SPIN1 promotes tumorigenesis by blocking the uL18-MDM2-p53 pathway

Ribosomal proteins (RPs) play important roles in modulating the MDM2-p53 pathway. However, less is known about the upstream regulators of the RPs. Here we identify SPIN1 (Spindlin 1) as a novel binding partner of human RPL5/uL18 that is important for this pathway. SPIN1 ablation activates p53, suppresses cell growth, reduces clonogenic ability, and induces apoptosis of cancer cells by sequestering uL18 in the nucleolus, preventing it from interacting with MDM2, and thereby alleviating uL18-mediated inhibition of MDM2 ubiquitin ligase activity towards p53. SPIN1 deficiency increases ribosome-free uL18 and uL5 (human RPL11), which are required for SPIN1 depletion-induced p53 activation. Analysis of cancer genomic databases suggests that SPIN1 is highly expressed in several human cancers, and its overexpression is positively correlated with poor prognosis in cancer patients. Altogether, our findings reveal that the oncogenic property of SPIN1 is highly attributed to its negative regulation of uL18, leading to p53 inactivation.

The well-documented tumor suppressor p53, referred as "the guardian of the genome", is 2 activated upon exposure to a myriad of cellular stresses. While loss of wild type p53 causes fatal 3 damages to cells, it is not surprising that the TP53 gene is mutated in more than 50% human 4 cancers, and the functions of p53 are often impeded through various mechanisms in the 5 remainder 1 . One predominant negative regulator of p53 is the E3 ubiquitin ligase MDM2, an 6 oncoprotein that conceals the N-terminal transcriptional activation (TA) domain of p53 2 and 7 deactivates this protein by either abrogating its transcriptional activity, or inducing its nuclear 8 export and ubiquitination 2-5 . A plethora of cellular stress could stabilize p53 by blocking the 9 MDM2-p53 feedback loop 6 . For example, p19 ARF inhibits MDM2-mediated p53 ubiquitination 10 and proteolysis by associating with MDM2 7 .

11
Another pathway is so-called the ribosomal proteins (RPs)-MDM2-p53 pathway 8,9 . 12 Accumulating evidence has continuingly verified this pathway as an emerging mechanism for 13 boosting p53 activation in response to ribosomal stress or nucleolar stress over the past decade 10-14 14 . Ribosomal stress is often triggered by aberrant ribosome biogenesis caused by nutrient 15 deprivation, inhibition of rRNA synthesis or malfunction of ribosomal proteins or nucleolar 16 proteins [8][9][10][11]15,16 . Earlier studies showed that disruption of ribosomal biogenesis induces 17 translocation of a series of ribosomal proteins, including uL18 (human RPL5), uL5 (human 18 RPL11), uL14 (human RPL23), eS7 (human S7) and uS11 (human S14) 17 , from the nucleolus to 19 the nucleoplasm and bind to MDM2, blocking its ability to ubiquitinate p53 and consequently 20 stabilizing p53 to maintain cellular homeostasis 12,18-23 . Although there are a few proteins that 21 have been identified to regulate this RPs-MDM2-p53 pathway, such as PICT-1 inhibition of uL5 22 24,25 and SRSF1 inhibition of uL18 26 , it still remains to determine if there are more proteins that 23 can regulate the RPs-MDM2-p53 pathway. In this present study, we identified SPIN1 as another 24 uL18 regulator. 25 SPIN1, a new member of the SPIN/SSTY family, was originally identified as a highly expressed 26 protein in ovarian cancer 27 . The oncogenic potential of SPIN1 was later supported by the 27 observation that overexpression of SPIN1 increases transformation and tumor growth ability of 28 NIH3T3 cells 28 . Signaling pathways responsible for SPIN1 functions include PI3K/Akt, Wnt 29 and RET that are highly relevant to tumorigenesis [29][30][31] . In addition, SPIN1 acts as a reader of 30 histone H3K4me3 and stimulates the transcription of ribosomal RNA-encoding genes 32-34 , 1 suggesting its role in rRNA synthesis. 2 In screening uL18-associated protein complexes using co-immunoprecipitation followed by mass 3 spectrometry, we identified SPIN1 as one of the potential uL18 binding proteins. We confirmed 4 the specific interaction of SPIN1 with uL18, but not with uL5 or uL14, and also found out that 5 by binding to uL18, SPIN1 prevents the inhibition of MDM2 by uL18 and promotes MDM2-6 mediated p53 ubiquitination and degradation. Also, SPIN1 knockdown induced ribosomal stress 7 by facilitating the release of ribosome-free uL18 or uL5, accompanying p53 activation. 8 Furthermore, SPIN1 knockdown inhibited cell proliferation and induced apoptosis in a 9 predominant p53-dependent manner in vitro and in vivo, consequently suppressing tumor growth 10 in a xenograft model. Therefore, these results for the first time demonstrate that SPIN1 can 11 regulate the RP-MDM2-p53 pathway by directly interacting with uL18, and suggest SPIN1 as a 12 potential molecule target in this pathway for developing anti-cancer therapy in the future.

15
SPIN1 interacts with uL18. 16 Our and others' studies previously demonstrated that uL18 can stabilize p53 by binding to 17 MDM2 and inhibiting its E3 ligase activity toward p53 19,35 . In order to identify potential 18 upstream regulators that may modulate the uL18-MDM2-p53 circuit, we performed co-19 immunoprecipitation (co-IP) using HEK293 cells that stably expressed Flag-uL18 with the anti-20 Flag antibody, and the co-immunoprecipitated proteins were cut out for mass spectrometric (MS) 21 analysis (Fig. 1A). The MS results not only revealed several previously described p53 regulatory 22 proteins, such as MYBBP1A, PRMT5 and SRSF1, as uL18 binding proteins (Table S1), but also 23 identified SPIN1 as a novel uL18-binding protein candidate that was previously shown to play a 24 role in tumorigenesis and rDNA transcription 30,34 .

25
Next, we confirmed the interaction between SPIN1 and uL18 by performing a series of 26 reciprocal co-IP assays. As expected, ectopic SPIN1 was specifically pulled down by ectopic 27 uL18 and vice versa in HCT116 p53-/cells ( Fig. 1B and 1C). Their interaction was also verified in 28 HEK293 cells (Supplementary Fig. S1). Also, we validated the interaction between endogenous 29 SPIN1 and uL18 in HEK293 cells using anti-SPIN1 antibody (Fig. 1D). Interestingly, only uL18, 30 but not uL5, was co-immunoprecipitated with SPIN1. In line with this result, when comparing 31 ectopic Flag-uL18 with Flag-uL5 and Flag-uL14, we found that only uL18, but not the other RPs, 1 could pull down Myc-SPIN1 (Fig. 1E), further bolstering the specific interaction between uL18 2 and SPIN1. Taken together, these results demonstrate that SPIN1 specifically binds to uL18, but 3 not uL5 or uL14, in cancer cells. Previous and recent studies showed that SPIN1 is a potential oncogene 29,30,36 , and uL18 can 8 stabilize p53 by binding to MDM2 19 . We therefore wondered if the interaction between SPIN1 9 and uL18 could confer any role to SPIN1 in regulation of the p53 pathway. First, we determined 10 if depletion of SPIN1 might affect p53-dependent cellular outcomes. Interestingly, we found that 11 knockdown of SPIN1 dramatically elevates p53 protein level in several wild type p53-containing 12 cells, including U2OS, H460 and HCT116 p53+/+ cells ( Fig. 2A), without affecting p53 mRNA 13 expression (Fig. 2B). Consistently, protein and mRNA levels of p53 target genes, such as p21 14 and PUMA, were also increased in response to SPIN1 knockdown ( Fig. 2A and 2B). Moreover, 15 the effects of SPIN1 siRNA on p53 activity were dose-dependent ( Supplementary Fig. S2). 16 Conversely, overexpression of SPIN1 in HCT116 p53+/+ decreased the protein levels of p53 and its 17 targets, such as p21 and PUMA, and the mRNA levels of these target genes, without affecting 18 p53 mRNA level ( Fig. 2C and 2D). 19 We next generated both HCT116 p53+/+ and HCT116 p53-/cell lines that express scramble shRNA 20 or SPIN1 shRNA to evaluate biological outcomes of SPIN1 knockdown. As illustrated in Fig. 2E, 21 the expression of p53 and some of its target genes were markedly induced when SPIN1 was 22 knocked down by its specific shRNA in HCT116 p53+/+ cells, but not in HCT116 p53-/cells. Using 23 these cell lines for cell viability assays, we observed that SPIN1 ablation more dramatically 24 represses the cell viability of HCT116 p53+/+ than that of HCT116 p53-/cells (Fig. 2F). In line with 25 this observation, SPIN1 depletion also led to more predominant reduction of HCT116 p53+/+ 26 colonies than that of HCT116 p53-/colonies, though both of the reductions were statistically 27 significant (Fig. 2G). Furthermore, percentage of cells undergoing apoptosis caused by SPIN1 28 shRNAs was much higher in HCT116 p53+/+ cells than in HCT116 p53-/cells, as measured by sub-29 G1 population (Fig. 2H). Notably, these effects were proportional to the efficiency of SPIN1 30 knockdown by two different shRNAs, indicating that the observed effects are highly related to 31 SPIN1 levels. Collectively, these data suggest that SPIN1 plays an oncogenic role by 1 predominantly inactivating the p53 pathway, although SPIN1 may also possess a p53-2 independent role in cancer cell growth and survival.
3 4 SPIN1 promotes p53 degradation by enhancing MDM2-mediated ubiquitination. 5 Since SPIN1 knockdown affected only the protein, but not the mRNA, levels of p53 ( Fig. 2A-6 2D), we next sought to determine the underlying mechanism. We first performed a 7 cycloheximide-chase experiment using HCT116 p53+/+ cells. As shown in Figures 3A and 3B, 8 knockdown of SPIN1 markedly prolonged p53's half-life from ~30 mins to ~60 mins, as 9 compared to scramble siRNA. Inversely, ectopic SPIN1 greatly shortened p53's half-life, from 10 ~4 0 mins to ~20 mins ( Fig. 3C and 3D). To further evaluate the effect of SPIN1 on MDM2-11 mediated p53 ubiquitination, which is the main mechanism responsible for p53 turnover 12,19,22,37 , 12 we then performed an in vivo ubiquitination assay by transfecting HCT116 p53-/cells with 13 plasmids indicated in Figure 3E. The results clearly showed that ectopic SPIN1 enhances 14 MDM2-mediated p53 ubiquitination in a dose-dependent manner. Consistently, co-transfection 15 of SPIN1 with MDM2 led to a stronger reduction of p53 protein levels, which was abrogated by 16 proteasome inhibitor MG132 (Fig. 3F). Interestingly, the induction of p53 degradation by SPIN1  where it binds to MDM2 19,38 , leading to stabilization of p53 and consequently p53-dependent 27 cell growth arrest, apoptosis or senescence. We then investigated if SPIN1 might regulate this 28 function of uL18, since SPIN1 could bind to uL18 (Figure 1), knockdown of SPIN1 led to p53 29 activation (Fig. 2), and SPIN1 inhibited MDM2-mediated p53 ubiquitination (Fig. 3). First, as 30 expected 19 , overexpression of uL18 induced the protein levels of p53 and its targets, such as p21 31 and MDM2, in wild type p53-containing U2OS cells (Fig. 4A). This induction of the p53 1 pathway by uL18 was markedly reduced by co-transfected SPIN1 (Fig. 4A). Since the effect of 2 uL18 on p53 is through uL18's interaction with MDM2 and consequent inhibition of its E3 3 ligase activity towards p53 19 , we tested if SPIN1 may affect uL18-MDM2 interaction. 4 Interestingly, our co-immunoprecipitation result showed that ectopic Myc-SPIN1 dramatically 5 reduces the amount of Flag-uL18 co-immunoprecipitated with HA-MDM2 in a dose-dependent 6 manner, although Myc-SPIN1 itself did not co-immunoprecipitate with HA-MDM2 ( Fig. 4B and 7 S3A). This effect was specific to the uL18-MDM2 interaction, as Myc-SPIN1 overexpression 8 did not alter the interactions between uL5 and MDM2 (Fig. 4C). Our immunofluorescence result 9 ( Fig. 4D and S3B) showed that SPIN1 and uL18 are clearly co-localized in the nucleolus, 10 suggesting that SPIN1 might sequester uL18 in the nucleolus and thus prevent it from binding 11 and inactivating MDM2 in the nucleoplasm. Taken together, these results demonstrate that 12 SPIN1 is a regulator of the uL18-MDM2-p53 pathway, acting by preventing uL18 from 13 interaction with MDM2.
14 15 SPIN1 depletion also causes ribosomal stress, activating p53. 16 Previous studies showed that SPIN1 could recognize H3K4 methylation and stimulate rRNA 17 gene expression, unveiling its role in rRNA synthesis 32,34 . Disruption of rRNA synthesis leads to 18 disassembly of ribosomal precursors and release of ribosome-free ribosomal proteins from the 19 nucleolus 16,19,22 . Based on these lines of information, we speculated that dysregulation of SPIN1 20 itself might also impact ribosome biogenesis, resulting in accumulating ribosome-free ribosomal 21 proteins to activate p53. To test this speculation, we first carried out a sucrose gradient 22 fractionation assay using scramble-and SPIN1-shRNA transfected HCT116 p53+/+ cells. The 23 collected fractions were subjected to Western blot (WB) analysis. As anticipated, the levels of 24 uL18 and uL5 in the soluble and ribosome-unbound fractions were markedly increased in 25 SPIN1-depletion cells, accompanying with elevated p53 and MDM2 protein levels (Fig. 5A).

26
Interestingly, the binding between endogenous uL18/uL5 and MDM2 increased upon SPIN1 27 knockdown, resembling ribosomal stress (Fig. 5B). Indeed, as expected, knockdown of SPIN1 28 reduced the expression of pre-rRNA and rRNA ( Supplementary Fig. S4). Moreover, as clearly 29 illustrated in Figure 5C, overexpression of SPIN1 compromised p53 activation induced by 30 actinomycin D or 5-Fu treatment, which was reported to trigger ribosomal stress that in turn 31 triggers the formation of RPs-MDM2 complex 10,19,23,39 . To further confirm the role of these free 1 forms of ribosomal proteins in SPIN1 ablation-induced p53 activation, we knocked down uL18 2 or uL5 using siRNA with or without SPIN1 depletion in U2OS cells. Strikingly, the reduction of 3 either uL18 or uL5 abrogated SPIN1 knockdown-induced p53 levels, as well as its target p21, as 4 compared to scramble siRNA-transfected cells ( Fig. 5D and 5E). Collectively, these data indicate 5 that knockdown of SPIN1 could also lead to ribosomal stress, releasing ribosome-free uL18 and 6 uL5, which are required for p53 activation induced by SPIN1 depletion. To understand the physical interaction between SPIN1 and uL18 in more details, we generated   28 To translate the above-described cellular functions of SPIN1 into more biological significance, 29 we established a xenograft tumor model by inoculating the aforementioned HCT116 (both 30 p53+/+ and p53-/-) cell lines that expressed scramble shRNA or SPIN1 shRNA into NOD/SCID 31 mice, and monitored tumor size for 18 days. As illustrated in Fig. 7A and 7B, SPIN1 knockdown 1 more markedly slowed down the growth of xenograft tumors derived from HCT116 p53+/+ cells 2 than that from HCT116 p53-/cells. Notably, SPIN1 depletion also reduced the growth of tumors 3 derived from HCT116 p53-/cells, suggesting that SPIN1 might possess a p53-independent 4 function required for cancer cell growth. In line with the tumor growth curve, the reduction of 5 tumor mass and weight by SPIN1 knockdown was more profound in HCT116 p53+/+ groups (~60% 6 reduction in weight) than that in HCT116 p53-/groups (~30% reduction in weight) ( Fig. 7C and   7 7D). To confirm our cell-based findings, we performed qRT-PCR and WB analysis using the 8 xenograft tumors. As expected, the mRNA levels of p21 and PUMA were significantly 9 upregulated upon SPIN1 knockdown in HCT116 p53+/+ , but not in HCT116 p53-/tumors ( Fig. 7E 10 and 7F). Consistently, the protein levels of p53 and its target PUMA were elevated in 11 HCT116 p53+/+ groups, but in HCT116 p53-/groups ( Fig. 7G and Supplementary Fig. S6). Taken 12 together, these results demonstrate that SPIN1 depletion retards tumor growth by mainly 13 activating p53, though SPIN1 might also possess p53-independent functions in regulation of cell 14 growth and survival.

15
The data presented above suggest that SPIN1 is required for tumorigenesis. Therefore, we further 16 searched some available genomic and gene expression database for SPIN1 expression in cancers.

17
Interestingly, our analysis of TCGA genome database 40,41 indicated that the SPIN1 gene is  The tumor suppressor p53 provides a critical brake on cancer development in response to 30 ribosomal stress, as impairing this ribosomal stress-uL18/uL5-p53 pathway could accelerate 31 tumorigenesis in c-Myc transgenic lymphoma mice 45 . However, it remains largely elusive 1 whether this pathway is subjected to the regulation by other yet unknown proteins. In our attempt 2 to understand molecular insights into this possible regulation, we identified SPIN1, the nucleolar 3 protein important for rRNA synthesis 34 , as a novel regulator of the uL18-MDM2-p53 pathway 4 through interplay with uL18 (Fig. 6E). Our studies as presented here provide the first line of 5 evidence for that SPIN1 acts as an upstream regulator of uL18's accessibility to MDM2 for p53 6 activation.

7
Using IP-MS analysis, we identified SPIN1 as a new uL18-associated protein (Fig. 1A). Our 8 biochemical and cellular experiments using co-IP and GST pull down assays further validated 9 the direct association of SPIN1 with uL18 ( Fig. 1B-1D; Fig. 6A-6D). Moreover, we found that 10 SPIN1 and uL18 co-localized in the nucleolus by immunofluorescence assay (Fig. 4D and S3B).

11
Remarkably, SPIN1 specifically binds to uL18, but not uL5 and uL14, as no binding was 12 detected between ectopic SPIN1 and uL5 or uL14 by co-IP (Fig. 1E). Interestingly, SPIN1 does 13 not appear to bind to MDM2, as it was not co-immunoprecipitated with MDM2 either (Fig. 4B).
14 Although our previous reports described a complex of uL18/uL5/uL14-MDM2 12 , our present 15 findings indicate that SPIN1 may work in a separate complex with uL18 differed from reported 16 RPs-MDM2 complexes. Also our results suggest that SPIN1 may retain uL18 in the nucleolus so 17 that the latter is unable to shuttle to the nucleoplasm and to inhibit MDM2 activity towards p53 presented here showed that SPIN1 depletion by its specific shRNA leads to the augment of the 25 p53-dependent cancer cell growth arrest and apoptosis. This is at least partly because SPIN1 can 26 promote MDM2-dependent ubiquitination and degradation of p53 (Fig. 3), which is highly likely 27 attributable to its capability to prevent uL18 from binding to MDM2 through retaining uL18 in 28 the nucleolus (Fig. 4E). 29 Also, knockdown of SPIN1 led to the increase of ribosome-free uL18 and uL5 levels, of the 30 uL18/uL5-MDM2 complex, and of p53 level and activity ( Fig. 5A and 5B). The activation of 31 p53 by knocking down SPIN1 is due to the ribosomal stress caused by the depletion of this 1 nucleolar protein, as SPIN1 is required for rRNA synthesis by RNA polymerase I 34 . Also, 2 consistent with these observations, overexpression of SPIN1 reduced the activation of p53 by 3 Actinomycin D treatment (Fig. 5C), whereas knockdown of uL18 or of uL5 impaired the 4 activation of p53 by SPIN1 knockdown (Fig. 5D and 5E). Several genes have been implicated to 5 modulate the RPs-MDM2-p53 pathway through interplay with ribosomal proteins 24,25,48-50 . In 6 particular, SRSF1 was identified as a component of the RP-MDM2-p53 complex, and could 7 stabilize p53 via uL18 26 . Different from their studies, SPIN1 specifically forms an independent 8 complex with uL18, but not MDM2 or other ribosomal proteins, such as uL5 or uL14, and acts 9 as a negative regulator of p53. Therefore, our present findings unveil a novel mechanism for 10 suppression of the uL18-MDM2-p53 pathway by SPIN1, whose depletion consequently leads to 11 p53-dependent cell growth inhibition and apoptosis. predominantly in wild type p53-containing cancer cells (Fig. 2). Remarkably, knockdown of 20 SPIN1 inhibited xenograft tumorigenesis derived from human colon cancer cells, which was 21 much more significantly in HCT116 p53+/+ cells than in HCT116 p53-/cells (Fig. 7). These results 22 demonstrate that SPIN1 can promote tumor growth and survival by inactivating p53 and its 23 pathway (Fig. 6E). 24 Intriguingly, we also found that SPIN1 ablation had a moderate inhibitory effect on cell growth 25 in p53-null HCT116 cells in vitro and in vivo as mentioned above (Fig. 2 and 7). These findings 26 suggest that SPIN1 might also possess p53-independent oncogenic effects, which might be 27 explained by two possible mechanisms. First, SPIN1 has been reported to execute its oncogenic Recent studies have demonstrated the role of SPIN1 in rRNA transcription 32,33 , which provides a 3 clue that dysregulation of SPIN1 may perturb ribosome biogenesis. In fact, in our current study, 4 we observed that SPIN1 deletion per se increases the levels of ribosome-free uL18 and uL5, 5 accompanying elevated p53 protein levels, which recapitulates the effects of ribosomal stress. 6 Our observation that p53 induction caused by SPIN1 depletion could be abrogated by 7 knockdown of either uL18 or uL5 further supports this hypothesis. Therefore, while it is 8 conceivable that SPIN1 counteracts p53 by blocking the interaction between uL18 and MDM2 as 9 discussed above, the mechanism by which disruption of SPIN1 causes ribosomal stress may be 10 also responsible for p53 activation.

11
In summary, our findings unveil SPIN1 as another novel and important regulator of the MDM2-12 p53 pathway by predominantly inhibiting the association of uL18 with MDM2 to modulate p53 13 activity (Fig. 6E) and provide molecular insights into the fine regulation of this pathway as well 14 as a potential target for the future development of an anti-cancer therapy.  Plasmids and antibodies.

31
The Myc-tagged SPIN1 plasmid was generated by inserting the full-length cDNA amplified by 1 PCR into the pcDNA3.1/Myc-His vector at EcoR I and Bam HI, using the following primers, 2 forward-CGGAATTCatgaagaccccattcggaaag; reverse-CGGGATCCggatgttttcaccaaaatcgtag.     Generating stable cell lines 31 Briefly, scramble shRNA or SPIN1 shRNAs purchased from Sigma were transfected into 1 HCT116 p53+/+ and HCT116 p53-/cells using TurboFect reagent. The cells were maintained at 37°C 2 in a 5% CO 2 humidified atmosphere for 48 h and were split to two aliquots, one for WB analysis 3 and the other for selection using final concentration of 2 µg/ml puromycin in growth medium. The Student's two-tailed t-test was used to determine mean difference among groups. P<0.05 22 was considered statistically significant, asterisks represent significance in the following way: *, 23 p<0.05; **, p<0.01. The term "n.s" indicates that no significant difference was found. All the 24 data are presented as mean ± SEM.   The mRNA levels of SPIN1, p53 and p53 target genes were detected in 6 tumors by RT-qPCR 8 (mean ± SEM, n=6). (g) The protein levels of SPIN1, p53 and p53 targets were detected in 6 9 tumors samples by WB analysis with indicated antibodies. *p<0.05, **p<0.01 by two-tailed t-10 test (d, e, f, g).