The Emerging Role of E3 Ubiquitin Ligase SMURF2 in the Regulation of Transcriptional Co-Repressor KAP1 in Untransformed and Cancer Cells and Tissues

Simple Summary KAP1 plays an essential role in different molecular and cellular processes central to carcinogenesis, disease progression, and treatment response, revealing both tumor promoting and anticancer functions. The mechanisms that control the steady-state levels of KAP1 and its protein abundance are not well known. Our findings show that SMURF2, a ubiquitously-expressed HECT-type E3 ubiquitin ligase with suggested anticancer activities, is capable to directly bind, ubiquitinate, and regulate KAP1 expression levels in non-cancerous and tumor cells and tissues. The data further show that SMURF2 has a significant influence on KAP1 interactome, regulating its protein–protein interactions and functions in a catalytically-dependent manner. These findings reveal SMURF2 as a pivotal regulator of KAP1, laying a foundation for the investigation of the role of the SMURF2–KAP1 axis in carcinogenic processes and therapeutic responses to anticancer treatment. Abstract KAP1 is an essential nuclear factor acting as a scaffold for protein complexes repressing transcription. KAP1 plays fundamental role in normal and cancer cell biology, affecting cell proliferation, DNA damage response, genome integrity maintenance, migration and invasion, as well as anti-viral and immune response. Despite the foregoing, the mechanisms regulating KAP1 cellular abundance are poorly understood. In this study, we identified the E3 ubiquitin ligase SMURF2 as an important regulator of KAP1. We show that SMURF2 directly interacts with KAP1 and ubiquitinates it in vitro and in the cellular environment in a catalytically-dependent manner. Interestingly, while in the examined untransformed cells, SMURF2 mostly exerted a negative impact on KAP1 expression, a phenomenon that was also monitored in certain Smurf2-ablated mouse tissues, in tumor cells SMURF2 stabilized KAP1. This stabilization relied on the unaltered E3 ubiquitin ligase function of SMURF2. Further investigations showed that SMURF2 regulates KAP1 post-translationally, interfering with its proteasomal degradation. The conducted immunohistochemical studies showed that the reciprocal relationship between the expression of SMURF2 and KAP1 also exists in human normal and breast cancer tissues and suggested that this relationship may be disrupted by the carcinogenic process. Finally, through stratifying KAP1 interactome in cells expressing either SMURF2 wild-type or its E3 ligase-dead form, we demonstrate that SMURF2 has a profound impact on KAP1 protein–protein interactions and the associated functions, adding an additional layer in the SMURF2-mediated regulation of KAP1. Cumulatively, these findings uncover SMURF2 as a novel regulator of KAP1, governing its protein expression, interactions, and functions.

Clinical investigations mostly suggested KAP1 as a tumor-promoting factor and showed that its overexpression is associated with more aggressive cancer phenotypes and poor patient survival [19]. These associations were documented in breast [20,21], cervical [22], ovarian [23], prostate [24], gastric, and liver cancer [25,26], as well in glioma patients [27,28]. Interestingly, higher expression levels of KAP1 were associated with better overall survival in patients with early stages of lung cancer [29]. Moreover, germline mutations and somatic inactivation of KAP1/TRIM28 were reported in Wilms tumor, the most common type of renal malignancy in childhood [30,31], suggesting the antitumorigenic role of KAP1 in these cancers. Despite these proceedings and the accumulating evidence pointing to KAP1 as a potential therapeutic target, the mechanisms regulating KAP1 protein abundance mostly remain elusive.
Here, we report SMURF2, the HECT-type E3 ubiquitin ligase and suggested tumor suppressor, as a novel regulator of KAP1. SMURF2 is a ubiquitously-expressed and evolutionary-conserved E3 ligase whose catalytic activity is implicated in the regulation of diverse molecular and cellular processes. These processes include the regulation of chromatin structure and nuclear shape, DDR and genomic integrity maintenance, gene expression, cell replication, migration, invasion, as well as cancer initiation, progression, and therapeutic response [32][33][34][35][36][37][38][39][40][41][42]. The data that we present in this study show that SMURF2 directly binds, ubiquitinates, and regulates the protein abundance of KAP1 in E3 ligase and cell-context-dependent manners, which is also evident in certain mouse and human normal and cancer tissues. Furthermore, we show that SMURF2 has a significant influence on KAP1 interactome, regulating its protein-protein interactions and functions in a catalyticallydependent manner.
Smurf2-ablated (Smurf2KO) and wild-type control C57BL/B6 mice were housed at the SPF animal facility according to the FELASA guidelines and an experimental protocol approved by the Animal Care and Use Committee of Bar-Ilan University (BIU).

Vectors and Constructs
N-terminal FLAG-tagged KAP1 (FLAG-KAP1) was constructed by PCR from GFP-KAP1 construct (a gift from Prof. Shiloh) using the following set of primers: 5 -aaccgaattcgcggcctccgcggcggc-3 (forward primer containing EcoRI site) and 5 -tatagtcgactcaggggccatcaccagggcca-3 (reverse primer containing SalI site). The PCR products were then separated and purified from 1% agarose gel, digested with the restriction enzymes EcoRI and SalI (NEB), and ligated into a pRK2-FLAG vector [33]. SMURF2-expressing constructs were generated as previously described [33,36]. All constructs were sequence verified.

Western Blot and Immunoprecipitation (IP)
For Western blot analysis, whole cell extracts were obtained by cell lysis in RIPA buffer (50 mM Tris-HCl (pH 7.8), 1% Nonidet P40 substitute, 150 mM NaCl, 0.1% SDS, 0.5% w/v sodium deoxycholate), supplemented with protease (Roche, Basel, Switzerland) and phosphatase inhibitors (Sigma-Aldrich, St. Louis, MO, USA) cocktails, with subsequent sample sonication (30% amplitude, 1 min on ice) and centrifugation (14,000 rpm, 15 min, 4 • C). The protein extracts from mouse tissues were prepared by tissue homogenization in RIPA buffer using TissueRuptor (Qiagen, Hilden, Germany) and sample sonication. The samples were then cleared by centrifugation. Protein concentration was determined using the Pierce ™ BCA assay kit (Thermo Fisher Scientific, Waltham, MA, USA). Samples were then resolved in SDS-PAGE and transferred to PVDF membrane (GE Healthcare, Chicago, IL, USA), followed by incubation with the indicated primary antibodies (Table S2) and corresponding horseradish peroxidase-conjugated secondary antibodies (Jackson Im-munoResearch Laboratories, West Grove, PA, USA, 1:10,000). The membranes were then developed using WesternBright ECL HRP substrate (Advansta, San Jose, CA, USA) and visualized in the Syngene G:BOX. Quantification of immunoblots was performed using Gel.Quant.NET, relative to the corresponding loading controls.
For IP and co-IP experiments, cells were lysed in 1% NP40 buffer (10 mM Tris (pH 7.5), 150 mM NaCl, 1% Nonidet P40 substitute, 5 mM EDTA, 10% glycerol). For endogenous co-IP, we used 0.5% NP40 buffer (10 mM Tris (pH 7.5), 150 mM NaCl, 0.5% Nonidet P40 substitute, 5 mM EDTA, 10% glycerol). Both buffers were supplemented with protease and phosphatase inhibitors. Samples were incubated on ice for 30 min, cleared by centrifugation, and protein concentration was assessed using the Pierce ™ BCA assay kit. For IP of endogenous SMURF2, we used anti-SMURF2 antibody (sc-25511, Santa Cruz, Heidelberg, Germany). Same amount of isotype IgG was used as a control. For IP of recombinant proteins, anti-FLAG M2 affinity gel (F2426, Sigma-Aldrich) and anti-MYC antibody (sc-40, Santa Cruz) were used. Immunoprecipitations were conducted overnight at 4 • C (on rotation). The following day, protein-G Sepharose beads (4 Fast Flow, GE Healthcare, Chicago, IL, USA) were added to the samples, following by sample incubation for additional 2 h on rotation at 4 • C. The beads were then thoroughly washed (at least three times) with an ice-cold lysis buffer and protein complexes eluted from the beads using 5× SDS gel-loading buffer (50 mM Tris-HCl (pH 8), 5 mM EDTA, 5% SDS, 50% glycerol, 50 mM DTT, 0.05% w/v bromophenol blue, 6% 2-mercaptoethanol) and boiled. The samples were then analyzed in Western blots.

In Vitro Protein Binding and Ubiquitination Assays
These assays were conducted as previously described [36][37][38], with some adjustments. Briefly, for in vitro binding assay, FLAG-KAP1 (produced using the TNT ® SP6 high-yield wheat germ expression system (L3260, Promega, Madison, WI, USA) was co-incubated with GST-SMURF2 (purified from bacteria). The reaction was performed in binding buffer (50 mM Tris-HCl (pH 7.5), 120 mM NaCl, 2 mM EDTA, and 0.1% Nonidet P40 substitute) at 37 • C for 15 min. GST-SMURF2 was then pulled down from the reaction using Glutathione Sepharose ™ 4B beads (GE Healthcare, Chicago, IL, USA), followed by extensive sample washing with an ice-cold binding buffer. The resulting SMURF2-KAP1 complexes were eluted from beads using 5× SDS gel-loading buffer and detected in immunoblots with anti-SMURF2 and anti-KAP1 antibodies.
For the analysis of FLAG-KAP1 ubiquitination in cells (in cellulo ubiquitination assay), HEK-293T cells were transiently transfected with FLAG-KAP1 and the indicated vectors, including HA-ubiquitin and MYC-SMURF2WT or catalytically-inactive SMURF2 (C716G; MYC-SMURF2CG). Cells were then lysed in RIPA buffer supplemented with the DUB inhibitor NEM (5 mM) or in 1% SDS, followed by immediate sample boiling at 95 • C for 10-15 min. Samples lysed in 1% SDS were then equilibrated with SDS-free RIPA-NEM buffer to reduce SDS concentration to 0.1%. Subsequently, all samples were sonicated (1 min, 30% amplitude) and FLAG-KAP1 pulled down using anti-FLAG M2 affinity resin. Beads were then extensively washed and the ubiquitination of FLAG-KAP1 analyzed by Western blotting with anti-HA antibody, as described above. For detection of the endogenous KAP1 ubiquitination, KAP1 was IPed from HEK-293T cells using anti-KAP1 antibody (ab109545, Abcam, Cambridge, UK), and then probed with anti-ubiquitin antibody-recognizing endogenous ubiquitin (#58395, Cell Signaling, Boston, MA, USA).
For PLA, U2OS cells growing on coverslips were transiently transfected with either MYC-SMURF2 or MYC empty vector. Next day, the PLA staining was performed using the Duolink ™ in situ Red Starter Kit Mouse/Rabbit (DUO92101, Sigma-Aldrich) with anti-MYC (#2278S, Cell Signaling, 1:200) and anti-KAP1 (MA1-2023, Invitrogen/Thermo Fisher, 1:500) antibodies. The fluorescent images were then visualized and captured using a LSM780 inverted confocal microscope (Zeiss, Jena, Germany) fitted with a Plan-Apochromat 63X/1.40 Oil DIC M27 objective. The images were analyzed using ZEN Blue (version 2.3 lite) and ImageJ (NIH) software tools. All comparative images in IF and PLA assays were obtained under identical microscope and camera settings.

KAP1 Interactome Analysis
The immunoprecipitated KAP1 protein complexes (purified through the FLAG-KAP1 pull-down) were eluted using 8 M urea buffer (8 M urea, 20 mM Tris (pH 7.5), 10 mM DTT, 100 mM NaCl) and subjected for mass spectrometry (MS) analysis at the Smoler Protein Research Center (Technion, Israel), where the samples were trypsin-digested and analyzed using LC-MS/MS in Q Exactive ™ Plus mass spectrometer (Thermo Fisher). The resulting peptides were identified by Discoverer ™ software against human proteome using two search algorithms: Sequest (Thermo Fisher) and Mascot (Matrix Sciences, Sherwood, OR, USA). All identified peptides were filtered with high confidence (false discovery rate (FDR) ≤ 1%), top rank, mass accuracy, and a minimum coverage of two peptides. An additional filter of two-fold change of the area under the peak of the peptides (a measure of a protein abundance) was used to enrich KAP1 interactors in the analyzed samples. The resulting KAP1 interactors were subjected to gene ontology (GO) analysis that was carried out using ToppFun suite bioinformatics tool [37,45]. GO terms with Benjamini and Hochberg-adjusted FDR (q-value FDR B&H) < 0.05 were considered as significant. Protein classification was performed using the PANTHER platform, and the protein-protein interaction network Cancers 2022, 14, 1607 6 of 20 reconstructed using the STRING tool [37,46,47]. Only interactors that are connected within the network were considered and k-means clustering method was applied to classify proteins into different categories. The proteomic data were also analyzed using QIAGEN Ingenuity pathway analysis (IPA) (QIAGEN Inc., https://digitalinsights.qiagen.com/IPA, accessed on 7 February 2022) [48]. This analysis was conducted on a total number of KAP1associated proteins identified at FDR ≤1%, with Fisher's exact p-value (−log(p-value)) set on ≥1.3. A negative fold change expression value was considered as downregulated, whereas a positive value was considered upregulated. Positive and negative z-scores were used to predict pathway activation or inhibition, respectively.

Quantitative Real Time PCR (qRT-PCR)
Total RNA was extracted from cells using RNeasy mini kit (Qiagen) according to the manufacturer's instructions. The cDNA was synthesized from total RNA with random primers using the High Capacity cDNA Reverse Transcription Kit (4368814, Applied Biosystems, Waltham, MA, USA). Equal amounts of synthesized cDNA (100 ng/reaction) were subjected to PCR analysis in ViiA ™ 7 Real-Time PCR System (Thermo Fisher) using Fast SYBR™ Green Master Mix (4385612, Applied Biosystems). KAP1 cDNA levels were calculated using 2 −∆∆CT and normalized to GAPDH gene expression. The sequence of primers used for qRT-PCR analysis is shown in Table S3. The primers were calibrated for their specificity to the exon-exon spanning region through the standard curve method.

Statistical Analyses
Two-tailed Student's t-test was used for statistical analysis of the data obtained in IF, PLA, IHC, qRT-PCR, and ubiquitination assays. Data with p-values less than 0.05 were considered as statistically significant.

KAP1 Is a Novel SMURF2 Interactor
In our preliminary studies aimed to characterize the interactome of SMURF2, we noted that KAP1 may be associated with SMURF2 [44]. This association was also suggested by the involvement of these proteins in similar molecular and cellular processes, including gene expression, DDR, and chromatin remodeling. To investigate the possibility that KAP1 is a bona fide interactor of SMURF2 and, potentially, its substrate, we conducted several lines of investigation. First, we co-expressed FLAG-KAP1 and MYC-SMURF2 (wild-type, SMURF2WT, or its catalytically-inactive form, SMURF2CG) in HEK-293T cells, immunoprecipitated FLAG-KAP1 with anti-FLAG resin and analyzed the samples using MS. The analysis showed that KAP1 is associated with both SMURF2WT and SMURF2CG ( Figure 1A). Particularly, we found 29 and 21 SMURF2-specific peptides (in SMURF2WT and SMURF2CG expressing samples, respectively) in a complex with KAP1 versus zero peptides in FLAG-KAP1 and MYC-Empty co-expressing cells ( Figure 1A, upper panel). The total number of KAP1 peptides identified in these samples was comparable: 99 in the MYC-Empty sample, 86 in SMURF2WT, and 94 in the SMURF2CG samples ( Figure 1A, bottom panel). Reciprocal co-IP analysis validated these findings and showed that KAP1 interacts with both the SMURF2WT and CG form, and vice versa ( Figure 1B). The binding detected between endogenous proteins provided further confirming evidence that KAP1 and SMURF2 are interacting partners ( Figure 1C). This interaction was not limited to HEK-293T cells but was also observed in other types of cells. Particularly, using U2OS cells, we demonstrate that SMURF2 and KAP1 co-immunoprecipitate and co-localize in both interphase and mitotic cells ( Figure 1D and Figure S1A-C). The conducted super-resolution STED microscopy analysis showed that these proteins interact within the nanometer scale range ( Figure 1E), suggesting a physical interaction between SMURF2 and KAP1. This finding was further corroborated by the proximity ligation assay that detected a direct interaction between these proteins in the cellular milieu ( Figure 1F). The direct interaction between SMURF2 and KAP1 was also evident in the in vitro binding assay, wherein we used purified GST-SMURF2 and FLAG-KAP1 proteins ( Figure 1G). evidence that KAP1 and SMURF2 are interacting partners ( Figure 1C). This interaction was not limited to HEK-293T cells but was also observed in other types of cells. Particularly, using U2OS cells, we demonstrate that SMURF2 and KAP1 co-immunoprecipitate and co-localize in both interphase and mitotic cells ( Figure 1D and Figure S1A-C). The conducted super-resolution STED microscopy analysis showed that these proteins interact within the nanometer scale range (Figure 1E), suggesting a physical interaction between SMURF2 and KAP1. This finding was further corroborated by the proximity ligation assay that detected a direct interaction between these proteins in the cellular milieu ( Figure 1F). The direct interaction between SMURF2 and KAP1 was also evident in the in vitro binding assay, wherein we used purified GST-SMURF2 and FLAG-KAP1 proteins ( Figure 1G).

KAP1 Is a Direct Ubiquitination Substrate of SMURF2
Subsequently, we proceeded to examine the possibility that SMURF2 operates as an E3 ubiquitin ligase of KAP1. To this end, we first conducted in cellulo ubiquitination assays, involving co-expression of FLAG-KAP1, MYC-SMURF2 (either WT or its inactive form), and HA-ubiquitin in HEK-293T cells. Cells transfected with the corresponding empty vectors served as additional controls. Following transfection, cells were lysed in RIPA buffer supplemented with the DUB inhibitor NEM and sonicated to ensure the complete extraction of KAP1 from the RIPA-resistant fractions. FLAG-KAP1 was then pulled-down, and its ubiquitination analyzed in immunoblots using anti-HA antibody, which specifically recognizes HA-ubiquitin conjugated to FLAG-KAP1 (Figure 2A, lanes 3-5 vs. 1, 2, and 6). The results presented in the figure also showed that in cells co-expressing FLAG-KAP1, HAubiquitin, and MYC-Empty vector, KAP1 predominantly exhibits a monoubiquitination pattern (Figure 2A, lane 3). The addition to the cells of a catalytically-active SMURF2 significantly increased the oligo/polyubiquitination of KAP1 (Figure 2A, lane 4), while the expression of SMURF2CG showed results highly similar to the control, MYC-Empty, sample ( Figure 2A, lane 5), suggesting that SMURF2 ubiquitinates KAP1 in an E3 ligase-dependent manner. These results were consistently observed in multiple independent experiments ( Figure 2B).
Similar findings were also obtained in cells lysed with 1% SDS, followed by immediate sample boiling and sonication ( Figure 2C,D). This stringent cell lysis condition permits more efficient deactivation of cellular DUBs, removing their possible interference with the ubiquitination of KAP1. It also enables the disruption of protein-protein interactions formed in the cellular milieu, eliminating the possibility that the observed ubiquitination of KAP1 emanates from its co-IPed partners. Nevertheless, independently of cell lysis conditions, we consistently observed SMURF2-mediated ubiquitination of KAP1 and its dependence on SMURF2 E3 ligase activity. Additionally, we detected the endogenous ubiquitination of KAP1 facilitated by the adventitious expression of SMURF2 ( Figure S2A).
Finally, to demonstrate that KAP1 is a direct ubiquitination target of SMURF2, we conducted the ubiquitination reconstitution assay (in vitro ubiquitination assay), using purified components of the ubiquitination cascade. These components included ubiquitinactivating enzyme E1, ubiquitin conjugating enzyme E2, GST-SMURF2 (either WT or its E3 ligase-dead form), HA-ubiquitin, and purified FLAG-KAP1 as a substrate. Following the incubation, FLAG-KAP1 was pulled-down from the reaction and its ubiquitination analyzed using anti-HA antibody. The results ( Figure 2E,F) show that SMURF2 is capable to directly ubiquitinate KAP1 in an E3 ligase-dependent manner, whereas SMURF2 mutant failed to produce this phenomenon, revealing results highly similar to the GST control. The purity of the proteins, which we produced and used in this study, was verified by Coomassie gel staining ( Figure S2B).

SMURF2 Positively Regulates the Protein Abundance of KAP1 in Cancer Cells
To determine the consequence of SMURF2 binding to KAP1 and its ubiquitination, we first analyzed the effect of the enforced expression of SMURF2 on the steady-state levels of KAP1. The results show that KAP1 protein levels were markedly increased in SMURF2WTexpressing cells, as compared to the cells transduced with either an empty vector or SMURF2 mutant ( Figure 3A), suggesting that SMURF2 upregulates KAP1 expression in a catalytically-dependent manner. Conversely, the knocking down of SMURF2 through RNAi dramatically decreased the expression levels of KAP1 ( Figure 3B). Similar results were also obtained following SMURF2 depletion using the CRISPR/Cas9 gene-editing tool (SMURF2 CR ), disrupting SMURF2 expression at the genetic level ( Figure 3C, left panels). Furthermore, we found that subsequent reconstitution of active SMURF2 into SMURF2 CR cells reversed this phenomenon and elevated the expression level of KAP1 diminished by SMURF2 depletion (Figure 3C, right panels). This finding suggests that loss of SMURF2 was solely responsible for the decreased cellular abundance of KAP1 in SMURF2 CR cells. The reduced protein levels of KAP1 following SMURF2 depletion were also observed in other cancer cell models, including cervix carcinoma HeLa cells, prostate carcinoma DU-145 cells, and breast adenocarcinoma MDA-MB-468 cells. This effect was monitored both after acute and stable SMURF2 knockdown using different approaches ( Figure 3D-F). Similar findings were also obtained in cells lysed with 1% SDS, followed by immediate sample boiling and sonication ( Figure 2C,D). This stringent cell lysis condition permits more efficient deactivation of cellular DUBs, removing their possible interference with the ubiquitination of KAP1. It also enables the disruption of protein-protein interactions formed in the cellular milieu, eliminating the possibility that the observed ubiquitination of KAP1 emanates from its co-IPed partners. Nevertheless, independently of cell lysis conditions, we consistently observed SMURF2-mediated ubiquitination of KAP1 and its dependence on SMURF2 E3 ligase activity. Additionally, we detected the endogenous ubiquitination of KAP1 facilitated by the adventitious expression of SMURF2 ( Figure S2A).

SMURF2 Negatively Regulates KAP1 Expression in Untransformed Human Cells and Mouse Tissues
As SMURF2 was reported to play a differential role in malignant and normal cells [32,44], we examined the effect of SMURF2 depletion on KAP1 expression in human untransformed cell lines. To this end, we knocked down SMURF2 in diploid lung IMR-90 fibroblasts and dermal BJ1 cells, which are known to maintain the untransformed phenotype typical of normal cells. The expression levels of KAP1 in these cells were then analyzed by Western blotting. The results ( Figure 4A) showed that in contrast to tumor cells SMURF2 knockdown in IMR-90 and BJ1 cells markedly increased KAP1 protein levels, suggesting that, in these cells, SMURF2 operates as a negative regulator of KAP1 expression. This effect was also observed in SMURF2KO (SMURF2 CR ) cells ( Figure 4B-D). Noteworthy, the data showed that SMURF2 exhibits its regulatory effect on KAP1 in a cell-context-dependent manner, as we did not detect any significant changes in KAP1 expression levels in SMURF2 knockdown MCF-10A mammary epithelial cells (data not shown). The context-dependent regulation of KAP1 by SMURF2 was also noted in Smurf2-ablated mouse tissues. Specifically, we found that Kap1 levels were significantly elevated in Smurf2KO liver tissues but showed no significant increase in other analyzed organs and tissues ( Figure 4E-G).

The SMURF2-KAP1 Relationship in Human Normal and Cancer Tissues
To relate our findings to the clinical settings, we examined the relationship between the protein expression of KAP1 and SMURF2 in different human normal and breast cancer tissues, using tissue microarrays (TMAs) and IHC analysis. Particularly, we stained and analyzed two different sets of TMAs: a multi-organ normal TMA (FDA999m, US Biomax), containing 32 types of normal tissues taken from three individuals, and breast cancer TMA (BR804a, US Biomax), containing 40 cases of breast cancer samples with 40 matched

The SMURF2-KAP1 Relationship in Human Normal and Cancer Tissues
To relate our findings to the clinical settings, we examined the relationship between the protein expression of KAP1 and SMURF2 in different human normal and breast cancer tissues, using tissue microarrays (TMAs) and IHC analysis. Particularly, we stained and analyzed two different sets of TMAs: a multi-organ normal TMA (FDA999m, US Biomax), containing 32 types of normal tissues taken from three individuals, and breast cancer TMA (BR804a, US Biomax), containing 40 cases of breast cancer samples with 40 matched adjacent normal tissues. The IHC analysis of normal tissues revealed that, despite the heterogenous expression of SMURF2 and KAP1 in different individuals,~50% of tissues were scored equally in terms of the SMURF2 and KAP1 staining intensity and the percentage of positive cells, meaning that tissues with a lower expression of SMURF2 showed a lower expression of KAP1, and vice versa ( Figure 5A). These tissues included the cerebellum, adrenal gland, pancreas, thyroid, and bone marrow, among others ( Figure 5B,C). In the lymph node, spleen, stomach, and prostate, SMURF2 and KAP1 were differentially scored: samples with a lower expression of SMURF2 showed a higher expression of KAP1, and samples with higher SMURF2 levels exhibited low KAP1 expression.
The IHC analysis of breast TMA showed that, in more than 80% of normal breast tissues, SMURF2 and KAP1 were equally scored in terms of staining intensity and the percentage of positive cells ( Figure 5D, upper panels). In cancer tissues, however, the SMURF2-KAP1 relationship was considerably shifted towards the differential score ( Figure 5D, bottom panels). Most of the cancer tissues showed either a lower expression of SMURF2 and higher expression of KAP1 ( Figure 5E) or the opposite: a higher expression of SMURF2 and lower expression of KAP1, suggesting that the carcinogenic process may change the relationship between the expression of these proteins. Noteworthy, and similar to the previous findings [20,21,44], the expression levels SMURF2 and KAP1 in tumors were considerably higher as compared to their normal matching counterparts ( Figure 5F). These IHC findings, together with our data showing that in certain types of cancer cells, including breast cancer, KAP1 is stabilized by SMURF2, may imply that elevated expression of KAP1 in certain tumors emanates from the heightened expression of SMURF2.

SMURF2 Has a Significant Influence on KAP1 Interactome, Regulating Its Protein Interactions and Downstream Functions in an E3 Ligase-Dependent Manner
To further investigate the impact of SMURF2 on KAP1, we then stratified the cellular interactome of KAP1 in cells expressing either active SMURF2 or its catalytically-deficient mutant. The analysis showed that KAP1 interacts with~600 proteins, with 534 proteins shared between the analyzed groups: SMURF2WT, SMURF2CG, and the control MYC-Empty ( Figure 6A). Applying an additional threshold of ≥2-fold change in the protein abundance reduced the number of KAP1 interactors to 105-148 proteins, with the lowest number of KAP1-interacting partners detected in SMURF2WT cells ( Figure 6B). Additionally, the results revealed that the enforced expression of SMURF2WT abolished KAP1 interactions with 56 proteins. Along with this finding, the MS data showed that the expression of both SMURF2WT and SMURF2CG forms increased KAP1 associations with its several known binding partners and promoted the formation of previously unreported protein-protein interactions of KAP1 (PXD029642). In total, 27 and 40 unique KAP1-interacting proteins were identified in cells expressing SMURF2WT and SMURF2CG, respectively ( Figure 6B).
The protein class analysis using the PANTHER classification system showed that the KAP1 binding partners in a control sample consisted nucleic acid metabolism proteins, gene-specific transcriptional regulators, translational proteins, protein-modifying enzymes, chromatin-binding or regulatory proteins, and scaffolds/adaptors, among several others ( Figure S4A). Similar classes of proteins were also detected in SMURF2-enriched cells, although with certain differences. For example, in cells with the enforced expression of SMURF2WT, we did not find as KAP1 interactors the scaffold/adaptor proteins, which were present in both control and SMURF2CG-expressing samples ( Figure S4B,C). Instead, SMURF2WT-expressing cells showed the enrichment of protein-binding activity modu-lators, suggesting that this group of proteins was enriched in KAP1 interactome by the expression of active SMURF2. adjacent normal tissues. The IHC analysis of normal tissues revealed that, despite the heterogenous expression of SMURF2 and KAP1 in different individuals, ~50% of tissues were scored equally in terms of the SMURF2 and KAP1 staining intensity and the percentage of positive cells, meaning that tissues with a lower expression of SMURF2 showed a lower expression of KAP1, and vice versa ( Figure 5A). These tissues included the cerebellum, adrenal gland, pancreas, thyroid, and bone marrow, among others ( Figure 5B,C). In the lymph node, spleen, stomach, and prostate, SMURF2 and KAP1 were differentially scored: samples with a lower expression of SMURF2 showed a higher expression of KAP1, and samples with higher SMURF2 levels exhibited low KAP1 expression. The reconstruction of the KAP1 protein-protein interaction map using the STRING tool showed that SMURF2 upregulates the molecular networks associated with the negative regulation of transcription, mRNA splicing, RNA binding, processing, protein translation and targeting, as well as heat shock protein binding ( Figure S5A). In parallel, its decreased KAP1 associations with molecular networks implicated in the regulation of gene expression, RNA binding, splicing, mRNA stability, as well as structural molecule activity and regulation of cell cycle phase transition ( Figure S5B). Moreover, the gene ontology (GO) analysis revealed that, in comparison to an empty control and catalytically-inactive SMURF2CG, SMURF2WT downregulated several key biological processes and molecular functions associated with the activities of KAP1 interactome. These included RNA binding and processing, viral transcription, post-transcriptional regulation of gene expression, and translation, among others ( Figure 6C,D and Table S4). Furthermore, the GO:cellular component analysis indicated that SMURF2WT markedly decreased the association of KAP1 interactors with the ribonucleoprotein complexes, cytosolic components, and ribosomes, while increasing the KAP1 association with proteins associated with the chromosome telomeric region ( Figure 6E). Additionally, the ingenuity pathway analysis (IPA) of the KAP1 interactome in SMURF2WT vs. SMURF2CG and the empty control samples revealed significant differences in the percentage of enrichment of the proteins involved in key signaling cellular pathways. These included EIF2 signaling, regulation of eIF4 and p70S6K signaling, mTOR signaling, the BAG2 signaling pathway, and others ( Figure 6F and Table S5). Most of these pathways were predicted to be downregulated by SMURF2WT, recapitulating findings of the GO analysis on the downregulation of biological processes and the molecular functions of KAP1 interactors by SMURF2. Taken together, these findings show that SMURF2 has a significant influence on KAP1 interactome, regulating its functions and associated signaling pathways in a catalytically-dependent manner.  Table S4. (F) KAP1 interactome analysis using IPA. Top ten signaling pathways with the highest −log(p-value) are shown. The detailed list/analysis of all signaling pathways predicted to be altered by SMURF2WT vs. SMURF2CG and control empty vector samples are shown in Table S5.
The protein class analysis using the PANTHER classification system showed that the KAP1 binding partners in a control sample consisted nucleic acid metabolism proteins, gene-specific transcriptional regulators, translational proteins, protein-modifying enzymes, chromatin-binding or regulatory proteins, and scaffolds/adaptors, among several others ( Figure S4A). Similar classes of proteins were also detected in SMURF2-enriched cells, although with certain differences. For example, in cells with the enforced expression of SMURF2WT, we did not find as KAP1 interactors the scaffold/adaptor proteins, which were present in both control and SMURF2CG-expressing samples ( Figure S4B,C). Instead, SMURF2WT-expressing cells showed the enrichment of protein-binding activity  Table S4. (F) KAP1 interactome analysis using IPA. Top ten signaling pathways with the highest −log(p-value) are shown. The detailed list/analysis of all signaling pathways predicted to be altered by SMURF2WT vs. SMURF2CG and control empty vector samples are shown in Table S5.

Discussion
Accumulating evidence shows that KAP1 is intrinsically engaged in several core biological processes pertinent to carcinogenesis, disease progression, therapeutic response, and patient survival, revealing both tumor-promoting and suppressing activities. Despite these proceedings, the KAP1 regulatory mechanisms, especially the mechanism regulating its cellular abundance, remain poorly characterized. Up-to-date, there are only two studies related to the regulation of protein abundance of KAP1 in cells. The first study showed that following DNA damage the phosphorylated and SUMOylated KAP1 is targeted for ubiquitin-mediated proteasomal degradation facilitated by RING-type E3 ubiquitin ligase RNF4 [49]. The second, more recent research, reported that Dcaf11, one of the central components in the cullin-4-based RING E3 ligase complex, targets KAP1 for ubiquitin-mediated degradation, thereby regulating telomere elongation in early embryos and embryonic stem cells [50].
In our study, we expanded the list of KAP1 post-translational regulators and identified the HECT-type E3 ubiquitin ligase SMURF2 as a pivotal regulator of KAP1. We showed that SMURF2 and KAP1 directly interact with each other, both in vitro and in the cellular environment (Figure 1), allowing a direct ubiquitin transfer from SMURF2 to KAP1 ( Figure 2). We further demonstrated that this phenomenon is relied on unaltered E3 ubiquitin ligase activity of SMURF2, leading to the stabilization of KAP1 in cancer cells ( Figure 3A). Depletion of SMURF2 using either the RNAi or CRISPR/Cas9-based approach profoundly diminished the protein expression of KAP1 ( Figure 3B-D), while unaffecting its mRNA expression levels ( Figure S3). Noteworthy, the re-expression of SMURF2 in SMURF2-depleted cells restored KAP1 protein levels ( Figure 3C), suggesting SMURF2 as one of the major factors controlling the stability of KAP1 through the proteasomal degradation pathway ( Figure 3G,H). Interestingly, while acting in cancer cells as a KAP1 stabilizing factor, in untransformed IMR90 and BJ1 cells, SMURF2 showed a negative impact on KAP1 expression ( Figure 4A-D), the phenomenon that was also observed in certain Smurf2KO tissues ( Figure 4E-G). This suggested that SMURF2 regulates KAP1 in cell/tissue-specific manner. The contextual nature of KAP1 regulation by SMURF2 was also evident from the examination of several other untransformed and cancer cell strains, including mammary epithelial MCF-10A cells, breast carcinoma MDA-MB-231 cells, and prostate adenocarcinoma PC3 cells. In all these cells, SMURF2 knockdown did not affect the steady-state levels of KAP1 (data not shown) in contrast to IMR-90 and BJ1 cells, as well as to the MDA-MB-468, HeLa, DU-145, and U2OS cells shown in this study (Figures 3  and 4). The context-dependent regulation of KAP1 by SMURF2 was also suggested by the IHC results obtained in TMAs ( Figure 5). One of the possible explanations for the differential regulation of KAP1 expression by SMURF2 in different types of cells and tissues could be the activation of compensatory mechanisms, including alterations in the cell transcriptional activity, affecting KAP1 levels. Indeed, SMURF2 was previously reported to act as a pleiotropic factor affecting gene expression at both epigenetic and post-translational levels in a cell-context-dependent manner [32,34,51]. Another potential explanation for the contextual regulation of KAP1 by SMURF2 could emanate from the differential expression in these cells of the deubiquitinating enzymes (DUBs). Indeed, the expression and activities of DUBs were extensively documented to impose a significant impact on protein degradation dynamics and proteome remodeling, both under normal conditions and in the disease states, including cancer [52,53]. In the case of SMURF2, at least two different DUBs-USP4 and USP15, shown to interfere with SMURF2-mediated ubiquitination of its protein targets [54,55], might be implicated. Additional possibilities may involve the ubiquitination of distinct lysine residues of KAP1 by SMURF2 and/or the formation of different ubiquitin chains on KAP1 in different types of cells and tissues. Further investigations of these possibilities will be needed for understanding the full details of the relationship between KAP1 and SMURF2. The subsequent studies should also determine the biological consequences of the SMURF2-mediated regulation of KAP1 interactome and its associated functions and signaling pathways uncovered in this work ( Figure 6). Particularly, the influ-ence of SMURF2-KAP1 on RNA biology, post-transcriptional regulation of gene expression, protein translation, structural molecule activity, regulation of the cell cycle, and telomere maintenance. Addressing these questions will allow better understanding of the role of the SMURF2-KAP1 complex in the biology of untransformed and malignant cells, promoting the development of pathway-oriented precision therapies targeting KAP1 in cancer.

Conclusions
In this work, we identified the HECT-type E3 ubiquitin ligase SMURF2 as an important cellular factor regulating the protein expression of KAP1 in different types of cells and tissues in a context-dependent manner. These findings lay a foundation for further investigations of the role of the SMURF2-KAP1 axis in the carcinogenic process and therapeutic response, with an ultimate goal to develop new, more effective treatment strategies based on the disruption in cancers of the dysregulated SMURF2/KAP1 module.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cancers14071607/s1, Figure S1: SMURF2 and KAP1 physically interact with each other, Figure S2: Endogenous ubiquitination assay and Coomassie gel staining of purified proteins used in this study, Figure S3: qRT-PCR analysis of KAP1 mRNA expression in SMURF2 knock down cells, Figure S4: PANTHER analysis of protein classes of KAP1 interactors, Figure S5: STRING analysis of protein-protein interaction network of KAP1 interactors affected by SMURF2, Figure S6: Uncropped Western Blot (WB) images, Table S1: siRNA sequences, Table S2: Primary antibodies used for immunoblotting, Table S3: qRT-PCR primers used in the study, Table  S4: Details of GO terms for KAP1-interacting proteins changed in abundance by SMURF2, Table S5: Details of the Ingenuity Pathway Analysis (IPA) of KAP1 interactome affected by SMURF2WT and its catalytically-inactive mutant (SMURF2CG).

Informed Consent Statement: Not applicable.
Data Availability Statement: The mass spectrometry proteomics data of KAP1 interactome have been deposited to the ProteomeXchange Consortium via the PRIDE [56] partner repository with the dataset identifier PXD029642, available online: https://www.ebi.ac.uk/pride/archive, accessed on 10 November 2021.