Malik Et Al - 2021

Article
SRSF protein kinase 1 modulates RAN translation

and suppresses CGG repeat toxicity
Indranil Malik1, Yi-Ju Tseng1,2,†, Shannon E Wright1,3,†, Kristina Zheng1, Prithika Ramaiyer1,
Katelyn M Green1,2 & Peter K Todd1,4,*
Abstract the fragile X gene, FMR1 (Hagerman et al, 2001). Normally healthy
humans harbor approximately 30 repeats, which get expanded to
Transcribed CGG repeat expansions cause neurodegeneration in 55–200 repeats in case of FXTAS (Hagerman & Hagerman, 2015).
Fragile X-associated tremor/ataxia syndrome (FXTAS). CGG repeat Patients develop imbalance, dementia, parkinsonism, and tremors
RNAs sequester RNA-binding proteins (RBPs) into nuclear foci and starting in their 50’s or 60’s (Jacquemont et al, 2003). Pathologi-
undergo repeat-associated non-AUG (RAN) translation into toxic cally, the condition is associated with diffuse neuronal loss and
peptides. To identify proteins involved in these processes, we brain atrophy as well as accumulation of ubiquitinated inclusions
employed a CGG repeat RNA-tagging system to capture repeat- within neurons and glia throughout the brain (Greco et al, 2002,
associated RBPs by mass spectrometry in mammalian cells. We 2006; Ariza et al, 2016). FXTAS is an inexorably progressive and
identified several SR (serine/arginine-rich) proteins that interact fatal condition without effective treatment. Thus, identifying targe-
selectively with CGG repeats basally and under cellular stress. table factor(s) involved in CGG repeat expansion-associated toxicity
These proteins modify toxicity in a Drosophila model of FXTAS. may help in the development of new therapeutics.
Pharmacologic inhibition of serine/arginine protein kinases CGG repeats are thought to elicit toxicity through two non-
(SRPKs), which alter SRSF protein phosphorylation, localization, exclusive mechanisms (Glineburg et al, 2018). Expanded repeat
and activity, directly inhibits RAN translation of CGG and GGGGCC RNAs can elicit gain-of-function (GOF) toxicity by sequestering
repeats (associated with C9orf72 ALS/FTD) and triggers repeat RNA essential RNA-binding proteins (RBPs) and forming RNA–RNA and
retention in the nucleus. Lowering SRPK expression suppressed RNA–protein complex condensates (Jazurek et al, 2016; Jain & Vale,
toxicity in both FXTAS and C9orf72 ALS/FTD model flies, and SRPK 2017; Glineburg et al, 2018). This pathologic process is best exem-
inhibitors suppressed CGG repeat toxicity in rodent neurons. plified by the sequestration of muscleblind (MBNL) proteins by CUG
Together, these findings demonstrate roles for CGG repeat RNA repeat RNA in myotonic dystrophy type 1 (DM1) (Taneja et al,
binding proteins in RAN translation and repeat toxicity and 1995; Miller et al, 2000). The repeat mRNA and MBNL protein
support further evaluation of SRPK inhibitors in modulating RAN avidly colocalize into RNA foci in both patient tissues and in model
translation associated with repeat expansion disorders. systems (Miller et al, 2000). Moreover, DM1 patients have splicing
defects and clinical phenotypes that mimic those seen with genetic
Keywords C9orf72 ALS; CGG repeats; FXTAS; RAN translation; SRSF ablation of MBNL1 and upregulation of MBNL1 suppresses relevant
Subject Categories Genetics, Gene Therapy & Genetic Disease; disease phenotypes in DM1 disease models (Mankodi et al, 2000;
Neuroscience; Translation & Protein Quality Kanadia et al, 2003, 2006; Wang et al, 2019). At CGG repeat expan-
DOI 10.15252/emmm.202114163 | Received 21 February 2021 | Revised 28 sions that cause FXTAS, in vitro RNA pull-down assays identified
August 2021 | Accepted 30 August 2021 | Published online 20 September 2021 Pur alpha, hnRNP A2/B1, Sam68, and DROSHA/DGCR8 as potential
EMBO Mol Med (2021) 13: e14163 repeat RNA targets (Jin et al, 2007; Sofola et al, 2007; Sellier et al,
2010, 2013). Overexpression of some of these factors in Drosophila
can suppress CGG repeat elicited phenotypes (Jin et al, 2007; Sofola
Introduction et al, 2007). However, direct manipulation of these factors has not
yet recapitulated (in their absence) or suppressed (in their overex-
Fragile X-associated Tremor/Ataxia Syndrome (FXTAS) is an age- pression) disease relevant phenotypes in rodent or human neuronal
related neurodegenerative disorder, which impacts approximately model systems.
1/5,000 people (Jacquemont et al, 2004; Tassone et al, 2012). It Repeat RNAs also support translational initiation in the absence
results from a transcribed CGG repeat expansion in the 5’ UTR of of AUG start codons through repeat-associated non-AUG (RAN)
1 Department of Neurology, University of Michigan, Ann Arbor, MI, USA

2 Cellular and Molecular Biology Graduate Program, University of Michigan, Ann Arbor, MI, USA
3 Neuroscience Graduate Program, University of Michigan, Ann Arbor, MI, USA
4 Ann Arbor Veterans Administration Healthcare, Ann Arbor, MI, USA
*Corresponding author. Tel: +01-734-615-5632; E-mail: [email protected]
†
These authors contributed equally to this work
ª 2021 The Authors. Published under the terms of the CC BY 4.0 license EMBO Molecular Medicine 13: e14163 | 2021 1 of 21
EMBO Molecular Medicine Indranil Malik et al
translation (Zu et al, 2011). RAN translation is supported by many containing reporter with PP7 viral stem-loops, which bind to viral
repeat expansions including CGG repeats associated with FXTAS coat protein with high affinity (Chao et al, 2008). By co-expressing
and GGGGCC repeats associated with ALS/FTD (Ash et al, 2013; an epitope-tagged coat-binding protein, PCP, we were able to isolate
Mori et al, 2013; Todd et al, 2013; Ba~nez-Coronel et al, 2015; Ishig- CGG repeat RNAs and associated RBPs. This modular system can be
uro et al, 2017; Zu et al, 2017; Soragni et al, 2018). RAN translation utilized in the context of cellular perturbations such as stress induc-
occurs across multiple reading frames of the same repeat to produce tion or drug treatment. Using this technique, we found that multiple
toxic repetitive polypeptides that accumulate in patient neurons and serine/arginine-rich splicing factor (SRSF) proteins interact with
tissues (Gendron et al, 2013; Zu et al, 2013, 2017; Krans et al, 2016; CGG repeat RNAs under normal and stress conditions. Genetic
Soragni et al, 2018). Existing experimental evidence for multiple targeting of serine/arginine proteins suppresses rough eye pheno-
repeat expansions suggests these RAN-translated peptides may play types and extends survival in a Drosophila model of FXTAS. More-
an active role in disease pathogenesis. For example, expression of over, genetic or chemical targeting of serine/arginine protein
dipeptide repeat products resulting from C9 ALS/FTD GGGGCC kinases (SRPK1) that regulate SRSF1 function and cellular distribu-
RAN translation are sufficient to elicit toxicity in model systems tion selectively suppress RAN translation and toxicity in fly and
even in the absence of repetitive RNA (May et al, 2014; Mizielinska rodent neuronal models of FXTAS through both direct effects on
et al, 2014; Wen et al, 2014; Jovicic et al, 2015; Lee et al, 2016; translation and through nuclear CGG repeat RNA retention. Taken
Zhang et al, 2018, 2019). together, these data present a novel approach to identify repeat
In FXTAS, RAN translation in the GGC reading frame generates a RNA-binding proteins in vivo and establish SR protein kinases as a
polyglycine-containing peptide termed FMRpolyG, which accumu- possible target to modulate RAN translation and repeat expansion-
lates into ubiquitinated inclusions in both model systems and associated toxicity.
patient tissues (Todd et al, 2013). Induced expression of repeats that
support FMRpolyG synthesis elicit toxicity in heterologous cells,
rodent neurons, flies, and transgenic mice (Todd et al, 2013; Buijsen Results
et al, 2014; Sellier et al, 2017). Moreover, sequence manipulations
that suppress RAN translation of FMRpolyG largely preclude CGG Development of a CGG repeat RNA-tagging system
repeat-associated toxicity in overexpression systems (Todd et al,
2013; Sellier et al, 2017). RAN translation is selectively activated by Cellular RNA-binding proteins (RBPs) play critical roles in RNA
cellular stress response pathways that typically preclude transla- gain-of-function toxicity in repeat expansion disorders. To identify
tional initiation, suggesting that specific translational factors or RBPs that interact with expanded CGG repeat RNAs and that may
alternative mechanisms may underlie RAN translation and its modulate RAN translation in a FXTAS disease model, we designed
contributions to repeat-associated toxicity (Green et al, 2017; Cheng an RNA-tagging system that allowed isolation and identification of
et al, 2018; Sonobe et al, 2018; Westergard et al, 2019). Indeed, CGG repeat RNA-binding proteins inside cells (Harlen & Church-
unbiased and targeted genetic approaches have identified potential man, 2017). To accomplish this, we modified previously character-
factors that preferentially modulate RAN translation including ribo- ized CGG RAN translation-specific nanoluciferase (nLuc) reporters
somal protein RPS25 and RNA helicase DDX3X (Cheng et al, 2019; by inserting PP7 viral stem-loops after the stop codon (Fig EV1A,
Linsalata et al, 2019; Yamada et al, 2019). Appendix Table S1 and Fig 1A) (Kearse et al, 2016). This reporter
One potential confounder from studies of repeat RNA binding system allows translation of the CGG RAN reporter, while keeping
proteins in FXTAS and other repeat expansion disorders to date is the PP7 stem-loop structures unperturbed to interact with the PP7
their reliance on in vitro repeat RNA capture methodologies coat protein (PCP) (Figs EV1B and 1B and C). While testing the
(Jazurek et al, 2016). As most RNAs come to interact with specific translation efficiency of this PP7-tagged reporter we found that,
RBPs during transcription, export, and/or translation as part of their consistent with previous findings, CGG RAN translation was signifi-
normal life cycle in the cell, we reasoned that critical factors cantly less efficient than AUG-driven canonical translation (Fig 1B)
involved in RAN translation, repeat RNA transport, and RNA foci (Kearse et al, 2016). Moreover, as expected, CGG RAN translation
formation/RBP sequestration might be missed with in vitro assays, was enhanced with activation of integrated stress response by thap-
which do not capture these interactions with great fidelity. It is also sigargin (TG) treatment (Fig 1C) (Green et al, 2017). Of note, addi-
possible that interactions of specific factors with CGG repeat RNA tion of PP7 stem-loops did not perturb the predicted secondary
may only occur in particular cellular states, such as after activation structure formed by the expanded CGG repeat RNA construct used
of cellular stress pathways. Dynamic RNA–protein and RNA–RNA in this study, indicating that the PP7 stem-loops should not preclude
interactions change under cellular stress (Van Treeck & Parker, RBPs from interacting with expanded CGG repeats (Appendix Fig
2018; Matheny et al, 2020) and repeat RNAs such as ALS/FTD- S1). This PP7-tagged construct was co-transfected with PP7 coat-
associated C9ORF72 GGGGCC repeats can partition into stress gran- binding protein containing a nuclear localization signal and a
ules and interact with specific proteins (Fay et al, 2017). Thus, 3xFLAG epitope tag (PCP-NLS-3xFLAG), which facilitated immuno-
capturing context-specific repeat RNA–protein interactions might precipitation (IP) using anti-FLAG antibody (Fig 1D).
reveal novel modulators of repeat RNA biology and pathogenesis. To quantitatively identify ribonucleoprotein complexes formed
To define the roles played by CGG repeat RNA binding proteins by the CGG-nLuc reporter, we used SILAC (stable isotope labeling
in both RAN translation and FXTAS pathogenesis in vivo, we devel- by amino acids in cell culture) with HEK293T cells grown in light
oped a repeat RNA-tagging system, which allows for the unbiased and medium amino acids before IP (Fig 1D) (Ong et al, 2002). We
identification of RNA-binding proteins inside cells (Harlen & used an AUG-nLuc –PP7 as control and a GGGGCC-nLuc-PP7 for
Churchman, 2017). We fused a pathogenic CGG repeat expansion comparative analysis (see Methods and for details). All these
2 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

Indranil Malik et al EMBO Molecular Medicine
A B C
D E
Figure 1. Repeat RNA-tagging system enables identification of CGG repeat specific RNA-binding proteins within cells.
A Schematic of PP7-tagged RNA reporters and PCP-NLS-FLAG constructs used in this study. Details of reporter sequences are described in Appendix Table S1.
B Relative protein expression from PP7-tagged CGG-nLuc reporters compared to AUG-driven reporters in HEK293T cells (n = 12 biological replicates).
C Expression of AUG-control and CGG-nLuc RAN translation reporters in HEK293T cells treated with the ER stress agent thapsigargin (TG, 2 lM) (n = 11 biological
replicates) normalized to vehicle (DMSO).
D Schematic of immunoprecipitation and mass spectrometry experiments aimed at identifying CGG repeat RNA-interacting proteins (see Materials and Methods for
details). Growth media, containing different isotopes of lysine and arginine for SILAC labeling, are indicated as light (Lys-0 and Arg-0) and medium (Lys-4 and Arg-
6).
E, F Log2 fold-change of the CGG-interacting protein enrichment compared AUG reporter (n = 2 independent experiments; error bars represent range between repeats)
under normal (E) and after integrated stress response activation by TG (F).
G List of top proteins enriched in CGG-PP7 RNA interaction compared to AUG in mass spectrometry experiments without or with TG treatment (TG+). Bar graphs
represent average of two biological replicates range. Enriched SR proteins are marked with blue dots.
Data information: For graphs in (B and C), error bar represents mean SD. Statistical analysis was performed using two-tailed Student’s t-test with Welch’s correction,
***P < 0.001; ****P < 0.0001.
Source data are available online for this figure.
ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 3 of 21

reporters have been characterized extensively for their RAN transla- with CGG repeat RNA during its synthesis, transport, and transla-
tion efficiency earlier (Kearse et al, 2016; Green et al, 2017; Linsa- tion (Dennis et al, 2003). As expected, previously identified CGG
lata et al, 2019). Here, we used these reporters as bait to capture interactors hnRNPA2B1 or Sam68 (KHDRBS1) interacted with our
specific proteins that may interact with these reporters and involved CGG reporter (Data Set EV1). Both AUG-initiated and CGG-repeat
in RNA export and/or translation processes. Importantly, for this constructs formed similar ribonucleoprotein complexes and exhib-
study we mainly analyzed the CGG interactome, as the GGGGCC ited significant overlap in their interactomes (Fig EV1C). However,
interactome will be the focus of a separate manuscript (see Methods several RBPs including multiple serine/arginine-rich domain (SR)
and Data Set EV1). For SILAC, AUG-nLuc-PP7 reporter was grown proteins preferentially interacted with CGG repeat reporters
in light and CGG-nLuc-PP7 reporter was grown in medium amino compared to the AUG reporter (Fig 1E–G).
acids containing HEK293T cells and both of them were co- Top RBPs that preferentially interact with CGG repeat reporter in
transfected with the 3xFLAG tagged PCP (Fig 1D). In parallel, to basal condition (no stress) include heterogeneous nuclear ribonucle-
determine RBPs that may interact with these reporters after ISR acti- oproteins hnRNPH, hnRNPC; poly(a) binding protein PABPN1,
vation, cells were treated with 2 lM TG 5 h before the IP. The PABPC1, and PABPC4; ribosomal proteins RPL18A, RPLP0, and
SILAC analysis provided more than 300 protein interactors with RPLP2 (Fig 1G). Stress-specific interactors include ubiquitin-60S
quantification of their binding preferences for AUG and CGG repor- ribosomal protein L40 (UBA52), hnRNPQ, PABPC1, LARP1, and
ters (Figs 1E and EV1C). Fewer interactors were identified after ISR YBX1 (Fig 1G). Interestingly, multiple SR proteins interact with the
activation with TG treatment (Figs 1F and EV1C). A reduction in CGG reporter basally and in response to cellular stress (Fig 1E–G).
RNA–protein interaction during integrated stress response might be SR proteins are a large family of RBPs consisting 12 structurally
a result of perturbation in normal RNA metabolism during cellular related proteins containing characteristic Arg/Ser-rich (RS) domains
stress (Bond, 2006). that influence mRNA splicing, export, stability, and translation
(Zhou & Fu, 2013). To validate some of these SR protein interac-
SR proteins selectively interact with CGG repeat RNAs tions, we took a two-pronged approach. First, we determined the
co-localization of a key SR protein, SRSF1, with CGG repeat RNA by
Gene ontology (GO) enrichment analysis of manually curated CGG- hybridization chain reaction (HCR) and immunocytochemistry
enriched protein interactors using the Database for Annotation, (ICC). HCR enabled the detection of CGG repeat RNA foci in trans-
Visualization and Integrated Discovery (DAVID) yielded top func- fected U2OS cells and showed co-localization of the RNA with
tional categories related to poly(A) RNA-binding, RNA-binding, SRSF1 (Fig 2A and B). In a parallel approach, we immunoprecipi-
ribonucleoprotein complexes, mRNA metabolic processes, RNA tated PP7-tagged reporter RNAs and immunoblotted for SRSF
processing, and translation (Fig EV1D), indicating that this RNA- proteins. SRSF1 interacted very strongly with PP7-tagged CGG
tagging IP successfully captured RBPs that may differentially interact repeat RNA reporter compared to a PP7-tag control or AUG reporter
B C
Figure 2. HCR-ICC detects co-localization of SRSF1 and CGG repeat RNA.

A Hybridization chain reaction (HCR) detection of CGG repeat RNA along with immunocytochemistry (ICC) to detect SRSF1-FLAG in U2OS cells. Expanded view of
merged channels shows the CGG RNA foci (arrowhead) and an example of SRSF1-CGG RNA foci co-localization (arrow). DAPI marks the nucleus. Scale bars are 10 µm.
B Co-localization analysis of CGG RNA (red traces) and SRSF1 (green traces) showing overlaps of CGG RNA and SRSF1. (a.u. = arbitrary unit).
C Co-immunoprecipitation of indicated SRSF proteins with PP7-tagged control sequence (partial nLUC sequence) to correct for non-specific binding, AUG, and CGG
reporter RNAs. SRSF1 and 2 specifically immunoprecipitated with PP7-tagged CGG reporter.

(Fig 2C). SRSF2 also interacted preferentially with the CGG RNA repeat elicited eye degeneration compared to an eGFP control
reporter, albeit less robustly than SRSF1 (Fig 2C). (Fig 3D). A recent study has shown that SRSF1 is required for
nuclear export of GGGGCC repeats and knockdown of SRSF1 modu-
SR proteins modulate CGG and GGGGCC repeat RNA toxicity lates GGGGCC repeat toxicity in a fly model (Hautbergue et al,
in Drosophila 2017). Consistent with this published work, in our mass spectrome-
try experiments multiple SRSF proteins interacted with the GGGGCC
To test whether any of the CGG repeat RNA interactors can modulate reporter (Data Set EV1). Thus, we asked if multiple SR proteins may
CGG RNA toxicity, we conducted a candidate-based screen using a modulate GGGGCC repeat toxicity in a Drosophila model. To this
Drosophila melanogaster model of FXTAS (Fig 3A and B) (Todd et al, end, we used a transgenic Drosophila model expressing UAS-driven
2013; Linsalata et al, 2019). This fly model carries an upstream 28 GGGGCC repeats that exhibits a rough-eye phenotype when
activation sequence (UAS)-driven 5’ UTR of human FMR1 with 90 expressed in fly eyes (He et al, 2020). We found that Drosophila
CGG repeats fused to EGFP in the +1 (FMRpolyG) reading frame. homologs of several SR proteins, including SRSF1, SRSF2, and 6,
Expression of this reporter in the fly eye via a GMR-GAL4 driver significantly reduced rough-eye phenotypes in the GGGGCC repeat
results in a rough-eye phenotype (Todd et al, 2013). We have expressing fly (Fig 3E and F). Moreover, knock down of SRSF1
previously used this fly model to screen for modifiers of FMR1 CGG significantly reduced severe rough eye phenotype observed at higher
RAN translation and identified several translation-associated factors temperature (29°C) that causes necrosis and shrinkage of GGGGCC
that modulate CGG repeat toxicity (Linsalata et al, 2019). For the repeat expressing fly eyes (Fig 3G).
modifier screen in this study, we selected a few candidates from the Ubiquitous expression of 90 CGG repeats after eclosion in adult
list of top CGG-interacting proteins (Fig 1G; Appendix Table S2) to flies using the Tub5-Geneswitch system significantly shortens lifes-
cross with GMR-GAL4 control and 90 CGG repeats expressing flies. pan (Todd et al, 2010; Linsalata et al, 2019). Conversely, we
Candidate modifiers with intrinsic toxicity were excluded from further observed that expression of siRNAs against SRSF1 under an indu-
analysis (Fig EV2A). We found that knocking down Drosophila cible Tub5-Geneswitch driver modestly increased the lifespan of
homologs of several SR proteins (SRSF1, 2, and 6) as well as transla- flies expressing 90 CGG repeats (Fig 3H). Similarly, SRSF1 knock-
tion initiation factor eIF3G, ribosomal protein RPLP0, and RNA down increased lifespan when CGG repeat was expressed selectively
helicase DHX30 significantly reduced CGG repeat RNA toxicity (Figs within neurons in adult flies under an inducible Geneswitch ElaV
3B and C, and EV2B). Factors whose knockdown significantly driver (Fig 3I). However, siRNAs against SRSF2 failed to enhance
enhanced CGG repeat toxicity in fly eye include fly homologs of survival of flies expressing 90 CGG repeats (Fig EV3A). Together,
heterogeneous nuclear ribonucleoproteins hnRNP H/F and hnRNPQ/ these results suggest that SRSF1 in particular plays a role in adult-
Syncrip; poly(a) binding protein (PABPN1); nuclear cap-binding onset CGG repeat-associated neurodegeneration in Drosophila. As
protein (NCBP1); and DExD-Box Helicase 39B (Figs 3B and EV2C). with CGG repeats, ubiquitous expression of siRNA against SRSF1
To investigate whether overexpression of SR proteins might led to a modest but statistically significant enhancement of survival
enhance CGG repeat toxicity, we developed a transgenic Drosophila of GGGGCC repeat expressing fly, but SRSF2 did not show any
model carrying UAS-driven dSF2 (Drosophila homolog of human significant change (Figs 3J and EV3B). Taken together, these results
SRSF1) and expressed this reporter in the fly eye using a GMR-GAL4 are consistent with the earlier findings related to SRSF1 and
driver. Expression of dSF2 alone did not lead to any eye abnormali- GGGGCC repeats (Hautbergue et al, 2017), while extending the
ties. However, co-expression of dSF2 significantly enhanced CGG results to a second disease-causing repeat expansion (CGG) and
Figure 3. SRSF proteins act as modifiers of CGG and GGGGCC repeat-associated toxicity in Drosophila.
A
B
Schematic of (CGG)90-EGFP construct and experimental outline for rough eye phenotype screening.
Quantitation of GMR-GAL4-driven uas-(CGG)90-EGFP eye phenotype with candidate modifiers (n ≥ 30 flies/genotype). siNTC = siRNA against a non-targeting
▸
control gene (mCherry). Different siRNA lines for the same target gene are numbered (#1 and #2). Error bars represent mean SD.
C Representative photographs of fly eyes expressing either GMR-GAL4 driver alone or with uas-(CGG)90-EGFP construct, with fly SRSF1 (dSF2) and SRSF2 (dSC35)
knockdown or disruptions (insertion).
D Representative photographs of fly eyes and quantitation (below) of rough eye scores with fly SRSF1 overexpression (dSF2 OE); n = 20–32/genotype. Error bars
represent mean SD.
E Representative photographs of fly eyes expressing GMR-GAL4-driven (GGGGCC)28-EGFP with indicated uas-siRNAs against fly SRSF genes in comparison with non-
targeting control (NTC) siRNA against LUC/luciferase.
F Quantitation of (GGGGCC)28-EGFP rough eye phenotype with SRSF modifiers (n ≥ 30 flies/genotype). Error bars represent mean SD.
G Representative photographs of fly eyes expressing GMR-GAL4-driven (GGGGCC)28-EGFP at 29°C along with the quantifications of necrosis and eye width. n = 28–
30/genotype. Error bars represent mean SD.
H, I Survival assays of flies expressing (CGG)90-EGFP under Tub5-GS (H) and ELAV-GS (I) drivers with control or SRSF1 siRNAs. Expression of (CGG)90-EGFP was initiated
with addition of drug starting 1 day post-eclosion and continued through experiment (Log-rank Mantel–Cox test; n = 98–101/genotype for Tub-GS and n = 120–
141/genotype for ELAV-GS flies); *P < 0.05, **P < 0.01.
J Survival assays of (GGGGCC)28-EGFP expressing fly under Tub5-GS driver (Log-rank Mantel–Cox test; n = 71–93/genotype) with control or SRSF1 siRNAs. **P < 0.01.
Data information: For eye scoring, target siRNA lines were compared to non-targeting control siRNA lines using a two-tailed Student’s t-test with Welch’s correction for
multiple comparisons. **P < 0.01; ***P < 0.001; ****P < 0.0001. Human orthologs of fly genes are used for labeling. Details of fly genes are described in
Appendix Table S2.

A B
C D
F G
H I J
Figure 3.

suggesting that other SR proteins can also be involved in repeat asked if inhibition of SRPK1 can modulate RAN translation in
expansion-associated toxicity. mammalian cells. We hypothesized that SRPK1 is required for SR
protein-mediated export of repeat RNAs into the cytoplasm
SRSF Protein Kinase 1 (SRPK1) modifies repeat RNA toxicity (Fig 5A). Thus, SRPK1 inhibition can possibly lead to a decrease in
in Drosophila cytoplasmic levels of repeat RNAs, resulting ultimately in a
decrease level of RAN translation. To monitor RAN translation in
Serine–arginine protein kinases (SRPKs 1–3) selectively phosphory- mammalian cells, we used our previously characterized nLuc-
late SR proteins and modulate their subcellular localization (Zhou & based CGG RAN translation reporter consisting of a 3xFLAG-
Fu, 2013). In addition, many non-splicing functions for SRPKs tagged nanoluciferase (nLuc-3XF) downstream of the 5’ UTR of
have been reported, including roles in tau phosphorylation and human FMR1 (Kearse et al, 2016). In order to test the effects of
Alzheimer’s disease (AD) pathogenesis (Hong et al, 2012) and in SRPK1 inhibition on RAN translation, we used two known chemi-
translational regulation (Brown et al, 2014). SRPKs exhibit highly cal inhibitors of SRPK1 (Fig 5A). First, we tested the effects of
tissue-specific expression profiles, suggesting that these kinases may SRPIN340, an ATP-competitive SRPK inhibitor, on RAN translation
have specialized functions (Wang et al, 1998; Nakagawa et al, by pre-treating HEK293T cells with SRPIN340 followed by trans-
2005). SR proteins are misregulated in multiple cancers, and phar- fecting CGG RAN translation reporters (Fukuhara et al, 2006).
macological targeting of the SRPK1-SR axis has been proposed as a SRPIN340 treatment at 50 lM led to a significant and selective
potential therapeutic approach (Amin et al, 2011; Mavrou et al, decrease in +1CGG RAN translation as detected by immunoblot
2015; Mavrou & Oltean, 2016). As multiple SR proteins modulate (Fig 5B). Similarly, SRPIN340 treatment at 30 lM inhibited CGG
both CGG and GGGGCC repeat RNA toxicities in our fly models, we RAN translation as measured by nanoluciferase assay, but this had
asked if SRPK1 may also affect repeat RNA toxicity in flies. To this no impact on AUG-nLuc expression (Fig 5C). The requirement of
end, first we tested the effects of altering expression of SRPK1 on lower concentration of SRPIN340 treatment for RAN inhibition in
CGG repeat RNA toxicity in our FXTAS fly model. Interestingly, we nanoluciferase assays compared to Western, is possibly due the
found that selective knockdown of the Drosophila homolog of differences in the sensitivity of detection between these assays.
SRPK1 (dSRPK1) reduced CGG repeat RNA toxicity and significantly The effects of SRPIN340 was not isolated to FMRpolyG synthesis:
improved the CGG repeat rough eye phenotype (Fig 4A and B). We showed that SRPK1 inhibition by SRPIN340 also suppressed
Next, we asked if altering SRPK1 expression can also reduce GGGGCC RAN translation (GA70) and +2CGG RAN translation
GGGGCC repeat RNA toxicity in our fly model. Indeed, expression (FMRpolyA) using previously published reporters (Fig 5D) (Kearse
of siRNA against SRPK1 (dSRPK1) reduced GGGGCC repeat toxicity et al, 2016; Green et al, 2017).
and significantly improved the associated rough eye phenotype As pharmacological agents may have off-target effects, we also
(Fig 4C and D). Additionally, ubiquitous expression of siRNAs evaluated the impact of a second SRPK inhibitor, SPHINX31 (Gam-
against dSRPK1 under an inducible Tub5-Geneswitch driver led to a mons et al, 2013; Batson et al, 2017). Similar to SRPIN340,
modest but statistically significant effect on lifespan of flies express- SPHINX31 significantly and selectively inhibited +1CGG, GGGGCC
ing GGGGCC repeats (Fig EV3C). We also confirmed that improve- and +2CGG RAN translation (Fig 5E and F). These results indicate
ment of CGG RNA toxicity in Drosophila occurs due to decreased that pharmacological inhibition of SRPK1 has a general inhibitory
expression of target genes by siRNA-mediated knockdown effect on RAN translation.
(Fig EV3D). Together, these results suggest that SRPK1 can modify
repeat expansion-associated toxicity, either directly through modu- SRPK1 inhibitors prohibit stress-induced increase in
lating SR proteins or through independent pathways. RAN translation
Amelioration of CGG RNA toxicity by SRSF1 and SRPK1 knock-
down might occur through altering the levels of RAN products in RAN protein levels increase under various stress conditions,
Drosophila. To test this possibility, first we assessed if any of the including ER stress (Green et al, 2017; Cheng et al, 2018; Sonobe
modifiers can decrease GFP inclusions (FMRpolyG-EGFP) in fly eyes. et al, 2018; Westergard et al, 2019). During integrated stress
A previous study from our laboratory has shown that RAN translation response (ISR), phosphorylation of eIF2a leads to a decrease in
of the CGG90-EGFP reporter can form GFP inclusions in Drosophila global translation, while RAN translation remains unperturbed.
eye, while expression of GFP alone do not form inclusions (Todd This creates a feed-forward loop that leads to production of more
et al, 2010). siRNA-mediated knockdown of either SRSF1 or SRPK1 RAN proteins, which contribute to neuronal dysfunction and
significantly reduced GFP inclusions in eye compared to control death. As SPRK1 inhibitors selectively suppressed RAN transla-
siRNA (Fig 4E and F). In addition, FMRpolyG-EGFP levels in adult tion, we wondered if SRPK1 inhibition might also impede stress-
flies expressing CGG repeats within neurons were modestly decreased induced enhancement of RAN translation. Pre-treating cells with
by siRNA-mediated knockdown of SRSF1 (Fig EV3E). Together, these 50 lM SRPIN340 led to a complete blockade of thapsigargin-
results suggest that SRSF1 and SRPK1 modify CGG repeat-associated induced enhancement of +1CGG RAN translation as detected by
toxicity in Drosophila through altering RAN protein levels. immunoblot and luciferase assays (Fig 6A and B). SRPIN340 also
suppressed stress-induced GGGGCC (GA70) and +2CGG RAN
SRPK1 inhibitors modulate RAN translation in translation (Fig 6B–D). Together, these results suggested that
cell-based reporters pharmacological inhibition of SRPK1 can prohibit both basal level
and stress-induced increase in RAN translation across at least two
Since genetic ablation of Drosophila homolog of SRPK1 altered RAN different repeats and at least two reading frames of the CGG
protein (FMRpolyG-EGFP) levels in a fly model of FXTAS, we next repeat.

A B
C D
E F
Figure 4. SRPK1 knockdown modifies CGG and GGGGCC repeat-associated toxicity in Drosophila.
A, B Representative (A) photographs of fly eyes and (B) quantitation with siRNA-mediated knockdown of SRPK1 (dSRPK1); n = 30–34/genotype. Error bars represent
mean SD.
C Representative photographs of fly eyes expressing GMR-GAL4-driven (GGGGCC)28-EGFP with siRNA-mediated knockdown of SRPK1 or disruption by insertion.
D Quantitation of rough eye phenotypes. t-test with Welch corrections for comparisons with the control; n = 31–34 flies/ genotype. Error bars represent mean SD.
E Representative external eye imaging for the detection of GFP aggregates caused by (CGG)90-EGFP transgene expression (top). Converted images used to quantify
total intensity of GFP puncta (bottom).
F Depletion of SRSF1 or SRPK1 by RNAi results in reduced (CGG)90-EGFP puncta compared to control siRNA as quantified by total intensity (a.u. = arbitrary unit).
n = 13–15 flies/genotype. Error bars represent mean SD.
Data information: Statistical analysis was performed using two-tailed Student’s t-test with Welch’s correction, ***P < 0.001; ****P < 0.0001.
To assess the mechanism by which SPRK1 inhibitors precluded did observe that 50 lM SRPIN340 treatment lead to a modest block-
stress-induced RAN translation, we asked whether SRPK1 inhibitors ade in the enhancement of phosphorylated eIF2a during ISR induc-
directly target cellular ISR pathways. To test this, we measured tion in the presence of +1CGG repeat RNA (Fig 6F). We conclude
eIF2a phosphorylation levels in TG-treated (stress-induced) cells in that although SRPIN340 does not directly target ISR pathways alone,
the presence or absence of SRPIN340. We did not observe any it may decrease the overall burden of cellular stress triggered by the
significant changes in the levels of phosphorylated eIF2a after ISR presence of +1CGG repeat RNA through inhibiting RAN protein
induction with TG in presence of SRPIN340 (Fig 6E). However, we production.

SRPK1 inhibitors trigger nuclear accumulation of CGG ribosomes. Alternatively, SRPK1 inhibition could directly impair
repeat RNAs RAN translation through either an SR protein dependent or indepen-
dent pathway. We tested both of these possibilities.
Genetic or pharmacological inhibition of SRPK1 could theoretically Inhibition of SRPK1 results in decreased phosphorylation of
reduce RAN translation through multiple mechanisms (Fig 5A). target SR proteins, particularly SRSF1, resulting in a reduction of
SRPK1 inhibition may alter SR protein-mediated export of repeat SRSF1 nuclear import (Zhou & Fu, 2013; Gonçalves et al,
RNAs, thus preventing their access to assembled and active 2014). Consistent with this, we found that SRPIN340 treatment
A B
C D
E F
Figure 5.

◀ Figure 5. SRPK1 inhibitors selectively suppress RAN translation.

A Schematic of SRPK1 signaling pathway that regulates subcellular SRSF1 localization, which in turns can impact export and translation of repeat RNAs in the
cytoplasm. SRPK1 may also act directly on protein translation pathways (dotted arrow). Known pharmacological compounds that inhibit SRPK1 can disrupt this
pathway.
B Anti-FLAG immunoblot of DMSO and SRPIN340 pre-treated HEK293T cells expressing AUG-nLuc-3xFLAG control or CGG-nLuc-3xFLAG RAN translation reporters.
b-Actin is used as a loading control. To prevent signal saturation, AUG-nLuc lysate was diluted 1:3 in sample buffer prior to loading (n = 3 biological replicates).
Schematics of the AUG-nLUC-3xFLAG and +1CGG(100)-nLuc-3xFLAG reporters presented on top.
C Relative expression of AUG-nLuc and CGG-nLuc reporters in HEK293T cells (n = 8–9 biological replicates) following treatment with DMSO and SRPIN340.
D Anti-FLAG immunoblot of DMSO and SRPIN340 pre-treated HEK293T cells expressing GGGGCC-nLuc-3xFLAG (GA70) and +2CGG-nLuc-3xFLAG (FMRpolyA) RAN
translation reporters (n = 3 biological replicates). Schematics of the GA70 (GGGGCCx70) and +2CGG reporters presented on top.
E Immunoblot of DMSO and SPHINX31 pre-treated HEK293T cells expressing AUG-nLuc-3xFLAG control or CGG-nLuc-3xFLAG RAN translation reporters (n = 3 biological
replicates).
F Anti-FLAG of DMSO and SPHINX31 pre-treated HEK293T cells expressing GGGGCC-nLuc-3xFLAG (GA70) and +2CGG-nLuc-3xFLAG (FMRpolyA) RAN translation reporters
(n = 3 biological replicates).
Data information: Error bars represent mean SD. Statistical analysis was performed using two-tailed Student’s t-test with Welch’s correction. *P < 0.05; **P < 0.01,
***P < 0.001, and ****P < 0.0001. To prevent over-exposure, the AUG-nLuc lysate was diluted 1:3 in sample buffer.
significantly decreased levels of phosphorylated SRSF1/2 in the mammalian cells and knockdown of SRPK1 in Drosophila
cytoplasm (Figs 7A and EV4A). Likewise, SRPIN340 treatment suppressed CGG repeat toxicity, we asked whether SRPK1 inhibitors
reduced nuclear SRSF1 protein levels (Fig 7B). If SRSF1 is required could mitigate CGG repeat toxicity in mammalian neurons. We
for nuclear export of CGG repeat RNAs, as described earlier for expressed +1(CGG)100-EGFP reporters encoding for +1 FMRpoly(G)
GGGGCC repeat RNAs (Hautbergue et al, 2017), then a decrease in in rat primary neurons along with an mApple reporter plasmid that
nuclear SRSF1 levels could impair export of CGG repeat RNAs. allowed for selective tracking of transfected cells using an auto-
Consistent with this prediction, treating cells with SRPIN340 trig- mated fluorescence microscopy assay system (Barmada et al, 2015;
gered accumulation of CGG repeat RNAs inside the nucleus as Linsalata et al, 2019). An AUG-driven EGFP reporter served as a
measured by HCR (Fig 7C). Together, these results suggest that inhi- control for transfection and exogenous protein expression-
bition of SRPK1 leads to impaired export and nuclear accumulation associated toxicity. Consistent with previous results (Linsalata et al,
of repeat RNAs, resulting in decrease production of RAN proteins in 2019), +1(CGG)100-EGFP expression markedly reduced neuronal
the cytoplasm (Fig 5A). survival compared to EGFP expression over 10 days (Figs 8A and
EV5A). We next treated neurons with SRPIN340 at a range of
SRPK inhibition directly impairs RAN translation from concentrations from 10 to 50 lM before +1(CGG)100-EGFP reporter
CGG repeats transfection (Fig EV5B). Neuronal survival rate was significantly
improved with SRPIN340 treatment at concentrations of 30, 40, and
Next, we asked if SRPK1 inhibition can directly affect RAN transla- 50 lM compared to DMSO treatment. However, SRPIN340 appeared
tion. We tested this by two means. First, we transfected HEK293T to have some neurotoxicity itself on EGFP-transfected neurons at
cells with in vitro transcribed +1CGG RAN reporter RNA in presence higher concentrations, with SRPIN340 treatment at 40 lM showing
or absence of SRPIN340 (Fig 7D). RNA transfection precludes the the most favorable and selective effects on neuronal survival (Figs
requirement of nuclear export of the RNA and reporter RNAs 8A and EV5A and B).
become readily available in the cytoplasm for translation. Interest- To confirm that this suppression of neurotoxicity by SRPIN340
ingly, SRPIN340 treatment led to a significant decrease in RAN treatment is not a drug-specific effect, we also tested SPHINX31
translation from +1CGG reporter RNAs, as detected by immunoblot effects on +1(CGG)100-EGFP-induced toxicity. Similar to SRPIN340,
and luciferase assay (Fig 7D and E). We also confirmed direct and SPHINX31 significantly improved +1(CGG)100-EGFP expressing
selective inhibition of RAN translation by SRPIN340 using an neuronal survival at all tested concentrations (Fig EV5C and D).
in vitro translation system. Addition of 50 lM SRPIN340 to a rabbit SPHINX31 elicited maximum suppression of neurotoxicity at
reticulocyte lysate (RRL) selectively inhibited translation of in vitro concentrations of 8 and 10 lM, consistent with our observation of
transcribed +1CGG reporter RNA without affecting AUG-nLUC inhibition of RAN translation in a similar range of concentrations.
reporter RNA translation (Fig 7F). However, SRPIN340 treatment At 8 lM concentration, SPHINX31 showed a promising effect on
effects on +1CGG RAN translation were smaller with RNA reporter neuronal survival against +1(CGG)100-EGFP reporter-induced toxic-
transfection compared to plasmid reporter transfection (Fig EV4B). ity over any intrinsic drug-associated toxicity (Fig 8B). Together,
Taken together, these results suggest that SRPK1 inhibition these findings suggest that pharmacological inhibition of SRPK1 can
suppresses RAN translation through at least two complementary suppress neurotoxicity of expanded CGG repeats through suppres-
mechanisms: nuclear RNA export and translation efficiency. sion of RAN translation.
Inhibition of SRPK1 enhances survival of (CGG)100

expressing neurons Discussion
As pharmacological inhibition of SRPK1 suppressed RAN translation Nucleotide repeat expansions as RNA form complex structures that
from both +1 FMRpoly(G) and +2 FMRpoly(A) translation frames in bind and sequester specific RNA binding proteins within different

A B
C D
E F
Figure 6. SRPK1 inhibition prevents stress-induced enhancement of RAN translation.

A Expression of +1CGG-nLuc-3xFLAG RAN translation reporters in HEK293T cells treated with 2 lM TG (for stress induction) analyzed by immunoblot (n = 3 biological
replicates). To evaluate effects of SRPK1 inhibition, cells were pre-treated with DMSO or SRPIN340 before reporter transfection.
B Relative expression of +1CGG-nLuc reporters in HEK293T cells (n = 6 biological replicates) following stress induction with 2 lM TG treatment. Values normalized to
vehicle (DMSO) treatment. As in (A), cells were pre-treated with DMSO or SRPIN340 before reporter transfection.
C, D Expression of GGGGCC-nLuc-3xFLAG (C) and +2CGG-nLuc-3xFLAG (D) RAN translation reporters in HEK293T cells treated with 2 lM TG (for stress induction)
analyzed by immunoblot (n = 3 biological replicates). To evaluate effects of SRPK1 inhibition, cells were pre-treated with DMSO or SRPIN340 before reporter
transfection.
E SRPK1 inhibition by SRPIN340 did not alter eIF2a phosphorylation in response to 2 lM TG (for stress induction) in HEK293T cells by immunoblot (n = 3 biological
replicates).
F SPRK1 inhibition suppressed eIF2a phosphorylation in HEK293T cells treated with 2 lM TG (for stress induction) in the presence of the +1CGG reporter. Cells were
pre-treated with DMSO or SRPIN340 before reporter transfection (n = 3 biological replicates).
Data information: Error bars represent mean SD. Two-tailed Student’s t-test with Welch’s correction. *P < 0.05; **P < 0.01; ***P < 0.001; and ****P < 0.0001.
cellular compartments (Krzyzosiak et al, 2012; Jazurek et al, 2016; cellular context into consideration (Jazurek et al, 2016). As such,
Ciesiolka et al, 2017). These interactions influence repeat RNA they have the potential to miss important interacting proteins that
stability, distribution, and translation efficiency, and alter the might be of lower overall abundance or require interaction within
behavior of the RNA binding proteins, which interact with repeat the context of specific cellular compartments or in vivo RNA struc-
RNAs. To date, most studies of repeat RNA–protein interactions tures. Here, we utilized an alternative method for identifying RNA–
have relied on in vitro capture methods that do not take mRNA RBP interactors that allows for capture of complexes that form

A B
C D
E F
Figure 7. SRPK1 inhibition alters CGG repeat RNA localization and directly inhibits RAN translation.
A Immunoblot of cytoplasmic phospho-SRSF1/2 after SRPIN340 treatment. GAPDH is used as the loading control. Error bars represent mean SD (n = 3 biological
replicates). Statistical analysis was performed using Student’s t-test with Welch’s correction. *P < 0.05. Full images of the same blot showing levels of all phospho SR
proteins (pan SRSF phosphorylation) after SRPIN340 treatment are presented in Fig EV4A.
B ICC images of FLAG-tagged SRSF1 after SRPIN340 treatment compared to vehicle (DMSO). Quantification shows the ratio of nuclear and cytoplasmic intensity of
SRSF1 signal (see methods for details). Error bars indicate mean 95% CI (n = 85–126 cells/condition). Statistical analysis was performed using t-test with Welch’s
correction, ****P < 0.0001.
C Nucleocytoplasmic distribution CGG repeat RNA after SRPIN340 treatment compared to vehicle (DMSO) as detected by HCR. Quantification shows the ratio of nuclear
and cytoplasmic intensity of CGG RNA signal (see methods for details) as parts of whole. Error bars indicate mean 95% CI (n = 124–151 cells/condition). Statistical
analysis was performed using t-test with Welch’s correction, ****P < 0.0001.
D Anti-FLAG immunoblot blot of DMSO and SRPIN340 pre-treated HEK293T cells transfected with in vitro transcribed CGG-nLuc-3xFLAG reporter RNA. b-Actin is used as
a loading control. Error bars represent mean SD (n = 6 biological replicates). Statistical analysis was performed using Student’s t-test with Welch’s correction.
*P < 0.05.
E Relative expression of in vitro transcribed CGG-nLuc reporter in HEK293T cells pre-treated with DMSO and SRPIN340. Error bars represent mean SD (n = 9
biological replicates).
F Expression of in vitro transcribed Aug-nLuc and CGG-nLuc reporters in rabbit reticulocyte lysate (RRL) in vitro translation system following pre-treatment with DMSO
or SRPIN340. Error bars represent mean SD (n = 9 biological replicates).
Data information: For E and F, statistical analysis was performed using Student’s t-test with Welch’s correction. *P < 0.05; **P < 0.01.

A B
Figure 8. SRPK1 inhibition mitigates (CGG)100 RNA toxicity in primary rat cortical neurons.
A, B Pharmacological targeting of SPRK1 with (A) 40 lM SRPIN340 or (B) 8 lM SPHINX31 reduced the cumulative risk of death in +1(CGG)100-EGFP (encoding
FMRpolyG) expressing neurons. n = # of neurons quantified for each condition; Cox proportional hazard analysis, ***P < 0.001.
within cells under both basal conditions and in response to cellular Because of the central importance of SR proteins in many cellular
stress. By applying this tool to CGG repeats that cause FXTAS, we functions, we pivoted to study whether altering the behavior of the
identified a series of novel interactors, including factors with dif- major SRSF kinase, SPRK, might serve as a better biological target
ferential interaction profiles after stress induction (Fig 1E–G). Evalu- in repeat expansion disorders. Consistent with this idea, we
ation of these interactors identified SRSF proteins as playing direct observed that SRPK inhibition, which influences SRSF1 distribution
roles in CGG repeat toxicity in model systems. Moreover, genetic and behavior, led to a marked nuclear retention of CGG repeat RNA
and pharmacological targeting of the major SRSF kinase (SRPK1) in the nucleus (Fig 7A–C). Thus, at least some of the observed
significantly impaired RAN translation at multiple GC-rich repeat effects of SRPK inhibition and presumably SRSF1 expression manip-
sequences and suppressed toxicity in Drosophila and CGG repeat ulation on both CGG repeat distribution and toxicity is mediated by
expressing rodent neurons (Figs 5–8). Taken together, these studies altered repeat RNA export. However, we also observed that SRPK
highlight the value of RNA-tagging for identification of RNA–RBP inhibition had a direct effect on RAN translation from CGG repeats,
interactions and suggest that SRPK1 may serve as a therapeutic as determined by in vitro translation assays and by RNA reporter
target worthy of further evaluation in repeat expansion disorders. transfections (Fig 7D–F). SRPK1 may have indirect effects on
Our screening identified multiple SR proteins that interact with protein translation through modulation of eIF4E phosphorylation as
CGG repeat RNA and loss of SR protein orthologues in Drosophila well as through impacts on AKT signaling (Zhou & Fu, 2013; Brown
suppressed CGG repeat toxicity (Figs 1–3). Though SR proteins play et al, 2014). Alternatively, inhibition of SRPKs can also affect trans-
a major role in splicing, they are also implicated in mRNA export, lation through altering the phosphorylation level of eIF4G, as
regulation of RNA stability, and translation (Jeong, 2017). SRSF1, 2, observed earlier for simultaneous knockdown of multiple SRPKs
and 9 have been previously shown to interact with other repeat (Hu et al, 2012). As such, the exact mechanism(s) by which SRPK
RNAs, with functional implications for GGGGCC repeat RNA associ- inhibition is acting to impair RAN translation and suppress CGG
ated with C9ALS (Sato et al, 2009; Donnelly et al, 2013; Lee et al, repeat toxicity will require further study.
2013; Cooper-Knock et al, 2014). Specifically, sense GGGGCC RNA Protein kinases are attractive targets for drug development.
has been shown to interact with both SRSF1 and SRSF2 and anti- Notably, SPRK1 inhibitors are actively pursued as potential anti-
sense C4G2 RNA has been shown to colocalize with SRSF2 (Cooper- cancer drugs (van Roosmalen et al, 2015; Mavrou & Oltean, 2016;
Knock et al, 2015; Hautbergue et al, 2017). Moreover, prior data Chandra et al, 2020). SRPK1 phosphorylates multiple serine resi-
suggested that SRSF1 is important for nuclear export of GGGGCC dues in the RS1 domain of SRSF1 to regulate its nuclear localization
repeat RNA (Hautbergue et al, 2017). Interestingly, analysis of (Zhou & Fu, 2013). Altered levels of SRSF1 have been reported in
C9ALS patient cerebellum samples has shown extensive alternative many cancers, where phosphorylation of SRSF1 plays a decisive role
splicing (AS) defects in transcripts targeted by hnRNPH1 and SRSF1, in alternative splicing of disease-associated transcripts (Anczuk ow
indicating that SRSF1 sequestration by GGGGCC RNA may account et al, 2015; Sheng et al, 2018). Furthermore, in some cancers
for this altered splicing (Prudencio et al, 2015; Conlon et al, 2016). misregulation of SRPK1 has been linked with cell proliferation,
Our data confirm many of these initial observations and extend migration, and angiogenesis (van Roosmalen et al, 2015). In light of
these results to CGG repeats. Together, they strongly indicate that these findings, our discovery of SRPIN340 and SPHINX31 as poten-
SR proteins play an important role in repeat RNA toxicity across tial inhibitors of RAN translation is worthy of further pursuit.
multiple repeats. Furthermore, our mechanistic studies strongly suggest that SRPK1

inhibition may affect RAN translation through multiple pathways Sigma A1978), 1:1,000 SRSF1 (Rabbit, Proteintech 12929-2-AP),
(Fig 7). This makes SRPK1 a robust candidate for targeting RAN 1:1,000 SRSF2 (Rabbit, Proteintech 20371-1-AP), 1:1,000 Anti-
inhibition that can be applied across multiple repeat expansion Phosphoepitope SR proteins, clone 1H4 mouse (MABE50, Milli-
disorders. However, there may be some intrinsic toxicity associated pore Sigma), GAPDH (mouse, Santa Cruz sc-32233), 1:1,000
with these compounds, given that we observed modest toxicity in eIF2a/EIF2S1 (phospho S51) (rabbit, Abcam ab32157), and
neurons expressing GFP alone, especially for SRPIN340 (Figs 8A 1:1,000 GFP (mouse, Roche/Sigma 11814460001) in 5% non-fat
and EV5). Interestingly, SPHINX31, which has been shown to dry milk. HRP-conjugated goat-anti-mouse (115-035-146) or goat-
inhibit SRPK1 more potently than SRPIN340, appeared less neuro- anti-rabbit (111-035-144) secondary antibodies (Jackson Immu-
toxic (Figs 8A and EV5) (Gammons et al, 2013), perhaps due to dif- noResearch Laboratories) were used at a 1:10,000 dilution in 5%
ferential effects of SRPIN340 on SRPK2 (Fukuhara et al, 2006). This non-fat dry milk.
indicates SPHINX31 to be a more favorable candidate for develop-
ment in this context. Plasmids
Besides SRSFs, our dual screens identified a number of other
intriguing RNA-binding proteins that may play a role in CGG repeat- NanoLuciferase (nLuc) reporters cloned into pcDNA3.1(+) vector
associated toxicity. For some of these, knockdown mitigates rough encoding AUG-nLuc-3xFLAG, +1CGG100-nLuc-3xFLAG (FMRpolyG),
eye phenotypes in flies, such as DHX30 and eIF3G, while for others +2CGG100-nLuc-3xFLAG (FMRpolyA), GGGGCC70-nLuc-3xFLAG
such as hnRNPQ, knockdown enhances rough eye phenotypes in a (GA70) sequences used in this paper are published earlier (Kearse
fashion more consistent with sequestration (Figs 1 and EV2B and et al, 2016; Green et al, 2017; Linsalata et al, 2019). 2x PP7 stem-
C). DHX30 is an ATP-dependent DEAD/H RNA helicase that has been loop sequence as described earlier (Coulon et al, 2014; Harlen &
implicated in translation regulation apoptosis-associated transcripts Churchman, 2017) was synthesized from GeneWitz. In order to make
(Rizzotto et al, 2020). A previous Drosophila screen identified the fly the RNA-tagging construct, pcDNA3.1(+) AUG-nLuc-3xFLAG and
homolog of DHX30 as a modest modifier of rough eye phenotype in pcDNA3.1(+) +1CGG100-nLuc-3xFLAG vectors were modified in two
this FXTAS fly model, which we also observed here (Linsalata et al, steps. First, nLuc reporter sequence was PCR modified to introduce a
2019). Given the roles of other DEAD-box helicases in RAN transla- stop codon after nLuc sequence and remove the 3xFLAG sequence.
tion, including DDX3, the selective identification of this factor as a This PCR product was cloned back into the original vectors to make
CGG repeat RNA interactor suggests the need for further study of pcDNA3.1(+) AUG-nLuc and pcDNA3.1(+) +1CGG100-nLuc vectors.
how it might alter repeat RNA behavior (Linsalata et al, 2019). Simi- Finally, 2× PP7 stem-loops were cloned into these vectors using ApaI
larly, the mammalian eIF3 complex protein eIF3F has already been restriction site to finally make—pcDNA3.1(+) AUG-nLuc-PP7 and
implicated in RAN translation at CAG and GGGGCC repeats (Ayhan pcDNA3.1(+) +1CGG90-nLuc-PP7 constructs. PCP-NLS sequence was
et al, 2018). While eIF3G is one of the core subunits, eIF3F a non- PCR amplified from a PCP containing plasmid to introduce SV40 NLS
core regulatory subunit of the eIF3 complex (Hinnebusch, 2006). sequence right after PCP. The PCP sequence-containing plasmid was
Identification of this complex in association with the repeat RNA generously gifted by Brittany Flores (Flores et al, 2019) and originally
suggests that it may play a role in its translation, especially given reported here (Yan et al, 2016). Then, a 3xFLAG sequence was ampli-
prior studies suggesting eIF3 may be critical in non-canonical initia- fied from the AUG-nLuc-3xFLAG plasmid and PCR sewed with the
tion events. Lastly, hnRNPQ is implicated as a key factor in control- PCP-NLS fragment to clone in pcDNA3.1(+) vector using KpnI and
ling the translation of FMRP from FMR1 transcripts through an IRES- ApaI restriction sites to finally make the pcDNA3.1(+) PCP-NLS-
mediated mechanism (Choi et al, 2019). Given recent studies linking 3xFLAG construct.
CGG repeats, RAN translation, and FMRP synthesis (Rodriguez et al, See Appendix Table S1 for reporter sequences used in this study.
2020), this factor will require further evaluation in RAN translation
assays at this and other repeat structures. Cell culture, drug treatments, and reporter assays
In sum, we developed an in-cell method for identifying repeat
RNA binding proteins and applied it to CGG repeats to reveal both HEK293T and U2OS cells were purchased from American Type
novel interactors and phenotypic modifiers associated with these Culture Collection (ATCC) and cultured in DMEM media supple-
repeat expansions. Future comparative studies with other repeat mented with 10% FBS. They were confirmed to be mycoplasma free
elements and in other systems and cellular compartments should in regular interval.
allow for a better understanding of repeat RNA–protein complex Luminescence assays were performed as described earlier with
formation and interactions in vivo and guide us in our understanding slight modifications (Green et al, 2017; Linsalata et al, 2019).
of the native functions of repeat elements, and how repeats as RNA Briefly, HEK293T cells were seeded in 96-well plates at a concentra-
cause disease and potentially what drug targets are likely to serve as tion of 2 × 104 cells/well and transfected ~24 h later at ~ 70% con-
effective therapeutics in these currently untreatable disorders. fluency with 25 ng nLuc reporter plasmid and 25 ng pGL4.13 Firefly
luciferase reporter plasmid using a ratio of 2:1 jetPRIME (Polyplus)
to DNA following manufacturer’s recommendation. ~24 h post-
Materials and Methods transfection cells were lysed with 70 ll Glo Lysis buffer (Promega)
by incubating for 5 min on a shaker at room temperature. Then,
Antibodies 25 ll of lysate was mixed with NanoGlo substrate diluted 1:50 in
NanoGlo buffer (Promega) and 25 ll of lysate was mixed with ONE-
For Western blots, following antibodies were used: FLAG-M2 at Glo luciferase assay buffer (Promega) in opaque 96-well plates.
1:1,000 dilution (mouse, Sigma F1804), 1:2,500 b-Actin (mouse, Reaction was allowed to continue for 5 min on a shaker in the dark.

Finally, luminescence was measured on a GloMax 96 Microplate (Lys-4) and L-arginine [13C6] HCl (Arg-6); while for heavy SILAC:
Luminometer. RNA transfections were performed with in vitro tran- L-lysine [13C6, 15N2]HCl (Lys-8) and L-arginine [13C6, 15N4]HCl
scribed reporter RNAs (nLuc and Firefly luciferase reporters) using (Arg-10) were added to the medium and filter-sterilized with a 0.2-
TransIT-mRNA transfection kit (Mirus), per manufacturer’s recom- µm filter. For transfection, cells were seeded at 2 × 106 cells/10 cm
mendation. ~24 h post-transfection luciferase assays performed as dish and transfected at ~70% confluency with 5.25 lg of PP7-tagged
described earlier. RNA bait plasmids [pcDNA3.1(+) AUG-nLuc-PP7, pcDNA3.1(+)
For Western blots, HEK293T cells were seeded in 24-well plates +1CGG100-nLuc-PP7, and pcDNA3.1(+) (GGGGCC)70-nLuc-PP7]
at a concentration of 1.5 × 105 cells/ml and transfected 24 h later at along with 0.75 lg of PCP-NLS-3xF plasmid, using jetPRIME
~70% confluency with 250 ng nLuc reporters using a ratio of 2:1 reagents. Light SILAC was used for AUG-nLuc-PP7, medium SILAC
jetPRIME as described earlier. 24-h post-transfection cells were lysed +1CGG100-nLuc-PP7, and heavy SILAC was used for 1(+)
in 300 ll of RIPA buffer containing protease inhibitor cocktail (GGGGCC)70-nLuc-PP7 reporter. For ISR activation, cells were
(cOmpleteTM Mini, Sigma) for 30 min at 4°C with occasional vortex- seeded similarly and transfected for 19 h, followed by 5 h of treat-
ing. Lysates were cleared by centrifugation at 14,000 rpm for ment with 2 lM Thapsigargin. Three 10-cm dishes were used per
10 min, mixed with 6× SDS sample buffer, and boiled at 90°C for condition. 24 h after transfection, cells were isolated by trypsiniza-
10 min before running on SDS–PAGE. If required, lysates were tion and washing with 1× PBS. Cell pellets were immediately flash-
stored at 80°C for future experiments. For immunoblot after drug frozen using liquid nitrogen and proceeded to IP. Cell pellets from
treatment or stress induction, HEK293T cells were seeded in 24-well three 10 cm plates were pooled together and lysed in 1 ml of NP-
plates at 1.5 × 105 cells/ml and treatments were performed as 40 buffer (supplemented with cOmpleteTM Mini protease inhibitor,
described earlier. For each experimental condition, at least three 1 mM PMSF, NEB Murine RNase inhibitor, and RNaseIN) by incu-
biological samples were run on 10% SDS–PAGE along with a stan- bating at 4°C for 30 min with occasional pipetting to mix. Lysates
dard curve for quantification of protein expression. Band intensities were cleared by centrifugation at 20,000 g for 10 min, and the
were measured using ImageJ and plotted using GraphPad Prism. supernatant was transferred into a new tube. Protein concentration
For SRPK1 inhibitors, HEK293T cells were plated as described for each sample was measured by BCA assay (23227, Thermo
above and pre-treated with SRPIN340 (Sigma 5042930001) and Fisher Scientific). For IP, 3 mg of total protein was used for each
SPHINX31 (BioVision, B2516-5) at desired concentrations for 8 and lysate. Lysates were first incubated with 40 ll of pre-washed
6 h before the transfection, respectively. Transfections and lumines- protein G beads for 30 min at 4°C to block any non-specific inter-
cence assays were performed as described above. For luminescence action, then incubated with pre-washed 40 ll of packed M2 FLAG
assays following ISR activation, HEK293T cells were seeded and beads (Sigma) rotating at 4°C for 4 h. Afterward, beads were
transfected as described before for 19 h, followed by 5 h of treat- washed with NP-40 lysis buffer for a total of four times, 3 min each
ment with 2 lM Thapsigargin (Thermo Fisher Scientific). All drugs at 4°C. Before the last wash, 20% of the IP was taken out and
were dissolved in DMSO and stored as recommended by the manu- saved for Western blot if needed. After the final wash with lysis
facturer. buffer, beads were transferred to a new tube and finally washed
Subcellular fractionation after SRPIN340 treatment was with 1× PBS (mixing by hand) and stored at 80°C until mass
performed as described earlier (Li et al, 2015). In brief, HEK293T spectrometry.
cells were grown on 6-well plates. 24 h after SRPIN340 treatment Mass spectrometry was performed by the proteomics resource
with desired concentration, cells were gently resuspended in ice- facility at the Department of Pathology, University of Michigan. In
cold 200 ll cytoplasmic extraction buffer (10 mM HEPES pH 7.6, brief, the beads were resuspended in 50 ll of 0.1 M ammonium
60 mM KCl, 1 mM EDTA, 0.15% NP-40, 0.5 mM DTT) supple- bicarbonate buffer (pH ~8). Cysteines were reduced by adding 50 ll
mented with protease inhibitor and incubated on ice for 5 min. of 10 mM DTT and incubating at 45°C for 30 min. An overnight
Then, cells were spun at 600 g for 4 min at 4°C. The supernatant digestion with 1 lg sequencing grade, modified trypsin was carried
(cytoplasmic fraction) was then transfer to a new tube. The pellet out at 37°C with constant shaking in a Thermomixer. Samples were
was then washed twice with the wash buffer (cytoplasmic extraction completely dried using vacufuge. Resulting peptides were dissolved
buffer lacking NP-40) by spinning at 600 g for 2 min each. The in 8 ll of 0.1% formic acid/2% acetonitrile solution, and 2 ll of the
pellet was then resuspended in 200 ll nuclear extraction buffer peptide solution was resolved on a nano-capillary reverse phase
(20 mM HEPES pH 7.9, 1.5 mM MgCl2, 430 mM NaCl, 0.2 mM column (Acclaim PepMap C18, 2 micron, 50 cm, Thermo Scientific)
EDTA, 25% glycerol, and protease inhibitor) and rotated on a rotor using a 0.1% formic acid/2% acetonitrile (Buffer A) and 0.1%
for 15 min at 4°C. Finally, both cytoplasmic and nuclear fractions formic acid/95% acetonitrile (Buffer B) gradient at 300 nl/min over
were cleared by centrifugation. a period of 180 min (2–22% buffer B in 110 min, 22–40% in
25 min, 40–90% in 5 min followed by holding at 90% buffer B for
Tagged RNA capture and mass spectrometry 5 min and equilibration with Buffer A for 25 min). Eluent was
directly introduced into Orbitrap Fusion tribrid mass spectrometer
For tagged-RNA immunoprecipitation (IP), HEK293T cells were (Thermo Scientific, San Jose CA) using an EasySpray source.
grown in SILAC medium for 5–6 passages before transfection. Proteins were identified by searching the MS/MS data against H
DMEM deficient in L-arginine and L-lysine is supplemented with Sapiens (UniProt; 20,145 reviewed entries; downloaded on 08-02-
10% dialyzed FBS and isotopes of lysine (final concentration 2017) using Proteome Discoverer (v2.1, Thermo Scientific). False
146 mg/l) and arginine (final concentration 84 mg/l) for triplex discovery rate (FDR) was determined using Percolator, and
SILAC labeling. For light SILAC: DMEM, L-lysine (Lys-0) and proteins/peptides with an FDR of ≤1% were retained for further
L-arginine (Arg-0); for medium SILAC: DMEM, L-lysine-4,4,5,5-d4 analysis.

Drosophila lines, rough eye screening, external eye fluorescent complete mini protease inhibitor (Roche) and centrifuged at
measurement, and survival 13,500 g for 10 min at 4°C to pellet cuticle and wing debris. The
supernatant was removed, mixed with 6× SDS sample buffer, and
All fly lines used here and their sources are listed in boiled at 90°C for 10 min before running on SDS–PAGE.
Appendix Table S2. Total RNA for qPCR analysis was isolated as described earlier
To make the dSF2 OE fly line, Drosophila dSF2 sequence was using TRIzol (Thermo Fisher) (Linsalata et al, 2019). For qRT–PCR
PCR amplified (dSF2 F 5’-CACCATGGGATCACGCAACGAGTGCCG- analysis, 10 lg of total RNA per sample was treated with 2 U of
3’ and dSF2 R 5’-ATAGTTAGAACGTGAGCGAGACCTGG-3’) was TURBO Dnase (Thermo Fisher) at 37°C for 30 min. DNase-treated
cloned into pEntry-TOPO vector (Thermo Fisher Scientific). The RNA was purified using RNA clean and concentrator-25 kit (Zymo
pENTR-dSF2 vector was recombined with Gateway plasmid pTWH Research). 500 ng of RNA was used to synthesize cDNA using
(Drosophila Genomics Resource Center, IN). The final vector was iScript cDNA synthesis kit (Bio-Rad). qPCR assays were performed
used for site-specific transgenesis using PhiC31 integrase technique using iQ SYBR Green Supermix (Bio-Rad) and an iQ5 qPCR system
(BestGene, CA). (Bio-Rad). Drosophila a-Tubulin transcript abundance used for
Flies were crossed and raised at 25°C on SY10 food supple- normalization of target transcripts using ΔΔCT method (Livak &
mented with dry yeast unless otherwise noted. For rough eye Schmittgen, 2001). Primers used for qPCR assays are listed on
screening, 5–6 virgin female flies expressing GMR-GAL4-driven Appendix Table S3.
UAS-FMR1 (CGG)90-EGFP reporters were crossed with male flies
carrying either UAS-driven siRNA against a candidate gene or a Hybridization Chain Reaction (HCR) and
germline mutation. For GGGGCC repeat RNA toxicity modifier immunocytochemistry (ICC)
phenotyping, a GMR-GAL4-driven UAS-(GGGGCC)28-EGFP reporter
containing fly was used. Rough eye phenotypes in F1 progenies were HCR v3.0 was performed as previously described (Glineburg et al,
scored at 1–2 days post-eclosion. A minimum of 30 flies (both male 2021). Briefly, U2OS cells were seeded at 5 × 104 cells/ml in the
and females) from two independent crosses was scored. For rough chamber and transfected with SRSF1-FLAG (Addgene #99021) and/
eye scores were given based on following eye abnormalities: (i) or CGG repeat plasmids. Transfection was performed with TransIT-
abnormal orientation of the bristles, (ii) supernumerary bristles, (iii) LT1 Transfection Reagent (Mirus, MIR-2304) in the ratio of 3:1 of
ommatidial fusion and disarray, (iv) presence of necrosis, and (v) reagent to plasmid. Transfected cells were washed twice with
collapse/shrinkage of the eye. For each category, three possible 1xPBS, then fixed in 4% PFA for 10 min at room temperature (RT).
scores were given: 1 (for presence of the abnormality), 3 (if the After fixation, cells were washed with 1xPBS twice before treating
abnormality affected >5% of the total eye), and 5 (if the abnormality with Turbo DNase for 30 min at 37°C incubator. Then, cells were
affected >50% of the total eye). Eye images were captured using a dehydrated overnight in 70% ethanol at 4°C and rehydrated with 1×
Leica M125 stereomicroscope and a Leica DFC425 digital camera. PBS for 1 h, prior to immunocytochemistry (ICC). For ICC, cells
For external eye fluorescent measurement, fly crosses were were permeabilized in 0.1% Triton X-100 in 1× PBS for 6 min and
performed as described above. Fluorescent images were taken at 1– block with 2% RNAse-free acetylated BSA in 1× PBS for 20 min at
2 days post-eclosion using a Leica M125 stereomicroscope with GFP RT. Cells were stained with primary antibody FLAG-M2 (1:100 dilu-
filter. All images were taken at the same exposure. GFP images were tion, Sigma# F1804) overnight at 4°C, then followed by three times
converted to grayscale, and total intensity was measured using of 5 min 1× PBS washes before staining with secondary antibody
ImageJ. (1:500 dilution, Alexa Fluor 488, Invitrogen# A11029) for 1 h at RT
For survival assays, flies carrying desirable repeat RNA reporter in dark. Then, the cells were fixed again with 4% PFA for 10 min at
and either a Tub5-GAL4 GeneSwitch or ElaV-GAL4 GeneSwitch RT followed by washing with 1× PBS 3 times each for 1 min before
driver were crossed with a modifier fly. F1 progenies were collected proceeding to hybridization chain reaction (HCR). HCR v3.0 was
1 day post-eclosion and placed on SY10 food supplemented with performed according to manufacturer’s instructions (Molecular
200 lM RU486 and flipped onto fresh RU486-containing food every Instrument). The initiator probe for CGG repeats and fluorophore
48 h. For survival, ~20 flies (equal male and females) from at least 647-labeled hairpin probes (B1H1 and B1H2) were synthesized by
three independent crosses maintained at 29°C. Number of deaths Molecular Instruments. Repeat RNA were detected by initiator probe
recorded every 48 h until expiration and plotted using GraphPad CGG at 8 nM and amplified by fluorophore 647 labeled hairpin
Prism. probes (B1H1 and B1H2) at 60 nM. Finally, cells were stained with
DAPI and stored at 4°C in dark until imaging.
Drosophila Western blotting, RNA isolation, and quantitative Imaging was performed using an oil 60× objective in Olympus
reverse-transcription PCR (qRT–PCR) FV1000 inverted laser scanning confocal microscope. For all experi-
ments, acquisition parameters were identical between conditions
Immunoblotting and qRT–PCRs were performed as described earlier within experiments. Cells were imaged in a series of Z-planes, and
with slight modifications (Linsalata et al, 2019). In brief, 1–2 days images were analyzed in ImageJ. Average intensity composite
post-eclosion flies carrying (CGG)90-EGFP and a GeneSwitch (Tub5 images were derived from raw image files. For nuclear and cytoplas-
or ELAV) driver were placed on 200 lM RU486-supplemented SY10 mic ratio analysis, signals for each channel were normalized prior
food with fresh RU486-supplemented food provided every 24 h, at to quantification. The background signal was first normalized to
29°C. For Western samples, flies carrying (CGG)90-EGFP with ELAV non-transfection group or DMSO control. Next, the ROI was applied
driver were maintained on RU486-supplemented food for 5 days. to the DAPI channels to specify the region of nucleus along with the
Flies were homogenized at 4°C in RIPA buffer supplemented with 647 channel, which captured CGG RNA amplified by HCR. CGG

RNA signal from the nucleus was calculated by the intensity in ROI
The paper explained
from DAPI channel in pixels, while the RNA signal from the cyto-
plasm was calculated by the total intensity from 647 channel Problem
subtracted by the RNA intensity in ROI from DAPI channel in pixels. Over 50 different nucleotide repeat expansions cause neurodevelop-
Finally, percentages of CGG RNA intensity in nucleus and cytoplasm mental and neurodegenerative disorders in humans. These repeats as
RNA interact with specific RNA binding proteins to elicit toxicity
were calculated. 20–25 views of images with a number of 85–151
through protein sequestration as well as repeat-associated non-AUG-
U2OS cells were counted for each condition. P-values were calcu- initiated translation. However, accurate catalogues of which RNA
lated Student’s t-test with Welch’s correction. binding proteins bind these repeats in different cellular compartments
and how they influence their biology is unclear.
RNA synthesis and rabbit reticulocyte lysate (RRL)
in vitro translation Results
Using an innovative in-cell CGG repeat capture technique, this study
identified SR proteins and SR protein kinases (SRPKs) as key regulators
For RNA transfections and in vitro translation assays using rabbit of repeat RNA behavior, RAN translation, and repeat toxicity across
reticulocyte lysate (RRL), RNAs were in vitro transcribed from model systems. Pharmacological targeting of SRPK1 alleviated toxicity
linearized plasmids containing nanoluciferase (AUG-nLuc and and enhanced survival in CGG repeat expressing neurons by both
CGG-nLuc) and firefly (FLuc) reporters described earlier (Green suppressing repeat RNA translocation to the cytoplasm and directly
et al, 2017). pcDNA3.1(+) nLuc reporter plasmids were linearized inhibiting RAN translation.
with PspOMI, and pCRII FLuc reporter plasmid was linearized
Impact
with HindIII-HF. In vitro RNA synthesis was performed using
This study establishes SRPK1 as a potential modifier of repeat toxicity
HiScribe T7 ARCA mRNA kit (with tailing; NEB) as per the manu- and RAN translation across multiple repeats. These results provide a
facturer’s instructions. Synthesized RNAs were purified using foundation for the development of new therapeutic approaches for
RNA Clean and Concentrator-25 kits (Zymo Research), and the additional repeat expansion disorders.
integrity and size of transcribed RNAs were confirmed by gel elec-
trophoresis before using in transfection assays or in RRL transla-
tion reactions.
For in vitro translation assays, Flexi Rabbit Reticulocyte Lysate Prediction of RNA secondary structures
System (Promega) was used as described earlier (Green et al, 2017).
In brief, 3 nM of in vitro transcribed mRNAs was incubated with RNA secondary structure prediction was performed using M Fold
30% RRL, 10 mM amino acid mix minus methionine, 10 mM amino web server (Zuker, 2003).
acid mix minus leucine, 0.5 mM MgOAc, 100 mM KCl 0.8 U/µl
Murine RNAse Inhibitor (NEB), and indicated amount of SRPIN304 Statistical methods
(or DMSO for control) at 30°C for 30 min. Reactions were then
diluted 1:7 in Glo Lysis Buffer (Promega) and incubated with Statistical analysis was performed using GraphPad Prism 7. For
NanoGlo Substrate freshly diluted 1:50 in NanoGlo Buffer comparison of nLuc reporter luciferase activity, Western blots, and
(Promega). Luminescence was measured using a GloMax 96 Micro- HCR/ICC analysis, two-tailed Student’s t-tests were performed with
plate Luminometer. Welch’s correction. For all experiments, specification of statistical
analysis and sample numbers (n) is provided with figure legends. A
Primary rat neuron drug treatment, transfection, and automated minimum of three independent biological samples (n > 3) with
fluorescence microscopy imaging technical replication of results from each sample. To avoid the
effects of subjective bias, fly screening was done by at least two
Rat embryonic cortical dissections from E20 Long–Evans rat pups of independent investigators using multiple crosses. Eye necrosis/
both sexes were performed as previously described (Malik et al, width measurements were done by blind data analysis.
2018; Flores et al, 2019). Dissociated cortical neurons were plated at
0.6 × 105 cells per well on poly-D-lysine-coated 96-well plate in
neuronal growth media (NGM, neurobasal A media, 2% B-27, 1% Data availability
Glutamax-1 (v:v)), and maintained at 37°C for 4 days before trans-
fection. On DIV 4, neurons were treated with SRPIN (10–50 µM) or Raw mass spectrometry data are provided in Data Set EV1. Mass spec-
SPHINX (2–10 µM) or DMSO 8 h before transfection. Neurons were trometry data are also deposited to the ProteomeXchange Consortium
then co-transfected 0.1 lg of pGW1-mCherry and either 0.1 lg of via the PRIDE partner repository with the dataset identifier PXD027000.
pGW1-GFP or 0.1 lg of pGW +1(CGG)100 GFP DNA per well of a The datasets produced in this study are available in the following
96-well culture plate, using Lipofectamine 2000 (Thermo Fisher databases: ProteomeXchange (http://proteomecentral.proteomexcha
Scientific). nge.org/cgi/GetDataset) using the data set identifier (PXD027000).
24 h after transfection, neurons were imaged at 24-h intervals for
10 days using an automated fluorescence microscopy platform Expanded View for this article is available online.
previously described (Arrasate et al, 2004; Barmada et al, 2015).
Images were processed using a custom code written in Python and Acknowledgements
ImageJ macro language and analyzed by cox proportional hazard The authors thank Todd laboratory members for helpful input and instruction
test using the survival package in R. on Drosophila biology. We thank Amy Krans for assistance with Drosophila

generation and cloning. We thank the Pletcher laboratory at the University of selective SRPK1 inhibitors as potential topical therapeutics for neovascular
Michigan for providing fly food. We thank the Barmada and Sutton laborato- eye disease. ACS Chem Biol 12: 825 – 832
ries for assistance with microscopy analysis and access to equipment. We Bond U (2006) Stressed out! Effects of environmental stress on mRNA
acknowledge the proteomics resource facility at the Department of Pathology, metabolism. FEMS Yeast Res 6: 160 – 170
University of Michigan. This work was funded by grants from the NIH Brown MC, Dobrikov MI, Gromeier M (2014) Mitogen-activated protein
(P50HD104463, R01NS099280, and R01NS086810 to PKT, NRSA F31NS113513 kinase-interacting kinase regulates mTOR/AKT signaling and controls the
to SEW, and NRSA F31NS100302 to KMG) and the VA (BLRD BX004842 to PKT) serine/arginine-rich protein kinase-responsive type 1 internal ribosome
and by private philanthropic support to PKT. IM was supported by an Alzhei- entry site-mediated translation and viral oncolysis. J Virol 88:
mer’s Association Research Fellowship. Y-JT was supported by the Cellular and 13149 – 13160
Molecular Biology Graduate program, University of Michigan. Buijsen RAM, Sellier C, Severijnen L-AWFM, Oulad-Abdelghani M, Verhagen
RFM, Berman RF, Charlet-Berguerand N, Willemsen R, Hukema RK (2014)
Author contributions FMRpolyG-positive inclusions in CNS and non-CNS organs of a fragile X
IM and PKT conceptualized the experiments. Indranil performed the formal premutation carrier with fragile X-associated tremor/ataxia syndrome.
analysis for most of the described assays and was responsible for developing the Acta Neuropathol Commun 2: 162
methodology along with PKT. Y-JT performed and analyzed the experiments Chandra A, Ananda H, Singh N, Qamar I (2020) Identification of a novel and
relevant to HCR and RNA–protein co-localization. SEW performed the studies potent small molecule inhibitor of SRPK1: mechanism of dual inhibition of
and analyzed the experiments relevant to neuronal survival. PR, KZ, and KMG SRPK1 for the inhibition of cancer progression. Aging (Albany NY) 12:
performed drosophila experiments and assisted with design and analysis of 163 – 180
these studies. IM wrote the initial draft of the manuscript along with PKT, and Chao JA, Patskovsky Y, Almo SC, Singer RH (2008) Structural basis for the
all authors reviewed and edited the manuscript. IM, SEW, KMG, and PKT were coevolution of a viral RNA–protein complex. Nat Struct Mol Biol 15: 103 – 105
responsible for obtaining funding for the work. PKT oversaw the project. Cheng W, Wang S, Mestre AA, Fu C, Makarem A, Xian F, Hayes LR, Lopez-
Gonzalez R, Drenner K, Jiang J et al (2018) C9ORF72 GGGGCC repeat-
Conflict of interest associated non-AUG translation is upregulated by stress through eIF2a
The authors declare that they have no conflict of interest. phosphorylation. Nat Commun 9: 51
Cheng W, Wang S, Zhang Z, Morgens DW, Hayes LR, Lee S, Portz B, Xie Y,
Nguyen BV, Haney MS et al (2019) CRISPR-Cas9 screens identify the RNA
References helicase DDX3X as a repressor of C9ORF72 (GGGGCC)n repeat-associated
non-AUG translation. Neuron 104: 885 – 898
Amin E, Oltean S, Hua J, Gammons M, Hamdollah-Zadeh M, Welsh G, Cheung Choi J-H, Kim S-H, Jeong Y-H, Kim SW, Min K-T, Kim K-T (2019) hnRNP Q
M-K, Ni L, Kase S, Rennel E et al (2011) WT1 mutants reveal SRPK1 to be regulates internal ribosome entry site-mediated fmr1 translation in
a downstream angiogenesis target by altering VEGF splicing. Cancer Cell neurons. Mol Cell Biol 39: e00371-18
20: 768 – 780 Ciesiolka A, Jazurek M, Drazkowska K, Krzyzosiak WJ (2017) Structural
Anczukow O, Akerman M, Clery A, Wu J, Shen C, Shirole N, Raimer A, Sun S, characteristics of simple RNA repeats associated with disease and their
Jensen M, Hua Y et al (2015) SRSF1-regulated alternative splicing in breast deleterious protein interactions. Front Cell Neurosci 11: 97
cancer. Mol Cell 60: 105 – 117 Conlon EG, Lu L, Sharma A, Yamazaki T, Tang T, Shneider NA, Manley JL
Ariza J, Rogers H, Monterrubio A, Reyes-Miranda A, Hagerman PJ, Martınez- (2016) The C9ORF72 GGGGCC expansion forms RNA G-quadruplex
Cerden ~o V (2016) A majority of FXTAS cases present with intranuclear inclusions and sequesters hnRNP H to disrupt splicing in ALS brains. eLife
inclusions within Purkinje cells. Cerebellum 15: 546 – 551 5: e17820
Arrasate M, Mitra S, Schweitzer ES, Segal MR, Finkbeiner S (2004) Inclusion Cooper-Knock J, Walsh MJ, Higginbottom A, Robin Highley J, Dickman MJ,
body formation reduces levels of mutant huntingtin and the risk of Edbauer D, Ince PG, Wharton SB, Wilson SA, Kirby J et al (2014)
neuronal death. Nature 431: 805 – 810 Sequestration of multiple RNA recognition motif-containing proteins by
Ash P, Bieniek K, Gendron T, Caulfield T, Lin W-L, DeJesus-Hernandez M, C9orf72 repeat expansions. Brain 137: 2040 – 2051
van Blitterswijk M, Jansen-West K, Paul J, Rademakers R et al (2013) Cooper-Knock J, Higginbottom A, Stopford MJ, Highley JR, Ince PG, Wharton
Unconventional translation of C9ORF72 GGGGCC expansion generates SB, Pickering-Brown S, Kirby J, Hautbergue GM, Shaw PJ (2015) Antisense
insoluble polypeptides specific to c9FTD/ALS. Neuron 77: 639 – 646 RNA foci in the motor neurons of C9ORF72-ALS patients are associated
Ayhan F, Perez BA, Shorrock HK, Zu T, Banez-Coronel M, Reid T, Furuya H, with TDP-43 proteinopathy. Acta Neuropathol 130: 63 – 75
Clark HB, Troncoso JC, Ross CA et al (2018) SCA8 RAN polySer protein Coulon A, Ferguson ML, de Turris V, Palangat M, Chow CC, Larson DR (2014)
preferentially accumulates in white matter regions and is regulated by Kinetic competition during the transcription cycle results in stochastic
eIF3F. EMBO J 37: e99023 RNA processing. eLife 3: e03939
~ez-Coronel M, Ayhan F, Tarabochia A, Zu T, Perez B, Tusi S, Pletnikova O,
Ban Dennis G, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, Lempicki RA
Borchelt D, Ross C, Margolis R et al (2015) RAN translation in Huntington (2003) DAVID: Database for Annotation, Visualization, and Integrated
disease. Neuron 88: 667 – 677 Discovery. Genome Biol 4: R60
Barmada SJ, Ju S, Arjun A, Batarse A, Archbold HC, Peisach D, Li X, Zhang Y, Donnelly C, Zhang P-W, Pham J, Haeusler A, Mistry N, Vidensky S, Daley E, Poth
Tank EMH, Qiu H et al (2015) Amelioration of toxicity in neuronal models E, Hoover B, Fines D et al (2013) RNA toxicity from the ALS/FTD C9ORF72
of amyotrophic lateral sclerosis by hUPF1. Proc Natl Acad Sci USA 112: expansion is mitigated by antisense intervention. Neuron 80: 415 – 428
7821 – 7826 Fay MM, Anderson PJ, Ivanov P (2017) ALS/FTD-Associated C9ORF72 repeat
Batson J, Toop HD, Redondo C, Babaei-Jadidi R, Chaikuad A, Wearmouth SF, RNA promotes phase transitions in vitro and in cells. Cell Rep 21:
Gibbons B, Allen C, Tallant C, Zhang J et al (2017) Development of potent, 3573 – 3584

Flores BN, Li X, Malik AM, Martinez J, Beg AA, Barmada SJ (2019) An Hu S-I, Katz M, Chin S, Qi X, Cruz J, Ibebunjo C, Zhao S, Chen A, Glass DJ
intramolecular salt bridge linking TDP43 RNA binding, protein stability, (2012) MNK2 Inhibits eIF4G activation through a pathway involving
and TDP43-dependent neurodegeneration. Cell Rep 27: 1133 – 1150 serine-arginine–rich protein kinase in skeletal muscle. Sci Signal 5: ra14
Fukuhara T, Hosoya T, Shimizu S, Sumi K, Oshiro T, Yoshinaka Y, Suzuki M, Ishiguro T, Sato N, Ueyama M, Fujikake N, Sellier C, Kanegami A, Tokuda E,
Yamamoto N, Herzenberg LA, Herzenberg LA et al (2006) Utilization of Zamiri B, Gall-Duncan T, Mirceta M et al (2017) Regulatory role of RNA
host SR protein kinases and RNA-splicing machinery during viral chaperone TDP-43 for RNA misfolding and repeat-associated translation in
replication. Proc Natl Acad Sci USA 103: 11329 – 11333 SCA31. Neuron 94: 108 – 124
Gammons MV, Fedorov O, Ivison D, Du C, Clark T, Hopkins C, Hagiwara M, Jacquemont S, Hagerman RJ, Leehey M, Grigsby J, Zhang L, Brunberg JA, Greco
Dick AD, Cox R, Harper SJ et al (2013) Topical antiangiogenic SRPK1 C, Des Portes V, Jardini T, Levine R et al (2003) Fragile X premutation
inhibitors reduce choroidal neovascularization in rodent models of tremor/ataxia syndrome: molecular, clinical, and neuroimaging correlates.
exudative AMD. Invest Ophthalmol Vis Sci 54: 6052 – 6062 Am J Hum Genet 72: 869 – 878
Gendron TF, Bieniek KF, Zhang Y-J, Jansen-West K, Ash PEA, Caulfield T, Jacquemont S, Hagerman RJ, Leehey MA, Hall DA, Levine RA, Brunberg JA,
Daughrity L, Dunmore JH, Castanedes-Casey M, Chew J et al (2013) Zhang L, Jardini T, Gane LW, Harris SW et al (2004) Penetrance of the
Antisense transcripts of the expanded C9ORF72 hexanucleotide repeat fragile X-associated tremor/ataxia syndrome in a premutation carrier
form nuclear RNA foci and undergo repeat-associated non-ATG translation population. JAMA 291: 460 – 469
in c9FTD/ALS. Acta Neuropathol 126: 829 – 844 Jain A, Vale RD (2017) RNA phase transitions in repeat expansion disorders.
Glineburg MR, Todd PK, Charlet-Berguerand N, Sellier C (2018) Repeat- Nature 546: 243 – 247
associated non-AUG (RAN) translation and other molecular mechanisms in Jazurek M, Ciesiolka A, Starega-Roslan J, Bilinska K, Krzyzosiak WJ (2016)
Fragile X Tremor Ataxia Syndrome. Brain Res 1693: 43 – 54 Identifying proteins that bind to specific RNAs - focus on simple repeat
Glineburg MR, Zhang Y, Krans A, Tank EM, Barmada SJ, Todd PK (2021) expansion diseases. Nucleic Acids Res 44: 9050 – 9070
Enhanced detection of expanded repeat mRNA foci with hybridization Jeong S (2017) SR proteins: binders, regulators, and connectors of RNA. Mol
chain reaction. Acta Neuropathol Commun 9: 73 Cells 40: 1 – 9
Gonçalves V, Henriques AFA, Henriques A, Pereira JFS, Pereira J, Neves Costa Jin P, Duan R, Qurashi A, Qin Y, Tian D, Rosser TC, Liu H, Feng Y, Warren ST
A, Moyer MP, Moita LF, Gama-Carvalho M, Matos P et al (2014) (2007) Pur a binds to rCGG repeats and modulates repeat-mediated
Phosphorylation of SRSF1 by SRPK1 regulates alternative splicing of neurodegeneration in a Drosophila model of Fragile X Tremor/Ataxia
tumor-related Rac1b in colorectal cells. RNA 20: 474 – 482 Syndrome. Neuron 55: 556 – 564
Greco CM, Berman RF, Martin RM, Tassone F, Schwartz PH, Chang A, Trapp Jovicic A, Mertens J, Boeynaems S, Bogaert E, Chai N, Yamada SB, Paul JW,
BD, Iwahashi C, Brunberg J, Grigsby J et al (2006) Neuropathology of Sun S, Herdy JR, Bieri G et al (2015) Modifiers of C9orf72 dipeptide repeat
fragile X-associated tremor/ataxia syndrome (FXTAS). Brain 129: 243 – 255 toxicity connect nucleocytoplasmic transport defects to FTD/ALS. Nat
Greco CM, Hagerman RJ, Tassone F, Chudley AE, Del Bigio MR, Jacquemont S, Neurosci 18: 1226 – 1229
Leehey M, Hagerman PJ (2002) Neuronal intranuclear inclusions in a new Kanadia RN, Johnstone KA, Mankodi A, Lungu C, Thornton CA, Esson D,
cerebellar tremor/ataxia syndrome among fragile X carriers. Brain 125: Timmers AM, Hauswirth WW, Swanson MS (2003) A muscleblind knockout
1760 – 1771 model for myotonic dystrophy. Science 302: 1978 – 1980
Green KM, Glineburg MR, Kearse MG, Flores BN, Linsalata AE, Fedak SJ, Kanadia RN, Shin J, Yuan Y, Beattie SG, Wheeler TM, Thornton CA, Swanson
Goldstrohm AC, Barmada SJ, Todd PK (2017) RAN translation at C9orf72- MS (2006) Reversal of RNA missplicing and myotonia after muscleblind
associated repeat expansions is selectively enhanced by the integrated overexpression in a mouse poly(CUG) model for myotonic dystrophy. Proc
stress response. Nat Commun 8: 2005 Natl Acad Sci USA 103: 11748 – 11753
Hagerman RJ, Leehey M, Heinrichs W, Tassone F, Wilson R, Hills J, Grigsby J, Kearse MG, Green KM, Krans A, Rodriguez CM, Linsalata AE, Goldstrohm AC,
Gage B, Hagerman PJ (2001) Intention tremor, Parkinsonism, and generalized Todd PK (2016) CGG repeat-associated non-AUG translation utilizes a cap-
brain atrophy in male carriers of fragile X. Neurology 57: 127 – 130 dependent scanning mechanism of initiation to produce toxic proteins.
Hagerman PJ, Hagerman RJ (2015) Fragile X–associated tremor/ataxia Mol Cell 62: 314 – 322
syndrome. Ann N Y Acad Sci 1338: 58 – 70 Krans A, Kearse MG, Todd PK (2016) Repeat-associated non-AUG translation
Harlen KM, Churchman LS (2017) Subgenic Pol II interactomes identify from antisense CCG repeats in fragile X tremor/ataxia syndrome. Ann
region-specific transcription elongation regulators. Mol Syst Biol 13: 900 Neurol 80: 871 – 881
Hautbergue GM, Castelli LM, Ferraiuolo L, Sanchez-Martinez A, Cooper-Knock Krzyzosiak WJ, Sobczak K, Wojciechowska M, Fiszer A, Mykowska A, Kozlowski
J, Higginbottom A, Lin Y-H, Bauer CS, Dodd JE, Myszczynska MA et al P (2012) Triplet repeat RNA structure and its role as pathogenic agent and
(2017) SRSF1-dependent nuclear export inhibition of C9ORF72 repeat therapeutic target. Nucleic Acids Res 40: 11 – 26
transcripts prevents neurodegeneration and associated motor deficits. Nat Lee Y-B, Chen H-J, Peres JN, Gomez-Deza J, Attig J, Stalekar M, Troakes C,
Commun 8: 16063 Nishimura A, Scotter E, Vance C et al (2013) Hexanucleotide repeats in
He F, Flores BN, Krans A, Frazer M, Natla S, Niraula S, Adefioye O, Barmada ALS/FTD form length-dependent RNA foci, sequester RNA binding proteins,
SJ, Todd PK (2020) The carboxyl termini of RAN translated GGGGCC and are neurotoxic. Cell Rep 5: 1178 – 1186
nucleotide repeat expansions modulate toxicity in models of ALS/FTD. Acta Lee K-H, Zhang P, Kim HJ, Mitrea DM, Sarkar M, Freibaum BD, Cika J,
Neuropathol Commun 8: 122 Coughlin M, Messing J, Molliex A et al (2016) C9orf72 dipeptide repeats
Hinnebusch AG (2006) eIF3: a versatile scaffold for translation initiation impair the assembly, dynamics, and function of membrane-less organelles.
complexes. Trends Biochem Sci 31: 553 – 562 Cell 167: 774 – 788
Hong Y, Chan CB, Kwon I-S, Li X, Song M, Lee H-P, Liu X, Sompol P, Jin P, Lee Li Y, Collins M, Geiser R, Bakkar N, Riascos D, Bowser R (2015) RBM45 homo-
H-G et al (2012) SRPK2 phosphorylates tau and mediates the cognitive oligomerization mediates association with ALS-linked proteins and stress
defects in Alzheimer’s disease. J Neurosci 32: 17262 – 17272 granules. Sci Rep 5: 14262

Linsalata AE, He F, Malik AM, Glineburg MR, Green KM, Natla S, Flores BN, for RAN translation and CGG repeats in regulating fragile X protein
Krans A, Archbold HC, Fedak SJ et al (2019) DDX3X and specific initiation synthesis. Nat Neurosci 23: 386 – 397
factors modulate FMR1 repeat-associated non-AUG-initiated translation. van Roosmalen W, Le Devedec SE, Golani O, Smid M, Pulyakhina I,
EMBO Rep 20: e47498 Timmermans AM, Look MP, Zi DI, Pont C, de Graauw M et al (2015)
Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using Tumor cell migration screen identifies SRPK1 as breast cancer metastasis
real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods determinant. J Clin Invest 125: 1648 – 1664
25: 402 – 408 Sato N, Amino T, Kobayashi K, Asakawa S, Ishiguro T, Tsunemi T, Takahashi
Malik AM, Miguez RA, Li X, Ho Y-S, Feldman EL, Barmada SJ (2018) Matrin 3- M, Matsuura T, Flanigan KM, Iwasaki S et al (2009) Spinocerebellar ataxia
dependent neurotoxicity is modified by nucleic acid binding and type 31 is associated with ‘inserted’ penta-nucleotide repeats containing
nucleocytoplasmic localization. eLife 7: e35977 (TGGAA)n. Am J Hum Genet 85: 544 – 557
Mankodi A, Logigian E, Callahan L, McClain C, White R, Henderson D, Krym Sellier C, Rau F, Liu Y, Tassone F, Hukema RK, Gattoni R, Schneider A, Richard
M, Thornton CA (2000) Myotonic dystrophy in transgenic mice expressing S, Willemsen R, Elliott DJ et al (2010) Sam68 sequestration and partial
an expanded CUG repeat. Science 289: 1769 – 1773 loss of function are associated with splicing alterations in FXTAS patients.
Matheny T, Treeck BV, Huynh TN, Parker R (2020) RNA partitioning into EMBO J 29: 1248 – 1261
stress granules is based on the summation of multiple interactions. RNA Sellier C, Freyermuth F, Tabet R, Tran T, He F, Ruffenach F, Alunni V, Moine
27: 174 – 189 H, Thibault C, Page A et al (2013) Sequestration of DROSHA and DGCR8 by
Mavrou A, Brakspear K, Hamdollah-Zadeh M, Damodaran G, Babaei-Jadidi expanded CGG RNA repeats alters microRNA processing in fragile X-
R, Oxley J, Gillatt DA, Ladomery MR, Harper SJ, Bates DO et al (2015) associated tremor/ataxia syndrome. Cell Rep 3: 869 – 880
Serine-arginine protein kinase 1 (SRPK1) inhibition as a potential novel Sellier C, Buijsen RAM, He F, Natla S, Jung L, Tropel P, Gaucherot A, Jacobs H,
targeted therapeutic strategy in prostate cancer. Oncogene 34: Meziane H, Vincent A et al (2017) Translation of expanded CGG repeats
4311 – 4319 into FMRpolyG is pathogenic and may contribute to fragile X tremor
Mavrou A, Oltean S (2016) SRPK1 inhibition in prostate cancer: a novel anti- ataxia syndrome. Neuron 93: 331 – 347
angiogenic treatment through modulation of VEGF alternative splicing. Sheng J, Zhao Q, Zhao J, Zhang W, Sun YU, Qin P, Lv Y, Bai LU, Yang Q, Chen
Pharmacol Res 107: 276 – 281 L et al (2018) SRSF1 modulates PTPMT1 alternative splicing to regulate
May S, Hornburg D, Schludi MH, Arzberger T, Rentzsch K, Schwenk BM, lung cancer cell radioresistance. EBioMedicine 38: 113 – 126
€sser FA, Mori K, Kremmer E, Banzhaf-Strathmann J et al (2014) C9orf72
Gra Sofola OA, Jin P, Qin Y, Duan R, Liu H, de Haro M, Nelson DL, Botas J (2007)
FTLD/ALS-associated Gly-Ala dipeptide repeat proteins cause neuronal RNA-binding proteins hnRNP A2/B1 and CUGBP1 suppress fragile X CGG
toxicity and Unc119 sequestration. Acta Neuropathol 128: 485 – 503 premutation repeat-induced neurodegeneration in a Drosophila model of
Miller JW, Urbinati CR, Teng-Umnuay P, Stenberg MG, Byrne BJ, Thornton CA, FXTAS. Neuron 55: 565 – 571
Swanson MS (2000) Recruitment of human muscleblind proteins to (CUG) Sonobe Y, Ghadge G, Masaki K, Sendoel A, Fuchs E, Roos RP (2018)
(n) expansions associated with myotonic dystrophy. EMBO J 19: Translation of dipeptide repeat proteins from the C9ORF72
4439 – 4448 expanded repeat is associated with cellular stress. Neurobiol Dis 116:
Mizielinska S, Gronke S, Niccoli T, Ridler CE, Clayton EL, Devoy A, Moens T, 155 – 165
Norona FE, Woollacott IOC, Pietrzyk J et al (2014) C9orf72 repeat Soragni E, Petrosyan L, Rinkoski TA, Wieben ED, Baratz KH, Fautsch MP,
expansions cause neurodegeneration in Drosophila through arginine-rich Gottesfeld JM (2018) Repeat-associated non-ATG (RAN) translation in
proteins. Science 345: 1192 – 1194 Fuchs’ endothelial corneal dystrophy. Invest Ophthalmol Vis Sci 59:
Mori K, Weng S-M, Arzberger T, May S, Rentzsch K, Kremmer E, Schmid B, 1888 – 1896
Kretzschmar HA, Cruts M, Van Broeckhoven C et al (2013) The C9orf72 Taneja KL, McCurrach M, Schalling M, Housman D, Singer RH (1995) Foci of
GGGGCC repeat is translated into aggregating dipeptide-repeat proteins in trinucleotide repeat transcripts in nuclei of myotonic dystrophy cells and
FTLD/ALS. Science 339: 1335 – 1338 tissues. J Cell Biol 128: 995 – 1002
Nakagawa O, Arnold M, Nakagawa M, Hamada H, Shelton JM, Kusano H, Tassone F, Iong KA, Tong T-H, Lo J, Gane LW, Berry-Kravis E, Nguyen D, Mu
Harris TM, Childs G, Campbell KP, Richardson JA et al (2005) LY, Laffin J, Bailey DB et al (2012) FMR1 CGG allele size and prevalence
Centronuclear myopathy in mice lacking a novel muscle-specific protein ascertained through newborn screening in the United States. Genome Med
kinase transcriptionally regulated by MEF2. Genes Dev 19: 2066 – 2077 4: 100
Ong S-E, Blagoev B, Kratchmarova I, Kristensen DB, Steen H, Pandey A, Mann Todd PK, Oh SY, Krans A, Pandey UB, Di Prospero NA, Min K-T, Taylor JP,
M (2002) Stable isotope labeling by amino acids in cell culture, SILAC, as a Paulson HL (2010) Histone deacetylases suppress CGG repeat-induced
simple and accurate approach to expression proteomics. Mol Cell neurodegeneration via transcriptional silencing in models of fragile X
Proteomics 1: 376 – 386 tremor ataxia syndrome. PLoS Genet 6: e1001240
Prudencio M, Belzil VV, Batra R, Ross CA, Gendron TF, Pregent LJ, Murray ME, Todd P, Oh S, Krans A, He F, Sellier C, Frazer M, Renoux A, Chen K-C,
Overstreet KK, Piazza-Johnston AE, Desaro P et al (2015) Distinct brain Scaglione K, Basrur V et al (2013) CGG repeat-associated translation
transcriptome profiles in C9orf72- associated and sporadic ALS. Nat mediates neurodegeneration in fragile X tremor ataxia syndrome. Neuron
Neurosci 18: 1175 – 1182 78: 440 – 455
Rizzotto D, Zaccara S, Rossi A, Galbraith MD, Andrysik Z, Pandey A, Sullivan Van Treeck B, Parker R (2018) Emerging roles for intermolecular RNA-RNA
KD, Quattrone A, Espinosa JM, Dassi E et al (2020) Nutlin-induced interactions in RNP assemblies. Cell 174: 791 – 802
apoptosis is specified by a translation program regulated by PCBP2 and Wang HY, Lin W, Dyck JA, Yeakley JM, Songyang Z, Cantley LC, Fu XD (1998)
DHX30. Cell Rep 30: 4355 – 4369 SRPK2: a differentially expressed SR protein-specific kinase involved in
Rodriguez CM, Wright SE, Kearse MG, Haenfler JM, Flores BN, Liu YU, Ifrim mediating the interaction and localization of pre-mRNA splicing factors in
MF, Glineburg MR, Krans A, Jafar-Nejad P et al (2020) A native function mammalian cells. J Cell Biol 140: 737 – 750

Wang ET, Treacy D, Eichinger K, Struck A, Estabrook J, Olafson H, Wang TT, Zhang Y-J, Guo L, Gonzales PK, Gendron TF, Wu Y, Jansen-West K, O’Raw AD,
Bhatt K, Westbrook T, Sedehizadeh S et al (2019) Transcriptome Pickles SR, Prudencio M, Carlomagno Y et al (2019) Heterochromatin
alterations in myotonic dystrophy skeletal muscle and heart. Hum Mol anomalies and double-stranded RNA accumulation underlie C9orf72 poly
Genet 28: 1312 – 1321 (PR) toxicity. Science 363: eaav2606
Wen X, Tan W, Westergard T, Krishnamurthy K, Markandaiah SS, Shi Y, Lin S, Zhou Z, Fu X-D (2013) Regulation of splicing by SR proteins and SR protein-
Shneider NA, Monaghan J, Pandey UB et al (2014) Antisense proline- specific kinases. Chromosoma 122: 191 – 207
arginine RAN dipeptides linked to C9ORF72-ALS/FTD form toxic nuclear Zu T, Gibbens B, Doty NS, Gomes-Pereira M, Huguet A, Stone MD, Margolis J,
aggregates that initiate in vitro and in vivo neuronal death. Neuron 84: Peterson M, Markowski TW, Ingram MAC et al (2011) Non-ATG-initiated
1213 – 1225 translation directed by microsatellite expansions. Proc Natl Acad Sci USA
Westergard T, McAvoy K, Russell K, Wen X, Pang Y, Morris B, Pasinelli P, Trotti 108: 260 – 265
D, Haeusler A (2019) Repeat-associated non-AUG translation in C9orf72- Zu T, Liu Y, Banez-Coronel M, Reid T, Pletnikova O, Lewis J, Miller TM, Harms
ALS/FTD is driven by neuronal excitation and stress. EMBO Mol Med 11: MB, Falchook AE, Subramony SH et al (2013) RAN proteins and RNA foci
e9423 from antisense transcripts in C9ORF72 ALS and frontotemporal dementia.
Yamada SB, Gendron TF, Niccoli T, Genuth NR, Grosely R, Shi Y, Glaria I, Proc Natl Acad Sci USA 110: E4968 – E4977
Kramer NJ, Nakayama L, Fang S et al (2019) RPS25 is required ~ez-Coronel M, Bubenik JL, Ayhan F, Ashizawa T, Xia
Zu T, Cleary JD, Liu Y, Ban
for efficient RAN translation of C9orf72 and other G, Clark HB, Yachnis AT et al (2017) RAN translation regulated by
neurodegenerative disease-associated nucleotide repeats. Nat Neurosci muscleblind proteins in myotonic dystrophy type 2. Neuron 95: 1292 – 1305
22: 1383 – 1388 Zuker M (2003) Mfold web server for nucleic acid folding and hybridization
Yan X, Hoek TA, Vale RD, Tanenbaum ME (2016) Dynamics of translation of prediction. Nucleic Acids Res 31: 3406 – 3415
single mRNA molecules in vivo. Cell 165: 976 – 989
Zhang Y-J, Gendron TF, Ebbert MTW, O’Raw AD, Yue M, Jansen-West K, License: This is an open access article under the
Zhang XU, Prudencio M, Chew J, Cook CN et al (2018) Poly(GR) impairs terms of the Creative Commons Attribution License,
protein translation and stress granule dynamics in C9orf72-associated which permits use, distribution and reproduction in
frontotemporal dementia and amyotrophic lateral sclerosis. Nat Med 24: any medium, provided the original work is properly
1136 – 1142 cited.

Malik Et Al - 2021

Uploaded by

Copyright:

Available Formats

Malik Et Al - 2021

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Malik Et Al - 2021

Uploaded by

Copyright:

Available Formats

Article

SRSF protein kinase 1 modulates RAN translation

1 Department of Neurology, University of Michigan, Ann Arbor, MI, USA

2 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 3 of 21

Figure 2. HCR-ICC detects co-localization of SRSF1 and CGG repeat RNA.

4 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 5 of 21

6 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 7 of 21

8 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 9 of 21

◀ Figure 5. SRPK1 inhibitors selectively suppress RAN translation.

Inhibition of SRPK1 enhances survival of (CGG)100

10 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

Figure 6. SRPK1 inhibition prevents stress-induced enhancement of RAN translation.

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 11 of 21

12 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 13 of 21

14 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 15 of 21

16 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 17 of 21

18 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 19 of 21

20 of 21 EMBO Molecular Medicine 13: e14163 | 2021 ª 2021 The Authors

ª 2021 The Authors EMBO Molecular Medicine 13: e14163 | 2021 21 of 21

You might also like