BioID-based intact cell interactome of the Kv1.3 potassium channel identifies a Kv1.3-STAT3-p53 cellular signaling pathway

Abstract

Kv1.3 is a multifunctional potassium channel implicated in multiple pathologies, including cancer. However, how it is involved in disease progression is not fully clear. We interrogated the interactome of Kv1.3 in intact cells using BioID proximity labeling, revealing that Kv1.3 interacts with STAT3- and p53-linked pathways. To prove the relevance of Kv1.3 and of its interactome in the context of tumorigenesis, we generated stable melanoma clones, in which ablation of Kv1.3 remodeled gene expression, reduced proliferation and colony formation, yielded fourfold smaller tumors, and decreased metastasis in vivo in comparison to WT cells. Kv1.3 deletion or pharmacological inhibition of mitochondrial Kv1.3 increased mitochondrial Reactive Oxygen Species release, decreased STAT3 phosphorylation, stabilized the p53 tumor suppressor, promoted metabolic switch, and altered the expression of several BioID-identified Kv1.3-networking proteins in tumor tissues. Collectively, our work revealed the tumor-promoting Kv1.3-interactome landscape, thus opening the way to target Kv1.3 not only as an ion-conducting entity but also as a signaling hub.

Fig. 1.. Characterization of cells stably expressing myc-BirA*-Kv1.3 fusion proteins.

(A) Cartoon of myc-BirA*-Kv1.3. BirA* is shown fused to the Kv1.3 N-terminal end. BirA* converts exogenously added biotin to highly reactive labile biotinyl-5′-AMP, allowing it to react with primary amines on vicinal/interacting proteins. The labeling radius is around 10 nm. (B) HEK293 cells were stably transfected with myc-BirA*-Kv1.3 or myc-BirA*. Following SDS-PAGE of cell lysates, myc-BirA*-Kv1.3 or myc-BirA* proteins were detected with anti-BirA. Fifty micrograms of protein per lane was loaded as indicated: parental HEK293 cells (control, untransfected cells); HEK293 cells transfected with the empty vector (myc-BirA* vector); two independent clones expressing myc-BirA*-Kv1.3. Vinculin: loading control. (C) Immunofluorescence analyses of HEK293 cells stably expressing myc-BirA*-Kv1.3. Localization of BirA*-fusion proteins detected at the PM with fluorescently labeled anti-myc (green). The PM PMCA (red) colocalize with the Kv1.3 fusion protein (see merge). DNA is labeled with DAPI (blue). Scale bar, 25 μm. A representative image of three independent experiments is shown. (D) Kv1.3 currents elicited by applying a train of depolarizing potentials in 20 mV stepwise increases up to +70 mV from a −70 mV holding potential (n = 3 for each). (E) Confocal images of HEK293 cells stably expressing myc-BirA*-Kv1.3 (clone #1) after 24 hours of incubation with exogenous biotin (50 μM). Myc-BirA*-Kv1.3 was detected with anti-myc (red) and biotinylated proteins with fluorescently labeled streptavidin (green). Scale bar, 25 μm. (F) HEK293 cells stably expressing myc-BirA*-Kv1.3, or parental and empty vector controls, were analyzed 24 hours after biotinylation induction with (+) or without (−) exogenous biotin supplementation. Biotinylated proteins were detected with streptavidin-HRP. Middle: the Kv1.3 fusion protein detected with anti-BirA. Bottom: Anti-actin was protein loading control (n = 3 independent experiments). (G) As in (F) in the presence of 50 μM biotin at different time points (n = 3). Loading control of the same blot is shown.

Fig. 2.. Pathway analysis of the interacting proteins.

(A) Treemap of the significantly enriched Reactome pathways. Each rectangle is a significant GO term that is joined into clusters of semantically related terms, visualized with different colors. The size of the rectangles reflects the P value of the GO term in cluster: Larger rectangles indicate higher significance. The genes belonging to the clusters highlighted with red circles were used to draw the schematic representation of the Kv1.3 interactome reported in the right part. (B and C) Dot plots reporting the top 20 significantly enriched Reactome pathways (B) and the 16 most significant KEGG pathways (C) obtained using the 527 hit proteins. On the x axis, the ratio between the number of hit proteins belonging to the pathway (count) and the total size of the pathway is reported. The dimension of the dot indicates the count (the larger the dot, the higher the count), and the color of the dot represents the adjusted P value. (D) Cluster analyses of hit proteins obtained using log2 FC of the plus/minus samples. (E) Dot plots reporting the top 20 significantly enriched GO terms for molecular function. The dimension of the dot represents the odds ratio and the color denotes the adjusted P value.

Fig. 3.. Generation and characterization of Kv1.3 KD B16 F10 cells.

(A) Left: PCR performed with the primers F1 and R2b on the B16F10 clones: the expected band for the WT is 2709 bp and the deleted band is expected to be approximately 1000 bp. The positive control was the B16F10 WT, and the negative control was water. Right: Western blot of WT and two independent KD lines (n = 3). Protein products with slightly different MWs are observable and are due to glycosylation [e.g., (95)]. (B) Proliferation rate of B16F10 cells and the Kv1.3 KD clones (n = 4, two-way ANOVA, P < 0.05). Parental B16F10 and Cas9-transfected cells displayed the same proliferation rate (not shown, n = 3). (C) Colony formation assay (n = 6 for each line) (one-way ANOVA, P < 0.05). The number and the dimension of colonies were measured with ImageJ (D) Representative confocal images of WT and Kv1.3 KD #1 and #2 B16F10 cells showing a fragmented mitochondrial network. Cells were stained for mitochondrial marker TOM-20 (magenta) and DAPI (cyan). Scale bars, 5 μm. (E) Mitochondrial ROS release was measured using Mitosox and FACS analysis. A representative experiment is shown on the left, while the right panel reports values normalized on WT cells’ Mitosox fluorescence signal. (F) Wound healing assay. Quantification as % of initial gap area (n = 4) (two-way ANOVA, P < 0.05). Images were analyzed with ImageJ and applied.

Fig. 4.. Kv1.3 expression affects STAT3 signaling.

(A) Representative confocal images of HEK cells and Kv1.3 KD #1 and #2 B16F10 cells transfected with either Kv1.3-YFP WT or PD Kv1.3-YFP (green), in which proximity ligation assay (PLA) was performed (red). Cells were also stained with DAPI (blue). (B) Kv1.3 coimmunoprecipitates with STAT3 in B16F10 KD #1 expressing Kv1.3-YFP. Immunoprecipitation for Kv1.3-YFP and STAT3, immunoblotted for STAT3 and GFP. (C) Analysis of expression and phosphorylation levels of STAT3 (Ser⁷²⁷) and total STAT in B16F10 WT and two B16F10 Kv1.3 KD clones. Analysis of expression and phosphorylation levels of STAT3 (Ser⁷²⁷) and total STAT in SKMEL28 WT (scramble, scr) and KD clones (siRNA-Kv1.3). (D) Analyses of Ser⁷²⁷ STAT3 phosphorylation level and of total STAT3 expression level after treatment with Margatoxin (Mgtx, 0.5 μM) and Stichodactyla toxin (Shk, 0.5 μM) and the mitochondria targeted Kv1.3 inhibitor PAPTP (0.5 and 1 μM) in B16F10 cells. Vinculin was used as the loading control. (E) Analyses of Ser⁷²⁷ and Tyr⁷⁰⁵ STAT3 phosphorylation level and of total STAT3 level after treatment with Margatoxin (0.5 μM) and Stichodactyla toxin (0.5 μM) and PAPTP (1 μM) in COLO357 cells (all blots have n = 3; *P < 0.05; **P < 0.001; ***P < 0.0001).

Fig. 5.. The Kv1.3-STAT3-p53 axis.

(A) Analysis of expression and phosphorylation levels of p53 (Ser³⁹²) and total p53 in B16F10 WT and two B16F10 Kv1.3 KD clones. (B) Analysis of p53 expression level after treatment with Margatoxin (Mgtx, 0.5 μM) and Stichodactyla toxin (Shk, 0.5 μM) and the mitochondria targeted Kv1.3 inhibitor PAPTP (0.5 and 1 μM) in B16F10 cells. Vinculin was used as the loading control (30 μg of protein from cell lysates per lane, n = 3). (C) Analysis of expression and phosphorylation levels of p53 (Ser³⁹²) and total p53 in B16F10 WT treated with PAPTP (0.5 μM) and mitoTEMPO (mT, 100 nM). (D) Analysis of expression and phosphorylation of p53 and STAT3 levels in HEK293 cells expressing Kv1.3-YFP. (E) Analysis of expression and phosphorylation of p53 and STAT3 levels in B16F10 KD cells expressing Kv1.3-YFP. For (D) and (E), see description of the procedure in Materials and Methods. (F) Analysis of expression and phosphorylation of p53 levels in B16F10 silenced for STAT3 (n = 3 for all blots; *P < 0.05; **P < 0.001; ***P < 0.0001).

Fig. 6.. Gene expression analysis in melanoma cells lacking Kv1.3.

(A) DEG analysis was performed comparing Kv1.3 KD and WT lines. A total of 2938 and 3524 differentially expressed genes (DEG) compared to WT were found for B16F10 KD Kv1.3 #1 and #2, respectively (FDR < 0.01). Top100 DEGs (with highest SD among all nine samples) are shown for all three replicates per cell line. (B) Transcript level of several p53-induced target genes and cell cycle progression regulatory genes. The color key indicates the log2-fold change relative to control in (B) to (D) and (F). (C) Transcript level of known pro- and anti-apoptotic players. (D) Transcript level of glycolytic genes. (E) Measurement of glycolytic capacity, i.e., the rate of increase in ECAR after the injection of oligomycin following glucose addition. Oligomycin inhibits mitochondrial ATP production and therefore shifts the energy production to glycolysis with increase in ECAR revealing maximum glycolytic capacity of the cells. The glycolytic reserve is the difference between glycolytic capacity and glycolysis rate (n = 3 independent experiments). (F) Transcript level of genes involved in purine and pyrimidine metabolism. (G) Heatmap of the transcript levels of 81 DEGs that encode protein partners of Kv1.3 identified by BioID. The shown 81 hits are shared by both clones.

Fig. 7.. In vivo tumor growth and metastatic spread in an orthotopic model using WT B16F10 cells and two independent Kv1.3 KD lines.

(A) Schematic view of the in vivo experiments. (B) Six animals from each group were injected with the indicated cell lines. The tumors were excised and are shown in the representative image. Tumor weight and volume are shown. (C) Representative photographs of lung metastasis removed after 21 days from the tail vein injection with the indicated cell lines in 9-week-old C57BL6/J mice. Quantification of metastases on the surface of the lung is shown. Data represent mean ± SD. n = 5 WT, n = 5 KD #1, n = 4 KD #2.*P < 0.05. (D) Phosphorylation array on tumor lysates reveals decreased tyrosine phosphorylation of the indicated proteins in the Kv1.3 KD tumors with respect to WT. Quantification of the dot intensity is shown. (E to J) SDS-PAGE followed by immunoblotting of B16F10 WT and two Kv1.3 KD clone-derived tumor lysates developed with the indicated antibodies. Quantification was performed on biological (n = 2 to 3 for each) and technical replicates (n ≥ 3 for each).

Fig. 8.. Cartoon of the proposed mechanisms accounting for reduced tumor growth.

Kv1.3 is required for tumor growth as it interacts with STAT3 and promotes its phosphorylation. In addition, lack or inhibition of the mitochondrial Kv1.3 leads to enhanced mitochondrial ROS release and to phosphorylation of p53 at Ser³⁹² that promotes its migration to mitochondria and its proapoptotic function. p53 stabilization affects cell cycle, apoptosis, and metabolism. See the Results and Discussion sections for explanation and discussion of further Kv1.3-linked pathways that may contribute to reduced tumor growth in the absence of the channel.