Transcriptome-wide RNA binding analysis of C9orf72 poly(PR) dipeptides

Resea ch A icle T ansc ip ome-wide RNA binding analysis o C9o 72 poly(PR) dipep ides Rubika Balend a 1,2, * , Igo Ruiz de los Mozos 3,4,5, * , Hana M Odeh 6 , Idoia Gla ia 1,2,7 , Ca melo Milio o 1,2 , Ka he ine M Wilson 1,2 , Agnieszka M Ule 5 , Ma ina Hallegge 3 , Lau a Masino 8 , S ephen Ma in 8 , Rickie Pa ani 3,5 , James Sho e 6 , Je nej Ule 3,5,9 , Ad ian M Isaacs 1,2 An in onic GGGGCC epea expansion in C9o 72 is a common gene ic cause o amyo ophic la e al scle osis and on o empo al demen ia. The epea s a e ansc ibed in bo h sense and an isense di ec ions o gene a e dis inc dipep ide epea p o eins, o which poly(GA), poly(GR), and poly(PR) ha e been implica ed in con ib- u ing o neu odegene a ion. Poly(PR) binding o RNA may con ibu e o oxici y, bu analysis o poly(PR)-RNA binding on a ansc ip ome- wide scale has no ye been ca ied ou . We he e o e pe o med c osslinking and immunop ecipi a ion (CLIP) analysis in human cells o iden i y he RNA binding si es o poly(PR). We ound ha poly(PR) binds o nea ly 600 RNAs, wi h he sequence GAAGA en iched a he binding si es. In i o expe imen s showed ha poly(GAAGA) RNA binds poly(PR) wi h highe a fini y han con ol RNA and induces he phase sepa a ion o poly(PR) in o condensa es. These da a indica e ha poly(PR) p e e en ially binds o poly(GAAGA)-con aining RNAs, which may ha e physiological consequences. DOI 10.26508/lsa.202201824 | Recei ed 11 No embe 2022 | Re ised 25 June 2023 | Accep ed 26 June 2023 | Published online 12 July 2023 In oduc ion A hexanucleo ide epea expansion in he C9o 72 gene is he mos common gene ic cause o amyo ophic la e al scle osis (ALS) and on o empo al demen ia (FTD) (DeJesus-He nandez e al, 2011;Ren on e al, 2011). Se e al mechanisms o oxici y ha e been implica ed in con ibu ing o he disease p ocess (Balend a & Isaacs, 2018). Dipep ide epea p o eins (DPRs) p oduced by epea -associa ed non-ATG (RAN) ansla ion a e likely o ep- esen an impo an oxic en i y (Ash e al, 2013;Mo i e al, 2013;Zu e al, 2013). Fi e di e en DPRs a e p oduced: poly(GA), poly(GP), poly(GR), poly(PA), and poly(PR). O hese, he a ginine-con aining DPRs, poly(GR) and poly(PR), a e he mos oxic in model sys ems (Moens e al, 2017). The common pa hological hallma k iden ified in he as majo i y o spo adic and gene ic ALS cases and a la ge p opo ion o FTD cases is mislocalisa ion and agg e- ga ion o he RNA- and DNA-binding p o ein TDP-43. This pa- hology is also ound in C9o 72 ALS and FTD (C9FTD/ALS) and is likely o be downs eam o DPR pa hology (Balend a & Isaacs, 2018). Se e al mechanisms ha e been a ibu ed o DPR pa hology and include nucleocy oplasmic a ficking dys unc ion, DNA damage, and ansla ional inhibi ion. A numbe o s udies ha e explo ed he e ec o he a ginine-con aining DPRs on memb aneless o ganelles, such as s ess g anules and nucleoli. In e ac ome s udies ha e confi med ha poly(PR) and poly(GR) bind o p o eins en iched in p ion-like low-complexi y domains (LCDs), many o which a e RNA-binding p o eins (RBPs) and cons i uen s o memb aneless o ganelles (Lee e al, 2016;Lin e al, 2016;Boeynaems e al, 2017; Ha mann e al, 2018;Moens e al, 2019;Odeh & Sho e , 2020). LCDs in RBPs acili a e he p ocess known as phase sepa a ion, by which memb aneless o ganelles a e o med, and his p ocess is p omo ed by he p esence o RNA (Molliex e al, 2015;Mu akami e al, 2015;Pa el e al, 2015;P o e e al, 2018). Mu a ions in TDP-43, FUS, and hnRNPA1 cause ALS/FTD, and hese mu a ions a e o en localised wi hin he LCDs o hese RBPs. These mu a ions inc ease he o ma ion o amyloid-like fib ils and dis u b phase sepa a ion dynamics. Poly(PR) and poly(GR) dis up he dynamics o phase sepa a ion in memb aneless o ganelles in cells and impai ansla ion (Lee e al, 2016;Boeynaems e al, 2017;Ha mann e al, 2018;Zhang e al, 2018;Moens e al, 2019;Whi e e al, 2019). These a ginine- ich DPRs can also unde go phase sepa a ion hemsel es in i o, which is dependen on anion cha ge, and he p esence o RNA dose-dependen ly inc eases he phase sepa a ion o pol- y(PR) (Boeynaems e al, 2017;Boeynaems e al, 2019). I is possible ha hese in e ac ions wi h RBPs and o he LCD-con aining p o eins a e pa ly media ed by in e ac ions o poly(PR) wi h 1 UK Demen ia Resea ch Ins i u e a UCL, London, UK 2 Depa men o Neu odegene a i e Disease, UCL Queen Squa e Ins i u e o Neu ology, London, UK 3 The F ancis C ick Ins i u e, London, UK 4 Depa men o Pe sonalized Medicine, NASERTIC, Go e nmen o Na a a, Pamplona, Spain 5 Depa men o Neu omuscula Diseases, UCL Queen Squa e Ins i u e o Neu ology, London, UK 6 Depa men o Biochemis y and Biophysics, Pe elman School o Medicine a he Uni e si y o Pennsyl ania, Philadelphia, PA, USA 7 Resea ch Suppo Se ice, Ins i u e o Ag obio echnology, CSIC-Go e nmen o Na a a, Mu il a, Spain 8 S uc u al Biology Science Technology Pla o m, The F ancis C ick Ins i u e, London, UK 9 UK Demen ia Resea ch Ins i u e a King’s College London, Mau ice Wohl Clinical Neu oscience Ins i u e, London, UK Co espondence: [email protected]; je [email protected] *Rubika Balend a and Igo Ruiz de los Mozos con ibu ed equally o his wo k ©2023Balend a e al. h ps://doi.o g/10.26508/lsa.202201824 ol 6 | no 9 | e202201824 1o 11 on 21 Decembe , 2023li e-science-alliance.o g Downloaded om h p://doi.o g/10.26508/lsa.202201824Published Online: 12 July, 2023 | Supp In o: RNA. An in e ac ome analysis o poly(GR)80 exp essed in hu- man emb yonic kidney cells e ealed ha i in e ac s wi h RBPs and ibosomal p o eins, including mi ochond ial ibosomal p o eins (Lopez-Gonzalez e al, 2016). Se e al in e ac ions we e abolished when samples we e ea ed wi h RNase A, sugges ing some we e RNA-media ed. Ano he s udy demons a ed ha poly(PR) in e ac s wi h mul iple DEAD-box RNA helicases, and ha his is dependen on RNA, sugges ing RNA media es he in e - ac ion (Suzuki e al, 2018). Poly(PR)20 pep ide, when applied ex- ogenously o human as ocy e cells in cul u e, leads o al e a ions in splicing o se e al mRNAs and a change in abundance o mRNAs encoding ibosomal p o eins in pa icula (Kwon e al, 2014), and some o hese RNAs a e bound di ec ly by poly(PR) (Kaneku a e al, 2016). Howe e , a ansc ip ome-wide analysis o poly(PR) binding o RNAs in he cellula con ex has no been in es iga ed ye . To achie e his goal, we used imp o ed iCLIP (iiCLIP), which enables quan i a i e iden ifica ion o p o ein–RNA c osslinking si es in i o (Lee e al, 2021 P ep in ), o in es iga e whe he he a ginine-con aining DPR poly(PR) binds o RNA wi h some se- quence specifici y in human cells. We show ha poly(PR) di ec ly c osslinks o RNA and shows en iched c osslinking on specific ansc ip s, including ALS- ele an mRNAs such as neu ofilamen medium chain (NEFM)andnucleolin(NCL). We u he show ha poly(PR) in e ac s wi h nanomola a fini y wi h GAAGA- con aining RNA, which also p omo es he phase sepa a ion o poly(PR). Resul s Poly(PR) iiCLIP e eals binding o specific RNAs To in es iga e ansc ip ome-wide DPR binding o RNA, we es ablished dena u ing pu ifica ion o DPR-RNA complexes o CLIP based on he p e iously es ablished app oach (Fig 1A–D) (Huppe z e al, 2014;Lee e al, 2021 P ep in ). We exp essed doxycycline-inducible iple FLAG- agged PR100 o GA100, o iple FLAG ag alone in human emb yonic kidney cells (HEK293Ts) (Figs 1B and S1). PR100-FLAG and GA100-FLAG we e bo h p esen in he nucleus and cy oplasm, wi h PR100-FLAG ha ing g ea e nuclea localisa ion han GA100-FLAG (Fig S1A and B). The e was no di - e ence in ans ec ion e ficiency o exp ession le el be ween PR100-FLAG and GA100-FLAG (Figs 1C and S1C), and FLAG was de ec ed by a do blo in FLAG-exp essing cells (Fig S1D–F). A e 24 h o exogenous exp ession, we used UV ligh o c osslink p o ein–RNA in e ac ions. To es o specific binding o poly(PR) o RNA, we used he con ol condi ions o poly(PR)-exp essing cells, which we e non-c osslinked, and FLAG-exp essing cells, which we e c osslinked. We subsequen ly immunop ecipi a ed he DPR-RNA complexes using he FLAG ag (Fig S2A and B). We hen employed he iiCLIP p o ocol o liga e an in a ed adap o o isualisa ion o he p o ein–RNA complexes, and ex ac ed and e e se- ansc ibed he RNA o gene a e cDNA lib a ies o high- h oughpu se- quencing (Lee e al, 2021 P ep in ). In a ed isualisa ion o he DPR-RNA complexes showedmuchs onge signalin he c osslinked PR100-FLAG cells (Fig 2A, lanes 3 and 4) compa ed wi h he non-c osslinked PR100-FLAG cells (Fig 2A, lane 5), he c oss- linked GA100-FLAG cells (Fig 2A, lanes 6 and 7), and he c osslinked FLAG-only cells (Fig 2A, lane 9)—and his di e ence is especially appa en o he di use signal ha usually ep esen s p o eins c osslinked o longe RNA agmen s (Fig 2A). This finding indi- ca es ha PR100 di ec ly c osslinks o RNA in human cells. Sequencing o he iiCLIP eads e ealed o e 1,200,000 unique cDNA c osslinking e en s in he c osslinked PR100-FLAG cells ac oss mul iple eplica es, wi h significan ly ewe in he con ol (<74,000) and GA100 (<120,000) condi ions (Fig 2B). PR100 c oss- linking e en s occu ed mos equen ly in in ons, in e genic egions, and he coding sequence, wi h addi ional signal in non- coding RNAs and 39and 59UTRs (Fig 2C). We iden ified 558 mRNAs wi h high le els o binding (≥200 c osslinking e en s) o PR100 as compa ed o he con ols o PR100-non-c osslinked and FLAG- alone samples (Table S1). Examples o genes wi h he highes numbe s o binding e en s (>1,000 c osslinking e en s) included X-inac i e specific ansc ip (XIST), me as asis- ela ed lung adenoca cinoma ansc ip 1 (MALAT1), NEFM,NCL,nuclea en iched abundan ansc ip 1 (NEAT1), and he e ogeneous nuclea ibonucleop o ein U (HNRNPU)(TableS1).XIST and NEAT1 we e in he op se en genes wi h he la ges numbe o c osslinking e en s (Table S1), in ag eemen wi h hei p e iously iden ified in e ac ion wi h poly(PR) h ough RNA-IP expe imen s (Suzuki e al, 2019). In addi ion, he pa kin gene (PRKN), mu a- ions in which cause Pa kinson’sdisease(Ki ada e al, 1998), had 266 c osslinking e en s. Gene On ology (GO) en ichmen analysis o he 558 mRNAs wi h he highes numbe o c osslinks o PR100 (Table S1) e ealed significan en ichmen in ol ing he bio- logical p ocesses o “RNA splicing,”“ egula ion o ch omosome o ganisa ion,”and “co alen ch oma in modifica ion”(Fig 2D and E). The e was also significan en ichmen in he cellula com- ponen s o “nuclea speckles,”“ch omosome egion,”and “cen o- me ic ch omosome egion”(Fig S3A and B), and in he molecula unc ions o “ATPase ac i i y,”“DNA-dependen ATPase ac i i y,”and “helicase ac i i y”(Fig S3C). Poly(PR) binds wi h high a fini y o poly(GAAGA) RNAs We analysed en ichmen o 5-me mo i s in ou PR100 iiCLIP da ase , which iden ified GAAGA as a highly en iched pen ame ic sequence (Fig 3A), exemplified in he NCL and NEFM ansc ip s (Fig 3C and D). This en ichmen appea ed specific o PR100, as en ichmen o 5-me mo i s in he GA100 iiCLIP da ase iden ified CCGGG as he mos en iched pen ame (Fig S4). AUAAU was a less- ep esen ed mo i (in he bo om 5% o all mo i s) su ounding he PR100 c osslinking si e (Fig 3B). To de e mine whe he he e was a di e en ial a fini y o poly(PR) o hese 5-me RNA sequences, we used biolaye in e e ome y o compa e he a fini y o pu ified PR20 and GP20 pep ides wi h bio inyla ed RNA oligonucleo ides con aining fi e epea s o GAAGA o AUAAU (Table 1). PR20 had as onge a fini y o he poly(GAAGA) RNA wi h an appa en K d o 2.6 ± 0.5 nM (Fig 4A, E, and I)compa edwi h hepoly(AUAAU) RNA, wi h an appa en K d o 11.1 ± 2.5 nM (Fig 4C and G). The e was no e idence o in e ac ion be ween he DPR GP20 and he pol- y(GAAGA) RNA, e en a 180- old highe concen a ions (3–25 μM T ansc ip ome-wide poly(PR)-RNA binding Balend a e al. h ps://doi.o g/10.26508/lsa.202201824 ol 6 | no 9 | e202201824 2o 11 o GP20 compa ed wi h up o 133 nM o PR20) (Fig 4K). We also examined he a fini y o pu ified GR20 pep ides o he poly(GAAGA) o poly(AUAAU) RNA. GR20 had a sligh ly s onge a fini y o he poly(GAAGA) RNA wi h an appa en K d o 3.6 ± 0.9 nM (Fig 4B, F, and J)compa edwi h hepoly(AUAAU)RNA,wi han appa en K d o 6.1 ± 0.6 nM (Fig 4D and H).Thedi e enceina - fini ies be ween GAAGA and he con ol RNA sequence was highe o poly(PR) han o poly(GR) (Fig 4I and J), which is consis en wi h he pen ame s being de i ed om poly(PR) iiCLIP da a. These expe imen s show ha bo h poly(PR) and poly(GR) ha e a high binding a fini y o he es ed RNAs, as expec ed because o hei posi i e cha ge. In e es ingly, his binding shows some sequence specifici y,asahighe a fini ywasobse edwi h heGAAGAmo i ha was mos en iched in he iiCLIP expe imen . Poly(GAAGA) RNA enhances poly(PR) and poly(GR) phase sepa a ion Because o he high a fini y o he poly(GAAGA) RNA sequence o poly(PR) and poly(GR), and i s en ichmen in poly(PR) binding si es in i o, we in es iga ed whe he he poly(GAAGA) RNA could in- fluence poly(PR) and poly(GR) phase sepa a ion. Poly(PR) and poly(GR) unde go phase sepa a ion in he p esence o polyanions, such as RNA (Boeynaems e al, 2017;Boeynaems e al, 2019;Hu en Figu e 1. C9o 72 dipep ide epea p o eins (DPRs) iiCLIP pipeline. (A) Diag am o he PR100-3xFLAG, GA100-3xFLAG, and 3xFLAG-alone cons uc s used in his s udy. (B) An i- FLAG immunoblo 24 h pos -induc ion o PR100- 3xFLAG, GA100-3xFLAG, and 3xFLAG in HEK293T cells. No e GA100 appea s mos ly agg ega ed as he majo i y is p esen a he op o he gel and FLAG alone is no isible because o i s low molecula weigh , so i s exp ession was confi med by a do blo (Fig S1D–F). (C) Quan ifica ion o he exp ession o PR100-FLAG and GA100-FLAG no malised o GAPDH exp ession. No di e ence was obse ed be ween hese condi ions. n = 3 eplica es pe condi ion. Ba s show he a e age and SD. P> 0.05, wo- ailed unpai ed es . (D) Summa y o he iiCLIP pipeline o in es iga ion o DPR-RNA di ec in e ac ion. T ansien ly ans ec ed HEK293Ts we e UV-c osslinked o s abilise DPR–p o ein in e ac ions, and cells we e lysed and diges ed wi h RNase. The FLAG ag was used o immunop ecipi a ion o DPR-RNA complexes. A p eadenyla ed, in a ed dye–labelled adap o was liga ed on o he 39end o he RNA. RNA was ex ac ed and e e se- ansc ibed, gene a ing cDNA lib a ies, which we e high- h oughpu - sequenced, and he da a we e analysed o de e mine si es o binding wi h nucleo ide specifici y ac oss he ansc ip ome. T ansc ip ome-wide poly(PR)-RNA binding Balend a e al. h ps://doi.o g/10.26508/lsa.202201824 ol 6 | no 9 | e202201824 3o 11 Figu e 2. iiCLIP e eals poly(PR) binds o RNA in human cells. (A) In a ed labelled p o ein–RNA complexes we e sepa a ed by SDS–PAGE and ans e ed on o a ni ocellulose memb ane. In lane 1, GFP-TDP-43-FLAG was un as a posi i e con ol, wi h a di use smea de ec ed abo e i s molecula weigh , ep esen ing he GFP-TDP-43-FLAG-RNA complexes. Lane M is he p o ein ladde ma ke . Lanes 3 and 4 a e c osslinked PR100-FLAG cells, which ha e a highe in ensi y han he PR100-FLAG–non-c osslinked cells (lane 5), he GA100-FLAG–c osslinked cells T ansc ip ome-wide poly(PR)-RNA binding Balend a e al. h ps://doi.o g/10.26508/lsa.202201824 ol 6 | no 9 | e202201824 4o 11 e al, 2020). Thus, we examined whe he poly(GAAGA) RNA had an e ec on poly(PR) and poly(GR) phase sepa a ion. In he absence o RNA, poly(PR) and poly(GR) do no unde go phase sepa a ion (Fig 5A, le ). Rema kably, he p esence o equimola poly(GAAGA) sig- nifican ly inc eased poly(PR) and poly(GR) phase sepa a ion, in- dica ed by a ma ked inc ease in u bidi y, and nume ous, small ound, anslucen condensa es (Fig 5A, middle, and Fig 5B). In con as , he poly(AUAAU) RNA induced ewe and la ge poly(PR) and poly(GR) condensa es, which we e mo phologically dis inc om hose o med in he p esence o poly(GAAGA) RNA (Fig 5A, igh ). Howe e , poly(AUAAU) RNA induced less phase sepa a ion han poly(GAAGA) RNA o bo h poly(PR) and poly(GR) (Fig 5B). These findings sugges ha highe a fini y RNA, such as poly(GAAGA), displays enhanced abili y o induce poly(PR) and poly(GR) condensa ion. Discussion In his s udy, we ha e in es iga ed whe he poly(PR) p oduced in C9FTD/ALS may di ec ly bind o RNA in human cells. Indeed, we demons a ed di ec and specific in e ac ions be ween he a ginine- ich DPR poly(PR) and GAAGA-con aining RNAs using a ansc ip ome-wide app oach. A ginine- ich DPRs ha e been shown o exe dele e ious e ec s on se e al cellula unc ions, which include nucleocy oplasmic anspo (F eibaum e al, 2015; Jo icic e al, 2015;Zhang e al, 2015;Boeynaems e al, 2016), phase ansi ion o cellula o ganelles (Lee e al, 2016;Lin e al, 2016; Boeynaems e al, 2017), p o eos asis (K ame e al, 2018), and RNA dys egula ion, which has p e iously been desc ibed in C9FTD/ALS models and pa ien cells and issues (Kwon e al, 2014;Kaneku a e al, 2016;Yin e al, 2017). Some o hese e ec s a e likely o be caused by he in e ac ions o a ginine- ich DPRs wi h o he p o- eins, and ou s udy sugges s ha hei RNA in e ac ions migh also con ibu e o hese e ec s. Using in i o s udies, we confi med ha poly(PR) binds RNA wi h nanomola a fini y, wi h a s onge appa en a fini y o he pol- y(GAAGA) as compa ed o poly(AUAAU) RNAs. Fu he o he dis- co e y ha poly(U) RNA p omo es he phase sepa a ion o poly(PR) (Boeynaems e al, 2017), i has been shown ha in a es ube, poly- A, poly- U, and poly- C RNA homopolyme s can p omo e he phase sepa a ion o poly(PR), bu poly- G does so o a lesse ex en (Boeynaems e al, 2019). In he homopolyme ic o m, he a fini y o he in e ac ion be ween poly(PR) and poly- A is he s onges , whe eas poly(PR) has an almos iden ical a fini y o poly- U and poly- C and he lowes a fini y o poly- G. This finding has been explained by he abili y o poly- G o o m G-quad uplex s uc u es, as opposed o o he homopolyme ic RNAs, which a e uns uc u ed. I has been hypo hesised ha base s acking in e ac ions associa ed wi h G-quad uplex o ma ion could compe e wi h he poly(PR) in e ac ion. In iguingly, in compa ison wi h hese p e ious find- ings, he pen ame ic sequence we ound o be he mos en iched binding o poly(PR) ansc ip ome-widein i ohasahighe G-con en (bu wi h insu ficien guanines o o m G-quad uplexes) han he leas equen ly bound pen ame , sugges ing ha RNA sequences con aining guanine can ha e a high a fini y o poly(PR) in he cellula con ex . Fu he mo e, i was demons a ed ha mixing homopolyme ic RNA molecules, which can make comple- men a y base pai s, can change he in e ac ions be ween RNA and poly(PR), possibly because o compe i ion be ween RNA base pai ing in e ac ions and RNA–pep ide in e ac ions (Boeynaems e al, 2019). O impo ance, adding o al HEK cell RNA dose- dependen ly amelio a es a nuclea impo pheno ype induced by adding poly(GR) and poly(PR) o cells (Hayes e al, 2020), sugges ing RNA may educe hese pheno ypes h ough high-a fini y in e ac- ions wi h hese DPRs. In iguingly, we ound ha RNA sequences ha igh ly bind o poly(PR) wi h high a fini y ha e an inc eased abili y o p omo e poly(PR) condensa e o ma ion. I would be o in e es in u u e s udies o de e mine whe he hese RNA-induced condensa es a e less oxic o cells and whe he poly(GAAGA) is able o al e he phase sepa a ion o poly(PR) wi hin cells. One appealing possibili y would be o use PR-specific RNA sequences, such as poly(GAAGA), as “bai s” o sa egua d he cell by seques- e ing poly(PR) om dele e ious in e ac ions. In ac , TNPO1, a nuclea impo ecep o , has been shown o play such a p o ec i e ole agains DPRs when o e exp essed (Hu en e al, 2020). Al hough i was ou in en ion o p o ide a ansc ip ome-wide da ase a he han o ocus on specific ansc ip sboundby poly(PR), we epo ha se e al in e es ing RNAs a e bound. These include he p e iously iden ified pa aspeckle long non- coding RNA NEAT1, o which poly(PR) binding was shown o lead o NEAT1 up- egula ion (Suzuki e al, 2019), and NCL, which en- codes he nucleolin p o ein, which is known o ha e a mo e dispe sed nuclea localisa ion in C9o 72 human issue and dis- ease models (Haeusle e al, 2014). As expec ed, based on p e ious s udies (Kwon e al, 2014), PR showed some punc a e nuclea s aining consis en wi h nucleola localisa ion and his is likely o g ea ly influence he RNAs ha a e ound o be bound o i . This s udy has also been pe o med in human cell lines using he o e exp ession o poly(PR), which limi s he disease ele ance o hese findings, and u u e s udies in neu onal models wi h mo e physiological exp ession le els would be o impo ance. I is in- c easingly ecognised ha RNA dys egula ion plays a majo ole in ALS/FTD and many gene ic causes o ALS/FTD a e in RBPs (Nussbache e al, 2019). Ou da ase o poly(PR)–RNA binding can now be used o hypo hesis-d i en in es iga ion o poly(PR) e - ec s on RNAs. Unde s anding he biology o hese in e ac ions may help o u he elucida e he unde lying ae iology o neu- odegene a ion in C9FTD/ALS. (lanes 6 and 7), he GA100-FLAG–non-c osslinked cells (lane 8), and he FLAG-c osslinked cells (lane 9). The FLAG ag consis s o 3XFLAG. (B) F equency o unique cDNAs, which ep esen indi idual c osslinking e en s iden ified by iiCLIP analysis. (C) Genomic loca ion o PR100 binding si es, in in ons, in e genic egions, and he coding sequence, wi h addi ional signal in non-coding RNAs and 59UTR and 39UTR segmen s. (D) Gene On ology gene se en ichmen analysis o PR100-c osslinked RNAs. Genes o RNAs bound in PR100 samples a e ep esen ed by hei Biological P ocess. The numbe o genes om he PR100-FLAG c osslinking da ase in each Gene On ology ca ego y is shown and colou -coded by an adjus ed P- alue. (E) Genes o RNAs bound in PR100 samples om he op h ee significan ca ego ies wi hin Biological P ocess (RNA splicing, egula ion o ch omosome o ganisa ion, and co alen ch oma in modifica ion) a e ep esen ed in a gene-concep ne wo k. The size o he ci cle o each Biological P ocess is p opo ional o he numbe o genes iden ified wi hin ha ca ego y. T ansc ip ome-wide poly(PR)-RNA binding Balend a e al. h ps://doi.o g/10.26508/lsa.202201824 ol 6 | no 9 | e202201824 5o 11 Figu e 3. Mo i en ichmen analysis o poly(PR)-RNA binding. (A) Mo i en ichmen analysis o he PR100 binding c osslinking si es e ealed he mos equen pen ame bound by PR100 was GAAGA (P=3.1×e −930 ). (B) Analysis o he posi ion o he pen ame ela i e o he c osslinking si e. GAAGA is en iched ups eam and downs eam o he c osslinking si e. The AUAAU pen ame has a lowe equency bo h ups eam and downs eam o he c osslinking si e. (C, D) NCL and NEFM, which a e ansc ip s highly bound by PR100 in he iiCLIP da ase (Table S1), ha e equen GANGA (GAAGA, GAGGA, o GACGA) mo i s in p oximi y o PR100 binding si es. The lowe pa o each panel indica es he posi ion o hese mo i s ela i e o he c osslinking si es wi hin he gene. Gene acks we e no malised by coun s pe million. Table 1. Oligonucleo ide sequences. Biolaye in e e ome y RNA oligonucleo ides GAAGA sequence 59-/5Biosg/ G A G A A G A G A A G A G A A G A G A A G A G A A-39 AUAAU sequence 59-/5Biosg/ A U A A U A U A A U A U A A U A U A A U A U A A U-39 T ansc ip ome-wide poly(PR)-RNA binding Balend a e al. h ps://doi.o g/10.26508/lsa.202201824 ol 6 | no 9 | e202201824 6o 11 Ma e ials and Me hods Cell lines HEK293Ts we e cul u ed in DMEM supplemen ed wi h 10% FBS, g own a 37°C wi h 5% CO 2 , and ou inely passaged. T ansien ans ec ions and iiCLIP p o ocol PR100 and GA100 (Mizielinska e al, 2014) we e cloned in o pcDNA5 Flp-In Exp ession ec o s wi h a 39 iple FLAG ag, gene a ing PR100-3xFLAG (PR100-FLAG) and GA100-3xFLAG (GA100-FLAG) pcDNA5 plasmids, wi h he 3xFLAG-only ec o (FLAG) also used as a con ol. Fo iiCLIP expe imen s, HEK293Ts we e g own a ≈80% confluency in 10-cm pla es and ansien ly ans ec ed wi h PR100-FLAG, GA100-FLAG, o FLAG pcDNA5 plasmids using Lip- o ec amine 2000. The exp ession o he cons uc s was induced by supplemen ing he media wi h 150 ng ml −1 o doxycycline o 24 h. The iiCLIP p o ocol was pe o med as p e iously desc ibed (Lee e al, 2021 P ep in ). T ansien ly ans ec ed cells induced o 24 h we e i adia ed wi h UV once wi h 160 mJ/cm 2 using a S a alinke 1800 a 254 nm. DNase was used a e cell lysis o emo e DNA. P o ein–RNA complexes we e liga ed o a p e- adenyla ed, in a ed dye–labelled adap o and pu ified. RNA was isola ed using p o einase K diges ion and e e se- ansc ibed in o cDNA. cDNA was subsequen ly pu ified and ci cula ised. Figu e 4. Poly(PR) and poly(GR) di ec ly bind o RNA wi h nanomola a fini y. Biolaye in e e ome y expe imen s measu ing binding o RNA o PR20, GR20, and GP20. (A, B, C, D) Associa ion phases o indi idual ep esen a i e expe imen s. As dissocia ion was ex emely slow (se e al hou s), i was only pa ially eco ded and i is no shown. (E, F, G, H) Plo s o no malised esponse e sus PR20 o GR20 concen a ion a e shown. (I, J) K d o poly(GAAGA) and PR20-GR20 in e ac ion was significan ly highe han poly(AUAAU) and PR20-GR20 in e ac ion. (I, J) Ba s show he a e age and SD o K d o h ee independen eplica e expe imen s o PR20 (I) and GR20 (J) wi h poly(GAAGA) o poly(AUAAU). **P< 0.01 and *P< 0.05, wo- ailed unpai ed es . (K) Biolaye in e e ome y expe imen s measu ing binding o poly(GAAGA) o GP20. 180- old highe concen a ions o GP20 (3.1–25 μM) we e used compa ed wi h PR20 and GR20 (2.1–133.3 nM) o confi m he e was no in e ac ion be ween GP20 and poly(GAAGA) RNA. T ansc ip ome-wide poly(PR)-RNA binding Balend a e al. h ps://doi.o g/10.26508/lsa.202201824 ol 6 | no 9 | e202201824 7o 11 iiCLIP analysis Mul iplexed cDNA lib a ies we e sequenced using Illumina HiSeq, gene a ing 100-n single-end eads. Sequenced eads we e p o- cessed by he iMaps so wa e package (h p://icoun .biolab.si/), and demul iplexed in o indi idual lib a ies based on hei ex- pe imen al ba codes. Unique molecula iden ifie nucleo ides we e used o dis inguish and collapse PCR duplica es. The ba code sequences and adap o s we e emo ed om he 59and 39ends. T immed sequences we e mapped o he human genome (build GRCh38, Gencode, e sion 27) wi h STAR aligne allowing wo misma ches (Langmead & Salzbe g, 2012). Uniquely mapping eads we e kep , and he p eceding aligned nucleo ide was assigned as he DPR-c osslinked si e. Significan c osslinking si es we e de- e mined by he iCoun False Disco e y Ra e (<0.05) algo i hm by weighing he en ichmen o c osslinks e sus shu fled andom posi ions (h ps://gi hub.com/ omazc/iCoun ). Fo subsequen anal- ysis, we se a h eshold o a leas 300,000 unique c osslinking e en s o each PR100-c osslinked sample, and n = 4 samples me his h eshold. Fo h ee o hese samples, p o ein–RNA complexes had been pu ified using SDS–PAGE and ans e ed on o ni ocellulose memb anes, as p e iously desc ibed (Lee e al, 2021 P ep in ). Fo one o hese samples, p o ein–RNA complexes had been pu ified using immunop ecipi a ion wi h beads. We analysed he numbe o c osslinking e en s o hese PR100-c osslinked samples and hei co esponding con ols: PR100-non-c osslinked, GA100-c osslinked, GA100-non-c osslinked, and FLAG-c osslinked. C osslinking e en s we e no malised by he o al numbe o c osslinks in he sample pe million (coun s pe million). Fo gene-le el analysis, genes we e iden ified ha con ain a leas 200 c osslinking e en s om PR100-FLAG iiCLIP, wi h <10% binding in he con ol condi ions o PR100-FLAG–non-c osslinked samples and FLAG-c osslinked samples (Table S1). These genes we e used o on ology en ich- men analysis pe o med wi h he R package clus e P ofile , compa ing agains all o he genes, using an FDR co ec ion and adjus ed P- aluecu -o o <0.01(Yu e al, 2012). En iched pen ame s we e calcula ed wi h DREME .5.4.1 (Bailey, 2011) using he 5 n up- s eam and 30 n downs eam o he significan c osslinking si es, compa ed wi h simila sequences collec ed om andom posi ions o he same genes ha did no o e lap wi h any significan c osslinking si es. Pen ame s we e chosen as hey ha e p o en in p e ious sys ema ic s udies mos use ul o dis inguish he se- quence binding specifici y o RBPs bo h o analysis o in i o binding specifici y om me hods such as RNA Bind-n-Seq (Dominguez e al, 2018) and o analysis o CLIP da a (Ku e e al,2022), and hus, he esul s o ou analyses can be mos easily compa ed wi h o he s udies. Immunoblo ing T ans ec ed HEK293Ts we e homogenised in lysis bu e (50 mM T is–HCl, pH 7.4, 100 mM NaCl, 1% Igepal CA-630, 0.1% SDS, and 0.5% sodium deoxychola e supplemen ed wi h cOmple e p o ease inhibi o cock ail [Roche]) and sonica ed in a Bio up o . Samples we e cen i uged a 16,000g o 10 min. Supe na an s we e loaded wi h NuPAGE LDS sample bu e and DTT, hen hea ed o 70°C o 5 min. Samples we e sepa a ed on NuPAGE 4–12% Bis-T is gels in MES unning bu e , hen ans e ed on o PVDF memb anes. A e blocking, memb anes we e incuba ed wi h an i-FLAG (F3165, 1:4,000; Sigma-Ald ich) o an i-GAPDH (2118S, 1:1,000; Cell Signaling Tech- nologies) ollowed by complemen a y seconda y an ibodies (LI-COR IRDye, 1:10,000). Specific binding was de ec ed wi h a LI-COR Odyssey CLx image . The in ensi y o bands was quan ified using Fiji–ImageJ so wa e. S a is ical analyses we e pe o med using G aphPad P ism 9. De ails a e gi en in he figu e legend. Fo do blo ing, supe na an s o cen i uged samples p epa ed as abo e we e do ed on o a ni ocellulose memb ane (18 o 9 μg o o al p o ein pe do ). A e blocking, memb anes we e incuba ed wi h an i-FLAG (F3165, 1:400; Sigma-Ald ich) o an i-GAPDH (2118S, 1:400; Cell Signaling Technologies) ollowed by complemen a y seconda y an ibodies (LI-COR IRDye, 1:10,000). Specific binding was de ec ed wi h a LI-COR Odyssey CLx image . Figu e 5. Poly(GAAGA) RNA enhances poly(PR) and poly(GR) phase sepa a ion. (A) Condensa ion o 20 μM PR20-FLAG o GR20-FLAG–alone (- RNA) o induc ion o equimola concen a ions o ei he poly(GAAGA) RNA o poly(AUAAU) RNA. Rep esen a i e images om h ee independen expe imen s we e aken using b igh -field mic oscopy. The black ba ep esen s 10 μm. (B) Tu bidi y (abso bance a 395 nm) measu emen s o 20 μM PR20-FLAG o GR20-FLAG wi h o wi hou RNA. Values ep esen he mean o h ee independen expe imen s ± SEM. One-way ANOVA (***P= 0.0001; ****P< 0.0001; and ns, no significan ). T ansc ip ome-wide poly(PR)-RNA binding Balend a e al. h ps://doi.o g/10.26508/lsa.202201824 ol 6 | no 9 | e202201824 8o 11 Immunofluo escence s aining T ans ec ed HEK293Ts g own on poly-D-lysine–coa ed Pe kinElme 96-well pla es we e fixed in 4% pa a o maldehyde in PBS o 15 min, hen washed h ee imes in PBS supplemen ed wi h 0.3% T i on X (PBST), and blocked wi h 5% BSA in PBST o 1 h. Cells we e in- cuba ed o e nigh a 4°C wi h a p ima y an ibody an i-FLAG (F3165, 1:500; Sigma-Ald ich) in 5% BSA in PBST. Cells we e washed in PBST h ee imes, hen incuba ed wi h a complemen a y Alexa Fluo seconda y an ibody (1:500) o 1 h a oom empe a u e. Cells we e washed once in PBST con aining DAPI o 10 min, hen wice mo e in PBST. Images we e acqui ed using a The mo Fishe Scien ific CX5 high- h oughpu imaging mic oscope wi h a 10x objec i e. Images we e analysed using p op ie a y onboa d so wa e. S a is ical ana- lyses we e pe o med using G aphPad P ism 9. De ails a e gi en in he figu e legend. Biolaye in e e ome y measu emen s Biolaye in e e ome y expe imen s we e pe o med on Fo eBio Oc e RED96 and Oc e R8 ins umen s (Sa o ius). Bio inyla ed RNAs we e syn hesised by In eg a ed DNA Technologies. PR20, GR20, and GP20 pep ides we e syn hesised by The mo Fishe Sci- en ific. Bio inyla ed RNA and poly-DPR pep ides we e dissol ed in T is–EDTA (10 mM T is–HCl and 1 mM disodium EDTA, pH 8.0) bu e solu ion wi h 150 mM NaCl, 0.1 mg/ml BSA, and 0.01% Tween-20 o educe non-specific in e ac ions. The assays we e ca ied ou a 25°C in a 96-well pla e and a sample olume o 200 μl. S ep a idin- coa ed biosenso s we e p e-equilib a ed, loaded wi h bio inyla ed RNAs, and exposed o p o ein concen a ions anging om 2.1 o 133 nM o PR20 and GR20 and om 3.1 o 25 μM o GP20. Equilib ium dissocia ion cons an s (K d ) o he RNA–p o ein in e ac ions we e de e mined by plo ing he ins umen esponse a equilib ium as a unc ion o p o ein concen a ion and fi ing he da a assuming a 1:1 in e ac ion, using non-linea leas squa es eg ession using Oc e BLI Analysis so wa e (Sa o ius). Oligonucleo ide sequences a e p o ided in Table 1. Biological iplica es we e pe o med using eshly p epa ed RNA and p o ein solu ions in independen ex- pe imen s. S a is ical analyses we e pe o med using G aphPad P ism 9. De ails a e gi en in he figu e legend. In i o poly(PR) and poly(GR) condensa ion assay A poly(PR) 20-me DPR wi h a C- e minal FLAG ag was pu chased om CSBio and e ified by mass spec ome y. The sequence o poly(PR) was as ollows: PRPRPRPRPRPRPRPRPRPRPRPRPRPRPRPRPRPRPRPRGSFEG- DYKDDDDK. A poly(GR) 20-me DPR wi h a C- e minal FLAG ag was pu chased om DGpep ides. The sequence o poly(GR) was as ollows: GRGRGRGRGRGRGRGRGRGRGRGRGRGRGRGRGRGRGRGRGSFEG- DYKDDDDK. Lyophilised powde was econs i u ed in 1X PBS, and snap- ozen in single-use 200 μM aliquo s and s o ed a −80°C. Poly(GAAGA) and poly(AUAAU) RNA sequences we e o de ed om IDT (sequences p o ided in Table 1). Lyophilised powde was econs i u ed in RNase- ee wa e , o a final s ock concen a ion o 100 μM. Aliquo s we e snap- ozen and s o ed a −20°C. Fo RNA-induced condensa e o ma ion, poly(PR) o poly(GR) and all RNAs we e fi s hawed on ice. Poly(PR) o poly(GR) was dilu ed o a final concen a ion o 20 μMin20mM Hepes–NaOH (pH 7.4), 150 mM NaCl, and 1mM DTT. Equimola amoun s (20 μM) o ei he poly(GAAGA) o poly(AUAAU) RNA we e added o poly(PR) o poly(GR) and incuba ed a oom empe a u e o 30 min. Phase-sepa a ed condensa es we e hen imaged by b igh -field mic oscopy (M5000; EVOS). Fo u bidi y measu emen s, abso - bance alues we e ead a an abso bance o 395 nm using TECAN (Safi e 2 ). S a is ical analyses we e pe o med using G aphPad P ism 8. De ails a e gi en in he figu e legend. Da a A ailabili y The iiCLIP sequencing da a a e a ailable on Gene Exp ession Omnibus wi h he accession numbe GSE212761. Supplemen a y In o ma ion Supplemen a y In o ma ion is a ailable a h ps://doi.o g/10.26508/lsa. 202201824 Acknowledgemen s The au ho s hank Michael Howell and he High-Th oughpu Sc eening Pla o m a he F ancis C ick Ins i u e o aluable assis ance. R Balend a is NIHR Academic Clinical Lec u e in Neu ology a UCL and has ecei ed unding om a Wellcome T us Resea ch T aining Fellowship [107196/Z/14/ Z] and he UCL Leona d Wol son Expe imen al Neu ology Cen e o his wo k. She was unded by an Academy o Medical Sciences S a e G an o Clinical Lec u e s (SGL027 1022). This wo k was unded by he Mo o Neu one Disease Associa ion ( o AM Isaacs), Alzheime ’s Resea ch UK (ARUK-PG2016A- 6; ARUK-EXT2019A-002) ( o AM Isaacs), he Eu opean Resea ch Council (ERC) unde he Eu opean Union’s Ho izon 2020 esea ch and inno a ion p o- g amme (648716—C9ND) ( o AM Isaacs), and he UK Demen ia Resea ch Ins i u e ( o AM Isaacs), which ecei es i s unding om UK DRI L d, unded by he UK Medical Resea ch Council, Alzheime ’s Socie y, and Alzheime ’s Resea ch UK. HM Odeh was suppo ed by an As aZeneca pos -doc o al ellowship and an Alzheime ’s Associa ion Resea ch Fellowship. J Sho e was suppo ed by ALSA, Ta ge ALS, AFTD, and he Packa d Founda ion o ALS Resea ch a JHU. Au ho Con ibu ions R Balend a: concep ualisa ion, esou ces, da a cu a ion, so wa e, o mal analysis, unding acquisi ion, alida ion, in es iga ion, isu- alisa ion, me hodology, p ojec adminis a ion, and w i ing—o iginal d a , e iew, and edi ing. I Ruiz de los Mozos: esou ces, da a cu a ion, so wa e, o mal analysis, alida ion, in es iga ion, isualisa ion, me hodology, p ojec adminis a ion, and w i ing—o iginal d a , e iew, and edi ing. HM Odeh: da a cu a ion, o mal analysis, alida ion, in es iga- ion, isualisa ion, me hodology, p ojec adminis a ion, and w i ing—o iginal d a , e iew, and edi ing. I Gla ia: in es iga ion and w i ing— e iew and edi ing. C Milio o: in es iga ion and w i ing— e iew and edi ing. KM Wilson: in es iga ion and w i ing— e iew and edi ing. AM Ule: in es iga ion, me hodology, and w i ing— e iew and edi ing. T ansc ip ome-wide poly(PR)-RNA binding Balend a e al. h ps://doi.o g/10.26508/lsa.202201824 ol 6 | no 9 | e202201824 9o 11