Transgenerational Epigenetic Inheritance: Prevalence, Mechanisms, and Implications for the Study of Heredity and Evolution

Abstract

This review describes new developments in the study of transgenerational epigenetic inheritance, a component of epigenetics. We start by examining the basic concepts of the field and the mechanisms that underlie epigenetic inheritance. We present a comprehensive review of transgenerational cellular epigenetic inheritance among different taxa in the form of a table, and discuss the data contained therein. The analysis of these data shows that epigenetic inheritance is ubiquitous and suggests lines of research that go beyond present approaches to the subject. We conclude by exploring some of the consequences of epigenetic inheritance for the study of evolution, while also pointing to the importance of recognizing and understanding epigenetic inheritance for practical and theoretical issues in biology.

Full text

TRANSGENERATIONAL EPIGENETIC INHERITANCE:
PREVALENCE, MECHANISMS, AND IMPLICATIONS FOR THE
STUDY OF HEREDITY AND EVOLUTION
Eva Jablonka
The Cohn Institute for the History and Philosophy of Science and Ideas, Tel-Aviv University, Tel-Aviv
69978, Israel
e-mail: jablonka@post.tau.ac.il
Gal Raz
The Graduate School of Medicine, Tel-Aviv University, Tel-Aviv 69978, Israel
e-mail: galraz@post.tau.ac.il
keywords
cell memory, epigenetics, induced heritable variations, Lamarckism,
microevolution, macroevolution
abstract
This review describes new developments in the study of transgenerational epigenetic inheritance, a
component of epigenetics. We start by examining the basic concepts of the field and the mechanisms
that underlie epigenetic inheritance. We present a comprehensive review of transgenerational cellular
epigenetic inheritance among different taxa in the form of a table, and discuss the data contained
therein. The analysis of these data shows that epigenetic inheritance is ubiquitous and suggests lines
of research that go beyond present approaches to the subject. We conclude by exploring some of the
consequences of epigenetic inheritance for the study of evolution, while also pointing to the importance
of recognizing and understanding epigenetic inheritance for practical and theoretical issues in biology.
Definitions: Epigenetics and
Epigenetic Inheritance
EPIGENETICS has become one of the
buzz words of biology in recent years.
Following the success of genome projects
in defining what genomes are, the empha-
sis has shifted to what they do, and there is
renewed interest in understanding the epi-
genetic processes of development. The
term “epigenetics,” however, has under-
gone many transformations since its origi-
nal definition by Waddington (see Wad-
dington 1968 for a discussion), reflecting
the changing foci of research in develop-
mental molecular biology since the second
half of the 20th century (Jablonka and
Lamb 2002, 2007c; Haig 2004; Holliday
2006). “Epigenetics” is therefore often em-
ployed loosely and inconsistently, and is
sometimes used as a synonym for “epige-
The Quarterly Review of Biology, June 2009, Vol. 84, No. 2
Copyright © 2009 by The University of Chicago Press. All rights reserved.
0033-5770/2009/8402-0001$15.00
Volume 84, No. 2June 2009THE QUARTERLY REVIEW OF BIOLOGY
131

netic inheritance.” To avoid misunder-
standings, we define both terms as they are
used in this review:
Epigenetics is the study of the processes
that underlie developmental plasticity and
canalization and that bring about persistent
developmental effects in both prokaryotes
and eukaryotes. At the cellular level, these
are the processes involved in cell determina-
tion and differentiation. At higher levels of
biological organization, epigenetic mecha-
nisms generate the context-dependent, self-
sustaining interactions between groups of
cells that lead to physiological and morpho-
logical persistence. The regulatory mecha-
nisms that establish and maintain variant cel-
lular and organismal states are known as
epigenetic control mechanisms,orepigenetic con-
trol systems (Nanney 1958).
Epigenetic inheritance is a component
of epigenetics. It occurs when phenotypic
variations that do not stem from variations
in DNA base sequences are transmitted to
subsequent generations of cells or organ-
isms. Many of the discoveries about epige-
netic inheritance between organisms are
derived from studies in developmental bi-
ology that look at inheritance in cell lin-
eages within an organism. Cell heredity in
mitotically dividing cells underlies the per-
sistence of determined states in multicellu-
lar organisms. That is, an individual’s kid-
ney stem cells and skin stem cells generally
breed true, even though their DNA se-
quences are identical and the developmen-
tal stimuli that led to the different cell
phenotypes are long gone. However, the
same cell heredity mechanisms that have
been found in cell lineages during develop-
ment are also observed when epigenetic in-
heritance occurs between generations of in-
dividuals. In single-celled organisms, such as
bacteria and asexually reproducing protists,
epigenetic inheritance leads to the clonal
persistence of induced and stochastically
generated phenotypic variations. In sexually
reproducing organisms, heritable epigenetic
variations in germline cells can result in the
transmission of developmentally induced
and stochastically generated phenotypes
from one generation of individuals to the
next through the gametes. Like “epigenet-
ics,” “epigenetic inheritance” is not always
consistently employed. It is used in both a
broad and a narrow sense.
Epigenetic inheritance in the broad sense is the
inheritance of developmental variations
that do not stem from differences in the
sequence of DNA or from persistent induc-
ing signals in the present environment. As
well as cell-to-cell transmission of epige-
netic variations in unicellular and multicel-
lular organisms (as will be explained be-
low), the definition covers body-to-body (or
soma-to-soma) information transference that
can take place through developmental inter-
actions between mother and offspring (e.g.,
Weaver et al. 2004), through social learning
(Avital and Jablonka 2000), and through
symbolic communication (Richerson and
Boyd 2005).
Cellular epigenetic inheritance is a narrower
aspect of epigenetic inheritance as dis-
cussed in the broad sense. It refers to epi-
genetic transmission in sexual or asexual
cell lineages, and the unit of this trans-
mission is the cell. Following Holliday
(1994, 2002, 2006), some biologists restrict
the term epigenetic inheritance solely to
the transmission of chromatin marks and
RNAs (e.g., Wu and Morris 2001). How-
ever, yeast geneticists use the term epige-
netic inheritance to refer to the inheri-
tance of protein conformations, such as
prions (e.g., Uptain and Lindquist 2001),
and the term is also used by biologists
studying self-sustaining loops and chroma-
tin inheritance in bacteria (e.g., Grandjean
et al. 1998; Laurent et al. 2005). We there-
fore define cellular epigenetic inheritance
as the transmission from mother cell to
daughter cell of variations that are not the
result of differences in DNA base sequence
and/or the present environment. Trans-
mission can be through chromatin marks,
through RNAs, through self-reconstructing
three-dimensional structures, and through
self-sustaining metabolic loops (Jablonka
et al. 1992; Jablonka and Lamb 1995, 2005,
2007a). It occurs following cell division in
prokaryotes, mitotic cell division in the
soma of eukaryotes, and sometimes follow-
ing the meiotic divisions in the germline.
The chromatin and RNA-mediated cellular
inheritance systems seem to play an impor-
tant role in inheritance through the germ-
line in both females and males.
132 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

In Figure 1, we illustrate the difference
between the broad and narrow sense of
epigenetic inheritance by showing the
main routes of between-generation trans-
mission in a sexually reproducing multicel-
lular organism. As our focus in this review
is on cellular transgenerational epigenetic inher-
itance, we are concentrating on the route of
between-generation transmission that in-
volves a single-cell “bottleneck,” i.e., trans-
mission through a gamete or a spore in
multicellular, sexually reproducing organ-
isms, or through a single sexual or asexual
cell in unicellular organisms. The environ-
ment may induce epigenetic variation by
directly affecting the germline or by affect-
ing germ cells through the mediation of
the soma, but, in either case, subsequent
transmission is through the germline. Al-
though the direct soma-to-soma transmis-
sion route of epigenetic variations is of
great importance for both development
and evolution (Jablonka and Lamb 2005,
2007a,b), its discussion is beyond the scope
of this review.
Mechanisms of Cellular Epigenetic
Inheritance (EISs)
Jablonka and Lamb (1989, 1995; Jablonka et
al. 1992) suggested that the different mecha-
nisms of epigenetic inheritance should be
understood and studied within a shared evo-
lutionary framework that incorporates the
developmental construction of heredity and
that acknowledges the Lamarckian aspects of
heredity and evolution. They called the pro-
cesses and mechanisms that underlie cellular
inheritance “epigenetic inheritance systems”
(abbreviated to EISs by Maynard Smith
[1990]). Cell heredity may occur when an
induced gene-product is diluted very slowly
by cell divisions, so that its concentration
remains above the threshold required for its
activity for several cell generations (Zachari-
oudakis et al. 2007), but such “memory” is
short-term. For more persistent memory and
cell heredity, autocatalysis is necessary, and
all the EISs we describe depend on mecha-
nisms that enable self-perpetuation. Four
types of cellular EISs are recognized today:
the EIS based on self-sustaining regulatory
loops, the EIS based on three-dimensional
Figure 1. Routes of Transmission of Epigenetic Variations in a Multicellular, Sexually Reproducing Organism
Route ashows the germline-to-germline transmission of induced epigenetic variations (e.g., chromatin marks). A variation can be induced in the germline and
can then be transmitted from one generation to the next, or it can first be induced in the soma, then affect the germline, and thereafter be inherited through the
germline. Route bshows soma-to-soma transmission (for example, through the transmission of symbionts and parasites, or through the self-perpetuating effects of
maternal behavior, social learning, and symbolic communication). A broad view of epigenetic inheritance encompasses both routes aand b, whereas the narrower,
cellular view includes only route a—transmission through a single-cell “bottleneck,” in this case, a gamete.
June 2009 133TRANSGENERATIONAL EPIGENETIC INHERITANCE

templating, the chromatin-marking EIS, and
the RNA-mediated EIS. Although these EISs
are usually seen as very different types of de-
velopmental mechanisms, all can contribute to
between-generation epigenetic inheritance.
Their dual nature as developmental mecha-
nisms and as inheritance systems means that
they can be studied from both perspectives.
The developmental perspective is currently
the dominant one, and the molecular basis
of epigenetic control systems is one of the
most intensely studied fields in biology today
(e.g., see collections of articles in Nature Re-
views Genetics 8(4) [Flintoft 2007], Nature
447(7143 Insight) [Campbell 2007], and Cell
128(4) [Marcus 2007], as well as Allis et al.
2007). In this review, we focus on the less
discussed aspect of EISs and epigenetic in-
heritance—the transmission of cellular epi-
genetic states from one generation of organ-
isms to the next. We therefore describe the
four EISs from a heredity-focused point of
view. All the examples to which we refer, as
well as many others, are presented here in
Table 1 and in our expanded table (available
online, with accompanying table references,
at The Quarterly Review of Biology homepage,
www.journals.uchicago.edu/QRB).
inheritance through self-sustaining
loops
Self-sustaining feedback loops are metabolic
circuits through which different patterns
of activity can be maintained, resulting in
alternative heritable cell phenotypes. The
first experimental studies of such loops
were those involving the bistability of the
lac operon of Escherichia coli (Novick and
Weiner 1957), and this system has since
been thoroughly analyzed at both the mo-
lecular and theoretical levels (Laurent
et al. 2005). These studies show that, when
inducer concentrations are low, genetically
identical cells can generate two alternative,
true-breeding, stable phenotypes.
Many other self-sustaining feedback loops
leading to alternative heritable phenotypes
have been described in bacteria and other
taxa (Dubnau and Losick 2006; Smits et al.
2006; Malagnac and Silar 2003, 2006). One
well-characterized example of positive regu-
lation leading to alternative cell phenotypes
is found in the fungus Candida albicans,
where an epigenetic switch underlies the
transition between white and opaque
cells—two states that are heritable for many
generations (Zordan et al. 2006). Self-
sustaining loops need not be based on tran-
scriptional regulation; they can also occur at
the post-translation stage, through protein
self-processing. An example is the enzyme
vacuolar protease B of Saccharomyces cerevisiae,
whose active form is necessary for its own
conversion from an inactive precursor to an
active state. On glycerol media, where the
precursor is synthesized in high amounts,
the self-processing of the enzyme is indefi-
nitely self-sustaining (Roberts and Wickner
2003). As noted by Wickner et al. (2004), it
is likely that other protein-modifying pro-
teins may also directly or indirectly affect
their own modification and behave as self-
sustaining, cell-transmissible loops.
structural inheritance:
spatial templating
Structural inheritance refers to the inher-
itance of alternative three-dimensional
(3-D) structures through spatial templat-
ing: a variant 3-D structure in a mother cell
guides the formation of a similar structure
for a daughter cell, leading to the transmis-
sion of the architectural variant (Nanney
1968). The study of cellular structural in-
heritance was initiated by investigations of
the inheritance of cortical variations in
ciliates. Beisson and Sonneborn (1965)
showed that an experimentally modified
organization of the cilia on Paramecium can
be inherited through many asexual and
sexual generations. The inheritance of cor-
tical variations induced by various physical
and chemical manipulations has also been
demonstrated for other ciliates (Nanney
1985; Frankel 1989; Grimes and Aufder-
heide 1991).
The propagation of prions is another
form of structural inheritance. The funda-
mental characteristic of a prion—a trans-
missible protein—is that it has a conforma-
tion that can initiate and sustain the re-
production of a similar conformation in
newly synthesized proteins. Prusiner (1998)
suggested that infectious proteins are the
134 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

causative agents of mammalian neurode-
generative diseases such as kuru, scrapie,
and mad cow disease and coined the term
prion (proteineceous infectious particles)
for such agents. He advanced and devel-
oped the concept of an infectious protein
whose propagation is based on the spatial-
templating of the variant prion conforma-
tion, which, in turn, converts a normally-
shaped protein into its own shape. The
concept was extended when Wickner (1994)
characterized the [URE3] variant in yeast
as a prion and explained its idiosyncratic
biochemical and hereditary characteristics
in terms of protein-based inheritance.
Since then, more prions have been identi-
fied in yeast and other fungi (Baxa et al.
2006; Tuite and Cox 2006). Although some
prions have deleterious effects, others
seem to have important biological func-
tions (Wickner et al. 2004; Shorter and
Lindquist 2005; Benkemoun and Saupe
2006).
One of the most interesting findings
about prions is that a single protein can
misfold into several different conforma-
tions that have specific growth dynamics,
stabilities, pathologies, and cross-species
infectivity (Chien et al. 2004). Different
prions may interact, leading to the forma-
tion of many different transmissible (cell-
heritable and infectious) phenotypes. There-
fore, unicellular organisms, which have the
same genotype and live in the same en-
vironment, can exhibit heritably differ-
ent morphologies and physiologies that
are the consequence of differently folded
identical proteins. The differences result
from differences in the developmental
histories of their ancestors.
Cavalier-Smith (2004) studied another
aspect of spatial templating, an aspect
associated with membrane reproduction.
The reproduction of most membranes, in-
cluding the plasma membrane, the endo-
plasmic reticulum, and the mitochondrial
membrane, requires the presence and tem-
plating of pre-existing membrane struc-
tures. Cavalier-Smith identified 18 types of
what he called “genetic membranes,” the
structures of which are cell-heritable
through these guided processes, and he
argued that the information embedded in
this “membranome” is as essential for the
construction of a cell as genomic informa-
tion. He suggested that crucial events in
the evolution of cells and major groups
were associated with heritable changes in
membranomes.
the chromatin-marking EIS
Chromatin, the stuff of chromosomes,
includes DNA and all the factors physically
associated with it: small chemical groups
covalently attached to DNA (e.g., methyl
groups), bound histone and nonhistone
proteins, and associated RNA molecules.
The organization of chromatin and chro-
mosomes, their localization in the nucleus,
and the dynamic interactions among the
various components of chromatin deter-
mine when, where, and to what extent
genes are transcribed, how DNA repair is
orchestrated, how different chromosomal
domains are organized, and how chromo-
somes, as units, behave during the vari-
ous phases of the cell cycle. Patterns of
chromatin are reconstructed following
DNA transcription and replication, and, al-
though the processes of reconstruction are
not well understood, there is evidence that
chromatin variations can be transmitted
between generations of individuals. The
study of the chromatin-marking EIS is
therefore crucial for the understanding of
both development and heredity.
DNA methylation, the best-understood
system of chromatin inheritance, is an epi-
genetic modification found in Eubacteria,
Archeabacteria, and Eukaryota. It is in-
volved in many important functions (for
general reviews see Casadesu´s and Low
2006; Vanyushin 2006a,b), including de-
fence against genomic parasites (Kidwell
and Lisch 1997), regulation and mainte-
nance of gene activity patterns (Barlow and
Bartolomei 2007; Li and Bird 2007), stabi-
lization of chromosomal structure (Karpen
and Hawley 2007), and DNA replication
and repair (Mortusewicz et al. 2005; Scher-
melleh et al. 2007). In eukaryotes, methyl-
ation usually occurs on the cytosines in
CpG doublets and also in CpNpG triplets
in plants. Since CpG and CpNpG se-
June 2009 135TRANSGENERATIONAL EPIGENETIC INHERITANCE

quences are palindromic, the two strands
of the duplex are mirror images of each
other. Following replication, two hemi-
methylated duplexes are formed, and
these hemimethylated regions are recog-
nized by maintenance DNA methylases,
which preferentially add methyls to non-
methylated C in the new strand (Allshire
and Selker 2007; Henderson and Jacobsen
2007; Li and Bird 2007). The replication of
methylation patterns is thus semi-conservative.
The fidelity of transmission in cell lineages
ranges from about 1% variation per cell
generation up to the very high fidelity of
10-6 variations per cell generation (Gene-
reux et al. 2005; Richards 2006).
Variations in a DNA methylation pattern
can also be inherited between generations,
and examples of this include paramuta-
tions in plants (interaction between alleles
that leads to a directed epigenetic herita-
ble change in one of the interacting alleles
[Brink 1973; Chandler 2007]); silencing
of foreign duplicated sequences in fungi
(Allshire and Selker 2007); variations in
telomeric, centromeric, and rRNA regions
(Karpen and Hawley 2007); and variations
in transgenes and endogenous genes. Spe-
cific examples of all these types of trans-
generational inheritance are presented in
Table 1.
How a DNA methylation pattern affects
a cell’s or an organism’s phenotype de-
pends on the way it interacts with the pro-
tein components of the chromosome,
which are also heritable. In eukaryotes, im-
portant variations in chromatin are associ-
ated with histones, the proteins that make
up the nucleosome core around which
DNA is wrapped. The dynamic nature of
histones, their variability, and their associ-
ation with every conceivable cellular func-
tion (see Berger 2007; Groth et al. 2007;
Blasco 2007; Morris and Moazed 2007)
makes understanding the inheritance of
their specific structure and organization
both challenging and urgent. Several mod-
els of this process have been constructed,
and according to all of them, the nucleo-
some variants and the post-transcriptional
modifications (PTMs) of the parental nu-
cleosomes are used as blueprints for the
restoration of the same nucleosome con-
figuration. These reconstructions may be
assisted by other chromatin components
such as DNA methylation patterns (Martin
and Zhang 2007), RNAs (Grewal and Jia
2007; Ringrose and Paro 2007), and by the
location of the chromosomal domains
within the nucleus (Misteli 2007).
In addition to DNA methylation patterns,
as well as histone variants and their PTMs,
chromatin has various enzymes that associate
dynamically with the histones and DNA
bound to them and that participate in their
regulatory and structural functions (Bantig-
nies and Cavalli 2006). The patterns of asso-
ciation of these nonhistone proteins with
other chromatin components can also be
reconstructed between generations of cells
and organisms (Ringrose and Paro 2004;
Bantignies and Cavalli 2006; Schuettengru-
ber et al. 2007). Recent data suggest that
RNA produced at regions that bind non-
histone chromatin proteins may lead to the
inheritance of the bound state through
RNA-DNA pairing (Grewal and Jia 2007;
Ringrose and Paro 2007).
the RNA-mediated EIS
During the last decade, it has become
apparent that RNA is central to the regu-
lation of cellular dynamics in eukaryotes
and is also involved in cell and organism
heredity. Gene silencing by small RNAs—
RNA interference (RNAi)—has been found
in all eukaryote phyla from yeast to man,
although a few species (e.g., budding
yeast) seem to have lost the capacity for
this. In all of the RNAi pathways, double-
stranded RNA molecules (dsRNA), which
trigger the process, are chopped into
shorter dsRNAs (usually between 21-30 nu-
cleotides long) by the enzyme Dicer. Af-
ter the original dsRNA is chopped, the re-
sulting siRNA (small interfering RNA) is
loaded onto a complex of proteins, one
strand of the duplex is removed, and the
other strand—the guide strand—directs si-
lencing. Silencing may occur through any
of the following mechanisms: (i) the siRNA
is loaded onto an enzyme complex that
interferes with the transcription or transla-
tion of mRNAs with a homologous se-
136 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

quence (Cullen 2004; Meister and Tuschl
2004); (ii) the siRNA is loaded onto an
enzyme complex that targets chromatin re-
gions with DNA that is homologous to the
siRNA, and alters chromatin into a silent
state (Matzke and Birchler 2005; Ekwall
2007; Huettel et al. 2007); or (iii) the
siRNA is loaded onto an enzymatic com-
plex that degrades and/or excises the DNA
sequences complementary to the siRNAs.
The latter processes of RNA-regulated
DNA rearrangement are being intensely
studied in ciliates (Meyer and Chalker
2007; Nowacki et al. 2008).
RNA can affect cell and organism hered-
ity in at least three different, non-mutually
exclusive ways. The first is the result of rep-
lication of siRNA through RNA-dependent
RNA polymerase. This replication is a two-
stage process, the details of which differ
between taxa, and it leads to the amplifica-
tion of siRNAs that act as repressors of
gene expression (Baulcombe 2007; Pak and
Fire 2007). These siRNAs are transmitted
to daughter cells during cell division and
can migrate to other cells as well (e.g.,
Palauqui et al. 1997; Himber et al. 2003).
The second way in which RNAs can af-
fect cell heredity is by guiding, targeting,
and assisting in transmitting variations in
chromatin structure that are reconstructed
and reproduced in daughter cells through
the chromatin-marking EIS (Matzke and
Birchler 2005; Ekwall 2007; Huettel et al.
2007). The third way is by targeting DNA
base sequences and guiding changes in
them that are then replicated by DNA poly-
merases (Meyer and Chalker 2007; Nowacki
et al. 2008). Heritable variations can be gen-
erated and perpetuated through all three
routes of RNA-mediated heredity, and the
formation of a particular dsRNA may be af-
fected by local conditions and may be devel-
opmentally regulated. Once formed, such
dsRNA may have long-term hereditary ef-
fects.
EISs and the germline
It is obvious from the foregoing descrip-
tions that the different EISs are often
mechanistically and functionally interre-
lated; therefore, the division of EISs into
four categories is somewhat arbitrary and
artificial. However, dividing EISs in this
manner highlights the variety and com-
plexity of the cellular mechanisms of in-
heritance. The details of how epigenetic
variations are transmitted through mitosis
in asexual clonal lineages remain a puzzle,
but in sexually reproducing organisms,
especially multicellular ones, the puzzle is
even greater.
In sexually reproducing organisms, epi-
genetic variations have to survive the com-
plex process of meiosis in order to be trans-
mitted to the next generation, and, in
multicellular organisms, they also have to
survive gametogenesis and early embryo-
genesis—two developmental stages that in-
volve significant restructuring of both cells
and chromatin. Although there is as yet no
evidence that prions and self-sustaining
loops are transmitted between generations
through sperm and egg, there is evidence
that chromatin marks and RNAs can be
transmitted in this manner, but it is not
clear how this occurs. It seems likely that
some footprints of chromatin marks re-
main and lead to the reconstruction of
ancestral states, or that some remnants of
ancestral states (including some RNAs) are
retained. Even in male vertebrates, where a
comprehensive replacement of histones by
protamines takes place during gametogen-
esis, the erasure of histone marks is not
complete. In the mouse, for example,
about 1% of the DNA remains wrapped
around histones, and two acetylated vari-
ants of H4 are maintained through sper-
miogenesis (Van der Heijden et al. 2006).
Van der Heijden and his colleagues (2006)
suggested that these histones, as well as
methylated cytosines in centromeric DNA,
are associated with the transgenerational
maintenance of the structure of centro-
meric heterochromatin. Chong et al. (2007)
reported that when male mice had muta-
tions in genes involved in epigenetic pro-
gramming—in a gene that encodes a chro-
matin remodeler protein, and in another
that encodes DNA methyltransferase—
there were phenotypic effects on their
offspring, even when they did not inherit
the defective gene. Extensive epigenetic
June 2009 137TRANSGENERATIONAL EPIGENETIC INHERITANCE

(methylation) variation has been found in
human germ cells (Flanagan et al. 2006),
but whether and to what extent this varia-
tion is passed between generations is not
known.
In addition to some chromatin marks,
certain RNAs may be transmitted through
the germline. Cells in the germline contain
small RNAs, known as Piwi-associated inter-
fering RNAs (piRNAs), that are important
for the proper maturation of germ cells
(Kim 2006). One function of piRNAs may
be the detection and silencing—or dele-
tion—of regions of unpaired DNA during
meiosis. Unpaired regions are targets of
the RNAi machinery; RNA transcribed
from them guides enzymatic complexes to
the unpaired regions, which are then de-
leted or heterochromatinized (Shiu et al.
2001; Bean et al. 2004; Turner et al. 2005;
Costa et al. 2006). Mammalian spermato-
cytes are filled with piRNAs, and similar
RNAs have been discovered in oocytes too.
The abundance of piRNAs suggests that they
may be transmitted to the next generation
and lead to transgenerational effects. The
transmission of an epigenetic modification
in male mice (induced in heterozygotes for a
variant Kit gene) is known to be mediated
through microRNAs with sequences partially
complementary to that of Kit RNA (Rassoul-
zadegan et al. 2006; see supplementary
material for details, available online at The
Quarterly Review of Biology homepage, www
.journals.uchicago.edu/QRB).
Transgenerational Epigenetic
Inheritance: Prevalence,
Distribution, and Induction
There is no recent review of cellular epi-
genetic inheritance between generations
that encompasses all four types of EISs and
their distributions across taxa. The only
comprehensive survey was made by Jab-
lonka and Lamb in 1995. Since then, many
more cases have been described, and our
understanding of the molecular mecha-
nisms underlying epigenetic inheritance
has deepened and expanded. The table we
present here (Table 1) brings together
over one hundred cases of inherited epige-
netic variations in bacteria, protists, fungi,
plants, and animals. We have included only
cases where there is convincing evidence
for epigenetic inheritance in the narrow
cellular sense, either through a single asex-
ual cell (as in bacteria, some protists, some
fungi, and some plants) or through a sex-
ually generated gamete or spore (most an-
imals, plants, and fungi). In most (al-
though not all) cases, molecular studies
have revealed the involvement of one or
more of the EISs, but a full molecular char-
acterization through all the reproductive
stages is not yet available for any organism.
More details about the cases are given as
supplementary material in an expanded ta-
ble form (available online at The Quarterly
Review of Biology homepage, www.journals
.uchicago.edu/QRB).
How common is epigenetic inheritance?
This question is often raised, and the an-
swer, based upon the data we have accu-
mulated, is that it may be ubiquitous. We
believe that epigenetic variants in every lo-
cus in the eukaryotic genome can be inher-
ited, but in what manner, for how long,
and under what conditions has yet to be
qualified. In other words, unlike the repli-
cation of DNA variations, which is largely
context insensitive, whether and for how
long a particular mark or cellular element
is transmitted between generations de-
pends on genomic, developmental, and ex-
ternal conditions. This does not mean that
conditions that allow epigenetic inheri-
tance are particularly rare. In fungi, for
example, the widespread occurrence of
epigenetic inheritance is generally ac-
knowledged. As long ago as 1949, Linde-
gren, reviewing data on inheritance in Neu-
rospora, said that two thirds of new variants
do not show Mendelian segregation and
are therefore discarded, and, in their re-
cent review of protein inheritance in fungi,
Benkemoun and Saupe (2006) comment-
ed: “Many of us working with filamentous
fungi know how often bizarre looking sectors or
segregates that defy Mendelism appear on our
plates. As a number of pioneering fungal genet-
icists have done in the past, maybe we should
have a closer look before putting them in the
autoclave” (p. 793). They suggested that
many of these “anomalies” may be caused
138 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

by some form of epigenetic inheritance. In
all organisms, chromatin inheritance can
theoretically occur at every locus in the
genome, and the double-stranded nature
of DNA provides a theoretical possibility
(when transcription occurs from both
strands) for every DNA segment to form
small dsRNA molecules that can lead to
heritable silencing. The potential of most
proteins to form
␤
sheets with spatial
templating properties (Baxa et al. 2006)
suggests that a prion-like transfer of con-
formation between cells may occur more
frequently than previously thought.
cases included in the table
In Table 1, we show only a small sample
of representative cases of epigenetic inher-
itance in allopolyploid plant hybrids be-
cause, in all cases that have been investi-
gated, allopolyploidy is accompanied by
extensive epigenetic changes, some of
which are inherited between generations.
In most of the cases presented in the table,
the molecular basis of the EIS has been
unravelled, but we included a few cases for
which the evidence for cellular epigenetic
inheritance seemed overwhelming, despite
the fact that molecular studies were lacking
(e.g., inheritance of the star phenotype in
silver foxes, and some cases of develop-
mentally induced variability and inheri-
tance in plants and fungi).
As our main goal in this review is to show
that variations in epigenetic marks can be
inherited for several generations, we did
not include classical cases of genomic im-
printing in the table. With genomic im-
printing, the epigenetic marks that are im-
posed on parental chromosomes during
oogenesis differ from those imposed dur-
ing spermatogenesis; therefore, in the off-
spring, a gene’s expression pattern de-
pends on whether it was inherited from the
father or from the mother (see Barlow and
Bartolomei 2007 for genomic imprinting
in mammals, Alleman and Doctor 2000 for
flowering plants, and for general reviews of
imprinting that include invertebrates, see
Jablonka and Lamb 1995; Haig 2002). By
definition, the genomic imprints charac-
terizing one sex are reversed when chro-
mosomes go through gametogenesis in the
opposite sex in the next generation; thus,
imprints are inherently reversible. How-
ever, molecular studies of imprinting have
been important in the history of epigenetic
inheritance because they have led to a gen-
eral recognition of the role of epigenetic
control mechanisms, such as changes in
DNA methylation in critical regions, and
have opened the way to unraveling cases
that were similar to genomic imprinting
but that were not parent-sex specific (e.g.,
the early studies by Hadchouel et al.
1987 and Allen et al. 1990, as discussed
in Jablonka and Lamb 1995). The only
imprinting-related case included in the ta-
ble is that of an imprint in a human grand-
mother that was not erased in her son and
that subsequently led to Prader-Willi syn-
drome in her grandchildren (Buiting et al.
2003). This is a clear case of an epigenetic
imprinting mark being transmitted as a re-
sult of a fault in resetting. As yet, we are not
aware of any studies showing that new vari-
ations in genomic imprints are transmitted
over several consecutive generations.
In most of the cases that we included in
the table, the epigenetic variations have
been transmitted for more than two gen-
erations, and we can rule out repeated
direct induction of the variation in each
generation. However, when the induced
parent carries an embryo, three genera-
tions may be necessary to confirm epige-
netic inheritance. For example, when a fe-
male mammal is exposed to an inducing
agent during pregnancy, direct induction
of the embryo’s germ cells has to be ex-
cluded; therefore, three generations of
transmission (from F0to F3) are required
to establish that epigenetic inheritance,
rather than direct induction of the embryo
and its germline, has occurred (Jirtle and
Skinner 2007). On the other hand, when a
male mammal is exposed to an inducer,
two generations of transmission (from F0
to F2) are sufficient to establish that there
has been epigenetic transmission through
the germline rather than through direct
induction. When the epigenetic effects
(e.g., patterns of methylation) in the F1
and F2generations are identical, and when
June 2009 139TRANSGENERATIONAL EPIGENETIC INHERITANCE

TABLE 1
Transgenerational epigenetic inheritance in prokaryotes and eukaryotes
Taxon Trait Locus/cellular system Stability Inducing conditions EIS Reference
Bacteria and their viruses
␭
phage of
Escherichia
coli
Lysogenic/lytic
cycle
Cl and Cro Stable Nutritional state of the
host, phage density
Self-sustaining loops Ptashne (2005)
Bacillus subtilis Inactivation of
chromosome
Whole chromosome Stable Polyethylenglycol (PEG)-
induced fusion
Chromatin-marking Grandjean et al. (1998)
Lack of cell wall Balance between peptidoglycan
synthesis and destruction
Stable on agar Experimental removal of
cell wall
Structural inheritance Landman (1991)
Sporulation Spo0A phosphorelay Stable Nutritional deprivation Self-sustaining loops Veening et al. (2005)
Natural
competence
(K state)
ComK activity 10%–20% in lab
strains, 1% in the
wild
Stochastic, elevated by
stress
Self-sustaining loops Maamar and Dubnau (2005)
Escherichia coli Utilization of
lactosea
Lac operon activity Stable under
conditions of low
inducer
concentration
Stochastic, growth in low
concentration of
inducer
Self-sustaining loops Cohn and Horibata
(1959a,b), Laurent et al.
(2005), Novick and
Weiner (1957), Ozbudak
et al. (2004)
Fluffy Agn43 Phase variation Probably oxidative stress GATC methylation Casadesu´ s and Low (2006)
Pili Pap operon 10-4 per generation,
10-3 reversion
Changed carbon source,
temperature, and
spontaneous
GATC methylation Hernday et al. (2002)
Growth rate
(persister type
II)b
Probably many genes 10-6 per generation,
10-1 per
generation
reversion
Spontaneous and
antibiotic treatment
Self-sustaining loops Balaban et al. (2004), Lewis
(2007), N. Balaban
(personal
communication)
Resistance to
antibiotics
(ampicillin,
tetracycline,
and nalidixic
acid)
Altered regulation of
␤
lactamase cryptic gene,
glutamate gene, and
decarboxylase gene; possible
involvement of DNA
methylase genes
3%–20% survival
(depending on
concentration of
antibiotic) and
50% reversion rate
Low, and successively
increased
concentrations of
antibiotics
Possibly self-sustaining
loops and /or DNA
methylation
Adam et al. (2008)
Pseudomonas
aeruginosa
Toxin injection TTSS system Stable Cell density Self-sustaining loops Filopon et al. (2006)
140 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

Synecococcus
elongates
(Cyanobacteria)
Circadian rhythm Regulatory loop involving key
KaiC protein
Several days Light and dark pulses Self-sustaining loops Kondo and Ishiura (2000)
Protists
Oxytricha
trifallax
(Ciliate)
Alteration of gene
order: aberrant
rearrangements
Genes that become unscrambled
in the somatic macronucleus
(i.e., most genes)
Stable through asexual
reproduction; at
least 3 generations
following sexual
reproduction
Experimental
manipulation
RNA-mediated
rearrangement
Nowacki et al. (2008)
Paramecium
aurelia
(Ciliate)
Serotypes
expressed
Cilia proteins Stable Changes in pH,
temperature, food
supply, and salinity
Self-sustaining loops Landman (1991)
Induced tolerance
to heat, salt,
and arsenic
Not specified Inherited for many
generations before
fading away
gradually
Exposure to high
temperature, high salt,
and arsenic
concentrations
Not known Jollos (1921); reviewed in
Jablonka et al. (1992)
Paramecium
tetraurelia,
Stylonchia
lemnae,
Tetrahymena
thermophila
(Ciliates)c
Various traits
related to
alternative
genetic
organization
patterns in the
macronucleus
In principal, any DNA
sequence in the genome
Stable through
macronucleus
reproduction,
both during
asexual and
following sexual
reproduction
Sequence comparison of
maternal (old) and
zygotic (new) nucleus,
based on non-
expressed sequences,
leads to the
inheritance of
maternal, nucleus-
guided gene
organization
Inherited DNA
rearrangement/
editing/ programming
of the macronucleus
mediated by small
RNAs and chromatin
modifications
Garnier et al. (2004),
Juranek et al. (2005),
Liu et al. (2004), Meyer
and Chalker (2007),
Taverna et al. (2002),
Yao et al. (2003)
Cortical
organization
Basal body and cortex proteins Stable in mitosis and
sometimes in
meiosis
Experimental
manipulation, stress
Structural inheritance
(guided assembly)
Grimes and Aufderheide
(1991)
Plasmodium
falciparum
(Malaria
parasite)
Telomere
inactivation
Telomere sequences Switch every ⬃15
generations
Spontaneous Chromatin-marking Roberts et al. (1992)
continued
June 2009 141TRANSGENERATIONAL EPIGENETIC INHERITANCE

TABLE 1
Continued
Taxon Trait Locus/cellular system Stability Inducing conditions EIS Reference
Tetrahymena
(Ciliate)
Increased insulin
binding and
production
following exposure
to diiodotyrosine
(T2)
Not specified Very stable—hundreds
of generations
Induction by insulin or
diiodotyrosine
Probably methylation;
treatment with 5-
azacytidine can
abolish response
Csaba (2008), Csaba et al.
(1999), Csaba and Kovacs
(1990, 1995), Csaba et al.
(1982a,b)
Volvox carteri Silencing of in vitro
methylated
transgene
C-ars transgene More than 100
generations
DNA transformation DNA methylation Babinger et al. (2007)
Fungi
Ascobolus
immerses
Transgene
inactivation
Any duplicated transgene Stable Pre-meiotically DNA methylation Martienssen and Colot
(2001), Rhounim et al.
(1992)
Candida
albicans d
Cell morphology,
ability to form
colonies on various
substrates, mating
properties
Master regulator WOR1
protein
Switches every ⬃
10,000 generations
Spontaneous, affected by
temperature
Self-sustaining loop Huang et al. (2006),
Malagnac and Silar
(2003), Zordan et al.
(2006)
Coprinus
cinereus
Methylation pattern Centromere-linked locus Stable when highly
methylated
Unknown Chromatin-marking,
DNA methylation
involved
Zolan and Pukkila (1986)
Podospora
anserina
(Filamentous
fungus)
Crippled Growth
(CG)
C—assumed to be a
transmissible self-sustaining
cascade factor involving a Map
kinase module
Stable mitotic
inheritance
Transformation of normal
cells to CG cells induced
by cytoduction,
promoting stationary
state and growth on
medium supplemented
with yeast extract
Self-sustaining loop Kicka et al (2006), Malagnac
and Silar (2006), Silar et
al. (1999)
[Het-s*] and [Het-s]
variants affecting
vegetative
incompatibility
het-s Stable; low frequency
[Het-s] converted
to inactive [Het-
s*], but the
reverse also
occurs, albeit at
low frequency
Spontaneous Structural inheritance
(prion)
Maddelein et al. (2002)
142 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

Saccharomyces
cerevisiae
Enhanced resistance
to starvation on
dextrose- minimal
plates and
suppression of
inability to sporulate
in mutants
[
␤
]-prion form of the PrB
vacuolar protease
Indefinite
propagation
Emerges in
deletion-mutants;
induction is increased
as a result of PrB over-
expression
Prion, through self-
sustaining loop
Roberts and Wickner
(2003)
Growth phenotype FLO genes near telomere Switch every 10 to 15
generations
Spontaneous Chromatin inheritance Halme et al. (2004)
Read through [PS] Stable Various Structural inheritance
(prion)
Tuite and Cox (2006)
Nitrogen catabolic
gene expression
[URE3] Stable Spontaneous and over
expression
Structural inheritance
(prion)
Benkemoun and Saupe
(2006)
Ascus formation [PIN]⫹Rnq1p Stable Spontaneous and over
expression of Sup35
Structural inheritance
(prion)
Benkemoun and Saupe
(2006)
Expression of
experimentally
modified GAL
network
GAL network Stable Low inducer
concentrations
Self-sustaining loop Acar et al. (2005)
Induction of GAL1
and GAL7 by
galactose
GAL genes Up to 7 generations Induction by galactose Slow dilution of
abundant regulatory
GAL1 protein
Zacharioudakis et al. (2007)
Anti-suppressor ISP⫹?Stable Spontaneous Probably based on
structural inheritance
Benkemoun and Saupe
(2006), Kunz and Ball
(1977), Tallo´czy et al.
(2000), Volkov et al.
(2002)
Glucosamine
resistance
GR?
Control of killer
virus expression
KI-d? (suspected prions)
Slow growth and
additional
requirement for
leucine
Structural alteration of the
mitochondrion
More than 100
generations in the
absence of leucine
Spontaneous inherited
loss of mitochondrial
DNA
Structural inheritance Lockshon (2002)
continued
June 2009 143TRANSGENERATIONAL EPIGENETIC INHERITANCE

TABLE 1
Continued
Taxon Trait Locus/cellular system Stability Inducing conditions EIS Reference
Schizosaccharomyces
pombe
Survival of mutant
strain lacking the
hcd region-
encoding, highly
conserved domain
of the essential
chaperone
calnexin
[cif] Calnexin
independence
inherited by 88.7%
of spores
Mating cells lacking the
gene coding for
calnexin with calnexin-
dependent cells
Structural inheritance,
probably prion
Collin et al. (2004)
Meiotic telomere
clustering and
chromatin structure
Interaction of telomeric and
subtelomeric regions with Taz1
Stable Normal; variation shown
by deleting telomere
sequences in
chromosomes
Chromatin-marking,
structural inheritance
Sadaie et al. (2003)
Reporter transgene
silencing and
mating type
switching
K⌬::ura4 (transgene inserted at
K-region)
At least 30 mitotic
generations after
meiosis
Transgenically induced Chromatin-organizing
factors probably
involved
Grewal and Klar (1996)
Plants
Antirrhinum
majus
(Snapdragon)
Flush/granulated
phenotype
(paramutation)
nivea locus Variegated in F1,
more stable in F2
Induced by crossing Transposition suggested;
chromatin-marking
plausible
Krebbers et al. (1987)
Arabidopsis
thaliana
(Mouseear
cress)
Reversion of dwarf
morphology
Cpr1–1 gene-
epiallele interacting with bal
Effect seen in F2
progeny but not in
F1
Induced by crossing Unknown Stokes and Richards (2002)
Size, rosette, and
petiole
abnormalities;
activity of GFP
reporter gene
Transgenic loci E82 and L91
interacting with COP1
endogene
At least 5
generations
Induced by crossing Involvement of RNA-
mediated DNA
methylation suggested
Qin and von Arnim (2002)
Delayed flowering fwa (gene-epiallele) Very stable EMS, fast neutron
treatment, or ddm1
induced (Wild type
DDM1 required to
maintain DNA
methylation)
DNA methylation,
histone H3
methylation (siRNAs
involved)
Kinoshita et al. (2007),
Lippman et al. (2004),
Soppe et al. (2000)
144 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

Dwarfism, constitutive
activation of
defense response
pathway
Bal locus At least 5
generations
ddm1 effect even when
segregated out
Probably DNA
methylation
Stokes et al. (2002)
Number of
reproductive organs
SUP At least 2
generations
Various mutagens DNA methylation Jacobsen and Meyerowitz
(1997)
Expression levels of
the retrotransposon
At2g10410
At2g10410 epiallele At least 8
generations
Spontaneous, found in
natural populations
DNA methylation Rangwala et al. (2006)
Increased levels of
homologous
recombination in
soma
Not specified At least 4 generations
in case of UV-C
radiation; at least 2
generations in case
of introduced
flagellin
Ultraviolet radiation or
introduction of
flagellin
Not specified Molinier et al. (2006),
J. Molinier (personal
communication)
Loss of hygromycin
resistance
(paramutation)
hpt transgene No effect in F1;F
2
affected
Transgene-induced Involves DNA
methylation
Scheid et al. (2003)
Transcriptional
activity of
transposable
elements
Various transposable elements At least 6
generations
ddm1 effect, even after
segregated out
DNA methylation,
chromatin
modifications; RNAi
probably involved
Lippman et al. (2003, 2004),
V. Colot (personal
communication)
Blue fluorescence,
tryptophan and
IAA deficiency,
morphological
abnormalities
PAI (gene-epialleles) At least 6
generations
T-DNA mutagenesis,
crossing; variation
found in natural
populations
DNA methylation; DNA/
DNA pairing is
suggested
Bender and Fink (1995),
Luff et al. (1999)
Arabidopsis
Interspecfic
hybridse
Many traits Many loci, both coding and
noncoding
Stability varies
according to locus,
but many loci
show stable
inheritance
Induced by
hybridization followed
by polyploidization
Chromatin-marking,
DNA methylation;
RNAi system probably
also involved
Comai et al. (2000),
Scheid et al. (2003)
continued
June 2009 145TRANSGENERATIONAL EPIGENETIC INHERITANCE

TABLE 1
Continued
Taxon Trait Locus/cellular system Stability Inducing conditions EIS Reference
Beta vulgaris
(Sugar beet)
Many traits in mitotic,
agamospermic, and
inbred lines
including: single or
multiple flower
initiation, self-
fertility,
polymorphism of
malic enzyme, and
variation in number
of chloroplasts
Mm,li,Rf1, and Me1 loci Variable, often 2 or
more generations
Mode of reproduction,
climatic conditions,
direction of cross
DNA methylation is
probably involved
Levites (2000), Levites and
Maletskii (1999),
Maletskii (1999)
Linaria
vulgaris
(Common
toadflax)
Flower symmetry Lcyc (gene-epiallele) At least 2
generations
Natural variation in
populations
DNA methylation Cubas et al. (1999),
J. Parker (personal
communication)
Linum
usitatissimum
(Flax)
Plant weight, height,
peroxidase isozyme
pattern, seed
capsule septa hair
number
r-DNA genes and repetitive
sequences
Stable Fertilizer and heat
regimes
Methylation and DNA-
repatterning
Cullis (2005)
Flowering age, main
stem height at
maturity, and
number of leaves
Not specified, but epimutations
in at least 3 independent,
nonrandom loci are assumed
to be involved
At least 9
generations
5-azaC treatment DNA methylation,
possibly associated
chromatin
remodeling as well
Fieldes and Amyot (1999),
Fieldes et al. (2005)
Lycopersicon
esculentum
(Tomato)
Inhibition of ripening
and development of
a colorless pericarp
KeSPL-CNR (gene epiallele) Very stable Spontaneous
epimutation
DNA methylation Manning et al. (2006)
Color variegation
(Paramutation)
Sulf epialleles Very stable X-rays; regeneration Chromatin inheritance
suggested
Hagemann (1969, 1993),
Hagemann and Berg
(1977), Wisman et al.
(1993), R. Hagemann
(personal
communication)
146 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

Melandrium
album
(White
campion)
Bisexuality Decrease in CG methylation in
many loci
2 successive
generations,
inherited through
the male parent
Treatment with 5-
azacytidine
DNA methylation Janousˇek et al. (1996)
Nicotiana
tabacum
(Tobacco)
Loss of kanamycin
resistance;
paramutation-like
effects
Nospro and nos At least 2
generations
Transgenic silencing in
doubly transformed
plants
DNA methylation Matzke and Matzke (1991),
Matzke et al. (1989)
Loss of hygromycin
resistance
Hpt transgene 2 generations Transgenic silencing DNA methylation Park et al. (1996)
Requirement of leaf
cell for cytokinin
Not specified Arises at 10-2 per cell
generation
Subculturing in media
containing successively
lower concentrations
of cytokinin
Unknown; DNA
methylation suggested
Meins (1986, 1989a,b),
Meins and Thomas
(2003)
Oryza sativa
(Rice)
CpG methylation
patterns show
inherited cultivar
specificity
Methylation state of cytosine in
various CCGG sites across
the genome
At least 6
generations
Crossing Nipponbare
and Kasalath cultivars5
DNA methylation Ashikawa (2001)
Various traits Methylation state in various
CCGG sites across the
genome
At least 2
generations
Introgression and selfing5DNA methylation Dong et al. (2006)
Kernel shape and
tiller number
S2,S3, and various unspecified
methylation sites
3–6 generations of
selfing
High-pressure treatment
given to seeds
DNA methylation Shen et al. (2006)
Induced dwarfism;
resistance to
pathogen
General change in DNA
methylation; Xa21G promoter
At least 9
generations
Induced by 5-azaC DNA methylation
involved
Akimoto et al. (2007),
Sano et al. (1990)
Petunia hybrida
(Petunia)
White-flowering An3 epiallele and the dTph1
transposon
At least 3
generations of self-
fertilization
Induced by
heterozygosity
Unknown; epigenetic
interaction between at
least 3 dTph1 copies
Van Houwelingen et al.
(1999)
Triticum
aestivum
(Wheat)
Cytosine methylation Glutenin gene At least 2
generations
Inbreeding Chromatin-marking,
DNA methylation
Flavell and O’Dell (1990)
Longer productive
spikes, larger
seeds, and other
quantitative traits
Hl1 and pc loci 57 generations Nicotinic acid Unknown Bogdanova (2003)
continued
June 2009 147TRANSGENERATIONAL EPIGENETIC INHERITANCE

TABLE 1
Continued
Taxon Trait Locus/cellular system Stability Inducing conditions EIS Reference
Triticum
interspecific
synthetic
hybridse
Many traits ⬃13% of the genome Stability depends on
locus; many very
stable
Hybridization and
polyploidization
Chromatin and DNA
methylation; RNAi
system probably
involved
Levy and Feldman (2004)
Triticale
(Stable
wheat-rye
hybrid)
Increase in stature,
number of tillers,
changed time of
ripening
Not specified At least 2
generations
5-azaC treatment Chromatin-marking;
DNA methylation
involved
Heslop-Harrison (1990)
Zea mays
(Maize)
Reduced
pigmentation
(paramutation)
B1 (gene-epiallele) Very stable Paramutagenic B⬘epiallele
arises spontaneously
from B-I allele at a
frequency of ⬃1% to
10%
Chromatin inheritance;
siRNA-directed
chromatin modification
is involved
Alleman et al. (2006),
Chandler (2007),
Chandler et al. (2000),
Coe (1966), Stam et al.
(2002)
Reduced
pigmentation
(paramutation)
r1 (complex loci) Variable Spontaneous, and
exposure to varied
light durations and
temperature
DNA methylation Chandler et al. (2000),
Mikula (1995),Walker
and Panavas (2001)
Reduced, light-
dependent
pigmentation
(paramutation)
pl (gene-epiallele) Varies from
metastable to very
stable inheritance
Spontaneous Chromatin-marking
suggested
Chandler et al. (2000),
Hollick and Chandler
(1998), Hollick et al.
(1995, 2000)
Reduced pericarp
color but dark
pigmentation at
the point of silk
attachment
(paramutation)
p1 (gene-epiallele) At least 5
generations
Spontaneous DNA methylation Cocciolone et al. (2001), Das
and Messing (1994),
Rajandeep et al. (2007),
Sidorenko and Peterson
(2001), Sekhon et al.
(2007)
Spotting of kernels MuDR transposable element Silencing effect of the
Muk locus on MuDR
was maintained over
at least 4
generations, even
when Muk had
segregated away
Crossing leading to
silencing of the
regulatory transposon
MuDR
DNA methylation;
siRNAs are involved
in maintaining
heritable methylation
states in Mu1 and
MuDR elements
Lisch et al. (2002), Slotkin
et al. (2003), Slotkin et
al. (2005)
148 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

Animals
Caenorhabditis
elegans
(Nematode)
Small and dumpy
appearance
RNAi of ceh-13 Over 40 generations Feeding with bacteria
expressing dsRNA
targeting ceh-13
Chromatin remodeling;
RNAi-mediated
Vastenhouw et al. (2006),
N. Vastenhouw
(personal
communication)
Silencing of green
fluorescent
protein (GFP)
RNAi of gfp transgene At least 40
generations
Feeding with bacteria
expressing dsRNA
targeting gfp
Chromatin remodeling;
RNAi-mediated
Vastenhouw et al. (2006),
N. Vastenhouw
(personal
communication)
Various effects, not
reported
RNAi of 13 genes At least 10
generations
Feeding with bacteria
expressing dsRNA
targeting the 13 genes
Chromatin remodeling;
RNAi-mediated
Vastenhouw et al. (2006),
N. Vastenhouw
(personal
communication)
Daphnia pulex
(Water
flea)
Expression of G6PD
S and N variants
G6PD locus or its regulator Spontaneous
reversion rate
between the 2
forms was 1 in 10
and1in2.
Spontaneous and
glucose induced;
presence of S form
related to stressful
conditions
Not known Ruvinsky et al. (1983 a,b,
1986)
Diaphanosoma
celebensis
(Cladoceran)
Timing of
reproduction and
number of
offspring
Not specified 2 generations Induced by the natural
estrogen E2
Not known; parallel
effect on induced
parents and their
offspring
Marcial and Hagiwara
(2007)
Drosophila
melanogaster
(Fruit fly)
Modifying ability of
Y chromosome
Imprintor gene interaction 11 generations Transient effect of
imprintor gene
Chromatin marks Dorn et al. (1993)
Ectopic outgrowth
in eyes
Kr (KrIf-1 allele), vtd3(TrxG
mutation)
At least 13
generations
Geldanamycin treatment
given to KrIf-1 strain,
or transient presence
of TrxG mutation vtd3
Chromatin-marking
involved
Ruden et al. (2003),
Sollars et al. (2003)
Eye-color Transgenic Fab-7 flanking lacZ
and mini-white reporter
transgenes
At least 4
generations
Transient presence of
GAL-4- protein
Probably chromatin
inheritance
Cavalli and Paro (1998,
1999)
continued
June 2009 149TRANSGENERATIONAL EPIGENETIC INHERITANCE

TABLE 1
Continued
Taxon Trait Locus/cellular system Stability Inducing conditions EIS Reference
Eye-color (due to de-
repression of mini-
white reporter gene
cloned downstream
to transgenic Fab-7)
Activation state of endogenous
and transgenic Fab-7 elements
containing CMM
At least 4 generations
(more than 4 years)
High temperature Probably chromatin
inheritance
Bantignies et al. (2003),
F. Bantignies (personal
communication)
Suppression of wing
deformations
De-repression of sd Not specified (stability
lower and harder to
detect)
High temperature Probably chromatin
inheritance
Susceptibility to
tumorigenesis
Probably several loci, including
heritable epigenetic variation
in the ftz promoter
Increased tumorigenicity
(2 generations);
modified ftz
methylation (at least 1
generation)
Crossing with hopTum-l and
Kr1mutants
DNA methylation,
chromatin inheritance
Xing et al. (2007)
Ephestia
kuehniella
(Moth)
Reversion of
shortened antennae
and associated
mating disadvantage
Suppressor
of sa (saWT)
Up to 5 generations;
incompletely
inherited from
mother but almost
fully inherited from
father
Exposure of larva and
pupa to lithium ions,
alternate electrical field,
or 25°C at late 5th instar
larval and pupal phases
Probably chromatin
inheritance
Pavelka and Kondelova
(2001)
Homo sapiens
(Human)
Cardiovascular
mortality and
diabetes
susceptibility
Imprinted tandem repeat
upstream of INS-IGF2-H19
region
At least 2 generations Food availability during
childhood growth period
Possibly methylation;
transmitted through
male germ line
Kaati et al. (2002, 2007),
Pembrey (2002)
Angelman and Prader-
Willi syndromes
15q11-q13 Inherited from paternal
grandmother (no
imprint erasure in the
father)
Spontaneous DNA methylation Buiting et al. (2003), Zoghbi
and Beaudet (2007)
Mus musculus
(Mouse)
Probability of
developing yellow
coat color and
obesity, as well as
susceptibility to
diabetes and cancer
Avy (gene-epiallele) and
probably other loci
Metastable (at least 2
generations of agouti
epigenotype);
cumulative, three-
generation effect on
obesity
Spontaneous; affected by
diet
Chromatin-marking,
including DNA
methylation
Blewitt et al. (2006), Cropley
et al. (2006), Morgan et al.
(1999), Waterland et al.
(2007), Wolff et al. (1998),
Waterland et al. (2008)
150 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

Reduced body
weight, reduced
level of proteins
involved in sexual
recognition, and
possibly higher
mortality between
birth and weaning
Not specified, but epimutation
is connected to Major
Urinary Protein (MUP) and
Olfactory Marker Protein
(OMP) genes
Preliminary results
suggesting
transmission of
the traits to B2
(the second
generation)
Induced by transfer of
mouse pronuclei at
the one-cell stage to
eggs of a different
genotype; traits are
transmitted to most of
the offspring through
male germ line
DNA methylation
assumed to be
involved
Roemer et al. (1997)
Probability of kinked
tail shape
Axin-fused (gene-epiallele) and
IAP transposable element
Spontaneous rate of
inactivation 6%; rate
of reactivation 1%
Spontaneous; influenced
by diet. Injection of
hydrocortisone during
spermiogenesis
reduces penetrance
Chromatin-marking
including DNA
methylation
Belyaev et al (1981a, 1983),
Rakyan et al. (2003),
Waterland et al. (2006),
D. Martin (personal
communication)
White-spotted tail
and feet
Kit (paramutation) 2 generations of
outbred crossing;
6 generations of
inbred crossing
between
paramutants
Transient presence of
Kittm1Alf mutation
RNA inheritance; RNAi
involved
Rassoulzadegan
et al. (2006),
M. Rassoulzadegan
(personal
communication)
Repression of
recombination of
the LoxP element
and concomitant
methylation
Transgenic LoxP and
surrounding chromosomal
sequences
Methylated state
maintained for at
least 3 generations
Transient presence of
Sycp1-Cre; exposure of
wild type to
recombinase activity
DNA methylation Rassoulzadegan et al.
(2002)
Genome stability Many At least 3
generations
Irradiation Chromatin methylation Barber et al. (2002),
Dubrova (2003)
Glucose intolerance Not specified 2 generations, some
effect in the 3rd
generation
Betel nut ingestion Not known Boucher et al. (1994)
Tendency to develop
tumors
Elevated expression of gene
coding for LF (an estrogen-
responsive protein) and C-fos
Apparent in the F1
and F2generations
Induced by
diethylstilbestrol
during F0pregnancy
DNA methylation
probably involved
Newbold et al (2006)
Cardiac hypertrophy The microRNA miR-1 At least 3
generations
Microinjection of miR-1
into fertilized eggs
RNA inheritance Wagner et al. (2008)
continued
June 2009 151TRANSGENERATIONAL EPIGENETIC INHERITANCE

TABLE 1
Continued
Taxon Trait Locus/cellular system Stability Inducing conditions EIS Reference
Myzus persicae
(Peach potato
aphid)
Loss of insecticide
resistance Probably amplified resistance
genes Stable inheritance of
lost resistance in
clones that have
amplified DNA
Induced by
DNA amplification DNA methylation
involved Field et al. (1989)
Rattus norvegicus
(Rat) Modified serotonin
content in immune
cellsg
Not specified 2 generations Intramuscular administration
of
␤
-endorphin during
19th day of pregnancy
Not specified Csaba et al. (2005)
Increased expression
of genes coding for
metabolic factors
Promoters of PPAR
␣
and GR in
liver; increased expression of
other RNAs
At least 2 generations Protein-restricted diet during
pregnancy DNA methylation Burdge et al. (2007),
G. Burdge (personal
communication)
Decreased
spermatogenic
capacity; elevated
incidence of tumor,
prostate, and kidney
diseases, serum
cholesterol levels,
and immune system
abnormalities;
premature aging
and male mating
disadvantage
Methylation state of 15
different DNA sequences.
Reduced expression of
ankyrin 28,Ncstn,Rab12,
Lrrn6a and NCAM1 found in
vinclozolin group as well as
increased expression of Fadd,
Pbm1b,snRP1c and Waspip
At least 3 generations;
transmission through
the male germ line
Vinclozolin or methoxychlor
treatment during gestation DNA methylation Anway et al. (2005, 2006a,
2006b), Chang et al.
(2006), Crews et al. (2007)
Altered glucose
homeostasis Not reported F0–F3generations Low-protein diet in F0, from
day 1 of pregnancy
through lactation
Not specified, DNA
methylation probably
involved in F1animals
Benyshek et al. (2006),
D. C. Benyshek (personal
communication)
Vulpes vulpes
(Fox) Piebald spotting Activation state of Star gene Star (semidominant
allele) activated in ⬃
1% of domesticated
animals; inherited for
more than 2
generations
Spontaneous in tamed foxes
raised in fur farms;
hormonal stress suggestedf
Possibly heritable
chromatin
modification
Belyaev et al. (1981b),
Trut et al. (2004)
Note: The species in each category are ordered alphabetically. References to the table are available online at The Quarterly Review of Biology homepage, www.journals.uchicago.edu/QRB.
aSimilar systems have been described for arabinose-utilization in E. coli (Khlebnikov et al. 2000) and the lactose operon in Salmonella enterica typhimurium (Tolker-Nielsen et al. 1998).
bProbably occurs in many pathogens (Lewis 2007).
cIt seems that all ciliates show cortical inheritance and guided assembly of cortical structures (Frankel 1989). Ciliates have a silent micronucleus and an active macronucleus, from which
noncoding sequences are excised and coding sequences are amplified. Following conjugation or autogamy, a new (zygotic) macronucleus is formed from the fused meiotic product of the
micronucleus, and the old, maternal macronucleus degenerates. The complex processes, guided by the maternal nucleus, that lead to the inheritance of patterns of chromosomal
rearrangements in the zygotic nucleus seem to be characteristic of all ciliate species (Meyer and Chalker 2007).
dSimilar phenomena are found in C. galborata,C. tropicalis,C. parpsilosis, and in the basiomycetes fungus Crystococcus neoformans.
eInterspecific plant hybrids, and sometimes hybrids between cultivars, display epigenetic variations that are heritable across generations. When hybridization is followed by polyploidiza-
tion, this seems to be a normal and invariant genomic response. In The Biological Journal of the Linnean Society 82(4) (Allen 2004), many examples of epigenomic effects of hybridization in
plants, including maize, wheat, rice, cotton, and sunflower, are reviewed.
fThere are similar, older studies of transgenerational effects following administration of hormones and drugs in a variety of mammals. These studies, which did not include molecular
data, were reviewed by Campbell and Perkins (1988).
152 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

molecular data show that the gametes of
the F1generation have acquired altered
epigenetic marks, the transmission of epi-
genetic variations rather than the direct
induction of such variations is a reasonable
assumption. Based on this, we included
some cases in our table in which only two
generations of transmission following in-
duction were reported. For instance, we
included a case reported by Burdge et al.
(2007) in which protein restriction during
a rat grandmother’s pregnancy led to iden-
tical somatic and epigenetic (methylation)
variations in her F1and F2descendants, as
well as a case reported by Newbold et al.
(2006) that showed that F0female rats ex-
posed to diethylstilbestrol had F1and F2
offspring with an increased susceptibility to
tumors associated with persistent changes
in the DNA methylation and expression of
an oncogene (see Ruden et al. 2005 for an
epigenetic model explaining these data).
The case of germline inherited epigenetic
variants (and their corresponding pheno-
types) that were acquired stochastically in
the Avy locus of the mouse was included
since transgenerational epigenetic effects
on coat color and obesity were reported.
However, this is a complex case that seems
to involve epigenetic modifications in the
Avy locus as well as in other loci in the ge-
nome (Blewitt et al. 2006; Cropley et al.
2006; Waterland et al. 2007, 2008). An-
other case we included was that reported
by Csaba et al. (2005) in which the modi-
fied serotonin content induced in the im-
mune cells of mice treated with endomor-
phin during pregnancy is subsequently
transmitted to their grandchildren. We did
not include the many cases of single gen-
eration inheritance, even though epige-
netic inheritance seems quite plausible in
some of them, as, for example, when a
predisposition to cancer in humans is re-
lated to an epimutation (an epigenetic he-
reditary abnormality in gene expression)
in a mismatch-repair gene, and this epimu-
tation is transmitted from mother to son
(Hitchins et al. 2007). We also excluded a
similar case in mice which showed that tu-
mor risk was increased following chro-
mium III chloride exposure and that this
risk was then transmitted from father to
son (Shiao et al. 2005). Other cases of
single-generation inheritance, such as pro-
tection against type I diabetes in humans
(where a paramutation-like process has
been reported) (Bennett et al. 1997), hy-
drostatic pressure-induced alterations in
DNA methylation in Japonica rice (Long et
al. 2006), and parallel alterations in gene
expression profiles in White Leghorn
chickens and their offspring following
stress in the parental generation (Lindqvist
et al. 2007), were also excluded, although
epigenetic inheritance may well have oc-
curred and may be revealed when subse-
quent generations are studied. However,
within the limitations imposed by the re-
search designs of the studies we reviewed
and the qualifications we have mentioned,
we believe that the table provides a fairly
exhaustive overview of the recognized cases
of cellular transgenerational epigenetic
inheritance that have been described in
English-language journals, although, inev-
itably, we likely missed some cases.
taxonomic distribution and inducing
conditions
The data in the table probably represent
the tip of a very large iceberg. What is
missing from the table is important, be-
cause the absences point to gaps that need
to be filled. For instance, there is no infor-
mation about epigenetic inheritance in the
kingdom Archea, and most phyla are not
represented. There are also few data di-
rectly addressing epigenetic inheritance in
viruses, although it may plausibly be as-
sumed that viruses exploit and use the epi-
genetic adaptations adopted by their hosts.
Data on epigenetic inheritance in chloro-
plasts and mitochondria are also very
scant. It is worth noting that the organisms
that show the greatest evidence for epige-
netic inheritance are the classical model
organisms of genetics—E. coli, yeast, Arabi-
dopsis, maize, rice, Caenorhabditis,Drosoph-
ila, the mouse, and the rat. However, a
systematic investigation of epigenetic in-
heritance in different conditions is not yet
available for any of these model organisms.
It is also worth noting that all the model
June 2009 153TRANSGENERATIONAL EPIGENETIC INHERITANCE

animals studied belong to taxa in which
the segregation between germline and
soma occurs early, and, therefore, epige-
netic inheritance may be more limited
than in the non-represented animal taxa
where segregation occurs late if it occurs at
all (Buss 1987; Jablonka and Lamb 1995).
Although the non-systematic way in which
the data were collected precludes general
conclusions, it seems as if epigenetic inher-
itance in multicellular organisms is most
common in plants and fungi. This is prob-
ably in part due to the lack of segregation
between soma and germline in these
groups that enables developmentally in-
duced epigenetic variations occurring in
somatic cells to be transferred to the ga-
metes when these somatic cells assume
germline functions. However, there are
two, additional considerations that may be
relevant to the difference between animals,
on the one hand, and plants and fungi, on
the other. First, the lack of nervous system-
directed mobility and activity in plants and
fungi means that they cannot adapt to
changing conditions behaviorally; if the
conditions experienced by offspring are
likely to be similar to those of their par-
ents, then inheriting epigenetic adapta-
tions from them is an alternative adaptive
strategy to behavior and is likely to be pos-
itively selected in plants and fungi (Jab-
lonka et al. 1995; Jablonka and Lamb 1995;
Lachmann and Jablonka 1996). Second,
mobility and CNS-dependent flexible learn-
ing in animals may often limit the predict-
ability of the environment in descendent
generations; therefore, wide-ranging stable
epigenetic cellular inheritance through
the germline may be selected against. In
general, it seems that the difference be-
tween the life strategies of plants and ani-
mals may account for the observation that
epigenetic inheritance in multicellular or-
ganisms is more common in plants and
fungi than among animals.
The relative importance—and some-
times even the very presence—of particu-
lar EISs in different taxa varies. Budding
yeast seems to lack the RNAi system, so
epigenetic inheritance based upon it is im-
possible, although other types of RNA-
mediated inheritance cannot be ruled out.
In some groups of animals, DNA methyl-
ation appears to have been lost (Regev et
al. 1998) and is not part of chromatin
marking in these organisms. Transgenera-
tional structural inheritance and inheri-
tance through self-sustaining loops seem to
be more common in unicellular organisms
and in fungi where horizontal transfer of
information through hyphal interaction is
common. It may be that the development
of a germline involves such drastic alter-
ations in cellular functions and structures
that self-sustaining loops and many cellular
structures are destabilized and disrupted.
The data presented in Table 1 lend sup-
port to this conjecture, although their pau-
city precludes decisive conclusions.
To understand why and when cellular
epigenetic variants are inherited, we need
to know the conditions that promote their
induction and stability in cell lineages.
Their developmental nature requires an
approach that is sensitive to context. How-
ever, our knowledge of the chromatin-
marking and RNA-mediated systems sug-
gests that certain parts of the genome may
exhibit chromatin- and RNA-mediated epi-
genetic inheritance more often than oth-
ers. Repetitive DNA sequences (especially
regions that code for RNAs that can form
dsRNA or stem-loop structures), DNA re-
gions where transcription is likely to start
from both complementary strands, and re-
peated chromosomal segments that pair
ectopically are all likely to exhibit RNA-
mediated epigenetic inheritance under a
wide range of conditions. These and other
repeated sequences that cooperatively bind
protein complexes and regions with CG dou-
blets are all likely candidates for multigen-
erational chromatin- and methylation-based
epigenetic inheritance.
The inducibility and transmissibility of epi-
genetic variants depend on developmental
conditions. Conditions of stress seem to be
particularly important as inducers of herita-
ble epigenetic variation, and lead to changes
in epigenetic and genetic organization that
are targeted to specific genomic sequences.
We mentioned earlier that the genomic
stresses of allopolyploidization and, to a
154 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

lesser extent, autopolyploidization lead
to epigenetic and genetic re-patterning
(Grant-Downton and Dickinson 2005,
2006; Rapp and Wendel 2005). A well-
known epigenetic phenomenon associated
with hybridization is nucleolar dominance,
or the expression in the hybrid of the
rRNA gene complex (NOR) from only one
parent. The preferential silencing of one
NOR is a large-scale gene-silencing phe-
nomenon associated with heritable DNA
methylation and repressive histone modifi-
cations (Pikaard 2000, 2003). However, the
epigenetic changes occurring in hybrids
are not restricted to rRNA genes. An ex-
ample we present in the table is that of the
genome-wide changes that occur in syn-
thetic wheat hybrids that were formed in
order to simulate the evolution of domes-
tic wheat. Levy and Feldman (2004) re-
viewed evidence showing that, in these hy-
brids, 13% of the genome undergoes
significant methylation changes, while
changes also occur in genome organiza-
tion (e.g., rearrangements and elimination
of some sequences). The methylation
changes affect both low copy numbers and
repetitive DNA sequences, and are associ-
ated with heritable transcriptional silenc-
ing. In addition, the activation of retroele-
ments leads to heritable alterations in gene
expression at other loci, thus resulting in
major changes in the profile of gene ex-
pression. The changes in the epigenetic
state of the genome are region and chro-
mosome specific: they are targeted to par-
ticular genomic sequences and reoccur,
with localized variations, upon repeated
formation of the same type of al-
lopolyploid. In species of the cordgrass
Spartina, genome-wide epigenetic and ge-
netic changes were observed in two re-
cently formed, morphologically different,
natural hybrids (and an allopolyploid), in
which 30% of the parental methylation pat-
terns were altered, in addition to similar
structural changes in the DNA sequences of
these two independently formed and genet-
ically similar hybrids (Salmon et al. 2005).
Many similar effects of allopolyploidization
have been reviewed in the extensive and
growing literature on plant polyploidy (e.g.,
see The Biological Journal of the Linnean Society
82(4) [Allen 2004]). The overall impression
gained from these studies is that heritable
epigenetic changes accompany the first
stages of allopolyploidization, and that the
types of repetitive sequences in the parental
species, the amount of divergence between
them (especially with regard to elements that
may be involved in epigenetic control), and
the direction of the cross all play important
roles in specifying the extent and nature of
epigenetic re-patterning.
Another form of genomic stress that may
lead to heritable variation is that associated
with a change in reproductive mode. For
example, the transition from sexual to
a-gametic reproduction in sugar beet leads
to heritable activation of some genes (Le-
vites 2000). DNA damage also leads to her-
itable epigenetic changes, and researchers
are beginning to uncover some of the fac-
tors that affect this response. Following
DNA repair, the epigenetic structure of the
repaired region is not fully reconstructed
and carries with it a repair-specific chroma-
tin signature that can be transmitted to
subsequent generations (Polo et al. 2006).
Moreover, the loading of the histone vari-
ant
␥
-H2AX (which is associated with re-
paired DNA segments) with cohesin leads
to sister-chromatid interactions that may
contribute to the radiation-induced ge-
nome instability that arises and is inherited
for several generations in the progeny of
damaged cells (Little 2003). This may be
the basis for the heritable genomic insta-
bility found in the offspring and grand-
offspring of male mice that were exposed
to irradiation (Dubrova 2003), as well as
for the increased level of recombination
seen for at least four generations after
UV-C irradiation of Arabidopsis plants (Mo-
linier et al. 2006).
Some of the cases included in Table 1
show that physiological stresses—for in-
stance, nutritional stresses imposed during
sensitive periods in the development of
flax (Cullis 2005)—can lead to both ge-
netic and epigenetic re-patterning, and
both types of re-patterning seem to be cor-
related and share a common mechanistic
basis. It seems likely that other cases in
June 2009 155TRANSGENERATIONAL EPIGENETIC INHERITANCE

which an environmental stressor has tar-
geted effects on genome organization,
such as the heat-induced changes in rRNA-
encoding DNA repeats in Brassica (Waters
and Schaal 1996), will also be found to
be associated with heritable epigenetic
changes in the genes or repeated elements
involved. In animals, alterations in hor-
monal balance, especially those occurring
over several generations, may also be fol-
lowed by epigenetic changes. This may be
the basis for the pattern of inheritance of
white spotting seen in domesticated silver
foxes (as we will discuss later) (Trut et al.
2004). However, stressful conditions may
not only affect the chromatin-marking and
the RNA-mediated EISs; if a new prion vari-
ant can be generated in stressful condi-
tions, it might cross previously existing spe-
cies barriers and have novel effects in its
new “host” species.
epigenetically-based similarities and
differences between generations:
inducing epigenetic variations
in the germline
Many of the studies of multicellular,
sexually reproducing organisms that we
present in Table 1 show that, as a result of
an inducing stimulus or of changed condi-
tions in the F0(parent) generation, similar
chromatin marks and similar phenotypes
are reconstructed in subsequent genera-
tions. However, the F0generation itself
may not show any phenotypic effects;
changes in epigenetic marks and associ-
ated somatic phenotypes may first appear
in the F1generation and may only then be
inherited by subsequent generations. For
example, in mice, the diethylstilbestrol-
induced increase in the probability of de-
veloping tumors appeared only in the F1
and F2generations (Newbold et al. 2006).
Understanding the inheritance of in-
duced variation in sexually reproducing
multicellular organisms is an important
topic of research, not least because, in the
past, much of the debate about the impor-
tance of acquired characters in evolution
revolved around this issue (Delage and
Goldsmith 1912). Traditionally, in multi-
cellular organisms with a germline, three
types of induced heritable effects have
been distinguished: direct induction, par-
allel induction, and somatic induction.
With direct induction, the germline is di-
rectly affected without any effect on the F0
parent’s soma, while with parallel induc-
tion, similar somatic phenotypic effects are
apparent in both the induced ancestor and
its descendants, but the induction events in
the somatic and germ lineages are inde-
pendent. Finally, with somatic induction, a
change is induced in the soma, and this
somatic effect causes a change in the germ-
line that reconstructs the somatically-
induced parental phenotype in the descen-
dants (Fothergill 1952; Jablonka and Lamb
1995). There is, however, a fourth possibil-
ity: an induced effect on the soma of the F0
generation may cause changes in the germ-
line, but the resulting somatic changes in
descendants are dissimilar from the effect
on the soma in the F0generation. This is a
case of parallel induction with nonparallel
effects. The different types of induced her-
itable effects are represented schematically
in Figure 2.
The literature shows that direct induc-
tion, parallel induction, and parallel induc-
tion with nonparallel or partially parallel
effects are common. In Table 1, many
of the cases of paramutation and of in-
duced, heritable, transposable element ac-
tivity can be classified as direct induction
or parallel induction, because the induc-
ing conditions directly affect events in the
germline (and sometimes, in parallel, in
the soma as well). Changes in hormone
dynamics that specifically target the germ-
line will either not affect somatic charac-
ters in the induced F0generation, or will
affect them in a way that is unrelated to the
effect seen in the F1. For example, in one
case, vinclozolin, an androgen suppressor,
induced testis disease in at least three gen-
erations of males following its administra-
tion to a pregnant female ancestor (Anway
et al. 2005, 2006a,b). This is an example of
direct induction with differing effects in
males and females. Parallel induction is
seen in the case in which nutritional and
temperature changes affected morphology
156 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

in the F0generation of flax, and similar
phenotypic effects were found in subse-
quent generations (Cullis 2005). Another
case of parallel induction is that seen in the
moth Ephestia kuehniella, where subjecting
the insect to modified temperature condi-
tions, lithium ion treatment, or an alterna-
tive electric field during the first half of its
pupal development resulted in the sup-
pression of a mutant short antennae in the
treated generation and five subsequent
generations (Pavelka and Koudelova´ 2001).
Also, glucose induction of the S form of the
G6PD enzyme in Daphnia pulex may be illus-
trative of a further possible case of parallel
induction (Ruvinsky et al. 1983a,b).
The distinction between direct and par-
allel induction is not always straightfor-
ward, as epigenetic changes may occur in
only some of the somatic cells of the in-
duced parent, thus resulting in a varie-
gated somatic phenotype. This can be de-
scribed as partial parallel induction. It can
be seen in some cases of paramutation in
plants where, when the induction rate is
high, sectors that are the result of paramu-
tation in somatic tissues can be seen in the
F0generation (Chandler et al. 2000).
Figure 2. Inducing an Inherited Effect
(A) Direct germline induction: An external stimulus induces a germline change from G0to G1with no effect on
the parental soma, which remains S0. The G1state is inherited, and leads to the development of an S1soma.
(B) Parallel induction: An external stimulus induces a change in the parent’s soma from S0to S1and in its
germline from G0to G1. The G1state is inherited and causes the development of an S1soma in descendants.
(C) Somatic induction: An external stimulus induces a change in the parent, altering its somatic phenotype from
S0to S1. The effect is transmitted from the S1soma to the germline, where G0is changed into G1;G
1is
consequently inherited and results in the development of an S1soma. (D) Parallel induction with nonparallel
effects: An external stimulus alters the soma from S0to S2, and the germline from G0to G1. The germline
modification is inherited and leads to the development of S1 soma in subsequent generations. With all four
types of induction, S1could have an effect on G1in all descendants of the original induced parents (not shown).
June 2009 157TRANSGENERATIONAL EPIGENETIC INHERITANCE

Several cases in the table can be inter-
preted as instances of somatic induction in
which heritable variation is induced in the
soma and is seen as a somatic character,
and the resulting effect is then transferred
from the soma to the germline. The ability
of small RNAs to move from cell to cell may
facilitate soma to germline information
transmission and may form the basis of
such inheritance. The most impressive ex-
ample of somatic induction via small RNAs
may be found in the case discussed by Vas-
tenhouw et al. (2006), in which C. elegans
were fed bacteria with DNA sequences cod-
ing for dsRNA, and the RNAs migrated
from the somatic cells of the nematode to its
germ cells, thus affecting subsequent gener-
ations. Steele et al. (1998) suggested another
route of transmission that is initiated by
RNA: the transfer of RNA transcripts from
the immune cells of mammals to their germ-
line, followed by reversed transcription and
incorporation of the reverse transcribed
DNA into the germline genome. Zhivotovsky
(2002) has modeled the conditions that
could lead to the evolution of such a system.
Another route of somatic induction is via
inducing conditions that affect the secre-
tion of hormones, which, in turn, affect the
germline. The effect of a hormone on the
somatic characters of the induced parent
(F0) and on its descendants may be similar,
but it is unlikely to be identical. This is
because even if the same genes are affected
in the soma of the adult F0animals as well
as in the F1and subsequent generations of
offspring, it is unlikely that the pattern of
activity of these genes will be the same. In
descendants, the induced epigenetic change
may be expressed during embryonic as well
as adult stages, whereas, in the F0animals,
the change is induced in adults. Only in
rare cases in which the pattern of timing
and spatial expression is identical in the F0
and subsequent generations—and is lim-
ited to the stage at which it was induced in
the F0—will identical somatic characters
occur in parent and offspring generations.
An additional complication is the sex of
the F0animals and their offspring, as sex-
limited effects will obviously be transmitted
in a sex-specific manner, and epigenetic
variations in an induced F0parent may
therefore be different for descendants of
the opposite sex. These considerations are
relevant to any kind of somatic induction,
no matter the mechanism behind it, so,
although partial similarity is likely, we ex-
pect that somatic induction leading to
identical phenotypes in the F0and subse-
quent generations will be rare.
The foregoing discussion suggests that
the traditional distinctions between direct,
parallel, and somatic induction do not sat-
isfactorily describe the possible interac-
tions between the soma and germline. The
similarity between the somatic characteris-
tics of an induced ancestor and its descen-
dants, which the traditional classification
highlights, is, of course, of interest, but we
think that the mode and mechanisms of
the induction of germline variations rather
than their effect (i.e., the similarity or lack
of similarity between the F0and the F1)
should be the focus of study. An important
aspect of the problem is whether external
inducing conditions affect the germline di-
rectly or whether their effect is mediated
through the soma. External environmental
conditions that are independent of the or-
ganism’s activity and its development can
have important heritable epigenetic ef-
fects, but when there are somatic mediat-
ing signals (e.g., RNAs or hormones) these
somatic signals may evolve to efficiently
communicate information to the germ-
line. We therefore expect that develop-
mentally-mediated somatic effects may
have adaptive consequences more often
than signals that act on the germline di-
rectly, and that when there are such
soma-mediated influences on the germ-
line, the effects on the somatic cells of
the F0will often differ from those that
will be seen in the next generations (Fig-
ure 2D).
The possibility of a somatic effect on the
germline that is mediated by hormones
was raised shortly after hormones were dis-
covered. The Austrian Lamarckian zoolo-
gist Paul Kammerer and the pioneer endo-
crinologist Eugen Steinach (1920) found
that exposing male rats to high tempera-
tures led to morphological and physiolog-
158 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

ical changes in their offspring and grand-
offspring. They suggested that the presence of
hormone-secreting interstitial cells adjacent
to germ cells in the gonads facilitated hor-
monal interactions between them, and
they claimed that heat produced a change
in hormone production in interstitial cells,
thus affecting germline cells and carrying
hereditary consequences (see Logan 2007
for a discussion of Kammerer and Stein-
ach’s work). Although the validity of this
claim and its possible interpretation in
terms of epigenetic inheritance is at present
unclear, the possibility that there are hor-
monal effects on epigenetic variation is no
longer considered heresy.
The experimental work of Vanyushin et
al. (2006a) has shown that methylation pat-
terns in the rat genome are controlled by
hydrocortisone dynamics, and that phyto-
hormones of different classes cause a de-
cline in global DNA methylation and the
repression of de novo methylation in plants.
Moreover, the evidence reviewed by Naz
and Sellamuthu (2006) suggests that, de-
spite doubts about some of the reported
information, there are 8 hormone recep-
tors and 16 cytokine/growth receptors in
mature ejaculated sperm, thereby allowing
for the possibility that hormones could
exert their effects on male gametes.
Hormone and neurotransmitter recep-
tors have also been found in oocytes and in
female germline cells. For example, the
oestrogen receptor (Wu et al. 1992), sero-
tonin receptor 5-HTID (Vesela´ et al. 2003),
Notch1 and Notch2 receptors (Cormier et
al. 2004), and the
␤
2-andrenoceptor (C
ˇi-
kosˇ et al. 2005) have all been detected in
oocytes, and the GH receptor has been
detected in fertilized eggs (Pantaleon et al.
1997; Ko¨lle et al. 2001). It is interesting to
note that the changes in expression of
some of the genes coding for these recep-
tors coincide temporally with early waves of
epigenetic re-programming during devel-
opment. The presence of hormone recep-
tors in gametes, and the modulation of
receptors’ synthesis during sensitive devel-
opmental periods when hormonal changes
occur, suggest that induced variations in
hormonal conditions may affect the epige-
netic state of genes within germline cells,
and these, in turn, can be transmitted to
the next generation.
The involvement of hormones in the in-
duction of heritable epigenetic variations
is no longer a mere speculation: several of
the mammalian examples presented in Ta-
ble 1 suggest that changes in hormonal
stimuli induce heritable epigenetic changes.
For example, the penetrance of the fused
phenotype is altered in the progeny of
mouse parents treated with hydrocortisone
(Belyaev et al. 1983). In silver foxes se-
lected for tame behavior, hormonal effects
in the serotonin system that controls ag-
gression seem to be involved in the herita-
ble activation of the star gene that leads to
white spotting (Belyaev et al. 1981a,b; Trut
et al. 2004; Popova 2006). The best inves-
tigated case of hormonally-mediated ef-
fects on epigenetic marks is that of the
transgenerational effect of the estrogenic
androgen disruptors vinclozolin and me-
thoxychlor on testes development in male
rats (Anway et al. 2005, 2006a,b; Chang et
al. 2006; Crews et al. 2007). With vinclozo-
lin, 15 different DNA sequences isolated
from sperm were shown to have altered
methylation patterns, and these patterns
were transmitted to the F1-F3generations
of offspring of treated F0females. Cru-
cially, vinclozolin was effective only when
administered between 8 and 15 days post
coitum, and had no effect when adminis-
tered later, between 15 and 20 days. This
sensitive developmental period coincides
with the epigenetic remethylation phase in
the male (Hajkova et al. 2002), thus sug-
gesting that the hormonal effect of andro-
gens is developmentally specific (limited to
this period of epigenetic reprogramming)
and is not a general toxic effect. If so,
modifications in these methylation pat-
terns in the soma of F0vinclozolin-treated
females are not expected. In plants, where
no epigenetic reprogramming phase simi-
lar to that in mammals is apparent, and the
germline is continuously produced during
development (Matzke and Scheid 2007),
changes in hormonal stimulations during
all phases of somatic development are
June 2009 159TRANSGENERATIONAL EPIGENETIC INHERITANCE

likely to affect epigenetic variations in the
germline.
epigenetic recall and other
directional changes in heritable
epigenetic marks
We defined cellular epigenetic inheri-
tance as the transmission from mother cell
to daughter cell of variations that are not
the result of differences in DNA sequence,
or of persistent inducing signals in the
cell’s environment. The examples pre-
sented in Table 1 support our assertion
that when epigenetic marks are inherited,
the same pattern of marks is more or less
faithfully reconstructed across generations.
If a particular mark—for example, a pat-
tern of 5 methylated cytosine sites—is in-
duced at a particular locus in the germline,
this pattern is then reconstructed (with a
certain error rate) in the descendants and
has similar phenotypic effects (Figure 3A).
Stable epigenetic inheritance is at one
extreme pole, and total reset to a single
default state—that of the uninduced par-
ents—is at the other pole of developmen-
tally influenced inheritance. However, the
developmental nature of epigenetic inher-
itance and our knowledge of the construc-
tion of chromatin marks suggest that we
consider other possibilities; the examples
in the table, therefore, represent only a
very small fraction of the types of epige-
netic hereditary phenomena that probably
exist.
The first possibility is that of partial re-
construction—an intermediate between the
two extreme poles of complete reset and
faithful reconstruction. For example, of
the 5 induced methylation sites, only 3 are
reconstructed in the offspring, and, in the
absence of the inducing stimulus, if the
phenotype of the offspring is the same as
that of an uninduced individual, this would
not be seen as a case of epigenetic inheri-
tance, even if the threshold for the devel-
opmental response was lowered or the
speed of reaction was enhanced in descen-
dants (see Figure 3B). The situation is sim-
ilar to that found with neural memory,
when the original stimulus that leads to the
initial learnt response is required to trigger
the response again, but because memory
traces remain, there is recall—facilitated
reconstruction of the learnt response upon
re-induction. We suggest that the inheri-
tance of some epigenetic memory traces
may lead to epigenetic recall—a facilitated
response in descendants that requires an
inducer. The inherited, partial epigenetic
patterns that facilitate a response are
called epigenetic engrams. (Engram is a term
that was invented by Richard Semon in
1904, to mean, roughly, “memory trace”;
see Schacter 2001.) In order to recognize
epigenetic recall, the kinetics of induced
responses in the F1and subsequent gener-
ations need to be studied. None of the
examples in Table 1 fulfils this require-
ment, and, to the best of our knowledge,
this kind of investigation—searching for
epigenetic engrams and for facilitated but
still inducer-dependent responses— has not
been part of the research program of epi-
genetics. Such responses, however, are
likely to be common, as the mechanisms
for them are all in place, and a system
enabling the reconstruction of epigenetic
engrams that allow recall would be selec-
tively advantageous in many conditions,
just as, in spite of the different timescale,
the evolution of neural sensitization has
been favored. Agrawal et al. (1999) studied
induced defenses against predators in wild
radish (Raphanus raphanistrum) and the wa-
ter flea (Daphnia cucullata), and showed
that induction in the parental generation
made offspring better adapted to preda-
tors than the offspring of uninduced par-
ents. They suggested that the persistent pa-
rental effect in radish plants may be either
a direct maternally-induced effect or the
result of more rapid induction of plant
defenses in the offspring of damaged
mothers. If the latter proves correct, it will
represent a case of epigenetic recall, un-
derlain by as yet uncharacterized epige-
netic engrams. (For an extended discus-
sion of the possibility of learning in cells
and non-neural organisms, see Ginsburg
and Jablonka 2008.)
In addition to epigenetic recall based on
partial reconstruction, we must also con-
sider the possibility that although induced
160 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

Figure 3. Types of Transgenerationally Inherited Epigenetic Effects
An inactive gene (gray rectangle) and its corresponding phenotype are depicted at the top of the figure. (A)
Epigenetic copying and corresponding phenotypic inheritance: An epigenetic mark consisting of 5 ⫹sites is induced
in the parent and affects the marking of the gene and the phenotype of the induced individual. (The “⫹” signs
indicate an altered methylation or histone modification site, and the inducer is indicated by a curved arrow.)
The epigenetic mark is reliably transmitted through the germline, thus leading to a modified heritable
morphology in the uninduced progeny (heritable site-states are indicated by “⫹” signs within the rectangle, and
the straight black arrows indicate transitions between generations). (B) Epigenetic recall: Partial inheritance of
the epigenetic pattern (represented by 3 internalized “⫹” signs) that was established in the induced parent
does not lead to modified morphology in progeny. However, the amount of inducer needed to re-establish the
full epigenetic pattern (5 ⫹pattern) and the corresponding induced phenotype is much smaller than in the
parent (the smaller curved arrow indicates low level stimulation). (C) Reactive but dissimilar effects of inherited
epigenetic patterns: (i) antagonistic - The parental epigenetic mark (5 ⫹signs) is inherited faithfully, but, in a
mismatched postnatal environment (triangle around progeny), it leads to a different phenotype in the
progeny. (ii) accumulative - Following recurrent induction in each generation, epigenetically modified and
“internalized” sites accumulate, and result in correspondingly more extreme phenotypes. When the epigenetic
pattern reaches a certain configuration (5 ⫹internalized sites), it is inherited even in the absence of the
inducer, and this is a form of epigenetic assimilation. (iii) lingering-fading - Following induction, the mark and
its corresponding morphology are established, but fade away gradually in subsequent generations in a nonin-
ducing environment.
June 2009 161TRANSGENERATIONAL EPIGENETIC INHERITANCE

marks in one generation may be faithfully
inherited, they might lead to a non-matching
yet predictable phenotype in the subse-
quent generation if the environments of
parent and progeny are drastically differ-
ent (Figure 3Ci). The offspring’s response
could be interpreted as a misfired predic-
tive response—the consequences of a strat-
egy that evolved when the parents’ and the
offspring’s conditions matched. The ef-
fects of such mismatches and their medical
significance have been discussed by Gluck-
man and Hanson (2005; Gluckman et al.
2007).
Another possibility worth considering
is that of directional changes in heritable
epigenetic marks over the course of gener-
ations. For instance, if inducing condi-
tions persist for several generations, epige-
netic marks may accumulate (Jablonka and
Lamb 2005). This could lead to a more
extreme phenotype (Figure 3Cii) and, pos-
sibly, to a greater fidelity of transmission.
Inducing conditions might endure due to
the persistence of external environmen-
tal factors (e.g., there is multi-generational
exposure to a chemical), continual trans-
mission through one sex (i.e., for several
generations a particular epiallele is trans-
mitted only through females, or only
through males), or continuous transmis-
sion through old parents, thus leading to
the Lansing effect (see Lamb 1994). The
opposite (Figure 3Ciii) may also occur;
that is, when induction in the parental gen-
eration is followed by non-inducing condi-
tions in the subsequent offspring genera-
tions, the induced epigenetic variations
may linger and gradually fade away, with
some marks being lost in each generation.
This might be the basis of the “lingering”
modifications described by Jollos (1921),
who found that following exposure to high
arsenic or salt concentrations, or to high
heat, paramecia showed heritable pheno-
types that slowly faded over many genera-
tions (Jablonka et al. 1992). We suggest
that the study of epigenetic engrams, and
the study of the kinetics of epigenetic
memory changes in different conditions,
will lead to an expansion of the research
agenda of epigenetics (Ginsburg and Jab-
lonka 2008).
Implications: Evolutionary,
Practical, and Theoretical
Given that epigenetic variations are of-
ten less stable than genetic variations, what
evolutionary significance do they hold? We
argue that a view of heredity that incorpo-
rates the transmission of epigenetic infor-
mation through cellular EISs presents chal-
lenges and opportunities to applied and
theoretical research in evolutionary biol-
ogy. Since, with few exceptions, the in-
corporation of epigenetic inheritance and
epigenetic control mechanisms into evolu-
tionary models and empirical studies is still
rare, our discussion is, inevitably, some-
what speculative.
implications for the study of
evolution
Heritable epigenetic variations and epi-
genetic control mechanisms are relevant
for the empirical and theoretical study of
evolution because they affect both the pro-
cesses of adaptation and of divergence
(Jablonka and Lamb 1995, 2005, 2006a,
2007b). Five types of effects are character-
ized: (i) evolutionary change occurring
through selection of epigenetic variants,
without involvement of genetic variation;
(ii) evolutionary change in which an initial
epigenetic modification guides the selec-
tion of correlated genetic variations; (iii)
evolutionary change stemming from the
direct effects of epigenetic variations and
epigenetic control mechanisms on the gen-
eration of local and systemic epigenomic
variations; (iv) evolutionary change resulting
from the constraints and affordances that
epigenetic inheritance imposes on develop-
ment; and (v) evolutionary change that leads
to new modes of epigenetic inheritance.
Evolution through Selection of
Epigenetic Variants
Adaptation can occur through the selec-
tion of heritable epialleles, without any ge-
netic change. This may be of particular
importance when populations are small
162 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

and lack genetic variability (e.g., in situa-
tions of intense inbreeding following isola-
tion or following changes in reproduc-
tive strategies). As the examples listed in
Table 1 indicate, epigenetic variants are
often induced when environmental condi-
tions change, so several individuals in the
population may acquire similar modifica-
tions at the same time. This means that ad-
aptation through the inheritance of newly
induced epigenetic variants may be very
rapid (Jablonka and Lamb 1995, 2005; Kus-
sell and Leibler 2005; Richards 2006; Boss-
dorf et al. 2008), thus leading to the accu-
mulation of epigenetic variations. Several of
the epigenetic variations presented in the
table are beneficial for their carriers, such as
increased epigenetically heritable antibiotic
resistance in bacteria (Adam et al. 2008) and
the switch between morphotypes in Candida
albicans (Zordan et al. 2006). Other cases,
such as increased mutability as a result of
radiation in mice (Dubrova 2003), increased
recombination rate in plants (Molinier et al.
2006), and alterations in flowering time,
color, and flower morphology (see table for
several specific examples and references) are
likely to be adaptive under some conditions;
therefore, positive selection of such variants
is plausible.
In order to assess the role of epigenetic
variation in microevolution, it is important
to evaluate the extent and heritability of
epigenetic variations in natural popula-
tions (Yi et al. 2004). In a programmatic
paper that outlines the framework for eco-
logical epigenetics, Bossdorf et al. (2008)
present some of the fundamental research
questions that need to be asked about epi-
genetic variations in natural populations,
regarding the extent and structure of epi-
genetic variation, its correlation with phe-
notypic variation, its inducibility, and its
effects on fitness.
The mechanisms of epigenetic control
may play an interesting role in structuring
epigenetic variation because they can coor-
dinate patterns of gene expression. Zuck-
erkandl and Cavalli (2007) believe that re-
peated sequences in “junk DNA” might be
carriers of epigenetic marks, and that
marks on these sequences can be commu-
nicated to other regions in the genome.
They suggested that an altered mark could
therefore result in coordinated hereditary
changes in the expression of several differ-
ent genes simultaneously, hence accelerat-
ing adaptive evolution.
Coordinated hereditary epigenetic changes
may have been involved in the process of do-
mestication. For instance, forty-six genera-
tions of selection for tameness in silver foxes
by Belyaev and his research group in No-
vosibirsk resulted in a complex of heritable
changes. The foxes became dog-like in their
behavior and displayed skeletal, hormonal,
and spotting changes, as well as altered tail
and ear posture, altered vocalizations, and
an increased number of supernumerary
chromosomes (Belyaev et al. 1981a,b; Trut et
al. 2004). Analysis of the pattern of inheri-
tance of white spotting revealed that spotting
behaved like a dominant or semi-dominant
trait, but the rate of appearance and disap-
pearance of the character was far too high
for new mutations to be a likely explanation.
These reversible changes could not be ex-
plained as an effect of inbreeding either,
because the coefficient of inbreeding was
only 0.03 (Trut et al. 2004). A probable ex-
planation is that the stress of domestication
and selection for tameness targeted genes
with large effects in the neuro-hormonal sys-
tem (Trut et al. 2004; Popova 2006) and may
have heritably reactivated some of them
(Belyaev 1981a,b). This epigenetic interpre-
tation, in terms of new epimutations rather
than new mutations, explains the high rate
of appearance and disappearance of some
phenotypes, and support for this comes
from the fact that at least two of the genes
(Agouti and C-kit) that seem to be involved in
the changes are known to have heritable epi-
genetic variants in mice (Trut et al. 2004).
The induction and selection of epigenetic
variations may also have been important in
the domestication of plants: ecological and
genomic stress conditions caused by mov-
ing plants to new conditions and crossing
divergent strains induce many epigenetic
variations, and selection of such variations
probably played a part in domestic plant
evolution.
June 2009 163TRANSGENERATIONAL EPIGENETIC INHERITANCE

Epigenetic Change Guiding the Selection
of Genetic Variations
The guiding role of development in evolu-
tion has been a subject of discussion ever
since the pioneering work of Wadding-
ton (1957, 1968, 1975) and Schmalhausen
(1949). Their basic idea was that selection
can lead to a change from a stimulus-
dependent to a stimulus-independent (or
less dependent) phenotypic response. The
process leading to the change from a phe-
notype whose expression was dependent
on an environmental inducer to a consti-
tutive expression was called genetic assim-
ilation (Waddington 1957; for recent dis-
cussion of the idea and its evaluation, see
Pigliucci et al. 2006; for suggestions em-
phasizing its role in the evolution of behav-
ior, see Avital and Jablonka 2000; Gottlieb
2002).
Ideas about the significance of develop-
mental plasticity have recently been strength-
ened and extended to provide a general
framework for evolutionary biology (Pigli-
ucci 2001; Schlichting and Pigliucci 1998).
West-Eberhard (2003) suggested that environ-
mentally-induced changes during develop-
ment guide the selection of genetic changes
that simulate, stabilize, and ameliorate any det-
rimental effects of induced developmental
changes. She called this developmental guid-
ing process, which includes but is not limited
to genetic assimilation, “genetic accommoda-
tion.” Induced epiallelic variations that are epi-
genetically inherited may enhance the effec-
tiveness of assimilation and accommodation
processes—something that is likely to be par-
ticularly important during conditions of stress
(Jablonka et al. 1992; Jablonka and Lamb
1995, 2005; Pa´l 1998; Sangster et al. 2004;
Badyaev 2005; Siegal and Bergman 2006). An
example showing the facilitating evolutionary
effects of epigenetic inheritance was provided
by True and Lindquist’s (2000) study, in which
they compared pairs of yeast strains differing
only in whether or not they carried [PSI⫹], the
prion form of a protein that is involved in
mRNA translation. By growing the pairs of
strains in a variety of conditions, they uncov-
ered strain-specific differences between them
in colony morphology and growth characteris-
tics. Since the presence of the [PSI⫹] prion
leads to the suppression of nonsense muta-
tions, the production of a variety of new pro-
tein products in the [PSI⫹] containing strains
(that arose because translation goes beyond
the normal endpoint of functional genes, or
because stop-codons in the middle of non-
functional genes are ignored) was increased
and was beneficial in some conditions. The
epigenetic, selectable variation that is gener-
ated in [PSI⫹] strains might enable a lineage to
adapt and “hold” the adaptation until genetic
changes take over; thus, the heritable epige-
netic variations in protein architecture pave
the way for genetic adaptation (True et al.
2004; Sangster et al. 2004). As a theoretical
model has shown, the adaptive effects of such a
system may lead to its evolution even if the
response is adaptive only once in a million
years (Masel and Bergman 2003). Selection of
the epigenetically-based variation generated by
this type of system would be particularly impor-
tant in asexual lineages, where the accumula-
tion of mutational changes may be slow.
Epigenetic inheritance-driven accommo-
dation has probably been important in chro-
mosome evolution as well. It may, for
example, have initiated the evolution of di-
morphic X and Y chromosomes. Jablonka
and Lamb (1995) suggested that the initial
epigenetic silencing of a sex-determining lo-
cus could have produced an epigenetic het-
eromorphism between chromosomes, which
led to pairing problems in meiosis and con-
sequent heterochromatinization and silenc-
ing of the homologous region. This would
have reduced recombination frequencies
and driven degeneration of the Y chro-
mosome; it could also have led to X-
chromosome imprinting and dosage com-
pensation in mammals (Jablonka 2004a).
Heritable epigenetic variations may also
play an important role in the evolution of
chromosomal structures such as centro-
meres. Henikoff et al. (2001) have proposed
that the rapid evolution of centromeric se-
quences and some centromere-associated
proteins may be driven by an epigenetically-
guided arms race (Talbert et al. 2004; Heni-
koff and Smith 2007). They suggested that
centromeres compete to enter the product
of female meiosis that will form a gamete
164 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

(rather than a polar body, which is a dead
end). Centromeres with DNA sequences that
result in more efficient spindle fiber attach-
ments out-compete others in the race to the
prospective egg, thus leading to centromere-
associated meiotic drive. However, meiotic
drive often has deleterious effects on the or-
ganism’s fitness, because deleterious genes
that are linked to the chromosome with the
driving locus—the “strong” centromere, in
this case—reduce individual fitness. The ill
effects of centromere-driven meiotic drive
are neutralized by the selection of alleles of
centromere protein genes with a lower bind-
ing affinity for the centromeric sequences;
hence, centromere-binding proteins evolve
rapidly and adaptively to counter the “self-
ish” centromere sequences. This evolution is
driven by the centromere DNA sequences
and their attached proteins. It is the func-
tional epigenotype that must remain stable,
whatever the specific identity of the DNA
sequence and the proteins at the centro-
meric regions.
An important consequence of the genera-
tion and evolution of different epigenotypes
across various populations is that, like ge-
netic variations, they may initiate reproduc-
tive isolation. Differences in chromatin struc-
ture that arise by chance or during local
adaptation may result in hybrid offspring
that either fail to develop normally or
are sterile because the two sets of parental
chromosomes carry incompatible chromatin
marks (Jablonka and Lamb 1995). For exam-
ple, incompatibility between parental marks
is thought to be the reason why hybrids be-
tween two species in the rodent genus Pero-
myscus develop abnormally (Vrana et al.
2000), and, in plants, crosses between par-
ents of different ploidies fail because of the
dysfunction of the hybrid endosperm, a tis-
sue exhibiting genomic imprinting (Sokolov
2006).
The Effects of Epigenetic Variations and
Epigenetic Control Mechanisms on the
Generation of Local and Systemic
Epigenomic Variation
Although in practice the biases imposed
by epigenetic variations, such as methyl-
ation marks, on the generation of local
changes in DNA are intertwined with epi-
genetic control mechanisms generating
systemic genomic changes, we need to dis-
cuss them separately, for heuristic reasons.
It has been known for some time that the
rates of mutation, transposition, and re-
combination are lower in condensed than
in open chromatin (Belyaev and Borodin
1982; Jablonka and Lamb 1995), and that
the movement of transposable elements,
which is recognized as a major cause of
genomic change (Kidwell and Lisch 1997),
is markedly influenced by various types of
internal (genetic) and external (environ-
mental) stresses. It is therefore clear that
epigenetic variations bias genetic changes.
However, the effect of epigenetic control
mechanisms can go beyond the more or
less localized mutational changes induced
by local chromatin variations. Zufall et al.
(2005) have suggested that developmen-
tally regulated genome rearrangements
brought about by epigenetic control mech-
anisms are an ancient feature of eu-
karyotes. If so, it is possible that, during
periods of stress, the same epigenetic con-
trol mechanisms cause global epigenomic
macro-variations that are inherited be-
tween generations and that lead to macro-
evolutionary changes (Jablonka and Lamb
2008; Lamm and Jablonka 2008). These
epigenetic control mechanisms may un-
derlie the systemic changes in the genome
that Goldschmidt (1940) believed drove
macroevolution. Goldschmidt proposed
that macroevolutionary changes are the re-
sult of large changes in the genome that
are based either on macromutations (mu-
tations in single genes that have very large
phenotypic effects) or on systemic muta-
tions (changes in the organization of the
genome, such as chromosomal rearrange-
ments). Goldschmidt’s ideas used to be
derided, but recent data from many bio-
logical fronts are changing this attitude
(Shapiro 1999; Bateman and DiMichele
2002; Fontdevila 2005). Sequence studies
have shown that during plant and animal
phylogeny, many developmental genes
have been duplicated and re-used (Gu et
al. 2004), and Rodin et al. (2005) have
suggested how epigenetic silencing may
June 2009 165TRANSGENERATIONAL EPIGENETIC INHERITANCE

play a role in this. Epigenetic control
mechanisms probably have a key role in
speciation through polyploidization and
hybridization, which are of central impor-
tance in plant evolution (Jorgensen 2004;
Rapp and Wendel 2005). As we noted ear-
lier, recent studies have shown that in
many naturally occurring and experimen-
tally constructed polyploids and hybrids,
DNA methylation patterns are dramatically
altered, and genes in some of the dupli-
cated chromosomes are heritably silenced.
Following auto- and allo-polyploidization,
there is a burst of selectable variation, with
all the opportunities for adaptation that
this provides. This evidence is very much in
line with the suggestions of McClintock
(1984), who argued that stress leads to a
reshaping of the genome.
Although we do not yet know how epi-
genetic control systems are involved in the
generation of such systemic mutations,
processes based on pairing, such as the
mechanisms seen in ciliates (in which pair-
ing with scanRNAs determines which se-
quences are degraded) and during meiotic
mis-pairing (in which unpaired regions are
deleted or heterochromatinized) may be
recruited under conditions of genomic
and ecological stress. Molinier et al. (2006,
and personal communication) showed that
exposing Arabidopsis to UV-C radiation in
one generation caused a heritable increase
in the recombination rate of the whole
population of irradiated plants for at least
four generations. This might be an exam-
ple of an induced, pairing-based systemic,
epigenetic change. Jablonka and Lamb
(1995, 2008) suggested that there may have
been selection for specific heritable epige-
netic responses based on pairing, which
are determined by the type of stress (e.g.,
direct radiation-induced damage to DNA,
nutritional stress, heat stress), its severity,
and its probability of reoccurrence.
Evolutionary Constraints and Affordances
Imposed by Epigenetic Inheritance
Cellular EISs were a precondition for the
evolution of complex multicellular organ-
isms with specialized cell lineages, because
cells in such lineages have to maintain and
transmit their determined states, even when
the conditions that initiated them are long
past. Since the cells that give rise to the next
generation of organisms need to have an
uncommitted state, and efficient EISs could
jeopardize this, EISs must have imposed a
strong constraint on the evolution of ontog-
eny. There are several features of develop-
ment that may be outcomes of selection to
prevent cells with inappropriate epigenetic
legacies from founding the next generation.
For example, the difficulty of reversing some
epigenetic states, the early segregation and
quiescent state of the germline of many ani-
mal groups, and the massive changes in
chromatin structure that occur during mei-
osis and gamete production, may all be the
result of selection against transmitting the
epigenetic “memories” associated with the
developmental changes and chance epimu-
tations that would prevent a zygote from
starting its development in a totipotent
epigenetic state (Jablonka and Lamb
1995, 2005). Recently, Pepper et al. (2007)
have suggested that serial differentiation—
the sequence of differentiation that starts
with self-renewing somatic stem cells and
proceeds through several non-self-renewing,
transient, amplifying cell stages before end-
ing with terminally differentiated cells—is
also a strategy that evolved to avoid the
somatic selection of selfish genetic and epi-
genetic variations.
Jablonka and Lamb (2006b) argued that
the constraints and affordances of epige-
netic control systems and epigenetic inher-
itance played a crucial role in all eight of
the major evolutionary transitions identi-
fied by Maynard Smith and Szathma´ry
(1995). For example, the transition from
independent genes to long chromosomes
was probably dependent on epigenetic in-
heritance based upon chromatin marking,
which maintains patterns of gene activity
following DNA replication. Epigenetic con-
trol mechanisms may also have been im-
portant in the transition from prokaryotes
to eukaryotes—a transition that was associ-
ated with processes of endosymbiogenesis.
It is likely that massive and heritable inac-
tivation of large parts of the symbiont-to-be
genome, as well as the employment of
166 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

mechanisms of structural inheritance that
enabled the integrity of the basic structure
of the symbiont-to-be membranes to per-
sist, were involved in this transition.
The Evolution of EISs: Their Origin and
Selection
In light of the growing volume of work
and the theoretical considerations that
suggest that nongenetic mechanisms of in-
formation transfer play key roles in evolu-
tion, the evolutionary origin of nongenetic
inheritance systems is of fundamental in-
terest. There have been some theoretical
and comparative studies that have ad-
dressed the evolution of EISs, but not all
aspects have been explored. The evolution-
ary origins of DNA methylation have been
considered, and several different hypothe-
ses for the advantages it conferred have
been suggested (Bestor 1990; Bird 1995;
Regev et al. 1998; Colot and Rossignol
1999; Mandrioli 2004). There are also sev-
eral comparative studies of histone evolu-
tion (Sandman et al. 1998; Felsenfeld and
Groudine 2003), and the evolution of
RNAi systems for defence against genomic
parasites and as regulators in an ancient
RNA world has been suggested (Cerutti
and Casas-Mollano 2006).
Specific developmental processes that
involve epigenetic inheritance, such as
genomic imprinting and X-chromosome
inactivation, have also been subjects of evo-
lutionary study (e.g., Jablonka and Lamb
1995; Lyon 1998; Haig 2002; Jablonka
2004a; Wolf and Hager 2006). However,
other developmental processes that de-
pend on EISs (paramutation and stress-
induced epigenomic alterations, for exam-
ple) have not yet received much attention
from evolutionary biologists, and the adap-
tive significance—if any—of epigenetic
mechanisms leading to systemic changes
during periods of genomic and ecological
stresses is at present an open question.
The stability of epigenetic transmission
is likely to be an evolved trait that depends
on the relative cost of error and the cost
of development (Rando and Verstrepen
2007). Epigenetic recall may be selectively
superior to full epigenetic inheritance in
environments that change every few gener-
ations because the cost of response-error
that occurs when memory is perfect is re-
duced, as is the cost of development-from-
scratch, which occurs when reset is com-
plete and full induction is required. The
transmission of epigenetic engrams that
lead to an inducer-requiring yet facilitated
response may therefore often be an opti-
mal compromise between the danger of a
tyrannically good memory, on the one
hand, and the expensive response-delay
that comes with ”forgetting” too thor-
oughly, on the other. Direct evidence for
epigenetic recall is needed, however, and
theoretical exploration through modeling
might point to biological systems with strat-
egies that would qualify them as good tar-
gets for empirical research.
Epigenetic inheritance should be fa-
vored in fluctuating environmental condi-
tions that last for more than one genera-
tion (but not for very long) and may be
particularly important in the type of envi-
ronments experienced by many microor-
ganisms (Lachmann and Jablonka 1996;
Balaban et al. 2004; Lewis 2007; Rando and
Verstrepen 2007). In such fluctuating en-
vironments, efficient epigenetic inheri-
tance is likely to evolve (i) if the parental
environment carries reliable information
about the offspring’s environment (Jab-
lonka et al. 1995), (ii) when the response
to induction is lengthy and incurs a very
high cost (Lachmann and Jablonka 1996),
and (iii) when recall is not an option or
incurs too high a cost.
theoretical and practical
implications
Incorporating epigenetic inheritance into
evolutionary theory extends the scope of evo-
lutionary thinking and leads to notions of
heredity and evolution that incorporate de-
velopment. Dobzhansky’s definition of evo-
lution as “a change in the genetic composi-
tion of populations” (1937, p.11) appears to
be too narrow because it does not incorpo-
rate all sources of heritable variations. Both
evolution and heredity need to be redefined.
Jablonka and Lamb (2007a,b,c) suggested
that evolution should be redefined as the
June 2009 167TRANSGENERATIONAL EPIGENETIC INHERITANCE

set of processes that lead to changes in the nature
and frequency of heritable types in a population,
and heredity as the developmental reconstruc-
tion processes that link ancestors and descen-
dants and lead to similarity between them.
These deliberately broad redefinitions al-
low evolutionary possibilities denied by the
“Modern Synthesis” version of evolutionary
theory, which states that variations are
blind, are genetic (nucleic acid-based), and
that saltational events do not significantly con-
tribute to evolutionary change (Mayr 1982).
The epigenetic perspective challenges all these
assumptions, and it seems that a new extended
theory, informed by developmental studies
and epigenetic inheritance, and incorporating
Darwinian, Lamarckian, and saltational frame-
works, is going to replace the Modern Synthe-
sis version of evolution (Jablonka and Lamb
2005, 2007c). We believe, therefore, that the
impact of epigenetics and epigenetic inheri-
tance on evolutionary theory and the philoso-
phy of biology will be profound.
As we noted earlier, it is now recognized
that epigenetic inheritance is relevant for
ecology, and new methods and approaches
to the research questions to which it points
should be developed (Bossdorf et al.
2008). The relevance of epigenetic varia-
tions to biodiversity in our rapidly chang-
ing world is also of obvious interest and
clearly has to be explored.
A discussion of the implications of epige-
netic studies for medicine is beyond the
scope of this review, but since epigenetic de-
fects can be transmitted between genera-
tions of cells and individuals, we direct the
reader’s attention to some recent reviews.
Baylin and Jones (2007) review the epigenet-
ics of cancer, and Zoghbi and Beaudet
(2007) review diseases caused by defects in
chromatin marking and imprinting. The epi-
genetic aspects of metabolic diseases and
their transgenerational effects are also being
intensely studied (see Bateson et al. 2004;
Gluckman and Hanson 2005; Gluckman et
al. 2007; Petronis 2004, 2006). The epidemi-
ological aspects of epigenetic inheritance
were reviewed by Jablonka (2004b), and the
importance of epigenetics for aging research
has been discussed by Vanyushin (1973),
Holliday (1984), Lamb (1994), and Issa
(2000). The recently reported ability of
pathogenic microorganisms to evolve herita-
ble epigenetic resistance to medication (e.g.,
antibiotics) may be of major medical impor-
tance (Adam et al 2008), and the relevance
of epigenetic inheritance for therapeutic
cloning and nuclear transplantation in ani-
mals, including humans, is self-evident (see
Jaenisch and Gurdon 2007).
Heredity is a fundamental property of liv-
ing organisms. It is therefore not surprising
that, in the beginning of the last century, the
rediscovery of Mendel’s laws and the chro-
mosomal mechanisms underlying them led
to profound changes in all branches of biol-
ogy. Today, at the dawn of the 21st century,
another aspect of heredity—epigenetic in-
heritance and the epigenetic control mech-
anisms underlying it—is being unravelled.
Like the early 20th-century discoveries, it,
too, is driving a great expansion and trans-
formation in our understanding of biology.
acknowledgments
We would like to thank the many colleagues who
generously provided us with greatly needed informa-
tion and feedback: Nathalie Q. Balaban, Daniel Beny-
shek, Renee Borges, Graham C. Burdge, Giacomo
Cavalli, Vincent Colot, Gyorgy Csaba, Scott Gilbert,
Janine Guespin-Michel, Rudolf Hagemann, Luisa
Hirschbein, David Martin, Marjorie Matzke, Fred
Meins, Jean Molinier, John Parker, Minoo Rassoulza-
degan, Wolf Reik, Yih-Horng Shiao, Michael Skinner,
Nadine Vastenhouw, Robert A. Waterland, and
Guangtian Zou. We also thank Maya Raz for her con-
tribution to the graphics. We are especially grateful to
Marion Lamb for reading (and re-reading) the whole
manuscript, and for her invaluable critical comments
and constructive suggestions. Finally, we would like to
thank Massimo Pigliucci and two anonymous review-
ers for their useful and critical comments.
REFERENCES
Adam M., Murali B., Glenn N. O., Potter S. S. 2008.
Epigenetic inheritance based evolution of antibi-
otic resistance in bacteria. BMC Evolutionary Biol-
ogy 8:52.
168 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

Agrawal A. A., Laforsch C., Tollrian R. 1999. Trans-
generational induction of defences in animals and
plants. Nature 401(6748):60–63.
Alleman M., Doctor J. 2000. Genomic imprinting in
plants: observations and evolutionary implica-
tions. Plant Molecular Biology 43(2–3):147–161.
Allen J. A., editor. 2004. Issue on plant polyploidy.
Biological Journal of the Linnean Society 82(4):409–
700.
Allen N. D., Norris M. L., Surani M. A. 1990. Epige-
netic control of transgene expression and imprint-
ing by genotype-specific modifiers. Cell 61(5):853–
861.
Allis C. D., Jenuwein T., Reinberg D., Caparros M.-L.,
editors. 2007. Epigenetics. Cold Spring Harbor
(NY): Cold Spring Harbor Laboratory Press.
Allshire R. C., Selker E. U. 2007. Fungal models for
epigenetic research: Schizosaccharomyces pombe and
Neurospora crassa. Pages 101–125 in Epigenetics, ed-
ited by C. D. Allis et al. Cold Spring Harbor (NY):
Cold Spring Harbor Laboratory Press.
Anway M. D., Cupp A. S., Uzumcu M., Skinner M. K.
2005. Epigenetic transgenerational actions of en-
docrine disruptors and male fertility. Science
308(5727):1466–1469.
Anway M. D., Leathers C., Skinner M. K. 2006a. En-
docrine disruptor vinclozolin induced epigenetic
transgenerational adult-onset disease. Endocrinol-
ogy 147(12):5515–5523.
Anway M. D., Memon M. A., Uzumcu M., Skinner
M. K. 2006b. Transgenerational effect of the en-
docrine disruptor vinclozolin on male spermato-
genesis. Journal of Andrology 27(6):868–879.
Avital E., Jablonka E. 2000. Animal Traditions: Behav-
ioural Inheritance in Evolution. Cambridge (UK):
Cambridge University Press.
Badyaev A. V. 2005. Stress-induced variation in evolu-
tion: from behavioural plasticity to genetic assim-
ilation. Proceedings of the Royal Society of London,
Series B: Biological Sciences 272(1566):877–886.
Balaban N. Q., Merrin Q. J., Chait R., Kowalik L.,
Leibler S. 2004. Bacterial persistence as a pheno-
typic switch. Science 305(5690):1622–1625.
Bantignies F., Cavalli G. 2006. Cellular memory and
dynamic regulation of polycomb group proteins.
Current Opinion in Cell Biology 18(3):275–283.
Barlow D. P., Bartolomei M. S. 2007. Genomic im-
printing in mammals. Pages 357–373 in Epigenetics,
edited by C. D. Allis et al. Cold Spring Harbor
(NY): Cold Spring Harbor Laboratory Press.
Bateman R. M., DiMichele W. A. 2002. Generating
and filtering major phenotypic novelties: neo-
Goldschmidtian saltation revisited. Pages 109–159
in Developmental Genetics and Plant Evolution, edited
by Q. C. B. Cronk et al. London (UK): Taylor &
Francis.
Bateson P., Barker D., Clutton-Brock T., Deb D.,
D’Udine B., Foley R. A., Gluckman P., et al. 2004.
Developmental plasticity and human health. Na-
ture 430(6998):419–421.
Baulcombe D. C. 2007. Amplified silencing. Science
315(5809):199–200.
Baxa U., Cassese T., Kajava A. V., Steven A. C. 2006.
Structure, function, and amyloididogenesis of fun-
gal prions: filament polymorphisms and prions
variants. Advances in Protein Chemistry 73:125–180
Baylin S. B., Jones P. A. 2007. Epigenetic determi-
nants of cancer. Pages 457–476 in Epigenetics, ed-
ited by C. D. Allis et al. Cold Spring Harbor (NY):
Cold Spring Harbor Laboratory Press.
Bean C. J., Schaner C. E., Kelly W. G. 2004. Meiotic
pairing and imprinted X chromatin assembly in
Caenorhabditis elegans.Nature Genetics 36(1):100–
105.
Beisson J., Sonneborn T. M. 1965. Cytoplasmic inher-
itance of the organization of the cell cortex in
Paramecium aurelia.Proceedings of the National Acad-
emy of Sciences USA 53(2):275–282.
Belyaev D. K., Borodin P. M. 1982. The influence of
stress on variation and its role in evolution. Biolo-
gisches Zentralblatt 100:705–714.
Belyaev D. K., Ruvinsky A. O., Borodin P. M. 1981a.
Inheritance of alternative states of the fused gene
in mice. Journal of Heredity 72:107–112.
Belyaev D. K., Ruvinsky A. O., Trut L. N. 1981b. In-
herited activation-inactivation of the star gene in
foxes: its bearing on the problem of domestica-
tion. Journal of Heredity 72(4):267–274.
Belyaev D. K., Ruvinsky A. O., Agulnik A. I., Agulnik
S. I. 1983. Effect of hydrocortisone on the fre-
quency and inheritance of the fused gene in mice.
Genetika 64(4):275–281.
Benkemoun L., Saupe S. J. 2006. Prion proteins as
genetic material in fungi. Fungal Genetics and Biol-
ogy 43(12):789–803.
Bennett S. T., Wilson A. J., Esposito L., Bouzekri N.,
Undlien D. E., Cucca F., Nistico` L., et al. 1997.
Insulin VNTR allele-specific effect in type 1 diabe-
tes depends on identity of untransmitted paternal
allele. Nature Genetics 17(3):350–352.
Berger S. L. 2007. The complex language of chromatin
regulation during transcription. Nature 447(7143):
407–412.
Bestor T. H. 1990. DNA methylation: evolution of
a bacterial immune function into a regulator
of gene expression and genome structure in
higher eukaryotes. Philosophical Transactions of the
Royal Society of London, Series B: Biological Sciences
326(1235):179–187.
Bird A. P. 1995. Gene number, noise reduction and
biological complexity. Trends in Genetics 11(3):94–
100.
Blasco M. A. 2007. The epigenetic regulation of mam-
June 2009 169TRANSGENERATIONAL EPIGENETIC INHERITANCE

malian telomeres. Nature Reviews Genetics 8(4):299–
309.
Blewitt M. E., Vickaryous N. K., Paldi A., Koseki H.,
Whitelaw E. 2006. Dynamic reprogramming of
DNA methylation at an epigenetically sensitive al-
lele in mice. PLoS Genetics 2(4):399–405.
Bossdorf O., Richards C. L., Pigliucci M. 2008. Epige-
netics for ecologists. Ecology Letters 11(2):106–115.
Brink R. A. 1973. Paramutation. Annual Review of Ge-
netics 7:129–152.
Buiting K., Gro
␤
S., Lich C., Gillessen-Kaesbach G.,
El-Maarri O., Horsthemke B. 2003. Epimutations
in Prader-Willi and Angelman syndromes: a mo-
lecular study of 136 patients with an imprinting
defect. American Journal of Human Genetics 72(3):
571–577.
Burdge G. C., Slater-Jefferies J., Torrens C., Phillips
E. S., Hanson M. A., Lillycrop K. A. 2007. Dietary
protein restriction of pregnant rats in the F0gen-
eration induces altered methylation of hepatic
gene promoters in the adult male offspring in the
F1and F2generations. British Journal of Nutrition
97(3):435–439.
Buss L. W. 1987. The Evolution of Individuality. Prince-
ton (NJ): Princeton University Press.
Campbell P., editor. 2007. Follow the curve: the forces
that shape cell membranes. Nature 447(7143 In-
sight):353–500.
Casadesu´s J., Low D. 2006. Epigenetic gene regulation
in the bacterial world. Microbiology and Molecular
Biology Reviews 70(3):830–856.
Cavalier-Smith T. 2004. The membranome and mem-
brane heredity in development and evolution.
Pages 335–351 in Organelles, Genomes and Eukaryote
Phylogeny, edited by R. P. Hirt and D. S. Horner.
Boca Raton (FL): CRC Press.
Cerutti H., Casas-Mollano J. A. 2006. On the origin
and functions of RNA-mediated silencing: from
protists to man. Current Genetics 50(2):81–99.
Chandler V. L. 2007. Paramutation: from maize to
mice. Cell 128(4):641–645.
Chandler V. L., Eggleston W. B., Dorweiler J. E. 2000.
Paramutation in maize. Plant Molecular Biology
43(2–3):121–145.
Chang H.-S., Anway M. D., Rekow S. S., Skinner M. K.
2006. Transgenerational epigenetic imprinting of
the male germline by endocrine disruptor expo-
sure during gonadal sex determination. Endocri-
nology 147(12):5524–5541.
Chien P., Weissman J. S., DePace A. H. 2004. Emerg-
ing principles of conformation-based prion inher-
itance. Annual Review of Biochemistry 73:617–656.
Chong S., Vickaryous N., Ashe A., Zamudio N.,
Youngson N., Hemley S., Stopka T., et al. 2007.
Modifiers of epigenetic reprogramming show pa-
ternal effects in the mouse. Nature Genetics 39(5):
614–622.
C
ˇikosˇS
ˇ., Vesela´ J., Il’kova´ G., Reha´k P., Czikkova´ S.,
Koppel J. 2005. Expression of beta adrenergic re-
ceptors in mouse oocytes and preimplantation
embryos. Molecular Reproduction and Development
71(2):145–153.
Colot V., Rossignol J. L. 1999. Eukaryotic DNA meth-
ylation as an evolutionary device. BioEssays 21(5):
402–411.
Cormier S., Vandormael-Pournin S., Babinet C.,
Cohen-Tannoudji M. 2004. Developmental ex-
pression of the Notch signaling pathway genes
during mouse preimplantation development. Gene
Expression Patterns 4(6):713–717.
Costa Y., Speed R. M., Gautier P., Semple C. A., Ma-
ratou K., Turner J. M. A., Cooke H. J. 2006. Mouse
MAELSTROM: the link between meiotic silencing
of unsynapsed chromatin and microRNA path-
way? Human Molecular Genetics 15(15):2324–2334.
Crews D., Gore A. C., Hsu T. S., Dangleben N. L.,
Spinetta M., Schallert T., Anway M. D., Skinner
M. K. 2007. Transgenerational epigenetic imprints
on mate preference. Proceedings of the National
Academy of Sciences USA 104(14):5942–5946.
Cropley J. E., Suter C. M., Beckman K. B., Martin
D. I. K. 2006. Germ-line epigenetic modification
of the murine Avy allele by nutritional supplemen-
tation. Proceedings of the National Academy of Sciences
USA 103(46):17308–17312.
Csaba G., Karabe´lyos C., Inczefi-Gonda A., Pallinger
E. 2005. Three-generation investigation on seroto-
nin content in rat immune cells long after
␤
-endorphin exposure in late pregnancy. Hormone
and Metabolic Research 37:172–177.
Cullen B. R. 2004. Transcription and processing
of human microRNA precursors. Molecular Cell
16(6):861–865.
Cullis C. A. 2005. Mechanisms and control of rapid
genomic changes in flax. Annals of Botany 95(1):
201–206.
Delage Y., Goldsmith M. 1912. The Theories of Evolution.
Trans. Andre´ Tridon. New York: B. W. Huebsch.
Dobzhansky T. 1937. Genetics and the Origin of Species.
New York: Columbia University Press.
Dubnau D., Losick R. 2006. Bistability in bacteria.
Molecular Microbiology 61(3):564–572.
Dubrova Y. E. 2003. Radiation-induced transgenera-
tional instability. Oncogene 22(45):7087–7093.
Ekwall K. 2007. ‘Arc’ escorts siRNAs in heterochro-
matin assembly. Nature Structural and Molecular Bi-
ology 14(3):178–179.
Felsenfeld G., Groudine M. 2003. Controlling the
double helix. Nature 421(6921):448–453.
Flanagan J. M., Popendikyte V., Pozdniakovaite N.,
Sobolev M., Assadzadeh A., Schumacher A., Zan-
geneh M., et al. 2006. Intra- and interindividual
epigenetic variation in human germ cells. Ameri-
can Journal of Human Genetics 79(1):67–84.
170 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

Flintoft L., editor. 2007. Focus on: epigenetics. Nature
Reviews Genetics 8(4):245–314.
Fontdevila A. 2005. Hybrid genome evolution by
transposition. Cytogenetic and Genome Research
110(1):49–55.
Fothergill P. G. 1952. Historical Aspects of Organic Evo-
lution. London (UK): Hollis and Carter.
Frankel J. 1989. Pattern Formation Ciliate Studies and
Models. New York: Oxford University Press.
Genereux D. P., Miner B. E., Bergstrom C. T., Laird
C. D. 2005. A population-epigenetic model to in-
fer site-specific methylation rates from double-
stranded DNA methylation patterns. Proceedings of
the National Academy of Sciences USA 102(16):5802–
5807.
Ginsburg S., Jablonka E. 2008. Epigenetic learning in
non-neural organisms. Journal of Biosciences 33(4).
Gluckman P., Hanson M. 2005. The Fetal Matrix: Evo-
lution, Development and Disease. Cambridge (UK):
Cambridge University Press.
Gluckman P. D., Hanson M. A., Beedle A. S. 2007.
Non-genomic transgenerational inheritance of
disease risk. BioEssays 29(2):145–154.
Goldschmidt R. B. 1940. The Material Basis of Evolution.
New Haven (CT): Yale University Press.
Gottlieb G. 2002. Developmental-behavioral initiation
of evolutionary change. Psychological Review 109(2):
211–218.
Grandjean V., Hauck Y., Beloin C., Le He´garat F.,
Hirschbein L. 1998. Chromosomal inactivation of
Bacillus subtilis exfusants: a prokaryotic model of
epigenetic regulation. Biological Chemistry 379(4–
5):553–557.
Grant-Downton R. T., Dickinson H. G. 2005. Epige-
netics and its implications for plant biology. 1.
The epigenetic network in plants. Annals of Botany
96(7):1143–1164.
Grant-Downton R. T., Dickinson H. G. 2006. Epige-
netics and its implications for plant biology. 2.
The ‘epigenetic epiphany’: epigenetics, evolution
and beyond. Annals of Botany 97(1):11–27.
Grewal S. I. S., Jia S. T. 2007. Heterochromatin revis-
ited. Nature Reviews Genetics 8(1):35–46.
Grimes G., Aufderheide K. 1991. Cellular aspects of
pattern formation: the problem of assembly.
Monographs in Developmental Biology 22:1–94.
Groth A., Rocha W., Verreault A., Almouzni G. 2007.
Chromatin challenges during DNA replication
and repair. Cell 128(4):721–733.
Gu Z., Rifkin S. A., White K. P., and Li W.-H. 2004.
Duplicate genes increase gene expression diversity
within and between species. Nature Genetics 36(6):
577–579.
Hadchouel M., Farza H., Simon D., Tiollais P., Pour-
cel C. 1987. Maternal inhibition of hepatitis B
surface antigen gene expression in transgenic
mice correlates with de novo methylation. Nature
329(6138):454–456.
Haig D. 2002. Genomic Imprinting and Kinship. New
Brunswick (NJ): Rutgers University Press.
Haig D. 2004. The (dual) origin of epigenetics. Cold
Spring Harbor Symposia on Quantitative Biology 69:
67–70.
Hajkova P., Erhardt S., Lane N., Haaf T., El-Maarri O.,
Reik W., Walter J., Surani M. A. 2002. Epigenetic
reprogramming in mouse primordial germ cells.
Mechanisms of Development 117(1–2):15–23.
Henderson I. R., Jacobsen S. E. 2007. Epigenetic in-
heritance in plants. Nature 447(7143):418–424.
Henikoff S., Ahmad K., Malik H. S. 2001. The centro-
mere paradox: stable inheritance with rapidly
evolving DNA. Science 293(5532):1098–1102.
Henikoff S., Smith M. M. 2007. Histone variants and
epigenetics. Pages 249–264 in Epigenetics, edited
by C. D. Allis et al. Cold Spring Harbor (NY): Cold
Spring Harbor Laboratory Press.
Himber C., Dunoyer P., Moissiard G., Ritzenthaler
C., Voinnet O. 2003. Transitivity-dependent and
-independent cell-to-cell movement of RNA si-
lencing. EMBO Journal 22(17):4523–4533.
Hitchins M. P., Wong J. J. L., Suthers G., Suter C. M.,
Martin D. I. K., Hawkins N. J., Ward R. L. 2007.
Brief report: inheritance of a cancer-associated
MLH1 germ-line epimutation. New England Journal
of Medicine 356(7):697–705.
Holliday R. 1984. The significance of DNA methyl-
ation in cellular aging. Pages 269–283 in Molecular
Biology of Aging, edited by A. D. Woodhead et al.
New York: Plenum Press.
Holliday R. 1994. Epigenetics—an overview. Develop-
mental Genetics 15:453–457.
Holliday R. 2002. Epigenetics come of age in the
twenty-first century. Journal of Genetics 81(1):1–4.
Holliday R. 2006. Epigenetics: a historical overview.
Epigenetics 1(2):76–80.
Huettel B., Kanno T., Daxinger L., Bucher E., van der
Winden J., Matzke A. J., Matzke M. 2007. RNA-
directed DNA methylation mediated by DRD1 and
Pol IVb: a versatile pathway for transcriptional gene
silencing in plants. Biochimica et Biophysica Acta-Gene
Structure and Expression 1769(5–6):358–374.
Issa J. P. 2000. CpG-island methylation in aging and
cancer. Current Topics in Microbiology and Immunol-
ogy 249:101–118.
Jablonka E. 2004a. The evolution of the peculiarities
of mammalian sex chromosomes: an epigenetic
view. BioEssays 26(12):1327–1332.
Jablonka E. 2004b. Epigenetic epidemiology. Interna-
tional Journal of Epidemiology 33(5):929–935.
Jablonka E., Lachmann M., Lamb M. J. 1992. Evi-
dence, mechanisms and models for the inheri-
tance of acquired characters. Journal of Theoretical
Biology 158(2):245–268.
June 2009 171TRANSGENERATIONAL EPIGENETIC INHERITANCE

Jablonka E., Lamb M. J. 1989. The inheritance of
acquired epigenetic variations. Journal of Theoreti-
cal Biology 139(1):69–83.
Jablonka E., Lamb M. J. 1995. Epigenetic Inheritance and
Evolution: The Lamarckian Dimension. Oxford (UK):
Oxford University Press.
Jablonka E., Lamb M. J. 2002. The changing concept
of epigenetics. Annals of the New York Academy of
Sciences 981:82–96.
Jablonka E., Lamb M. J. 2005. Evolution in Four Dimen-
sions: Genetic, Epigenetic, Behavioral, and Symbolic
Variation in the History of Life. Cambridge (MA):
MIT Press.
Jablonka E., Lamb M. J. 2006a. Evolutionary epigenet-
ics. Pages 252–264 in Evolutionary Genetics: Concepts
and Case Studies, edited by C. W. Fox and J. B. Wolf.
New York: Oxford University Press.
Jablonka E., Lamb M. J. 2006b. The evolution of
information in the major transitions. Journal of
Theoretical Biology 239(2):236–246.
Jablonka E., Lamb M. J. 2007a. Pre´cis of Evolution in
Four Dimensions.Behavioral and Brain Sciences 30:
353–365.
Jablonka E., Lamb M. J. 2007b. Bridging the gap: the
developmental aspects of evolution. Behavioral and
Brain Sciences 30:378–392.
Jablonka E., Lamb M. J. 2007c. The expanded evolu-
tionary synthesis—a response to Godfrey-Smith,
Haig, and West-Eberhard. Biology and Philosophy
22(3):453–472.
Jablonka E., Lamb M. J. 2008. The epigenome in
evolution: beyond the modern synthesis. VOGis
Herald 12:242–254.
Jablonka E., Oborny B., Molnar I., Kisdi E., Hofbauer
J., Czaran T. 1995. The adaptive advantage of phe-
notypic memory in changing environments. Philo-
sophical Transactions of the Royal Society of London,
Series B: Biological Sciences 350(1332):133–141.
Jaenisch R., Gurdon J. 2007. Nuclear transplantation
and the reprogramming of the genome. Pages
415–434 in Epigenetics, edited by C. D. Allis et al.
Cold Spring Harbor (NY): Cold Spring Harbor
Laboratory Press.
Jirtle R. L., Skinner M. K. 2007. Environmental epi-
genomics and disease susceptibility. Nature Reviews
Genetics 8(4):253–262.
Jollos V. Experimentelle Protistenstudien 1. 1921. Un-
tersuchungen u¨ber Variabilita¨t und Vererbung
bei Infusorien. Archiv fu¨r Protistenkunde 43:1–222.
Jorgensen R. A. 2004. Restructuring the genome in
response to adaptive challenge: McClintock’s bold
conjecture revisited. Cold Spring Harbor Symposia on
Quantitative Biology 69:349–354.
Karpen G. H., Hawley R. S. 2007. Epigenetic regula-
tion of chromosome inheritance. Pages 265–289
in Epigenetics, edited by C. D. Allis et al. Cold
Spring Harbor (NY): Cold Spring Harbor Labora-
tory Press.
Kidwell M. G., Lisch D. 1997. Transposable elements
as sources of variation in animals and plants. Pro-
ceedings of the National Academy of Sciences USA
94(15):7704–7711.
Kim V. N. 2006. Small RNAs just got bigger: Piwi-
interacting RNAs (piRNAs) in mammalian testes.
Genes and Development 20(15):1993–1997.
Ko¨lle S., Stojkovic M., Prelle K., Waters M., Wolf E.,
Sinowatz F. 2001. Growth hormone (GH)/GH re-
ceptor expression and GH-mediated effects dur-
ing early bovine embryogenesis. Biology of Repro-
duction 64:1826–1834.
Kussell E., Leibler S. 2005. Phenotypic diversity, pop-
ulation growth, and information in fluctuating en-
vironments. Science 309(5743):2075–2078.
Lachmann M., Jablonka E. 1996. The inheritance of
phenotypes: an adaptation to fluctuating environ-
ments. Journal of Theoretical Biology 181:1–9.
Lamb M. J. 1994. Epigenetic inheritance and aging.
Reviews in Clinical Gerontology 4(2):97–105.
Lamm E., Jablonka E. 2008. The nurture of nature:
hereditary plasticity in evolution. Philosophical Psy-
chology 21(3):305–319.
Laurent M., Charvin G., Guespin-Michel J. 2005. Bi-
stability and hysteresis in epigenetic regulation of
the lactose operon. Since Delbruck, a long series
of ignored models. Cellular and Molecular Biology
51(7):583–594.
Levites E. V. 2000. Epigenetic variability as a source of
biodiversity and a factor of evolution. Pages 73–75
in Biodiversity and Dynamics of Ecosystems in North
Eurasia, Volume 1, Part 3. Novosibirsk (Russia):
Institute of Cytology and Genetics SB RAS.
Levy A. A., Feldman M. 2004. Genetic and epigenetic
reprogramming of the wheat genome upon al-
lopolyploidization. Biological Journal of the Linnean
Society 82(4):607–613.
Li E., Bird A. 2007. DNA methylation in mammals.
Pages 341–356 in Epigenetics, edited by C. D. Allis
et al. Cold Spring Harbor (NY): Cold Spring Har-
bor Laboratory Press.
Lindegren C. C. 1949. The Yeast Cell, its Genetics and
Cytology. Saint Louis (MO): Educational Publishers.
Lindqvist C., Janczak A. M., Na¨tt D., Baranowska I.,
Lindqvist N., Wichman A., Lundeberg J., Lind-
berg J., Torjesen P. A., Jensen P. 2007. Transmis-
sion of stress-induced learning impairment and
associated brain gene expression from parents to
offspring in chickens. PLoS ONE 2(4):e364.
Little J. B. 2003. Genomic instability and bystander
effects: a historical perspective. Oncogene 22(45):
6978–6987.
Logan C. A. 2007. Overheated rats, race, and the
double gland: Paul Kammerer, endocrinology and
172 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

the problem of somatic induction. Journal of the
History of Biology 40(4):683–725.
Long L., Lin X., Zhai J., Kou H., Yang W., Liu B. 2006.
Heritable alteration in DNA methylation pattern
occurred specifically at mobile elements in rice
plants following hydrostatic pressurization. Bio-
chemical and Biophysical Research Communications
340(2):369–376.
Lyon M. F. 1998. X-chromosome inactivation: a re-
peat hypothesis. Cytogenetics and Cell Genetics 80:
133–137.
Malagnac F., Silar P. 2003. Non-Mendelian determi-
nants of morphology in fungi. Current Opinion in
Microbiology 6(6):641–645.
Malagnac F., Silar P. 2006. Regulation, cell differen-
tiation and protein-based inheritance. Cell Cycle
5(22):2584–2587.
Mandrioli M. 2004. Epigenetic tinkering and evolu-
tion: is there any continuity in the role of cytosine
methylation from invertebrates to vertebrates? Cel-
lular and Molecular Life Sciences 61(19–20):2425–
2427.
Marcus E., editor. 2007. Issue on epigenetics. Cell
128(4):627–802.
Martin C., Zhang Y. 2007. Mechanisms of epigenetic
inheritance. Current Opinion in Cell Biology 19(3):
266–272.
Masel J., Bergman A. 2003. The evolution of the evolv-
ability properties of the yeast prion [PSI⫹]. Evolu-
tion 57(7):1498–1512.
Matzke M. A., Birchler J. A. 2005. RNAi-mediated
pathways in the nucleus. Nature Reviews Genetics
6(1):24–35.
Matzke M. A., Scheid O. M. 2007. Epigenetic regula-
tion in plants. Pages 167–189 in Epigenetics, edited
by C. D. Allis et al. Cold Spring Harbor (NY): Cold
Spring Harbor Laboratory Press.
Maynard Smith J. 1990. Models of a dual inheritance
system. Journal of Theoretical Biology 143(1):41–53.
Maynard Smith J., Szathma´ry E. 1995. The Major Tran-
sitions in Evolution. Oxford (UK) and New York:
W. H. Freeman Spektrum.
Mayr E. 1982. The Growth of Biological Thought: Diver-
sity, Evolution, and Inheritance. Cambridge (MA):
Belknap Press.
McClintock B. 1984. The significance of responses of
the genome to challenge. Science 226(4676):792–
801.
Meister G., Tuschl T. 2004. Mechanisms of gene silenc-
ing by double-stranded RNA. Nature 431(7006):343–
349.
Meyer E., Chalker D. L. 2007. Epigenetics of ciliates.
Pages 127–150 in Epigenetics, edited by C. D. Allis
et al. Cold Spring Harbor (NY): Cold Spring Har-
bor Laboratory Press.
Misteli T. 2007. Beyond the sequence: cellular orga-
nization of genome function. Cell 128:787–800.
Molinier J., Ries G., Zipfel C., Hohn B. 2006. Trans-
generation memory of stress in plants. Nature 442:
1046–1049.
Morris C. A., Moazed D. 2007. Centromere assembly
and propagation. Cell 128(4):647–650.
Mortusewicz O., Schermelleh L., Walter J., Cardoso
M. C., Leonhardt H. 2005. Recruitment of DNA
methyltransferase I to DNA repair sites. Proceedings
of the National Academy of Sciences USA 102(25):
8905–8909.
Nanney D. L. 1958. Epigenetic control systems. Pro-
ceedings of the National Academy of Sciences USA
44(7):712–717.
Nanney D. L. 1968. Cortical patterns in cellular mor-
phogenesis: differences in cortical patterns in cil-
iates may be hereditary, but independent of genic
differences. Science 160(3827):496–502.
Nanney D. L. 1985. Heredity without genes: ciliate
explorations of clonal heredity. Trends in Genetics
1:295–298.
Naz R. K., Sellamuthu R. 2006. Receptors in sperma-
tozoa: are they real? Journal of Andrology 27(5):
627–636.
Newbold R. R., Padilla-Banks E., Jefferson W. N. 2006.
Adverse effects of the model environmental estro-
gen diethylstilbestrol are transmitted to subse-
quent generations. Endocrinology 147:S11–S17.
Novick A., Weiner M. 1957. Enzyme induction as an
all-or-none phenomenon. Proceedings of the Na-
tional Academy of Sciences USA 43(7):553–566.
Nowacki M., Vijayan V., Zhou Y., Schotanus K., Doak
T. G., Landweber L. F. 2008. RNA-mediated epige-
netic programming of a genome-rearrangement
pathway. Nature 451(7175):153–159.
Pak J., Fire A. 2007. Distinct populations of primary
and secondary effectors during RNAi in C. elegans.
Science 315(5809):241–244.
Pa´l C. 1998. Plasticity, memory and the adaptive land-
scape of the genotype. Proceedings of the Royal Society
of London, Series B: Biological Sciences 265(1403):
1319–1323.
Palauqui J. C., Elmayan T., Pollien J. M., Vaucheret H.
1997. Systemic acquired silencing: transgene-
specific post-transcriptional silencing is transmit-
ted by grafting from silenced stocks to non-
silenced scions. EMBO Journal 16(15):4738–4745.
Pantaleon M., Whiteside E. J., Harvey M. B., Barnard
R. T., Waters M. J., Kaye P. L. 1997. Functional
growth hormone (GH) receptors and GH are ex-
pressed by preimplantation mouse embryos: a role
for GN in early embryogenesis? Proceedings of the
National Academy of Sciences USA 94(10):5125–5130.
Pavelka J., Koudelova´ J. 2001. Inheritance of a
temperature-modified phenotype of the short an-
tennae (sa) mutation in a moth, Ephestia kuehniella
(Lepidoptera: Pyralidae). Journal of Heredity 92(3):
234–242.
June 2009 173TRANSGENERATIONAL EPIGENETIC INHERITANCE

Pepper J. W., Sprouffske K., Maley C. C. 2007. Animal
cell differentiation patterns suppress somatic evo-
lution. PloS Computational Biology 3(12):e250.
Petronis A. 2004. The origin of schizophrenia: genetic
thesis, epigenetic antithesis, and resolving synthe-
sis. Biological Psychiatry 55(10):965–970.
Petronis A. 2006 Epigenetics and twins: three variations
on the theme. Trends in Genetics 22(7):347–350.
Pigliucci M. 2001. Phenotypic Plasticity: Beyond Nature
and Nurture. Baltimore (MD): John Hopkins Uni-
versity Press.
Pigliucci M., Murren C. J., Schlichting C. D. 2006.
Phenotypic plasticity and evolution by genetic as-
similation. Journal of Experimental Biology 209(12):
2362–2367.
Pikaard C. S. 2000. The epigenetics of nucleolar dom-
inance. Trends in Genetics 16(11):495–500.
Pikaard C. S. 2003. Nucleolar dominance. Pages 399–
402 in Nature Encyclopedia of the Human Genome,
Volume 4, edited by D. N. Cooper. London (UK):
Nature Publishing Group.
Polo S. E., Roche D., Almouzni G. 2006. New histone
incorporation marks sites of UV repair in human
cells. Cell 127(3):481–493.
Popova N. K. 2006. From genes to aggressive behav-
ior: the role of serotonergic system. BioEssays
28(5):495–503.
Prusiner S. B. 1998. Prions. Proceedings of the National
Academy of Sciences USA 95(23):13363–13383.
Rando O. J., Verstrepen K. J. 2007. Timescales of
genetic and epigenetic inheritance. Cell 128(4):
655–668.
Rapp R. A., Wendel J. F. 2005. Epigenetics and plant
evolution. New Phytologist 168(1):81–91.
Rassoulzadegan M., Grandjean V., Gounon P., Vin-
cent S., Gillot I., Cuzin F. 2006. RNA-mediated
non-mendelian inheritance of an epigenetic
change in the mouse. Nature 441(7092):469–474.
Regev A., Lamb M. J., Jablonka E. 1998. The role of
DNA methylation in invertebrates: developmental
regulation or genome defense? Molecular Biology
and Evolution 15(7):880–891.
Richards E. J. 2006. Inherited epigenetic variation—
revisiting soft inheritance. Nature Reviews Genetics
7(5):395–401.
Richerson P. J., Boyd R. 2005. Not by Genes Alone: How
Culture Transformed Human Evolution. Chicago
(IL): University of Chicago Press.
Ringrose L., Paro R. 2004. Epigenetic regulation of
cellular memory by the polycomb and trithorax
group proteins. Annual Review of Genetics 38:413–
443.
Ringrose L., Paro R. 2007. Polycomb/trithorax re-
sponse elements and epigenetic memory of cell
identity. Development 134(2):223–232.
Roberts B. T., Wickner R. B. 2003. Heritable activity:
a prion that propagates by covalent autoactivation.
Genes and Development 17(17):2083–2087.
Rodin S. N., Parkhomchuk D. V., Riggs A. D. 2005.
Epigenetic changes and repositioning determine
the evolutionary fate of duplicated genes. Biochem-
istry (Moscow) 70(5):559–567.
Ruden D. M., Xiao L., Garfinkel M. D., Lu X. 2005.
Hsp90 and environmental impacts on epigenetic
states: a model for the trans-generational effects of
diethylstibesterol on uterine development and
cancer. Human Molecular Genetics 14(Review Issue
1):R149–R155.
Ruvinsky A. O., Lobkov Y. I., Belyaev D. K. 1983a.
Spontaneous and induced activation of genes af-
fecting the phenotypic expression of glucose
6-phosphate dehydrogenase in Daphnia pulex.1.
Intraclonal variations in the electrophoretic mo-
bility of G6PD. Molecular & General Genetics 189(3):
485–489.
Ruvinsky A. O., Lobkov Y. I., Belyaev D. K. 1983b.
Spontaneous and induced activation of genes af-
fecting the phenotypic expression of glucose
6-phosphate dehydrogenase in Daphnia pulex 2.
Glucose-induced changes in the electrophoretic
mobility of G6PD. Molecular & General Genetics 189:
490–494.
Salmon A., Ainouche M. L., Wendel J. F. 2005. Ge-
netic and epigenetic consequences of recent hy-
bridization and polyploidy in Spartina (Poaceae).
Molecular Ecology 14(4):1163–1175.
Sandman K., Pereira S. L., Reeve J. N. 1998. Diversity
of prokaryotic chromosomal proteins and the or-
igin of the nucleosome. Cellular and Molecular Life
Sciences 54(12):1350–1364.
Sangster T. A., Lindquist S., Queitsch C. 2004. Under
cover: causes, effects and implications of Hsp90-
mediated genetic capacitance. BioEssays 26(4):348–
362.
Sano H., Kamada I., Youssefian S., Katsumi M.,
Wabiko H. 1990. A single treatment of rice seed-
lings with 5-azacytidine induces heritable dwarf-
ism and undermethylation of genomic DNA. Mo-
lecular and General Genetics 220(3):441–447.
Schacter D. L. 2001. Forgotten Ideas, Neglected Pioneers:
Richard Semon and the Story of Memory. Philadelphia
(PA): Psychology Press.
Schermelleh L., Haemmer A., Spada F., Ro¨sing N.,
Meilinger D., Rothbauer U., Cardoso M., Leon-
hardt H. 2007. Dynamics of Dnmt1 interaction
with the replication machinery and its role in
postreplicative maintenance of DNA methylation.
Nucleic Acids Research 35(13):4301–4312.
Schlichting C. D., Pigliucci M. 1998 Phenotypic Evolu-
tion: A Reaction Norm Perspective. Sunderland (MA):
Sinauer.
Schmalhausen I. I. 1949. Factors of Evolution: The The-
174 Volume 84THE QUARTERLY REVIEW OF BIOLOGY

ory of Stabilizing Selection. Philadelphia (PA): Black-
iston.
Schuettengruber B., Chourrout D., Vervoort M., Le-
blanc B., Cavalli G. 2007. Genome regulation by
polycomb and trithorax proteins. Cell 128(4):735–
745.
Shapiro J. A. 1999. Genome system architecture and
natural genetic engineering in evolution. Annals of
the New York Academy of Sciences 870:23–35.
Shiao Y. H., Crawford E. B., Anderson L. M., Patel P.,
Ko K. 2005. Allele-specific germ cell epimutation
in the spacer promoter of the 45S ribosomal RNA
gene after Cr(III) exposure. Toxicology and Applied
Pharmacology 205(3):290–296.
Shiu P. K. T., Raju N. B., Zickler D., Metzenberg R. L.
2001. Meiotic silencing by unpaired DNA. Cell
107(7):905–916.
Shorter J., Lindquist S. 2005. Prions as adaptive con-
duits of memory and inheritance. Nature Reviews
Genetics 6(6):435–450.
Siegal M. L., Bergman A. 2006. Canalization. Pages
235–251 in Evolutionary Genetics, edited by C. W.
Fox and J. B. Wolf. New York: Oxford University
Press.
Smits W. K., Kuipers O. P., Veening J. W. 2006. Phe-
notypic variation in bacteria: the role of feedback
regulation. Nature Reviews Microbiology 4(4):259–
271.
Sokolov V. A. 2006. Imprinting in plants. Russian Jour-
nal of Genetics 42(9):1043–1052.
Steele E. J., Lindley R. A., Blanden R. V. 1998.
Lamarck’s Signature: How Retrogenes are Changing
Darwin’s Natural Selection Paradigm. Reading (MA):
Perseus Books.
Talbert P. B., Bryson T., Henikoff S. 2004. Adaptive
evolution of centromere proteins in plants and
animals. Journal of Biology 3:18.
True H. L., Lindquist S. L. 2000. A yeast prion pro-
vides a mechanism for genetic variation and phe-
notypic diversity. Nature 407(6803):477–483.
True H. L., Berlin I., Lindquist S. L. 2004. Epigenetic
regulation of translation reveals hidden genetic
variation to produce complex traits. Nature
431(7005):184–187.
Trut L. N., Plyusnina I. Z., Oskina I. N. 2004. An
experiment on fox domestication and debatable
issues of evolution of the dog. Russian Journal of
Genetics 40(6):644–655.
Tuite M. F., Cox B. S. 2006. The [PSI⫹] prion of yeast:
a problem of inheritance. Methods 39(1):9–22.
Turner J. M. A., Mahadevaiah S. K., Fernandez-
Capetillo O., Nussenzweig A., Xu X. L., Deng
C.-X., Burgoyne P. S. 2005. Silencing of unsyn-
apsed meiotic chromosomes in the mouse. Nature
Genetics 37(1):41–47.
Uptain S. M., Lindquist S. 2002. Prions as protein-
based genetic elements. Annual Review of Microbi-
ology 56:703–741.
Van der Heijden G. W., Derijck A., Ramos L., Giele
M., Van der Vlag J., de Boer P. 2006. Transmission
of modified nucleosomes from the mouse male
germline to the zygote and subsequent remodel-
ing of paternal chromatin. Developmental Biology
298(2):458–469.
Vanyushin B. F. 2006a. DNA methylation and epige-
netics. Russian Journal of Genetics 42(9):985–997.
Vanyushin B. F. 2006b. DNA methylation in plants.
Current Topics in Microbiology and Immunology 301:
67–122.
Vanyushin B. F., Nemirovsky L. E., Klimenko V. V.,
Vasiliev V. K., Belozersky A. N. 1973. The
5-methylcytosine in DNA of rats. Tissue and age
specificity and the changes induced by hydrocor-
tisone and other agents. Gerontologia 19(3):138–
152.
Vastenhouw N. L., Brunschwig K., Okihara K. L., Mu¨l-
ler F., Tijsterman M., Plasterk R. H. A. 2006. Long-
term gene silencing by RNAi. Nature 442(7105):
882.
Vesela´ J., Reha´k P., Mihalik J., Czikkova´ S., Pokorny´ J.,
Koppel J. 2003. Expression of serotonin receptors
in mouse oocytes and preimplantation embryos.
Physiological Research 52(2):223–228.
Vrana P. B., Fossella J. A., Matteson P., del Rio T.,
O’Neill M. J., Tilghman S. M. 2000. Genetic and
epigenetic incompatibilities underlie hybrid dys-
genesis in Peromyscus.Nature Genetics 25(1):120–
124.
Waddington C. H. 1957. The Strategy of the Genes: A
Discussion of Some Aspects of Theoretical Biology. Lon-
don (UK): Allen & Unwin.
Waddington C. H. 1968. The basic ideas of biology.
Pages 1–32 in Towards a Theoretical Biology: Prolego-
mena, edited by C. H. Waddington. Edinburgh
(UK): Edinburgh University Press.
Waddington C. H. 1975. The Evolution of an Evolution-
ist. Edinburgh (UK): Edinburgh University Press.
Waterland R. A., Travisano M., Tahiliani K. G. 2007.
Diet-induced hypermethylation at agouti viable yel-
low is not inherited transgenerationally through
the female. FASEB Journal 21(12):3380–3385.
Waterland R. A., Travisano M., Tahiliani K. G.,
Rached M. T., Mirza S. 2008. Methyl donor sup-
plementation prevents transgenerational amplifi-
cation of obesity. International Journal of Obesity 32:
1373–1379.
Waters E. R., Schaal B. A. 1996. Heat shock induces a
loss of rRNA-encoding DNA repeats in Brassica
nigra.Proceedings of the National Academy of Sciences
USA 93(4):1449–1452.
Weaver I. C. G., Cervoni N., Champagne F. A.,
D’Alessio A. C., Sharma S., Seckl J. R., Dymov S.,
Szyf M., Meaney M. J. 2004. Epigenetic program-
June 2009 175TRANSGENERATIONAL EPIGENETIC INHERITANCE

ming by maternal behavior. Nature Neuroscience
7(8):847–854.
West-Eberhard M. J. 2003. Developmental Plasticity and
Evolution. New York: Oxford University Press.
Wickner R. B. 1994. [URE3] as an altered URE2
protein: evidence for a prion analog in Saccharo-
myces cerevisiae.Science 264(5158):566–569.
Wickner R. B., Edskes H. K., Ross E. D., Pierce M. M.,
Baxa U., Brachmann A., Shewmaker F. 2004.
Prion genetics: new rules for a new kind of gene.
Annual Review of Genetics 38:681–707.
Wolf J. B., Hager R. 2006. A maternal–offspring co-
adaptation theory for the evolution of genomic
imprinting. PLoS Biology 4(12):2238–2243.
Wu C.-T., Morris J. R. 2001. Genes, genetics, and
epigenetics: a correspondence. Science 293(5532):
1103–1105.
Wu T. C. J., Wang L., Wan Y.-J. Y. 1992. Expression of
estrogen receptor gene in mouse oocyte and dur-
ing embryogenesis. Molecular Reproduction and De-
velopment 33(4):407–412.
Yi H., Riddle N. C., Stokes T. L., Woo H.-R., Richards
E. J. 2004. Induced and natural epigenetic varia-
tion. Cold Spring Harbor Symposia on Quantitative
Biology 69:155–160.
Zacharioudakis I., Gligoris T., Tzamarias D. 2007. A
yeast catabolic enzyme controls transcriptional
memory. Current Biology 17(23):2041–2046.
Zaratiegui M., Irvine D. V., Martienssen R. A. 2007. Non
coding RNAs and gene silencing. Cell 128(4):763–776.
Zhivotovsky L. A. 2002. A model of the early evolution
of soma-to-germline feedback. Journal of Theoretical
Biology 216(1):51–57.
Zoghbi H. Y., Beaudet A. L. 2007. Epigenetics and
human disease. Pages 435–456 in Epigenetics, ed-
ited by C. D. Allis et al. Cold Spring Harbor (NY):
Cold Spring Harbor Laboratory Press.
Zordan R. E., Galgoczy D. J., Johnson A. D. 2006. Epi-
genetic properties of white-opaque switching in Can-
dida albicans are based on a self-sustaining transcrip-
tional feedback loop. Proceedings of the National
Academy of Sciences USA 103(34):12807–12812.
Zuckerkandl E., Cavalli G. 2007. Combinatorial epi-
genetics, “junk DNA,” and the evolution of com-
plex organisms. Gene 390(1–2):232–242.
Zufall R. A., Robinson T., Katz L. A. 2005. Evolution
of developmentally regulated genome rearrange-
ments in eukaryotes. Journal of Experimental Zoology,
Part B: Molecular and Developmental Evolution
304B(5):448–455.
176 Volume 84THE QUARTERLY REVIEW OF BIOLOGY