Type I Prmts and PRMT5 Inversely Regulate Post-Transcriptional Intron Detention

bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1 Type I PRMTs and PRMT5 Inversely Regulate Post-Transcriptional Intron Detention

2 Maxim I. Maron1, Alyssa D. Casill2, Varun Gupta3, Simone Sidoli1, Charles C. Query3, Matthew

3 J. Gamble2,3, and David Shechter1*

5 1Department of Biochemistry, Albert Einstein College of Medicine, Bronx, NY 10461

6 2Department of Molecular Pharmacology, Albert Einstein College of Medicine, Bronx, NY 10461

7 3Department of Cell Biology, Albert Einstein College of Medicine, Bronx, NY 10461

8 *For correspondence: [email protected]; 718-430-4120

9 ABSTRACT

10 Protein arginine methyltransferases (PRMTs) are required for the regulation of RNA processing

11 factors. Type I enzymes catalyze mono- and asymmetric dimethylation; Type II enzymes cata-

12 lyze mono- and symmetric dimethylation. To understand the specific mechanisms of PRMT ac-

13 tivity in splicing regulation, we inhibited Type I and II PRMTs and probed their transcriptomic

14 consequences. Using the newly developed SKaTER-seq method, analysis of co-transcriptional

15 splicing revealed that PRMT inhibition resulted in slower splicing rates. Surprisingly, altered co-

16 transcriptional splicing kinetics correlated poorly with ultimate changes in alternative splicing of

17 polyadenylated RNA—particularly intron retention (RI). Investigation of RI following inhibition of

18 nascent transcription demonstrated that PRMTs inversely regulate RI post-transcriptionally.

19 Subsequent proteomic analysis of chromatin-associated polyadenylated RNA identified aberrant

20 binding of the Type I substrate, CHTOP, and the Type II substrate, SmB. Targeted mutagenesis

21 of all methylarginine sites in SmD3, SmB, and SmD1 recapitulated splicing changes seen with

22 Type II PRMT inhibition. Conversely, mutagenesis of all methylarginine sites in CHTOP recapit-

23 ulated the splicing changes seen with Type I PRMT inhibition. Closer examination of subcellular

24 fractions indicated that RI were isolated to the nucleoplasm and chromatin. Together, these data

25 demonstrate that PRMTs regulate the post-transcriptional processing of nuclear, detained in-

26 trons through Sm and CHTOP arginine methylation.

27 INTRODUCTION

28 The mammalian genome encodes nine protein arginine methyltransferases (PRMTs 1-9;

29 PRMT4 is also known as CARM1). Arginine methylation is critical in regulating signal transduc-

30 tion, gene expression, and splicing (reviewed by Guccione and Richard 2019; Lorton and

31 Shechter 2019). For instance, an important function of PRMT5 along with its cofactors pICln and

32 MEP50 (also known as WDR77), is assembly of small nuclear ribonucleoproteins (snRNPs)—

33 core components of the spliceosomal machinery (Meister et al. 2001; Neuenkirchen et al. 2015).

34 This includes both non-enzymatic chaperoning of Sm proteins via PRMT5/pICln following their

35 translation and post-translational methylation of SmD1 (SNRPD1), SmD3 (SNRPD3), and

36 SmB/B’ (SNRPB), by PRMT5-MEP50 (Friesen et al. 2001; Meister et al. 2001; Paknia et al.

37 2016). Following their methylation, these Sm proteins are delivered to SMN where they are

38 bound to small nuclear RNAs (snRNAs) in preparation for further processing and eventual nu-

39 clear import (Boisvert et al. 2002; Meister and Fischer 2002; Pellizzoni et al. 2002). Disruption

40 of PRMT5 leads to numerous splicing defects, primarily exon skipping (SE) and intron retention

41 (RI) (Boisvert et al. 2002; Bezzi et al. 2013; Koh et al. 2015; Braun et al. 2017; Fedoriw et al.

42 2019; Fong et al. 2019; Radzisheuskaya et al. 2019; Tan et al. 2019; Li et al. 2021; Sachamitr

43 et al. 2021). Intron retention is a highly prevalent alternative splicing event in tumors and is

44 ubiquitous across cancer types (Dvinge and Bradley 2015). Therefore, understanding the mech-

45 anisms that govern RI and the connection to PRMTs is of great interest to further advance our

46 understanding of this disease.

47 RNA splicing can occur during transcriptional elongation or after the transcript has been

48 cleaved and released from polymerase (reviewed by Neugebauer 2019). Although the im-

49 portance of PRMT5 in preserving splicing integrity is clear, whether PRMT5 exerts its influence

50 over co- or post-transcriptional splicing is still unknown. Previous work has indirectly implicated

51 PRMT5 in the regulation of post-transcriptional splicing in Arabidopsis and as a regulator of

52 detained introns (DI)—unspliced introns in poly(A) RNA that remain nuclear and are removed

53 prior to cytoplasmic export (Braun et al. 2017; Jia et al. 2020). DI are highly conserved, enriched

54 in RNA processing factors, and their removal has been proposed as part of a feedback mecha-

55 nism to control protein expression during differentiation and cell stress (Yap et al. 2012; Wong

56 et al. 2013; Braunschweig et al. 2014; Boutz et al. 2015; Pimentel et al. 2016). Sm proteins have

57 been implicated in regulating DI as PRMT5 inhibition leads to loss of Sm methylation and a

58 simultaneous increase in DI (Braun et al. 2017). Nonetheless, a direct connection of DI or RI

59 levels to Sm protein function and arginine methylation has yet to be demonstrated.

60 The mechanistic role of Type I PRMTs (PRMT1-4, 6, 8) in splicing is still poorly understood.

61 Recent reports demonstrate that there are consequences on SE following Type I PRMT inhibi-

62 tion, with many of the proteins altered in methylation having a potential role in RNA post-tran-

63 scriptional regulation (Fedoriw et al. 2019; Fong et al. 2019). Consistent with a role in post-

64 transcriptional processing, the primary Type I methyltransferase—PRMT1—has been implicated

65 in the regulation of RNA export through methylation of the Transcription and Export (TREX)

66 components—Aly and REF export factor (ALYREF) and chromatin target of PRMT1 (CHTOP)

67 (Hung et al. 2010; Van Dijk et al. 2010; Chang et al. 2013). Loss of CHTOP leads to an accu-

68 mulation of nuclear polyadenylated transcripts due to a block in RNA export (Chang et al. 2013).

69 Furthermore, ALYREF and CHTOP are both deposited co-transcriptionally along RNA and are

70 bound to RI (Viphakone et al. 2019). As the nuclear export of RNA is intimately linked to splic-

71 ing—dissociation of splicing factors is a prerequisite for successful export—regulation of this

72 process is an attractive means to control RI (Luo and Reed 1999). Despite this, the role of nu-

73 clear export factors and the impact of Type I PRMTs on regulating RI remains unexplored.

74 Here, we test both the co- and post-transcriptional consequences of PRMT5 or Type I PRMT

75 inhibition (PRMTi). We also examine which factors are responsible for PRMTi-mediated splicing

76 consequences. As both Type I PRMTs and PRMT5 have been extensively reported to be re-

77 quired for the pathogenicity of lung cancer (Hwang et al. 2021)—yet their role in the transcription

78 and splicing of this disease remains largely unstudied—we used A549 human alveolar adeno-

79 carcinoma cells as our model. Using GSK591 (also known as EPZ015866 or GSK3203591), a

80 potent and selective inhibitor of PRMT5 (Duncan et al. 2016), and MS023, a potent pan-Type I

81 inhibitor (Eram et al. 2016), we demonstrate that PRMTs inversely regulate RI post-transcrip-

82 tionally through Sm and CHTOP arginine methylation.

83 RESULTS

84 Type I PRMT or PRMT5 inhibition promotes changes in alternative splicing

85 PRMTs consume S-adenosyl methionine (SAM) and produce S-adenosyl homocysteine

86 (SAH) to catalyze the post-translational methylation of either one or both terminal nitrogen atoms

87 of the guanidino group of arginine (Figure 1a) (Gary and Clarke 1998). All PRMTs can generate

88 monomethyl arginine (Rme1). Type I PRMTs further catalyze the formation of asymmetric

89 NG,NG-dimethylarginine (Rme2a); Type II PRMTs (PRMT5 and 9) form symmetric NG,N’G-dime-

90 thylarginine (Rme2s). PRMT5 is the primary Type II methyltransferase (Yang et al. 2015). As

91 previous reports indicated that lengthy treatment with PRMTi promotes aberrant RNA splicing,

92 we wanted to determine whether alternative splicing differences with PRMTi occurred as early

93 as day two and, if so, how they evolved over time (Bezzi et al. 2013; Fong et al. 2019;

94 Radzisheuskaya et al. 2019; Tan et al. 2019; Li et al. 2021; Sachamitr et al. 2021). Therefore,

95 we performed poly(A)-RNA sequencing on A549 cells treated with DMSO, GSK591, MS023, or

96 both inhibitors in combination for two-, four-, and seven-days. Sequencing at seven days for co-

97 treatment was not performed as this resulted in significant cell death (not shown). Using replicate

98 Multivariate Analysis of Transcript Splicing (rMATS) (Shen et al. 2014) to identify alternative

99 splicing events, at day two we observed significant differences in RI with PRMTi (FDR < 0.05)

100 relative to DMSO (Figure 1b). Furthermore, whereas RI were increased with GSK591- and co-

101 treatment—signified by a positive difference in percent spliced in (+ΔPSI or +ΔΨ)—RI were de-

102 creased with MS023 treatment (-ΔΨ) relative to DMSO.

103 We next intersected all common RI between treatment conditions in A549 at day two (Figure

104 1c). We found 239 common introns. Strikingly, whereas GSK591 and co-treatment with GSK591

105 and MS023 had increased inclusion, for the same introns MS023 alone resulted in decreased

106 inclusion relative to DMSO (Figure 1c). To confirm and quantify the presence of PRMT-regu-

107 lated RI by RT-qPCR, we selected three genes, GAS5, ANKZF1, and NOP2. These were iden-

108 tified by rMATS as containing inversely regulated RI by either PRMT5 or Type I PRMTi (Figure

109 1d). Notably, while the expression of GAS5 and NOP2 was increased with GSK591, it was un-

110 changed with MS023. Using GAS5 intron 9, ANKZF1 Intron 9, and NOP2 Intron 14 primers that

111 spanned the intron-exon boundaries and normalized by exonic primers, at day two following

112 PRMTi we confirmed the presence of the RI (Figure 1e). Consistent with our rMATS data, GAS5

113 intron 9, ANKZF1 intron 9, and NOP2 intron 14 had increased retention upon GSK591 treatment

114 (P = 0.141, 0.016, 0.001, respectively). Furthermore, despite decreased dynamic range inherent

115 to quantifying intronic signal loss, we observed less retention with MS023 treatment (P < 0.0001,

116 0.064, 0.040, respectively) (Figure 1e).

117 To confirm that these RI were specific to PRMTi and not a consequence of off-target effects

118 of GSK591 or MS023 we generated A549 cells expressing dCas9-KRAB-MeCP2 (Yeo et al.

119 2018). Independent transduction of two unique guide RNAs (gRNA) for four days targeting ei-

120 ther the major Type I methyltransferases—PRMT1 and PRMT4—or the Type II methyltransfer-

121 ase, PRMT5 achieved efficient knockdown (Suppl. Figure 1a). Furthermore, we recapitulated

122 increased RI in GAS5 intron 9, ANKZF1 intron 9, and NOP2 intron 14 by RT-qPCR following

123 knockdown of PRMT5 with both unique gRNAs (Suppl. Figure 1b). We also observed re-

124 duced RI in GAS5 intron 9, ANKZF1 intron 9, but not NOP2 intron 14 when analyzing RI fol-

125 lowing transduction with the stronger PRMT1 gRNA (Suppl. Figure 1b). Knockdown of

126 PRMT4 only reduced RI in GAS5 intron 9 with one gRNA, but not in ANKZF1 intron 9 or NOP2

127 intron 14. Together, these data support that the RI seen following Type I or PRMT5i are spe-

128 cific, yet the consequences of MS023 treatment on RI are likely more robust than CRISPR in-

129 terference due to the inhibition of multiple Type I enzymes and the ability of PRMTs to scav-

130 enge each other’s substrates (Dhar et al. 2013; Eram et al. 2016).

131 PRMTi-mediated RI are conserved across cell types and share common characteristics

132 We next asked whether the RI in our data were common to other datasets in which PRMT

133 activity was perturbed. To accomplish this, we used rMATS on publicly available data (Braun et

134 al. 2017; Fedoriw et al. 2019; Fong et al. 2019; Radzisheuskaya et al. 2019). In these experi-

135 ments–conducted in a variety of cell lines from diseases including acute myeloid leukemia (THP-

136 1), chronic myeloid leukemia (K562), pancreatic adenocarcinoma (PANC03.27), and glioblas-

137 toma (U87)–arginine methylation was inhibited via PRMT knockdown using CRISPRi or with

138 PRMTi. As demonstrated by the high odds ratio (log2(OR) > 6) between all the datasets, we

139 showed that there was a highly significant (Fisher’s exact adjusted P < 1e-05) overlap in RI

140 (Suppl. Figure 2a). This further supports that the consequences of PRMTi on RI are not off-

141 target effects of the inhibitors and are also common to multiple cell types.

142 RI have previously been reported to have common characteristics such as being shorter,

143 closer to transcription end site (TES), and having reduced splice site strength (Braunschweig et

144 al. 2014). To determine the common characteristics of the PRMT-regulated RI, we analyzed

145 their length, distances to the TES, and sequences. We found that the RI were significantly closer

146 to the TES and shorter when compared to the genomic distribution of A549 expressed introns

147 (P < 2.2e-16) (Figure 1f and 1g). Moreover, in analyzing the probability of nucleotide distribution

148 at the 5’ and 3’ splice sites we noted both a preference for guanine three nucleotides downstream

149 of the 5’ splice site and increased frequency of cytosine in the polypyrimidine tract (Suppl. Fig-

150 ure 2b). This is consistent with previous literature demonstrating that RI have common features

151 contributing to their persistence in poly(A) RNA (Bezzi et al. 2013; Braunschweig et al. 2014;

152 Braun et al. 2017; Tan et al. 2019).

153 PRMT-dependent changes in co-transcriptional splicing do not reflect changes in mRNA

154 The inverse relationship on RI by Type I or PRMT5i in addition to their TES-proximal location,

155 shorter length, and weaker splice sites led us to investigate the kinetics of co-transcriptional

156 splicing. To accomplish this, we used Splicing Kinetics and Transcript Elongation Rates by Se-

157 quencing (SKaTER seq) (Casill et al. 2021). As splicing changes are present as early as two

158 days following GSK591 or MS023 treatment, we used this time point of PRMTi for our analysis.

159 Briefly, SKaTER seq uses a three-hour 5,6-dichloro-1-β-D-ribofuranosylbenzimidazole (DRB)

160 treatment to synchronize transcription, followed by a rapid wash-out to allow productive elonga-

161 tion to commence. Once RNA pol II begins elongating, starting at 10 minutes nascent RNA is

162 collected every five minutes until 35 minutes post-DRB washout (Figure 2a). Nascent RNA

163 isolated via a 1 M urea wash of chromatin and an additional poly(A)-RNA depletion is then se-

164 quenced. The rate of nascent RNA formation—including: (1) RNA pol II initiation and pause-

165 release (spawn) rate, (2) elongation rate, (3) splicing rate, and (4) transcript cleavage rate—is

166 then calculated by using a comprehensive model that determines the rates that best fit the se-

167 quencing coverage (Casill et al. 2021).

168 To assess the accuracy of the rates determined by the SKaTER model, we used the spawn,

169 elongation, splicing, and cleavage rates to simulate a predicted poly(A)-RNA cassette exon Ψ

170 and compared these results with our poly(A)-RNA sequencing. We successfully predicted cas-

171 sette exon Ψ detected in poly(A) RNA in DMSO (Spearman’s correlation coefficient (ρ) = 0.54),

172 GSK591 (ρ = 0.61), and MS023 (ρ = 0.55) (P < 2.2e-16) treatment conditions (Figure 2b). Next,

173 we compared spawn rate to poly(A)-RNA sequencing transcripts per million (TPM). As predicted,

174 we observed a strongly significant (P < 2.2e-16) correlation between RNA pol II spawn rate and

175 TPM in DMSO-, GSK591-, and MS023-treated cells (ρ = 0.50, 0.51, 0.51, respectively) (Suppl.

176 Figure 3a) (Casill et al. 2021). Consistent with previously published reports, a comparison of

177 splicing rates within each condition confirmed that constitutive introns splice faster than cassette

178 exons as well as alternative 5’ and 3’ splice sites (Suppl. Figure 3b) (Pandya-Jones et al. 2013).

179 As poly(A)-RNA sequencing revealed opposing alternative splicing changes—specifically in-

180 creased RI with GSK591 and decreased RI with MS023—we next asked how PRMTi affected

181 the global distribution of splicing rates relative to DMSO. Surprisingly, GSK591 treatment did not

182 significantly change the median of this distribution (Wilcoxon rank-sum test, P > 0.05), while

183 MS023 treatment resulted in significantly slower global splicing rates relative to DMSO (P < 2.2e-

184 16) (Figure 2c). As our primary interest was in RI, we compared the splicing rates of introns that

185 were retained with PRMTi to the same introns in DMSO-treated cells. We found no significant

186 difference in the median of GSK591-treated cells compared to DMSO (P = 0.53), while MS023

187 treatment similarly led to slower splicing rates (P = 0.07) (Figure 2d). This result was incon-

188 sistent with the changes seen in poly(A) RNA as a slower splicing rate should increase RI.

189 Therefore, the co-transcriptional splicing rate suggested that changes in RI following PRMTi

190 were not due to co-transcriptional splicing.

191 RI share unique characteristics that limit their co-transcriptional removal

192 We next asked how RI splicing rates compared to non-RI. We observed that introns that

193 tended to be retained in GSK591- or MS023-treated cells were slower to splice (P = 0.01 and P

194 = 0.04, respectively) when compared to the genomic distribution (Figure 2e). As we previously

195 found that RI were more likely to be closer to the TES and had slower splicing rates (Figure 1f

196 and 1g), we asked whether intron position correlated with splicing rate. Indeed, we determined

197 that introns located closer to the TES had slower splicing rates (Suppl. Figure 4a). When ex-

198 amining additional characteristics of RI, we noted that they had a significantly higher GC fraction

199 in both GSK591- (0.52, P = 1.63e-15) and MS023-treated cells (0.50, P = 3.89e-6) relative to

200 the global median (0.41) (Figure 2f). As GC content influences polymerase elongation, we an-

201 alyzed the intronic elongation rate of RI in comparison to non-RI. Consistent with an increased

202 GC content, we found that RI had slower elongation rates relative to non-RI (P = 0.07 and 0.02

203 for GSK591 and MS023, respectively) (Figure 2g). Next, as previous reports implicated poly-

204 merase elongation rate with co-transcriptional splicing outcomes—splice site availability dictates

205 splicing decisions—we wanted to understand if the slower elongation rate seen with RI compen-

206 sated for their shorter size (Fong et al. 2014). To accomplish this, we assessed how long RNA

207 pol II took to transcribe RI compared to non-RI. We found that despite their slower elongation

208 rate, RI transcription was completed significantly faster compared to non-RI in both GSK591-

209 and MS023-treated cells (P = 6.22e-17 and P = 8.12e-6, respectively) (Figure 2h). Given the

210 proximity of RI to the TES, this likely further decreased their ability to be removed co-transcrip-

211 tionally. Taken together, these results support the conclusion that RI in GSK591- and MS023-

212 treated cells share unique characteristics that increase their probability of post-transcriptional

213 processing.

214 PRMTs regulate RI post-transcriptionally

215 The paradox that RI are decreased with Type I PRMTi—despite their slower co-transcriptional

216 splicing—led us to hypothesize that PRMTs exert their control over splicing post-transcription-

217 ally. To test this hypothesis, we used the elongation, splicing, and cleavage rates determined by

218 SKaTER seq to calculate the probability that a transcript will be cleaved from RNA pol II prior to

219 a given intron being spliced. Consistent with most splicing being co-transcriptional, we found

220 that the median probability of cleavage prior to splicing was 9.7% in DMSO-treated cells and

221 8.7% in GSK591-treated cells, yet the probability with MS023 treatment was significantly higher

222 at 18.1% (Figure 3a). The global distribution was significantly reduced with GSK591 treatment

223 (P = 0.002) and increased with MS023 treatment (P < 2.2e-16) when compared to DMSO (Fig-

224 ure 3a). We next analyzed the probability of cleavage prior to splicing for RI. Consistent with the

225 hypothesis that PRMTs regulate splicing of RI post-transcriptionally, the median probability of

226 transcript cleavage prior to splicing was significantly higher for RI (50.6% and 63.2%, respec-

227 tively) when compared to non-RI (11.7% and 23.1%, respectively; P = 5.74e-10 and 0.0002)

228 (Figure 3b). Furthermore, intron position was strongly predictive of whether transcript cleavage

229 was likely to occur prior to splicing: TES proximal introns had a higher probability of cleavage

230 prior to their being spliced (Suppl. Figure 4b). All told, the proximity of RI to the TES, their

231 shorter length, and their slower splicing rate likely drives their decreased probability of being

232 spliced co-transcriptionally.

233 The slowing of co-transcriptional splicing following PRMTi suggests that there would be an

234 increase in observed RI. Paradoxically, poly(A)-RNA sequencing revealed that although RI are

235 increased in GSK591-treated cells, they are reduced in MS023-treated cells. This led us to hy-

236 pothesize that MS023 promotes more efficient post-transcriptional intron loss, whereas the op-

237 posite is true for GSK591. To test this model, we pre-treated A549 cells for two days with

238 GSK591 or MS023 and then blocked transcription using actinomycin D (ActD) for 60 minutes.

239 We then performed poly(A)-RNA sequencing to determine the post-transcriptional conse-

240 quences of PRMTi on RI (Figure 3c). Genome browser tracks demonstrated reduced RI and

241 flanking exon signal intensity following 60 minutes of ActD treatment (Figure 3d). Further quan-

242 tification of RI common to both GSK591- and MS023-treated cells relative to their -ActD con-

243 trols—normalized to the abundance of their flanking exons—demonstrated a negative median

244 log2 fold change in all conditions, indicating less RI abundance in the ActD-treated samples.

245 Moreover, consistent with a post-transcriptional mechanism, we observed significantly less in-

246 tron loss in GSK591-treated cells (median = -0.106) compared to DMSO- (-0.220) or MS023-

247 treated cells (-0.275) (P = 0.002 relative to DMSO) (Figure 3e). The RI loss following ActD in

248 MS023-treated cells was not significantly different than DMSO, although the median was more

249 negative suggesting a trend toward increased decay (P = 0.36) (Figure 3e).

250 We next validated these changes in RI following treatment with ActD using RT-qPCR for our

251 three candidate introns described above (Figure 1e). GSK591 treatment resulted in significantly

252 slower intron decay relative to DMSO for GAS5 intron 9 and NOP2 intron 14 but not ANKZF1

253 intron 9 after 60 minutes of ActD treatment (P = 0.06, 0.0005, 0.74, respectively) (Figure 3f).

254 Conversely, intron decay was faster with MS023 treatment relative to DMSO for GAS5 intron 9

255 but not ANKZF1 intron 9 or NOP2 intron 14 (P = 0.0043, 0.27, 0.98, respectively) (Figure 3f).

256 As the changes in RI following transcriptional inhibition with ActD reflect those of PRMTi alone—

257 increased RI with GSK591 and decreased RI with MS023—these results support that PRMTs

258 regulate RI post-transcriptionally.

259 PRMTi alters binding of RNA processing factors to chromatin-associated poly(A) RNA

260 We next sought to identify the factors responsible for mediating the post-transcriptional con-

261 sequences of PRMTi. Previous reports have highlighted that delayed co-transcriptional splicing

262 leads to chromatin retention of poly(A) transcripts (Brody et al. 2011; Pandya-Jones et al. 2013;

263 Yeom et al. 2021). Therefore, we analyzed differences in proteins bound to nuclear, and more

264 specifically, chromatin-associated poly(A) RNA. To accomplish this, we performed UV-crosslink-

265 ing of A549 cells two days after treatment with PRMTi. We then isolated the chromatin fraction,

266 and after a high-salt wash, sheared the material with a light, brief sonication. This was followed

267 by poly(A) enrichment using oligo(dT) beads. To control for non-specific interactions, we per-

268 formed high stringency washes and added an excess of free poly(A) to our negative control.

269 After elution of the bound poly(A) RNA, we digested the RNA and analyzed the remaining ma-

270 terial with liquid chromatography coupled online with tandem mass spectrometry (LC-MS/MS)

271 (Figure 4a).

272 We identified 1,832 unique proteins in the input chromatin across the three conditions (Suppl.

273 Table 1). Following GSK591 treatment, 118 had significantly altered abundance with 70 in-

274 creased and 48 decreased (P < 0.05). In the MS023-treated input chromatin, we observed 255

275 differentially enriched proteins with 150 increasing in abundance and 105 decreasing (P < 0.05).

276 The poly(A) enriched fraction—after background subtraction—contained 1,251 unique proteins.

277 Of these, 32 were differentially bound in the GSK591-treated samples with 24 increased and 8

278 decreased (P < 0.05). In the MS023-treated cells, there were 55 proteins with altered binding—

279 37 increased and 18 decreased (P < 0.05).

280 To identify the biological processes most affected by these inhibitors, we performed over rep-

281 resentation analysis of the top 200 enriched proteins (Suppl. Figures 5a and 5b). In the chro-

282 matin fraction of both GSK591- and MS023-treated cells—consistent with the gross aberrations

283 in splicing following treatment with these inhibitors—we observed a significant enrichment of

284 gene ontology terms for RNA splicing and ribonucleoprotein complex biogenesis (Padj < 0.05)

285 (Suppl. Figure 5a). Interestingly, in GSK591-treated cells, there was also a unique enrichment

286 for the regulation of protein localization to Cajal bodies (Padj < 0.05). Conversely, in MS023-

287 treated cells we observed a highly significant enrichment for nuclear RNA export and the regu-

288 lation of RNA catabolism (Padj < 0.05). When looking specifically at the poly(A)-enriched frac-

289 tion, we noted similar themes as those in the input chromatin (Suppl. Figure 5b).

290 PRMTi perturbs binding of nuclear export factors and snRNPs to poly(A) RNA

291 Of the proteins that were differentially bound to the input chromatin and poly(A) RNA following

292 treatment with PRMTi, we were interested in candidates that were most likely to mediate post-

293 transcriptional RI. To accomplish this, we took into consideration the enriched gene ontology

294 categories described above and highlighted the most significant factors involved in RNA splicing,

295 transport, and degradation relative to their log2 fold change (Figure 4b and 4c). In the input

296 chromatin of GSK591-treated cells, we observed increased signal in the snRNP chaperone,

297 SMN, as well as the snRNP component, SNRPB (SmB/B’) (Figure 4b). We also observed en-

298 richment of SNRPB in the poly(A)-enriched fraction (Figure 4c). Previous reports have docu-

299 mented changes in exon skipping following SNRPB knockdown—including within its own

300 transcript—that resemble changes seen with PRMT5 knockdown (Saltzman et al. 2011; Bezzi

301 et al. 2013). Together with the fundamental role of snRNPs in splicing, this suggested that

302 SNRPB may also be involved in the regulation of GSK591-mediated RI.

303 In the input chromatin of MS023-treated cells we observed an increase of the TREX compo-

304 nents ALYREF, DDX39B, and POLDIP3 relative to DMSO (Figure 4b). ALYREF and DDX39B

305 are recruited to pre-mRNA co-transcriptionally and are crucial for the eventual hand-over of the

306 mRNA to the nuclear export receptor, nuclear RNA export factor 1 (NXF1) (Heath et al. 2016).

307 When looking specifically at the poly(A) RNA fraction of MS023-treated cells, we noted an in-

308 crease in CHTOP, also a component of the TREX complex (Figure 4c) (Heath et al. 2016). Both

309 CHTOP and ALYREF have been shown to be direct targets of PRMT1—the major Type I me-

310 thyltransferase and target of MS023 (Hung et al. 2010; Van Dijk et al. 2010). In both cases,

311 arginine methylation of ALYREF and CHTOP was demonstrated to weaken their binding to RNA

312 and promote interaction with NXF1—consistent with their increase in both the input chromatin

313 and poly(A) RNA fractions following MS023-mediated inhibition of Type I PRMT activity (Hung

314 et al. 2010; Chang et al. 2013).

315 To further refine the factors likely to be responsible for PRMTi-dependent RI, we compared

316 these data with proteins that contained methylarginine in our PTMScan methylome data (Maron

317 et al. 2021). When we intersected the arginine methylome with the differentially enriched proteins

318 in our input and poly(A) enriched fractions (P < 0.1), there were 88 and 23 unique methylarginine

319 containing proteins, respectively. We then analyzed the top 15 differentially enriched proteins in

320 each fraction. In the input, we noted that ALYREF and SNRPB were both modified with

321 methylarginine (Figure 4d). Furthermore, when looking at the poly(A) fraction, we again ob-

322 served SNRPB as well as CHTOP (Figure 4e). The prior identification of ALYREF, CHTOP, and

323 SNRPB as methylarginine containing proteins suggested that these bona fide PRMT substrates

324 may be potential mediators of PRMT-dependent RI (Brahms et al. 2001; Hung et al. 2010; Van

325 Dijk et al. 2010; Maron et al. 2021).

326 To validate changes in CHTOP and SNRPB chromatin-association following PRMTi, we iso-

327 lated the chromatin fraction and probed for CHTOP and SNRPB. Consistent with abundant GR

328 repeats—the preferred substrate motif for PRMT1 and PRMT5—present in CHTOP, we ob-

329 served faster CHTOP gel migration consistent with hypomethylation in MS023-treated cells (Fig-

330 ure 4f). Furthermore, in line with the proteomic analysis of input chromatin following GSK591

331 treatment, there was increased signal intensity of SNRPB (Figure 4f). To further confirm the

332 changes in methylarginine on CHTOP and SNRPB, we performed an immunoprecipitation for

333 these proteins following PRMTi (Figure 4g). Subsequent analysis of methylarginine confirmed

334 loss of CHTOP Rme2a with MS023 treatment and SNRPB Rme2s with GSK591 treatment.

335 Taken together, the differential enrichment of ALYREF, DDX39B, and POLDIP3 on chromatin,

336 along with CHTOP on poly(A) RNA following treatment with MS023—in addition to their depend-

337 ence on Rme2a—presented the possibility that the TREX complex may have an important role

338 in the post-transcriptional consequences of Type I PRMTs on RI. Moreover, as snRNPs are

339 fundamental in splicing, increased SNRPB in both the input and poly(A) fractions following treat-

340 ment with GSK591—as well as the loss of Rme2s—prompted us to further evaluate the role of

341 SNRPB in the regulation of RI.

342 Knockdown of CHTOP and SNRPB recapitulates PRMTi-mediated RI

343 To address the question of whether CHTOP, ALYREF, or SNRPB were involved in regulation

344 of RI, we first checked whether there was any publicly available RNA-seq data in which these

345 proteins were perturbed. We found two independent datasets where SNRPB was knocked down

346 in U251 glioblastoma cells or HeLa cells (Saltzman et al. 2011; Correa et al. 2016). We also

347 identified two independent datasets where CHTOP or ALYREF were knocked down in HEK293T

348 cells or HeLa cells, respectively (Fan et al. 2019; Viphakone et al. 2019). Strikingly, after per-

349 forming rMATS on these datasets, we observed that both SNRPB knockdown (SNRPBkd) ex-

350 periments strongly recapitulated the increase in RI seen with GSK591 treatment (Figure 5a).

351 Conversely, although ALYREF knockdown did not significantly affect RI levels (Suppl. Figure

352 6a), CHTOP knockdown (CHTOPkd) resulted in a global decrease in RI inclusion paralleling that

353 seen with MS023 treatment (Figure 5a). To understand if the RI were common across datasets,

354 we intersected either the SNRPBkd or CHTOPkd RI with PRMTi-dependent RI (Figures 5b and

355 5c). There were 535 mutual RI between GSK591 and both the SNRPBkd datasets, most of

356 which had the same effect on ΔΨ (log2 Odds Ratio 7.84 for GBM and 7.56 for HeLa, P < 1e-

357 300) (Figure 5b). The opposite was true when comparing GSK591 to CHTOPkd—we observed

358 620 shared RI (log2 Odds Ratio 7.49, P < 1e-300) that largely contrasted in their ΔΨ (Figure

359 5b). MS023 treatment had the inverse effect of GSK591 when compared to SNRPBkd and

360 CHTOPkd. There were 164 overlapping RI between both SNRPBkd datasets (log2 Odds Ratio

361 6.68 for GBM and 6.38 for HeLa, P < 7e-290) with the majority having opposite ΔΨ values;

362 CHTOPkd and MS023 treatment shared 201 RI (log2 Odds Ratio 646, P < 2e-279), with the

363 majority having coincident ΔΨ values (Figure 5c). Consistent with PRMTs regulating a common

364 group of RI, SNRPBkd and CHTOPkd were significantly correlated with either GSK591 or

365 MS023 treatment (Figure 5d). Remarkably, the correlations that were positive for GSK591 treat-

366 ment were negative for MS023 treatment, reflecting the changes seen in poly(A)-RNA seq and

367 consistent with SNRPB and CHTOP mediating the effects of PRMTi on RI.

368 Sm arginine mutants increase RI

369 To address the question of whether Sm arginine methylation was directly involved in the reg-

370 ulation of RI, we designed a single vector containing SNRPD3, SNRPB, and SNRPD1—the Sm

371 targets of PRMT5—as either wildtype, R-to-A, or R-to-K mutants separated by 2a-self cleaving

372 peptides (P2A) and modified with FLAG- (SNRPD3 and SNRPB) or V5-affinity (SNRPD1) tags

373 at their N-termini (Figure 5e). To identify the appropriate arginines to mutate, we utilized our

374 PTMScan data as well as a previously published high-resolution crystal structure of the

375 U4/U6.U5 tri-snRNP (Charenton et al. 2019; Maron et al. 2021). As we recently showed that

376 PRMTs prefer to methylate intrinsically disordered regions (IDRs), we mutated all 29 arginines

377 within Sm IDRs and within GR repeats (Figure 5e) (Maron et al. 2021). We then transduced

378 A549 cells with the constructs and performed RT-qPCR for GAS5 intron 9, ANKZF1 intron 9, or

379 NOP2 intron 14. We achieved similar expression of the exogenous Sm proteins relative to the

380 endogenous between wildtype and mutant constructs at the level of RNA (Suppl. Figure 6b).

381 However, likely owing to the meticulous control of total Sm levels by the cell, we were unable to

382 achieve ample overexpression with the wildtype vector (Suppl. Figure 6c) (Prusty et al. 2017).

383 We observed a migratory shift toward a lower molecular weight in the R-to-A mutants when

384 compared to the wildtype or R-to-K mutants (Suppl. Figure 6c). We also observed a strong

385 Rme2s signal on the FLAG-SmB wildtype protein that was absent on the R-to-A and R-to-K

386 mutants (Suppl. Figure 6c). Furthermore, and consistent with a role for Sm methylarginine in

387 regulating RI levels, following Sm R-to-A mutagenesis, we observed increased inclusion in all

388 three RI candidates (P < 0.05) (Figure 5f). We also detected a significant increase in GAS5

389 intron 9 (P < 0.001) with the Sm R-to-K mutants and a trend toward increased RI in ANKZF1

390 intron 9 (P = 0.23) and NOP2 intron 14 (P = 0.16). Thus, mutagenesis of Sm methylarginine

391 sites increased RI in our candidate transcripts, yet the greater effect of the R-to-A mutants sug-

392 gests that the charge of arginine itself may also play an important role in regulating RI levels.

393 CHTOP arginine mutants decrease RI

394 We performed similar experiments with CHTOP whereby we mutated 30 arginines present

395 within the centrally located “GAR” motif—the preferred PRMT1 substrate recognition motif—to

396 either alanine or lysine (Figure 5g). This region of CHTOP has been previously shown to be

397 required for PRMT1-catalyzed methylarginine (Van Dijk et al. 2010). We transduced A549 cells

398 with the constructs and performed RT-qPCR for GAS5 intron 9, ANKZF1 intron 9, or NOP2 intron

399 14. We achieved similar expression of the mutant CHTOP proteins when compared to the

400 wildtype at the level of RNA and protein (Suppl. Figure 6c and 6d). Interestingly, when trans-

401 ducing cells with the wildtype CHTOP, we noted a gross downregulation of the endogenous

402 transcript (Suppl. Figure 6e). CHTOP has been previously reported to control its own expres-

403 sion as part of an autoregulatory loop (Izumikawa et al. 2016). We did not observe this compen-

404 sation with either the R-to-A or R-to-K mutants. Consistent with the gel shift in CHTOP seen

405 following MS023 treatment, we observed a similar change in the R-to-A mutant when compared

406 to the wildtype or R-to-K mutants. Moreover, supporting a role for CHTOP methylarginine in

407 Type I PRMT-dependent RI inclusion, in both R-to-A and R-to-K mutants we observed de-

408 creased inclusion in GAS5 intron 9 (P < 0.001) (Figure 5h). We also saw a trend toward de-

409 creased inclusion of ANKZF1 intron 9 (P = 0.23 and 0.26) and significantly decreased inclusion

410 of NOP2 intron 14 (P < 0.0001) in the R-to-K mutant, but not the R-to-A mutant. Taken together,

411 these experiments support that CHTOP methylarginine is involved in regulating RI levels.

412 PRMT-regulated RI are detained within the nucleoplasm and chromatin

413 PRMT5 has been proposed to specifically regulate DI—introns that persist in poly(A) RNA but

414 remain nuclear (Braun et al. 2017). However, direct experimental evidence—namely fractiona-

415 tion of subcellular compartments to identify the location of RI—has been lacking. To address the

416 question of whether the RI in our data are nuclear and therefore DI, we first ran rMATS on pub-

417 licly available ENCODE poly(A)-RNA seq from cytoplasmic and nuclear fractions of A549 cells

418 (ENCSR000CTL and ENCSR000CTM, respectively). We observed an enrichment of RI within

419 the nuclear fraction—97% of significant RI events (FDR < 0.05) had a +ΔΨ, where ΔΨ is the

420 difference in Ψ between the nuclear and cytoplasmic compartments. We then intersected these

421 data with the PRMTi-mediated RI and noted that there was a significant overlap with both

422 GSK591- (log2 OR 8.89, P < 1e-300) and MS023-treatment (log2 OR 7.75, P < 1e-300) when

423 compared to all A549 expressed introns (Figure 6a). We also observed that GAS5 intron 9,

424 ANKZF1 intron 9, and NOP2 Intron 14 were significantly increased in nuclear poly(A) RNA

425 (GAS5 and ANKZF1 not shown) (Figure 6b).

426 To validate that these RI remain nuclear following PRMTi, and to further resolve their lo-

427 calization within the nucleus, we treated A549 cells with PRMTi and isolated cytoplasmic, nucle-

428 oplasmic, and chromatin fractions. We then quantified levels of GAS5 intron 9, ANKZF1 intron

429 9, and NOP2 intron 14 or non-RI within the same poly(A) transcripts (Figure 6c). In accordance

430 with the RI being localized to the nucleus, in DMSO-, GSK591-, and MS023-treated cells, we

431 observed a significant increase in nucleoplasmic and chromatin signal compared to the cyto-

432 plasm (P < 0.05) (Figure 6c). Importantly, there was no significant difference between cytoplas-

433 mic or nuclear enrichment of the non-RI tested. Together, these data support that these RI are

434 indeed localized to the nucleoplasm and chromatin fractions and are thus DI. Therefore, we

435 conclude that Type I PRMTs and PRMT5 regulate post-transcriptional DI through modulation of

436 snRNP and CHTOP arginine methylation (Figure 7a). We also propose the following model for

437 PRMT-dependent DI regulation: (1) snRNP Rme2s is required for post-transcriptional maturation

438 of DI-containing transcripts—in the absence of Rme2s these snRNPs are bound to transcript but

439 the DI are sequestered from splicing or degradation (Figure 7b), (2) CHTOP Rme2a is neces-

440 sary for DI-containing transcripts to progress through the TREX-mediated nuclear export path-

441 way; CHTOP lacking Rme2a inhibits this progression but does not protect the mRNA, which

442 together with the consequent increase in nuclear residence time promotes splicing or degrada-

443 tion (Figure 7c).

444 DISCUSSION

445 In this study, our goal was to characterize the consequences of PRMTi on RI. Many reports

446 have documented that PRMT5 has a crucial role in splicing (Boisvert et al. 2002; Bezzi et al.

447 2013; Koh et al. 2015; Braun et al. 2017; Fedoriw et al. 2019; Fong et al. 2019; Radzisheuskaya

448 et al. 2019; Tan et al. 2019; Li et al. 2021; Sachamitr et al. 2021). However, the role of Type I

449 PRMTs in splicing has only recently been appreciated (Fedoriw et al. 2019; Fong et al. 2019).

450 This is the first study to report that Type I PRMTi decreases DI. Interestingly, in the presence of

451 both Type I and PRMT5i, DI were increased. This suggests that PRMT5 catalyzed Rme2s is

452 dominant in DI regulation. This is consistent with our model, whereby snRNP binding occurs

453 upstream of CHTOP-mediated intron detention and protects RNA from post-transcriptional splic-

454 ing or degradation (Figure 7a).

455 Furthermore, whether PRMTs regulate splicing co- or post-transcriptionally has not previously

456 been explored. We found that Type I PRMTi promoted slower co-transcriptional splicing, while

457 paradoxically resulting in less DI. While understanding the effects of Type I PRMTi on global

458 splicing is beyond this current work, one possibility may be that given the largely negative effect

459 arginine methylation has on protein-RNA binding (Blackwell and Ceman 2012), the absence of

460 Rme2a may inhibit or delay dissociation of RNA processing factors from nascent RNA thereby

461 leading to slower splicing. We and others have found that PRMT-dependent DI share character-

462 istics such as higher GC content, non-canonical splice sites, and location closer to the TES that

463 makes their co-transcriptional removal less likely (Wong et al. 2013; Braunschweig et al. 2014;

464 Boutz et al. 2015; Braun et al. 2017). Consistently, when modeling co-transcriptional kinetics of

465 DI compared to the global distribution, DI had slower splicing and elongation rates and took less

466 time to transcribe. Together with their TES-proximal location, this combination is likely to de-

467 crease the probability that splicing will be completed co-transcriptionally. These characteristics

468 were true for Type I PRMT inhibited cells as well—despite having less DI overall—indicating

469 their reliance on a post-transcriptional mechanism. As has been suggested by others, DI are

470 likely an evolutionarily conserved class of introns used for fine tuning the transcriptome

471 (Braunschweig et al. 2014; Boutz et al. 2015; Pimentel et al. 2016). This is further supported by

472 our comparison of DI in publicly available datasets, which—despite using diverse cell models

473 and methods of inhibiting PRMTs—had a strong overlap with the DI present in our data. The

474 same characteristics that lead these introns to become detained are likely what make them more

475 sensitive to perturbation of splicing components. This could lead them to be co-opted in disease

476 such as cancer where they have been shown to inactivate tumor suppressors and RNA export

477 proteins (Dvinge and Bradley 2015; Jung et al. 2015).

478 Indeed, when using actinomycin D to study post-transcriptional DI processing, PRMT5i re-

479 sulted in slower DI loss compared to DMSO or Type I PRMTi, whereas Type I PRMTi had faster

480 DI loss relative to DMSO and PRMT5i. Thus, the post-transcriptional—rather than co-transcrip-

481 tional—consequences reflect those seen in poly(A) RNA. Although little is known about how

482 post-transcriptional splicing is accomplished, previous work has demonstrated that post-tran-

483 scriptional spliceosomes are retained in nuclear speckles (Girard et al. 2012). Furthermore,

484 CHTOP and the TREX complex have also been found to localize within nuclear speckles (Dias

485 et al. 2010; Chang et al. 2013). Perturbation of either splicing factors or TREX components re-

486 sulted in accumulation of poly(A) RNA in nuclear speckles (Dias et al. 2010; Chang et al. 2013).

487 It is therefore possible that the two types of PRMTi coalesce at the level of this membraneless

488 organelle. Consistent with this is the observation that SRRM1 was enriched in the poly(A) frac-

489 tion in PRMT5i, whereas SRRM2 was enriched in the poly(A) fraction in Type I PRMTi (Figure

490 5b). Many SR-repeat proteins—including SRRM1 and SRRM2—are localized to nuclear speck-

491 les (Ilik et al. 2020). Given the importance of arginine methylation in regulating liquid-liquid phase

492 separation (LLPS) and the role of LLPS in maintaining nuclear speckle integrity—as well as other

493 membraneless organelles such as Cajal bodies—PRMTi may result in failed coordination and

494 spatial localization of RNA processing factors leading to aberrant post-transcriptional processing

495 (Courchaine et al. 2021; Liao and Regev 2021).

496 As delayed splicing has been shown to result in chromatin retention of transcripts (Brody et

497 al. 2011; Pandya-Jones et al. 2013; Yeom et al. 2021), we assayed chromatin-associated

498 poly(A) RNA to identify factors that regulate PRMTi-dependent DI. By interrogating changes in

499 proteins bound to chromatin-associated poly(A) RNA—and their methylarginine levels—we

500 identified snRNPs and the TREX complex component, CHTOP, as key-mediators of DI regula-

501 tion. Interestingly, CHTOP contains a retained intron within its own transcript that is used to

502 autoregulate its expression by nonsense-mediated decay (Izumikawa et al. 2016). This is ac-

503 complished by antagonizing intron excision by hnRNPH1 when CHTOP—using its “GAR”-mo-

504 tif—binds to its own transcript (Izumikawa et al. 2016). Therefore, it is possible that a similar

505 mechanism may be occurring globally whereby introns that are slow to splice are bound by

506 CHTOP—which has been shown to bind nascent RNA co-transcriptionally—and this prevents

507 further binding by hnRNP proteins and intron excision (Izumikawa et al. 2016; Viphakone et al.

508 2019).

509 Another mechanism may be explained by the role of CHTOP in alternative polyadenylation

510 (APA) (Viphakone et al. 2019). Previous work demonstrated that CHTOP specifically binds to

511 the 3’ end of transcripts and CHTOPkd promotes changes in APA (Viphakone et al. 2019). The

512 length of poly(A) is a driver of exosome-mediated decay, whereby abnormally long poly(A)

513 tails—seen in detained intron containing transcripts—drive exosome-mediated decay (Bresson

514 and Conrad 2013; Bresson et al. 2015; Muniz et al. 2015). Hyperadenylation occurs through

515 poly(A) binding protein nuclear 1 (PABPN1) dependent stimulation of poly(A) polymerases

516 (Bresson and Conrad 2013). In MS023-treated cells, PABPN1 was specifically enriched in the

517 chromatin and poly(A) fraction (P < 0.1) (Supplemental Table 1). Thus, loss of CHTOP Rme2a

518 may promote hyperadenylation and consequent exosomal degradation of DI-containing tran-

519 scripts.

520 As PRMTs have thousands of diverse substrates—many of which are involved in RNA home-

521 ostasis—the effect of PRMTs on splicing integrity is likely to be multifactorial (Guccione and

522 Richard 2019; Lorton and Shechter 2019). Here, we highlighted both snRNPs and CHTOP as

523 key mediators of PRMTi-regulated DI. However, given the importance of both CHTOP and

524 snRNPs in global RNA processing, why only a subset of alternative splicing events and tran-

525 scripts are affected by their perturbation remains to be known. Importantly, there are likely many

526 other factors involved in regulating PRMT-dependent alternative splicing. Future work will be

527 needed to characterize the methylarginine dependent mechanisms of post-transcriptional RNA

528 processing.

529 MATERIALS AND METHODS

530 Cell Culture

531 A549 and HEK293T cells were cultured in DMEM (Corning) supplemented with 10% FBS

532 (Hyclone), 100 µg/mL streptomycin and 100 I.U./mL penicillin (Corning) and maintained at 37

533 °C with humidity and 5% CO2. For this study, fresh cells were purchased from ATCC and tested

534 routinely for Mycoplasma (PCR, FWD primer: ACTCCTACGGGAGGCAGCAGT, REV primer:

535 TGCACCATCTGTCACTCTGTTAACCTC).

536 RT-qPCR

537 RNA purification was performed using RNeasy Plus (Qiagen). Isolated RNA was reverse tran-

538 scribed with Moloney murine leukemia virus (MMLV) reverse transcriptase (Invitrogen) and ol-

539 igo(dT) primers. LightCycler 480 Sybr Green I (Roche) master mix was used to quantitate cDNA

540 with a LightCycler 480 (Roche). An initial 95°C for 5 minutes was followed by 45 cycles of am-

541 plification using the following settings: 95°C for 15 s, 60°C for 1 minute. Primer sequences can

542 be found in Supplemental Table 2.

543 Poly(A)-RNA sequencing

544 RNA was extracted using RNeasy® Mini Kit (Qiagen) following the manufacturer’s protocol.

545 RNA quantitation and quality control were accomplished using the Bioanalyzer 2100 (Agilent

546 Technologies). Stranded RNA seq libraries were constructed by Novogene Genetics US. The

547 barcoded libraries were sequenced by Novogene on an Illumina platform using 150nt paired end

548 libraries generating ~30-40 million reads per replicate. Reads were trimmed and aligned to the

549 human genome (hg19) with Spliced Transcripts Alignment to a Reference (STAR) (Dobin et al.

550 2013). Alternative splicing events were determined using Replicate Multivariate Analysis of Tran-

551 script Splicing (rMATS, version 4.0.2) (Shen et al. 2014). Expression was determined using Kal-

552 listo (Bray et al. 2016). IGV (Broad Institute) was used as the genome browser. Graphs pertain-

553 ing to RNA seq were created using Gviz (Hahne and Ivanek 2016) in R (4.0.2) and assembled

554 in Adobe Illustrator 2020.

555 Splicing Kinetics and Transcript Elongation Rates by Sequencing

556 SKaTER seq was performed as described (Casill et al. 2021). Briefly, A549 cells were grown

557 with 0.01% DMSO, 1 µM GSK591 (Cayman), or 1 µM MS023 (Cayman) for two days followed

558 by addition of 100 µM DRB (Cayman). DRB-containing media was removed, and the cells were

559 incubated at 37 °C until the indicated time point. The cells were washed once with 4°C PBS, and

560 lysed by addition of 1 mL CL buffer (25 mM Tris pH 7.9 at 4°C, 150 mM NaCl, 0.1 mM EDTA,

561 0.1% Triton X-100, 1 mM DTT) supplemented with protease inhibitor (Thermo) and Drosophila

562 melanogaster S2 cell spike-in. Next, lysate was centrifuged at 845 x g for 5 minutes at 4 °C. The

563 pellet resuspended in 1 mL CL buffer without S2 spike-in and incubated on ice for five minutes.

564 Repeat centrifugation was performed. The supernatant was removed, and cells resuspended in

565 100 µL GR buffer (20 mM Tris pH 7.9, 75 mM NaCl, 0.5 mM EDTA, 50% glycerol, 0.85 mM DTT)

566 followed by addition of 1.1 mL NL buffer (20 mM HEPES pH 7.6, 300 mM NaCl, 7.5 mM MgCl2,

567 1% NP-40, 1 mM DTT, 1 M Urea). Following a 15-minute incubation, the lysate was spun at

568 16,000 x g for 10 minutes and the resulting chromatin pellet was resuspended and stored in

569 TRIzol (Thermo, 15596026) at -80 °C. RNA isolation was followed by poly(A) depletion using

570 the NEBNext Poly(A) mRNA magnetic isolation module (NEB, E7490L). RNA quantitation and

571 quality control were accomplished using the Bioanalyzer 2100 (Agilent Technologies). Stranded

572 RNA-seq libraries were prepared using the KAPA RNA HyperPrep Kit with RiboErase (HMR)

573 and KAPA Unique Dual-Indexed Adapters (Roche) according to instructions provided by the

574 manufacturer. The barcoded paired-end libraries were sequenced by Novogene using a No-

575 vaSeq S4, generating ~70 million reads per replicate.

576 Actinomycin D post-transcriptional processing

577 A549 cells were grown in the presence of 0.01% DMSO, 1 µM GSK591 (Cayman), or 1 µM

578 MS023 (Cayman) for two days. Following a two-day incubation, the media was removed and

579 replaced with media containing 5 µg/mL actinomycin D (Sigma) with 0.01% DMSO, 1 µM

580 GSK591, or 1 µM MS023 for 60 minutes. RNA was isolated using RNeasy® Mini Kit (Qiagen)

581 and poly(A)-RNA sequencing was performed as above.

582 Chromatin associated poly(A)-RNA enrichment and LC-MS/MS

583 Poly(A)-RNA isolation was performed with modifications to a previously described protocol

584 (Iadevaia et al. 2018). A549 cells were grown in the presence of 0.01% DMSO, 1 µM GSK591

585 (Cayman), or 1 µM MS023 (Cayman) for two days. Cells were washed with 4 °C PBS and irra-

586 diated on ice with 100 mJ cm-2 in a UV Stratalinker 1800. Cells were centrifuged at 500 x g for

587 10 minutes at 4 °C. Chromatin was isolated with nuclear lysis buffer (NLB; 10 mM Tris-HCl pH

588 7.5 at 4°C, 0.1% NP-40, 400 mM KCl, 1 mM DTT) supplemented with 40 U/mL RNaseOUT

589 (Thermo), protease inhibitor (Thermo), and phosphatase inhibitor (Thermo). The chromatin pel-

590 let was resuspended in NLB and sonicated for five seconds at 20% amplitude with a probe-tip

591 sonicator using a 1/8” tip. The sonicate was centrifuged at 10,000 x g for 10 minutes and the

592 soluble material transferred to a low-adhesion RNase-free microcentrifuge tube. An aliquot from

593 each sample was saved to serve as the unenriched control. The samples were split into two

594 separate tubes, one of which received 10 µg of competitor 25-nt poly(A) RNA. Magnetic oligo-

595 (dT) beads (NEB) were equilibrated in NLB and added to the enrichments. The samples were

596 vortexed at room temperature for 10 minutes. The beads were then captured on a magnetic

597 column, and the supernatant transferred to fresh tube for additional rounds of depletion. The

598 beads were washed once with buffer A (10 mM Tris pH 7.5, 600 mM KCl, 1 mM EDTA, 0.1%

599 Triton X-100), followed by buffer B (10 mM Tris pH 7.5, 600 mM KCl, 1 mM EDTA) and lastly

600 buffer C (10 mM Tris pH 7.5, 200 mM KCl, 1 mM EDTA). The RNA was eluted by incubating the

601 beads in 10 µL of 10 mM Tris pH 7.5 at 80 °C for two minutes, capture of magnetic beads using

602 a magnetic column, and quickly transferring the supernatant to a new tube. The beads were then

603 used for two additional rounds of poly(A)-RNA capture.

604 The (un)enriched proteome were treated with Benzonase (Sigma) and then digested with

605 trypsin (Promega) prior to performing LC-MS/MS. Proteins were resuspended in 5% SDS and 5

606 mM DTT and left incubating for 1 hour at room temperature. Samples were then alkylated with

607 20 mM iodoacetamide in the dark for 30 minutes. Afterward, phosphoric acid was added to the

608 sample at a final concentration of 1.2%. Samples were diluted in six volumes of binding buffer

609 (90% methanol and 10 mM ammonium bicarbonate, pH 8.0). After gentle mixing, the protein

610 solution was loaded to an S-trap filter (Protifi) 96-well plate and spun at 500 x g for 30 sec. The

611 sample was washed twice with binding buffer. Finally, 100 ng of sequencing grade trypsin

612 (Promega), diluted in 50 mM ammonium bicarbonate, was added into the S-trap filter and sam-

613 ples were digested at 37 oC for 18 hours. Peptides were eluted in three steps: (i) 40 µl of 50 mM

614 ammonium bicarbonate, (ii) 40 µl of 0.1% TFA and (iii) 40 µl of 60% acetonitrile and 0.1% TFA.

615 The peptide solution was pooled, spun at 1,000 x g for 30 sec and dried in a vacuum centrifuge.

616 Prior to mass spectrometry analysis, samples were desalted using a 96-well plate filter

617 (Orochem) packed with 1 mg of Oasis HLB C-18 resin (Waters). Briefly, the samples were re-

618 suspended in 100 µL of 0.1% TFA and loaded onto the HLB resin, which was previously equili-

619 brated using 100 µL of the same buffer. After washing with 100 µL of 0.1% TFA, the samples

620 were eluted with a buffer containing 70 µL of 60% acetonitrile and 0.1% TFA and then dried in a

621 vacuum centrifuge. For LC-MS/MS acquisition, samples were resuspended in 10 µL of 0.1%

622 TFA and loaded onto a Dionex RSLC Ultimate 300 (Thermo Scientific), coupled online with an

623 Orbitrap Fusion Lumos (Thermo Scientific). Chromatographic separation was performed with a

624 two-column system, consisting of a C-18 trap cartridge (300 µm ID, 5 mm length) and a picofrit

625 analytical column (75 µm ID, 25 cm length) packed in-house with reversed-phase Repro-Sil Pur

626 C18-AQ 3 µm resin. Peptides were separated using a 60-minute gradient from 4-30% buffer B

627 (buffer A: 0.1% formic acid, buffer B: 80% acetonitrile + 0.1% formic acid) at a flow rate of 300

628 nL/min. The mass spectrometer was set to acquire spectra in a data-dependent acquisition

629 (DDA) mode. The full MS scan was set to 300-1200 m/z in the orbitrap with a resolution of

630 120,000 (at 200 m/z) and an AGC target of 5 x 105. MS/MS was performed in the ion trap using

631 the top speed mode (2 secs), an AGC target of 104 and an HCD collision energy of 35. Raw files

632 were searched using Proteome Discoverer software (v2.4, Thermo Scientific) using SEQUEST

633 search engine with the SwissProt human database. The search for total proteome included var-

634 iable modification of N-terminal acetylation and fixed modification of carbamidomethyl cysteine.

635 Trypsin was specified as the digestive enzyme with up to 2 missed cleavages allowed. Mass

636 tolerance was set to 10 pm for precursor ions and 0.2 Da for product ions. Peptide and protein

637 false discovery rate was set to 1%.

638 Data were transformed, normalized and statistics were applied as described previously

639 (Aguilan et al. 2020). Briefly, the data were log2 transformed and normalized by the average of

640 the data distribution. Statistical analysis was performed by using a two-tail heteroscedastic t-

641 test.

642 Immunoprecipitation

643 A549 cells were grown with 0.01% DMSO, 1 µM GSK591 (Cayman), or 1 µM MS023 (Cay-

644 man) for two days. Cells were harvested using trypsin (Corning) and washed once with 4 °C

645 PBS supplemented with PRMTi. Cells were then resuspended in RIPA buffer (1% NP-40, 150

646 mM NaCl, 50 mM Tris-HCl pH 8 at 4 °C, 0.25% sodium deoxycholate, 0.1% SDS, and 1 mM

647 EDTA) supplemented with 40 U/mL RNaseOUT (Thermo) and protease inhibitor (Thermo). Ly-

648 sates were incubated on ice for 10 minutes followed by sonication for five seconds at 20% am-

649 plitude with a probe-tip sonicator using a 1/8” tip. Lysates were then spun at 10,000 x g for 10

650 minutes at 4°C. Supernatants were transferred to new low-adhesion RNase-free microcentrifuge

651 tubes and normalized to the same protein concentration using bicinchoninic acid (Pierce). Pri-

652 mary antibody targeting CHTOP (LSBio, LS-C193506) or SNRPB (ProteinTech, 16807-1-AP)

653 was added followed by incubation overnight at 4°C with gentle rotation. The next morning, Pro-

654 tein G agarose (Millipore-Sigma 16-201) was equilibrated in lysis buffer and added to the lysates

655 at 4 °C with gentle rotation. The beads were washed three times with lysis buffer followed by

656 resuspension in 1x Laemmli buffer for western blotting with CHTOP (as above), SNRPB (as

657 above), Rme2s (CST, 13222), and Rme2a (CST, 13522) antibodies.

658 CRISPRi and expression of methylarginine mutants

659 The dCas9-KRAB-MeCP2 expression vector (VB900120-5303pyt) was purchased from Vec-

660 torBuilder (Santa Clara, CA). The PRMT-targeting gRNA parent vector (VB210119-1169qwd)

661 was purchased from VectorBuilder and custom gRNA sequences (Supplemental Table 2)

662 derived from CRISPick (Doench et al. 2016) were cloned as described previously (Ran et al.

663 2013) with the exception that BfuA1 (NEB; R0701S) was used to digest the parent vector. Sm

664 wildtype (VB210103-1026dg), R-to-A (VB210103-1027ttz), and R-to-K (VB210317-1185yxr) as

665 well as CHTOP wildtype (VB210427-1238rhx), R-to-A (VB210427-1241jrw), and R-to-K

666 (VB210427-1242gjh) expression vectors were cloned by VectorBuilder. Lentiviral particles con-

667 taining expression vectors were produced in HEK293T cells using calcium phosphate transfec-

668 tion. Transduction of A549 cells was accomplished by combining lentivirus with polybrene con-

669 taining media (4 µg/mL) and centrifuging at 30 °C for 90 minutes at 500 x g followed by incuba-

670 tion for 24 hours at 37 °C, 5% CO2 with humidity. Following the 24-hour incubation, lentiviral

671 containing media was removed and complete DMEM containing appropriate selection antibiotic

672 (Cayman) was added. A549 cells expressing dCas9-KRAB-MeCP2 were selected using 2 µg/mL

673 puromycin and then maintained in culture with 1 µg/mL puromycin. For PRMT knockdown and

674 Sm expression experiments, A549 cells were selected with 10 µg/mL blasticidin for 72 hours.

675 For CHTOP expression experiments, A549 cells were selected for 96 hours with 1 mg/mL G418.

676 Primers used in RT-qPCR experiments can be found in Supplemental Table 2. Lysis for west-

677 ern blotting was accomplished as above with FLAG (Sigma, F1804), Rme2s (as above), and

678 GAPDH (Abcam, ab9484) antibodies.

679 Cell fractionation

680 A549 cells were grown with 0.01% DMSO, 1 µM GSK591 (Cayman), or 1 µM MS023 (Cay-

681 man) for two days. Cells were harvested using trypsin (Corning) and washed once with 4 °C

682 PBS supplemented with PRMTi. Cells were resuspended in Hypotonic Lysis Buffer (10 mM Tris-

683 Cl pH 8, 0.1% NP-40, 1 mM KCl, 1.5 mM MgCl2, supplemented with protease inhibitor and 40

684 U/mL RNaseOUT) and rotated for 30 minutes at 4 °C. Cells were then centrifuged at 10,000 x g

685 for 10 minutes at 4 °C and the supernatant was kept as the cytosolic fraction. The nuclear pellet

686 was washed once with hypotonic buffer to reduce carry over then resuspendned in NLB (as

687 above) and rotated for 30 minutes at 4 °C. After centrifugation at 10,000 x g for 10 minutes the

688 supernatant was kept as the nucleoplasm and the remaining material was washed once with

689 NLB and kept as the chromatin fraction. RNA from each fraction was isolated using TRIzol

690 (Thermo) and RT-qPCR was performed as above. For western analysis, the chromatin fraction

691 was resuspended in NLB then sheared with sonication. This was followed by addition of Laemmli

692 buffer to 1x and blotting with H3 (Abcam, ab1791), CHTOP, and SNRPB antibodies (as above).

693 Statistical analysis

694 All western blots were performed independently at least twice. RT-qPCR was performed at

695 least three-times with independent biological replicates. Statistical analyses were performed us-

696 ing either Prism software (Version 8.3.1, GraphPad) or R (version 4.0.2). To compare general

697 distributions, the Kolmogorov-Smirnov test was used. To compare median distribution changes

698 the Wilcoxon rank-sum test was used. To account for differences in sample size between global

699 and RI distributions, random sampling from the global population equivalent to the number of RI

700 within the tested condition was performed. This process was repeated one thousand times after

701 which the median P was reported. To compare means where only two groups exist, independent

702 t-test was performed. To compare means where two categorical variables exist, two-way

703 ANOVA with post-hoc Dunnett’s test was performed. GeneOverlap with Fisher’s exact test was

704 used to determine odds ratios of RI overlap between different RNA-seq datasets (Shen 2020).

705 All code used to generate data in this manuscript can be found here:

706 https://github.com/Shechterlab/PRMTsRegulatePostTranscriptionalDI.

707 ACKNOWLEDGMENTS

708 This work was supported by the National Institutes of Health [R01GM108646 to D.S.,

709 GM57829 to C.C.Q., R01GM134379 to M.J.G], the American Lung Association Discovery Award

710 LCD-564723 [D.S] and the HIRSCHL/MONIQUE WEILL-CAULIER TRUSTS to M.J.G and also

711 to D.S..

712 We thank Dr. Martin Krzywinski (http://mkweb.bcgsc.ca/) for help with data visualization.

713 CONFLICT OF INTEREST

714 The authors declare that they have no conflicts of interest with the contents of this article. The

715 content is solely the responsibility of the authors and does not necessarily represent the official

716 views of the National Institutes of Health.

717 AUTHOR CONTRIBUTIONS

718 M.I.M. conceived and designed the project, performed biochemical and bioinformatic

719 experiments, interpreted resulting data, and wrote the manuscript. A.D.C. performed SKaTER-

720 seq analysis. V.G. performed bioinformatic analysis. S.S. performed LC-MS/MS analysis. M.J.G.

721 and C.C.Q. helped with experimental design and data interpretation. D.S. supervised the study

722 and helped with experimental design, data interpretation, and manuscript writing. All authors

723 reviewed the manuscript.

724 ACCESSION NUMBERS

725 Raw data for RNA seq and SKaTER seq is deposited under GEO (GSExxxxxx). Accessions

726 for publicly available RNA-seq data used in this study can be found in Supplemental Table 3.

727 Raw data for chromatin-associated poly(A) LC-MS/MS is deposited under Chorus (xxxx).

728 REFERENCES 729 Aguilan JT, Kulej K, Sidoli S. 2020. Guide for protein fold change and p-value calculation for 730 non-experts in proteomics. Mol Omics 16: 573-582.

731 Bezzi M, Teo SX, Muller J, Mok WC, Sahu SK, Vardy LA, Bonday ZQ, Guccione E. 2013. 732 Regulation of constitutive and alternative splicing by PRMT5 reveals a role for Mdm4 733 pre-mRNA in sensing defects in the spliceosomal machinery. Genes & Development 27: 734 1903-1916.

735 Blackwell E, Ceman S. 2012. Arginine methylation of RNA-binding proteins regulates cell 736 function and differentiation. Molecular Reproduction and Development 79: 163-175.

737 Boisvert FO-M, CôTé J, Boulanger M-C, CléRoux P, Bachand FO, Autexier C, Richard SP. 738 2002. Symmetrical dimethylarginine methylation is required for the localization of SMN 739 in Cajal bodies and pre-mRNA splicing. Journal of Cell Biology 159: 957-969.

740 Boutz PL, Bhutkar A, Sharp PA. 2015. Detained introns are a novel, widespread class of post- 741 transcriptionally spliced introns. Genes & Development 29: 63-80.

742 Brahms H, Meheus L, De Brabandere V, Fischer U, Lührmann R. 2001. Symmetrical 743 dimethylation of arginine residues in spliceosomal Sm protein B/B′ and the Sm-like 744 protein LSm4, and their interaction with the SMN protein. 7: 1531-1542.

745 Braun CJ, Stanciu M, Boutz PL, Patterson JC, Calligaris D, Higuchi F, Neupane R, Fenoglio S, 746 Cahill DP, Wakimoto H et al. 2017. Coordinated Splicing of Regulatory Detained Introns 747 within Oncogenic Transcripts Creates an Exploitable Vulnerability in Malignant Glioma. 748 Cancer Cell 32: 411-426.e411.

749 Braunschweig U, Barbosa-Morais NL, Pan Q, Nachman EN, Alipanahi B, Gonatopoulos- 750 Pournatzis T, Frey B, Irimia M, Blencowe BJ. 2014. Widespread intron retention in 751 mammals functionally tunes transcriptomes. Genome Research 24: 1774-1786.

752 Bray NL, Pimentel H, Melsted P, Pachter L. 2016. Near-optimal probabilistic RNA-seq 753 quantification. Nature Biotechnology 34: 525-527.

754 Bresson SM, Conrad NK. 2013. The Human Nuclear Poly(A)-Binding Protein Promotes RNA 755 Hyperadenylation and Decay. PLoS Genetics 9: e1003893.

756 Bresson SM, Hunter OV, Hunter AC, Conrad NK. 2015. Canonical Poly(A) Polymerase Activity 757 Promotes the Decay of a Wide Variety of Mammalian Nuclear RNAs. 11: e1005610.

758 Brody Y, Neufeld N, Bieberstein N, Causse SZ, Böhnlein E-M, Neugebauer KM, Darzacq X, 759 Shav-Tal Y. 2011. The In Vivo Kinetics of RNA Polymerase II Elongation during Co- 760 Transcriptional Splicing. PLoS Biology 9: e1000573.

761 Casill AD, Haimowitz AJ, Kosmyna B, Query CC, Ye K, Gamble MJ. 2021. Spatial organization 762 of transcript elongation and splicing kinetics. Cold Spring Harbor Laboratory.

763 Chang C-T, Hautbergue GM, Walsh MJ, Viphakone N, Van Dijk TB, Philipsen S, Wilson SA. 764 2013. Chtop is a component of the dynamic TREX mRNA export complex. The EMBO 765 Journal 32: 473-486.

766 Charenton C, Wilkinson ME, Nagai K. 2019. Mechanism of 5ʹ splice site transfer for human 767 spliceosome activation. Science 364: 362-367.

768 Correa BR, De Araujo PR, Qiao M, Burns SC, Chen C, Schlegel R, Agarwal S, Galante PAF, 769 Penalva LOF. 2016. Functional genomics analyses of RNA-binding proteins reveal the 770 splicing regulator SNRPB as an oncogenic candidate in glioblastoma. Genome Biology 771 17.

772 Courchaine EM, Barentine AES, Straube K, Lee D-R, Bewersdorf J, Neugebauer KM. 2021. 773 DMA-tudor interaction modules control the specificity of in vivo condensates. Cell.

774 Dhar S, Vemulapalli V, Patananan AN, Huang GL, Di Lorenzo A, Richard S, Comb MJ, Guo A, 775 Clarke SG, Bedford MT. 2013. Loss of the major Type I arginine methyltransferase 776 PRMT1 causes substrate scavenging by other PRMTs. Scientific Reports 3.

777 Dias AP, Dufu K, Lei H, Reed R. 2010. A role for TREX components in the release of spliced 778 mRNA from nuclear speckle domains. Nature Communications 1: 97.

779 Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, 780 Gingeras TR. 2013. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29: 15- 781 21.

782 Doench JG, Fusi N, Sullender M, Hegde M, Vaimberg EW, Donovan KF, Smith I, Tothova Z, 783 Wilen C, Orchard R et al. 2016. Optimized sgRNA design to maximize activity and 784 minimize off-target effects of CRISPR-Cas9. Nature Biotechnology 34: 184-191.

785 Duncan KW, Rioux N, Boriack-Sjodin PA, Munchhof MJ, Reiter LA, Majer CR, Jin L, Johnston 786 LD, Chan-Penebre E, Kuplast KG et al. 2016. Structure and Property Guided Design in 787 the Identification of PRMT5 Tool Compound EPZ015666. ACS Medicinal Chemistry 788 Letters 7: 162-166.

789 Dvinge H, Bradley RK. 2015. Widespread intron retention diversifies most cancer 790 transcriptomes. Genome Medicine 7.

791 Eram MS, Shen Y, Szewczyk MM, Wu H, Senisterra G, Li F, Butler KV, Kaniskan HÜ, Speed 792 BA, Dela Seña C et al. 2016. A Potent, Selective, and Cell-Active Inhibitor of Human 793 Type I Protein Arginine Methyltransferases. ACS Chemical Biology 11: 772-781.

794 Fan J, Wang K, Du X, Wang J, Chen S, Wang Y, Shi M, Zhang L, Wu X, Zheng D et al. 2019. 795 ALYREF links 3′-end processing to nuclear export of non-polyadenylated mRNA s. The 796 EMBO Journal 38: e99910.

797 Fedoriw A, Rajapurkar SR, O'Brien S, Gerhart SV, Mitchell LH, Adams ND, Rioux N, Lingaraj 798 T, Ribich SA, Pappalardi MB et al. 2019. Anti-tumor Activity of the Type I PRMT 799 Inhibitor, GSK3368715, Synergizes with PRMT5 Inhibition through MTAP Loss. Cancer 800 Cell.

801 Fong JY, Pignata L, Goy P-A, Kawabata KC, Lee SC-W, Koh CM, Musiani D, Massignani E, 802 Kotini AG, Penson A et al. 2019. Therapeutic Targeting of RNA Splicing Catalysis 803 through Inhibition of Protein Arginine Methylation. Cancer Cell 36: 194-209.e199.

804 Fong N, Kim H, Zhou Y, Ji X, Qiu J, Saldi T, Diener K, Jones K, Fu X-D, Bentley DL. 2014. 805 Pre-mRNA splicing is facilitated by an optimal RNA polymerase II elongation rate. 806 Genes & Development 28: 2663-2676.

807 Friesen WJ, Paushkin S, Wyce A, Massenet S, Pesiridis GS, Van Duyne G, Rappsilber J, 808 Mann M, Dreyfuss G. 2001. The Methylosome, a 20S Complex Containing JBP1 and 809 pICln, Produces Dimethylarginine-Modified Sm Proteins. 21: 8289-8300.

810 Gary JD, Clarke S. 1998. RNA and Protein Interactions Modulated by Protein Arginine 811 Methylation. pp. 65-131. Elsevier.

812 Girard C, Will CL, Peng J, Makarov EM, Kastner B, Lemm I, Urlaub H, Hartmuth K, Lührmann 813 R. 2012. Post-transcriptional spliceosomes are retained in nuclear speckles until 814 splicing completion. Nature Communications 3: 994.

815 Guccione E, Richard S. 2019. The regulation, functions and clinical relevance of arginine 816 methylation. Nature Reviews Molecular Cell Biology.

817 Hahne F, Ivanek R. 2016. Visualizing Genomic Data Using Gviz and Bioconductor. Methods 818 Mol Biol 1418: 335-351.

819 Heath CG, Viphakone N, Wilson SA. 2016. The role of TREX in gene expression and disease. 820 Biochemical Journal 473: 2911-2935.

821 Hung M-L, Hautbergue GM, Snijders APL, Dickman MJ, Wilson SA. 2010. Arginine 822 methylation of REF/ALY promotes efficient handover of mRNA to TAP/NXF1. Nucleic 823 Acids Research 38: 3351-3361.

824 Hwang JW, Cho Y, Bae G-U, Kim S-N, Kim YK. 2021. Protein arginine methyltransferases: 825 promising targets for cancer therapy. Experimental & Molecular Medicine 53: 788-808.

826 Iadevaia V, Matia-González AM, Gerber AP. 2018. An Oligonucleotide-based Tandem RNA 827 Isolation Procedure to Recover Eukaryotic mRNA-Protein Complexes. Journal of 828 Visualized Experiments.

829 Ilik İA, Malszycki M, Lübke AK, Schade C, Meierhofer D, Aktaş T. 2020. SON and SRRM2 are 830 essential for nuclear speckle formation. eLife 9.

831 Izumikawa K, Yoshikawa H, Ishikawa H, Nobe Y, Yamauchi Y, Philipsen S, Simpson RJ, Isobe 832 T, Takahashi N. 2016. Chtop (Chromatin target of Prmt1) auto-regulates its expression 833 level via intron retention and nonsense-mediated decay of its own mRNA. Nucleic Acids 834 Research: gkw831.

835 Jia J, Long Y, Zhang H, Li Z, Liu Z, Zhao Y, Lu D, Jin X, Deng X, Xia R et al. 2020. Post- 836 transcriptional splicing of nascent RNA contributes to widespread intron retention in 837 plants. Nat Plants.

838 Jung H, Lee D, Lee J, Park D, Kim YJ, Park W-Y, Hong D, Park PJ, Lee E. 2015. Intron 839 retention is a widespread mechanism of tumor-suppressor inactivation. Nature Genetics 840 47: 1242-1248.

841 Koh CM, Bezzi M, Low DHP, Ang WX, Teo SX, Gay FPH, Al-Haddawi M, Tan SY, Osato M, 842 Sabò A et al. 2015. MYC regulates the core pre-mRNA splicing machinery as an 843 essential step in lymphomagenesis. Nature 523: 96-100.

844 Li W-J, He Y-H, Yang J-J, Hu G-S, Lin Y-A, Ran T, Peng B-L, Xie B-L, Huang M-F, Gao X et 845 al. 2021. Profiling PRMT methylome reveals roles of hnRNPA1 arginine methylation in 846 RNA splicing and cell growth. Nature Communications 12.

847 Liao SE, Regev O. 2021. Splicing at the phase-separated nuclear speckle interface: a model. 848 Nucleic Acids Research 49: 636-645.

849 Lorton BM, Shechter D. 2019. Cellular consequences of arginine methylation. Cellular and 850 Molecular Life Sciences.

851 Luo MJ, Reed R. 1999. Splicing is required for rapid and efficient mRNA export in metazoans. 852 Proceedings of the National Academy of Sciences 96: 14937-14942.

853 Maron MI, Lehman SM, Gayatri S, Deangelo JD, Hegde S, Lorton B, Sun Y, Bai DL, Sidoli S, 854 Gupta V et al. 2021. Independent transcriptomic and proteomic regulation by type I and 855 II protein arginine methyltransferases. iScience: 102971.

856 Meister G, Eggert C, Bühler D, Brahms H, Kambach C, Fischer U. 2001. Methylation of Sm 857 proteins by a complex containing PRMT5 and the putative U snRNP assembly factor 858 pICln. Current Biology 11: 1990-1994.

859 Meister G, Fischer U. 2002. Assisted RNP assembly: SMN and PRMT5 complexes cooperate 860 in the formation of spliceosomal UsnRNPs. The EMBO Journal 21: 5853-5863.

861 Muniz L, Davidson L, West S. 2015. Poly(A) Polymerase and the Nuclear Poly(A) Binding 862 Protein, PABPN1, Coordinate the Splicing and Degradation of a Subset of Human Pre- 863 mRNAs. Molecular and Cellular Biology 35: 2218-2230.

864 Neuenkirchen N, Englbrecht C, Ohmer J, Ziegenhals T, Chari A, Fischer U. 2015. 865 Reconstitution of the human U snRNP assembly machinery reveals stepwise Sm 866 protein organization. The EMBO Journal 34: 1925-1941.

867 Neugebauer KM. 2019. Nascent RNA and the Coordination of Splicing with Transcription. Cold 868 Spring Harbor Perspectives in Biology 11: a032227.

869 Paknia E, Chari A, Stark H, Fischer U. 2016. The Ribosome Cooperates with the Assembly 870 Chaperone pICln to Initiate Formation of snRNPs. Cell Reports 16: 3103-3112.

871 Pandya-Jones A, Bhatt DM, Lin CH, Tong AJ, Smale ST, Black DL. 2013. Splicing kinetics and 872 transcript release from the chromatin compartment limit the rate of Lipid A-induced gene 873 expression. RNA 19: 811-827.

874 Pellizzoni L, Yong J, Dreyfuss G. 2002. Essential Role for the SMN Complex in the Specificity 875 of snRNP Assembly. Science 298: 1775-1779.

876 Pimentel H, Parra M, Gee SL, Mohandas N, Pachter L, Conboy JG. 2016. A dynamic intron 877 retention program enriched in RNA processing genes regulates gene expression during 878 terminal erythropoiesis. Nucleic Acids Research 44: 838-851.

879 Prusty AB, Meduri R, Prusty BK, Vanselow J, Schlosser A, Fischer U. 2017. Impaired 880 spliceosomal UsnRNP assembly leads to Sm mRNA down-regulation and Sm protein 881 degradation. Journal of Cell Biology 216: 2391-2407.

882 Radzisheuskaya A, Shliaha PV, Grinev V, Lorenzini E, Kovalchuk S, Shlyueva D, Gorshkov V, 883 Hendrickson RC, Jensen ON, Helin K. 2019. PRMT5 methylome profiling uncovers a

884 direct link to splicing regulation in acute myeloid leukemia. Nature Structural & 885 Molecular Biology 26: 999-1012.

886 Ran FA, Hsu PD, Wright J, Agarwala V, Scott DA, Zhang F. 2013. Genome engineering using 887 the CRISPR-Cas9 system. Nature Protocols 8: 2281-2308.

888 Sachamitr P, Ho JC, Ciamponi FE, Ba-Alawi W, Coutinho FJ, Guilhamon P, Kushida MM, 889 Cavalli FMG, Lee L, Rastegar N et al. 2021. PRMT5 inhibition disrupts splicing and 890 stemness in glioblastoma. Nature Communications 12.

891 Saltzman AL, Pan Q, Blencowe BJ. 2011. Regulation of alternative splicing by the core 892 spliceosomal machinery. Genes & Development 25: 373-384.

893 Shen L. 2020. GeneOverlap: Test and visualize gene overlaps. http://shenlab- 894 sinaigithubio/shenlab-sinai/.

895 Shen S, Park JW, Lu Z-X, Lin L, Henry MD, Wu YN, Zhou Q, Xing Y. 2014. rMATS: Robust 896 and flexible detection of differential alternative splicing from replicate RNA-Seq data. 897 Proceedings of the National Academy of Sciences 111: E5593-E5601.

898 Tan DQ, Li Y, Yang C, Li J, Tan SH, Chin DWL, Nakamura-Ishizu A, Yang H, Suda T. 2019. 899 PRMT5 Modulates Splicing for Genome Integrity and Preserves Proteostasis of 900 Hematopoietic Stem Cells. Cell Reports 26: 2316-2328.e2316.

901 Van Dijk TB, Gillemans N, Stein C, Fanis P, Demmers J, Van De Corput MT, Essers J, 902 Grosveld F, Bauer U-M, Philipsen S. 2010. Friend of Prmt1, a Novel Chromatin Target 903 of Protein Arginine Methyltransferases. Molecular and Cellular Biology 30: 260-272.

904 Viphakone N, Sudbery I, Griffith L, Heath CG, Sims D, Wilson SA. 2019. Co-transcriptional 905 Loading of RNA Export Factors Shapes the Human Transcriptome. Molecular Cell 75: 906 310-323.e318.

907 Wong JJ, Ritchie W, Olivia, Selbach M, Jason, Huang Y, Gao D, Pinello N, Gonzalez M, 908 Baidya K et al. 2013. Orchestrated Intron Retention Regulates Normal Granulocyte 909 Differentiation. Cell 154: 583-595.

910 Yang Y, Hadjikyriacou A, Xia Z, Gayatri S, Kim D, Zurita-Lopez C, Kelly R, Guo A, Li W, Clarke 911 SG et al. 2015. PRMT9 is a Type II methyltransferase that methylates the splicing factor 912 SAP145. 6: 6428.

913 Yap K, Lim ZQ, Khandelia P, Friedman B, Makeyev EV. 2012. Coordinated regulation of 914 neuronal mRNA steady-state levels through developmentally controlled intron retention. 915 Genes & Development 26: 1209-1223.

916 Yeo NC, Chavez A, Lance-Byrne A, Chan Y, Menn D, Milanova D, Kuo C-C, Guo X, Sharma 917 S, Tung A et al. 2018. An enhanced CRISPR repressor for targeted mammalian gene 918 regulation. Nature Methods 15: 611-616.

919 Yeom K-H, Pan Z, Lin C-H, Lim HY, Xiao W, Xing Y, Black DL. 2021. Tracking pre-mRNA 920 maturation across subcellular compartments identifies developmental gene regulation 921 through intron retention and nuclear anchoring. Genome Research 31: 1106-1119.

922

923 Figure Legends

924 Figure 1 – Type I PRMTs and PRMT5 inversely regulate intron retention 925 a. Overview of protein arginine methyltransferases and their catalyzed reactions. 926 b. Comparison of ΔΨ for retained introns (RI) following PRMTi where ΔΨ = Ψ (PRMTi) – Ψ 927 (DMSO). Significance determined using Kolmogorov-Smirnov test; * < 0.05, **** < 0.0001, 928 ns = not significant. 929 c. Comparison of ΔΨ z-score for common RI after two-day treatment with PRMTi. 930 d. Genome browser track of poly(A)-RNA seq aligned reads for GAS5, ANKZF1, and NOP2. 931 Yellow shading denotes RI. Scale (0.1 kb) indicated in lower right corner. 932 e. RT-qPCR of RI highlighted in panel d. Data are represented as mean ± SD. Significance 933 determined using Student’s t-test; * < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001, ns = not 934 significant. 935 f-g. Comparison of RI and A549-expressed intron distance to the transcription-end site (TES) (f) 936 or intron length (g) in log10(kb). Dashed line indicates genomic median; solid line within 937 boxplot is condition-specific median. Significance determined using Wilcoxon rank-sum 938 test; **** < 0.0001.

939 Figure 2 – RI share unique characteristics and are independent of PRMT-regulated co- 940 transcriptional splicing 941 a. Histogram of SKaTER-seq aligned reads across WASL. 942 b. Correlation of SKaTER-seq model simulated cassette exon Ψ versus poly(A)-RNA seq Ψ. 943 c. Distribution of splicing rates for common genomic introns. Dashed line indicates DMSO 944 median; solid line within boxplot is condition-specific median. Significance determined us- 945 ing Wilcoxon rank-sum test; **** < 0.0001, ns = not significant. 946 d. Distribution of splicing rates for common retained introns as in panel c. 947 e-h. Distribution of splicing rates (e), GC fraction (f), intronic elongation rate (g), and time to 948 transcribe (h) for RI (orange) and A549 expressed introns (dark gray) within the same 949 condition. Dashed line indicates genomic median; solid line within boxplot is condition- 950 specific median. Significance determined using Wilcoxon rank-sum test; * < 0.05, **** < 951 0.0001.

952 Figure 3 – Type I PRMTs and PRMT5 regulate RI post-transcriptionally 953 a. Probability of cleavage prior to splicing for common introns relative to DMSO. Dashed line 954 indicates DMSO median; solid line within boxplot is condition-specific median. Significance 955 determined using Wilcoxon rank-sum test; ** < 0.01, **** < 0.0001. 956 b. Probability of cleavage prior to splicing for RI (orange) and A549 expressed introns (dark 957 gray) within the same condition. Dashed line indicates genomic median; solid line within 958 boxplot is condition-specific median. Significance determined using Wilcoxon rank-sum 959 test; *** < 0.001, **** < 0.0001. 960 c. Overview of Actinomycin D (ActD) poly(A)-RNA sequencing protocol. 961 d. Genome browser track of poly(A)-RNA seq aligned reads (-) and (+) ActD for NOP2 962 e. Distribution of log2(+ActD/-ActD) abundance for common RI between GSK591 and MS023 963 relative to DMSO. Dashed line indicates DMSO median; solid line within boxplot is condi- 964 tion-specific median. Significance determined using Wilcoxon rank-sum test; ** < 0.01, *** 965 < 0.001.

966 f. RT-qPCR of RI (-) and (+) ActD relative to DMSO. Data are represented as mean ± SD. 967 Significance determined using Student’s t-test; ** < 0.01, *** < 0.001, ns = not significant.

968 Figure 4 – PRMTi promotes aberrant binding of RNA processing factors to chromatin- 969 associated poly(A) RNA 970 a. Overview of chromatin-associated poly(A)-RNA LC-MS/MS experiment. 971 b-c. Dot plot of proteins bound to chromatin (b) or chromatin-associated poly(A) RNA (c) relative 972 to DMSO. Circle size is proportional to -log2(P). Colored values denote factors with ontology 973 pertaining to RNA splicing, transport, or degradation. The names of the top 15 significant 974 proteins are labeled. Significance determined using a heteroscedastic t-test. 975 d-e. Venn diagram comparing proteins containing methylarginine (PTMscan) and those that 976 were differentially enriched in the chromatin (d) or chromatin-associated poly(A) (e) frac- 977 tions following PRMTi. 978 f. Western blot of chromatin following two-day treatment with DMSO, GSK591, or MS023. 979 g. Immunoprecipitation and analysis of CHTOP and SNRPB methylarginine following treat- 980 ment with DMSO, GSK591, or MS023 for two days.

981 Figure 5 – Sm and CHTOP methylarginine mediates PRMT-dependent RI 982 a. Comparison of ΔΨ for RI following SNRPB knockdown (kd), GSK591, MS023, and 983 CHTOPkd where ΔΨ = Ψ (Treatment) – Ψ (Control). 984 b-c. Comparison of ΔΨ z-score for common RI between SNRPBkd and CHTOPkd with either 985 GSK591 (b) or MS023 (c). 986 d. Spearman rank correlation of ΔΨ for common RI between SNRPBkd and CHTOPkd with 987 either GSK591 or MS023. 988 e. Schematic of Sm expression constructs where lollipops represent individual methylarg- 989 inines. 990 f. RT-qPCR of RI following transduction with Sm R-to-A or R-to-K mutants relative to wildtype 991 (WT). Data are represented as mean ± SD. Significance determined using Student’s t-test; 992 * < 0.05, *** < 0.001, **** < 0.0001. 993 g. Schematic of CHTOP expression construct as in panel e. 994 h. RT-qPCR of RI following transduction with CHTOP R-to-A or R-to-K mutants relative to 995 WT. Data are represented as mean ± SD. Significance determined using Student’s t-test; 996 *** < 0.001, **** < 0.0001, ns = not significant.

997 Figure 6 – PRMT-dependent RI are localized to the nucleoplasm and chromatin 998 a. Scatter plot of ΔΨ for common RI in A549 nuclear/cytoplasmic fractions and GSK591 (left, 999 green) or MS023 (right, purple) where ΔΨ = Ψ (Nuclear/PRMTi) – Ψ (Cytoplasm/DMSO). 1000 b. Genome browser track of poly(A)-RNA seq aligned reads for NOP2 in A549 cytoplasmic or 1001 nuclear fractions. 1002 c. RT-qPCR of RI or non-RI from cytoplasmic, nuclear, or chromatin fractions of A549 cells 1003 treated with DMSO, GSK591, or MS023 for two days. Data are represented as mean ± SD. 1004 Significance determined using two-way analysis of variance with Tukey’s multiple compar- 1005 isons test; * < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001, ns = not significant.

1006 Figure 7 – Type I PRMTs and PRMT5 regulate post-transcriptional intron detention 1007 through Sm and CHTOP arginine methylation 1008 a-c. Model figures representing an overview of Type I- and PRMT5-dependent regulation of 1009 detained introns (DI) (a). PRMT5-catalyzed Sm Rme2s is necessary for productive splicing 1010 and maturation of DI; absence of Rme2s results in detention of the parent transcript (b). 1011 Type I PRMT-catalyzed CHTOP Rme2a is required for appropriate polyadenylation and 1012 nuclear export of DI; absence of Rme2a increases nuclear residence time resulting in in- 1013 creased degradation or splicing (c). 1014

1015 SUPPLEMENTAL FIGURE LEGENDS

1016 Supplemental Figure 1 – PRMTkd recapitulates RI inclusion seen with PRMTi 1017 a-b. RT-qPCR of PRMT expression (a) or RI (b) 96 hours after transduction of A549 cells with 1018 gRNAs targeting the indicated PRMT. Data are represented as mean ± SD. Significance 1019 determined using Student’s t-test; * < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001.

1020 Supplemental Figure 2 – PRMTs regulate a conserved class of RI that share unique fea- 1021 tures 1022 a. Matrix comparing the log2(odds ratio) and significance as determined by the Fisher’s Exact 1023 Test of overlapping RI from indicated experimental models. 1024 b. Web logo diagram of nucleotide distribution probability of A549 expressed or retained in- 1025 trons.

1026 Supplemental Figure 3 – Transcript spawn rate correlates with expression and alterna- 1027 tive splicing events are slower than constitutive ones 1028 a. Correlation of RNA pol II initiation and pause-release (spawn rate) with poly(A)-RNA seq 1029 transcripts per million (TPM). x-axis is spawn rate in tertiles; y-axis is poly(A)-RNA seq 1030 ln(TPM). 1031 b. Distribution of splicing rate versus splicing event as determined by SKaTER-seq modeling. 1032 A5SS = alternative 5’ splice site, A3SS = alternative 3’ splice site. Dashed line indicates 1033 constitutive intron median; solid line within boxplot is event-specific median. Significance 1034 determined using Kolmogorov-Smirnov test; * < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001, 1035 ns = not significant.

1036 Supplemental Figure 4 – Splicing rate and probability of cleavage prior to splicing corre- 1037 late with intron position 1038 a. Correlation of splicing rate and intron position in cells treated with DMSO, GSK591, or 1039 MS023 for two-days. 1040 b. Cumulative distribution functions comparing the probability of transcript cleavage prior to 1041 intron splicing and intron position. Color range indicates intron position where light blue is 1042 closer to transcription start site and dark blue closer to transcription end site.

1043 Supplemental Figure 5 – PRMTi alters association of RNA splicing, localization, and pro- 1044 cessing factors to chromatin and chromatin-associated poly(A) RNA 1045 a-b. Enriched biological processes found in input chromatin (a) and chromatin-associated 1046 poly(A) RNA (b) in A549 cells treated with GSK591 or MS023 relative to DMSO.

1047 Supplemental Figure 6 – CHTOP and Sm arginine mutants phenocopy PRMTi 1048 a. Comparison of ΔΨ for RI following ALYREFkd and CHTOPkd where ΔΨ = Ψ (Treatment) 1049 – Ψ (Control). 1050 b. RT-qPCR of Exogenous Sm R-to-A and R-to-K mutants normalized to wildtype expression 1051 vector. Data are represented as mean ± SD. 1052 c. Western blot of total cell extract from A549 cells expressing CHTOP or Sm wildtype, R-to- 1053 A, or R-to-K mutants.

1054 d. RT-qPCR of Exogenous CHTOP R-to-A and R-to-K mutants normalized to wildtype ex- 1055 pression vector. Data are represented as mean ± SD. 1056 e. RT-qPCR of Endogenous CHTOP in A549 cells expressing CHTOP wildtype, R-to-A, or R- 1057 to-K mutants. Data are represented as mean ± SD. 1058

1059 Supplemental Table 1 – Chromatin-Associated Poly(A) RNA Enrichment 1060 Enrichment and relative abundance of proteins in chromatin or chromatin-associated 1061 poly(A) RNA following treatment with DMSO, GSK591, or MS023 for two days.

1062 Supplemental Table 2 – Primer Sequences for RT-qPCR and gRNAs

1063 Supplemental Table 3 – Publicly available data used in this study

bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyrightMaron holder for et. this al.preprint Figure 1 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made a available underd aCC-BY 4.0 International license.

GAS5 ANKZF1 NOP2 chr1:173,833,393-173,834,422 chr2:220,098,427-220,100,321 chr12:6,669,254-6,669,976 Type I PRMT1,2,3,4,6, and 8 Type I exon: 11 10 9 8 8 9 10 11 15 14 13 Type II PRMT5,9 CH3 Rme2a H2N N CH3 30000 5000 20000

Type I and II HN Rme1 MS023 CH3 0 H2N NH2 H2N NH 0 0 N 30000 5000 20000 HN HN H SAM O SAH CH3 CH3 Type II HN NH Rme2s 0 0 0 N N H H HN 30000 5000 20000 O O GSK591

N 0 0 0 H 0.1 kb O b Retained Intron Levels e Relative Intron Abundance Day 2 n = 1,660

GSK591 ns DMSO 1,860 **** GSK591 GSK591+ MS023 **** MS023 589 3 MS023 *** 2 Day 4 n = 1,969 * GSK591 * 2,021 **** 1 0.14 GSK591+ MS023 **** 786 MS023 0 Intron Inclusion Day 7 n = 1,486 / DMSO) log 2(PRMTi -1 *

GSK591 **** 0.06 776 MS023 -2 **** GAS5 ANKZF1 NOP2 Intron 9 Intron 9 Intron 14 -1 0 1 Less Included ΔΨ More Included c Day 2 Common Retained Intron Abundance f RI vs Genomic Intron Distance to TES (kb) g RI vs Genomic Intron Length (kb) GSK591+ GSK591 MS023 MS023 Intron Inclusion Relative to Control A549 Expressed Introns (z-score) **** **** **** GSK591 2 RI GSK591+MS023 2 2 1 MS023 0 n = 168 **** **** **** −1 −2 0 0 (Intron Length) 10 (Distance to TES) 10 log log

39 −2 −2

32 n = 1,660 1860 589 160,595 n = 1,660 1860 589 160,595 bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyrightMaron holder for et. this al.preprint Figure 2 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made a b available under aCC-BY 4.0 International license. Simluated vs Actual WASL chr7:123,319,997-123,391,057 Casette Exon Inclusion 200 1.00 DMSO GSK591 -DRB 0 MS023

+DRB 0.75

10 0.50 15 RNA−seq Ψ 20 0.25 25 Post-DRB (min) 30 0.00

35 Bottom Middle Top SKaTER-seq simulated Ψ (tertile)

c d e Genome vs RI Splicing Rate Genomic Splicing Rate RI Splicing Rate

ns DMSO vs GSK591 DMSO vs MS023 GSK591 MS023 **** DMSO **** GSK591 ns 0.07 * * 1 1 1 MS023 Genome RI

0 0 0 10 (events/min) 10 (events/min) (events/min) 10

−1 −1 −1 Splicing Rate log Splicing Rate log −2 −2 −2 Splicing Rate log

n = 16,991 (introns) n = 158 39 n = 31,759 242 32,084 72 f g h Genome vs RI GC Fraction Genome vs RI Intronic Elongation Rate Genome vs RI Time to Transcribe GSK591 MS023 GSK591 MS023 GSK591 MS023 **** **** * 0.07 **** **** Genome 2 0.8 RI

10 10 (min) 1 0.6

0 0.4 5

Intron GC Fraction −1 0.2

Time to Transcribe Intron log Transcribe to Time −2 0.0 Intronic Elongation Rate (bp/min * 10 ³ ) n = 31,759 242 32,084 72 0 n = 25,005 181 24,641 59 n = 25,005 181 26,641 59 bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyrightMaron holder for et. this al.preprint Figure 3 a (which wasGenome not certified by peer review) isb the author/funder, who Genomehas granted vs RIbioRxiv a license to display the cpreprint in perpetuity. It is made Probability of Cleavage Prior to Splicing availableProbability under aCC-BY of Cleavage 4.0 International Prior to Splicing license. A549 Cells GSK591 MS023 DMSO or PRMTi ** **** Genome (48h) **** **** *** DMSO RI Treated A549 Cells 0.9 GSK591 0.9 MS023 Actinomycin D

0.6 Transcription 0.6 1 hour

Post-Transcriptional Processing Probability Probability 0.3 0.3 Processed Poly(A)-RNA

0.0 0.0 n = 8,253 (per condition) 18,220 126 21,458 37 Poly(A)-RNA Sequencing d e NOP2 chr12:6,669,263-6,669,738 Post ActD Intron Decay exon:15 14 15 14 15 14 ** 0.36 2 DMSO 15000 15000 15000 5000 *** GSK591 MS023 10000 10000 10000 -ActD 0 5000 5000 5000

0 0 0 0 15000 15000 15000 5000 −2 log 2FC(+ActD / -ActD) 10000 10000 10000 +ActD 5000 5000 5000 −4

0 0 0 0 median = -0.220 -0.106 -0.275 0.1 kb n = 252 (introns) f Relative Intron Abundance GSK591 1.5 *** MS023

1.0 ns 0.06 ns ns 0.5

0.0

-0.5

log 2FC(+ActD / -ActD Rel. DMSO) -1.0 ** GAS5 ANZKF1 NOP2 Intron 9 Intron 9 Intron 14 bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyrightMaron holder for et. this al.preprint Figure 4 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Chromatin Poly(A) Enrichment a b c UV-crosslinking Splicing, Transport, Degradation Splicing, Transport, Degradation Chromatin Isolation -log2(P) 10 -log2(P) GSK591 MS023 GSK591 MS023 5 3 254 nm 10 6 15 9 4 12 A AA A A A A 5 A A A ALYREF A A A SRRM2 A SRSF11 HNRNPA1 HNRNPA3 RBMXL1 SNRPB SNRPB SMN1 FUS HNRNPA3 POLDIP3 SARNP NUDT21 PTBP1 USP39 DDX39B RBMXL1 POLR2G USP39 HNRNPA1 Poly(A) Enrichment PCBP1 DHX9 CSTF2T PARN SMNDC1 POLDIP3 PTBP1 HNRNPL SRRM1 LSM6 HNRNPK CPSF4 CHTOP SF3B1 PABPC4 ZMAT2 -Competitor +Competitor SART1 EXOSC9 HNRNPA0 POLR2C RBM25 RAE1 HNRNPL HNRNPK 0 0 NUP43 SNW1 Bead T Bead T EIF4B

T T

FYTTD1

T T NUP43 PRPF8

T T T T T T

A A

A T T A

SART1

PPIL4

PPIL4 T T

A T T

SNRNP40

UPF1

A T T

T T

A log 2 FC(PRMTi / DMSO) log 2 FC(PRMTi / DMSO)

A SART1

CRNKL1 PRPF8

−5 4

Mass Spectrometry NUP205 −10

d e Top 15 enriched Top 15 enriched PTMscan Input methylarginine containing PTMscan Poly(A) methylarginine containing (Methyl Arginine) Chromatin proteins (Methyl Arginine) Enriched proteins TAF15 EWSR1 AHNAK2 UBAP2L YBX3 CHTOP

Relative Abundance HNRNPA3 SERBP1 SNRPB HNRNPA3 RBM3 SERBP1 497 302 562 69 m/z 88 UBAP2L KHDRBS1 RBM3 HNRNPA1 TAF15 RPS2 ALYREF FUS RPS2 23 SNRPB FBL EIF4H HNRNPA1 GAR1 HNRNPL EIF4B EWSR1 G3BP1 f g

48h PRMTi Chromatin 2% Input IgG IP: CHTOP IP: SNRPB

kDa DMSO GSK591 MS023 kDa DMSO GSK591 MS023 DMSO GSK591 MS023 DMSO GSK591 MS023 DMSO GSK591 MS023

CHTOP 25 SNRPB 25

25 Rme2s 25 SNRPB

15 H3

25 CHTOP

25 Rme2a (short)

25 Rme2a (long) bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyright Maronholder for et.this preprintal. Figure 5 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. a Retained Intron Levels b Common Retained Intron Abundance c Common Retained Intron Abundance

n = 1,802 SNRPBkd SNRPBkd CHTOPkd SNRPBkd CHTOPkd (GBM) GSK591 GSK591 MS023 MS023

1,309 SNRPBkd

(HeLa) n = 535 n = 620 n = 164 n = 201

1,660 GSK591 (A549) Intron Inclusion Relative to Control d RI correlation (z-score) 591 589 23 MS023 2

(A549) GS K MS 0 1 SNRPB (GBM) 0.61 -0.17 1,148 0 CHTOPkd −1 (HEK293T) −2 SNRPB (HeLa) 0.51 -0.22 -1 0 1 Less Included ΔΨ More Included CHTOP -0.44 0.31

e g Arginine Arginine

WT SmD3 SmB SmD1 WT CHTOP

P2A P2A Alanine Alanine Lysine Lysine R A/K R A/K SmD3 SmB SmD1 CHTOP (30) (29) P2A P2A f Relative Intron Abundance h Relative Intron Abundance **** *** Sm-R A CHTOP R A 0.23 2.0 *** Sm-R K 0.5 CHTOP R K n o * 0.16 i s u 1.0 l

c 0.0 n I n o r

t ****

0.0 n

I -0.5 Intron Inclusion

log 2 (Sm-Mut Rel. WT) 0.23 ns *** 0.26 (CHTOP-Mut Rel. WT) log 2 (CHTOP-Mut -1.0 -1.0 *** GAS5 ANZKF1 NOP2 GAS5 ANZKF1 NOP2 Intron 9 Intron 9 Intron 14 Intron 9 Intron 9 Intron 14 bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyright Maronholder for et.this preprintal. Figure 6 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. a b

1.0 NOP2 chr12:6,669,263-6,670,236 exon:15 14 13 12 15 14 13 12 0.5

Δ Ψ (A549) Cytoplasm Nuclear

More Nuclear 15000 15000 0.0 10000 10000

−0.5 5000 5000

Nuclear/Cytoplasm 0 0 −1.0 More Cytoplasmic 0.1 kb −1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0 Less Included More Included Less Included More Included GSK591 ΔΨ (A549) MS023 ΔΨ (A549)

c GAS5 ANKZF1 NOP2 Cyto Nuc Chr Cyto Nuc Chr Cyto Nuc Chr Cyto Nuc Chr Cyto Nuc Chr Cyto Nuc Chr 1.5 * DMSO ns ** GSK591 1.0 ns ** ns MS023 ** ns **** 0.5 ns *** **** ns ns ns ns Parent Transcript ns ns Abundance Relative to 0.0 Intron 9 (RI) Intron 4 (non-RI) Intron 9 (RI) Intron 12 (non-RI) Intron 14 (RI) Intron 13 (non-RI) bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyright Maronholder for et.this preprintal. Figure 7 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Cytoplasm a AA Nucleus A A A transcript

A A

A intron Export snRNPs

Type I PRMT inhibition RNA export AA A AA

IDR with Rme2s Post- Degradation Transcriptional Splicing IDR with Rme2a

PRMT5 inhibition exosome

Nuclear Speckle

Nuclear and Chromatin Detention AAA A A traffic jam AAA A AAA A A Chromatin A Nuclear pore

b B/B’ m S PRMT5 3 SAM Productive splicing and export D AAA m A A S SAH D1 m S PRMT5 inhibition

Sm snRNA Detention and delayed AAA A A splicing or degradation snRNP

c Productive polyadenylation and export CHTOP PRMT1/4 Type I PRMT inhibition

SAM AAAA A A Detention and enhanced Export SAH AAAA A splicing or degradation Factors bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. TheMaron copyright et. holder al. for this preprintSuppl. Figure 1 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made a available under aCC-BY 4.0 International license.

2.0 PRMT1 PRMT4 PRMT5 0.07 Scrambled 1.5 PRMT1 gRNA #1

1.0 PRMT1 gRNA #2 ** *** **** *** PRMT4 gRNA #1 0.5 **** PRMT4 gRNA #2 Relative Expression

(PRMTkd/Scrambled) **** 0.0 PRMT5 gRNA #1 PRMT5 gRNA #2 b GAS5 Intron 9 ANKZF1 Intron 9 NOP2 Intron 13

1.5 *

1.0 0.07 0.07 *** 0.06 *** 0.5

0.0

Intron Inclusion -0.5 * * **

log 2(PRMTkd/Scrambled) -1.0 bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. TheMaron copyright et. holder al. for this preprintSuppl. Figure 2 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made a Odds Ratio of RI Overlap available under aCC-BY 4.0 International license. (A549) (K562) GSK591 (A549) EPZ015666 (U87) GSK591 (K562) sgPRMT5 (THP1) MS023 (A549) MS023 (K562) GSK591+MS023 GSK591+MS023 GSK591 (PANC03.27) GSK712 (PANC03.27) GSK591+GSK712 (PANC03.27) 1e-300 2e−48 1e−176 1e-300 1e-300 2e−58 1e-300 2e−192 1e-300 1e-300

GSK712 (PANC03.27) 4e−283 5e−16 3e−58 2e−97 5e−225 3e−22 1e-300 2e−64 3e−297

GSK591 (PANC03.27) 1e-300 1e−50 4e−174 1e-300 3e−255 2e−46 1e-300 2e−146 GSK591+MS023 (K562) 9e−232 2e−30 2e−202 7e−79 1e−66 5e−94 6e−256 GSK591+MS023 (A549) 1e-300 8e−76 4e−257 1e-300 1e-300 3e−103

MS023 (K562) 1e−89 1e−05 3e−80 8e−35 5e−21

MS023 (A549) 1e-300 4e−17 5e−80 5e−148

sgPRMT5 (THP1) 1e-300 1e−33 5e−109 Padj

GSK591 (K562) 3e−234 5e−31

8e−95 6 8 10 EPZ015666 (U87) log2(Odds Ratio)

5’ SS 3’ SS 1.0 C TTTTTTTT T TTT 0.5 TTTT T C Genomic GT AAAAAACCCCCCTTTTTTT TTTTT C GAA A AT CC T T AGA CCC A G GA A CCCC AAAGGCCCCCCC AC G AC GA CC A GG GA C TT AAAAA GG A T ACA CT CC C C CAC GG A 0.0 T G G GC GGGGGGGGGGGGGGGAAAAAAAGG GT TCCGG

probability 1.0 GT AG

0.5 Retained Introns

0.0 -3 1 2 6 -30 -25 -20 -15 -10 -5 -2 -1 5 bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Maron et. al. Suppl. Figure 3 a b Spawn Rate vs TPM Alternative Splicing Event vs Splicing Rate DMSO GSK591 MS023 Spawn Rate ns ns * ns 10.0 DMSO 5 0.07 0.09 * ** ** GSK591 **** **** **** **** **** **** MS023 * ns *

7.5

0 10 (events/min) 5.0

−5 poly(A) RNA-seq TPM ln(x) 2.5 Splicing Rate log

0.0 −10 Bottom Middle Top A5SS constitutive casette A3SS A5SS constitutive casette A3SS A5SS constitutive casette A3SS SKaTER−seq Rate (tertile) intron exon intron exon intron exon n = 112 32,577 471 159 104 28,135 449 145 94 27,882 350 104 bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. TheMaron copyright et.holder al. for this preprintSuppl. Figure 4 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made a Splicing Rate vs Intron Positionavailable under aCC-BY 4.0 International license.

1 DMSO GSK591 MS023

0 (events/min) 10 −1

Splicing Rate log −2

first second internal penultimate last Intron Position b Probability of Cleavage Prior to Splicig vs Intron Position DMSO GSK591 MS023

1.00

0.75 Intron Position first

second 0.50 internal

penultimate Cumulative Fraction 0.25 last

0.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Probability of Cleavage Prior to Splicing bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyrightMaron holder et. al. for this preprintSuppl. Figure 5 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. a b Input poly(A) enriched GSK591 MS023 GSK591 MS023 RNA splicing p.adjust RNA splicing, via transesterification reactions with bulged adenosine p.adjust ribonucleoprotein complex biogenesis 0.03 mRNA splicing, via spliceosome 0.0015 ribonucleoprotein complex assembly 0.02 RNA splicing, via transesterification reactions 0.0010 ribonucleoprotein complex subunit organization 0.01 ribonucleoprotein complex biogenesis 0.0005 regulation of RNA splicing viral gene expression protein localization to plasma membrane GeneRatio GeneRatio telomere maintenance 0.05 nuclear−transcribed mRNA catabolic process, NMD 0.10 nucleobase−containing compound transport telomere organization 0.10 0.15 rRNA processing RNA localization cell−cell junction assembly 0.15 cotranslational protein targeting to membrane rRNA metabolic process translational initiation regulation of protein localization to Cajal body ribonucleoprotein complex localization positive regulation of protein localization to Cajal body nucleocytoplasmic transport protein localization to nuclear body protein−containing complex localization protein localization to Cajal body ribonucleoprotein complex assembly RNA localization rRNA metabolic process regulation of mRNA metabolic process ribonucleoprotein complex subunit organization RNA catabolic process regulation of mRNA metabolic process mRNA catabolic process viral gene expression rRNA processing RNA export from nucleus viral transcription protein localization to endoplasmic reticulum mRNA catabolic process translational initiation protein targeting to ER regulation of nucleobase−containing compound transport nucleic acid transport nuclear export RNA transport bioRxiv preprint doi: https://doi.org/10.1101/2021.08.20.457069; this version posted August 20, 2021. The copyrightMaron holder et. al. for this preprintSuppl. Figure 6 (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made a d available under aCC-BY 4.0 InternationalRelative license Exogenous. CHTOP Expression

n = 55 1.0 ALYREFkd CHTOP WT CHTOP R A (HEK293T) P O CHTOP R K

1,148 CH T s 0.5 CHTOPkd u o

(HEK293T) n e g o x E 0.0 -1 0 1

Less Included ΔΨ More Included (Expression Relative to Control) b e Relative Exogenous Sm Expression Relative Endogenous CHTOP Expression )

2.0 n 0.008 i t c A Sm WT (

1.5 n 0.006 Sm o Sm R A i s s u s o Sm R K e r

n 1.0 0.004 p e x g E o x e E 0.5 v 0.002 i t a Endogenous CHTOP l e

0.0 R 0.000 (Expression Relative to Control) c CHTOP Sm A549

kDa WT RA RK WT RA RK -

- Uncleaved P2A 35

25 - SmB FLAG

- SmD3 15

Rme2s 25 FLAG-SmB Endogenous SmB/B’

GAPDH 35 100 70

DB71 25