Histone H3K9 methylation is dispensable for Caenorhabditis elegans development but suppresses RNA:DNA hybrid-associated repeat instability

Abstract

Histone H3 lysine 9 (H3K9) methylation is a conserved modification that generally represses transcription. In Caenorhabditis elegans it is enriched on silent tissue-specific genes and repetitive elements. In met-2 set-25 double mutants, which lack all H3K9 methylation (H3K9me), embryos differentiate normally, although mutant adults are sterile owing to extensive DNA-damage-driven apoptosis in the germ line. Transposons and simple repeats are derepressed in both germline and somatic tissues. This unprogrammed transcription correlates with increased rates of repeat-specific insertions and deletions, copy number variation, R loops and enhanced sensitivity to replication stress. We propose that H3K9me2 or H3K9me3 stabilizes and protects repeat-rich genomes by suppressing transcription-induced replication stress.

Full text

© 2016Nature America, Inc. All rights reserved.
Nature GeNetics ADVANCE ONLINE PUBLICATION 1
ARTICLES
H3K9 is a common target of methylation in vivo and can carry one,
two or three methyl groups. H3K9me2 or H3K9me3 mark transcrip-
tionally silent heterochromatin in most eukaryotes1–3 . In mammals,
insects and Schizosaccharomyces pombe, H3K9 methylation is highly
enriched at telomeres, pericentric heterochromatin and interspersed
repetitive elements (REs)4–7.
Ligands that recognize methylated H3K9, such as heterochromatin
protein 1 (HP1), mediate transcriptional repression of reporter genes
and chromatin compaction near centromeres2,8. H3K9me is also
implicated in the silencing of genes both during development9,10 and
in pathological states. For instance, tumor-suppressor genes have been
found to be transcriptionally silenced by mistargeted H3K9me in can-
cers11,12, and H3K9me marks triplet repeat sequences, whose expan-
sion has debilitating consequences in syndromes such as Huntington’s
or Fragile X13,14. Nonetheless, by reducing levels of H3K9me the effi-
ciency of somatic cell reprogramming can be increased15,16.
It has been difficult to study the function of H3K9me-mediated
repression in complex organisms for several reasons. First, there are
at least eight documented and partially redundant H3K9 histone
methyltransferases (HMTs) in mammals (SUV39h1, SUV39h2, G9a,
SETDB1, SETDB2, PRDM2, PRDM3 and PRDM16 in mice). Second,
the vast majority of H3K9 methylation is found on extended stretches
of REs that cannot be accurately mapped by standard deep sequenc-
ing techniques17. In some cases the disruption of individual H3K9me
HMTs is embryonically lethal, owing in part to compromised mitotic
chromosome segregation18–20. The loss of SUV39h1, SUv39h2 or their
homologs also results in mitotic defects, aneuploidy and chromosomal
rearrangements in mice, flies and fission yeast7,21,22. This may have
masked phenotypes arising from the loss of H3K9me in transcrip-
tional repression during development.
The holocentric nematode C. elegans has only two, nonredundant
H3K9me-depositing HMTs, MET-2 and SET-25 (refs. 23,24). Here we
exploited the finding that mutants lacking both HMTs have no detect-
able H3K9 methylation24, and yet produce viable embryos, to study how
the loss of this histone modification impacts a multicellular organism.
RESULTS
Loss of H3K9me did not impair embryonic differentiation into
adult tissues
The HMT MET-2, which catalyzes the mono- and di-methylation
of H3K9, is the homolog of mammalian SETDB1, also known as
ESET23. SET-25, on the other hand, shares considerable SET domain
homology with SUV39h1, SUV39h2 and G9a enzymes, and it is the
only C. elegans enzyme that trimethylates H3K9 (ref. 24). To con-
firm that met-2 set-25 double mutant worms lack H3K9 methylation
throughout development, we performed immunofluorescence analy-
sis at all stages of worm development (Fig. 1a). We found no detect-
able H3K9me2 or me3 in met-2 set-25 embryos, second-stage larvae
(L2) or gonads of adult worms, confirming our earlier mass spec-
troscopic analysis of total histones isolated from mutant embryos or
larvae24. Histone acetylation and other common methylation marks
(Supplementar y Fig. 1) remained intact24. Despite this complete
absence of H3K9me, the met-2 set-25 mutant embryos developed
into viable adults.
Histone H3K9 methylation is dispensable for
Caenorhabditis elegans development but suppresses
RNA:DNA hybrid-associated repeat instability
Peter Zeller1,2,4, Jan Padeken1,4, Robin van Schendel3, Veronique Kalck1, Marcel Tijsterman3 & Susan M Gasser1,2
Histone H3 lysine 9 (H3K9) methylation is a conserved modification that generally represses transcription. In Caenorhabditis
elegans it is enriched on silent tissue-specific genes and repetitive elements. In met-2 set-25 double mutants, which lack all H3K9
methylation (H3K9me), embryos differentiate normally, although mutant adults are sterile owing to extensive DNA-damage-
driven apoptosis in the germ line. Transposons and simple repeats are derepressed in both germline and somatic tissues. This
unprogrammed transcription correlates with increased rates of repeat-specific insertions and deletions, copy number variation,
R loops and enhanced sensitivity to replication stress. We propose that H3K9me2 or H3K9me3 stabilizes and protects repeat-rich
genomes by suppressing transcription-induced replication stress.
1Friedrich Miescher Institute for Biomedical Research, Basel, Switzerland. 2Faculty of Natural Sciences, University of Basel, Basel, Switzerland. 3Department of
Human Genetics, Leiden University Medical Center, Leiden, the Netherlands. 4These authors contributed equally to this work. Correspondence should be addressed
to S.M.G. (susan.gasser@fmi.ch).
Received 28 June; accepted 22 August; published online 26 September 2016; doi:10.1038/ng.3672

© 2016Nature America, Inc. All rights reserved.
2 ADVANCE ONLINE PUBLICATION Nature GeNetics
ARTICLES
To monitor the kinetics of somatic development, we compared the
timing of wild-type (N2) and met-2 set-25 organisms as they transi-
tioned from the first larvae stage (L1) to the L1 stage of the next gen-
eration. This is a highly synchronous cycle that takes 3 d in wild-type
strains grown at 20 °C (Fig. 1b). In contrast to wild-type worms, 52% of
the met-2 set-25 mutants showed stochastic delays in stage transitions,
even though most mutant embryos reached adulthood (88% became
mature adults; Fig. 1b). These delays were more pronounced at 25 °C
than at 20 °C and were not restricted to one specific stage (Fig. 1b).
Nonetheless, only 2% of the adult offspring displayed a grossly irregu-
lar morphology at 20 °C, i.e., ‘dumpy’ appearance, partially defective
cuticles or bursting as adults (Fig. 1c). Such aberrant morphologies
were below detection level (<0.1%) in wild-type populations25.
Chromosome missegregation has been suggested to be a main cause
for the phenotypes observed in mutants for H3K9 HMTs in other organ-
isms7,19,26. Using histone H2B fusion to GFP (H2B-GFP), we tracked
the frequency of mitotic chromosome bridges or lagging chromosomes
in wild-type and met-2 set-25 embryos. The frequency of defective
mitoses at either 20 °C or 25 °C was similar in wild-type and mutant
embryos (Fig. 1d). Moreover, the duration of mitosis was identical,
which argues against any mutant-specific spindle checkpoint activation
(Fig. 1d). To monitor meiotic chromosome missegregation, we followed
H2B-GFP-tagged oocytes undergoing meiosis in gonads. Thanks to the
chromosome condensation and enlarged nuclei that occur in diakine-
sis, we could determine bivalent chromosome number per cell. Again
there was no detectable difference between met-2 set-25 and wild-type
oocytes at either temperature (Fig. 1e). Thus, we excluded aneuploidy
and spindle checkpoint activation as triggers for the developmental
delay or aberrant morphologies of H3K9me-deficient worms.
Temperature-dependent sterility of met-2 set-25 mutant
Brood sizes were notably smaller upon propagation of the double
HMT mutant, and worms became completely sterile after two genera-
tions at 26 °C (Supplementary Fig. 2a). We determined the number of
viable progeny of met-2 set-25 vs. wild-type worms under controlled
growth conditions at 15 °C, 20 °C and 25 °C. Although brood size
was equal between the met-2 set-25 and wild-type worms at 15 °C,
mutant adults had significantly fewer viable progeny at both 20 °C and
Day 1
Day 2
Day 3
Day 1
Day 2
Day 3
met-2 set-25wt
Dead
L1–L2
L3
L4
Young adult
Eggs laid
L1 (F1)
Developmental stage
25 °C
20 °C
Day 1
Day 2
Day 3
Day 1
Day 2
Day 3
Dead
L1–L2
L3
L4
Young adult
Eggs laid
L1 (F1)
Developmental stage
# Chr. per oocyte
In diakinesis
1
2
3
4
5
6
wt
Diakinesis
Diplotene
20 µm
Diakinesis
Diplotene
20 µm
met-2 set-25
1
2
3
4
5
6
20 °C 25 °C
Mean duration (min) 9 ±1 10 ± 19 ± 1 9 ± 1
% of proper mitosis 98 9797 100
20 °C 25 °C
H2B-GFP
5 min 6 min 5 min 6 min
met-2 set-25wt
H2B-GFP
Dumpy Cuticle def.BurstNo phenotype
met-2 set-25
2%98%
wtmet-2 set-25
H3K9me2
H3K9me2
H3K9me3
H3K9me3
Gonads
H3K9me2
H3K9me2
H3K9me3
H3K9me3
Embryos
H3K9me2
H3K9me2
H3K9me3
H3K9me3
L2 larvae
met-2 set-25wt
met-2 set-25wt
20 °C
25 °C
met-2 set-25wt
5.9
±
0.3 5.7
±
0.6
5.7
±
0.1 5.6
±
0.2
a b
c
d e
Figure 1 Worms lacking H3K9me were viable but showed stochastically delayed development. (a) Immunofluorescence images using H3K9me2- and
H3K9me3-specific antibodies on wild-type (wt) and met-2 set-25 strains at indicated developmental stages. H3K9me2 and H3K9me3 signals are in
green, and DAPI in blue. Scale bars, 5 µm. (b) met-2 set-25 mutation provoked stochastic delays in development from L1 larval stage into fertile adults.
Developmental progress of singled mutant and wt L1 larvae monitored every 24 h for 3 d at 20 °C and 25 °C (N (number of biological replicates) = 3,
n (number of animals per replica) = 50). (c) Example images of worm morphologies arising in met-2 set-25 cultures and their frequencies (N = 4,
n = 50). Scale bar, 100 µm. (d,e) H3K9me2/H3K9me3 was not essential for chromosome segregation in C. elegans. Images (d) from time-lapse
(∆t = 1 min) movies of mitotic cells in embryos expressing H2B-GFP in which mitotic defects were scored (wt, 20 °C n= 34; wt, 25 °C n = 50;
met-2 set-25, 20 °C n = 45; and met-2 set-25, 25 °C n = 36). Scale bars, 3 µm. Duration of mitosis reflects minutes from the beginning of
chromosome condensation until completion of telophase. The number of bivalent chromosomes (e) in wt and met-2 set-25 worms expressing H2B-GFP
counted in oocytes undergoing diakinesis (N = 3; wt, 20 °C n = 57; met-2 set-25, 20 °C n = 51; wt, 25 °C n = 50; and met-2 set-25, 25 °C n = 50).
Mean and s.d. are shown. Insets, nucleus of an oocyte in diakinesis.

© 2016Nature America, Inc. All rights reserved.
Nature GeNetics ADVANCE ONLINE PUBLICATION 3
ARTICLES
25 °C (Fig. 2a). A similar temperature-dependent loss of fertility has
been observed for mutants of the PIWI pathway27,28 (Supplementary
Fig. 2b), a germline-specific small RNA pathway that helps to silence
transposable elements29.
Gonad development per se was not impaired in the met-2 set-25
mutant (Supplementary Fig. 2c). However, by scoring the expres-
sion of the CED-1::GFP phagocytic receptor, which accumulates
on the plasma membrane of apoptotic cells30, we detected a high
level of germline apoptosis (Fig. 2b). The level increased when we
grew worms at 25 °C. In the double mutant an average of 30 cells
per germ line were positive for CED-1 at 20 °C (wild-type: 10 cells),
and over 80 cells per germ line at 25 °C (wild-type: 18 cells; Fig. 2b).
Consistently, RNA sequencing (RNA-seq) of met-2 set-25 gonads
showed an increase in mRNA from various other apoptosis-specific
genes31 (Supplementary Fig. 2d).
Although C. elegans germline cells are known to be particularly sensi-
tive to DNA damage, germline apoptosis can have multiple causes32. To
see whether apoptosis in H3K9me-deficient gonads is caused by DNA
damage, we deleted the mammalian p53 homolog, CEP-1, and scored
CED-1::GFP distribution at 20 °C and 25 °C (ref. 33). In the met-2 set-25
cep-1 triple mutant and in the strain lacking cep-1 alone, we detected only
background levels of germline apoptosis at both temperatures (Fig. 2b).
This strongly suggests that the germline apoptosis seen in the absence
of H3K9me stemmed from DNA damage. The met-2 set-25 cep-1
triple mutant was synthetic sterile, as expected (Fig. 2c). Of embryos
laid at 20 °C, hatching rate dropped from above 95% in the met-2 set-25
mutant to below 80% when coupled with cep-1 (Supplementary
Fig. 2e). This is likely due to an increase in DNA damage in the mutant,
because the number of RAD-51 foci per cell, a marker of processed
breaks, increased significantly (P < 0.001, two-sided Wilcoxon signed
rank test), as did the number of cells in the mitotic zone of the germ
line with RAD-51 foci (3.4% in the wild type and 14.6% in the double
mutant; Supplementary Fig. 2f). This suggests that germline cells incur
enhanced levels of damage in the absence of H3K9me.
H3K9me2 marks REs, whereas H3K9me3 marks REs and
silent genes
To understand the link between the loss of H3K9 methylation and the
observed increase in DNA damage, we first reexamined the sequences
reported to be bound by histones bearing H3K9me2 and H3K9me3.
We performed chromatin immunoprecipitation followed by high-
throughput sequencing (ChIP-seq) experiments, not unlike those
reported by the modENCODE consortium34,35. We found a tenfold
enrichment of both H3K9me2 and H3K9me3 along the distal arms of
the five worm autosomes in early embryos (Supplementary Fig. 3a).
We did not observe this distribution for other repressive marks, such
as H3K27me3, nor for the active mark, H3K4me3. Chromosome
arms were similarly enriched for all types of REs (Supplementary
Fig. 3a36). A detailed analysis of the distribution of H3K9me2 versus
H3K9me3 in embryos showed that a high proportion of H3K9me2
was on REs (~34% of all H3K9me2), whereas H3K9me3 was present
equally on exons and REs (~26% each, Fig. 3a).
Distinct classes of repetitive DNA constitute large fractions of the
genomes of complex organisms. These include DNA or RNA trans-
posons, which can generate copies of themselves and integrate into the
genome, as well as simple repeats, such as tandemly arranged micro-
or minisatellites (Fig. 3b). Unlike transposons, these latter repeats
lack open reading frames (ORFs) and regulatory sequences. Worm
genomes contain all classes of REs, although DNA (rather than RNA)
transposons are the most abundant transposable elements37. Short
repetitive sequences are not found as megabase blocks of pericentric
satellite sequence in worms, but as short clusters distributed along the
chromosome. As a consequence, 87% of the C. elegans REs, or roughly
~60,000 discrete elements, can be uniquely mapped to individual sites
of the genome by standard next-generation sequencing.
Plotting the enrichment of H3K9me2 and H3K9me3 on all REs
in embryos, we found that 24.3% of REs were exclusively enriched
for H3K9me2, and 18.1% had either both marks or exclusively
H3K9me3 (Fig. 3c). This revealed that 42.4% of mappable REs were
enriched for H3K9me, with H3K9me2 and H3K9me3 distributed
differentially over the three repeat classes. RNA transposons were
most strongly correlated with H3K9me3 (58.5%, with 5.7% bear-
ing H3K9me2 only); tandem or simple repeats were more likely to
carry H3K9me2 alone (31.6%), and DNA transposons fell into two
groups: 25.5% were uniquely dimethylated whereas 17.7% carried
H3K9me3 (Fig. 3c).
In embryos H3K9me3 was enriched on transcriptionally silent
genes (12.0%), where it coated entire ORFs of loci (Fig. 3d and
0
100
200
300
400
15 °C
# of viable progeny per worm
wt
met-2 set-25
wt
met-2 set-25
wt
met-2 set-25
wt
wt
met-2 set-25
met-2 set-25
wt
wt
met-2 set-25
met-2 set-25
20 °C 25 °C
d-1 Sheath cell engulng
apoptotic cell
CED-1::GFP
*
*
CED-1::GFP
wtmet-2 set-25
0
20
40
60
80
100
120
# of CED-1 positive cells
*** ***
*** ***
n.s.
20 °C 25 °C
cep-1
20 °C 25 °C 20 °C 25 °C
# of viable progeny per worm
# of viable progeny per worm
CEP-1 + +– – + +– –
n.s. **** ** ****
0
100
200
300
400
0
100
200
300
400
wt
met-2 set-25
wt
met-2 set-25
a b c
Figure 2 DNA-damage-checkpoint-dependent increase of apoptotic cells in the germ line of met-2 set-25 worms. (a) Number of viable progeny of wt
and met-2 set-25 mutant per worm at 15 °C, 20 °C and 25 °C (N = 3, n = 75). (b) Example image of a gonad and the quantification of the number
of apoptotic cells in worms expressing the apoptosis marker CED-1::GFP in wt and met-2 set-25 background, with and without CEP-1. Apoptosis rate
was determined as the number of cells fully engulfed by CED-1::GFP per gonad arm. CED-1 is a phagocytic receptor, which translocates to the plasma
membrane during apoptosis. Asterisks indicate gonad tip and boxes mark enlarged section in the overview image (N = 3, n = 75). Scale bar, 10 µm.
(c) Number of viable progeny per worm of wt and met-2 set-25 with or without CEP-1. At both 20 °C and 25 °C (N = 3, n = 75) cep-1 and met-2 set-25
showed a synthetic loss of viable progeny. Boxplots show median, boxes 50% and whiskers 90% of the group. Two-sided Wilcoxon signed-rank test: n.s.
indicates not significant, **P < 0.005, ***P < 0.0001 and ****P < 0.00005.

© 2016Nature America, Inc. All rights reserved.
4 ADVANCE ONLINE PUBLICATION Nature GeNetics
ARTICLES
Supplementary Fig. 3b), and it was depleted from active genes
(1.8%; Fig. 3d). Among the H3K9me3-bound genes were many that
were expressed only in terminally differentiated tissues and a large
fraction of pseudogenes (Fig. 3d and Supplementar y Fig. 3c–e).
Loss of H3K9me led to the derepression of genes and REs
To determine whether loss of H3K9me affects transcription, we
performed RNA-seq on RNA isolated from either gonads or early
embryos of wild-type and met-2 set-25 strains, grown at either 20 °C
or 25 °C. In embryos cultured at 20 °C, we observed the reproducible
derepression (>2-fold compared to wild-type) of 308 genes. Of these
72.2% (234) were marked by H3K9me in wild-type cells, and are
therefore likely to be regulated directly by MET-2 and/or SET-25
(Fig. 4a and Supplementary Fig. 4a). This set of derepressed genes
was only a subset (~9.7%) of all genes bearing H3K9me, arguing that
the loss of H3K9me is not always sufficient to activate transcription.
Derepression of genes was also temperature-sensitive, with 2.2-fold
more genes being upregulated at 25 °C, including 83.8% of those
% of all H3K9me2-enriched regions
Promoter
Exons
Introns
Intergenic
REs
0 10 20 30 40
Promoter
Exons
Introns
Intergenic
REs
0 10 20 30 40
Tandem repeats
Microsatellites Minisatellites
18.1% 17.7% 58.5% 11.3%
31.6%
5.7%25.5%24.3%
12.0% 1.8% 17.9% 27.9%
5.8%6.6%4.3%6.5%
Total repetitive elements DNA transposons RNA transposons Tandem repeats
Total genes Embryonically
expressed genes
Embryonically
non-expressed genes
Pseudogenes
RNA transposons
LTR retrotransposon Non-LTR retrotransposons
LTR gag pro pol env LTR IVP EN RT A(n)
DNA transposons
TIR TIR
Cut and paste DNA transposons
Transposase
IP H3K9me3 – input (log2)
% of genome
(C. elegans)8.4% 0.8% 3%
Sequence
annotations
IP H3K9me3 – input (log2)
IP H3K9me3 – input (log2)
IP H3K9me3 – input (log2)
IP H3K9me3 – input (log2)
IP H3K9me3 – input (log2)
IP H3K9me3 – input (log2)
IP H3K9me3 – input (log2)
IP H3K9me2 – input (log2) IP H3K9me2 – input (log2) IP H3K9me2 – input (log2)IP H3K9me2 – input (log2)
IP H3K9me2 – input (log2) IP H3K9me2 – input (log2) IP H3K9me2 – input (log2) IP H3K9me2 – input (log2)
% of all H3K9me3-enriched regions
Signal density
Uniquely
H3K9me2
pos (%)
H3K9me2
and me3
pos (%)
0
0
IP – input (log2)
IP – input (log2)
Sequence
annotations
6
4
2
0
–2
–4
–6
6
4
2
0
–2
–4
–6
6
4
2
0
–2
–4
–6
6
4
2
0
–2
–4
–6
–4 –2 0 2 4 6 8 –4 –2 0 2 4 6 8 –4 –2 0 2 4 6 8 –4 –2 0 2 4 6 8
6
4
2
0
–2
–4
–6
6
4
2
0
–2
–4
–6
6
4
2
0
–2
–4
–6
6
4
2
0
–2
–4
–6
–4 –2 0 2 4 6 8 –4 –2 0 2 4 6 8–4 –2 0 2 4 6 8 –4 –2 0 2 4 6 8
a
b
c
d
Figure 3 Differential enrichment of H3K9me2 and H3K9me3 on repeat element classes and gene types. (a) Percentage of H3K9me2 and H3K9me3
domains covering promoters, exons, introns, unique intergenic sequences or REs (N = 2). H3K9me positive regions were determined from genomic bins
of sequences recovered after CHIP-seq using H3K9me2- or H3K9me3-specific antibodies with IP/input > 0. (b) Schematic representation of the three
major repeat classes. DNA transposons encode a single transposase, which catalyzes all the steps of transposition, flanked by two terminal inverted
repeats (TIRs). RNA transposons are either long terminal repeat (LTR) or non-LTR retrotransposon types. As derivatives of ancient retrovirus infections
LTR retrotransposons encode gag (structural proteins of the virus core), pol (reverse transcriptase, integrase), pro (protease) and env (envelope).
Non-LTR transposons encode a reverse transcriptase (RT) and an endonuclease (EN). Retrotransposon flanking regions in both cases supply promoter
elements. Tandem repeats are short, noncoding sequence stretches that are repeated in a head-to-tail fashion. (c) High-density scatterplots show
the enrichment of H3K9me2 and H3K9me3 on REs based on CHIP-seq data. IP, immunoprecipitation. RNA transposons were heavily enriched for
H3K9me3 (58.5%), whereas 31.6% of tandem repeats had only H3K9me2. Lines indicate the quadrants of single-positive, double-positive and
double-negative elements. (d) High-density scatterplots of the H3K9me2 and H3K9me3 enrichment on genes. Nonexpressed genes and pseudogenes
were enriched for H3K9me3.

© 2016Nature America, Inc. All rights reserved.
Nature GeNetics ADVANCE ONLINE PUBLICATION 5
ARTICLES
298
131
31
31
7
155
41
Gonads (20 °C)
Embryos
(25 °C)
Embryos
(20 °C)
– 3 – 2 – 1 0 1 2 3
– 3
– 2
– 1
0
1
2
3
– 3
– 2
– 1
0
1
2
3
– 3
– 2
– 1
0
1
2
3
Embryos (20 °C)
Fold change (log2) met-2 set-25 rep1
Fold change (log2)
met-2 set-25 rep2
Embryos (25 °C)
Fold change (log2) met-2 set-25 rep1
Fold change (log2)
met-2 set-25 rep2
Gonads (20 °C)
Fold change (log2) met-2 set-25 rep1
Fold change (log2)
met-2 set-25 rep2
14
44
8
4
5
31
8
Gonads (20 °C)
Embryos
(25 °C)
Embryos
(20 °C)
20 °C
25 °C
20 °C
Gonads
HelitronY3_CE
Helitron2_CE
PAL8D_CE
IR1_CE
TC4
PALTA2_CE
TC5
PAL5A_CE
LONGPAL2
NPALTA1_CE
TIR9TA1B_CE
PALTTTAAA3
TIR9TA1_CE
PALTA3_CE
IR4
PALTA1_CE
Tc3
PAL8C_2
TIR43YW1_CE
Mariner4_CE
Mariner5_CE
CEMUDR2
TIR9TA1A_CE
Tc1
CEMUDR1
NTC2A
CERP4
CELE11
RC14
MIRAGE1
Tc6
Turmoil1
SINE1_CE
CELE45
CER17−I_CE
CER2−LTR_CE
CER16−I_CE
LTR1_CE
CER6−I_CE
CER8−LTR_CE
CER12−LTR_CE
CER10−LTR_CE
CER13−I_CE
CER5−I_CE
CER7−LTR_CE
CER17−LTR_CE
CER15−1−I_CE
CER9−LTR_CE
CER12−1−LTR_CE
CER9−I_CE
CER11−I_CE
CER10−I_CE
CER15−1−LTR_CE
CER8−I_CE
CER16−2−I_CE
CER4−I_CE
LTR2_CE
CER16−LTR_CE
CER15−I_CE
CER7−I_CE
CER2−1−LTR_CE
CER2−I_CE
CER15−LTR_CE
CER3−1−LTR_CE
CER3−I_CE
LINE2C_CE
LINE2H_CE
LINE2C1_CE
RTE1
20 °C
25 °C
20 °C
(CGGG)n
(TCC)n
(CCG)n
(AATTG)n
(CCTAG)n
(TCGTG)n
(CCTA)n
(ATTG)n
(CTAGGG)n
(TATAA)n
(GGA)n
(CGATA)n
(GAAA)n
(CATG)n
(AAATG)n
(TAAA)n
(TTGG)n
(CAGA)n
(CTACT)n
(CAGAG)n
(CAAT)n
(CAG)n
(TGG)n
(GGGAA)n
(TATTG)n
(GAA)n
(TGAG)n
(ATG)n
(CCAA)n
(ACTCG)n
(TGAA)n
(ATCTG)n
(TTATG)n
CeRep54
(CATTT)n
(TTCA)n
(CCA)n
(TTG)n
(CATCC)n
(TCG)n
(CGGGA)n
(TCCCG)n
(GAATG)n
(TCCA)n
(CATTC)n
(GATTG)n
(CAATC)n
(GGATG)n
(TGGA)n
(CGA)n
MSAT1_CE
– 4 – 2 0 2 4
log2 fold change in expression
– 3 – 2 – 1 0 1 2 3 – 3 – 2 – 1 0 1 2 3
Gonads
20 °C
25 °C
20 °C
Gonads
Embryos
DNA transposons
RNA transposons
Tandem repeats
– 4 – 2 0 2 4
– 4
– 2
0
2
4
– 4
– 2
0
2
4
– 4
– 2
0
2
4
Embryos (20 °C)
Fold change (log2) met-2 set-25 rep1
Fold change (log2)
met-2 set-25 rep2
Embryos (25 °C) Gonads (20 °C)
Fold change (log2) met-2 set-25 rep1
Fold change (log2)
met-2 set-25 rep2
Fold change (log2) met-2 set-25 rep1
Fold change (log2)
met-2 set-25 rep2
Embryos Embryos
20% 37.6% 14.4%
1.2% 2.6% 1%
RE subfamilies
Genes
– 4 – 2 0 2 4 – 4 – 2 0 2 4
a b
c d
e
Figure 4 Temperature-dependent derepression of subsets of genes and repeat families in embryos and gonads in met-2 set-25 worms. (a) Fold change
(log2, met-2 set-25/wt) in gene expression of two replicas (rep1 and rep2) of RNA-seq data from embryos (20 °C and 25 °C) and from isolated gonads
(20 °C) for each strain. The genes marked in red were consistently >2-fold upregulated (P < 0.05, FDR < 0.1), and % of total genes is indicated.
(b) Venn diagram of the derepressed genes shows that genes affected in gonads were distinct from those upregulated in early embryos, and that
derepression was temperature-enhanced in embryos (N = 3). (c) Scatterplot of the expression changes of H3K9me-enriched RE subfamilies in
met-2 set-25 embryos and gonads compared to wt. The REs marked in red were >1.5-fold derepressed in both replicas (P < 0.05, FDR < 0.1). (d) Venn
diagram of the derepressed REs shows that subfamilies affected in gonads were partially distinct from subfamilies derepressed in embryos, and that
RE upregulation was temperature-enhanced in embryos (N = 3). (e) Heat map of the expression changes in all significantly affected RE subfamilies
(P < 0.05, FDR < 0.05) in embryos (20 °C and 25 °C) and gonads, sorted by repeat class (N = 3).

© 2016Nature America, Inc. All rights reserved.
6 ADVANCE ONLINE PUBLICATION Nature GeNetics
ARTICLES
already derepressed at 20 °C (Fig. 4a,b). Transcription in gonads was
elevated by the loss of H3K9me (210 genes). The affected genes were
largely distinct from those derepressed in somatic cells (37.6% over-
lap; Fig. 4a,b), arguing that transcription factor availability is critical
for transcriptional activity in the absence of repressive chromatin.
No essential regulators of meiosis were misregulated.
Given that REs were enriched for H3K9me in wild-type worms,
we next examined expression changes for REs, which we analyzed as
thoc-2
α-RNA:DNA hybrid
DAPI
α-RNA:DNA hybrid
wt
% of embryos with
RNA:DNA hybrids
0
20
40
60
80
100
15 252015 252015 2520
wt met-2 set-25
% of input
0
1
2
3
4
5
12 20 °C
α-RNA:DNA hybrid
ChrI: 13.15 mb
CELE14A
MSAT1
CER5
CER5
MSAT1
CER5
met-2 set-25
0.0
0.2
0.4
0.0
0.2
0.4
wt
DRIP-seq – RNase H (log2)
Genes
repeats
met-2 set-25
% of input
+ RNase H
0
2
4
6
8
10
wt met-2 set-25 thoc-2wt met-2 set-25 thoc-2
20 °C
Signal intensity
/background (AU)
4 2 1
DNA (µg) 4 2 1 421
0
2
α-RNA:DNA hybrid
+RNase H
Mean DRIP (log2)/
RNaseH control
0 1–1
Total RE
DNA
RNA
Tandem repeats
met-2 set-25 wt
Exons
Transposons
H3K9me
+
Derepressed
in met-2 set-25
All
H3K9me
+
Derepressed
in met-2 set-25
Temp. (°C)
20 °C
α-RNA:DNA hybrid
0
0.1
0.2
0.3
0.4
0.5
CEMUDR-1
Tc3
Tc4
CER10-I_CE
Helitron 2
MSAT-1
(TATCG)n
unc-119
lmn-1
eef-1A.1
CEMUDR-1
Tc3
Tc4
CER10-I_CE
Helitron 2
MSAT-1
(TATCG)n
unc-119
lmn-1
eef-1A.1
ab
c
d e
f
All
Figure 5 met-2 set-25 worms accumulate RNA:DNA hybrids at repeat elements. (a) Quantification of multiple dot blots against RNA:DNA hybrids
(antibody S9.6, HB-8730, ATCC, n = 3) in genomic DNA isolated from gravid adults of wt, met-2 set-25 and thoc-2 strains grown at 20 °C. 4 µg,
2 µg and 1 µg of nucleic acids were loaded for each strain. Where indicated, genomic DNA was treated with RNase H before blotting (mean + s.e.m.;
N = 3). (b) Immunofluorescence (IF) images of isogenic wt, met-2 set-25 and thoc-2 mutant embryos grown at 20 °C, stained with antibody S9.6 to
visualize RNA:DNA hybrids (green); DAPI is in blue. Scale bar, 5 µm. (c) Quantification of IF signals after S9.6 staining from embryos of indicated
strains, grown at 15 °C, 20 °C or 25 °C (mean + s.e.m.; N = 3, n = 15). (d,e) Genome-wide distribution of R loops determined by DRIP with antibody
S9.6, followed by qPCR or deep sequencing of recovered DNA. DRIP-qPCR (d) for seven repeat subfamilies upregulated in met-2 set-25 worms and
for three control loci (unc-119, lmn-1 and eef-1A.1), that were not upregulated (mean and s.e.m.; N = 2). Heat map of an S9.6 DRIP-seq experiment
(e) showing mean log2 enrichment over the corresponding controls treated with RNase H (samples normalized to total number of reads). Loci were
segregated based on indicated sequence criteria, and were further subgrouped based on the presence of H3K9me and response to the met-2 set-25
mutations (N = 1). (f) DRIP-seq example showing the R-loop signal over a RE cluster. The IP signal was normalized to the input and the RNase H
control values were subtracted.

© 2016Nature America, Inc. All rights reserved.
Nature GeNetics ADVANCE ONLINE PUBLICATION 7
ARTICLES
subfamilies. We characterized ~84% of all annotated repeats (300 sub-
families), and excluded only very-low-complexity repeat sequences or
elements with a single annotated occurrence. In met-2 set-25 mutant
embryos at 20 °C, 20% of the H3K9me-enriched repeat subfamilies
were derepressed by at least 1.5-fold, and at 25 °C this value increased
to 37.6% (Fig. 4c and Supplementary Fig. 4b). Gonads isolated from
double mutant adults showed an increase of transcription in 14.4%
of all H3K9me repeat subfamilies (Fig. 4c). This lower number of
derepressed REs may reflect germline-specific redundant silencing
by the PIWI pathway38,39. Indeed, different REs were upregulated
in gonads and somatic cells (Fig. 4d and Supplementar y Fig. 4c),
with tandem repeats being distinctly underrepresented in the germ
line (Fig. 4e). We note that each class of repeats includes REs that
were not derepressed by loss of H3K9me, which may reflect either
the existence of other, H3K9me-independent silencing pathways,
or a requirement for transcription factors that are tissue-specific or
developmental-stage-specific.
We asked whether the transcriptional landscape of genes surround-
ing a RE might influence its expression upon loss of H3K9me. This
is particularly relevant for simple tandem repeats, which lack recog-
nizable promoter or enhancer sequences40. To our surprise, ~50%
of the derepressed tandem repeats were not in the proximity of an
upregulated gene (data not shown).
H3K9me-deficient worms accumulated R loops
We next examined the relationship between aberrant RE transcrip-
tion and the obser ved DNA damage. Perturbation of the replication
fork is a major driver of DNA lesions41, and a substantial obstacle for
its progression is the transcription machinery, in particular when
stalled by RNA:DNA hybrids (R loops)42–45. In fission yeast, R loops
are enriched at repetitive sequences, such as transposons, telomeres
or the rDNA46, and correlated with genetic instability47,48. We there-
fore checked whether the met-2 set-25 double mutant accumulated
R loops, using multiple approaches based on an antibody specific for
RNA:DNA hybrids (S9.6, gift of P. Pasero49).
We detected an accumulation of R loops in met-2 set-25 worms
that was not detectable in wild-type worm DNA by performing a dot
blot analysis of genomic DNA. We also detected significant R-loop
occurrence by immunostaining of mutant, but not wild-type, embryos
(P < 0.001, Student`s t-test; Fig. 5a–c and Supplementary Fig. 5a).
The level was roughly similar to that scored in a mutant strain deficient
for the Tho-Trex complex (thoc-2), in which RNA:DNA hybrids accu-
mulate owing to impaired RNA processing and export (Fig. 5a)50,51.
To test for antibody specificity, we treated the isolated DNA with
RNase H before blotting, to specifically degrade RNA:DNA het-
eroduplexes. Quantification showed that 60% of the signal (met-2
set-25, loading 4 µg; Fig. 5a) was lost after treatment with RNase
H. Consistent with the elevated level of RE transcription at higher
temperatures, the level of R loops increased with temperature, both in
the dot blot analysis of adult worm DNA, as well as in the immunos-
taining of embryos (Fig. 5c and Supplementary Fig. 5a). The thoc-2
mutant, on the other hand, reached R-loop saturation even at 15 °C.
To examine formation of R loops in a sequence-dependent manner,
we immunoprecipitated RNA:DNA hybrids from wild-type or met-2
RNA transposons
0
10
20
30
40
Genomic copy number
(met-2 set-25/wt)
DNA transposons
Tandem repeats Exons
wt
met-2
set-25
25 °C
0
10
20
30
40
Genomic copy number
(met-2 set-25/wt)
0
250
500
750
1,000
MSAT1
lmn-1
CEREP58
(TATCG)n
Genomic copy number
(norm. to unc-119)
0
250
500
750
1,000
0
20
40
60
80
100
0
20
40
60
80
100
% adults 3 d post HU pulse
0 2.5 5 10 20
wt met-2set-25
wt met-2set-25
300 5 10 20 40
HU
γ-IR
HU (mM)
% adults 3 d post IR
γ-IR (Gy)
×8Singling for
12 generations
Whole genome
sequencing
HELITRON
MSAT1
CEREP58
wt
met-2 set-25
wt
met-2 set-25
wt
met-2 set-25
wt
met-2 set-25
wt
met-2 set-25
wt
met-2 set-25
Genomic copy number
(norm. to unc-119)
HELITRON2
HELITRONY4
a
b
c
d e
(TATCG)n
Figure 6 The met-2 set-25 strain is hydroxyurea sensitive and accumulates mutations in repeat elements and reactivated transposable elements. (a) Worms
lacking MET-2 and SET-25 were hypersensitive to hydroxyurea (HU). Synchronized populations of wt and mutant L1 larvae were exposed to 20 mM HU for
16 h, and then the numbers of worms that develop into adults after 3 d were quantified. Statistical analysis shows significant differential loss of viability
between met-2 set-25 and wt worms (Wilcoxon test P < 0.0001 at all doses; N = 6, n = 25). (b) Synchronized populations of wt and mutant L1 larvae were
exposed to sublethal doses of gamma irradiation (γ-IR), and the numbers of worms that developed into adults after 3 d were quantified. Only at 0 Gy was
there enhanced met-2 set-25 lethality (Wilcoxon test P < 0.005; N = 3, n = 15). Boxplots show median, 50% boxes and 90% whiskers. (c) Wt and met-2
set-25 worms were singled for 12 generations at 25 °C, generating 8 substrains per genotype. Genomic DNA of each substrain was sequenced, and only
mutations unique to one of the substrains were considered. (d) Copy number of REs and exons in met-2 set-25 determined by genome sequencing and
sorted according to repeat class. REs analyzed by qPCR are indicated in red. (e) Analysis of the copy number of REs by qPCR from 8 met-2 set-25 and 8 wt
substrains cultured as described in c, shows an increase in CNV of DNA transposons, simple repeats but not RNA transposons or single-copy genes.

© 2016Nature America, Inc. All rights reserved.
8 ADVANCE ONLINE PUBLICATION Nature GeNetics
ARTICLES
set-25 embryos followed by deep sequencing (DRIP-seq) or qPCR
(DRIP-qPCR). By qPCR, we found that specific repeat elements that
were derepressed in the absence of H3K9me, were enriched for R
loops fourfold to ninefold in mutant over wild-type strains. This was
not the case for low- or moderate-level transcribed genes (unc-119
or lmn-1), nor was there a met-2 set-25-dependent increase in DRIP
for a highly transcribed gene (eef-A.1), although the levels of R loops
did increase at highly transcribed genes in both wild-type and met-2
set-25 strains (Fig. 5d). As proof that the antibody was specific for
RNA:DNA hybrids, we note that the DRIP-qPCR signal was highly
sensitive to treatment with RNase H (Fig. 5d).
On a genome-wide level (DRIP-seq), we detected the most pro-
nounced enrichment of RNA:DNA hybrids in met-2 set-25 embryos
on REs that were derepressed in the double mutant (Fig. 5e,f and
Supplementary Fig. 5e). RNA:DNA hybrids were particularly
enriched on transcribed DNA transposons and tandem repeats but
not on RNA transposons (Fig. 5e). Confirming R-loop mapping in
other organisms, we observed RNA:DNA hybrids more frequently
on highly transcribed genes, telomeres and the rDNA locus, even
in wild-type cells (Supplementary Fig. 5b–d)52,53, yet these signals
showed no further increase in the met-2 set-25 mutant.
This high level of RNA:DNA hybrids suggests the presence of rep-
lication stress in met-2 set-25 worms. To monitor their sensitivity to
fork stalling, we exposed worms to hydroxyurea, a DNA replication
inhibitor that reversibly inhibits ribonucleotide reductase, thereby
depleting deoxynucleotide pools and exacerbating replication fork
stalling54. L1 larvae exposed to 20 mM hydroxyurea for 16 h and
allowed to recover for 3 d in absence of the inhibitor, yielded 95 ± 3%
(mean ± s.d.) viability (resumption of development), whereas only
43 ± 11% of the met-2 set-25 larvae survived hydroxyurea exposure
(Fig. 6a). This hypersensitivity was specific to agents causing replica-
tion stress, as treating similarly staged larvae with ionizing radiation
did not differentially affect wild-type and met-2 set-25 strains (Fig. 6b).
Thus hydroxyurea hypersensitivity correlated with the accumula-
tion of R loops, and suggests that both the developmental delays and
sterility detected in H3K9me-deficient worms reflect collisions of
replication with unscheduled transcription.
In the absence of H3K9me, mutations accumulated in REs
Replication stress and formation of R loops have been correlated
with both fork instability and double-strand break hotspots in
yeast44,46,55,5 6. To determine whether genomes of H3K9me-defi-
cient worms accumulate mutations at elevated rates, we singled 8
wild-type and 8 met-2 set-25 worms for 12 generations at 25 °C,
thereby creating 8 individual substrains per genotype. Sequencing of
the genome of each substrain revealed mutations exclusively in one
of the 16 genomes (Fig. 6c). This allowed us to score the number,
nature and location of changes accumulated owing to the met-2
set-25 mutation.
We note that the rate and nature of single nucleotide variants
(SNVs) did not differ between wild-type and met-2 set-25 worms,
which allowed us to exclude generation time as a confounding factor
in the analysis (Supplementary Fig. 6a). However, 6 of the 8 met-2
set-25 sub-strains acquired at least one insertion or deletion (indel)
(with a total of 9 different obser ved indels; Supplementary Table 1).
In contrast, only one wild-type strain incurred small deletions (3-base
pair (bp) and 5-bp). The average indel in the met-2 set-25 substrains
covered 5.3 kilobases (kb) (the largest being 33.5 kb), all met-2 set-25
indels occurred at sites enriched for H3K9me3, and 8 of the 9
met-2 set-25 indels occurred in REs whose majority showed enhanced
transcription upon loss of H3K9me.
Confirming the existence of large and stable germline changes, we
detected a 10-kb inversion flanked on one side by a 1-kb deletion by
whole genome sequencing and PCR of met-2 set-25 worms that had
been cultivated for several months. The inversion was immediately
adjacent to an excised Tc3 transposon, and opposite the inversion was
a de novo Tc3 transposon insertion unique to the cultivated H3K9me-
deficient strain (Supplementary Fig. 6b−d). The excised Tc3 element
carried H3K9me3 in the wild-type strain, and was transcriptionally
activated in met-2 set-25.
< 1/3
1/3 – 2/3 > 2/3
phsp16.2
gfp lacZ
(
(
588 bp 744 bp
SNV
Indel
Euchromatic array
None
Heat shock
β-gal
Heterochromatic array
0
wt
12 h 24 h
% of worms
msh-6
Euchromatic array
0
Heterochromatic array
% of input
None< 1/3
1/3 – 2/3> 2/3
Categories
20 °C
12 h 24 h
% of worms
ChIP:
ChIP:
Translation
gfp lacZ
(
(
Out of frame
Phsp16.2
gfp lacZ
(
(
Translation
Mutation
establishes frame
β-gal
Phsp16.2
~300 copy = heterochromatic array
Low-copy = euchromatic array
12 h
24 h L2
L1
L1
460 bp
1.2
0.8
0.4
0
% of input
1.2
0.8
0.4
H3K9me3 H3K9me2 IgG
H3K9me3 H3K9me2 IgG
Tc4
Tc4
Tc4
lmn-1
lmn-1
lmn-1
Tc4
Tc4
Tc4
lmn-1
lmn-1
lmn-1
100
80
60
40
20
0
100
80
60
40
20
met-2
set-25
wt msh-6met-2
set-25
wt msh-6met-2
set-25
wt msh-6met-2
set-25
a
b
c
d
e
gfp
gfp
gfp
gfp
gfp
gfp
Figure 7 Somatic accumulation of indels leading to frameshift mutations in met-2 set-25 mutant larvae. (a) Schematic of reporter to monitor mutation
frequency in single cells in two different chromatin contexts. A lacZ construct containing a frame-shift mutation under the control of a heat shock
promoter was integrated either as a high copy (~200−300 copies), or low-copy array (~20 copies). The frameshift prevents lacZ translation, which
can be reestablished by mutation. (b) To quantify the accumulation of mutations, L1 larvae were released into development for 12 h or 24 h at 20 °C
before a heat shock and subsequent β-gal staining. (c) ChIP-qPCR monitored enrichment of H3K9me2/H3K9me3 on the reporter array by PCR for
gfp. H3K9me was recovered on the heterochromatic (high-copy) array but not the euchromatic (low-copy) array. The genomic copy of lmn-1 and Tc4
served as negative and positive controls. (d) Genomic DNA of the heterochromatic array was isolated from either met-2 set-25, or wt worms grown for
24 h. Indicated fragments were PCR amplified, subcloned and sequenced by Sanger sequencing. Indels and SNVs that restore the ORF are indicated
by triangles and dots, respectively in indicated fragments of the construct that were sequenced (N = 3, n = 50). (e) High frequency of LacZ frameshift
mutations was recovered in the heterochromatic reporter in met-2 set-25 and msh-6 worms, but not in the euchromatic reporter. Results were
categorized according to the proportion of β-gal positive cells per worm (mean and s.e.m.; N = 3, n = 50). Scale bar, 50 µm.

© 2016Nature America, Inc. All rights reserved.
Nature GeNetics ADVANCE ONLINE PUBLICATION 9
ARTICLES
We next checked wild-type and met-2 set-25 genomes for copy
number variations (CNVs) in repeat families with multiple members.
Two DNA transposons (HELITRON2 and HELITRONY4) and two
tandem repeats (MSAT1 and (TATCG)n) showed high CNV uniquely
in the met-2 set-25 substrains. In contrast, the RNA transposon
CEREP58 and the single-copy gene lmn-1, like telomeric repeats and
the rDNA, remained stable (Fig. 6d,e and Supplementary Fig. 6e,f)57.
RNA transposons, which also failed to accumulate R loops, did not
show CNV. We conclude that met-2 set-25 germ lines accumulated
indels at sites bearing H3K9me in wild-type strains as well as changes
in DNA-transposon and tandem-repeat copy number.
A reporter incurred frequent indels in H3K9me-deficient
somatic cells
This sequence analysis monitored stable germline changes in the
worm population, and selected against any mutation that would per-
turb meiotic genome transmission. To visualize the mutation rate in
somatic cells, we used a heterochromatic reporter with a lacZ gene
placed out of frame to the ATG start codon, generating multiple pre-
mature stop codons in the first 100 bp of the transcript. Insertions
or deletions between the ATG and the ORF are necessary to enable
the translation of the lacZ mRNA into a functional β-galactosidase
enzyme (Fig. 7a)58. This allowed us to compare mutation rates of
wild-type and met-2 set-25 worms by microscopy, following a colori-
metric stain for heat-shock-induced β-galactosidase expression. By
comparing two time points during somatic development (12 h and
24 h after L1) we could differentiate mutations that might have been
present in the fertilized egg from mutations incurred during somatic
development (Fig. 7b). To compare the mutation rate of repetitive het-
erochromatic and unique euchromatic sequences, we made use of the
observation that transgenes integrated as high–copy number arrays
induce the formation of H3K9me-containing heterochromatin59
(i.e., enriched for H3K9me2 and H3K9me3; Fig. 7c). We compared
this reporter with the same reporter construct integrated as a low–
copy number array, which remains unmethylated and euchromatic.
We classified phenotypes by the extent of β-galactosidase expression
on a worm-by-worm basis, and sequenced the constructs amplified
from worms at the 24-h time point (Fig. 7d,e).
In the wild-type background after 24 h of cultivation, the hete-
rochromatic reporter produced functional β-galactosidase in only
around 3 ± 2% (s.e.m.) of the worms (Fig. 7e; >1/3 expressing in-frame
lacZ). In the met-2 set-25 mutant, the fraction of worms expressing
in-frame lacZ increased to 78 ± 8% (>1/3 staining blue, P = 0.01). In
contrast, the euchromatic reporter did not express in-frame lacZ in
met-2 set-25 worms (1 ± 1%, any level of in-frame lacZ). We used a
mutant of the mismatch repair machinery msh-6 as a positive con-
trol60. The met-2 set-25 mutant primarily showed an increase in the
β-galactosidase-positive phenotype by 24 h, and not by 12 h, unlike
the msh-6 mutant (Fig. 7e), suggesting that the met-2 set-25-induced
mutations occurred during differentiation. The types of mutations
monitored were confirmed by batch-wise cloning of single reporter
units and Sanger sequencing. We indeed detected small insertions and
deletions in the met-2 set-25 worms, enabling in-frame translation
of β-galactosidase (Fig. 7d). Thus, like the germline changes scored
by genome sequencing, sequences with H3K9me in wild-type back-
grounds accumulated indels at high rates during somatic cell division
in H3K9me-deficient worms.
DISCUSSION
H3K9 methylation is the defining histone modification for herit-
ably silent chromatin and is conserved as such from fission yeast to
humans. C. elegans mutants lacking H3K9me are viable, despite the
enrichment of H3K9me2/H3K9me3 on silent tissue-specific genes, on
pseudogenes and on RE. In contrast to the case in other species18,19,61,
we found no defects in chromosome segregation upon loss of
H3K9me. However, we observed a temperature-dependent sterility,
which coincided with an increase in DNA-damage-induced apoptosis
and stochastic delays in development. Correlating with these pheno-
types, we detected derepression of ~20% RE, from all repeat classes, a
value that increased at elevated temperatures. Expression of these RE
was accompanied by the accumulation of RNA:DNA hybrids, CNV
and a hypersensitivity to replication stress. This correlation suggests
that it is either the transcription of the repetitive sequence alone, or
transcription coupled with the inherent pairing nature of repeats,
that generates insertions and deletions within REs in the absence of
H3K9me. Of note, DNA transposons and tandem repeats showed
higher levels of R loops and CNV than RNA transposons, although
all classes were derepressed upon loss of H3K9me.
The damage incurred in the germ line leads to extensive apop-
tosis and Rad51 focus accumulation, suggesting that these cells
accumulated double-strand breaks as well as indels. There may be
additional sources of damage in the germ line, other than those that
correlate with replication-fork-associated damage, R loops and indels
scored by genome sequencing. We note that RNA polymerase-DNA
polymerase collision has been reported to generate fragile sites
of breakage44,62, which in worm germline cells would provoke an
apoptotic response32,63.
It is likely that the genomic mutations we detected in the met-2
set-25 strain arise from replication fork perturbation. This can be
triggered by enhanced stalling of the replication fork generated by R-
loop formation, which in turn allows hairpin or fold-back structures
to form in repeats as they are being replicated. Hairpin or fold-back
structures can also arise from breaks in the single-stranded DNA that
accumulate either at R loops or behind the fork, owing to perturbed
coordination between leading- and lagging-strand polymerases
(Fig. 8). The passage of the replication fork through REs itself can
lead to hairpin structures41. However, we propose that in met-2 set-25
cells, unprogrammed transcription of REs enhances R loops, which
may in turn enhance aberrant structures to such a degree that the
cellular machineries that normally relieve such stress, can no longer
cope with their abundance.
Repetitive element
H3K9me
RNAP II
Replication fork
Wild-type
met-2 set-25
Repetitive element
Secondary
structures
RNAP II
Replication fork
R loops
Collision
Figure 8 Transcribed REs in H3K9me-deficient strains can exacerbate
replication stress provoking genomic instability. A model illustrates
how the loss of H3K9me could lead to the formation of secondary DNA
structures that engender replication stress specifically at heterochromatic
repeats, to perturb genome integrity.

© 2016Nature America, Inc. All rights reserved.
10 ADVANCE ONLINE PUBLICATION Nature GeNetics
ARTICLES
Unscheduled collisions of the replication and transcription machin-
eries appear to generate breaks as well as other forms of genome insta-
bility42–46,62. Damage is often attributed to the presence of RNA:DNA
hybrids64–67, yet torsional stress, which can arise from high levels
of bidirectional transcription68,69, may also contribute to genomic
instability. We consider it notable not only that the derepression of
RE generated genomic mutations and R loops, but that both these
events mapped to REs that are normally marked by H3K9me in wild-
type cells, and which became derepressed in a temperature-enhanced
manner in the met-2 set-25 mutant. We propose that the crucial role
of H3K9 methylation in suppressing transcription on a genome-wide
level is not to program cell differentiation, but to stabilize repetitive
sequences that accumulate in higher eukaryotic genomes.
Several studies have suggested the use of inhibitors for H3K9me
HMTs in the treatment of cancer (for example, lung, prostate, hepa-
tocellular and pancreatic cancer)19,61, and preclinical studies have
been considered promising so far70. These same inhibitors have been
used to show that hypomethylation of H3K9 increases the rate of
induced pluripotent stem cell generation15,16. We argue that there
are clear drawbacks to such therapies, given the genomic instability
provoked by loss of H3K9me shown here. Whereas mammals addi-
tionally silence through meCpG, it has been documented that DNA
methylation can be targeted by H3K9me or its HMTs71. Thus, the
findings presented in this study are likely to have implications for
protocols that attempt to manipulate the mammalian epigenome.
URLs. http:/ /w ww.bioconductor.org /p ackages/3.1/bioc/html/
QuasR.html.
METHODS
Methods and any associated references are available in the online
version of the paper.
Accession codes. All data from this study have been deposited in the
Sequence Read Archive (SRA) under accession SRP080806.
Note: Any Supplementary Information and Source Data files are available in the
online version of the paper.
ACKNOWLEDGMENTS
A number of strains were provided by the Caenorhabditis Genetics Center
(CGC), which is funded by NIH Office of Research Infrastructure Programs (P40
OD010440). We thank R. Ciosk and P. Pasero for reagents and materials, I. Katiç and
members of the Friedrich Miescher Institute Genomics and Microscopy facilities
for advice and discussion, and P. Ginno and L. Constantino for advice on R-loop
detection. J.P. is supported by a long-term EMBO fellowship. S.M.G. thanks the Swiss
National Science Foundation as well as the Novartis Research Foundation for support.
AUTHOR CONTRIBUTIONS
P.Z. and J.P. planned and executed most experiments, evaluated results and wrote
the paper; S.M.G. planned experiments, evaluated results and wrote the paper;
R.v.S. and M.T. helped with evaluation of genome mutations and provided the
LacZ mutagenesis assay; and V.K. provided invaluable technical help.
COMPETING FINANCIAL INTERESTS
The authors declare no competing financial interests.
Reprints and permissions information is available online at http://www.nature.com/
reprints/index.html.
1. Bannister, A.J. et al. Selective recognition of methylated lysine 9 on histone H3
by the HP1 chromo domain. Nature 410, 120–124 (2001).
2. Lachner, M., O’Carroll, D., Rea, S., Mechtler, K. & Jenuwein, T. Methylation of
histone H3 lysine 9 creates a binding site for HP1 proteins. Nature 410, 116–120
(2001).
3. Nakayama, J., Rice, J.C., Strahl, B.D., Allis, C.D. & Grewal, S.I. Role of histone
H3 lysine 9 methylation in epigenetic control of heterochromatin assembly. Science
292, 110–113 (2001).
4. Allshire, R.C., Javerzat, J.P., Redhead, N.J. & Cranston, G. Position effect variegation
at fission yeast centromeres. Cell 76, 157–169 (1994).
5. Cam, H.P. et al. Comprehensive analysis of heterochromatin- and RNAi-mediated
epigenetic control of the fission yeast genome. Nat. Genet. 37, 809–819
(2005).
6. Lehnertz, B. et al. Suv39h-mediated histone H3 lysine 9 methylation directs DNA
methylation to major satellite repeats at pericentric heterochromatin. Curr. Biol.
13, 1192–1200 (2003).
7. Peters, A.H. et al. Loss of the Suv39h histone methyltransferases impairs mammalian
heterochromatin and genome stability. Cell 107, 323–337 (2001).
8. Eissenberg, J.C. et al. Mutation in a heterochromatin-specific chromosomal protein
is associated with suppression of position–effect variegation in Drosophila
melanogaster. Proc. Natl. Acad. Sci. USA 87, 9923–9927 (1990).
9. Epsztejn-Litman, S. et al. De novo DNA methylation promoted by G9a prevents
reprogramming of embryonically silenced genes. Nat. Struct. Mol. Biol. 15,
1176–1183 (2008).
10. Wissmann, M. et al. Cooperative demethylation by JMJD2C and LSD1 promotes
androgen receptor–dependent gene expression. Nat. Cell Biol. 9, 347–353
(2007).
11. Chen, M.W. et al. H3K9 histone methyltransferase G9a promotes lung cancer
invasion and metastasis by silencing the cell adhesion molecule Ep-CAM. Cancer
Res. 70, 7830–7840 (2010).
12. Hua, K.T. et al. The H3K9 methyltransferase G9a is a marker of aggressive ovarian
cancer that promotes peritoneal metastasis. Mol. Cancer 13, 189 (2014).
13. Kumari, D. & Usdin, K. The distribution of repressive histone modifications on
silenced FMR1 alleles provides clues to the mechanism of gene silencing in fragile
X syndrome. Hum. Mol. Genet. 19, 4634–4642 (2010).
14. Lee, J. et al. Epigenetic regulation of cholinergic receptor M1 (CHRM1) by histone
H3K9me3 impairs Ca2+ signaling in Huntington’s disease. Acta Neuropathol. 125,
727–739 (2013).
15. Loh, Y.H., Zhang, W., Chen, X., George, J. & Ng, H.H. Jmjd1a and Jmjd2c histone
H3 Lys 9 demethylases regulate self-renewal in embryonic stem cells. Genes Dev.
21, 2545–2557 (2007).
16. Shi, Y. et al. A combined chemical and genetic approach for the generation of
induced pluripotent stem cells. Cell Stem Cell 2, 525–528 (2008).
17. Treangen, T.J. & Salzberg, S.L. Repetitive DNA and next-generation sequencing:
computational challenges and solutions. Nat. Rev. Genet. 13, 36–46 (2011).
18. Dodge, J.E., Kang, Y.K., Beppu, H., Lei, H. & Li, E. Histone H3-K9 methyltransferase
ESET is essential for early development. Mol. Cell. Biol. 24, 2478–2486
(2004).
19. Kondo, Y. et al. Downregulation of histone H3 lysine 9 methyltransferase
G9a induces centrosome disruption and chromosome instability in cancer cells.
PLoS One 3, e2037 (2008).
20. Tachibana, M. et al. G9a histone methyltransferase plays a dominant role in
euchromatic histone H3 lysine 9 methylation and is essential for early embryogenesis.
Genes Dev. 16, 1779–1791 (2002).
21. Mellone, B.G. et al. Centromere silencing and function in fission yeast is
governed by the amino terminus of histone H3. Curr. Biol. 13, 1748–1757
(2003).
22. Peng, J.C. & Karpen, G.H. H3K9 methylation and RNA interference regulate
nucleolar organization and repeated DNA stability. Nat. Cell Biol. 9, 25–35
(2007).
23. Andersen, E.C. & Horvitz, H.R. Two C. elegans histone methyltransferases
repress lin-3 EGF transcription to inhibit vulval development. Development 134,
2991–2999 (2007).
24. Towbin, B.D. et al. Step-wise methylation of histone H3K9 positions heterochromatin
at the nuclear periphery. Cell 150, 934–947 (2012).
25. Harris, J. et al. Mutator phenotype of Caenorhabditis elegans DNA damage
checkpoint mutants. Genetics 174, 601–616 (2006).
26. Bernard, P. et al. Requirement of heterochromatin for cohesion at centromeres.
Science 294, 2539–2542 (2001).
27. Batista, P.J. et al. PRG-1 and 21U-RNAs interact to form the piRNA complex
required for fertility in C. elegans. Mol. Cell 31, 67–78 (2008).
28. Ketting, R.F., Haverkamp, T.H., van Luenen, H.G. & Plasterk, R.H. Mut-7 of
C. elegans, required for transposon silencing and RNA interference, is a homolog
of Werner syndrome helicase and RNaseD. Cell 99, 133–141 (1999).
29. Sijen, T. & Plasterk, R.H. Transposon silencing in the Caenorhabditis elegans germ
line by natural RNAi. Nature 426, 310–314 (2003).
30. Zhou, Z., Hartwieg, E. & Horvitz, H.R. CED-1 is a transmembrane receptor that
mediates cell corpse engulfment in C. elegans. Cell 104, 43–56 (2001).
31. Greiss, S., Schumacher, B., Grandien, K., Rothblatt, J. & Gartner, A. Transcriptional
profiling in C. elegans suggests DNA damage dependent apoptosis as an ancient
function of the p53 family. BMC Genomics 9, 334 (2008).
32. Gartner, A., Milstein, S., Ahmed, S., Hodgkin, J. & Hengartner, M.O. A conserved
checkpoint pathway mediates DNA damage–induced apoptosis and cell cycle arrest
in C. elegans. Mol. Cell 5, 435–443 (2000).
33. Schumacher, B., Hofmann, K., Boulton, S. & Gartner, A. The C. elegans homolog
of the p53 tumor suppressor is required for DNA damage–induced apoptosis.
Curr. Biol. 11, 1722–1727 (2001).
34. Gerstein, M.B. et al. Integrative analysis of the Caenorhabditis elegans genome by
the modENCODE project. Science 330, 1775–1787 (2010).
35. Liu, T. et al. Broad chromosomal domains of histone modification patterns in
C. elegans. Genome Res. 21, 227–236 (2011).

© 2016Nature America, Inc. All rights reserved.
Nature GeNetics ADVANCE ONLINE PUBLICATION 11
ARTICLES
36. Cangiano, G. & La Volpe, A. Repetitive DNA sequences located in the terminal
portion of the Caenorhabditis elegans chromosomes. Nucleic Acids Res. 21,
1133–1139 (1993).
37. Padeken, J., Zeller, P. & Gasser, S.M. Repeat DNA in genome organization and
stability. Curr. Opin. Genet. Dev. 31, 12–19 (2015).
38. Das, P.P. et al. Piwi and piRNAs act upstream of an endogenous siRNA pathway
to suppress Tc3 transposon mobility in the Caenorhabditis elegans germline. Mol.
Cell 31, 79–90 (2008).
39. Tabara, H. et al. The rde-1 gene, RNA interference, and transposon silencing in
C. elegans. Cell 99, 123–132 (1999).
40. Wicker, T. et al. A unified classification system for eukaryotic transposable elements.
Nat. Rev. Genet. 8, 973–982 (2007).
41. Aguilera, A. & García-Muse, T. Causes of genome instability. Annu. Rev. Genet. 47,
1–32 (2013).
42. Azvolinsky, A., Giresi, P.G., Lieb, J.D. & Zakian, V.A. Highly transcribed RNA
polymerase II genes are impediments to replication fork progression in Saccharomyces
cerevisiae. Mol. Cell 34, 722–734 (2009).
43. Deshpande, A.M. & Newlon, C.S. DNA replication fork pause sites dependent on
transcription. Science 272, 1030–1033 (1996).
44. Hoffman, E.A., McCulley, A., Haarer, B., Arnak, R. & Feng, W. Break-seq reveals
hydroxyurea-induced chromosome fragility as a result of unscheduled conflict
between DNA replication and transcription. Genome Res. 25, 402–412 (2015).
45. Merrikh, H., Machón, C., Grainger, W.H., Grossman, A.D. & Soultanas, P. Co-
directional replication–transcription conflicts lead to replication restart. Nature 470,
554–557 (2011).
46. Wahba, L., Amon, J.D., Koshland, D. & Vuica-Ross, M. RNase H and multiple RNA
biogenesis factors cooperate to prevent RNA:DNA hybrids from generating genome
instability. Mol. Cell 44, 978–988 (2011).
47. Nakamori, M., Pearson, C.E. & Thornton, C.A. Bidirectional transcription stimulates
expansion and contraction of expanded (CTG)*(CAG) repeats. Hum. Mol. Genet.
20, 580–588 (2011).
48. Prak, E.T. & Kazazian, H.H. Jr. Mobile elements and the human genome. Nat. Rev.
Genet. 1, 134–144 (2000).
49. Boguslawski, S.J. et al. Characterization of monoclonal antibody to DNA.RNA and
its application to immunodetection of hybrids. J. Immunol. Methods 89, 123–130
(1986).
50. Castellano-Pozo, M., García-Muse, T. & Aguilera, A. The Caenorhabditis elegans
THO complex is required for the mitotic cell cycle and development. PLoS One 7,
e52447 (2012).
51. Castellano-Pozo, M. et al. R loops are linked to histone H3 S10 phosphorylation
and chromatin condensation. Mol. Cell 52, 583–590 (2013).
52. Balk, B. et al. Telomeric RNA–DNA hybrids affect telomere-length dynamics and
senescence. Nat. Struct. Mol. Biol. 20, 1199–1205 (2013).
53. Wahba, L., Costantino, L., Tan, F.J., Zimmer, A. & Koshland, D. S1-DRIP-seq
identifies high expression and polyA tracts as major contributors to R-loop formation.
Genes Dev. 30, 1327–1338 (2016).
54. Rosenkranz, H.S. & Levy, J.A. Hydroxyurea: a specific inhibitor of deoxyribonucleic
acid synthesis. Biochim. Biophys. Acta 95, 181–183 (1965).
55. Gan, W. et al. R-loop-mediated genomic instability is caused by impairment of
replication fork progression. Genes Dev. 25, 2041–2056 (2011).
56. Zhang, H. & Freudenreich, C.H. An AT-rich sequence in human common fragile site
FRA16D causes fork stalling and chromosome breakage in S. cerevisiae. Mol. Cell
27, 367–379 (2007).
57. Collins, J., Saari, B. & Anderson, P. Activation of a transposable element in the
germ line but not the soma of Caenorhabditis elegans. Nature 328, 726–728
(1987).
58. Pothof, J. et al. Identification of genes that protect the C. elegans genome against
mutations by genome-wide RNAi. Genes Dev. 17, 443–448 (2003).
59. Towbin, B.D., Meister, P., Pike, B.L. & Gasser, S.M. Repetitive transgenes in
C. elegans accumulate heterochromatic marks and are sequestered at the nuclear
envelope in a copy-number- and lamin-dependent manner. Cold Spring Harb. Symp.
Quant. Biol. 75, 555–565 (2010).
60. Tijsterman, M., Pothof, J. & Plasterk, R.H.A. Frequent germline mutations and
somatic repeat instability in DNA mismatch-repair-deficient Caenorhabditis elegans.
Genetics 161, 651–660 (2002).
61. Wagner, T. & Jung, M. New lysine methyltransferase drug targets in cancer.
Nat. Biotechnol. 30, 622–623 (2012).
62. Helmrich, A., Ballarino, M. & Tora, L. Collisions between replication and transcription
complexes cause common fragile site instability at the longest human genes.
Mol. Cell 44, 966–977 (2011).
63. Vermezovic, J., Stergiou, L., Hengartner, M.O. & d’Adda di Fagagna, F. Differential
regulation of DNA damage response activation between somatic and germline cells
in Caenorhabditis elegans. Cell Death Differ. 19, 1847–1855 (2012).
64. Chan, Y.A., Hieter, P. & Stirling, P.C. Mechanisms of genome instability induced
by RNA-processing defects. Trends Genet. 30, 245–253 (2014).
65. El Hage, A., Webb, S., Kerr, A. & Tollervey, D. Genome-wide distribution of RNA–
DNA hybrids identifies RNase H targets in tRNA genes, retrotransposons and
mitochondria. PLoS Genet. 10, e1004716 (2014).
66. Kilchert, C., Wittmann, S. & Vasiljeva, L. The regulation and functions of the nuclear
RNA exosome complex. Nat. Rev. Mol. Cell Biol. 17, 227–239 (2016).
67. Skourti-Stathaki, K., Proudfoot, N.J. & Gromak, N. Human senataxin resolves RNA/
DNA hybrids formed at transcriptional pause sites to promote Xrn2-dependent
termination. Mol. Cell 42, 794–805 (2011).
68. Kim, N. & Jinks-Robertson, S. Transcription as a source of genome instability.
Nat. Rev. Genet. 13, 204–214 (2012).
69. Kouzine, F., Levens, D. & Baranello, L. DNA topology and transcription. Nucleus
5, 195–202 (2014).
70. Yuan, Y. et al. Gossypol and an HMT G9a inhibitor act in synergy to induce cell
death in pancreatic cancer cells. Cell Death Dis. 4, e690 (2013).
71. Liu, S. et al. Setdb1 is required for germline development and silencing of
H3K9me3-marked endogenous retroviruses in primordial germ cells. Genes Dev.
28, 2041–2055 (2014).

© 2016Nature America, Inc. All rights reserved.
Nature GeNetics doi:10.1038/ng.3672
ONLINE METHODS
C. elegans cultures and strains. Supplementary Table 2 lists the strains used
in this study. Strains were made by backcrossing deletion alleles and reporter
strains obtained from the C. elegans knockout consortium to the GW638 strain
(met-2(n4256) III; set-25(n5021) III) at least five times. Worms were grown
at 20 °C, except where specifically indicated.
Immunofluorescence analysis, antibodies and live microscopy, includ-
ing apoptosis assay. IF analysis was carried out as previously described24
by freeze-cracking and fixation in 1% paraformaldehyde followed by short
postfixation in methanol (for embryos and gonads72) or methanol followed by
acetone (for larval stages). Staining was performed in PBS with 0.1% TritonX-
100 and 2% milk powder. For live-cell imaging, larvae were mounted on
slides coated with 2% agarose. Microscopy was carried out on a spinning disc
confocal microscope (SD1, W1, Visitron, Puchheim). Stacks of images were
analyzed using ImageJ.
Antibodies used in this study were mouse anti-H3K9me2, MABI0317
(MBL73), mouse anti-H3K9me3, MABI0318 (MBL73), mouse anti-RNA:DNA
hybrid S9.6, hybridoma HB-8730 (ATCC)49, rabbit anti-pan-acetyl H4, 06-866
(Merck Millipore) and rabbit anti-RAD-51, 29480002 (Novu Biologics).
Developmental timing, progeny size and hatching rate. Worms of indicated
genotype were synchronized through bleaching and were then singled onto
plates containing OP50 bacteria. For the developmental timing their stage
was determined every 24 h. In order to determine the progeny size, adults
were transferred to fresh plates once a day for three days to keep generations
separate and their complete progeny size was determined after their hatching
at the indicated temperature. To determine the hatching rate singled worms
were transferred every 8 h to freshly seeded plates. The number of laid embryos
was determined directly after transfer, the number of hatched animals was
determined on day 3. If not otherwise indicated, worms were grown at the
experimental temperature (transferred from 20 °C) for at least two generations
before the experiments.
Chromatin immunoprecipitation experiments. Early embryonic progeny
was harvested after synchronization (60–65 h depending on each strain) for
wt and met-2 set-25 mutant strains in two independent biological replicates.
H3K9me2 and H3K9me3 ChIP was performed as previously described74 using
the antibodies mentioned above. In brief, 40 µg of chromatin was incubated
overnight with 3–6 µg of antibody coupled to Dynabeads Sheep Anti-Mouse
IgG (Invitrogen) or Dynabeads Sheep Anti-Rabbit IgG (Invitrogen), in FA
buffer (50 mM HEPES/KOH pH 7.5, 1 mM EDTA, 1% Triton X-100, 0.1%
sodium deoxycholate, 150 mM NaCl)) containing 1% SDS. Chromatin-
antibody complexes were washed with the following buffers: 3 × 5 min FA
buffer; 5 min FA buffer with 1 M NaCl; 10 min FA buffer with 500 mM NaCl;
5min with TEL buffer (0.25 M LiCl, 1% NP-40, 1% sodium deoxycholate,
1 mM EDTA, 10 mM Tris-HCl, pH 8.0) and twice for 5 min with TE. Complexes
were eluted at 65 °C in 100 µl of elution buffer (1% SDS in TE with 250 mM
NaCl) for 15 min. Both input and IP samples were incubated with 20 µg of
RNAse A for 30 min at 37 °C and 20 µg of proteinase K for 1 h at 55 °C.
Crosslinks were reversed overnight at 65 °C. DNA was purified using a Zymo
DNA purification column (Zymo Research).
Library preparation and analysis. Libraries were prepared from chromatin
IP and genomic DNA samples using the NEBNext ultra DNA library prep kit
for Illumina (NEB # 7370) and the NEBNext Multiplex Oligos for Illumina
(NEB # E7335), according to the manufacturer’s recommendations. No size
selection was performed during sample preparation and the libraries were
indexed and amplified using 12 PCR cycles, using the recommended condi-
tions. After further purification with Agencourt AmPure XP beads (Beckman
# A63881), the librar y size distribution and concentrations were determined
using a BioAnalyzer 2100 (Agilent technologies) and Qubit (Invitrogen)
instrument, respectively. The final pools were prepared by mixing equimolar
amounts of all individually indexed libraries and then sequenced on a HiSeq
2500 (Illumina) in rapid mode (Paired-End 50). Processing of the LEM-2
ChIP-seq data, all paired-end ChIP-seq data (2 × 50 bp) were mapped to the
C. elegans genome (ce6) with the R package QuasR75 using the included aligner
bowtie76. Definitions of REs were taken from Repbase77. Repeat subfamilies
were built to allow assignment of multimapping reads to all REs and collapsing
single elements according to their Repbase ID into families.
Read density along the genome was calculated by tiling the genome into
200-bp windows (non-overlapping) and counting the number of sequence frag-
ments within each window, using the qCount function of the QuasR package
(see URLs). To compensate for differences in the read depths of the various
libraries, we divided each sample by the total number of mapped reads and
multiplied by the average library size. Log2 expression levels were calculated
after adding a pseudocount of 1 (y = log2(x + 1)). ChIP-seq signals are dis-
played as average enrichment of IP − input (log2).
RNA expression experiments (RNA-seq and qPCR). For embryos and
larvae, RNA was isolated by freeze cracking (four times) followed by phenol-
chloroform extraction and isopropanol precipitation. Total RNA was depleted
for rRNA using Ribo-Zero Gold kit from Epicentre before library production
using Total RNA Sequencing ScriptSeq kit. Gonad RNA was extracted from
50 prepared gonads per replica using the Arcturus pico pure RNA isolation kit
followed by library production using the Total RNA-seq NuGen Ovation kit.
50-bp single-end sequencing was done on an Illumina HiSeq 2500. Processing
of the RNA-seq data, gene and repeat expression levels from RNA-seq data
were quantified as described previously24 using WormBase (WS190) annota-
tion for coding transcripts and Repbase annotations for REs. Primers used for
qPCR experiments are listed in Supplementary Table 3.
Mutation sequencing experiments. Worms were grown at 25 °C for 1 month,
and singled every second generation. Afterward worms were expanded on
peptone-rich plates (20 cm) per replica and mixed staged worms were har-
vested for genomic DNA isolation. DNA was extracted by a standard protocol
digesting worms with proteinase K, followed by phenol/chloroform extraction
and RNAse treatment. 50-bp paired-end sequencing was done on an Illumina
HiSeq 2500. Reads were mapped to the WS190 genome using BWA78 and
converted into BAM files using samtools79. Breakpoints were identified with
Pindel80 and SNVs with samtools.
Southern blot. Southern blot was performed following a standard protocol
using a digoxigenin-labeled probe produced by PCR with primers listed in
Supplementar y Table 3.
LacZ mutator assay and cloning for somatic mutations. LacZ mutator assay
was adapted from ref. 58. Worms were synchronized and grown for indicated
durations on Dh5α containing plates. After a heat shock (heterochromatic
array: 5 h (2 h at 33 °C, 1 h at 20 °C and 2 h at 33 °C), euchromatic array: 1 h
20 min (20 min at 33 °C, 10 min at 20 °C, 20 min at 33 °C, 10 min at 20 °C
and 20 min at 33 °C) and 2 h recovery at 20 °C, worms were stained for
β-galactosidase expression. To identify somatic mutations the indicated regions
of the reporter were amplified using a Q5 proofreading polymerase (NEB) and
primers listed in Supplementary Table 3. PCR products were batch clones
into pCR2.1-TOPO sequencing vector (Invitrogen) and Sanger sequencing
was performed on 20 clones per replica and region.
DNA damage sensitivity assays. Assays were previously described81. Recovery
from an hydroxyurea (HU) pulse was monitored by soaking L1 larvae in M9
buffer containing indicated concentrations of HU and OP50 bacteria for 16 h
before washing and plating on fresh OP50 plates. At day 3, the percentage of
viable adults was quantified. To quantify IR sensitivity worms were irradiated
(CellRad, Faxitron) at the L1 stage. At day 3, the percentage of viable adults
was quantified.
R-loop detection. For dot plots, genomic DNA was isolated using phenol-
chloroform extraction followed by ethanol precipitation. DNA concentra-
tions were determined using Nanodrop and the indicated amount of DNA
was resuspended to a final volume of 50 µl in nuclease-free water after either
a 1 h incubation with 5 µl of RNase H (NEB; +RNAse H), or a 1 h mock
incubation at 37 °C, and spotted directly onto a nylon GeneScreen Plus mem-
brane (NEF988; PerkinElmer) using a Bio-Dot Microfiltration Apparatus
(Bio-Rad). The membrane was UV-crosslinked and blocked with 5% milk in

© 2016Nature America, Inc. All rights reserved.
Nature GeNetics
doi:10.1038/ng.3672
1 × PBS/0.1% Tween-20 before incubation with primary and secondary antibod-
ies. The mouse S9.6 antibody (HB-8730, ATCC, gift of P. Pasero, Montpellier)
was used at a 1:500 dilution, and a 10,000× dilution of goat anti-mouse HRP
(Bio-Rad) was used as the secondary. The HRP signal was developed with
Clarity Western ECL Substrate (Bio-R ad). Imaging was performed using
ImageQuant LAS4000 mini und analyzed using ImageJ. Immunofluorescence
staining was performed according to the steps described for the C. elegans
larval stages using 4XSSC-T (0.1% Tween-20) instead of PBS-T. R loops
were stained with the S9.6 antibody, diluted 1:100 in SSC-T and 3% BSA
overnight at 4 °C.
DNA:RNA hybridization. Embryos were lysed by bead beating (MP
BIOMEDICALS FastPrep-24 5G Instrument) in G2 buffer (80 mM guanidine
HCl, 30 mM Tris pH 8.0, 30 mM 5% Tween, 0.5% TritonX). Genomic DNA
was isolated by proteinase K digestion, followed by purification using genomic
tips (500/G, QIAGEN). DNA was digested over night with AseI and BstUI at
37 °C (ref. 82). For RNase H control samples, RNase H was added in parallel to
digestion. 5 µg of digested DNA per IP was incubated with 10 µl of S9.6 anti-
body overnight in binding buffer (10 mM NaPO4, 140 mM NaCl, 0.5% Triton).
Bound DNA fragments were recovered with 50 µl of Protein-A Dynabeads
(Invitrogen), followed by four washes with binding buffer and proteinase K
treatment. Samples were purified using DNA Clean & Concentrator-5 (Zymo
Research) columns. Samples were sonicated to ~400-bp fragments before
library preparation, as described above.
72. Meister, P., Towbin, B.D., Pike, B.L., Ponti, A. & Gasser, S.M. The spatial dynamics
of tissue-specific promoters during C. elegans development. Genes Dev. 24,
766–782 (2010).
73. Kimura, H., Hayashi-Takanaka, Y., Goto, Y., Takizawa, N. & Nozaki, N. The
organization of histone H3 modifications as revealed by a panel of specific
monoclonal antibodies. Cell Struct. Funct. 33, 61–73 (2008).
74. Ikegami, K., Egelhofer, T.A., Strome, S. & Lieb, J.D. Caenorhabditis elegans
chromosome arms are anchored to the nuclear membrane via discontinuous
association with LEM-2. Genome Biol. 11, R120 (2010).
75. Gaidatzis, D., Lerch, A., Hahne, F. & Stadler, M.B. QuasR: quantification and
annotation of short reads in R. Bioinformatics 31, 1130–1132 (2015).
76. Langmead, B., Trapnell, C., Pop, M. & Salzberg, S.L. Ultrafast and memory-efficient
alignment of short DNA sequences to the human genome. Genome Biol. 10, R25
(2009).
77. Jurka, J. et al. Repbase Update, a database of eukaryotic repetitive elements.
Cytogenet. Genome Res. 110, 462–467 (2005).
78. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows–Wheeler
transform. Bioinformatics 25, 1754–1760 (2009).
79. Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics
25, 2078–2079 (2009).
80. Ye, K., Schulz, M.H., Long, Q., Apweiler, R. & Ning, Z. Pindel: a pattern growth
approach to detect break points of large deletions and medium sized insertions
from paired-end short reads. Bioinformatics 25, 2865–2871 (2009).
81. Craig, A.L., Moser, S.C., Bailly, A.P. & Gartner, A. Methods for studying the DNA
damage response in the Caenorhabdatis elegans germ line. Methods Cell Biol. 107,
321–352 (2012).
82. Ginno, P.A., Lott, P.L., Christensen, H.C., Korf, I. & Chédin, F. R-loop formation is
a distinctive characteristic of unmethylated human CpG island promoters. Mol. Cell
45, 814–825 (2012).