of environmental information
in C. elegans
Adam Klosin,1,2 Eduard Casas,3 Cristina Hidalgo-Carcedo,1,2
Tanya Vavouri,3,4 Ben Lehner1,2,5*
The environment experienced by an animal can sometimes influence gene expression
for one or a few subsequent generations. Here, we report the observation that a
temperature-induced change in expression from a Caenorhabditis elegans heterochromatic
gene array can endure for at least 14 generations. Inheritance is primarily in cis with the
locus, occurs through both oocytes and sperm, and is associated with altered
trimethylation of histone H3 lysine 9 (H3K9me3) before the onset of zygotic transcription.
Expression profiling reveals that temperature-induced expression from endogenous
repressed repeats can also be inherited for multiple generations. Long-lasting epigenetic
memory of environmental change is therefore possible in this animal.
Resident animals are not the only ones sub- ject to their environment; their progeny can also be affected (1–11). For example, star- vation or exposure to high temperature in Caenorhabditis elegans can lead to altered
small RNA transmission and putative target mRNA
expression for up to three generations (12, 13), and
a few temperature-induced expression changes
have been detected for two generations in animals with an inactive nuclear RNA interference
(RNAi) pathway (14). In contrast, gene silencing initiated by exogenous double-stranded RNA
(dsRNA) or piwi-interacting RNAs (piRNAs) can
sometimes be stably inherited between generations (15–19).
When we subjected C. elegans to high temperature (25°C), expression from daf-21 (Hsp90)
promoter::fluorescent protein constructs was strongly elevated (fig. S1). Expression from a single-copy
transgene was still elevated in the progeny of
animals transferred to 20°C after five generations at 25°C but not in their descendants (fig.
S1A). In contrast, expression from an integrated
multicopy array took 14 generations to return to
basal levels after the temperature was reduced
after 5 generations at 25°C (Fig. 1A and fig. S1).
A single generation of growth at 25°C was sufficient to generate a seven-generation memory of
increased expression (fig. S1C). Multigeneration
inheritance of temperature-induced expression
and a transgene-dependent phenotype was also
observed with other high-copy arrays (table S1).
mRNA transcribed from a daf-21 promoter array is first detected in wild-type (WT) worms at
the 16-cell stage of development; this confirms
no maternal supply of mRNA to the embryo (fig.
S2) (20). Expression differences inherited from
parents reared at different temperatures or sorted
according to their expression were apparent from
the onset of zygotic transcription (Fig. 1B and fig.
S3), and genetic crosses demonstrated inheritance
through both oocytes (Fig. 1C) and sperm (Fig. 1D).
The array is therefore inherited in an inactive state
but poised for a specific level of activation that
reflects expression in the previous generation.
To distinguish whether inheritance occurs in
cis with the DNA locus or in trans—for example,
in the cytoplasm—we crossed worms with high
and low expression to each other and then crossed
the resulting F1 male progeny to WT hermaphrodites (fig. S4) (20). The bimodal distribution of
expression in the F2 progeny indicates that the
major mode of inheritance is in cis with the locus
(Fig. 1E) (21).
To investigate chromatin modifications as potential mediators of this inheritance, we quantified histone modifications on the array in early
embryos developing at 20°C whose grandparents
had developed at either 16° or 25°C (Fig. 2A).
Embryos whose grandparents developed at 25°C
had less of the repressive histone modification
H3K9me3 on the array than embryos whose
grandparents developed at 16°C (Fig. 2, A and B).
This difference was apparent in early embryos
before the onset of zygotic transcription, indicating that the altered chromatin is not a secondary
response to altered transcription in the embryo
(Fig. 2, A and B). No differences were observed in
the Polycomb-associated repressive modification
trimethylated histone 3 lysine 27 (H3K27me3) or
in H3K36me3 and H3K4me2, two modifications
associated with active chromatin (Fig. 2B and fig.
S5). The differences in H3K9me3 were maintained
in late embryos after the onset of transcription
No mRNA expression from the array was detected in the adult germ line (fig. S7). However,
H3K9me3 was reduced on the array in the germline nuclei of adults that had been transferred
from 16° to 25°C as embryos (Fig. 2, C and D,
and fig. S8). Therefore, high temperature during germline development results in depletion
of H3K9me3 from the array, even though there
is no production of stable transcripts in this
The putative histone methyltransferase, SET-
25, is responsible for all detectable H3K9me3 in
C. elegans embryos (22) (fig. S5B), colocalizes with
H3K9me3-enriched transgenic arrays within embryonic nuclei (22), and is required for the maintenance of piRNA-initiated stable gene silencing
(15). Inactivating set-25 increased expression from
the array, with no difference in expression between animals maintained at 20° or 25°C (Fig. 3,
A and B). Hence, the repression of the array at
low temperature requires SET-25. Moreover, no
differences in expression were observed between
the F1 offspring of set-25 hermaphrodites mated
with male animals transmitting an array with
either high or low expression (Fig. 3B). In contrast, the inactivation of seven other small RNA
pathway or chromatin components (including
a Polycomb mutant mes-2) showed no obvious defects in the transmission of the expression memory
(fig. S9). Even after >20 generations of growth at
a constant temperature, substantial variation in
transgene expression is observed in both WT
and set-25 mutant populations (Fig. 3C). In WT
animals, these differences are transmitted to the
next generation (Fig. 3C), but this is not the case
in set-25 mutants (Fig. 3C).
Our results suggest a simple model for how the
transgene array shows memory of high-temperature
exposure that endures for many generations
(fig. S10). High temperature inhibits SET-25–
mediated repression in the germ line, causing
loss of H3K9me3 from the array. This dere-pressed chromatin is transmitted to subsequent
generations, resulting in increased expression
when transcription initiates in somatic lineages.
Over multiple generations of growth at low temperature, repression is gradually restored by
heterochromatin remodeling in each germline
cycle. This is consistent with previously reported
gradual quantitative intergenerational changes
in H3K9me3 following a temperature change at
some loci (14).
We tested whether this model predicts the behavior of endogenous loci in the genome by sequencing RNA from set-25 mutants and WT
animals at 20° and 25°C and from WT animals
three generations after a change from 25° to 20°C.
For protein-coding genes, derepression in set-25
mutants provided weak prediction of increased
expression at high temperature (fig. S11), consistent with a larger contribution from other regulators,
such as specific transcription factors. Derepression
in set-25 mutants was, however, a better predictor
of increased expression at high temperature for
multiple classes of repetitive elements and also
for pseudogenes (Fig. 4A and figs. S11 to S14),
consistent with impaired SET-25 activity’s making
320 21 APRIL 2017 • VOL 356 ISSUE 6335 sciencemag.org SCIENCE
1EMBL-CRG Systems Biology Unit, Centre for Genomic
Regulation (CRG), European Molecular Biology Organization,
08003 Barcelona, Spain. 2Universitat Pompeu Fabra, 08003
Barcelona, Spain. 3Program for Predictive and Personalized
Medicine of Cancer, Institute Germans Trias I Pujol, Campus
Can Ruti, 08916 Badalona, Barcelona, Spain. 4Josep Carreras
Leukaemia Research Institute, 08916 Badalona, Barcelona,
Spain. 5Institució Catalana de Recerca i Estudis Avançats,
08010 Barcelona, Spain.
*Corresponding author. Email: firstname.lastname@example.org (B.L.);