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