RETRACTED: Arginine methylation of SHANK2 by PRMT7 promotes human breast cancer metastasis through activating endosomal FAK signalling

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Reviewer #1:

The current manuscript by Liu et al., attempts to link PRMT7 activity concomitant with Shank2 di-methylation, to the regulation of migration (through FAK signaling) and proliferation (through the regulation of H-Ras ubiquitylation).The biochemical data connecting SHANK2 di-methylation to the association with cytoskeletal elements (FAK/cortactin) and H-RAS is extensive but in certain areas lacks critical controls. Overall, the data support the notion that methylation of Shank2 is important and that this modification may be mediated by PRMT7. In this way, the study very much mirrors other published papers wherein PRMT7 has been shown to methylate a protein (like eIF2alpha for example) leading to phenotypic alterations. The functional assays suggesting that SHANK2 methylation may mediate a switch between invasion and proliferation are however very weak and contradictory in some ways. The clinical correlations are also likely underpowered, making it difficult to make conclusions. Very little is known about Shank2 and the idea that methylation (by PRMT7) could refine its binding partners and function is novel. However, further experimentation is needed to further establish functional links as described.

Major points:

1) Reviewer #1 and 2’s major concern: All proteomics data sets must be shown in order to ascertain how robust and specific the interactions chosen actually are.

We thank reviewers for pointing this out. As shown in the Figure 1A and Figure 5B, we added the results of FlagPRMT7 and HA-shank2 mass spectrometry respectively, which included PRMT7 associated proteins/peptides and shank2 associated proteins/peptides.

2) Reviewer #1 and 3’s major concern: The studies outlined in Figure 3 did not actually test whether Shank2 R240 methylation promotes cell migration through activating FAK/cortactin signaling. Similarly, the results presented for proliferation in in Figure 6 did not link the interaction of Shank2 with HRas to its regulation of proliferation. As it stands, the results are purely correlative; such that migration or proliferation assays would need to be designed to test the consequences of these interactions specifically. Indeed, it is likely that Shank2 R240 affects many factors, which may contribute to migration and proliferation.

We thank reviewers for the suggestion and performed a series of new experiments in MDA-MB-231 cells and Balb/c nude mice to address this issue. In brief, we used FAK inhibitor GSK2256 to analyze if shank2 R240 methylation promoted cell migration through activating FAK/cortactin signaling (Figure 7FJ-L). Furthermore, we also tested whether shank2 R240 methylation promoted tumour metastasis through triggering FAK/cortactin signalling activation by using FAK inhibitor (Figure 8A). All the results suggested that shank2 R240 methylation promoted tumour metastasis by activating FAK/cortactin signaling.

3) Reviewer #1’s major concern: In Figure 2, all of the samples have methylated Shank2, despite varying levels of Prmt7. Moreover, some samples have very little Shank2. Hence, it is unclear whether the methylated levels would be biologically significant. Indeed, the percentage of methylated protein was not estimated in any assay and probably should be. The protein atlas is not a fantastic way to analyze this type of thing and patient tissues should instead be analyzed. Finally, the RT-PCR analysis should be expanded to include details regarding expression levels in patients with different types of breast cancer and/or differing outcomes. Overall, as it stands the clinical correlations are relatively weak.

We greatly appreciate the reviewer for raising the critical question about the clinical correlations between shank2 R240 methylation and patients with different types of breast cancer. We re-analyzed the methylation level of shank2 in different types of breast cancer tissues. In the Figure 3A-D and supplement table, we tested the paracancer and cancer tissues of 27 patients and found that the level of PRMT7 and methylation of shank2 was higher in luminal B Her2⁽⁺⁾ and Triple-negative breast cancer samples than that in normal, luminal A and luminal B Her2^(-) samples, which implicated a positive correlation between shank2 methylation and high metastatic potential.

4) Reviewer #1 and 3’s major concern: Throughout the study, critical controls are missing. For example, in Figure 3, the non-treated controls are needed and in Figure 7, all treatments need to be compared to shShank alone as well as the untreated cells.

We thank the reviewer for the comment. We have supplemented the control and shshank2 group in full text.

5) Reviewer #1’s major concern: The mechanisms by which demethylation of Shnk2 affects function were not sufficiently explored and yet were extrapolated with the data presented in Figure 4. The authors should specifically interrogate how the two methyl groups at R270 affect binding to the chosen proteins. Which parts of Shank2 are binding to the other proteins and how or why is this altered upon di-methylation?

We greatly appreciate the reviewer’s suggestion. We presented two methyl groups at shank2 R240 data to analyze the closed/open conformation of the SPN domain and ANK domain. Structural analyze indicated that shank2 R240 methylation disrupted SPN-ANK domain blockade to “open” shank2, providing more chances for insertion of partner proteins (Figure 4).

6) Reviewer #1 and 3’s major concern: The functional assays suggesting that SHANK2 methylation may mediate a switch between invasion and proliferation are however very weak and contradictory in some ways. It is unclear whether the phenotype of increased tumor growth is dependent on activated RAS-MAPK pathway. The author should have conducted similar experiment using MEK/ERK inhibitor in vivo, for example.

We greatly appreciate the reviewers for raising the issue about shank2 R240 methylation mediated proliferation and invasion switching. It’s true that proliferation and invasion are two complicated process. In the present manuscript, we just focused on the role of PRMT7 mediated shank2 R240 methylation in promoting invasion and metastasis of breast cancer.

Due to some studies have shown that RAS activation can activate FAK, and FAK activation can also activate RAS. To simplify the complex problems, we decided to remove the results that shank2 R240 methylation inhibited tumour growth by mono-ubiquitinating RAS, which we are stilling tracing.

In addition, we performed new experiments in endocytosis and FAK activation to address the mechanism of shank2 methylation triggered breast cancer metastasis. According to the results of shank2 affinity mass spectrometry, we found that methylated shank2 can interact with a kind of endocytosis associated protein, such as clathrin, AP2, dynamin2 and EEA1 etc. Indeed, PRMT7 mediated shank2 methylation reinforced tumour metastasis through activating endosomal FAK signaling (Figure 7D and Figure 8A). Taken together, the results presented indicated that shank2 R240 methylation promoted breast cancer metastasis through activating endosomal FAK signalling.

[Editors’ note: what follows is the authors’ response to the second round of review.]

Essential revisions:

1) Please carefully edit both the Introduction and the Discussion section. There are many grammatical and spelling errors in these sections. Also, make sure that the Discussion section refers to the correct Figures. In addition, please check that the proper tense is used throughout the paper.

We sincerely thank the reviewer for pointing out our errors. We went through the manuscript carefully and corrected all the grammatical and spelling errors we found. Also, we adjusted the corresponding figures in the Discussion section.

2) For patient material, include an ethics statement. Also, please include a table outlining the characteristics of each patient (including parameters such as stage and subtype).

We added the ethics statement in Materials and methods section. Meanwhile, Supplementary file 1 and Supplementary file 2 have now been supplied to provide the information regarding the tumour subtypes, pathological analysis and malignancy of the breast cancer patients.

3) In Figure 2E, please include the PRMT7 blots showing increasing expression.

We added the blots (Flag-PRMT7) in Figure 2E in the revised manuscript.

4) It seems that PRMT7 may just be increased in all breast cancers, and that the correlation to metastasis is overstated. Please use ANOVA to analyse data in Figures 3A and 3B. Also, please include an analysis of a larger data set (ie TCGA) for overall and progression free survival. If there are no correlations, then it may just be that this factor in increased in cancer in general, which would be fine.

We thank the reviewer for this valuable and constructive suggestion. According to Figure 3B, we conducted an optical density analysis of the bands. Through the optical density values, we calculated the ratios of PRMT7 to β-actin and di-methylated SHANK2 to SHANK2. We found that the level of PRMT7 and PRMT7 mediated-SHANK2 di-methylation is higher in Luminal B, HER2⁺ and triple-negative breast cancer compared with normal breast tissue or Luminal A breast cancer (Figure 3B, C). Therefore, our data implicate a positive correlation between PRMT7 expression and high metastatic potential.

To further demonstrate the relationship between PRMT7 and metastasis, we analyzed the correlation between PRMT7 and the stage of breast cancer patients through TCGA Breast (BRCA) database. Among 1075 breast cancer patients, 847 had low PRMT7 expression (79%) and 228 had high PRMT7 expression (21%). In addition, we found that 89 out of 847 PRMT7 low-expression breast cancer patients were stage III or IV (10.5%), while 63 out of 228 PRMT7 high-expression breast cancer patients were stage III or IV (27.6%) (Figure 3—figure supplement 1). These data indicate that the expression of PRMT7 is positively correlated with the proportion of patients with stage III or IV breast cancer, which predicts poor prognosis.

5) The IF images in Figure 5 are not ideal, and not the best way to conclude co-localisation. Please comment on weaknesses in discussion, noting that other more precise assays (like PLA) may be warranted.

We agree with the reviewer on this issue. Our IF images showed that SHANK2, Talin, FAK, cortactin, EEA1 and Rab5 were localized at the same location in the cells. Consistently, our IP assays also showed that SHANK2, Talin, FAK and cortactin were in the same complex. We agree that PLA is important for analyzing the precise interaction between SHANK2 and endosomal proteins. However, PLA experiment is suitable for proving the binding between two proteins, while our results showed that SHANK2 methylation affected the combination of SHANK2 protein complex. It is not clear which of the two proteins directly bind to each other. Therefore, we did not carry out related work for the time being, and the follow-up work will continue to be studied.

6) Please use ANOVA to analyse Figures 8B and 8D.

We thank the reviewer for the suggestion. We have incorporated the analyses into the revised manuscript in Figure 8B, 8C and 8D.

7) Please show all images related to 8E in the supplement. Also include how stage and staining correlates with subtype.

We showed all the immunochemical images of breast cancer tissues and statistic of stage and staining correlates with subtype related to Figure 8E in Figure 8—figure supplement 1 in the revised manuscript.

8) The metastasis assay employed is more of a lung colony assay and misses many aspects that require invasion. Please discuss the limitations of this assay.

We thank the reviewer for pointing this out. Because we found that MDA-MB-231 cells have a higher level of methylation of SHANK2 compared with normal mammary epithelial cell (MCF10A) or low metastatic potential breast cancer cells (MCF7, BT474 and T47D), we examined the effect of SHANK2 methylation on migration/invasion ability and lung metastasis using MDA-MB-231 cells.

Given that breast cancer MDA-MB-231 cells have a lower ability of lung metastasis after tumour formation in situ of nude mice (Raquel et al., 2006); we have to choose tail vein injection to detect the effect of SHANK2 methylation on lung metastasis. Meanwhile, many studies also used mouse tail vein injection to detect lung metastasis using MDA-MB-231 cells (Xiaolong et al., 2017; Kyoungwha et al., 2019; Mark et al., 2019). Consistently, our lung colony assay results indicate that SHANK2 R240 methylation promotes lung metastasis of MDA-MB-231 cells.

9) Introduction – The authors state that PRMT7 is capable of producing SDMA and give two references to support this assertion. However, structurally it has been shown that PRMT7 is only capable of catalyzing MMA (summarized in Jain and Clarke, 2019). Furthermore, PRMT5, the main PRMT responsible for SDMA is a common contaminant in Flag IPs (Nishioka et al., 2003). Additionally, alteration of PRMT7 expression can also affect cellular SDMA levels (summarized in Jain and Clarke, 2019), thus making the use of SDMA antibodies tricky when determining and analyzing PRMT7 and its substrates.

We greatly appreciate the reviewer asking a critical question about PRMT7 mediated SDMA. Studies reported that PRMT7 catalyzed histone H4 mono-methylation in Trypanosoma brucei (Wang et al., 2014; Tamar et al., 2018). On the other hand, previous studies also showed that PRMT7 mediated p38 and GLI2 di-methylation, but did not mediate mono-methylation in skeletal muscle cells (Jeong et al., 2020; Vuong et al., 2020). Similarly, PRMT7 also affected sm di-methylation but not mono-methylation in HeLa cells (Graydon et al., 2007). Our mass spectrometry results indicated that SHANK2 only occurred SDMA in MDA-MB-231 cells with high level of PRMT7. Meanwhile, our results demonstrated that overexpression of PRMT7 resulted in increased symmetric arginine di-methylation of SHANK2 in HEK293T cells (low level of PRMT7), while SHANK2 di-methylation level was reduced upon PRMT7 depletion (Figure 1G-I). Meanwhile, when increasing doses of PRMT7 in HEK293T cells were overexpressed, SHANK2 R240 di-methylation level was increased accordingly (Figure 2E). To further confirm PRMT7 directly methylated SHANK2, we used purified GST-PRMT7 and HA-SHANK2 for in vitro methylation assay followed by incubation with SAM (the methyl donor). Apparently, incubation of recombinant PRMT7 and SHANK2 gave rise to a remarkable increase in methylation of SHANK2 in the presence of SAM (Figure 1J). Data obtained indicate that PRMT7 is required for SHANK2 di-methylation.

Previously studies reported that PRMT5 and PRMT7 were in the same complex (Gonsalvez et al., 2007). Our mass spectrometry results also showed that PRMT5 and PRMT7 were in the same complex. To determine the role of PRMT5 in PRMT7-mediated SHANK2 R240 di-methylation, we examined the effect of PRMT5 overexpression on SHANK2 R240 di-methylation in HEK293T cells. We found that dosage dependent expression of PRMT5 did not increase the di-methylation of SHANK2 R240 (Figure 2—figure supplement 1A, B). Furthermore, we examined the effect of PRMT5 inhibitors (GSK591) on SHANK2 di-methylation. We found that even inhibiting PRMT5 activity, overexpression of PRMT7 could still increase the SHANK2 di-methylation level both in HEK293T and MDA-MB-231 cells (Figure 2—figure supplement 1C, D). Thus, although PRMT5 and PRMT7 exist in the same complex, our data indicate that the di-methylation of SHANK2 R240 is mainly mediated by PRMT7.

10) Figure 1E – The molecular of each construct should be included in the figure.

We have added molecular tags of each construct in Figure 1E in the revised manuscript.

11) Figure 1G – Can the specificity of the SDMA band corresponding to shank2 be confirmed? This can be done by including the full, uncropped SDMA blot for an experiment were the authors use a shank2 shRNA, for ex. Figure 2C. Also, can the authors speculate why methylation of shank2 wasn't detected with the MMA antibody?

We thank the reviewer for the comment. According to Figure 1J and H, we found that SHANK2 only occurred as SDMA but not MMA and ADMA in the presence of PRMT7. Meanwhile, our mass spectrometry results also indicated that SHANK2 only occurred as SDMA in MDA-MB-231 cells. Together, we speculate that PRMT7 may regulate its spatial conformation by combining with other cofactors, resulting in di-methylation of its substrate instead of mono-methylation modification.

In addition, we analyzed the SDMA of SHANK2 in full uncropped SDMA blot in shCtrl or shPRMT7 MDA-MB-231 cells in the revised manuscript as shown below, indicating that PRMT7 di-methylated the R240 residue of SHANK2.

12) Figure 1J – In text and figure, the authors state that a HA-tagged shank construct was used, however in figure legend and Materials and methods section, it is stated that a Flag-tagged shank construct was used. Please clarify.

We apologize for the mistake. We corrected the mistake in Figure legend 1J and the Materials and methods section in the revised manuscript.

13) Figure 5E – PRMT7 staining is not detectable in this image. Please address this.

Given that the extremely low level of PRMT7 in MCF7 cells, therefore PRMT7 staining is not detectable in MCF7-Vector cells. We have indicated this issue in the revised manuscript.

14) Subsection “Shank2 R240 methylation recruits FAK/dynamin2/talin complex to endosome” – A statement should be included stating the R240F is a methyl-mimic.

We stated R240F as a methyl-mimic in the revised manuscript in subsection “SHANK2 R240 methylation recruits FAK/dynamin2/talin complex to endosome” in the revised manuscript.

15) Figure 7 – Can the authors comment on why expression of R240F didn't increase the migration/invasive phenotype if methylation of shank by PRMT7 is supposed to be responsible for mediating these processes?

Indeed, our results indicate that SHANK2 R240F increased the capability of migration and invasion in MDA-MB-231 cells (Figure 7B, C, E, F, H, I). In addition, we also compared SHANK2 240K and SHANK2 240F, and found that SHANK2 240F had stronger migration and invasion capabilities than SHANK2 240K.

16) Discussion section – It is stated shank2 interacts with Met, NRP1, and EGFR, however these results are not included in table in Figure 5A. Likewise, LAMP1, TOM70, and Lamin A/C are not included in the results. Please remove the statement.

In fact, we did find these proteins in our mass spectrometry. We have included these proteins in Figure 5—figure supplement 1A in the revised manuscript. These data indicate that SHANK2 could bind to receptor proteins and different organelles, such as lysosome, mitochondria, ER, Golgi, centrosome, exosome and nuclear.

17) In Figure 1A, the number of PRMT7 peptides (only 11) identified upon PRMT7 immunoprecipitation seems quite low, suggesting a poor enrichment in PRMT7. Is the table a selected list of peptides identified or the most abundant peptides detected? Please address in paper.

We thank the reviewer for pointing this out. Indeed, peptides shown in Figure 1A are part of peptides we selected. Most of our mass spectrometry results are proteins in the cytoplasm (Figure 1A). Consistently, the previous studies proved that PRMT7 was mainly existed in cytoplasm (Ferreira et al., 2020). Therefore, we selected the relevant proteins in the cytoplasm for further research.

18) Performing FLAG-purification to study PRMT7 might generate important contaminations from PRMT5 (and probably PRMT5-associated proteins) since 146 out of 156 (94%) reported experiments in the CRAPome (crapome.org; Mellacheruvu et al., (2013)). Were appropriate controls with Mock flag-purification performed? If yes, the peptide counts in control conditions should also be provided. Knowing that PRMT5 interacts with PRMT7 (Gonsalvez et al., 2007), that both enzymes generate symmetric di-methylation, and that PRMT5 binds to FLAG peptide (Mellacheruvu et al., 2013), the use of FLAG-tagged PRMT7 as a bait appears as an unsound strategy. Please discuss caveats.

We greatly appreciate the reviewer for asking a critical question about the interaction between PRMT7 and PRMT5. To determine how PRMT7 orchestrated the metastasis of breast cancer, we performed FLAG protein purification and mass spectrometry identification in PRMT7-MCF7 cells. Indeed, our mass spectrometry results found that PRMT5 and PRMT7 were in the same complex. According to the results in Figure 2—figure supplement 1, although PRMT5 and PRMT7 may exist in the same complex and work together, the di-methylation of SHANK2 R240 is mainly mediated by PRMT7.

On the other hand, we did compare the purification results of FLAG group and FLAG-PRMT7 group by silver staining before performing mass spectrometry on the purified FLAG-PRMT7. However, we only compared the differences in protein bands between the two groups, and did not perform mass spectrometry analysis on the control bands.

19) In Figure 2C, all the conditions should be shown in the Shank2 IP section (not only in the input). Furthermore, this approach is rather unspecific using anti-SHANK2 instead of tagging the mutant and wt protein. Difficult to assess whether the exogenous protein of endogenous is IPed in this Figure (although SHANK2 levels are restored in depleted cells).

We thank the reviewer for the comment. We re-performed the methylation modification of HA-tag SHANK2 in the revised manuscript. indicating that PRMT7 di-methylated the R240 residue of SHANK2.

20) In Figure 5B, SHANK2 was not identified in anti-SHANK2-immunoprecipitated material. This suggests the lack of enrichment of SHANK2 in the IPed material and probably the non-sepcific purification of PRMT7 and other proteins in the process. Please address this. Furthermore, please include mock anti-HA purified controls.

We thank the reviewer for the comment. In fact, our mass spectrometry results were also enriched with peptides containing SHANK2, and we added the information about SHANK2 peptides in the revised manuscript. Together, our data indicated that SHANK2 R240 di-methylation catalyzed by PRMT7 was in favour of the assembly of a multi-protein complex to facilitate endosome formation.

Furthermore, we included the mock anti-HA purified controls in the revised manuscript.

21) SHANK2 protein is referred as "shank2"? Please use standard gene/protein nomenclature.

We adjusted the SHANK2 gene to “SHANK2” and SHANK2 protein to “SHANK2” in the revise manuscript.

22) The size of some figures is inappropriate (Figure 3G and H (inlays), immunofluorescence images in Figure 5; inlays in Figure 7; Figure 8A. Please increase sizes.

We adjusted the size of the images in these figures in the revised manuscript.

23) The Abstract should be improved. The authors should provide introductory sentences leading to a clear rationale before describing results.

We greatly appreciate the reviewer’s suggestion. We have rewritten the Abstract in the revised manuscript.

References:

Chongyuan, Wang., Yuwei, Z., Tamar, B.C., Lei, L., et al. (2014). Structural determinants for the strict mono-methylation activity by Trypanosoma brucei protein arginine methyltransferase 7, Structure 22:756–768.

Graydon, B.G., Liping, T., Jason, K.O., François, M.B., et al. (2007). Two distinct arginine methyltransferases are required for biogenesis of Sm-class ribonucleoproteins, J. Cell Biol. 178:733–740.

Kyoungwha, P., Jinah, P., Sung, G.A., Jihee, L., et al. (2019). RNF208, an estrogen-inducible E3 ligase, targets soluble Vimentin to suppress metastasis in triple-negative breast cancers. Nat Commun 10: 5805.

Raquel, M., Shan, M., Yuval, S., Christina, R.L., et al. (2006). Highly Efficacious Nontoxic Preclinical Treatment for Advanced Metastatic Breast Cancer Using Combination Oral UFT-cyclophosphamide Metronomic Chemotherapy. Cancer Res 66:3386-91.

Tamar, B.C., Abhishek, T., Owen, M.P., Nicole, I., et al. (2018). Phe71 in type III trypanosomal protein arginine methyltransferase 7 (TbPRMT7) restricts the enzyme to mono-methylation. Biochemistry 57:1349–1359.

Tiago, R.F., Adam, A.D., Ewan, P., Eliza, V.C.A., et al. (2020). PRMT7 regulates RNA-binding capacity and protein stability in Leishmania parasites. Nucleic Acids Res 48: 5511-5526.

Xiaolong, T., Lei, S., Ni, X., Zuojun, L., Minxian, Qian., et al. (2017). SIRT7 antagonizes TGF-β signaling and inhibits breast cancer metastasis. Nat Commun 8: 318.

Mark, E., Nandini, M., Todd, M.G., Yong, W., et al. (2019). Bone vascular niche E-selectin induces mesenchymal-epithelial transition and Wnt activation in cancer cells to promote bone metastasis. Nat Cell Biol 21: 627-639.