The PRMT7-dependent methylation of shank2 modulates invasion-proliferation switching during breast cancer metastasis

Invasiveness of cancer cells is associated with proliferation inhibition in multiple types of cancers. Here, we identified the pivotal roles of Arginine methyltransferase PRMT7 in promoting invasion and attenuating proliferation of breast cancer cells. PRMT7 exerted its functions through binding to the scaffold protein shank2 to induce the di-methylation of shank2 at R240. Shank2 R240 methylation exposed ANK domain by disrupting its SPN-ANK domain blockade. Moreover, shank2 R240 methylation rendered recruitment of FAK that elicited the FAK auto-phosphorylation, which consequently augmented the shank2-dependent migration and invasion of breast cancer cells. On the other hand, the shank2 R240 methylation impeded proliferation of breast cancer cells by antagonizing the Ras-Raf binding via tethering the mono-ubiquitinated H-Ras. These findings characterize the PRMT7-dependent shank2 methylation as a key player in mediating reciprocal switching between invasion and proliferation, also point to the value of shank2 R240 methylation as a target for tumour metastasis treatment strategies.

To explore the mechanism of PRMT7 action in tumour metastatic progression, we used the 148 Flag-tagged PRMT7 fusion protein as a bait in mass spectrometry to pick up the possible 149 regulatory factors that are directly regulated by PRMT7 (data not shown). Among these 150 PRMT7-interacting proteins, we chose shank2 as a target for a more detailed study, since it 151 plays a critical role in cytoskeleton organization. Although shank1 and shank3 have been 152 reported to inhibit cancer cell migration, whether shank2 plays a role in cancer progression is 153 unknown. We first conducted a co-immunoprecipitation assay to confirm the binding between 154 PRMT7 and shank2 in vitro ( Figure 1A,B). We then incubated the purified GST-tagged 155 PRMT7 with shank2 protein, and the results showed that they could directly interact with each 156 other ( Figure 1C). To map the shank2 regions that bind PRMT7, we expressed a series of 157 truncated GST-shank2 ( Figure 1D), and we detected that the shank2 ANK, PDZ and SAM 158 domains were responsible for the interaction between PRMT7 and shank2 ( Figure 1E). By 159 using in vivo methylation assay, we found that overexpression of PRMT7 resulted in 160 increased symmetric arginine di-methylation of shank2 ( Figure 1F). Meanwhile, shank2 161 di-methylation was reduced upon PRMT7 depletion, and the reduced di-methylation was

164
To further confirm PRMT7 directly methylates shank2, we used purified GST-PRMT7 and 165 Flag-hank2 for in vitro methylation assay followed by incubation with SAM (the methyl donor).

166
Apparently, incubation of recombinant PRMT7 and shank2 gave rise to a remarkable 167 increase in methylation of shank2 in the presence of SAM, while shank2 methylation was 168 undetectable in the absence of SAM or PRMT7 ( Figure 1I). These results suggest that 169 PRMT7 physically interacts with and methylates shank2. Previous reports showed that the expression of shank2 in head and neck cancer tissues was 173 higher than that in adjacent tissues, and this upregulation was correlated with the survival 174 rate and prognosis of patients (Qin et al., 2016). According to the Human Protein Atlas 175 database, the expression of shank2 in breast cancer samples is higher than that in the 176 adjacent tissues (Figure 2A). We also observed that shank2 mRNA was highly expressed in  189 that overexpressed Rat HA-shank2 (a gift from Dr. Min Goo Lee). We detected that the 190 arginine 240 residue (R240) of shank2 was di-methylated ( Figure 3A). Using sequence 191 homology comparison, we found that the R240 residue is evolutionarily conserved ( Figure   192 3B). Then we constructed a shank2 mutation with R240 residue replaced by lysine (shank2 193 R240K) to test the effect of shank2 R240 methylation. We found that shank2 di-methylation 194 was decreased in shank2 R240K, whereas shank2 WT did not affect the di-methylation level 195 ( Figure 3C,D). We then overexpressed increasing doses of PRMT7 in HEK293T cells,and 196 we found that shank2 R240 di-methylation level was increased accordingly ( Figure 3E, shank2 to methylate the R240 of the protein.

199
Next, we intended to study the effects of shank2 di-methylation on breast cancer cells. To explore the mechanism of shank2 function in breast cancer cell migration and invasion,

218
we studied the conformational changes of wild-type shank2 (WT) and its mutant (R240K) by 219 using molecular dynamics simulations. As shown in Figure 4A, the SPN domain in shank2 220 R240K mutant needs more energy to reach an "open" state than that in shank2 WT,

221
according to the free energy profile. The trajectory for the window with the global minimum 222 free energy (the distance between SPN and ANK is 31.5 Å, the SPN and ANK domains were 223 represented in different colours in Figure 4B) in WT and the one (the distance between SPN 224 and ANK is 30 Å) in R240K were analysed. The average number of hydrogen bonds between 9 the two domains was 0.96±0.05 in each frame of WT. The most hydrogen bonds were formed 226 in N100/R227 and E119/K260 for the locations of these residues ( Figure 4C). After R240 was 227 mutated to K, two domains formed 1.57±0.02 hydrogen bonds in D97/K229, Y140/H192, 228 N100/R227, Y140/N190 and S256/E119 for the locations of mentioned residues ( Figure 4D).

229
Thus, structural analyses suggest that shank2 R240K mutants not only possess a more 230 stable energy fluctuation, but also form more hydrogen bonds, causing the loss of 231 appropriate space for insertion of partner proteins.

381
In conclusion, data presented in this report outline a working model in which PRMT7 where all the cell lines were characterized by DNA finger printing and isozyme detection. Cells

394
were immediately expanded and frozen such that they could be revived every 3 to 4 months.

395
MCF10A cells were cultured as previously described (Geng et al., 2017). BT549 cells were   which were designed to target human genes, were described below.

417
Western blotting and Immunoprecipitation was performed as described previously (Geng et al.,

424
The β-actin was used as an internal control. The sequences of the primers used in this study 425 are described below.  These experiments were essentially performed as described previously (Qian et al., 2013).

433
The procedure was performed basically as described previously (Gil-Henn et al., 2013).