Proteome-wide Identification of Poly(ADP-Ribosyl)ation Targets in Different Genotoxic Stress Responses

Please cite this article in press as: Jungmichel et al., Proteome-wide Identification of Poly(ADP-Ribosyl)ation Targets in Different Genotoxic Stress Re-sponses, Molecular Cell (2013), http://dx.doi.org/10.1016/j.molcel.2013.08.026

Molecular Cell

Resource

Proteome-wide Identificationof Poly(ADP-Ribosyl)ation Targetsin Different Genotoxic Stress ResponsesStephanie Jungmichel,1 Florian Rosenthal,3,4 Matthias Altmeyer,2 Jiri Lukas,2 Michael O. Hottiger,3

and Michael L. Nielsen1,*1Department of Proteomics2Chromosome Stability and Dynamics Group, Department of Disease BiologyThe Novo Nordisk Foundation Center for Protein Research, University of Copenhagen, 2200 Copenhagen, Denmark3Institute of Veterinary Biochemistry and Molecular Biology4Life Science Zurich Graduate School

University of Zurich, 8057 Zurich, Switzerland*Correspondence: [email protected]

http://dx.doi.org/10.1016/j.molcel.2013.08.026

SUMMARY

Poly(ADP-ribos)ylation (PARylation) is a reversibleposttranslational modification found in higher eu-karyotes. However, little is known about PARylationacceptor proteins. Here, we describe a sensitiveproteomics approach based on high-accuracy quan-titative mass spectrometry for the identification ofPARylated proteins induced under different cellularstress conditions. While confirming the majority ofknown PARylated substrates, our screen identifiesnumerous additional PARylation targets. In vivoand in vitro validation of acceptor proteins confirmsthat our methodology targets covalent PARylation.Nuclear proteins encompassing nucleic acid bindingproperties are prominently PARylated upon geno-toxic stress, consistent with the nuclear localizationof ARTD1/PARP1 and ARTD2/PARP2. Distinct differ-ences in proteins becoming PARylated upon variousgenotoxic insults are observed, exemplified by thePARylation of RNA-processing factors THRAP3 andTAF15 under oxidative stress. High-content imagingreveals that PARylation affects the nuclear relocali-zation of THRAP3 and TAF15, demonstrating thepotential of our approach to uncover hitherto unap-preciated processes being controlled by specificgenotoxic-stress-induced PARylation.

INTRODUCTION

Mammalian cells are constantly exposed to genotoxic stress

and, therefore, have developed sophisticated mechanisms for

detecting and signaling the presence of damaged DNA to

accomplish efficient DNA repair processes. One of the earliest

cellular responses following exposure to genotoxic stress hap-

pens through the reversible posttranslational modification

(PTM) poly(ADP-ribosyl)ation (PARylation) (Lukas et al., 2011).

The activation of poly(ADP-ribose) (PAR) polymerases (PARPs),

now also referred to as ADP-ribosyltransferases with diphteria

toxin homology (ARTDs), entails the rapid synthesis of long,

branched PAR chains from nicotinamide adenine dinucleotide

(NAD+) that can lead to a transient 10- to 500-fold increase of

cellular PAR levels (Hassa et al., 2006). PAR polymers are sug-

gested to play a key role in the regulation of chromatin structure

modulation, DNA repair, transcription, and cell death (Luo and

Kraus, 2012). The importance of PAR is emphasized by the

fact that knockout mice for Artd1/Parp1 or Artd2/Parp2 are

hypersensitive to DNA-damaging agents and show increased

genomic instability after genotoxic stress (Hassa et al., 2006).

Although PAR formation was identified 50 years ago, surprisingly

little is known about the molecular targets of PARylation and

which processes these specifically regulate. Presently, only a

limited number of in vivo PARylated proteins (PARP substrates)

are reported in the literature and public databases, primarily

because of the absence of ‘‘unbiased‘‘ technologies for detect-

ing PARylated proteins on a global scale.

Mass spectrometry (MS)-based proteomics has emerged as a

key technology for the comprehensive identification of PTM

substrates and site-specific mapping of various types of PTMs.

Here, we describe a quantitative proteomics approach for

assessing the extent of PARylated proteins under various types

of genotoxic stress by performing pull-downs with a PAR binding

domain on stable isotope labeling by amino acids in cell culture

(SILAC)-encoded mammalian cells (Ong et al., 2002).

We decided to employ the wild-type Af1521 macrodomain

(Af1521_wt) as a PAR-binding module in our experimental setup

because of its strong affinity towards ADP-ribose (Kd = 0.13 mM)

and the availability of a characterized binding-deficient mutant

that completely abolishes PAR binding (Dani et al., 2009; Karras

et al., 2005). In brief, the binding-defective mutant (Af1521_mut)

is used for pull-down in ‘‘light’’ SILAC lysates, and Af1521_wt is

used in both ‘‘medium’’ and ‘‘heavy’’ SILAC lysates (Figure 1A).

Protein eluates from each pull-down are combined and digested

to peptides, and PARylated proteins are identified by liquid

chromatography tandem MS (LC-MS/MS). This enables the

identification of covalently PARylated proteins by a quantitative

comparison of light andmedium SILAC states while concurrently

Molecular Cell 52, 1–14, October 24, 2013 ª2013 Elsevier Inc. 1

mailto:[email protected]


(legend on next page)

Molecular Cell

Proteome-wide Identification of PARylated proteins

2 Molecular Cell 52, 1–14, October 24, 2013 ª2013 Elsevier Inc.


Molecular Cell



allowing for the investigation of stress-dependent modulations

of protein PARylation in the heavy SILAC state (Figures 1A and

1B).

Using the established approach, we identify a large number of

PARylated proteins induced under various types of genotoxic

stress and confirm several PARylated protein targets through

biochemical in vitro and in vivo assays. Moreover, the presented

method identifies the majority of already-known targets involved

in DNA repair, demonstrating the ability to directly recognize

PARylated substrates. A comparison of the different types of

genotoxic stress revealed that oxidative and alkylation stress

induces PARylation of a large number of proteins involved in

RNA metabolism, thereby positioning PARylation as an impor-

tant functional link between DNA and RNAmetabolic processes.

To emphasize this connection, we find the pre-mRNA-splicing

protein THRAP3 to be PARylated under oxidative damage.

THRAP3 has previously been reported to play a role in the

DNA damage response (DDR) in a manner that parallels tran-

scription inhibition (Beli et al., 2012). Combining our high-resolu-

tion MS data with automated high-content microscopy, we

demonstrate that PARylation of THRAP3 affects its cellular local-

ization and is required for colocalizing with splicing factors in

nuclear speckles, an effect that is enhanced upon transcriptional

inhibition. Similarly, we show that the PARylation of RNA-binding

protein TAF15 partially prevents the protein from forming

nucleolar caps under transcriptional inhibition. Collectively, the

presented data set provides a valuable resource of PARylated

proteins in human cells and describes how PARylation affects

cellular localization of RNA-associated proteins under specific

genotoxic stress conditions.

RESULTS

Establishing a Proteomic Approach for the Identificationof Genotoxic-Stress-Dependent Protein PARylationFor the affinity enrichment of PARylated proteins, we purified

Af1521_wt and Af1521_mut macrodomains as GST fusion pro-

teins and verified their binding properties towards PARylated

substrates by western blot (WB) (Figures S1A and S1B available

online). Recently, Af1521was also found to be amono(ADP-ribo-

syl) hydrolase (Jankevicius et al., 2013; Rosenthal et al., 2013).

However, no activity was reported against PARylated proteins,

which are the primary targets of this study. Nevertheless, we vali-

dated that no hydrolyase activity was exerted by Af1521 against

Figure 1. Proteome-wide Identification of PARylated Proteins in Respo

(A) A schematic representation of the SILAC-based enrichment strategy. U2OS c

were treated with genotoxic stress. Separate pull-downs of SILAC-encoded lysate

SILAC state and a PAR-binding GST-Af1521 wild-type macrodomain for medium

high-resolution LC-MS/MS.

(B) Quantitative comparison of tryptic peptide abundances.

(C) SILAC-based enrichment of PARylated proteins in the absence of exogenous

SILAC ratios from the reverse experiment were plotted against each other. Re

processes in comparison to annotated GO genes across the entire human geno

(D) U2OS cell lysate preparation in the absence or presence of 3-AB in the lysis bu

with 3-AB prior to cell lysis (lane 4). Pull-downs were performed with GST-Af152

(E) Hierarchical clustering of proteins derived from SILAC-based U2OS lysate pu

buffer. Functional GO analysis of regulated proteins in the absence of PARPi rev

(F) U2OS cells were transfected with PARG siRNA for 72 hr and treated with H2O

poly-, oligo- or mono-(ADP-ribosyl)ated substrates under the

experimental conditions used for our pull-down studies (Fig-

ure S1C), assuring no interference of catalytic activity with our

setup.

Next, we evaluated the performance of Af1521_wt and

Af1521_mut domains in a quantitative SILAC setup (Figure S1D).

High technical reproducibility of the experimental setup was

confirmed by a replicate ‘‘reverse’’ experiment, where medium

and heavy SILAC labels were swapped (Figure S1E and S1F).

Comparing proteins differentially bound between Af1521_wt

and Af1521_mut pull-downs revealed a large fraction of proteins

with a stronger affinity for Af1521_wt (Figure 1C). Gene ontology

(GO) enrichment analysis identified these as factors primarily

involved in DNA repair processes.

Thoughour results confirm thatAf1521_wt canbeemployedas

a PAR-binding module in a SILAC-based manner, we were

intrigued by the enrichment of PARylated proteins in the absence

of genotoxic stress (Figure 1C). The basal in vivo activity of

ARTD1/PARP1 has been described as low, stimulated through

genotoxic stress, and regulated by the opposing actions of

PARG (Bonicalzi et al., 2005). However, in vitro PAR formation

by ARTD1/PARP1 only requires the presence of NAD+ and DNA

fragments (Altmeyer et al., 2009). Thus, the shearing of DNA dur-

ing cell lysis could cause unphysiological PAR formation despite

fast and gentle sample handling. To investigate this, we tested

the presence of PARP inhibitors (PARPi) during sample prepara-

tion in a triple-encoded SILAC experiment with the Af1521_wt

domain used as ‘‘bait’’ in all three SILAC states (Figure S2A).

For light and medium SILAC cells, PARPi (PJ-34 or 3-AB) was

added to the lysis buffer, whereas heavy SILAC cells were left un-

treated during the entire experimental procedure. Additionally,

light SILAC cells were pretreated with PARPi for 2 hr prior to

cell lysis. After MS analysis, a substantial fraction of proteins

was found to be regulated more than 4-fold in the heavy SILAC

sample (complete absence of PARP inhibitor; Figure S2B, H/L

ratios). Contrary to this, very little ADP ribosylation was detected

in cells exposed to PARPi in the lysis buffer (Figure S2B, M/L

ratios), supporting the notion that postlysis mechanical shearing

stress triggers ARTD enzymatic activity (Beneke et al., 2012).WB

analysis of the same pull-down samples with PAR and ARTD1/

PARP1 antibodies corroborated these results (Figures 1D and

S2C). Moreover, hierarchical clustering of PARylated proteins

affected by the sample preparation procedure revealed that

DNA repair pathways are mainly affected by these processes

nse to Genotoxic Stress with SILAC-Based Quantification

ells were grown in light, medium, or heavy SILAC medium. Heavy-labeled cells

s were performed with a PAR-binding-defective Af1521mutant (G42E) for light

and heavy SILAC states. Eluates were resolved by SDS-PAGE and analyzed by

genotoxic stress. Logarithmized H/M SILAC ratios from the forward and M/H

gulated proteins reveal strong enrichment of proteins involved in DNA repair

me (indicated p values < 0.005).

ffer to prevent unphysiological activation of protein PARylation and cells treated

1_wt and analyzed by WB with the indicated antibodies.

ll-down in the absence (�) or presence (+) of 3-AB (5 mM) or PJ-34 in the lysis

eals strong enrichment of PARylated proteins involved in DNA repair.

2 and analyzed by WB with PAR antibody.


+si

PAR

G

+PJ

-34

(inly

sis

buffe

r)+

AD

P-H

PD(in

lysi

sbu

ffer)

F

A B

C

- UVH2O2 MMS

β-actin

PAR

IR

phospho-p53(Ser15)

H2O

2

MM

S UV IR

-4

-2

0

2

4

-4

-2

0

2

4H

2O2

MM

S UV IR

24%Nucleus

Other/Unknown76%

D

E

+ DNA damage(H/L ratio)

- DNA damage(M/L ratio)

Log 2

SIL

AC

ratio

s

G

Log 2

SIL

AC

ratio

s

0.6

0.8

1

0

0.2

0.4

0.6

0.8

1

-8 -4 0 4 8

M/LSILAC ratios

H/L SILAC ratios

p = 9.69e-35

Den

sity

Fold change

+ DNA damage- DNA damage

UV

IRMM

S

H2 O

2

DNA and RNA metabolic processes:

- DNA repair- transcription

RNA metabolic processes

-3 -2 -1 0 1 2 3

SILAC-Experiment Treatment Light (L) Medium (M) Heavy (H)

1 H2O2 Af1521_mut Af1521_wt Af1521_wt + H2O2

2 MMS Af1521_mut Af1521_wt Af1521_wt + MMS

3 UV Af1521_mut Af1521_wt Af1521_wt + UV

4 IR Af1521_mut Af1521_wt Af1521_wt + IR

4 3 21 2 4 31


Molecular Cell




Molecular Cell



(Figure 1E). On the basis of these findings, we included PARPi in

all sample preparations and set out to investigate the extent of

PARylation upon the induction of various types of genotoxic

stress. WB analysis of cells exposed to oxidative stress by

hydrogen peroxide (H2O2) only showed a relatively mild increase

in PAR levels (Figure 1F), which could be related to the short half-

life of PAR polymers (Alvarez-Gonzalez and Althaus, 1989). Thus,

in order to sustain PAR levels throughout the pull-down proce-

dure, we decided to reduce PARG activity by siRNA-dependent

depletion (Figure 1F) and additionally included PARG inhibitors

in the lysis buffer for all following experiments.

Proteomic Investigation of the Differential Impact ofVarious Types of Genotoxic Stress on ProteinPARylationNext, having established a robust SILAC-based workflow for the

enrichment of PARylated proteins, we investigated the cellular

PARylation response to various genotoxic insults. To this end,

we performed four SILAC experiments investigating PARylated

proteins upon treatment with different genotoxic agents (H2O2,

methyl methane sulfonate [MMS], UV radiation, and ionizing ra-

diation [IR]; Figure 2A).

Only heavy SILAC-labeled cells were exposed to genotoxic

agents; hence, PARylated substrates should exhibit an increase

in heavy/light (H/L) SILAC ratio in comparison to the medium/

light (M/L) ratios (cells not exposed to genotoxic stress). To vali-

date this, we plotted the logarithmized M/L and H/L SILAC ratios

of quantified proteins derived from the SILAC experiment

exposed to H2O2 insult (Figure 2B). The data showed that the

distribution of SILAC ratios of quantified proteins is indeed

shifted towards higher H/L SILAC ratios in comparison to the

M/L counterparts, demonstrating significantly increased (p =

9.69 3 10�35) protein PARylation upon genotoxic stress (Fig-

ure 2B). Moreover, given that light andmedium SILAC conditions

are essentially the same for all four SILAC experiments, they

constitute an internal control condition between pull-downs, as

confirmed by correlation analyses (Figure S3). In addition, box

plot analysis of all M/L ratios demonstrated identical distribu-

tions across all investigated samples (Figure 2C), whereas box

plot analysis of genotoxic-stress-induced samples (H/L ratios)

revealed much broader distributions, confirming that protein

PARylation is induced in all applied treatments (Figure 2D).

Strikingly, the degree of regulation was dependent upon the

Figure 2. Quantitative Analysis of DNA-Damage-Induced Protein PARy

(A) Four biological SILAC experiments were performed in order to investigate the

experiments, cells were treated in heavy SILAC state with H2O2 (1 mM, 10 min),

downs were performed with siPARG-treated U2OS cells and 40 mM PJ-34 and 1

(B) Significantly upregulated PARylated proteins in response to oxidative stress. L

representing untreated and treated cells. M/L SILAC ratios cumulate around 1 a

responding to induced protein PARylation. Statistical significance was calculate

(C) Box plot analysis of logarithmizedM/L SILAC ratios from all four SILAC experim

genotoxic stress (whiskers with minimum and maximum 1.5 interquartile range [

(D) Box plot analysis of logarithmized H/L SILAC ratios. Significant regulation rev

with minimum and maximum 1.5 IQR).

(E) SiPARG-transfected cells were treated with DNA-damaging agents and ana

follows the distribution of SILAC ratios in (D).

(F) A heat map of quantified proteins from the combined pull-downs in the absenc

stress are clustered together.

(G) GO term annotation enrichment for cellular distribution of significantly upregu

source of exogenous damage, and H2O2 showed the strongest

response, followed by MMS, UV, and IR in a steadily decreasing

manner (Figure 2D). WB analysis of cellular PARylation gener-

ated by the different genotoxic stress treatments revealed an

identical pattern (Figure 2E), confirming the direct association

between PARylated proteins andmeasured SILAC ratios. In sup-

port of the varying degree of PARylation caused by different gen-

otoxic stress treatments, a heat map clustered H2O2 and MMS

samples together, demonstrating that these chemical agents

cause the greatest degree of PARylation (Figure 2F). In order to

determine which proteins are significantly enriched in each

pull-down experiment, we used a ‘‘significant outlier’’ strategy

based on the Student’s t test to establish statistical significance

(p < 0.01). In total, 165 proteins could be determined as signifi-

cantly upregulated in the four experiments. However, these

measurements take only a normal distribution of the data set

into account, whereas PARylation during the H2O2 and MMS

treatments affects a much larger proportion of proteins in com-

parison to UV or IR damage (Figures 2D and 2E). This causes

an overall augmentation of the H/L ratios disproportioned to

the background binders, which can lead to an underestimation

of significancemeasurements that are based on normal distribu-

tions. As a consequence, we included PARylated substrates

from the H2O2 and MMS experiments if they exhibited at least

2.5-fold upregulation in their H/L ratios, concomitantly assuring

stringency for determining true interaction partners. Altogether,

we identified 235 proteins from the four experiments as signifi-

cantly enriched under the investigated genotoxic stress condi-

tions (Table S2).

Investigating the cellular distribution of the identified sub-

strates by GO enrichment analysis revealed that 76% of the

PARylated proteins belong to the nucleus (Figure 2G), signifying

a strong enrichment of nuclear PARylation targets considering

the pull-downs were performed on whole-cell lysates. This is in

full agreement with the nuclear localization of ARTD1/PARP1

and ARTD2/PARP2, which, together, account for the majority

of genotoxic stress-induced PAR formation in cells.

Proving Direct ADP Ribosylation of Target ProteinsTo select candidates for further investigation of PARylation-

dependent processes, we compared our data set to previously

published PARylation targets involved in DNA repair. We identi-

fied several of the proteins proposed to be PARylated, including

lation

impact of different DNA-damaging agents on protein PARylation. In the single

MMS (10 mM, 1 hr), UV (40 J/m2, 1 hr), or IR (10 Gy, 1 hr) respectively. All pull-

mM PARG inhibitor ADP-HPD in the lysis buffer.

ogarithmizedM/L and H/L SILAC ratios from the H2O2 experiment were plotted,

nd follow a normal distribution, whereas H/L ratios are shifted upwardly, cor-

d with a Wilcoxon rank-sum test.

ents (see A). No regulation of PARylated proteins is observed in the absence of

IQR]).

eals a strong induction of protein PARylation upon genotoxic stress (whiskers

lyzed by WB with the indicated antibodies. The extent of protein PARylation

e (�) or presence (+) of DNA damage. PARylated proteins induced by genotoxic

lated proteins in all four SILAC experiments.


25013010070

55

35

25

Ctrl PNKPCETN2

MPGMECP2

NUSAP1

POLR2E

CDK2XPC

Coomassie

32P-NAD

ARTD1/PARP1 ARTD2/PARP2

A

siARTD1/PARP1 siARTD2/PARP2siControl

- UVH2O2 MMS IR - UVH2O2 MMS IR - UVH2O2 MMS IR

β-actin

PAR

phospho-p53 (Ser15)

ARTD1/PARP1

C

191

97

97

51

39

Ctrl PNKPCETN2

MPGMECP2

NUSAP1

POLR2E

CDK2XPC

51

64ARTD2/PARP2

19618125

Total number of targets from SILAC screen

Targets from SILAC screen:DNA repair

Published targets: DNA repair

ARTD1ARTD2

FACT140XRCC5XRCC6PCNARPA1RPA2

XRCC1TOP2

HMGB1Histones

B

25013010070

55

35

25

10070

55

35

25

250130

10070

55

35

25

250130

*

**


Molecular Cell




Molecular Cell



ARTD1/PARP1, FACT140 (SUPT16H), XRCC5, XRCC6, PCNA,

RPA1, RPA2, XRCC1, TOP2, and HMGB1, strengthening the

reliability of our SILAC approach (Figure 3A). Cell-type-specific

protein expression or the requirement for specific genotoxic

conditions could account for not identifying previously sug-

gested PARylated proteins. Besides confirming the majority of

known PARylated proteins involved in DNA metabolism, we

additionally found 18 PARylation targets specifically described

to play a role in DNA-repair-related processes (Figure 3A).

As an additional demonstration that our established method

directly targets PARylated proteins, we biochemically verified

covalent PAR modification of purified proteins by in vitro PARy-

lation assays. Recombinant protein substrates were incubated

with purified ARTD1/PARP1 or ARTD2/PARP2 enzymes in the

presence of [32P]-NAD+ and a DNA fragment in order to measure

the incorporation of NAD+ radioactivity by autoradiography. The

activation of ARTD1/PARP1 and ARTD2/PARP2 was confirmed

by the detection of their strong automodification in the upper

part of the gel lanes (Figure 3B). Using ARTD1/PARP1, we could

detect a strong PARylation of PNKP, MPG (MID1), MECP2,

NUSAP1 and POLR2E, and, to a lesser degree, CETN2 and

CDK2. In vitro PARylation of XPC was difficult to confirm, given

that the protein runs with an almost similar size as ARTD1/

PARP1 (103 versus 111 kDa). A similar problem arises for

ARTD2/PARP2, which, at a mass of 64 kDa, further limits the

detection range. Still, MPG is validated as PARylated in vitro

by ARTD2/PARP2 (Figure 3B). Several of the identified PARy-

lated proteins, including PNKP, MPG (MID1), NUSAP1, CETN2,

and CDK2, have also been previously assigned with a functional

role in DNA repair processes, indicative of PARylation influ-

encing their mode of operation in DNA metabolism.

Notably, the derived in vitro results are in agreement with

ARTD1/PARP1 being responsible for the majority of cellular

PAR, whereas ARTD2/PARP2 accounts only for up to 15% of

PAR formation. This characteristic seems to be reflected not

only in the amount of PAR formed but also the number of actual

targets. In support of ARTD1/PARP1 being the primary enzyme

responsible for the identified PARylated substrates, knockdown

of ARTD1/PARP1 resulted in a strong decline of the PAR signal.

In contrast, siARTD2/PARP2-treated cells hardly showed any

difference in comparison to cells treated with control small inter-

fering RNA (siRNA) (Figure 3C). Notably, we detected a reduced

signal for phospho-p53 (Ser15) for all investigated genotoxic

stress treatments following ARTD1/PARP1 downregulation.

This was in agreement with previous reports showing comprised

DDR signaling for PAR-inhibited cells (Haince et al., 2007), high-

lighting the role of PAR as an early genotoxic-stress-sensory

molecule.

Figure 3. Identification and In Vitro Validation of Protein PARylation Ta

(A) Overlap of total number of PARylated proteins identified in this study with prot

DNA repair on the basis of GO term annotation.

(B) In vitro PARylation of identified protein targets in the SILAC screen. Purified fu

with recombinantly expressed proteins in the presence of [32P]-NAD and double-s

Coomassie blue (bottom) and autoradiography (AR, top). *, ARTD1/PARP1; **, A

ARTD1/PARP1.

(C) ARTD1/PARP1 accounts for the majority of protein PARylation in response to

control, ARTD1/PARP1, or ARTD2/PARP2 siRNA. Cells were treated with DNA

indicated antibodies.

Targets for PARylation In Vivo Are Involved in DifferentCellular ProcessesTo corroborate the in vitro validation of PARylated substrates

with in vivo data, we verified a number of in vitro identified PAR-

ylation targets by WB with specific antibodies after Af1521_wt

pull-down. In response to H2O2-induced stress, we found

MPG, PNKP, MECP2, XPC, and NUSAP1 to be enriched,

whereas NF-kB p65 served as a negative control (Figure 4A).

Pretreatment of the cells with PJ-34 completely abolished the

signal, confirming that enrichment occurs in a PARP-dependent

manner.

Furthermore, to demonstrate the in vivo PARylation of targets,

we immunoprecipitated GFP-tagged proteins stably expressed

under the control of their endogenous promoters (Poser et al.,

2008) and blotted them with PAR antibody. In order to cover

the functional diversity of targets in our data set in the best way

possible, we selected proteins regulating different nuclear pro-

cesses. We verified our findings for TAF15, which belongs to

the TLS/FUS, EWS, TAF15 (TET) protein family of RNA- and

DNA-binding proteins suggested to play a role in transcription

and splicing (Bertolotti et al., 1996; Jobert et al., 2009). Using

GFP-TAF15-expressing HeLa cells, we could confirm the results

of the Af1521 pull-down by WB. Similarly, GFP-TAF15 was vali-

dated to be PARylated by GFP immunoprecipitation in a geno-

toxic-stress-dependent and PARP-inhibitor-repressed manner

(Figure 4B). Strikingly, we foundall threemembers of the TETpro-

tein family to be PARylated under genotoxic stress conditions

(TableS2), suggesting a common role for PARylation in regulating

the function of these proteins in response to genotoxic stress.

Another RNA-associated factor that we discovered to

be PARylated in a genotoxic-stress-dependent manner was

THRAP3 (TRAP150), which controls mRNA splicing and nuclear

mRNA degradation and was recently assigned a role in the DDR

(Beli et al., 2012; Lee et al., 2010). In support of our SILAC data,

we found GFP-THRAP3 to be enriched in the Af1521 pull-down

in a PARP-dependent manner, and conversely, validated

THRAP3 PARylation upon GFP immunoprecipitation (Figure 4B).

The ATPase SMARCA5 (SNF2h) is a chromatin remodeler

belonging to the SWI/SNF family of proteins. It has recently

been linked to genotoxic stress signaling in a PAR-dependent

manner although no direct in vivo PARylation of SMARCA5

was detected after DNA damage (Smeenk et al., 2013), which

could be the consequence of the transient nature or low degree

of PARylation of SMARCA5. In the background of reduced PARG

activity, we could confirm the direct in vivo PARylation of

SMARCA5 (Figure 4B). Recently, ARTD1/PARP1 was demon-

strated to PARylate several components of a purified TIP5 com-

plex, hereby contributing to the repression of ribosomal RNA

rgets

eins that were previously observed to be PARylated and assigned with a role in

ll-length human ARTD1/PARP1 (left) and ARTD2/PARP2 (right) were incubated

tranded DNA oligomer. Samples were resolved by SDS-PAGE and analyzed by

RTD2/PARP2; Ctrl, control including Histone H1 and zinc finger 1 and 2 from

genotoxic stress. U2OS cells were cotransfected with PARG siRNA and either

-damaging agents as described in Figure 2A and analyzed by WB with the


191

97

191

97

97

191

191

97

191

191

- +

H2O2

PAR

XPC

GFP

PJ-34

-- ++

- +-- ++

- +-- ++

PD: Af1521_wt

*

*

*

TAF15-GFP

A

MECP2

NUSAP1

PNKP

B

MPG

97

64

64

39

51

28

64

51

51

NFkB

- +-- ++

10 % Input

191

97PAR

PAR

GFP

GFP

PAR

H2O2

PJ-34- +-- ++

PAR

GFP

SMARCA5-GFP

GFP

PAR

GFP

PAR

THRAP3-GFP

GFP

PAR

GFP

PAR

PD: Af1521_wt IP: GFP10% Input

PAR

GFP

97

191

97

191

97

191

97

191

191

97

191

191

97

191

97

97

191

Figure 4. In Vivo PARylation of Protein Tar-

gets

(A) siPARG-transfected U2OS cells were incu-

bated with PJ-34 (40 mM, 1.5 hr) or left untreated

prior to treatment with H2O2 (1 mM, 10 min) as

indicated. Cell lysates were pulled down with

GST-Af1521_wt and analyzed by WB with the

indicated antibodies in order to detect enriched

PARylated proteins. NF-kB antibody was used as

negative control.

(B) HeLa cells stably expressing TAF15, THRAP3,

or SMARCA5 as GFP fusion proteins were trans-

fected with PARG siRNA and treated with PJ-34

and H2O2 as in (A). Lysates were subjected to

Af1521_wt pull-down or GFP immunoprecipitation

and subsequently analyzed by WB with PAR and

GFP antibodies (*, PARylated GFP target protein).

Molecular Cell



transcription and the establishment of silent ribosomal DNA

chromatin during cell division (Guetg et al., 2012). SMARCA5

was found to be part of this complex, also referred to as nucle-

olar remodeling complex NoRC, indicating that PARylation of

SMARCA5 could regulate more processes than the DDR. Inter-

estingly, in our genotoxic stress screen, we also identify H2O2-

andMMS-dependent PARylation of UBTF (UBF1), another factor

involved in the regulation of nucleolar chromatin, supporting the


idea of PARylation as a means of coordi-

nating nucleolar function in response to

genotoxic stress. In general, chromatin

remodelers have recently been recog-

nized as important signaling coordinators

in genome stability pathways (Papami-

chos-Chronakis and Peterson, 2013).

Compelling evidence shows that PARyla-

tion governs access and activities of

some of these factors at sites of DNA

damage, as recently demonstrated for

CHD1L (ALC1) and CHD4 (Ahel et al.,

2009; Chou et al., 2010). Intriguingly, we

observed a strong PARylation of CHD1L

in all four genotoxic stress treatments,

whereas CHD4was significantly modified

after MMS treatment, indicating that the

direct PARylation of these proteins may

contribute to the regulation of their func-

tions (Table S2).

Comparative Analysis of theGenotoxic-Stress-InducedPARylome Reveals DistinctRegulation of RNA Metabolism inResponse to Oxidative andAlkylating DamageFunctional GO analysis of regulated

proteins revealed that DNA metabolic

processes are significantly enriched (Fig-

ure 5A). A considerable portion of these

factors is involved in DNA repair (p <

9 3 10�5), including many known PARylated candidates (Fig-

ure S4A). Strikingly, more than 60% of all significantly regulated

proteins also comprise nucleic-acid-binding activity (Figure 5B).

Although the proportion of RNA-binding proteins in the complete

data set was less in comparison to DNA-binding proteins, their

enrichment in relation to the total number of RNA-binding

proteins in the genome was more than 2-fold higher. This sug-

gests that posttranscriptional processes such as splicing,

A

C

E

B

72 11

28 32

5

22 3

1033 3

0

6 5

3 1

H2O2

IR MMSUV

RNA metabolic processes

H2O2

MMS

DNA metabolism

D

0 10 20 30 40 50 60 70

Nucleic acidbinding

RNA binding

DNA binding

2.4e-41

1.8e-32

1.5e-17

PARylated proteins/ SILAC screenGO genes/total genome

Proteins involved [%]

CHD1L

PCNA

RFC4

TOP2A

NUSAP1

PARP1

UHRF1HMGB2

CDK2

MPG

RPA1NFIC XRCC5

H2AFXHMGA1

RBBP4

HIST1H2BJ

HIST1H2BKHIST1H2AK

HIST1H1C

CDCA5

RFC1

XRCC1

LIG3

XRCC6

SUPT16H

PNKP

SFPQ

NONO

RFC3

SSRP1

XPC

TOP1

CPSF3

HNRNPU

CPSF2

ELAVL1

SRSF1

HNRNPA0

HNRNPR

GTF2I

SNRPD1

FBL

SNRPD2SNRNP70

SNRPN

KHDRBS1

SMNDC1

HNRNPA2B1

FUS

FIP1L1

NONO

RBM3

HNRNPA1

PABPN1

SNRPA

RBMX

SFPQ

SRSF7

CPSF4

HNRNPUL1

THOC4

DIMT1L

SUPT16H

CPSF1

POLR2E

PARP1

NFIC

THRAP3

SRSF1

RBMX

SNRPA

HNRNPU

FUS

KHDRBS1

TRA2B

HNRNPA2B1

HNRNPA1

NOP56

LOC653884

FBL

UTP15

SRSF7

HNRNPR

SUPT16H

THOC4

POLR2H

HNRNPC

NFIC

MKI67IP

NONO

RBM3

CPSF4

CPSF2

ELAVL1FIP1L1

PARP1

HNRNPA0

CPSF1THRAP3

0 10 20 30 40

DNA repair

Chromosomeorganization

DNA replication

Transcription

RNA processing

RNA splicing

Genome H2O2 MMS IR UV

DN

A m

etab

olis

mR

NA

met

abol

ism

Proteins involved [%]

Figure 5. Functional Consequences of

Distinct Types of DNA Damage Treatment

on Protein PARylation

(A) Protein interaction networks of significantly

PARylated proteins from all SILAC experiments

grouped into DNA metabolism by GO term anno-

tation for biological process (p < 8 3 10�14).

Network interaction data were extracted from the

STRING database and visualized with Cytoscape.

(B) InterPro domain annotation associated with the

molecular function of significantly regulated pro-

teins from all SILAC experiments in comparison to

annotated GO genes in the entire genome (indi-

cated p values < 1.53 10�17 ). Strong enrichments

for RNA- and DNA-binding proteins are observed.

(C) A Venn diagram demonstrating the overlap of

PARylated proteins from the individual DNA dam-

age experiments (as described in Figure 2A).

(D) GO enrichment analysis reveals specific

enrichment of RNA metabolic processes in H2O2

and MMS experiment, whereas proteins involved

in DNA metabolic processes are equally identified

across all DNA damage experiments. Analysis was

performed with GO term annotations for biological

processes of significantly regulated protein sets in

each DNA damage SILAC experiment (specific p

values are listed in Figure S4C).

(E) Protein interaction networks of significantly

PARylated proteins from the H2O2 (top) and MMS

(bottom) experiments grouped into RNA meta-

bolism by GO term annotation for biological pro-

cess (p < 2 310�16).

Molecular Cell



polyadenylation, mRNA stability and transport, and translation

are readily controlled by PARylation-induced genotoxic stress.

To assess the specific functional consequences of PARylation

under different types of genotoxic stress, we compared the

significantly regulated protein population of each single geno-

toxic stress experiment to all others (Figure 5C and Table S2).

In doing so, we identified ten proteins to be PARylated in all

four genotoxic stress conditions (ARTD1/PARP1, RPA1, MPG,

TAF15, FUS, RBMX, ALC1, DTX2, RUNX1, and ZNF384). Inter-

estingly, the dissection of the single genotoxic-stress-specific

data sets revealed that H2O2 and MMS treatment particularly

affect proteins involved in RNA metabolic processes (Figures

5D and S4C). This observation is also reflected by the overrepre-

sentation of RNA-binding proteins in H2O2 and MMS samples,

whereas the enrichment of regulated DNA-binding proteins is

equal for all genotoxic stress conditions (Figure S4B). A func-

tional network analysis showed that a large number of PARylated

substrates involved in RNA metabolic processes are intercon-

Molecular Cell 52, 1–

nected (Figure 5E). Whether this connec-

tivity is characteristic of PARylated pro-

teins involved in RNA metabolism

remains to be elucidated, given that pro-

teins in RNA metabolic processes are

generally strongly interconnected. Never-

theless, the majority of proteins involved

in RNA metabolic processes were not

known to be targeted for PARylation,

and only few proteins were previously observed to be modified.

However, emerging evidence indicates that PARylation plays an

important role in RNA processes during genotoxic stress. In fact,

in a recent siRNA screen, the RNA-binding protein RBMX was

identified as a positive regulator of homologues recombination

and found to localize to sites of DNA damage in a PARylation-

dependent manner (Adamson et al., 2012). Notably, we identify

RBMX as a PARylation target upon all applied genotoxic stress

treatments in our SILAC screen, supporting the concept of

ARTD1/PARP1-dependent regulation of RBMX (Table S1).

High-Content Imaging Analysis of PARylation-Dependent Regulation of Stress-Induced SubnuclearRelocalization of PARylated Proteins TAF15 andTHRAP3To investigate the role of PARylation in RNA metabolic pro-

cesses during genotoxic stress, we conducted detailed analyses

of TAF15 and THRAP3, known to be involved in transcription and

14, October 24, 2013 ª2013 Elsevier Inc. 9

ActD

ActD+ H2O2

ActD+ PJ-34+ H2O2

TAF15-GFP DAPImerge

ActD+ MMS

B

ActD+ PJ-34+ MMS

A

Nor

mal

ized

inte

nsity

TA

F15

caps

/ nu

cleu

s [A

U]

0

0.02

0.04

0.06

0.08

0.10

0.12

0.14

Ctrl H2O2 MMS

ActD

ActD + PJ-34

ActD + ABT-888

Automated nuclei detection Automated TAF15 cap detection

Fibrillarin

ActD+ PJ-34

C

Figure 6. PARylation-Dependent Relocali-

zation of TAF15 from Perinucleolar Caps in

Response to Genotoxic Stress

(A) In order to induce nucleolar cap formation,

HeLa cells were treated with 5 mg/ml ActD for up to

3 hr in the presence or absence of 40 mM PJ-34.

Prior to formaldehyde fixation, cells were exposed

to 2 mM H2O2 for 1 hr or 1 mM MMS for 2 hr and

additionally pre-extracted for visualization of

nucleolar caps. Immunostaining was performed

with TAF15 and Fibrillarin antibodies.

(B) Exemplary images taken from HCI analysis

illustrating automated software-assisted detection

of DAPI-stained nuclei and immunostained TAF15

caps.

(C) Quantification of TAF15 cap intensities per

nucleus of the conditions represented in (A) and

additional conditions with ABT-888.

Molecular Cell



splicing, respectively, and reported to localize to specific sub-

structures in the nucleus (Bertolotti et al., 1996; Jobert et al.,

2009; Lee et al., 2010). Dynamic spatiotemporal reorganization

of subnuclear structures and their constituents is an important

means for efficiently coordinating nuclear processes and cellular

stress responses (Boulon et al., 2010). For instance, the tran-

scriptional inhibition of RNA polymerase I and II with Actinomycin

D (ActD) has been shown to induce the segregation of nucleolar

components, resulting in the accumulation of a specific set of

proteins in perinucleolar caps (Shav-Tal et al., 2005). All TET fam-

ily members are known to accumulate in these stress-induced

cap structures (Jobert et al., 2009; Shav-Tal et al., 2005). This


led us to further investigate whether PAR-

ylation would contribute to the regulation

of nucleolar cap formation of TAF15 upon

genotoxic stress. We found TAF15 to be

enriched in cap-like structures upon

transcriptional inhibition with ActD in a

PARP-independent manner (Figure 6A

and S5A). In the absence of ActD, no

cap formation was induced after expo-

sure to genotoxic stress and/or PARPi

(Figure S5B). Strikingly, nucleolar cap

formation was abrogated when ActD-

treated cells were exposed to H2O2 or

MMS, indicating that oxidative and alkyl-

ation stress may regulate the reorganiza-

tion of nucleolar structures (Figures 6A

and S5A). Interestingly, the presence of

PJ-34 partially rescued the formation of

TAF15 containing cap-like structures

upon H2O2 and MMS treatment, indi-

cating that PARylation may contribute to

the redistribution of TAF15 from nucleolar

caps. In order to conduct an unbiased

and quantitative analysis for the stress-

dependent regulation of cap structure

formation, we applied high-content imag-

ing (HCI), thus enabling automated, inten-

sity-based detection of nuclei and immunofluorescence-stained

TAF15 caps of more than 1,000 cells per investigated condition

(Figure 6B). The quantitative measurements corroborated the

observed phenotypes for TAF15 caps, whose formation was

diminished upon genotoxic stress in a PAR-dependent manner,

as shown by the application of two different PARPi (PJ-34 and

ABT-888; Figure 6C).

To assess the applicability of our resource data for an unre-

lated biological regulation, we investigated the stress-depen-

dent subcellular relocalization of splicing factor THRAP3.

THRAP3 has previously been reported to be functionally linked

to its localization in nuclear speckles (Lee et al., 2010) and is

A

B

C D

Figure 7. PARylation-Dependent Accumu-

lation of THRAP3 in Nuclear Speckles in

Response to Genotoxic Stress

(A) Stable THRAP3-GFP-expressing HeLa cells

were treated with ActD and H2O2 and immuno-

stained with rabbit polyclonal GFP and mouse

monoclonal SC35 antibodies.

(B) THRAP3-GFP cells were left untreated or

treated with 5 mg/ml ActD for 3 hr in the presence

or absence of 40 mM PJ-34. Prior to methanol

fixation, cells were exposed to 2 mM H2O2 for the

indicated time points and immunostained with

mouse monoclonal GFP antibody and mouse

monoclonal SC35 antibodies.

(C) Quantification of THRAP3 speckle intensities

per nucleus of the conditions represented in (A)

and additional conditions with ABT-888.

(D) Quantification of THRAP3 speckle intensities

per nucleus of cells treated with siRNAs against

ARTD1/PARP1 or ARTD2/PARP2 in the presence

or absence of ActD and/or 2 mM H2O2.

Molecular Cell



known to function in the storage, assembly, and modification of

splicing factors. Hereby, a cell-type-specific basal exchange of

splicing factors occurs between speckles and nucleoplasm

that can bemodulated upon stress conditions such as transcrip-

tional inhibition. In contrast to the cytoplasm, compartmentaliza-

tion within the nucleus is not dependent on diffusion barriers

such as membranes. Thus, the reorganization of subnuclear

structures are often controlled by PTMs such as phosphorylation

and dephosphorylation events (Lamond and Spector, 2003). As

a result, we surmised that the PARylation of THRAP3 could

constitute a hitherto unappreciated mechanism for regulating

the stress-dependent assembly of factors in nuclear speckles.

First, we analyzed the cellular localization of GFP-tagged

THRAP3 and detected a strong accumulation into speckle-like

Molecular Cell 52, 1–14

structures upon H2O2 treatment with a

maximum response after 1 hr (Fig-

ure S6A). The localization of THRAP3

into nuclear speckles, as verified by co-

localization with the speckle marker

SC35 (Figure 7A), was more pronounced

under transcriptional inhibition with ActD.

These results possibly reflect a substanti-

ated inhibition of transcription, whereas

incubation with ActD alone did not result

in speckle formation, although THRAP3-

containing structures appeared more

spherical in comparison to untreated

cells (Figure 7B). Intriguingly, preincuba-

tion with PJ-34 prevented the oxidative-

stress-dependent accumulation of

THRAP3 in nuclear speckles (Figure 7B),

demonstrating that this stress-related

function of THRAP3 is regulated in a

PARylation-dependent manner. To

corroborate these results, we performed

an extended HCI analysis and quantified

the appearance of speckles in a

large-scale setup for each investigated condition (Figure S6B).

Accordingly, THRAP3 accumulation in speckles was fully abro-

gated in the presence of PJ-34, and the same phenotype was

observed for a more specific PARPi (ABT-888) (Figure 7C). To

investigate the PARP specificity of the observed THRAP3

accumulation, we performed HCI experiments with siRNAs to

knock down PARP1 and PARP2, respectively, revealing that

PAR-dependent THRAP3 accumulation is strongly dependent

on ARTD1/PARP1, but not ARTD2/PARP2 (Figures 7D, S6C,

and S6D). Notably, the weaker PAR dependence of the TAF15

phenotype in comparison to THRAP3 correlates with the

substoichiometric in vivo PARylation levels observed for

these proteins (Figure 4B). Collectively, the presented results

demonstrate that PARylation events regulate the spatiotemporal

, October 24, 2013 ª2013 Elsevier Inc. 11

Molecular Cell



dynamics of nuclear factors THRAP3 and TAF15, hereby coordi-

nating RNA metabolic processes in response to genotoxic

stress.

DISCUSSION

Here, we describe a SILAC-based enrichment strategy for high-

confidence in vivo detection of PARylated proteins under various

types of genotoxic stress. Substantial efforts were made to

establish an optimized technical protocol, including the impor-

tant use of PARP inhibitors during sample preparation to circum-

vent artificial PARylation of proteins, which otherwise could lead

to overestimation or misinterpretation of the quantitative data. In

strong support of the acquired data, we prove covalent PARyla-

tion of several targets by in vitro and in vivo approaches. Intrigu-

ingly, functional analysis of the data set establishes the extensive

regulation of processes related to not only DNA metabolism but,

significantly, also RNA metabolism, including transcription, RNA

splicing, and transport. We specifically uncover a role of PARy-

lation for THRAP3 and TAF15 in the coordination and assembly

of subnuclear structures, thereby facilitating the efficient regula-

tion of cellular responses to genotoxic stress.

The activity of ARTD1/PARP1 has long been known to have

an impact on base excision repair (BER) and single-strand break

repair (SSBR) (Dantzer et al., 2000; Fisher et al., 2007), although

its specific role in these processes has remained under debate.

In our screen, we identified several proteins of the BER and

SSBR machinery to be PARylated (ARTD1/PARP1, XRCC1,

MPG, LIG3, and PCNA) extending the list of previously known

targets by the DNA glycosylase MPG and DNA ligase LIG3, indi-

cating that direct PARylation of repair factors could constitute an

additional level of regulating repair efficiency. PNKP, another

identified and in vitro validated PARylation target, has been

described to serve a critical role in processing broken DNA

strands for allowing an efficient repair process (Weinfeld et al.,

2011), indicating that PARylation might affect the enzymatic

activity of its substrate proteins. Additionally, ARTD enzymes

have been demonstrated to play important roles in other DNA

repair pathways, such as nucleotide excision repair, mismatch

repair, and nonhomologous end-joining, supported by our find-

ings of regulated pathway-specific proteins such as XPC and

CETN2, MSH2 and MSH6, and XRCC5 and XRCC6, respec-

tively. Whether the PARylation of pathway-specific proteins

regulates their function positively or negatively and how specific

types of DNA lesions are involved in these processes remain to

be elucidated.

Aside from the direct PARylation of target proteins, cellular re-

sponses to genotoxic stress are also regulated by the recruit-

ment of PAR-binding proteins. Recent proteomic studies

focused on the identification and regulation of noncovalently

PAR-binding proteins by employing nondenaturing affinity purifi-

cation methods (Gagne et al., 2012). In contrast to this, we spe-

cifically determined the extent of covalently PARylated proteins

under various types of genotoxic stress using denaturing enrich-

ment methods. In support of our method targeting specifically

PARylated substrates, we did not identify PAR-binding factors

such as macro H2A.1, APLF, CHFR, and RNF146, which have

been reported to rapidly accumulate at sites of DNA damage


through their PAR-binding motifs (Kang et al., 2011; Li et al.,

2010; Liu et al., 2013; Timinszky et al., 2009). Nevertheless, it

cannot be excluded that PAR-binding proteins also become

subject to direct PARylation, which, according to our data set,

appears to be the case for ALC1, C6orf130, and Trip12, suggest-

ing a complex interplay between PAR-binding activities and the

PARylation of substrate proteins.

We found that PARylation affects a large number of proteins

involved in RNA metabolic processes, prominently in response

to oxidative and alkylation stress in comparison to UV or IR. It

has only recently become apparent that PARylation provides

an important link between the DDR and RNAmetabolism (Adam-

son et al., 2012; Paulsen et al., 2009). Furthermore, this connec-

tion is supported by our functional analysis of the pre-mRNA

splicing factor THRAP3, which we show to be PARylated in vivo

under oxidative damage. By applying quantitative HCI analysis,

we show that PARylation of THRAP3 affects its stress-depen-

dent cellular localization and is required for its colocalization

with splicing factors in nuclear speckles upon transcriptional

inhibition.

Considering that RNA-binding proteins are enriched in our

data set, we investigated the in vivo PARylation of TAF15 and

established a role for PARylation in coordinating its subnuclear

distribution and function. TAF15 is a member of the TET protein

family and contributes to the control of transcription, splicing,

RNA transport, and DNA repair processes (Tan and Manley,

2009). The TET proteins are frequently involved in genetic

translocations in sarcomas, thereby causing the inappropriate

transcriptional activation of target genes (Tan and Manley,

2009). Moreover, the depletion of FUS or EWS leads to

increased genomic instability and sensitivity to IR (Kuroda

et al., 2000; Li et al., 2007). In fact, recent studies have demon-

strated that Ewing’s sarcoma cells and xenografts are highly

sensitive towards ARTD1/PARP1 inhibition (Brenner et al.,

2012; Garnett et al., 2012). Suggested mechanisms for the

increased sensitivity included both the enforced accumulation

of DNA damage and ARTD1/PARP1-dependent positive feed-

back loops specifically of EWS-FLI1 fusion proteins for tran-

scriptional activation (Brenner et al., 2012). Direct PARylation

of TET family members provides another means for regulating

the function and activity of these proteins and its transloca-

tion-based fusion products.

Previously, PARylation has received much attention as a

potential target in cancer therapy, given that PARP inhibitors

showed promising results in early clinical trials (Fong et al.,

2009). Because our results confirm the PARylation of proteins

involved in RNA metabolic processes, further elucidation of

the regulatory role of PARylation in these processes upon

genotoxic stress might provide valuable information for future

strategies for effectively treating specific cancer subtypes with

PARP inhibitors. Although recent genome-wide siRNA screens

established a functional intersection of RNA processing with

DNA repair (Adamson et al., 2012; Paulsen et al., 2009), very

few RNA-processing factors have been examined in detail for

their link to the DDR. Therefore, our data infer that the

PARylation of RNA-processing factors upon genotoxic insult

may provide an additional layer for the participation of RNA-

binding proteins in the DDR.

Molecular Cell



EXPERIMENTAL PROCEDURES

Enrichment of PARylated Proteins

U2OS cells were grown in SILAC Dulbecco’s modified Eagle’s medium

(Invitrogen) supplemented with 10% dialyzed fetal bovine serum, sodium

pyruvate, L-glutamine, penicillin and streptomycin, and either L-lysine and

L-arginine, L-lysine 4,4,5,5-D4 and L-arginine–U-13C6, or L-lysine-U-13C6-

15N2 and L-arginine–U-13C6-15N4 (Cambridge Isotope Laboratories) (Ong

et al., 2002). The following cells were exposed to genotoxic agents and har-

vested at the indicated time points unless otherwise stated: IR (10 Gy, 1 hr),

UV (40 J/m2, 1 hr), MMS (10 mM, 1 hr), and H2O2 (1 mM in PBS, 10 min). Equal

protein amounts were incubated with GST-Af1521_wt or GST-Af1521_mut for

2 hr at 4�C, and bound complexes were eluted in Laemmli buffer. Eluates were

combined and resolved on 4%–20% SDS-PAGE, and gel slices were excised

and subsequently digested with Trypsin (T6567, Sigma-Aldrich).

Mass Spectrometric Analysis

All MS experiments were performed on a nano-flow high-performance liquid

chromatography system connected to anOrbitrapQ Exactivemass spectrom-

eter. Each peptide fractionwas autosampled and separated on a 15 cmanalyt-

ical column (75 mm inner diameter) packed with 3 mm C18 beads with a 2 hr

gradient ranging from 5%–40% acetonitrile in 0.5% acetic acid at a flow rate

of 250 nl/min. The Q Exactive mass spectrometer was operated in data-

dependent acquisition mode, and all samples were analyzed with the previ-

ously described ‘‘sensitive’’ acquisition method (Kelstrup et al., 2012). All

raw data analysis was performed with MaxQuant (Cox and Mann, 2008)

version 1.3.0.5 supported by the Andromeda peptide search engine (Cox

et al., 2011).

SUPPLEMENTAL INFORMATION

Supplemental Information contains Supplemental Experimental Procedures,

six figures, and two tables and can be found with this article online at http://

dx.doi.org/10.1016/j.molcel.2013.08.026.

ACKNOWLEDGEMENTS

We thank members of the Novo Nordisk Foundation Center for Protein

Research (NNF-CPR) for fruitful discussions and valuable comments. The

work carried out in this study was supported in part by the NNF-CPR, the

Lundbeck Foundation, the European Union 7th Framework Programme

PRIME-XS, grant agreement 262067, and EURAtrans, grant agreement

HEALTH-F4-2010-241504. This work was supported in part by the Swiss Na-

tional Science Foundation grant 310030B_138667 and the Kanton of Zurich

(both to M.O.H.). The Prestige Antibodies were a kind donation from Jan

Mulder.

Received: April 12, 2013

Revised: June 28, 2013

Accepted: August 13, 2013

Published: September 19, 2013

REFERENCES

Adamson, B., Smogorzewska, A., Sigoillot, F.D., King, R.W., and Elledge, S.J.

(2012). A genome-wide homologous recombination screen identifies the RNA-

binding protein RBMX as a component of the DNA-damage response. Nat.

Cell Biol. 14, 318–328.

Ahel, D., Horejsı, Z., Wiechens, N., Polo, S.E., Garcia-Wilson, E., Ahel, I., Flynn,

H., Skehel, M., West, S.C., Jackson, S.P., et al. (2009). Poly(ADP-ribose)-

dependent regulation of DNA repair by the chromatin remodeling enzyme

ALC1. Science 325, 1240–1243.

Altmeyer, M., Messner, S., Hassa, P.O., Fey, M., and Hottiger, M.O. (2009).

Molecular mechanism of poly(ADP-ribosyl)ation by PARP1 and identification

of lysine residues as ADP-ribose acceptor sites. Nucleic Acids Res. 37,

3723–3738.

Alvarez-Gonzalez, R., and Althaus, F.R. (1989). Poly(ADP-ribose) catabolism in

mammalian cells exposed to DNA-damaging agents. Mutat. Res. 218, 67–74.

Beli, P., Lukashchuk, N., Wagner, S.A., Weinert, B.T., Olsen, J.V., Baskcomb,

L., Mann, M., Jackson, S.P., and Choudhary, C. (2012). Proteomic investiga-

tions reveal a role for RNA processing factor THRAP3 in the DNA damage

response. Mol. Cell 46, 212–225.

Beneke, S., Meyer, K., Holtz, A., Huttner, K., and Burkle, A. (2012). Chromatin

composition is changed by poly(ADP-ribosyl)ation during chromatin immuno-

precipitation. PLoS ONE 7, e32914.

Bertolotti, A., Lutz, Y., Heard, D.J., Chambon, P., and Tora, L. (1996). hTAF(II)

68, a novel RNA/ssDNA-binding protein with homology to the pro-oncopro-

teins TLS/FUS and EWS is associated with both TFIID and RNA polymerase

II. EMBO J. 15, 5022–5031.

Bonicalzi, M.E., Haince, J.F., Droit, A., and Poirier, G.G. (2005). Regulation of

poly(ADP-ribose) metabolism by poly(ADP-ribose) glycohydrolase: where and

when? Cell. Mol. Life Sci. 62, 739–750.

Boulon, S., Westman, B.J., Hutten, S., Boisvert, F.M., and Lamond, A.I. (2010).

The nucleolus under stress. Mol. Cell 40, 216–227.

Brenner, J.C., Feng, F.Y., Han, S., Patel, S., Goyal, S.V., Bou-Maroun, L.M.,

Liu, M., Lonigro, R., Prensner, J.R., Tomlins, S.A., and Chinnaiyan, A.M.

(2012). PARP-1 inhibition as a targeted strategy to treat Ewing’s sarcoma.

Cancer Res. 72, 1608–1613.

Chou, D.M., Adamson, B., Dephoure, N.E., Tan, X., Nottke, A.C., Hurov, K.E.,

Gygi, S.P., Colaiacovo, M.P., and Elledge, S.J. (2010). A chromatin localization

screen reveals poly (ADP ribose)-regulated recruitment of the repressive poly-

comb and NuRD complexes to sites of DNA damage. Proc. Natl. Acad. Sci.

USA 107, 18475–18480.

Cox, J., and Mann, M. (2008). MaxQuant enables high peptide identification

rates, individualized p.p.b.-range mass accuracies and proteome-wide pro-

tein quantification. Nat. Biotechnol. 26, 1367–1372.

Cox, J., Neuhauser, N., Michalski, A., Scheltema, R.A., Olsen, J.V., and Mann,

M. (2011). Andromeda: a peptide search engine integrated into the MaxQuant

environment. J. Proteome Res. 10, 1794–1805.

Dani, N., Stilla, A., Marchegiani, A., Tamburro, A., Till, S., Ladurner, A.G.,

Corda, D., and Di Girolamo, M. (2009). Combining affinity purification by

ADP-ribose-binding macro domains with mass spectrometry to define the

mammalian ADP-ribosyl proteome. Proc. Natl. Acad. Sci. USA 106, 4243–

4248.

Dantzer, F., de La Rubia, G., Menissier-De Murcia, J., Hostomsky, Z., de

Murcia, G., and Schreiber, V. (2000). Base excision repair is impaired in

mammalian cells lacking Poly(ADP-ribose) polymerase-1. Biochemistry 39,

7559–7569.

Fisher, A.E., Hochegger, H., Takeda, S., and Caldecott, K.W. (2007).

Poly(ADP-ribose) polymerase 1 accelerates single-strand break repair in con-

cert with poly(ADP-ribose) glycohydrolase. Mol. Cell. Biol. 27, 5597–5605.

Fong, P.C., Boss, D.S., Yap, T.A., Tutt, A., Wu, P., Mergui-Roelvink, M.,

Mortimer, P., Swaisland, H., Lau, A., O’Connor, M.J., et al. (2009). Inhibition

of poly(ADP-ribose) polymerase in tumors from BRCA mutation carriers.

N. Engl. J. Med. 361, 123–134.

Gagne, J.P., Pic, E., Isabelle, M., Krietsch, J., Ethier, C., Paquet, E., Kelly, I.,

Boutin, M., Moon, K.M., Foster, L.J., and Poirier, G.G. (2012). Quantitative pro-

teomics profiling of the poly(ADP-ribose)-related response to genotoxic

stress. Nucleic Acids Res. 40, 7788–7805.

Garnett, M.J., Edelman, E.J., Heidorn, S.J., Greenman, C.D., Dastur, A., Lau,

K.W., Greninger, P., Thompson, I.R., Luo, X., Soares, J., et al. (2012).

Systematic identification of genomicmarkers of drug sensitivity in cancer cells.

Nature 483, 570–575.

Guetg, C., Scheifele, F., Rosenthal, F., Hottiger, M.O., and Santoro, R. (2012).

Inheritance of silent rDNA chromatin is mediated by PARP1 via noncoding

RNA. Mol. Cell 45, 790–800.

Haince, J.F., Kozlov, S., Dawson, V.L., Dawson, T.M., Hendzel, M.J., Lavin,

M.F., and Poirier, G.G. (2007). Ataxia telangiectasia mutated (ATM) signaling

network is modulated by a novel poly(ADP-ribose)-dependent pathway in




Molecular Cell



the early response to DNA-damaging agents. J. Biol. Chem. 282, 16441–

16453.

Hassa, P.O., Haenni, S.S., Elser, M., and Hottiger, M.O. (2006). Nuclear ADP-

ribosylation reactions in mammalian cells: where are we today and where are

we going? Microbiol. Mol. Biol. Rev. 70, 789–829.

Jankevicius, G., Hassler, M., Golia, B., Rybin, V., Zacharias, M., Timinszky, G.,

and Ladurner, A.G. (2013). A family of macrodomain proteins reverses cellular

mono-ADP-ribosylation. Nat. Struct. Mol. Biol. 20, 508–514.

Jobert, L., Pinzon, N., Van Herreweghe, E., Jady, B.E., Guialis, A., Kiss, T., and

Tora, L. (2009). Human U1 snRNA forms a new chromatin-associated snRNP

with TAF15. EMBO Rep. 10, 494–500.

Kang, H.C., Lee, Y.I., Shin, J.H., Andrabi, S.A., Chi, Z., Gagne, J.P., Lee, Y., Ko,

H.S., Lee, B.D., Poirier, G.G., et al. (2011). Iduna is a poly(ADP-ribose) (PAR)-

dependent E3 ubiquitin ligase that regulates DNA damage. Proc. Natl. Acad.

Sci. USA 108, 14103–14108.

Karras, G.I., Kustatscher, G., Buhecha, H.R., Allen, M.D., Pugieux, C., Sait, F.,

Bycroft, M., and Ladurner, A.G. (2005). The macro domain is an ADP-ribose

binding module. EMBO J. 24, 1911–1920.

Kelstrup, C.D., Young, C., Lavallee, R., Nielsen, M.L., and Olsen, J.V. (2012).

Optimized Fast and Sensitive Acquisition Methods for Shotgun Proteomics

on a Quadrupole Orbitrap Mass Spectrometer. J. Proteome Res. Published

online May 10, 2012. http://dx.doi.org/10.1021/pr3000249.

Kuroda, M., Sok, J., Webb, L., Baechtold, H., Urano, F., Yin, Y., Chung, P., de

Rooij, D.G., Akhmedov, A., Ashley, T., and Ron, D. (2000). Male sterility and

enhanced radiation sensitivity in TLS(-/-) mice. EMBO J. 19, 453–462.

Lamond, A.I., and Spector, D.L. (2003). Nuclear speckles: a model for nuclear

organelles. Nat. Rev. Mol. Cell Biol. 4, 605–612.

Lee, K.M., Hsu, IaW., and Tarn, W.Y. (2010). TRAP150 activates pre-mRNA

splicing and promotes nuclear mRNA degradation. Nucleic Acids Res. 38,

3340–3350.

Li, H., Watford,W., Li, C., Parmelee, A., Bryant, M.A., Deng, C., O’Shea, J., and

Lee, S.B. (2007). Ewing sarcoma gene EWS is essential for meiosis and B

lymphocyte development. J. Clin. Invest. 117, 1314–1323.

Li, G.Y., McCulloch, R.D., Fenton, A.L., Cheung, M., Meng, L., Ikura, M., and

Koch, C.A. (2010). Structure and identification of ADP-ribose recognition

motifs of APLF and role in the DNA damage response. Proc. Natl. Acad. Sci.

USA 107, 9129–9134.

Liu, C., Wu, J., Paudyal, S.C., You, Z., and Yu, X. (2013). CHFR is important for

the first wave of ubiquitination at DNA damage sites. Nucleic Acids Res. 41,

1698–1710.


Lukas, J., Lukas, C., and Bartek, J. (2011). More than just a focus: The chro-

matin response to DNA damage and its role in genome integrity maintenance.

Nat. Cell Biol. 13, 1161–1169.

Luo, X., and Kraus, W.L. (2012). On PAR with PARP: cellular stress signaling

through poly(ADP-ribose) and PARP-1. Genes Dev. 26, 417–432.

Ong, S.E., Blagoev, B., Kratchmarova, I., Kristensen, D.B., Steen, H., Pandey,

A., and Mann, M. (2002). Stable isotope labeling by amino acids in cell culture,

SILAC, as a simple and accurate approach to expression proteomics. Mol.

Cell. Proteomics 1, 376–386.

Papamichos-Chronakis, M., and Peterson, C.L. (2013). Chromatin and the

genome integrity network. Nat. Rev. Genet. 14, 62–75.

Paulsen, R.D., Soni, D.V., Wollman, R., Hahn, A.T., Yee, M.C., Guan, A.,

Hesley, J.A., Miller, S.C., Cromwell, E.F., Solow-Cordero, D.E., et al. (2009).

A genome-wide siRNA screen reveals diverse cellular processes and path-

ways that mediate genome stability. Mol. Cell 35, 228–239.

Poser, I., Sarov, M., Hutchins, J.R., Heriche, J.K., Toyoda, Y., Pozniakovsky,

A., Weigl, D., Nitzsche, A., Hegemann, B., Bird, A.W., et al. (2008). BAC

TransgeneOmics: a high-throughputmethod for exploration of protein function

in mammals. Nat. Methods 5, 409–415.

Rosenthal, F., Feijs, K.L., Frugier, E., Bonalli, M., Forst, A.H., Imhof, R.,Winkler,

H.C., Fischer, D., Caflisch, A., Hassa, P.O., et al. (2013). Macrodomain-con-

taining proteins are new mono-ADP-ribosylhydrolases. Nat. Struct. Mol.

Biol. 20, 502–507.

Shav-Tal, Y., Blechman, J., Darzacq, X., Montagna, C., Dye, B.T., Patton, J.G.,

Singer, R.H., and Zipori, D. (2005). Dynamic sorting of nuclear components

into distinct nucleolar caps during transcriptional inhibition. Mol. Biol. Cell

16, 2395–2413.

Smeenk, G., Wiegant, W.W., Marteijn, J.A., Luijsterburg, M.S., Sroczynski, N.,

Costelloe, T., Romeijn, R.J., Pastink, A., Mailand, N., Vermeulen, W., and van

Attikum, H. (2013). Poly(ADP-ribosyl)ation links the chromatin remodeler

SMARCA5/SNF2H to RNF168-dependent DNA damage signaling. J. Cell

Sci. 126, 889–903.

Tan, A.Y., and Manley, J.L. (2009). The TET family of proteins: functions and

roles in disease. J. Mol. Cell Biol. 1, 82–92.

Timinszky, G., Till, S., Hassa, P.O., Hothorn, M., Kustatscher, G., Nijmeijer, B.,

Colombelli, J., Altmeyer, M., Stelzer, E.H., Scheffzek, K., et al. (2009). Amacro-

domain-containing histone rearranges chromatin upon sensing PARP1 activa-

tion. Nat. Struct. Mol. Biol. 16, 923–929.

Weinfeld, M., Mani, R.S., Abdou, I., Aceytuno, R.D., and Glover, J.N. (2011).

Tidying up loose ends: the role of polynucleotide kinase/phosphatase in

DNA strand break repair. Trends Biochem. Sci. 36, 262–271.

http://dx.doi.org/10.1021/pr3000249

Documents

Proteome-wide Identification of Poly(ADP-Ribosyl)ation Targets in Different Genotoxic Stress Responses