO (dimethyl sulfoxide) or non-specific mouse immunoglobulin of the same isotype as mAb56-IgG (immunoglobulin G) κ2b. The results are presented in Figure 4B-E and Videos S5-S10. Analysis of the time-lapse series showed that both AncoA4 and mAb56 significantly reduced all MT behavioral parameters: growth speed, dynamicity, and growth length (left grey and middle turquoise panels, Figure 4C-E). The effects were not additive: mAb56 and AnCoA4 combination treatment resulted in a smaller reduction in growth speed (right purple panel, Figure 4C). MT dynamicity was reduced by 10.7% in the cells exposed to AnCoA4 (left grey panel, Figure 4D), and by 8.7% in cells treated with mAb56 (middle turquoise panel, Figure 4D). Simultaneous treatment of the cells with AnCoA4 and mAb56 decreased MT dynamicity slightly less than each blocker alone (right purple panel, Figure 4D). The effect was similar on growth length, which was shortened by 14.1% in AnCoA4 (left grey panel, Figure 4E), 4.5% in mAb56 (middle turquoise panel, Figure 4E), and 9% in the combination (right purple panel, Figure 4E).
In summary, AnCoA4 produced significant reductions of all parameters of MT dynamics, and mAb56 still induced significant but smaller effects than AnCoA4. The combined treatment showed no additivity, suggesting a shared mechanism for both blockers and, therefore, a common pathway influenced by Kv10.1 and ORAI1. This observation would be compatible with MT growth regulation by Ca 2+ entering the cell through ORAI1 in a Kv10.1-dependent manner. It is essential to point out that the changes induced by channel blockade are opposite to those caused by channel knockdown. Possible reasons for this difference are discussed below.

Kv10.1 Regulates Intracellular [Ca 2+ ] Oscillations
We have shown that Kv10.1 and ORAI1 are in proximity, and that block of currents of either of them results in altered MT behavior during prometaphase, raising the possibility that Kv10.1 activity induces changes in cytosolic [Ca 2+ ] through the functional interaction with ORAI1.
We tested the effect of blockade of ORAI1 or Kv10.1, and knockdown of Kv10.1, on the behavior of intracellular [Ca 2+ ] during the G2/M transition in HeLa cells. As above, ORAI1 block was achieved by AnCoA4 treatment (using DMSO as a control); Kv10.1 was blocked by mAb56 (mouse IgG κ2b as a control); and Kv10.1 was transiently downregulated by KCNH1-siRNA (non-targeting Scr-siRNA-control). Before imaging, the cells were loaded with the intracellular Ca 2+ indicator Fluo-4AM, and the fluorescent signal was recorded over 120 min, which roughly corresponds to the entire duration of M-phase ( Figure 5A, Videos S11-S16). The recorded signal was then normalized for photobleaching ( Figure 5B). The area under the curve (AUC) and the width and the height of local maxima in the spectrum were determined, as shown in Figure 5C. The mean AUC, which indicates the relative intracellular Ca 2+ concentration, was 16.7% smaller in AnCoA4-treated cells (left grey panel, Figure 5D). The block of Kv10.1 currents also reduced the relative amount of intracellular Ca 2+ , although by only 3.8% (middle turquoise panel, Figure 5D). In comparison, the average intracellular [Ca 2+ ] was elevated by 10% in Kv10.1 KD cells (right blue panel, Figure 5D). Changes both in width ( Figure 5E) and in the prominence of the peaks ( Figure 5F) can explain differences in AUCs. The block of either of the channels resulted in narrower peaks (left grey and middle turquoise panels, Figure 5E).
In contrast, the height of the peaks was not significantly affected (left grey and middle turquoise panels, Figure 5F). Opposingly, Kv10.1 silencing produced more prominent spikes (right blue panel, Figure 5F), but did not change the duration of the peaks (right blue panel, Figure 5G).
Cancers 2020, 12, x FOR PEER REVIEW 11 of 26 panels, Figure 5F). Opposingly, Kv10.1 silencing produced more prominent spikes (right blue panel, Figure 5F), but did not change the duration of the peaks (right blue panel, Figure 5G). (Videos S11-S16) Ca 2+ imaging was performed at 7.5 h after initiation of the cell cycle from the G1/S border using HeLa cells preloaded with Fluo-4AM. Time-lapse images were taken every 5 s over 120 Calcium signals are not only encoded in the absolute magnitude of changes but also the frequency and amplitude of calcium oscillations [94]. Therefore, using the same normalized recordings, we calculated the changes in the frequency of intracellular Ca 2+ oscillations converting the time domain of the signal into the frequency domain by Fast Fourier transform ( Figure 5G). The block of ORAI1 significantly suppressed calcium oscillations. The absolute frequencies were already smaller in AnCoA4-treated cells ( Figure 5H). Considering the relative power of each frequency, the average frequency of Ca 2+ oscillations was 62.5% lower in ORAI1 blocked cells (3.45 mHz) than in the DMSO-treated cells (9.19 mHz; left grey panel, Figure 5I). As shown in Figure 5H,I (middle turquoise panel), the antibody-mediated block of Kv10.1 did not affect the frequency of oscillations, while Kv10.1 downregulation increased the average frequency of Ca 2+ oscillations (right blue panel, Figure 5H). The weighted average frequency was higher in Kv10.1 KD (6.86 mHz) compared to the control group (6.32 mHz; right blue panel, Figure 5I).
As already mentioned, MT growth is affected by Ca 2+ ions: Tubulin polymerization is inhibited in the presence of high [Ca 2+ ] in vitro [37,38]. Therefore, Ca 2+ oscillations would induce transitions from MT growth to MT shrinkage. A decrease of Ca 2+ entry and oscillations would result in reduced dynamicity, while faster oscillations would accelerate the switching between MT growth and shrinkage. The block of ORAI1 decreases both the total cytosolic [Ca 2+ ] and the frequency of cytosolic Ca 2+ oscillations. The block of Kv10.1 has a similar effect, albeit less intense. The block of either of the channels shortens the duration of Ca 2+ peaks, and thereby the absolute amount of Ca 2+ ions entering the cell.
In contrast, siRNA-mediated downregulation of Kv10.1 has opposite effects on intracellular Ca 2+ , substantially increasing intracellular [Ca 2+ ] and frequency of Ca 2+ oscillations. Therefore, Kv10.1 current is required for regulating the duration of Ca 2+ spikes, while the frequency is dependent not on the current but the physical presence of Kv10.1. Thus, our data strongly suggest that the origin of changes in MT dynamics under the different treatments reported here correlate to the effects on intracellular Ca 2+ dynamics. Additionally, the mitotic spindles of Kv10.1 and ORAI1-inhibited cells appeared smaller and deteriorated ( Figure 4A), which is in good agreement with the previously described effect of Ca 2+ chelation on the mitotic spindle [39].
Differences between the effect of functional blockade and knockdown of the target are known for other proteins, like AURKB [95]. Our current observation of opposite effects of functional inhibition and depletion of Kv10.1 on MT dynamics and intracellular Ca 2+ oscillations is also reminiscent of our previous results on the role of Kv10.1 in primary cilium homeostasis [21]. While the potassium current is necessary for primary cilium disassembly, the overexpression of the channel prevents reciliation in a way independent of the canonical function of the channel, since a non-permeant mutant also produces the effect, as does the C-terminal cytoplasmic domain without the transmembrane segments (see Figure EV3 in [21]). Similarly, abolishing the ability of Kv10.1 to conduct ions through mutagenesis is not enough to completely inhibit its tumor-promoting activity [96]. Furthermore, a non-conducting Kv10.1 splice variant (E65, found in cancer cell lines), activates CDK1 when injected into Xenopus oocytes [97]. The depletion of Kv10.1 would, thus decrease both the full length and the spliced forms of the channel and prevent E65-mediated CDK1 activation. In contrast, this effect would not occur when the channel function is inhibited.
Since canonic store-operated Ca 2+ entry is not active during mitosis [98], the possible source of Ca 2+ in dividing cells is SICE. Both membrane expression and activity of ORAI1 are dependent on SPCA2 (Secretory Pathway Ca 2+ -ATPase) [99][100][101]. Recently, it has been shown that SPCA2 also enhances Kv10.1 membrane expression, suggesting a new model of SICE [92]. In this scenario, it is tempting to speculate that the supramolecular complex between SPCA2, ORAI1, and Kv10.1 participates in providing the Ca 2+ required for MT growth. Kv10.1, which is inhibited by Ca 2+ -calmodulin [102,103], would provide the negative control segment of the trio, terminating the signal.

Cell Cycle Synchronization
Depending on the experiment, HeLa cells were synchronized at various stages of the cell cycle. When needed, 4 h before the synchronization, the cells were transfected with the corresponding siRNAs.
Changes in proteins of interest over time were studied in HeLa cells synchronized at the G1/S border by growing the cells for two periods of 16 h with excess thymidine (2 mM; Sigma-Aldrich, Munich, Germany) and 8 h release in-between. The cells blocked at G1/S were then transferred into fresh growth medium, and cell samples were collected every 2 h between 0 to 6 h and 10 to 14 h, and every hour between 6 to 10 h (G2/M phase). The samples were then used for immunoblot experiments.
For metaphase and anaphase synchronization, the cells were treated only once with 2 mM thymidine for 16 h, then rinsed twice with DPBS (Dulbecco's phosphate-buffered saline; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and released in fresh growth medium for 4 h. Subsequently, the cells were exposed to 2 ng/mL nocodazole for 4 h and washed again with DPBS as before. The cells were harvested either at this stage (for PLK1 activity) or synchronized further. For the anaphase, the cells were released for 40 min into a standard growth medium. In the case of metaphase, the cells were further incubated with 15 µM MG132 for 2 h at 37 • C and 5% CO 2 .
Monopolar spindles were produced by treating asynchronously growing HeLa cells with 2.5 µM DME (dimethylenastron; Calbiochem, Sigma-Aldrich, Munich, Germany) for 4 h at 37 • C and 5% CO 2 . The cells were then either immediately fixed and used for immunofluorescence or imaged within two hours.
For intracellular calcium imaging, HeLa cells were synchronized at the G2/M border by a double thymidine protocol followed by a release into a fresh growth medium for 7.5 h.

SDS-PAGE and Immunoblotting
Low and mid-range molecular weight proteins were resolved in 4-12% NuPAGE Novex Bis-Tris Mini Gels (Invitrogen, Carlsbad, CA, USA) at 200 V in NuPAGE MOPS electrophoresis buffer (Invitrogen, Carlsbad, CA, USA) and blotted onto 0.2 µm pore size nitrocellulose membrane (Invitrogen, Carlsbad, CA, USA) at 50 V for 2 h at RT in 10 mM NaHCO 3 , 3 mM Na 2 CO 3 , 0.01% SDS (w:v), and 20% methanol (v:v) transfer buffer. High molecular weight proteins were separated in 3-8% NuPAGE Novex Tris-acetate Gels (Invitrogen, Carlsbad, CA, USA) in Tris-acetate electrophoresis buffer (Invitrogen, Carlsbad, CA, USA) at 150 V, and transferred onto a nitrocellulose membrane (0.45 µm, GE Healthcare, Chicago, IL, USA) by 10 mM NaHCO 3 , 3 mM Na 2 CO 3 , 0.01% SDS, and 15% methanol transfer buffer at 35 V and 4 • C, overnight. After transfer, the membranes were rinsed with distilled water and treated with Western Blot Signal Enhancer (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer's instructions. Non-specific binding was blocked for 90 min with either 0.1% casein (Roche, Penzberg, Germany) or 5% BSA (Sigma-Aldrich, Munich, Germany) in TBS (Tris-buffered saline, 20 mM Tris, 150 mM NaCl) with 0.1% Tween 20. The membranes were then incubated with the appropriate primary antibodies (Table 2) in blocking buffer overnight at 4 • C. After washing with TBST, the membranes were exposed to HRP (horseradish peroxidase)-coupled secondary antibody (Table 2) in blocking buffer for 45 min. Proteins were detected by incubating with a chemiluminescent HRP substrate (Millipore, Billerica, MA, USA) for 5 min and protected from light. The signal was detected with a ChemiDoc XRS system (Bio-Rad Laboratories, Feldkirchen, Germany) controlled by the Quantity One 1-D Analysis software v4.6.9 (Bio-Rad Laboratories, Hercules, CA, USA). Ten images were acquired between 1 s and 10 min of exposure or until saturation was reached. In some cases, the membranes were stripped using Restore Plus Western Blot stripping buffer (Thermo Scientific, Rockford, IL, USA) for 30 min and reused after blocking if no signal was detected in a 5 min exposure after incubating with HRP chemiluminescence substrate. Changes in expression of the protein of interest were estimated by densitometry analysis using the ImageLab software version 4.1 (Bio-Rad Laboratories, Hercules, CA, USA) using the intensity profile of each lane and normalizing the volume intensity of the bands to the corresponding value for actin.

Immunoprecipitation
The scarce expression of Kv10.1 in HeLa and hTERT RPE1 cells made it necessary that the efficiency of knockdown was tested by immunoprecipitation with an anti-Kv10.1 monoclonal antibody (Kv10.1-33.mAb), and recovery with protein G magnetic beads (New England BioLabs, Frankfurt am Main, Germany) following the manufacturer's instructions ( Figure S3). The beads were washed three times with 500 µL of cold immunoprecipitation buffer (0.1% Triton X-100, 50 mM Tris-HCl, 300 mM NaCl, 5 mM EDTA and protease inhibitor cocktail) and Kv10.1 was eluted at 70 • C for 10 min in NuPAGE LDS Sample buffer (1:4, v:v).

Single-Cell Gel Electrophoresis
DNA lesions were detected by alkaline single-cell gel electrophoresis (comet assay) [105]. Cells were plated in a six-well plate, transfected with siRNA one day later and assayed after 48 h. All the steps were carried out under dim light and at 4 • C. The assay was performed in clean CometSlide 2 Well (Trevigen, Gaithersburg, MD, USA) slides coated with 0.5% normal melting point agarose, on which 15,000 cells/well in 1% low melting point agarose were layered. The slides were incubated in lysis buffer (2.5 M NaCl; 0.1 M EDTA; 10 mM Tris, 10% DMSO, v:v; 1% Triton X-100, v:v; pH, 10) for 1 h, placed in cold electrophoresis buffer (1 mM Na 2 EDTA; 300 mM NaOH; pH > 13) and denatured for 20 min. The electrophoresis was carried out for 20 min at 13 V (0.6 V/cm across the platform) and 300 mA. The slides were neutralized (0.4 M Tris; pH, 7.5) for 10 min and then fixed in ethanol, air-dried overnight, and stained with SYBR Gold (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. Ten positions in each well of the still-wet slides were imaged on a Nikon TiE-Andor spinning disk microscope (Oxford Instruments, Abingdon, UK) equipped with a PLAN APO VC Air 20× 0.75 NA objective (Nikon Instruments, Amsterdam, Netherlands). Images were acquired with an Andor DU-888 X-9984 camera controlled with the NIS-Elements AR software v.4.50 (Nikon Instruments, Amsterdam, Netherlands). Maximum intensity projections of z-stack images with 0.2 µm z-step size were analyzed using CASPLab open-source software [106].

Immunofluorescence
To visualize mitotic spindles and kinetochores, cells seeded on coverslips were fixed with 2% p-formaldehyde, 60 mM PIPES, 27 mM HEPES, 10 mM EGTA, 4 mM MgSO 4 , pH, 7.0 for 5 min at RT, followed by methanol at −20 • C for 5 min. After blocking for 90 min in 10% BSA in DPBST (0.1% Tween-20), the cells were incubated for 2 h at RT with anti-α-tubulin, anti-CENP-B (centromere protein B), or anti-γ-tubulin (see Table 3 for dilutions and specifications), followed by a one-hour incubation with the secondary antibodies conjugated to Alexa Fluor-488, -546 or -633, respectively. Nuclei were counterstained with Hoechst 33342 for 5 min at RT, washed three times with DPBST for 5 min at RT, and mounted using Prolong Gold antifade reagent (Invitrogen). The slides were imaged on a Leica TCS SP5 II confocal laser scanning microscope (Leica Microsystems, Wetzlar, Germany) equipped with HCX PL APO Oil 40× 1.25 NA objective. Images were acquired with a z-optical spacing of 0.2 µm using LAS AF (Leica Microsystems, Wetzlar, Germany) software, and processed with the image analysis software ImageJ2 [107][108][109] (U. S. National Institutes of Health, Bethesda, MD, USA) and Imaris (Bitplane, Oxford Instruments, Belfast, UK).

Determination of Fan-Shaped Mitotic Spindles
The occurrence of asymmetric (fan-shaped) mitotic spindles was determined in HeLa cells synchronized in prometaphase with DME 48 h post-siRNA transfection. Defective mitotic spindles were analyzed by immunofluorescence microscopy detecting microtubules and DNA. The occurrence of fan-shaped mitotic spindles was defined as a percentage of the defective spindles from the total number of monopolar mitotic spindles.

Determination of Lagging Chromosomes and Defective Metaphases
To detect lagging chromosomes in anaphase, HeLa cells were transfected with Scror KCNH1-siRNAs, and the synchronization with thymidine-nocodazole was started 4 h later. Cells were visualized by immunofluorescence microscopy detecting CENP-B-positive chromosomes.
Only chromosomes that were both CENP-B-positive and separated from the two pole-oriented chromosome masses were counted as "lagging chromosomes".
The defective metaphases were defined in the cells synchronized at metaphase and stained for α-tubulin, CENP-B, Hoechst 33342. Using Hoechst 33342 and CENP-B staining, the alignment of chromosomes was analyzed.

Determination of Mitotic Spindle Angle
Mitotic spindle angle and the distance between the poles were measured using Imaris v9.0 in HeLa cells synchronized at metaphase and stained for γand α-tubulin. Two spheres with a diameter of 1.5 µm around the maximum intensity of γ-tubulin were generated at the centrosomes. A third sphere was positioned as a projection of one of the centrosomes at the z-level of the opposite centrosome.

Live Cell Imaging
All time-lapse imaging experiments were performed on a Nikon TiE-Andor spinning disk confocal microscope equipped with a PLAN APO-TIRF 100× 1.49 NA oil immersion objective. Images were acquired with an Andor DU-888 X-9984 camera controlled with NIS-Elements software.

Assessment of Microtubule Dynamics
Microtubule plus-end assembly rates were determined by tracking EB3-tdTomato in living cells. HeLa and hTERT RPE1 cells were transfected with pEB3-tdTomato alone or in combination with siRNAs. 48 h later, the cells were treated with 2.5 µM DME in a phenol-red free medium for 3.5 h, and transferred to the environmental control chamber of the microscope (37 • C and 5% CO 2 ) and further incubated until a total incubation time of 4 h. Throughout the entire experiment, the temperature was maintained constant at 37 • C. Images were acquired with 500 ms intervals for 1 min. EB3-tdTomato was excited using the 561 nm laser line, and the emission was collected with a 600/52 emission filter. The MT dynamics were analyzed with U-Track 2.2.0 following the authors' guidelines [110][111][112].

Intracellular Calcium Imaging in HeLa Cells
Fluctuations in cytosolic [Ca 2+ ] were monitored in HeLa cells synchronized at the G2/M phase using Fluo-4AM (Invitrogen, Carlsbad, CA, USA) [113]. After 6.5 h of release from the second block with thymidine, the cells were incubated with 10 µM Fluo-4AM in phenol-red free DMEM medium for 50 min, and then the medium was replaced, and image acquisition was performed at 37 • C and 5% CO 2 . Image acquisition started at 60 min after the addition of Fluo4-AM at five positions per well for 2 h every 10 s.
In each recorded image, ten cells were manually selected as regions of interest (ROIs; in total, 50 cells per condition), and three background ROIs of constant size were identified. The intensity values in each ROI were measured over time with the NIS-Elements software. The corrected total cell fluorescence (CTCF) for each ROI and time point was calculated as a difference of the cell integrated intensity and the product of multiplication of the area of the selected cell and the mean background fluorescence. The generated CTCF values were further analyzed using MATLAB. In the time domain, a linear function was fitted to the collected data to compensate for the baseline drift resulting from photobleaching. The baseline was then subtracted from the measured values, generating normalized CTCFs. Additionally, to make the collected data comparable between experiments, the values were expressed as a percent change to the respective time-point "0". The area under the curve (AUC), prominence, and full width at half maximum (FWHM) were calculated using trapz and findpeaks functions. Further, the time domain was transformed into a frequency domain by applying a Fast Fourier Transformation. The two-sided, and then the single-sided spectra were computed. The frequencies were sorted after power spectrum density (PSD) in descending order, and the first 15 values were used to calculate the weighted average frequency, defined as a mean of the product of frequency and the respective PSD.

Statistical Analysis
All data were analyzed using Prism software v8.4.3 (GraphPad Software LLC, San Diego, CA, USA). The analysis used for each experiment is indicated in the corresponding Figure legend. Data are shown as the median with interquartile range or the mean with standard error (SEM). p-value < 0.05 was considered statistically significant.

Discussion
Our current understanding of the Kv10.1 implication in the G2/M phase involves both canonical (ion-channel activity) and non-canonical (massive intracellular domains and the splice variants present in the cell) mechanisms ( Figure 6). On the one hand, Kv10.1/ORAI1 tandem triggers changes in intracellular Ca 2+ behavior. The fluctuations in [Ca 2+ ] can directly affect the MT growth and change the activity of many of the proteins, which can translate the information encoded in the frequency and amplitude of Ca 2+ oscillations [94]. Considering the centrosomal localization of Kv10.1 [21], the decoding would take place at the centrosomes-a hub for a myriad of cell cycle regulators, including AURKA (Aurora kinase A) and PLK1 [114]. Ca 2+ /CaM transiently activates both these kinases during the G2/M transition [115,116]. It is, therefore, appealing to speculate that AURKA and PLK1 might react to Kv10.1/ORAI1 driven changes in cytosolic Ca 2+ oscillations. On the other hand, the physical presence of Kv10.1 both at the centrosomes and the split-forms in the cytosol hold a high potential of interacting and controlling the activity of many enzymes involved in mitosis. It is tempting to speculate that Kv10.1 can interfere with the activation of CDK1, AURKA, and PLK1 independently of its conductance. As already mentioned, the spliced form of the Some proteins sense such oscillatory changes in cytosolic Ca 2+ . AURKA and PLK1 hold a high potential for such sensory proteins: both kinases are activated upon binding of Ca 2+ /CaM. Additionally, Kv10.1 split variant E65 directly activates CDK1 (low-left panel). As a result of the described changes, the MT assembly rate is modulated. Therefore, Kv10.1, through its canonic ion-channel function and protein-protein interactions, regulates the proper entry and progression through M phase (low-right panel). (K + -potassium ions; Ca 2+ -calcium ions; CaM-calmodulin; E65-split-form of Kv10.1; CDK1-cyclin-dependent kinase 1; PLK1-Polo-like kinase 1; AURKA-Aurora kinase A; MAPs-microtubule-associated proteins).
On the other hand, the physical presence of Kv10.1 both at the centrosomes and the split-forms in the cytosol hold a high potential of interacting and controlling the activity of many enzymes involved in mitosis. It is tempting to speculate that Kv10.1 can interfere with the activation of CDK1, AURKA, and PLK1 independently of its conductance. As already mentioned, the spliced form of the channel E65 activates CDK1 in Xenopus oocytes [97]. Interestingly, a positive feedback loop exists between AURKA, CDK1, and PLK1 [117][118][119][120][121]. All these three kinases are implicated in microtubule growth and SAC regulation [122][123][124]. Therefore, any imbalance in this system during mitosis can affect MT dynamicity, the positioning, or even formation of the mitotic spindle. In either way, SAC response will be altered, delaying, or speeding up the metaphase-to-anaphase transition.

Conclusions
The tumor specificity and ease of access of Kv10.1 make it an attractive candidate for cancer treatment. However, the clinical development of Kv10.1-directed therapies has faced difficulties due to the incomplete understanding of the underlying molecular mechanisms. Here, we attempted to bridge that gap and broaden our very limited knowledge on the implication of ion channels in cell growth. Furthermore, understanding the molecular mechanisms causing the different effects of functional inhibition or gene-silencing mediated downregulation of Kv10.1 can help to tailor cancer treatment to tumor phenotype. Therefore, although the current work in cell lines cannot be directly translated to the clinical scenario, we are persuaded that Kv10.1 will be a relevant oncological target in the future.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6694/12/9/2409/s1. Figure S1: Kv10.1 downregulation activates DNA damage-induced checkpoint in the absence of DNA damage in HeLa cells. Figure S2: Kv10.1 downregulation leads to increased microtubule plus end growth rates in hTERT RPE1 cells. Figure S3: Efficacy of siRNA-mediated downregulation of Kv10.1 in HeLa cells. Videos S1-S16: https://zenodo.org/record/3951101#.XxXVqS9h1R0. Videos S1-S10: Time-lapse images of a mitotic HeLa cell expressing EB3-tdTomato synchronized in prophase with DME-protocol and analyzed with U-Track software are given (duration, 1 min; interval, 500 ms; growth tracks are marked in red). Videos S1 and S2: A representative time-lapse image of a cell transfected with Scror KCNH1-siRNA (Kv10.1 downregulation). Videos S3 and S4: A representative time-lapse image of a cell transfected with pEmpty (plasmid backbone only) or pKCNH1-WT (wild-type Kv10.1 overexpression). Videos S5 and S6: A representative time-lapse image of a cell treated with DMSO (dimethyl sulfoxide) or AnCoA4 (ORAI1 conductance block). Videos S7 and S8: A representative time-lapse image of a cell treated with IgG (immunoglobulin G) or mAb56 (Kv10.1 conductance block). Videos S9 and S10: A representative time-lapse image of a cell treated with mAb56 + DMSO or mAb56 + AnCoA4 (simultaneous block of Kv10.1 and ORAI1). Videos S11-S16: Time-lapse images of HeLa cells synchronized at G2/M with double pulse thymidine protocol, preloaded with an intracellular calcium indicator Fluo-4AM are given (duration, from 10 min to 120 min after the start of the imaging; interval, 5 s; "Red Fire" LUT is applied showing the changes in Fluo-4 intensity). Videos S11 and S12: A representative time-lapse image of a cell treated with DMSO or AnCoA4 (ORAI1 conductance block). Videos S13 and S14: A representative time-lapse image of a cell treated with IgG or mAb56 (Kv10.1 conductance block). Videos S15 and S16: A representative time-lapse image of a cell transfected with Scror KCNH1-siRNA (Kv10.1 downregulation).