FBXO7 ubiquitinates PRMT1 to suppress serine synthesis and tumor growth in hepatocellular carcinoma

Cancer cells are often addicted to serine synthesis to support growth. How serine synthesis is regulated in cancer is not well understood. We recently demonstrated protein arginine methyltransferase 1 (PRMT1) is upregulated in hepatocellular carcinoma (HCC) to methylate and activate phosphoglycerate dehydrogenase (PHGDH), thereby promoting serine synthesis. However, the mechanisms underlying PRMT1 upregulation and regulation of PRMT1-PHGDH axis remain unclear. Here, we show the E3 ubiquitin ligase F-box-only protein 7 (FBXO7) inhibits serine synthesis in HCC by binding PRMT1, inducing lysine 37 ubiquitination, and promoting proteosomal degradation of PRMT1. FBXO7-mediated PRMT1 downregulation cripples PHGDH arginine methylation and activation, resulting in impaired serine synthesis, accumulation of reactive oxygen species (ROS), and inhibition of HCC cell growth. Notably, FBXO7 is significantly downregulated in human HCC tissues, and inversely associated with PRMT1 protein and PHGDH methylation level. Overall, our study provides mechanistic insights into the regulation of cancer serine synthesis by FBXO7-PRMT1-PHGDH axis, and will facilitate the development of serine-targeting strategies for cancer therapy.

The study unveils a novel mechanism involving FBXO7, an E3 ubiquitin ligase, which modulates PRMT1 ubiquitylation.The data provides solid biochemical evidence of a novel regulation by FBXO7 of PRMT1 stability.Nonetheless, a key weakness lies in the limited biological significance of the observed effects, both in in vivo tumor growth and in vitro cell proliferation, which appear marginal and inconclusive.This raises questions regarding the broader biological implications of the described biochemical axis.
In addition to this, several specific points should be addressed: -The biological consequences of FBXO7 are not clear.A thorough analysis of the biological implications of the observed GSH and ROS imbalances within the biological system would be essential.Questions that need addressing include whether FBXO7-overexpressing cells exhibit cell death and cell cycle arrest due to ROS accumulation, whether overexpression of a PRMT1 K37R mutant can rescue these effects, and whether supplementation of antioxidants in vitro can restore the observed effects.
-In addition to the point above, addressing the biological significance in vivo, there will still be uncertainty in the in vivo effects, who appeared generally marginal.
The study lacks a clear elucidation of the causative relationships within the FBXO7/PRMT1 axis.Assessing how FBXO7 overexpression impacts tumor growth and whether PRMT1 K37R can rescue the effects of FBXO7 overexpression in vivo is critical.
-A notable gap in the study is the absence of a comprehensive metabolic analysis to directly assess the effects of FBXO7.Understanding whether the alterations in serine/glycine biosynthesis result from direct or indirect consequences on alternative metabolic pathways is crucial.Conducting global metabolomics and metabolic tracing experiments, especially with marked glucose in serine/glycine-deprived and non-deprived media, could provide valuable insights into global metabolic changes and possibly open to broader biological significance.
-Investigating the effects of PRMT1 K37R on PHGDH methylation and activity and whether it preserves or restores the effects of FBXO7 overexpression would also provide stronger evidence in support of the described axis.
In conclusion, while the study introduces a novel regulatory aspect of FBXO7 on PRMT1 ubiquitylation, its limited biological significance and gaps in the mechanistic understanding warrant further investigation.

Reviewer #3 (Remarks to the Author):
In this paper on serine synthesis regulation in cancer, Luo et al. describe how F-box-only protein 7 (FBXO7), an E3 ubiquitin ligase, inhibits serine synthesis by promoting the ubiquitin-proteasome degradation of protein arginine methyltransferase 1 (PRMT1) in hepatocellular carcinoma (HCC).Previously, they had already shown that PRMT1 is upregulated in HCC.PRMT1 methylates phosphoglycerate dehydrogenase (PHGDH) at arginine 236, thereby activating PHGDH and promoting serine synthesis.Nevertheless, the mechanisms underlying overexpression of PRMT1 and regulation of the PRMT1-PHGDH axis in HCC remains unknown.
Here, the authors unravel the details of this mechanism.In a series of elegant experiments including all relevant controls, the authors convincingly show that FBXO7 directly interacts with PRMT1, leading to the lysine 37 ubiquitination and subsequent degradation of PRMT1.FBXO7-mediated PRMT1 downregulation then decreases the extent to which PHGDH is methylated on its target arginine residue, resulting in impaired serine synthesis, accumulation of reactive oxygen species (ROS) levels, and inhibition of HCC cell growth.
The authors have gathered a substantial amount of evidence to support these claims, including in vitro biochemical experiments and clinical data, further strengthening their hypothesis.
Overall, the design of the experiments and the levels of detail are very solid.The data convincingly show that FBXO7 suppresses PHGDH R236 methylation and catalytic activity by downregulating PRMT1 in HCC cells.
I have no major comments, only a few minor comments listed below in bullet points: overexpressed in HCC to promote tumor growth […]".Is it really overexpressed or is it just degraded to a lesser extent upon expression?As the mRNA levels seem not to change, I assume it is the latter.
-P5: Although the authors have used a smart in silico screening to identify interaction partners that ultimately provided them the link to FBXO7, I'm wondering why the authors have not performed an unbiased interaction screen based on affinity purification coupled to mass spectrometric identification of interaction partners themselves.The advantage would be that PRMT1 interactions in the relevant cell system (Huh7 and PLC/PRF/5) could be studied this way.I'm not sure whether BioGRID contains interactome data for this specific cell line.
-P7: "[…] as shown in Figure 2A, the upregulated protein level of PRMT1 caused by FBXO7 KD was not observed in MG132-treated HCC cells […]".The Western blots for the PLC/PRF/5 cells are convincing in this respect, but the Huh7 less so.It's hard to see the lack of upregulated levels in the MG132 treated FBXO7 KD cells.Have the authors considered more quantitative approaches to determine PRMT1 and FBXO7 levels, such as quantitative targeted mass spectrometry (PRM)?-P8: In Fig 3C , although the high MW smear as observed in the WT lane is certainly not observed in the K37R cells, the intensity of the bands / smear in the HA IB seems to be increased in the K37R mutant with respect to the WT (minus FLAG-FBXO7).Do the authors have an explanation for this?Also, even though the ubiquitination of the protein results in a smear on the blot, distinct bands are observed.Is it possible, based on these results, to hypothesize on the substrate being mono-or poly-ubiquitinated?-In the fragmentation mass spectrum in Fig 3F , no (or, at most, noise intensity level) fragment ion peaks are observed around the Lysine residue with the diGly remnant.However, given the fact that the K-D bond is not cleaved and the total peptide mass corresponds to that of the modified peptides, the data convincingly shows that K37 is ubiquitinated.Was the PHGDH R236 methylation site also confirmed by MS, like the K37 ubiquitination site of PRMT1?-In Fig 4G, the statistical (in)significance indicators for the shPRMT1 shScramble vs shFBXO7 are missing (idem for Figs 5A and B).
-Have the authors considered including experiments with a mutant PHGDH that lacks the R236 target methylation site?If so, why were these not included in the manuscript?-The model in Fig 7M suggests that low levels of FBXO7 lead to ROS accumulation (I suppose that the spiked balloon with the text 'ROS' means 'accumulation of ROS').However, in Fig 5H, the authors show that the depletion of FBXO7 causes a reduction of ROS levels and that the depletion of PRMT1 leads to increased ROS levels.Also, ectopic FBXO7 expression promotes ROS accumulation in control cells and has no obvious effect in PRMT1 KD cells (Fig 5I).This suggests that FBXO7 induces oxidative stress by downregulating PRMT1 in HCC cells.Therefore, the description of the results and the cartoon in  In our study, shFBXO7 #2 (targeting 3'UTR) was used in Fig. 2c, 3c, and 3e, as well as Fig. 4,Fig. 5,and Fig. 6.The knockdown effectiveness of shFBXO7 #2 was shown in Fig. 2a and 2c.We also repeated Fig. 3c and detected FBXO7 level in whole cell lysates to demonstrate the efficiency of FBXO7 KD as well as FBXO7 rescue in Huh7 cells.
Similarly, the shPRMT1 was also corrected as shPRMT1 #2 (targeting 3'UTR) in the related figures.2. Figure 3F is uninterpretable as it is too small with no guide in the text or legend as to which peak / fragment we should be looking at.Have they conducted quantitative MS in any of their conditions to determine the relative change in ubiquitination, as a second modality other than immunoblotting?
Response: We apologize for the small size of Fig. 3f with no guide.This image has been optimized to make the b-ions and y-ions clearer (Fig. 3f).The ubiquitinated lysine 37 (K37, y5 1+ ion with an additional shift of 114.1 Da) was highlighted.We also performed parallel reaction monitoring (PRM)-based quantitative proteomics approach to quantitate PRMT1 K37 ubiquitination level following FBXO7 overexpression.Unfortunately, we failed to identify K37 ubiquitination of PRMT1 protein during PRM analysis probably due to the relatively low abundance of PRMT1 protein for PRM proteomics analysis.Because of the quantitative purpose, antibodybased affinity purification and enrichment of PRMT1 protein is not allowed in PRM analysis (Fig. R1, workflow 3 for PRM analysis).Compared with the workflows for ubiquitination site identification with antibody enrichment of PRMT1 protein (Fig. R1, workflow 1 and workflow 2), relatively low PRMT1 protein level was subjected to PRM proteomics analysis, which might lead to the failure for the recognition of PRMT1 K37 ubiquitination by mass spectrometry.
In addition, our in vivo ubiquitination assay combined with lysine mutation convincingly demonstrated that FBXO7 mediates PRMT1 K37 ubiquitination and promotes ubiquitin-mediated degradation of PRMT1 in HCC cells (Fig. 2 and 3).Response: We apologize for this confusing diagram.According to your kind suggestion, we have simplified this schematic model to make it clearer that FBXO7 promotes the ubiquitin-mediated degradation of PRMT1, leading to decreased PHGDH methylation.
In addition, we have corrected 'FBOX7' as 'FBXO7' (Supplementary Fig. 8  5. Could they conduct further analysis of publicly-available datasets to determine which transcriptional subtype of HCC has elevated expression of FBXO7 or PRMT1?One could hypothesize that this would be limited to one sub-group of patients and this data would suggest potential stratification of subsequent trials targeting this pathway. Response: Thanks for your valuable suggestions.Given that the tumor suppressor gene TP53 was mutated in 31% of HCC patients, and p53 target gene expression signature (p53 signature) correlating with patient survival was recently developed (Cancer Genome Atlas Research Network.Cell.2017;169(7):1327-41).We thus stratified 186 HCC patients in TCGA database by p53 signature, which showed that patients with low p53 signature exhibited reduced overall survival (Supplementary Fig. 7c).Interestingly, the p53 signature was positively correlated with the mRNA level of FBXO7 (Supplementary Fig. 7d), suggesting that FBXO7 expression may be linked with the sub-group of patients with different p53 signature.Therefore, therapeutic strategies targeting FBXO7-PRMT1-PHGDH axis might be applicable for HCC patients with low p53 signature, which requires further investigation (Page 12, Lines 27-29; Page 13, Lines 1-6 in the revised manuscript with track changes).Minor points 1. Demonstration of dose-dependency of proteosome inhibitors only checked with one drug.I would suggest that the authors confirm with a second agent such as lactacystin or bortezomib.
Response: Thanks for your kind suggestion.We used bortezomib to confirm the effect of proteasomal inhibitors on PRMT1 protein level.In line with the observation in MG132-treated cells, a dose-dependent increase of PRMT1 protein level was also observed following bortezomib treatment (Supplementary Fig. 1c) (Page 5, Lines 4-7 in the revised manuscript with track changes).In addition, immunoblotting analysis using two different commercial PRMT1 antibodies demonstrated that the upregulated protein level of PRMT1 caused by FBXO7 KD was not observed in bortezomib-treated HCC cells (Supplementary Fig. 3h) (Page 6, Lines 23-28; and Page 7, Line 1 in the revised manuscript with track changes).These data suggest that FBXO7-mediated downregulation of PRMT1 protein level is attributed to ubiquitin-proteasome degradation.
a key enzyme in serine synthesis.Here they investigate the regulation of PRMT1.The E3 ubiquitin ligase FBXO7 promotes PRMT1 degradation, inhibiting serine synthesis.FBXO7 interacts with PRMT1, leading to its ubiquitination and degradation, disrupting serine synthesis.This downregulation of PRMT1 results in reduced serine synthesis, elevated reactive oxygen species, and impaired HCC cell growth.Clinical data support the clinical relevance indicating that FBXO7 is downregulated in HCC tissues, inversely correlated with PRMT1 expression and PHGDH methylation.
The study unveils a novel mechanism involving FBXO7, an E3 ubiquitin ligase, which modulates PRMT1 ubiquitylation.The data provides solid biochemical evidence of a novel regulation by FBXO7 of PRMT1 stability.Nonetheless, a key weakness lies in the limited biological significance of the observed effects, both in in vivo tumor growth and in vitro cell proliferation, which appear marginal and inconclusive.This raises questions regarding the broader biological implications of the described biochemical axis.
Response: Thank you for your critical comments.To strengthen our conclusion and broaden the biological significance regarding FBXO7-mediated PRMT1 K37 ubiquitination, we stably expressed FLAG-FBXO7 and GFP-PRMT1-WT/K37R in Huh7 and PLC/PRF/5 cells.We found that FLAG-FBXO7 overexpression reduced PHGDH methylation and activity, suppressed serine synthesis, and induced ROS accumulation, leading to cell cycle arrest, apoptosis induction, as well as growth inhibition of HCC cells in vitro.In addition, the increase of oxidative stress, induction of apoptosis, and inhibition of growth of HCC cells caused by FBXO7 overexpression were further confirmed in HCC tumor xenografts in vivo.All of these observations in FBXO7-overexpressing cells were markedly restored by PRMT1 K37R mutant, but not its wild-type.Moreover, ROS accumulation, apoptosis induction, and growth inhibition caused by FBXO7 overexpression could be reversed by treatment with the antioxidant N-acetyl cysteine (NAC).These related results and figures can be found in detail in the following responses.We believe that the revised version of this study will provide broader biological implications of the described biochemical axis.
In addition to this, several specific points should be addressed: -The biological consequences of FBXO7 are not clear.A thorough analysis of the biological implications of the observed GSH and ROS imbalances within the biological system would be essential.Questions that need addressing include whether FBXO7overexpressing cells exhibit cell death and cell cycle arrest due to ROS accumulation, whether overexpression of a PRMT1 K37R mutant can rescue these effects, and whether supplementation of antioxidants in vitro can restore the observed effects.
Response: Thank you for your critical comments.We stably expressed FLAG-FBXO7 and GFP-PRMT1-WT/K37R in Huh7 and PLC/PRF/5 cells to examine the biological consequences of FBXO7-mediated PRMT1 ubiquitination and degradation.As shown in Fig. 5f, FBXO7 overexpression promoted ROS accumulation in Huh7 and PLC/PRF/5 cells, which was significantly compromised by introducing K37R mutant, but not wild-type, of PRMT1.In addition, ROS accumulation in FBXO7overexpressing cells with serine/glycine starvation was markedly attenuated by NAC (Fig. 5g).These data demonstrate that FBXO7 promotes ROS accumulation and oxidative stress by downregulating PRMT1 in HCC cells (Page 10, Lines 16-20 in the revised manuscript with track changes).We further determined the role of FBXO7-mediated PRMT1 ubiquitination and degradation in HCC growth.The suppressed growth of HCC cells caused by FBXO7 overexpression was significantly rescued by K37R mutant of PRMT1, but not wildtype PRMT1 (Fig. 6b).Moreover, FBXO7 overexpression led to cell cycle arrest (Supplementary Fig. 6b) and apoptosis induction (as evidenced by increased cleavage of caspase 3; Supplementary Fig. 6c) in HCC cells grown in serine/glycine-free (-SG) medium, which was prominently restored by K37R mutant of PRMT1, but not wildtype PRMT1.Notably, the decreased cell growth and increased caspase 3 cleavage in FBXO7-overexpressing cells with serine/glycine starvation were markedly recovered by NAC treatment (Fig. 6c, d).These data indicate that FBXO7 induces apoptosis and inhibits the growth of HCC cells by downregulating PRMT1 and promoting ROS accumulation (Page 11, Lines 5-13 in the revised manuscript with track changes).-In addition to the point above, addressing the biological significance in vivo, there will still be uncertainty in the in vivo effects, who appeared generally marginal.
The study lacks a clear elucidation of the causative relationships within the FBXO7/PRMT1 axis.Assessing how FBXO7 overexpression impacts tumor growth and whether PRMT1 K37R can rescue the effects of FBXO7 overexpression in vivo is critical.
Response: Thank you for your critical comments.Following your valuable suggestions, we generated a mouse xenograft model by subcutaneously inoculating Huh7 cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R into nude mice, and then examined the effect of FBXO7-mediated PRMT1 ubiquitination in tumor growth.
As shown in Fig. 6g, h, overexpression of FBXO7 inhibited the growth of xenografts from mice fed with a +SG diet.A further decrease in tumor size and weight was observed when mice were fed with a -SG diet (Fig. 6g, h).The inhibition effect of FLAG-FBXO7 on tumor growth was markedly recovered by K37R mutant of PRMT1, but not wild-type PRMT1 (Fig. 6g, h).Notably, tumor xenograft with FBXO7 overexpression showed decreased intensity of Ki-67 staining, accompanied by increased intensity of 8-oxo-dG (a marker of oxidative stress) and cleaved caspase 3 (an apoptosis marker) staining, all of which were obviously rescued by expressing PRMT1 K37R mutant, but not PRMT1 WT (Fig. 6i, Supplementary Fig. 6g).Taken   -Investigating the effects of PRMT1 K37R on PHGDH methylation and activity and whether it preserves or restores the effects of FBXO7 overexpression would also provide stronger evidence in support of the described axis.
Response: Thank you for your kind suggestions.We examined PHGDH methylation level and PHGDH activity in HCC cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R.As shown in Fig. 4d, enforced expression of FBXO7 led to obviously decreased PHGDH methylation level.Overexpressing K37R mutant, but not wild-type PRMT1, rescued the decreased level of PHGDH methylation in FBXO7overexpressing cells.In addition, FBXO7 overexpression resulted in reduced PHGDH activity in HCC cells, which was markedly restored by K37R mutant of PRMT1, but not its wild-type (Fig. 4h).These results demonstrate that FBXO7 suppresses PHGDH methylation and activity by downregulating PRMT1 in HCC cells (Page 8, Lines 26-28; and Page 9, Lines 6-8 in the revised manuscript).In conclusion, while the study introduces a novel regulatory aspect of FBXO7 on PRMT1 ubiquitylation, its limited biological significance and gaps in the mechanistic understanding warrant further investigation.

Reviewer #3 (Remarks to the Author)
In this paper on serine synthesis regulation in cancer, Luo et al. describe how F-boxonly protein 7 (FBXO7), an E3 ubiquitin ligase, inhibits serine synthesis by promoting the ubiquitin-proteasome degradation of protein arginine methyltransferase 1 (PRMT1) in hepatocellular carcinoma (HCC).Previously, they had already shown that PRMT1 is upregulated in HCC.PRMT1 methylates phosphoglycerate dehydrogenase (PHGDH) at arginine 236, thereby activating PHGDH and promoting serine synthesis.
Nevertheless, the mechanisms underlying overexpression of PRMT1 and regulation of the PRMT1-PHGDH axis in HCC remains unknown.
Here, the authors unravel the details of this mechanism.In a series of elegant experiments including all relevant controls, the authors convincingly show that FBXO7 directly interacts with PRMT1, leading to the lysine 37 ubiquitination and subsequent degradation of PRMT1.FBXO7-mediated PRMT1 downregulation then decreases the extent to which PHGDH is methylated on its target arginine residue, resulting in impaired serine synthesis, accumulation of reactive oxygen species (ROS) levels, and inhibition of HCC cell growth.
The authors have gathered a substantial amount of evidence to support these claims, including in vitro biochemical experiments and clinical data, further strengthening their hypothesis.
Overall, the design of the experiments and the levels of detail are very solid.The data convincingly show that FBXO7 suppresses PHGDH R236 methylation and catalytic activity by downregulating PRMT1 in HCC cells.
I have no major comments, only a few minor comments listed below in bullet points: overexpressed in HCC to promote tumor growth […]".Is it really overexpressed or is it just degraded to a lesser extent upon expression?As the mRNA levels seem not to change, I assume it is the latter.
Response: Thanks for your valuable comments.In this study, we found that PRMT1 is upregulated in HCC due to a lesser extent of degradation.Therefore, we have corrected the statement "PRMT1 is overexpressed" as "PRMT1 is upregulated" (Page 2, Line 8; Page 3, Lines 26-27; Page4, Line 3; Page 4, Lines 24-26; Page 8, Line 12; Page 14, Line 18; Page 15, Line 5 in the revised manuscript with track changes).
-P5: Although the authors have used a smart in silico screening to identify interaction partners that ultimately provided them the link to FBXO7, I'm wondering why the authors have not performed an unbiased interaction screen based on affinity purification coupled to mass spectrometric identification of interaction partners themselves.The advantage would be that PRMT1 interactions in the relevant cell system (Huh7 and PLC/PRF/5) could be studied this way.I'm not sure whether BioGRID contains interactome data for this specific cell line.
Response: Following your kind suggestion, we performed label-free quantitative proteomics analysis to identify the interaction partners of PRMT1 in Huh7 cells.In line with in silico screening results, we found that FBXO7 was the only E3 ligase among the top 10 PRMT1-interacting candidates (Supplementary Fig. 2a; Page 5, Lines 19-22 in the revised manuscript with track changes).We therefore selected FBXO7 for subsequent co-IP validation.To this end, we repeated the immunoblotting analysis and found that the upregulated protein level of PRMT1 caused by FBXO7 KD was indeed not observed in MG132treated Huh7 cells (Fig. 2a).This observation was further confirmed using another proteasomal inhibitor, bortezomib, in both Huh7 and PLC/PRF/5 cells (Supplementary Fig. 3h).Moreover, although the PRMT1 level was not determined by PRM, we used two different commercial PRMT1 antibodies (Abcam, ab190892; Proteintech, 11279-1-AP) for immunoblotting analysis to exclude the effect of nonspecific recognition of PRMT1 antibodies.Consistently, the upregulation of PRMT1 protein level caused by FBXO7 KD was not observed in MG132-or bortezomib-treated HCC cells, as evidenced by immunoblotting analysis using two different commercial PRMT1 antibodies (Supplementary Fig. Fig.2e, Expression of FLAG-FBXO7 WT, but not its enzymatically dead ΔF-box mutant, prominently increased the ubiquitination level of GFP-PRMT1 in Huh7 cells (Page 7, Lines 9-13 in the revised manuscript with track changes).

Figure 3
Figure 3. c GFP-PRMT1 (WT or indicated KR mutants) was co-expressed with HA-Ub in FBXO7 KD Huh7 cells rescued with or without FLAG-FBXO7.After MG132 (25 μM, 6 h) treatment, IP was performed with GFP antibody, followed by immunoblotting with indicated antibodies.

Figure R1 .
Figure R1.Schematics of the workflows for identifying the ubiquitination sites of PRMT1 by mass spectrometry analysis (workflow 1 and workflow 2), and for quantitating PRMT1 K37
3. The experiment that is missing from figure4is the demonstration that overexpression of the K37R mutant of PRMT1 prevents the reduction in R236 methylation and enzymatic function of PHGDH associated with FBXO7 over-expression.Can they demonstrate this by blotting or functional analysis??Response: Following your valuable suggestions, we examined PHGDH methylation level and PHGDH activity in HCC cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R.As shown in Fig.4d, enforced expression of FBXO7 led to obviously decreased PHGDH methylation level.Overexpressing K37R mutant, but not wild-type PRMT1, obviously rescued the decreased level of PHGDH methylation in FBXO7-overexpressing cells.In addition, FBXO7 overexpression resulted in reduced PHGDH activity in HCC cells, which was markedly restored by K37R mutant of PRMT1, but not its wild-type (Fig.4h).These results demonstrate that FBXO7 suppresses PHGDH methylation and activity by downregulating PRMT1 in HCC cells (Page 8, Lines 26-28; and Page 9, Lines 6-8 in the revised manuscript).

Figure 4 .
Figure 4. d PHGDH was immunoprecipitated in cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R, followed by immunoblotting with indicated antibodies.h Endogenous PHGDH was immunoprecipitated in cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R, followed by measurement of PHGDH activity.
in the revised version).Supplementary Fig. 8. Schematic model for the mechanism of FBXO7-PRMT1-PHGDH axis in regulating serine synthesis, oxidative stress, and HCC growth.The E3 ubiquitin ligase FBXO7 promotes PRMT1 degradation by lysine 37 ubiquitination, leading to reduction of PHGDH arginine methylation and catalytic activity, suppression of serine synthesis, accumulation of ROS, and inhibition of HCC growth.

Supplementary Figure 7
. c Overall survival of HCC patients based on p53 signature (n = 186 samples) in TCGA database.d Pearson correlation test analyzing the relationship between FBXO7 level and p53 signature in HCC tissues (n = 186 samples).

Figure 5 .
Figure 5. f ROS levels in cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R cultured in CM or -SG medium.g ROS levels in FLAG-FBXO7-overexpressing cells cultured in -SG medium, in the presence or absence of NAC (5 mM).

Figure 6 .b
Figure 6.b Growth rates of cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R grown in CM or -SG medium.c Growth rates of FLAG-FBXO7-overexpressing cells grown in -SG medium in the presence or absence of NAC.d Immunoblotting analysis of cleaved caspase 3 and caspase 3 in FLAG-FBXO7-overexpressing cells grown in -SG medium in the presence or absence of NAC.

Figure 6
Figure 6.g, h Huh7 cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R were subcutaneously inoculated into nude mice fed with a +SG or -SG diet.Tumor image (g) and weight (h) were shown.i, Representative images of IHC staining for Ki-67, 8-oxo-dG, and cleaved caspase 3 in tumor xenografts in g.Scale bars, 50 μm.

Figure 5 .a
Figure 5. a Schematic of U-[ 13 C]-glucose incorporation into serine and glycine in cells.b, c Incorporation of U-[ 13 C]-glucose carbon into serine (b) and glycine (c) detected by LC-MS/MS in cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R.

Figure 4
Figure 4. d PHGDH was immunoprecipitated in cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R, followed by immunoblotting with indicated antibodies.h Endogenous PHGDH was immunoprecipitated in cells overexpressing FLAG-FBXO7 and GFP-PRMT1-WT or K37R, followed by measurement of PHGDH activity.

Figure R3 .
Figure R3.Immunoblotting analysis of PRMT1 and FBXO7 in FBXO7 KD cells treated with 3g, h) (Page 6, Lines 23-28; and Page 7, Line 1 in the revised manuscript with track changes).These data indicate that FBXO7-mediated downregulation of PRMT1 protein level is attributed to ubiquitin-proteasome degradation.