Cooperation between PRMT1 and PRMT6 drives lung cancer health disparities among Black/African American men

Summary Lung cancer is the third most common cancer with Black/AA men showing higher risk and poorer outcomes than NHW men. Lung cancer disparities are multifactorial, driven by tobacco exposure, inequities in care access, upstream health determinants, and molecular determinants including biological and genetic factors. Elevated expressions of protein arginine methyltransferases (PRMTs) correlating with poorer prognosis have been observed in many cancers. Most importantly, our study shows that PRMT6 displays higher expression in lung cancer tissues of Black/AA men compared to NHW men. In this study, we investigated the underlying mechanism of PRMT6 and its cooperation with PRMT1 to form a heteromer as a driver of lung cancer. Disrupting PRMT1/PRMT6 heteromer by a competitive peptide reduced proliferation in non-small cell lung cancer cell lines and patient-derived organoids, therefore, giving rise to a more strategic approach in the treatment of Black/AA men with lung cancer and to eliminate cancer health disparities.


INTRODUCTION
Lung cancer is the leading cause of cancer-related death in the world.Non-small cell lung cancer (NSCLC) accounting for approximately 85% of all lung cancer types includes lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), and large cell carcinoma (LCC).Despite declines in incidence over the last decade, the five-year survival rate for NSCLC is approximately 26%. 1 While survival has increased, Black/AA men in the United States have disproportionally higher incidence and mortality rates of lung cancer compared to NHW men. 2,3Lung cancer disparities in Black/AA men are multifactorial, driven by more than tobacco exposure, instead involving complex and poorly understood interactions between inequities in access to care, upstream determinants of health, and community stress. 2,4Additionally, biological factors are also believed to play critical roles in driving the disparities. 5,6][9][10][11] These findings suggest that other complex genetic and/or epigenetic mechanisms may contribute to the differences in outcomes between Black/AA men and NHW men with lung cancer.Therefore, identifying biomarkers and/or critical genes to serve as early detection or novel therapeutics could reduce lung cancer incidence and mortality rates among Black/AA men.
Arginine methylation is a post-translational modification catalyzed by protein arginine methyltransferases (PRMTs).PRMTs are a family of enzymes known to methylate arginine residues on histones and non-histone proteins, resulting in type I asymmetrical dimethylarginine catalyzed by PRMT1, PRMT2, PRMT3, PRMT4, PRMT6, and PRMT8, type II symmetrical dimethylarginine catalyzed by PRMT5 and PRMT9, and type III monomethylarginine catalyzed by PRMT7. 12Subsequently, the methylation of downstream protein substrates affects transcription, splicing, translation, signal transduction, DNA damage and repair, and cancer development. 13,14

RESULTS
Higher PRMT6 expression is detected in lung adenocarcinoma of Black/AA men compared to NHW men To determine whether type I PRMTs play a role in lung cancer health disparities, we analyzed the expression levels in NSCLC using TCGA datasets and the association of expression with prognosis and outcomes.Consistent with previous studies, 27,28 we found that PRMT1, PRMT4, and PRMT6 were up-regulated in LUAD and LUSC (Figures S1A-S1F) and high expression of these PRMTs was associated with shorter survival in LUAD (Figures S1G-S1I), but not in LUSC (Figures S1J-S1L) analyzed by Kaplan-Meier Plotter. 31There is no difference in PRMT1 and PRMT4 expression in LUAD between Black/AA and NHW in both men and women (Figures S2A-S2D).However, higher PRMT6 expression in LUAD of Black/AA men compared to NHW men (Figure S2E) was observed, but not in women (Figure S2F).No differential expression of PRMT6 was observed in LUSC of Black/AA men and women versus NHW men and women (Figures S2G and S2H).To further validate the observation, we analyzed the expression of PRMT1, PRMT4, and PRMT6 in fresh frozen lung tumor tissue and found elevated expression of all three PRMTs in tumor tissue compared to adjacent uninvolved tissue (Figures 1A and 1B).Most importantly, PRMT6 expression was higher in LUAD tissue of Black/AA men compared to NHW men (Figures 1C and 1D).PRMT6, but not PRMT1, mRNA levels were also higher in LUAD tissue of Black/AA men compared to NHW men (Figures 1E and 1F).Taken together, our results show that, among type I PRMTs, PRMT1, PRMT4, and PRMT6 are up-regulated and associated with worse outcomes in patients with LUAD.Higher PRMT6 levels in LUAD of Black/AA men may contribute to lung cancer health disparities observed in this particular cohort.

PRMT1 and PRMT6 form a heteromer
PRMTs can function independently or cooperate with other PRMTs. 28,29To test the cooperation between PRMT6 and PRMT1 or PRMT4, we analyzed their interactions in transfected cells by coimmunoprecipitation assays and found that PRMT6 interacted with PRMT1, but not with PRMT4 (Figures 2A and 2B).Using recombinant proteins expressed in E. coli, we showed a direct protein-protein interaction between PRMT1 and PRMT6 (Figures 2C and 2D).To examine the interaction in living cells, we performed bimolecular fluorescence complementation (BiFC) assays, 32 where two non-fluorescent fragments of a fluorescent protein form a fluorescent complex when fused to two proteins that interact with each other.We fused the N-terminal fragment of YFP (YFPn) or the C-terminal fragment of YFP (YFPc) to the N-terminus of PRMT1 or PRMT6 and detected the formation of PRMT1/PRMT1 homocomplex, PRMT6/PRMT6 homocomplex, and PRMT1/PRMT6 heterocomplex in BiFC assays (Figure 2E).All transfected proteins were expressed in the cells (Figure 2F).Finally, we demonstrated that endogenous PRMT1 and PRMT6 form a heterocomplex (Figure 2G).

A peptide inhibitor disrupts the formation of PRMT1/PRMT6 heteromer
To demonstrate the importance of the PRMT1/PRMT6 heteromer, we designed several peptides corresponding to amino acid sequences in the loops of PRMT1 or PRMT6 connecting b4 and aD involved in cofactor and substrate binding. 33These peptides were fused with a cellpenetrating peptide from the transactivator of transcription (TAT) of human immunodeficiency virus and tested for the ability to inhibit cell growth.We found that peptide #1, referred to as TAT-1/6i, reduced cell growth by 50% compared to the control peptide, TAT (Figure S3).We next established GST pulldown assays and detected interactions between GST-PRMT1 and HA-PRMT1, GST-PRMT6 and HA-PRMT6, GST-PRMT1 and HA-PRMT6, and GST-PRMT6 and HA-PRMT1 (Figures 3A-3D); no interactions were detected using GST as a control (Figure S4).Incubation with TAT-1/6i only disrupted the formation of the PRMT1/PRMT6 heteromer (Figures 3C and 3D), but not PRMT1/PRMT1 and PRMT6/PRMT6 homocomplexes (Figures 3A and 3B).We next examined whether TAT-1/6i disrupted PRMT1/PRMT6 heteromer in cells using BiFC assays.H1299 cells expressing both YFPc-HA-PRMT1 and YFPn-FLAG-PRMT6 were treated with TAT or TAT-1/6i and analyzed by immunofluorescence and YFP signals.While both YFPc-HA-PRMT1 and YFPn-FLAG-PRMT6 were expressed in the same transfected cells, treatment with TAT-1/6i resulted in a significant reduction in the number of YFP-positive cells (Figure 3E) and YFP signals (Figure 3F) compared to TAT treatment.These results demonstrate that TAT-1/6i disrupts PRMT1/PRMT6 heteromer.

Disruption of the PRMT1/PRMT6 heteromer reduces cell proliferation
To test the specificity and activity in vitro, we treated H1299 and H2122 cells with TAT-1/6i and analyzed arginine methylation on the substrates of PRMT1 or PRMT6, histone 4 (H4) or histone 3 (H3), respectively.No differences in the levels of asymmetrical dimethylation on H4R3 (H4R3me2a) and H3R2 (H3R2me2a) were detected by treatment with TAT-1/6i compared to TAT.In contrast, treatment with MS023, a broader type I PRMT inhibitor, reduced the levels of H4R3me2a and H3R2me2a (Figures 4A and 4B).Treatment with TAT-1/6i did not affect the arginine methylation of G3BP1, a substrate of PRMT1, 34 and increased the arginine methylation of PTEN, a substrate of PRMT6 35 (Figure 4C).These results indicate that arginine methylation activities of PRMT1 and PRMT6 toward their substrates were not affected by TAT-1/6i.We further validated the specificity of TAT-1/6i in cell proliferation using several NSCLC cell lines with varied expressions of PRMT1 and PRMT6 (Figure 4D).TAT-1/6i reduced the growth in cell lines that express both PRMT1 and PRMT6 (H1299 and H2122), but did not affect the growth in cell lines that lack either PRMT1 (H1299 PRMT1 KO) or PRMT6 (A549 and H1299 PRMT6 KO; Figure 4E).Furthermore, the expression of HA-PRMT6 in A549 cells rendered their sensitivity toward TAT-1/6i on cell proliferation (Figures 4F and 4G).
We then established PDXOs and PDOs from LUAD tissue to determine the effect of disruption of PRMT1/PRMT6 heteromer on cell growth.Figure 5A showed that the PDXOs retained the tumor organization and expression of LUAD markers, Napsin A, thyroid transcription factor 1 (TTF-1), and cytokeratin 7 (CK7), as compared to tissue from the original PDX.By contrast, no staining with anti-p63 (a LUSC marker) was detected.Both PRMT1 and PRMT6 were endogenously expressed in PDXOs and PDOs analyzed by immunoblotting (Figure 5B).Treatment of both PDXOs and PDOs with TAT-1/6i induced cell death determined by the microscopic observation of shrinkage of organoids (Figure 5C) and resulted in a significant decrease in cell viability (Figure 5D).We examined a possible mechanism of cell death and did not detect (E and F) mRNA levels of PRMT1 (E) and PRMT6 (F) in LUAD tissue from NHW men and Black/AA men were analyzed by qRT-PCR.Bars indicate the mean of all samples (two-tailed t-test for p value).See also Figures S1 and S2.
an induction of cleaved PARP and cleaved caspase 3, well-known apoptosis markers (data not shown), by the peptide suggesting that other pathways, e.g., autophagic cell death or necrosis, resulted in the observation in Figure 5C.By contrast, treatment with TAT had no effect.These results demonstrate that the disruption of the PRMT1/PRMT6 heteromer inhibits the growth of LC PDXOs and PDOs likely resulting from non-apoptotic cell death pathways.

PRMT1/PRMT6 heteromer catalyzes the arginine methylation of ILF2
We previously demonstrated that the association of ILF2 with PRMT6 potentiated lung tumor progression. 36We therefore examined whether the PRMT1/PRMT6 heteromer could catalyze the arginine methylation of ILF2.ILF2 was immunoprecipitated from H1299, PRMT1 KO, or PRMT6 KO cells followed by immunoblotting using anti-asymmetric dimethylarginine (ADMA) antibodies.ILF2 methylation was significantly reduced in PRMT1 KO and PRMT6 KO cells compared to the parental cells (Figure 6A).We next performed in vitro methylation assays to analyze the abilities of ectopically expressed HA-tagged PRMT1 or PRMT6 to catalyze ILF2 methylation.HA-PRMT1 or HA-PRMT6 immunoprecipitated from H1299 cells efficiently introduced ADMA on ILF2.By contrast, HA-PRMT1 immunoprecipitated from PRMT6 KO cells or HA-PRMT6 immunoprecipitated from PRMT1 KO cells failed to methylate ILF2 (Figures 6B and 6C).To examine the role of catalytic activities of PRMT1 and PRMT6 in the methylation of ILF2, we expressed a PRMT1 mutant, PRMT1 AAA,37 or a PRMT6 mutant, PRMT6 KLA,38 which lack catalytic activity, in PRMT1 KO or PRMT6 KO cells.Arginine methylation of ILF2 by PRMT1 mutant or PRMT6 was significantly reduced (Figures 6D and 6E) indicating that the enzymatic activities of both PRMT1 and PRMT6 are required.Furthermore, treatment of NSCLC cells and LC PDOs with TAT-1/6i reduced arginine methylation of ILF2 (Figure 6F) indicating that PRMT1/PRMT6 heteromer is responsible for ILF2 methylation.We further examined the arginine methylation of ILF2 by PRMT1 and PRMT6 using recombinant proteins expressed in E. coli.Although PRMT1 alone, but not PRMT6 alone, could methylate ILF2, the presence of both PRMT1 and PRMT6 strongly methylated ILF2 (Figure 6G).More importantly, the arginine methylation of ILF2 introduced by PRMT1/PRMT6 heteromer was significantly decreased by the treatment of TAT-1/6i (Figure 6H).
We next examined the recruitment of ILF2 into the PRMT1/PRMT6 heteromer by co-immunoprecipitation assays.While ILF2 was efficiently coprecipitated with FLAG-PRMT1 or FLAG-PRMT6 in H1299 cells, the coprecipitation was dramatically reduced in the absence of PRMT6 or PRMT1, respectively (Figures 6I and 6J).Similarly, coprecipitation of PRMT1 or PRMT6 with FLAG-ILF2 was significantly reduced in the absence of PRMT6 or PRMT1, respectively (Figures 6K and 6L).These results demonstrate the ability of the PRMT1/PRMT6 heteromer to recruit/recognize and therefore catalyze arginine methylation on a new class of substrates, e.g., ILF2.

PRMT1/PRMT6 heteromer positively regulates ILF2 expression
We examined the effect of the PRMT1/PRMT6 heteromer on ILF2 expression and detected a significant reduction in ILF2 levels in PRMT1 KO and PRMT6 KO cells (Figure 7A).Disruption of PRMT1/PRMT6 heteromer resulted in a reduction of ILF2 expression in H1299 and H2122 cells, but not in A549 cells lacking PRMT6 expression (Figure 7B), and in LC PDOs (Figure 7C).Expression of HA-PRMT6 in A549 cells increased ILF2 expression (Figure 7D) and rendered their sensitivity to TAT-1/6i-induced ILF2 inhibition (Figure 7E).We next tested whether the change in ILF2 expression resulted from protein stability.ILF2 stability was significantly decreased in PRMT1 KO and PRMT6 KO cells compared to H1299 cells (Figure 7F).To determine the effect of arginine methylation on ILF2 expression, we identified methylation sites by mass spectrometry.ILF2 containing five potential RG sites at the N-terminus (Figure S5A) was incubated with both His-PRMT1 and GST-PRMT6 in a methylation assay.We identified a peptide with dimethylated R2, a peptide with dimethylated R5, a peptide with dimethylated R7, and a peptide containing R9 with a monomethyl group (Figures S5B-S5F).We were unable to find a peptide containing R16.These results suggest that R2, R5, and R7 are dimethylated, and R9 is likely dimethylated as we detected a monomethyl group which is an intermediate product of type I PRMTs.Although we could not detect whether all four Rs are dimethylated within a peptide, our results suggested that they are likely simultaneously dimethylated or randomly dimethylated by PRMT1/PRMT6 heteromer.To confirm this, we mutated all four arginines to lysines and examined the effect on methylation by PRMT1/PRMT6 heteromer.ILF2 carrying mutations (ILF2 R2/5/7/9K ) blocked the methylation by the heteromer (Figures S5G and S5H) and exhibited a shorter half-life compared to wild-type ILF2 (Figure 7G).Together, our results demonstrate that arginine methylation by PRMT1/PRMT6 heteromer positively regulates ILF2 expression resulting from increased protein stability.

Higher ILF2 expression is detected in lung adenocarcinoma of Black/AA men
Previous studies have shown that the increased expression of ILF2 promotes cell proliferation and is associated with worse overall survival in patients with NSCLC. 39,40Using TCGA datasets and Kaplan-Meier Plotter, we found the upregulation of ILF2 in LUAD and an association of higher ILF2 expression with poorer survival (Figures S6A and S6B).We did not observe higher ILF2 mRNA levels in LUAD of Black/AA men versus NHW men (Figure S6C).However, higher ILF2 protein levels were detected in fresh frozen LUAD tissue of Black/AA men versus NHW men (Figures S6D and S6E).These results further support the findings that PRMT1/PRMT6 heteromer positively drives ILF2 expression in LUAD of Black/AA men via control of protein stability.

ILF2 is essential for cell proliferation
To investigate whether the effect of PRMT1/PRMT6 heteromer on cell proliferation acts through ILF2, we re-expressed HA-PRMT1 in PRMT1 KO cells and FLAG-PRMT6 in PRMT6 KO cells.Re-expression of PRMT1 or PRMT6 in the respective knockout cell lines restored ILF2 levels (Figures S7A and S7B) and cell proliferation (Figure S7C).We next examined the role of ILF2 in cell growth.Knockdown of ILF2 by siRNAs reduced cell growth in PRMT1 KO, PRMT1 KO with the re-expression of PRMT1, PRMT6 KO, and PRMT6 KO with the re-expression of PRMT6 (Figures S7D-S7F).Taken together, our results suggest that the increased expression of ILF2 by the PRMT1/PRMT6 heteromer promotes cell proliferation and likely contributes to lung cancer health disparities in Black/AA men.

Arginine methylation of ILF2 by PRMT1/PRMT6 heteromer counteracts ubiquitination by CRL4 CRBN E3 ligase
To determine the underlying mechanism of ILF2 regulation by PRMT1/PRMT6 heteromer, we treated H1299, PRMT1 KO, and PRMT6 KO cells with MG132, a proteasome inhibitor, and observed that ILF2 levels were restored in PRMT1 KO and PRMT6 KO cells.Meanwhile, the levels of ILF2 were not affected in H1299 cells (Figure 8A).MG132 treatment also increased the levels of p21 subjected to degradation by proteasomes.These results suggest that ILF2 is undergoing significant proteasome-mediated degradation in the absence of PRMT1 or PRMT6.We next measured the levels of ubiquitination of ILF2 in H1299, PRMT1 KO, and PRMT6 KO cells.Ubiquitination of ILF2 was increased in both PRMT1 KO and PRMT6 KO cells compared to H1299 cells (Figures 8B and 8C).Additionally, ubiquitination of mutant ILF2 (ILF2 R2/5//7/9K ) was increased compared to wild-type ILF2 (Figure 8D), indicating that the methylation of ILF2 prevents the ubiquitination of ILF2.It has been shown that the CUL4-RBX1-DDB1-CRBN (CRL4 CRBN ) E3 ubiquitin ligase complex regulates ILF2 turnover. 41Thus, we examined the effect of the PRMT1/PRMT6 heteromer on the recruitment of ILF2 into CRL4 CRBN complex.The levels of ILF2 coprecipitated with CUL4A was increased in the absence of PRMT1 or PRMT6 (Figure 8E).In addition, the association of ILF2 R2/5/7/9K with CUL4A was significantly increased compared to wild-type ILF2 (Figure 8F).Altogether, our results demonstrate that ILF2 is degraded by a ubiquitination-mediated process and arginine methylation by PRMT1/PRMT6 heteromer counteracts the action of CRL4 CRBN complex on ILF2 (Figure 8G).

DISCUSSION
Here, we reported that, among the members of type I PRMT family, levels of PRMT1, PRMT4, and PRMT6 were elevated in NSCLC, and high expression was associated with poorer patient survival.Furthermore, PRMT6 expression was significantly increased in LUAD tissue of Black/AA men compared to NHW men.We also demonstrated that PRMT1 and PRMT6 form a heteromer essential for cell proliferation.Together, these results and our previous findings that mice with the lung-specific expression of PRMT6 develop more lung tumors 36 strongly suggest that elevated PRMT6 expression in cooperation with PRMT1 drives lung cancer development and likely contributes to health disparities in Black/AA men.However, we cannot completely rule out that PRMT6 alone via the formation of homodimers contributes to the observed disparities in Black/AA men.It has been shown that PRMT1 is co-purified with PRMT6 29 and PRMT6 is a substrate of PRMT1, in which enzymatic activity of PRMT6 is decreased by methylation. 30However, the significance of this interaction contributing to cell proliferation and cancer development is still unknown.We demonstrated that PRMT1/PRMT6 heteromer played a critical role in cell proliferation and perhaps lung cancer development.Although the formation of the PRMT1/PRMT6 heteromer is via direct protein-protein interactions, the topology of the complex is still unknown.While PRMT1 and PRMT6 can form a heterodimer, we speculate that they can also form a heterotetramer containing a PRMT1 homodimer and a PRMT6 homodimer and/or an oligomer containing more than one PRMT1 homodimer and one PRMT6 homodimer.This is based on our findings that a competitive peptide corresponding to the loop region of PRMT1 connecting the b4 and aD, which is not essential for the formation of PRMT homodimer mediated between the dimerization arm (a4-6 helices) and helices aY/Z and a1/2 of the Rossmann domain, 42 disrupts the formation of PRMT1/PRMT6 heteromer.
Type I PRMTs recognize and catalyze the arginine methylation of various substrates, including histones and non-histone proteins. 12,13We demonstrated that the PRMT1/PRMT6 heteromer recruits and catalyzes the methylation of a novel substrate, e.g., ILF2, which is weakly recognized by either PRMT1 or PRMT6 homodimers (Figure 6).Although the details of substrate recognition by the PRMT1/PRMT6 heteromer are currently unknown, our data implicate that the heteromer catalyzes arginine methylation on a battery of proteins playing critical roles in a variety of cellular processes.Therefore, it is of interest to identify these substrates and characterize their functions in lung cancer development and their role in health disparities.
Our results indicate that the enzymatic activities of both PRMT1 and PRMT6 are essential for catalyzing the arginine methylation of ILF2.The data showed that no methylation of ILF2 by the complex containing PRMT6 and catalytically inactive PRMT1 was detected (Figure 6D).In contrast, some methylation by the complex containing PRMT1 and catalytically inactive PRMT6 was detected (Figure 6E).Using recombinant PRMT1 and PRMT6, we showed that PRMT1 alone could methylate ILF2, but not PRMT6.However, working together as a heteromer yielded a (B) The ability of HA-PRMT6 to catalyze methylation on ILF2 is markedly reduced in the absence of PRMT1.HA-PRMT6 expressed in H1299 or PRMT1 KO cells was immunoprecipitated with anti-HA.The precipitates were incubated with His-ILF2 in methylation assays followed by anti-ADMA or anti-HA immunoblotting.
(C) The ability of HA-PRMT1 to catalyze arginine methylation on ILF2 is markedly reduced in the absence of PRMT6.HA-PRMT1 expressed in H1299 or PRMT6 KO cells was immunoprecipitated with anti-HA.The precipitates were incubated with His-ILF2 in methylation assays followed by anti-ADMA or anti-HA immunoblotting.(D) Catalytic activity of PRMT1 is required for the methylation of ILF2.FLAG-PRMT1 expressed in H1299 or PRMT1 KO cells or FLAG-PRMT1 AAA , a catalytically inactive PRMT1, expressed in PRMT1 KO cells were immunoprecipitated with anti-FLAG.The precipitates were incubated with His-ILF2 in methylation assays followed by anti-ADMA, anti-FLAG, or anti-PRMT6 immunoblotting.
(E) Catalytic activity of PRMT6 is required for the methylation of ILF2.FLAG-PRMT6 expressed in H1299 or PRMT6 KO cells or FLAG-PRMT6 KLA , a catalytically inactive PRMT6, expressed in PRMT6 KO cells were immunoprecipitated with anti-FLAG.The precipitates were incubated with His-ILF2 in methylation assays followed by anti-ADMA, anti-FLAG, or anti-PRMT1 immunoblotting.
Coomassie blue staining of the membrane after immunoblotting is shown.
(H) Both His-PRMT1 and GST-PRMT6 were incubated with TAT (500 nM) or TAT-1/6i (500 nM) in room temperature for 25 min, followed by the addition of His-ILF2 and methylation assays.The reactions were analyzed by anti-ADMA immunoblotting.Coomassie blue staining of the membrane after immunoblotting is shown.
(I) Coprecipitation of ILF2 with FLAG-PRMT1 is reduced in the absence of PRMT6.FLAG-PRMT1 expressed in H1299 or PRMT6 KO cells was immunoprecipitated with anti-FLAG followed by anti-ILF2 or anti-FLAG immunoblotting.
(J) Coprecipitation of ILF2 with FLAG-PRMT6 is reduced in the absence of PRMT1.FLAG-PRMT6 expressed in H1299 or PRMT1 KO cells was immunoprecipitated with anti-FLAG followed by anti-ILF2 or anti-FLAG immunoblotting.
(K) Coprecipitation of PRMT1 with FLAG-ILF2 is reduced in the absence of PRMT6.FLAG-ILF2 expressed in H1299 or PRMT6 KO cells was immunoprecipitated with anti-FLAG followed by anti-PRMT1 or anti-FLAG immunoblotting.
(L) Coprecipitation of PRMT6 with FLAG-ILF2 is reduced in the absence of PRMT1.FLAG-ILF2 expressed in H1299 or PRMT1 KO cells was immunoprecipitated with anti-FLAG followed by anti-PRMT6 or anti-FLAG immunoblotting.
synergistic affect toward increased ILF2 methylation (Figure 6G).Currently, the exact mechanism by which PRMT1/PRMT6 heteromer catalyzes ILF2 methylation is unclear.It is possible that the methylation of ILF2 by PRMT1 is further activated by PRMT6-mediated methylation of PRMT1 although Cao et al. showed that PRMT6 was methylated by PRMT1. 30Alternatively, the methylation of ILF2 at specific arginine sites within the RG motifs by PRMT1 could render further methylation at distinct arginine sites by PRMT6.It has been previously shown that ILF2 interacts with CRBN, a known substrate receptor, the domains required for the interaction need further investigation.We speculate that the interaction between the N-terminus of ILF2 and CRBN is required for recruitment into the CRL4 CRBN complex to undergo ubiquitination.We postulate that methylation by PRMT1/PRMT6 heteromer will prevent the interaction between ILF2 and CRBN, thereby inhibiting the ubiquitination-mediated degradation of ILF2 (Figure 8G).Although the dysregulation of ILF2 is involved in multiple cancer types, [24][25][26] its role in lung cancer is unclear.Our results showed higher ILF2 levels in lung tumor samples compared to the matched uninvolved lungs 36 and in lung cancer tissues of Black/AA men compared to NHW men (Figure S6D).We further demonstrated that PRMT1/PRMT6 heteromer's effect on cell proliferation acted through ILF2 (Figure S7), suggesting a positive effect of ILF2 on cell growth.Thus, it is of interest to determine the role of ILF2 in lung cancer development and in mediating the function of the PRMT1/PRMT6 heteromer in vivo.
To demonstrate the proof of principle for the targeting of the PRMT1/PRMT6 heteromer in lung cancer, we showed that disrupting the complex reduced cell viability in LC PDOs from a Caucasian male (PDXO) and a Black/AA female (PDO).Since higher PRMT6 expression, thus higher PRMT1/PRMT6 heteromer, is found in LUAD of Black/AA men, targeting the complex will be most beneficial for this cohort.It is also of interest to compare the effectiveness of the competitive peptide in LUAD PDOs derived from Black/AA men versus NHW men.
In summary, our study identified a unique role of the PRMT1/PRMT6 heteromer in driving lung cancer development which likely contributes to cancer health disparities in Black/AA men.Targeting the PRMT1/PRMT6 heteromer could create a new class of PRMT inhibitors in treating lung cancer to overcome the current limitations of PRMT1 or PRMT6 inhibitors, thereby opening up new horizons for efficacious PRMT1/ PRMT6-targeted agents that are expected to eliminate lung cancer health disparities.

Limitations of the study
In this study, we found that higher PRMT6 expression in LUAD of Black/AA men versus NHW men using a limited sample number.This observation needs to be further validated using a larger cohort.The higher PRMT6 expression correlated with worse outcomes in Black/AA men versus NHW men should be also further investigated.Although our data indicate that disrupting PRMT1/PRMT6 heteromer reduces cell proliferation using NSCLC cell lines, it remains uncertain whether this is also occurring in vivo.Thus, the next step is to utilize genetically engineered mouse models to demonstrate the role of PRMT1/PRMT6 heteromer in lung cancer.While we demonstrate that PRMT1/PRMT6 heteromer recognizes and catalyzes arginine methylation on a new class of substrates, we still do not know the list of substrates and the mechanism of recognition and catalyzation of the substrates.Furthermore, the underlying mechanism by which the heteromer drives lung (G) Decreased stability of HA-ILF2 R2/5/7/9K .HA-tagged wild-type ILF2 or ILF2 R2/5/7/9K was expressed in H1299 cells.Their stabilities were measured as in (D).Halflife of ILF2 was calculated and presented as mean G SD (n = 3).See also Figures S5-S7.
cancer and causes worse outcomes in Black/AA men with lung cancer is still unclear.Our study using NSCLC cell lines indicates that ILF2 plays a critical in cell proliferation and is likely a downstream effector of PRMT1/PRMT6 heteromer.However, the in vivo role of ILF2 in driving lung cancer and whether the function of PRMT1/PRMT6 heteromer is mediated via ILF2 still await further investigation.We demonstrate that the methylation of ILF2 by PRMT1/PRMT6 heteromer counteracts the action of ubiquitination by CRL4 CRBN complex.However, the mechanism by which methylation affects ubiquitination and whether CRL4 CRBN directly catalyzes the ubiquitination of ILF2 are not provided in the study.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:  (D) H1299 cells were co-transfected with a control vector or vectors expressing FLAG-ILF2 or FLAG-ILF2 R2/5/7/9K and a vector expressing HA-ubiquitin.Cell extracts were subjected to anti-FLAG immunoprecipitation followed by anti-HA or anti-FLAG immunoblotting.
(E) H1299, PRMT1 KO, or PRMT6 KO cells were co-transfected with vectors expressing FLAG-CUL4A and HA-ILF2.Cell extracts were subjected to anti-FLAG immunoprecipitation followed by anti-FLAG or anti-HA immunoblotting.(F) H1299 cells were co-transfected with a control vector or vectors expressing FLAG-ILF2 or FLAG-ILF2 R2/5/7/9K and a vector expressing HA-CUL4A.Cell extracts were subjected to anti-HA immunoprecipitation followed by anti-HA or anti-FLAG immunoblotting.

Gene editing of PRMT1 and PRMT6
Guide RNAs targeting the first exons of PRMT1 (CATGATGCAGTTCGCGGCCT) and PRMT6 (GAAAAGAAAGCTTGAGTCGG) were used to generate PRMT1 knockout (KO) and PRMT6 KO. 36 H1299 cells were transfected with the guide plasmids using Lipofectamine 2000 reagent (Invitrogen).48 h after transfection, cells were trypsinized, diluted to a concentration of 1 cell/100mL, seeded into 96-well plates (1 cell/well), and grown for 3 weeks.Single clones were isolated and assayed by immunoblot analysis for PRMT1 and PRMT6 expression.Multiple clones that showed complete loss of PRMT6 expression were selected for downstream analysis.

Lentiviral production and transduction
HEK293T cells (8 3 10 6 cells) were seeded onto a 100-mm culture dish.Cells were transfected 16 h later with a lentiviral vector expressing a respective gene (12 mg), pMD.G (2 mg), and psPAX2 (8 mg) using Lipofectamine 2000.Medium was replaced with fresh growth medium (7 mL) containing sodium butyrate (10 mM) 16 h later.Medium containing lentivirus was collected 24 h later, centrifuged, filtered using a 0.45 mm syringe filter, and saved at À80 C. Cells were further incubated in fresh growth medium (7 mL) containing sodium butyrate (5 mM), and medium was collected 24 h later, filtered, and saved at À80 C. To transduce cells with lentiviral vectors expressing HA-PRMT1 or FLAG-PRMT6, PRMT1 KO or PRMT6 KO cells (3 3 10 5 cells/well) were seeded onto a 6-well dish and incubated with 1 mL of growth medium containing lentivirus and polybrene (8 mg/mL) for 16 h.The transduced cells were selected for 7 days in the presence of hygromycin B (400 mg/mL) for HA-PRMT1 or puromycin (0.75-1 mg/mL) for FLAG-PRMT6.

Protein expression and purification in E. coli
Single colony of transformed BL21 (for GST fusion) or OrigamiTM2 (DE3) (for His fusion) expressing His-or GST-hPRMT1, GST-hPRMT6, or His-ILF2 was inoculated to 2 mL of growth medium and incubated overnight at 37 C with shaking.The next day, the culture was diluted 1:50 to fresh medium and continued to grow for 2 h followed by the addition of 0.5 mM of isopropyl-b-D-thiogalactoside (IPTG).The culture was incubated at 25 C with shaking for additional 6 h and harvested.The pellets were re-suspended in 10 mL of lysis buffer [125 mM Tris-HCl (pH8.0), 150 mM sodium chloride, 1 mM PMSF, 250 mg/mL lysozyme, and 0.5% Triton X-100].Genomic DNA was shared by sonication followed by centrifugation.Protein purification was carried out per Thermo Scientific procedures using Pierce Glutathione Superflow Agarose for GST fusion proteins or HisPur Ni-NTA Resin for His fusion proteins.

Cell proliferation assays
Cell proliferation was measured by Sulforhodamine B (SRB) assays. 49Briefly, cells (1,500-2,500 cells/well) were cultured in a 96-well plate and harvested at different time points.Cells were fixed with 10% trichloroacetic acid for 30 min at 4 C and subjected to SRB staining.Briefly, 70 mL of 0.4% (w/v) SRB in 1% acetic acid solution were added to the cells, followed by incubation at room temperature for 20 min.The cells were washed with 1% acetic acid to remove unbound SRB.Bound SRB was solubilized with 200 mL of 10 mM unbuffered Tris solution.Absorbance was read at 492 nm subtracting the background measurement at 620 nm.Cell growth rate was calculated by normalizing the readings of each time point with 0 time point, which was set at 1, to control for plating of equal number of cells among different treatments.

Immunoprecipitation, pulldown, and immunoblot analyses
Extracts from NSCLC cells or fresh frozen lung tissue samples were prepared in NP-40 lysis buffer [150 mM NaCl, 50 mM Tris-HCl (pH 8), and 1% NP-40].Proteins were either immunoprecipitated with specific antibodies using G resin (GenScript, L00209) or anti-FLAG M2 affinity gel (Sigma, A2220), or pulled down using Pierce Glutathione Superflow Agarose (ThermoFisher, 25236) or HisPur Ni-NTA Resin (ThermoFisher, 88221) at 4 C overnight.The precipitates were extensively washed with NP-40 lysis buffer, resolved on SDS-PAGE gels, and analyzed by immunoblotting. 49 vitro methylation assays In vitro methylation assay was conducted using either bacterially produced recombinant proteins or immunoprecipitates. 50Briefly, H1299 cells were transfected transiently with HA (FLAG)-PRMTs (10 mg) in 100 mm culture dishes.Post 48 h of transfection, cells were lysed in NP40 lysis buffer, followed by the immunoprecipitation methods from this study.The bacterial recombinant proteins were generated according to the protocols in this study.Subsequently, immunoprecipitates, containing bound PRMTs or recombinant PRMTs, were incubated with His-ILF2 or His-ILF2 R2/5/7/9K (1-2 mg) in a reaction buffer composed of 50 mM Tris (pH 8.5), 20 mM KCl, 10 mM MgCl 2 , 1 mM b-mercaptoethanol, 100 mM sucrose, and S-adenosyl-L-methionine (final concentration: 500 mM) at 30 C for 1-1.5 h.Reactions were stopped by adding 6X SDS sample buffer, followed by denaturation through incubation at 95 C for 10 min.Samples were resolved via SDS-PAGE and subjected to immunoblotting using the anti-ADMA antibody (Cell Signaling, 13522).

RNA isolation and quantitative RT-PCR
Total RNA from lung tissues was extracted using Trizol reagent (Invitrogen).1 mg of total RNA was reverse transcribed using random primers.Real-time PCR was performed using the Brilliant II SYBR Green QPCR Master Mix (Agilent, 600828) and the Bio-Rad CFX96 qPCR detection system.The primers used for PCR experiments are listed in the key resources table.

Mass spectrometry
His-ILF2 was incubated with both His-PRMT1 and GST-PRMT6 in a methylation assay.The reaction was resolved by SDS-PAGE followed by Coomassie blue staining.The band corresponding to ILF2 was excised from the gel, cut into equal size cubes (approximately 1 mm 3 ), and transferred to a siliconized tube.The gel pieces were washed with 80% acetonitrile (CAN) for 10 min, followed by a H 2 O wash for 10 min.Gel pieces were dehydrated with 50% ACN for 5 min and vacuum centrifuged for 20 min.Samples were rehydrated with 20 mM DTT and incubated for 1 h at 56 C, followed by incubation with 60 mM iodoacetimide for 45 min in the dark.The pieces were dehydrated with 50% ACN for 10 min and vacuum centrifuged for 20 min.To rehydrate the gel cubes, samples were incubated for 10 min in 20 mL of 0.1% Protease Max (PM) trypsin (12.5 ng/mL) in 100 mM ammonium bicarbonate (AB).The gel pieces were then overlaid with 30 mL of 0.1% PM in 100 mM AB and allowed to digest for 2 h at 37 C.The digests were collected into fresh siliconized tubes, centrifuged at 14,000 rpm for 5 min, transferred to a fresh tube, and stored for subsequent mass spectrometry analysis.For LC-MS/MS, samples were analyzed using a Q Exactive HFX (Thermo Scientific) coupled to an Easy nLC 1200 (Thermo Scientific).The LC-MS system consisted of a Thermo Electron Q-Exactive HF-X mass spectrometer with an Easyspray Ion source connected to an Acclaim PepMap 75 mm 3 2 cm nanoviper C18 3 mm 3 100 A ˚pre-column in series with an Acclaim PepMap RSLC 75 mm 3 50 cm C18 2 mm bead size (Thermo Scientific).Peptides were injected onto the column and then eluted from the column with an 80% ACN and 0.1% formic acid gradient at a flow rate of 0.3 mL/min over 1 h.The nano-spray ion source was operated at 1.9 kV.The digests were analyzed using the rapid switching capability of the instrument thus acquiring a full scan mass spectrum to determine peptide molecular weights followed by product spectra (25 High Energy C-trap Dissociation HCD spectra).This mode of analysis produces approximately 50,000 MS/MS spectra of ions ranging in abundance over several orders of magnitude.Not all MS/MS spectra are derived from peptides.The data were analyzed by database searching using the Sequest HT search algorithm with semi-tryptic enzyme digest using a custom database containing ILF2 downloaded from Swiss Prot.The data were processed with Proteome Discoverer 3.0.Searches including post-translational modifications (PTMs) of oxidation of methionine and carbamidomethyl cysteine.The number of missed cleavages for a tryptic search was set to 7. Modifications per peptide were set to 8. PTM search was set for methyl and di-methyl.

Figure 1 .
Figure 1.Expression of PRMT1, PRMT4, and PRMT6 is elevated in NSCLC tissue, and higher PRMT6 levels are detected in LUAD tissue of Black/AA men compared to NHW men (A) Extracts of uninvolved lung tissue and tumor tissue from the same patient were analyzed by immunoblotting using anti-PRMT1, anti-PRMT4, anti-PRMT6, or anti-vinculin.A portion of Ponceau S staining for equal protein loading is shown.(B) Quantification of signals in (A) is shown.Bars indicate the mean of all samples (two-tailed t-test for p value).(C) Extracts of LUAD tissue from NHW men and Black/AA men were analyzed by immunoblotting using anti-PRMT1, anti-PRMT4, anti-PRMT6, or anti-vinculin.A portion of Ponceau S staining for equal protein loading is shown.Vinculin blots here are the same ones in Figure S6D.(D) Quantification of signals in (C) is shown.Bars indicate the mean of all samples (two-tailed t-test for p value).(E and F) mRNA levels of PRMT1 (E) and PRMT6 (F) in LUAD tissue from NHW men and Black/AA men were analyzed by qRT-PCR.Bars indicate the mean of all samples (two-tailed t-test for p value).See also Figures S1 and S2.

Figure 2 .
Figure 2. PRMT1 interacts with PRMT6 (A) H1299 cells were co-transfected with a control vector or a vector expressing FLAG-PRMT6 and vectors expressing HA-PRMT1 or HA-PRMT4.Cell extracts were subjected to anti-FLAG immunoprecipitation.The immunoprecipitates were analyzed by anti-HA or anti-FLAG immunoblotting.(B) H1299 cells were transfected with a control vector or a vector expressing FLAG-PRMT6.Cell extracts were subjected to anti-FLAG immunoprecipitation.The immunoprecipitates were analyzed by anti-PRMT1, anti-PRMT4, or anti-FLAG immunoblotting.(C) GST or GST-PRMT6 was incubated with His-PRMT1.The reactions were subjected to GST pulldown followed by anti-PRMT1 or anti-GST immunoblotting.(D) His-PRMT1 incubated with GST-PRMT6 was subjected to Ni-NTA pulldown followed by anti-PRMT6 or anti-PRMT1 immunoblotting.(E) H1299 cells were co-transfected with vectors expressing YFPn-PRMT1 and YFPc-PRMT1, YFPn-PRMT6 and YFPc-PRMT6, YFPn-PRMT6 and YFPc-PRMT1, YFPn-PRMT1 and YFPc-MS2, or YFPn-PRMT6 and YFPc-MS2.YFP signal was analyzed by fluorescence microscopy.Scale bar, 25 mm.(F) Expression of transfected proteins in (E) was analyzed by anti-FLAG or anti-HA immunoblotting.(G) Extracts of H1299 or PRMT1 KO cells were subjected to anti-PRMT6 immunoprecipitation.The immunoprecipitates were analyzed by anti-PRMT1 or anti-PRMT6 immunoblotting.

Figure 3 .
Figure 3. Development of a peptide inhibitor of PRMT1/PRMT6 heteromer (A) GST-PRMT1 incubated with cell extracts containing HA-PRMT1 in the presence of TAT (20 mM) or TAT-1/6i (20 mM) was subjected to GST pulldown followed by anti-HA or anti-GST immunoblotting.(B) GST-PRMT6 incubated with cell extracts containing HA-PRMT6 in the presence of TAT (20 mM) or TAT-1/6i (20 mM) was subjected to GST pulldown followed by anti-HA or anti-GST immunoblotting.(C) GST-PRMT1 incubated with cell extracts containing HA-PRMT6 in the presence of TAT (20 mM) or TAT-1/6i (20 mM) was subjected to GST pulldown followed by anti-HA or anti-GST immunoblotting.(D) GST-PRMT6 incubated with cell extracts containing HA-PRMT1 in the presence of TAT (20 mM) or TAT-1/6i (20 mM) was subjected to GST pulldown followed by anti-HA or anti-GST immunoblotting.See also Figures S3 and S4.(E and F) H1299 cells were co-transfected with vectors expressing YFPc-HA-PRMT1 and YFPn-FLAG-PRMT6 followed by treatment with TAT (200 mM) or TAT-1/6i (200 mM) for 24 h.Cells were subjected to anti-HA and anti-FLAG immunofluorescence.Signals were analyzed by fluorescence microscopy using 103 magnification (E, Scale bar, 100 mm) or 403 magnification (F, Scale bar, 20 mm).

Figure 8 .
Figure 8. PRMT1/PRMT6-mediated methylation counteracts the ubiquitination of ILF2 (A) H1299, PRMT1 KO, or PRMT6 KO cells were treated with MG132 (5 mM) for 8 h.Levels of ILF2, p21, and GAPDH were analyzed by immunoblotting.(Band C) H1299 or PRMT1 KO cells (B) and H1299 or PRMT6 KO cells (C) were transfected with a vector expressing HA-ubiquitin.Cell extracts were subjected to anti-ILF2 immunoprecipitation followed by anti-HA or anti-ILF2 immunoblotting.(D) H1299 cells were co-transfected with a control vector or vectors expressing FLAG-ILF2 or FLAG-ILF2 R2/5/7/9K and a vector expressing HA-ubiquitin.Cell extracts were subjected to anti-FLAG immunoprecipitation followed by anti-HA or anti-FLAG immunoblotting.(E) H1299, PRMT1 KO, or PRMT6 KO cells were co-transfected with vectors expressing FLAG-CUL4A and HA-ILF2.Cell extracts were subjected to anti-FLAG immunoprecipitation followed by anti-FLAG or anti-HA immunoblotting.(F) H1299 cells were co-transfected with a control vector or vectors expressing FLAG-ILF2 or FLAG-ILF2 R2/5/7/9K and a vector expressing HA-CUL4A.Cell extracts were subjected to anti-HA immunoprecipitation followed by anti-HA or anti-FLAG immunoblotting.(G) Illustration of CRL4 CRBN complex and PRMT1/PRMT6 heteromer in the regulation of ILF2 expression.

TABLE
d RESOURCE AVAILABILITY B Lead contact B Materials availability B Data and code availability d EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS B Lung cancer specimens B Cell lines and cell culture d METHOD DETAILS B LC patient-derived organoids (PDOs) Data were collected from at least two independent, replicate experiments.All data are presented as the mean G SD. Statistical significance (p value) was calculated by two-tailed t-test or one-way ANOVA (*, p < 0.05; **, p < 0.005; ***, p < 0.0005; ****, p < 0.00005; *****, p < 0.000005).