Ubiquitin-Related Modifier 1 (URM-1) Modulates Cx43 in Breast Cancer Cell Lines

Gap-junction-forming connexins are exquisitely regulated by post-translational modifications (PTMs). In particular, the PTM of connexin 43 (Cx43), a tumor suppressor protein, regulates its turnover and activity. Here, we investigated the interaction of Cx43 with the ubiquitin-related modifier 1 (URM-1) protein and its impact on tumor progression in two breast cancer cell lines, highly metastatic triple-negative MDA-MB-231 and luminal breast cancer MCF-7 cell lines. To evaluate the subsequent modulation of Cx43 levels, URM-1 was downregulated in these cells. The transcriptional levels of epithelial-to-mesenchymal transition (EMT) markers and the metastatic phenotype were assessed. We demonstrated that Cx43 co-localizes and interacts with URM-1, and URMylated Cx43 was accentuated upon cellular stress. The significant upregulation of small ubiquitin-like modifier-1 (SUMO-1) was also observed. In cells with downregulated URM-1, Cx43 expression significantly decreased, and SUMOylation by SUMO-1 was affected. The concomitant expression of EMT markers increased, leading to increased proliferation, migration, and invasion potential. Inversely, the upregulation of URM-1 increased Cx43 expression and reversed EMT-induced processes, underpinning a role for this PTM in the observed phenotypes. This study proposes that the URMylation of Cx43 in breast cancer cells regulates its tumor suppression properties and contributes to breast cancer cell malignancy.


Introduction
Connexins are integral membrane proteins that form gap junctions, enabling the direct cytoplasmic exchange of ions and low-molecular-weight metabolites between adjacent cells [1]. Connexins facilitate direct intercellular communication, which is a pivotal feature for the development, function, and homeostasis of tissues and organs [2,3]. An increasing number of gap-junction-independent functions are also attributed to these proteins [4]. Connexins are extensively regulated at the transcriptional, post-transcriptional, and posttranslational levels, leading to the modulation of their trafficking, stability, and activity [5]. Most connexins harbor multiple consensus sites for phosphorylation, S-nitrosylation, ubiquitylation, and SUMOylation. Among connexin proteins, Cx43 is the most widely studied connexin. We have previously demonstrated the involvement of Cx43 in the metastatic potential of triple-negative breast cancer [6][7][8][9][10] and in inflammation [6,11]. PTMs play To further confirm the PLA data, immunofluorescence and co-immunoprecipitation assays of Cx43 and URM-1 were conducted. For that purpose, MDA-Cx43D cells were exposed to the tumor-promoting phorbol ester (PMA), which, upon Protein Kinase C activation, regulates the trafficking, assembly, degradation, and channel gating of Cx43 gap junctions [29][30][31]. PMA is established to induce a fast and transient block of Cx43 gap junctional communication in many cell types [32][33][34][35]. This is often followed by the accelerated endocytosis and degradation of Cx43 gap junctions [36,37]. PMA also blocks this assembly of gap junctions [38]. The immunofluorescence assay revealed that the treatment of MDA-Cx43D with PMA affects Cx43 localization and induces the formation of annular junctions as early as 2 h, and this effect was sustained 24 h post-PMA treatment ( Figure 1D). Moreover, co-immunoprecipitation, 2 or 24 h post-PMA induction, indicates that Cx43 URMylation and SUMOylation by SUMO-1 occur as early as 2 h and increase upon exposure to PMA ( Figure 1E). This result demonstrates that Cx43 internalization is accompanied by PTMs. MCF-7 cells were also exposed to PMA and analyzed for Cx43 and URM-1 levels. Western blot analysis showed decreased Cx43 levels 3 h post-PMA treatment and increased URM-1 levels within 1 h of PMA treatment, which subsequently decreased (Supplementary Figure S1B). Collectively, these results indicate that URM-1 can interact with and modify Cx43, similar to other ubiquitin-like small protein modifiers. To further confirm the PLA data, immunofluorescence and co-immunoprecipitation assays of Cx43 and URM-1 were conducted. For that purpose, MDA-Cx43D cells were exposed to the tumor-promoting phorbol ester (PMA), which, upon Protein Kinase C ac
increased URM-1 levels within 1 h of PMA treatment, which subsequently d (Supplementary Figure S1B). Collectively, these results indicate that URM-1 can with and modify Cx43, similar to other ubiquitin-like small protein modifiers.

Loss of URM-1 Promotes EMT in Breast Cancer Cell Lines
Given the tumor suppressor role of Cx43 in breast cancer and the apparent modulation of Cx43 by URM-1 levels in breast cancer cell lines, cells were evaluated for changes in the levels of EMT markers upon URM-1 downregulation. The transcriptional levels of Snail, VEGF, MMP-2, MMP-9, N-cadherin, and E-cadherin were evaluated by qPCR. Gene expression analysis showed that MDA-Cx43D shURM cells expressed significantly higher levels of VEGF and MMP-2 (p < 0.01, Figure 3A). The levels of Snail, VEGF, and MMP-2 were also significantly increased in MCF-7 shURM cells (p < 0.05, p < 0.05, and p < 0.01, respectively, Figure 3B). Both MDA-Cx43D shURM and MCF-7 shURM cells expressed higher levels of MMP-9 and the mesenchymal marker N-cadherin and did not change or decreased the level of the epithelial marker E-cadherin, but these changes did not reach the statistical significance threshold ( Figure 3A,B). Functionally, these findings were evaluated using a real-time cell analyzer (RTCA). The proliferation, migration, and invasion potential increased among cells downregulating URM-1, as reflected by the increased cell index. In particular, the loss of URM-1 rendered MDA-Cx43D shURM cells significantly more invasive (p < 0.05, Figure 3C) and significantly increased the migration and invasion capacity of MCF-7 shURM cells (p < 0.05, Figure 3D). URM-1 upregulation in MDA-MB-231 cells was reflected by a parallel increase in the Cx43 expression level in MDA-MB-231 URM cells (p < 0.001, Figure 3E). The evaluation of EMT markers in these cells by qPCR showed that upon the upregulation of URM-1, the transcript levels of Snail, Twist, and VEGF (p < 0.01) and of N-cadherin and MMP-9 (p < 0.05) significantly decreased ( Figure 3E). In addition, the mRNA levels of the epithelial marker E-cadherin greatly increased (p < 0.01) ( Figure 3E). These changes translated to morphological alterations; light microscopy images show that MDA-MB-231 cells upregulating URM-1 display a less mesenchymal phenotype (fibroblastic in appearance and present as spindle-shaped flattened cells) and a more epithelial phenotype (monolayers with sheet-like appearance) ( Figure 3F).
To determine the implication of Cx43 URMylation on the functionality of connexins in forming functional gap junctions between breast cancer cells, two modalities were adopted: the in vitro wound healing assay and dye transfer. In the first modality, we examined cell migration in response to a mechanical scratch wound in the presence or absence of URMylated Cx43. Images of scratch areas at time points of 0 and 48 h are illustrated in Figure 4A. While the scratch was completely closed after 48 h in MDA-MB-231, the overexpression of Cx43 in these cells induced the partial recovery of the scratch, implicating Cx43 in the migration capacity of MDA-MB-231. Interestingly, URM downregulation in MDA-Cx43D restored the migration capacity of these cells, and the scratch completely closed, as in the parental MDA-MD-231 at 48 h ( Figure 4A). In MCF-7 cells, the recovery of the scratch was partial, but the downregulation of URM significantly increased wound recovery after 48 h (p < 0.05) ( Figure 4A). In the dye transfer assay, calcein-labeled MDA-Cx43D or MDA-Cx43D ShURM cells were co-cultured with unlabeled MDA-Cx43D or MDA-Cx43D ShURM, respectively, at a 1:1 ratio for 1 h. Similarly, labeled MCF-7 or MCF-7 ShURM cells were co-cultured with unlabeled MCF-7; then, dye transfer between the tested cells was evaluated by flow cytometry, and the mean fluorescence intensity (MFI) was quantified. We observed a clear shift in fluorescence intensity in both labeled breast cancer cell lines ( Figure 4B). The downregulation of URM in either cell line reduced or blocked the shift in fluorescence, emphasizing that dye transfer occurred through URMylated Cx43 forming gap junctions ( Figure 4B). Collectively, these data implicate URM-1 in the breast cancer cell malignant phenotype, presumably due to the loss of Cx43 in cells downregulating URM-1. The upregulation of URM-1 seems to, at least partially, reverse EMT in breast cancer cells. Moreover, the URMylation of Cx43 is required for its functionality and the formation of functional gap junctions. cultured with unlabeled MDA-Cx43D or MDA-Cx43D ShURM, respectively, at a 1:1 ratio for 1 h. Similarly, labeled MCF-7 or MCF-7 ShURM cells were co-cultured with unlabeled MCF-7; then, dye transfer between the tested cells was evaluated by flow cytometry, and the mean fluorescence intensity (MFI) was quantified. We observed a clear shift in fluorescence intensity in both labeled breast cancer cell lines ( Figure 4B). The downregulation of URM in either cell line reduced or blocked the shift in fluorescence, emphasizing that dye transfer occurred through URMylated Cx43 forming gap junctions ( Figure 4B). Collectively, these data implicate URM-1 in the breast cancer cell malignant phenotype, presumably due to the loss of Cx43 in cells downregulating URM-1. The upregulation of URM-1 seems to, at least partially, reverse EMT in breast cancer cells. Moreover, the URMylation of Cx43 is required for its functionality and the formation of functional gap junctions.

Discussion
Post-translational modifications modulate protein expression, assembly, and trafficking and determine their definitive functions [39]. The connexin family of proteins, including Cx43, is post-translationally modified, leading to the opening or closing of the channel or insertion into specialized lipid rafts, which induces dramatic functional changes [40]. Since Cx43 is modified by ubiquitylation [41,42] and SUMOyaltion [43], we explored whether it is also modified by another ubiquitin-like molecule, URM-1, known to rapidly conjugate to multiple target proteins in response to oxidative stress [41][42][43].
After demonstrating the co-localization of Cx43 with Ub, URM-1, and SUMO-1 in MDA-Cx43D and MCF-7 cells and the interaction of Cx43 with each of these PTMs, the effect of PMA-induced stress was investigated in an attempt to determine a sequence of posttranslational modifications of Cx43 or an interplay between ubiquitylation, SUMOylation, and URMylation. MDA-Cx43D cells were exposed to PMA, which induces cellular stress through the PKC-dependent internalization of Cx43 and hence hinders Cx43-mediated intercellular communication [29,30]. Enhanced URMylation by URM-1 was followed by enhanced SUMOylation by SUMO-1 upon PMA treatment, in line with a previous study reporting a substantial increase in SUMO conjugation upon exposure to stress [44]. Collectively, our data suggest that the URMylation of Cx43 by URM-1 may be a prerequisite to SUMOylation by SUMO-1. In fact, a study showed that mutations in the lysine residues of an oncoprotein, Tax, compromise its URMylation by URM-1 [28]. It is conceivable that the URMylation of Cx43 occurs on conserved lysine residues, which requires further investigation.
Cells downregulating URM-1 showed lower transcriptional and translational levels of Cx43. Connexins in general and the loss of Cx43 in particular are closely associated with cancer progression and metastasis [7,8,10,16,21,23]. We have previously shown that upregulating Cx43 in MDA-MB-231 attenuates its malignant phenotype [7]. Since downregulating URM-1 was associated with decreased levels of Cx43, we investigated EMT markers in MDA-Cx43D shURM and MCF-7 shURM cell lines. The levels of VEGF, MMP-2, and MMP-9 were upregulated in these cell lines compared to their Cx43-expressing counterparts. These findings translated to functional alterations, with the loss of URM-1 (associated with the loss of Cx43) enhancing the proliferation, migration, and invasion profiles of breast cancer cells. Importantly, while MCF-7 cells are usually poorly invasive, the loss of URM-1 enhanced their malignant profile. In addition, upregulating URM-1 was associated with the significant and considerable reversal of EMT, inducing MDA-MB-231 cells to acquire a more epithelial phenotype.
Collectively, our study unveils a novel interaction of URM-1, a ubiquitin-related modifier, with Cx43, a widely recognized gap junction protein with a tumor suppressor role in breast cancer, and demonstrates that the loss of URM-1 promotes EMT, while the upregulation of URM-1 drives EMT reversal in breast cancer.

Analysis of Gene Expression
RNA extraction and reverse transcription. Total RNA was extracted with RNeasy ® Plus mini kit (Qiagen, Germany) according to the manufacturer's instructions and served as the template for cDNA synthesis with Random Hexamers and SuperScript ® II Reverse Transcriptase (ThermoFisher Scientific, MA, USA).
Gene expression analysis. Quantitative real-time polymerase chain reaction (qPCR) was performed in a CFX96™ Real-Time PCR Detection System (Bio-Rad, Hercules, California, USA) using the qPCR Sybr green Master Mix. Primer sequences are listed in Table 1. The relative fold change in gene expression was calculated using the ∆∆ Cq method after normalization to the housekeeping gene gapdh [46].
Co-immunoprecipitation. Cells were washed in ice-cold phosphate-buffered saline (PBS) supplemented with 10 mM N-ethylmaleimide (NEM) prior to lysis at pH 8 in 2% SDS and 50 mM Tris. Lysates were then 10-fold diluted in an IP buffer comprising 50 mM Tris-HCl (pH 8), 200 mM NaCl, 0.1 mM EDTA, 0.5% NP-40, 10% glycerol, and a cocktail of protease inhibitors (Roche). Afterward, 1 µg of goat anti-human Cx43 (C-20) was added for 12 h at 4 • C, followed by the addition of protein A-agarose (Sigma, Mannheim, Germany) for 2 h. Beads were washed in IP buffer three times prior to elution of immunoprecipitated proteins. The latter were then solubilized in 2× Laemmli buffer and denatured using β-mercaptoethanol (5%), followed by heating at 95 • C for 10 min. Proteins were then loaded onto a 10% polyacrylamide gel for Western blotting.

Imaging
Cell morphology and confluence were observed, and images were acquired using Cell Observer and Zen software (Zen Lite 2009, Carl Zeiss, Germany). Immunofluorescence micrographs were visualized and imaged using a Laser Scanning Confocal Microscope (LSM 710, Carl Zeiss, Oberkochen, Germany).

Real-Time Cell Analyzer (RTCA) Assay for Migration, Invasion, and Proliferation
To assess the effect of URM-1 loss on migration, invasion, and proliferation of cell lines, quantitative analysis was performed using RTCA (×CELLigence RTCA[A2]DP, Roche, Penzberg, Germany), as previously described [47]. For migration and invasion assays, cells were re-suspended in serum-free media and seeded at a density of 20,000 cells/well of RTCA CIM-plates coated with Matrigel ® for the invasion assay and without Matrigel ® for the migration assay. Cells were allowed to adhere for 24 h. Invasion was monitored by recording cell impedance every 15 min for a minimum of 18 h. For the proliferation assay, cells were seeded in an E-plate at a density of 7000 cells/well with an additional 120 µL of medium containing 10% serum. Proliferation was monitored by recording cell impedance every 60 min for 24 h. Experiments were performed twice in triplicate for the proliferation assay and in duplicate for migration and invasion assays. The RTCA software generated a survival curve that evaluates the cell survival or cell index (CI). CI directly correlates with cell number. Data were expressed as bar graphs of CI % of control.

In vitro Wound Healing Assay
MDA-MD231, MDA-Cx43D, MDA-Cx43D ShURM, MCF-7, or MCF-7 ShURM cells (3 × 10 5 cells/well) were seeded in 24-well plates to grow in a monolayer for 24 h. Then, a sterile 20-200 µL pipette tip was held vertically to scratch a cross in each well. The detached cells were removed by washing with 500 µL of PBS and shaking at 500 rpm for 5 min. Afterward, 500 µL of fresh medium with or without diluted samples was added and incubated for 72 h. Before image acquisition, the plate was washed with 500 µL of pre-warmed PBS and gently shaken for 30 s. Then, the pre-warmed medium or sample was added again, and pictures were taken. The scratch closure was monitored and imaged in 24 h intervals. The analysis of the scratch images was performed using ImageJ. The percentage of open wound area was plotted over time for each concentration. Data are presented as mean ± SD. Three to six replicates were included in the analysis, and an unpaired Student's t-test was performed. Significance was considered at p < 0.05.

Dye Transfer Assays
MCF-7, MCF-7 ShURM, MDA-Cx43D, and MDA-Cx43D ShURM were seeded in 6well plates at a density of 25 × 10 3 per cm 2 for 24 h. Cells were labeled with 1 µM Calcein for 1 h, washed, and incubated with serum-free medium for 30-60 min to allow intercellular esterase to convert non-fluorescent Calcein to a green-fluorescent form. Labeled MCF-7 or MCF-7 ShURM cells were co-cultured with unlabeled MCF-7. Similarly, labeled MDA-Cx43D or MDA-Cx43D cells were co-cultured with unlabeled MDA-Cx43D for 1 h. The co-cultures were incubated in complete media at 37 • C and 5% CO 2 for 1 h. Non-adherent cells were then removed by washing, and the adherent cells were detached by trypsinization and re-suspended in phosphate-buffered saline (PBS) containing 2% formaldehyde to be analyzed by flow cytometry. Dye transfer was evaluated by measuring mean fluorescence intensity (MFI). Mean fluorescence intensities of both the labeled and unlabeled populations were evaluated using dot plots based on fluorescence of unlabeled and calcein-labeled control samples.

Statistical Analysis
Experiments were performed three times, unless specified otherwise in the figure legends. Results of quantitative data are reported as the mean ± standard deviation (SD). Statistical significance was determined using Student's t-test. The p value was determined, and a p value of 0.05 was set as the significance threshold. Microsoft Excel was used for statistical analysis.