Download Parental Legacy Determines Methylation and Expression of an

Cell, Vol. 50, 719-727,
August
29, 1987, Copyright
0 1997 by Cell Press
Parental Legacy Determines Methylation
and Expression of an Autosomal Transgene:
A Molecular Mechanism for Parental Imprinting
Judith L. Swain: Timothy A. Stewart,t*
and Philip Ledert§
* Department of Medicine
Duke University Medical Center
Durham, North Carolina 27710
r Department of Genetics
Harvard Medical School
§and Howard Hughes Medical Institute
Boston, Massachusetts 02115
We have created a transgenic
mouse strain in which
an autosomal transgene bearing elements of the RSV
LTR and a translocated
c-myc gene obeys very unusual rules. If the transgene is inherited from the male
parent, it is expressed In the heart and no other tissue.
If it is inherited from the female parent, it is not expressed at all. This pattern of expression
correlates
precisely with a parentally imprinted methylation
state
evident in all tissues. Methylation
of the transgene
is
acquired by its passage through the female parent and
eliminated
during gametogenesis
in the male. These
observations
provide direct molecular evidence that
autosomal gene expression can depend upon the sex
of the parent from which the gene is inherited. They
also provide a plausible mechanism for understanding
parental imprinting
that may be relevant to the failure
of parthenogenesis
in mammals, the apparent nonMendelian
behavior of some autosomal
genes, and
the role of methylation
in gene regulation.
nal and maternal genomes has been obtained through
genetic experiments. In these, the effects of parental inheritance of specific regions of the genome have been
evaluated using intercrosses between mice carrying either Robertsonian
or reciprocal translocations
of nonhomologous
chromosomes
(Cattanach and Kirk, 1965;
Cattanach, 1966; Searle and Beechey, 1976; Searle and
Beechey, 1965; Solter, 1967; Surani, 1967). The implication of these studies is that both maternal and paternal
contributions of specific chromosomal regions, and thus
specific genes, are necessary for normal embryonic development. These genes must somehow be “imprinted”
during inheritance from the parent. The distinctive information imparted by passage of these genes through the
male and female parents evidently allows them to act collaboratively in the embryo. This concept of differential expression of paternally and maternally derived genes might
explain the failure of successful parthenogenesis
in
mammals.
The molecular basis of parental imprinting is unknown
and difficult to understand in mammals. Thus it was particularly fortunate that, in the course of experiments designed to determine the in vivo effect of an activated c-myc
oncogene, we produced a transgenic mouse in which the
transgene is autosomally inherited, while its expression
(which occurs exclusively in the heart) is strictly dictated
by its parental origin. Furthermore, this pattern of expression correlates with one of two global patterns of methylation that are differentially imprinted on the transgene during passage through the male and female parents. The
study of this unique strain of transgenic mice forms the basis for this report.
Introduction
Results
There is growing evidence, largely from the mouse, that
the expression of certain genes is determined by whether
the gene is inherited from the male or female parent. This
concept, called parental imprinting,
is supported
by
studies demonstrating that both the paternal and maternal
genomes are necessary for normal embryonic development (Cattanach and Kirk, 1965; Cattanach, 1966; McGrath
and Solter, 1964a; McGrath and Solter, 1966; Mann and
Love&Badge, 1964; Renard and Babinet, 1966; Surani et
al., 1964; Surani et al., 1966a; Surani et al., 1966b; Surani,
1967). For example, nuclear transplantation
experiments
in fertilized mouse oocytes indicate that the two parental
genomes play complementary
roles during embryogenesis. The paternal genome appears to be relatively more
important for the development of the extraembryonic
tissues, while the maternal genome assumes greater importance for the development of the embryo proper (Barton
et al., 1964; Surani et al., 1964).
Additional support for the nonequivalence
of the pater-
Inheritance
of the Transgene
The recombinant plasmid used to produce the strain of
transgenic mice described in these studies is shown in
Figure 1. In addition to plasmid sequences, the construct
consists of a major portion of the Rous sarcoma virus
(RSV) LTR followed by a fusion gene cloned from the SW7
mouse plasmacytoma
cell line. The latter contains the
c-myc gene translocated into the immunoglobulin
locus.
Specifically, the plasmacytoma-derived
fragment contains the a constant and switch regions of the immunoglobulin heavy chain locus (but not the heavy chain enhancer) and the c-rnyc gene in which the 5’ portion of the
untranslated first exon has been deleted.
Genomic DNAs from the founder and over 700 descendant mice were subjected to restriction fragment pattern
analysis using a probe derived from a portion of the immunoglobulin gene (identified as the a constant region [Cal
in Figure 1). The results of these analyses indicated that
approximately 10 copies of the construct integrated into a
single genomic site where they are arranged in a head-totail, tandem array. Pedigree analyses indicated that the
transgene (barring unlikelyx-linked
pseudoautosomal
be-
Summary
*Present address: Genentech,
460 Point San Bruno
San Francisco,
California 94090.
Boulevard,
South
Cell
720
A
Figure 1. Diagrammatic
Representation
of the RSV-S107 Fusion Gene
Injected to Produce the Strain of liansgenic
Mouse Analyzed in This
Study
An EcoRl fragment from the mouse 5107 plasmacytoma
(Kirsch et al.,
1961) containing
immunoglobulin
and c-myc sequences
was ligated
into a derivative of the pRSVCAT plasmid (Gorman et al., 1962). The
final construct
consisted of a 2.2 kb fragment of pBR322, a 437 bp fragment of the 3’ RSV LTR. and a 17 kb fragment containing the immunoglobulin heavy chain and truncated
c-myc genes. Vertical hatching
withi
identifies coding sequences
he immunoglobulin
region. Disequences
within the c-myc gene.
agonal hatching identifies cod’
,mc(p’
The probe used for Souther?-analysis
of genomic DNA from the transgenie animals is illustrated and consists of a 1.4 kb EcoRI-Xbal
frag
ment from the CI constant region of the immunoglobulin
gene. Ca, a
constant
region; R, EcoRI;.X, Xbal.
B
c-MYC
EXON 1
havior) is inherited in an autosomal fashion; that is, it is
neither X- or Y-linked since male-to-male and female-tofemale transmission of the gene occurs. A small number
of transgenic carriers show a different restriction fragment
pattern, suggesting that an internal rearrangement
and
deletion occurs at low frequency (<0.05%). This rearrangement, which is thereafter stably inherited, may involve a recombination
in the immunoglobulin
switch
region.
Myocardial-Specific
Expression
Since these experiments were initiated to assess the oncogenic effect of the activated c-myc gene in a variety of
organ systems (Stewart et al., 1984), it was necessary to
identify the organs in which transgene expression occurred. This was done using RNAase protection analysis,
which revealed that expression of the transgene was
limited to the heart. An example of such an analysis in
which RNA samples were taken from the heart, thymus,
salivary gland, lung, liver, pancreas, kidney, spleen, brain,
and skeletal muscle of a male carrier is shown in Figure
2A. As indicated in the pedigree, the analyzed mouse inherited the transgene from its father.
The RNAase protection analysis was performed using
a probe derived from the normal, murine c-myc cDNA that
detects products of the endogenous
c-myc gene as well
as a unique product corresponding
to the transgene. The
probe (Figure 26) extends from a Notl site in the 3’ portion
of exon 1 to a Pstl site in the 5’ portion of exon 2 and thus,
when annealed to total cellular RNA and digested with
RNAase, should yield a protected fragment 365 bases in
length corresponding
to the normal, processed c-myc
mRNA. It should also yield fragments 225 and 145 bases
in length, corresponding to the unspliced first and second
c-myc
exons present in endogenous
c-myc
precursor
mRNA. As expected, these three fragments were observed in all organs tested, whether the animal was (as
shown in Figure 2A) or was not (data not shown) a transgenie carrier.
In contrast to the normal c-myc gene, the c-myc se-
I
Not 1
cDNA
probe
I
c-MYC
EXON 2
365BP
F
PS, 1
EXON 1 2’25BP
TRANSGENE
175BP
EXON 2 145BP
Figure
2. RNAase
Protection
Analysis
(A) RNAase protection analysis of RNA prepared from various organs
of an RSV-S107 transgenic
mouse illustrating myocardial-specific
expression of the transgene.
The pedigree of the analyzed propositus
is
shown above the figure. Twenty micrograms
of total RNA was incubated with the antisense mRNA probe shown in (B), and RNAase
protection
analysis
was performed
as described
in Experimental
Procedures.
Protected
fragments
of 365 bp corresponding
to
processed
endogenous
c-myc mRNA and 225 and 145 bp fragments
corresponding
to exons 1 and 2 of unprocessed
c-rnF are detected in
all tissues.
A unique 175 bp fragment
that corresponds
to the
processed
transgene is detected in the heart. In addition, the amount
of the 145 bp fragment (corresponding
to exon 2 of c-myc) is increased
in the heart compared
with other organs. (B) Illustration of the antisense mRNA probe used in the RNAase protection
analysis. The
probe (open rectangle) encompassed
a 365 bp Notl-Pstl fragment of
the 3’portion of exon 1 and the 5’portion of exon 2 of the c-myc cDNA.
The protected
fragments
observed
are illustrated
by hatched rectangles.
quence incorporated into the transgene (see Figure 1) carries a truncated first exon. Its transcription is initiated 5’to
this truncated first exon, and its spliced transcript protects
a unique fragment 175 bases in length when annealed to
the cDNA probe shown in Figure 2B. The unspliced
precursor of the transcript is a 145 base fragment of the
probe corresponding to the second exon of c-myc. This latter fragment is indistinguishable
from that protected by
endogenous, unprocessed c-myc mRNA. Under the conditions of the assay, the small (30 bp), protected fragment
corresponding
to the truncated first exon of c-myc is not
detectable. Thus, protection of a 175 base fragment is diagnostic for expression of the transgene. It is observed
only in the RNA derived from the heart (see Figure 2A).
Parental-Specific
721
365
bp-
Gene
-
Expression
12345678
-
va
e.
*.
am
145bp-
Figure 3. RNAase Protection Analysis of RNA Extracted
of Littermates
Resulting from a Mating of Heterozygous
Carriers
from Hearts
Transgenic
Carrier offspring
(heterozygotic
and homozygotic
not distinguished)
are illustrated by half-filled, half-stippled
symbols; noncarriers,
by open
symbols. RNAase protection analysis was performed
as described
in
Figure 2A using the probe illustrated in Figure 28. Although offspring
1 through 7 carry the transgene,
only offspring 3 and 7 display the 175
bp protected fragment indicative of transgene
expression.
The 365,
225, and 145 base long fragments corresponding
to processed
and unprocessed
c-myc mRNA (Figure 28) are present in hearts from both
carrier and noncarrier
animals.
The expression of this fusion gene in tissue of mesodermal origin is consistent with the transgene expression of
other strains of mice inheriting RSV fusion genes (Overbeek, et al., 1986). Thus, the tissue specificity is likely to
be a property of the RSV LTR rather than the site of integration, though this is certainly not proven. With respect
to the physiologic effect of transgene expression, preliminary indications are that the animals do not have an increased incidence of tumor development in any organ including the heart.
Since both the endogenous c-myc gene and the transgene can yield an unspliced transcript that protects a 145
base fragment, the increased amount of this fragment
seen in the heart could theoretically arise from either or
both genes. Note that this product is present in amounts
greater than the sum of the unique 175 base transgenic
fragment and the 145 base fragment present in the other
organs that do not express the transgene (Figure 2A). Assuming that this transcript largely arises from the transgene, it indicates that the relative amount of unspliced
transgenic mRNA precursor is greater than spliced tran-
script. Alternatively, the transgene may give rise to more
than one transcript, for example, one initiating 5’ to the
truncated first exon and a second initiating within the first
intron. On Northern analysis only a single transcript of 2.3
kb is detected (data not shown), suggesting that transcription initiation starts within the immunoglobulin-derived
DNA segment. But transcription at a cryptic promoter in
the first intron of c-myc might be indistinguishable
from
transcript initiation within the immunoglobulin
gene switch
region. Thus, more than one transcription site may be
active in the transgene.
Expression
of the Ttansgene Depends upon
Parental lnherltance
In the initial sample of carrier mice assayed, not all expressed the transgene. This variation in expression was
at first puzzling and suggested that some additional
genetic element(s) governed expression. The initial analysis was further complicated by the fact that the expression
phenotype, requiring myocardial analysis, could not be
tested until the carrier animal had passed the transgene
to several litters to ensure that all possible genotypes were
captured. A typical pedigree showing the genotype and
phenotype of a litter produced by mating male and female
heterozygotes is given in Figure 3. Although 7 of the 8 offspring inherited at least one transgenic locus (as determined by analysis of their DNAs), only 2 of the 7 carriers,
a male and a female, expressed it. In both cases, as with
all other expressing mice, expression was restricted to the
myocardium.
To understand the genetic rules that govern expression,
a larger group of systematic matings was established; Fl,
F2, and further descendant litters were secured and analyzed with respect to genotype and phenotype. Through
the analysis of RNA from the hearts of 77 transgenic carriers, the following pattern of expression emerged (Table 1).
None of the 42 carrier offspring (males and females) of
carrier females mated with noncarrier males expressed
the transgene. In contrast, each of the 20 carrier offspring
(males and females) of noncarrier females mated with carrier males expressed the transgene. Of the 15 carrier offspring from matings between male and female carrier
mice (heterozygote matings), expression was detected in
only 6. In cases in which complete organ analyses were
carried out, expression was always restricted to the myocardium.
Table 1. Influence of Parental Origin of the Transgene
on Its
Subsequent
Expression
in Hearts of Transgenic
Offspring
Status of Transgene
in TG Offspring
Parent
Expression
Female
Male
Expressed
Not Expressed
TG
Non-TG
TG
Non-TG
TG
TG
0
20
6
42
0
9
TG, transgenic
carrier.
Non-TG, nontransgenic
carrier.
Cell
722
Table 2. Influence of Parental Origin
DNA Methylation
Status in Transgenic
Methylation
Pattern
in TG Offspring
Parent
on Its
of Transgene
Female
Male
Undermethylated
Methylated
Combined
TG
Non-TG
TG
Non-TG
TG
TG
0
50
10
59
0
55
0
0
19
TG, transgenic
carrier.
Non-TG. non-transgenic
12Kb-
of the Transgene
Offspring
carrier.
8.OKb-
4.3Kb3.OKb2.4Kb-
1.4KbFigure 4. Southern
rier and Transgenic
Analysisof
DNAfrom
Carrier Matings
Offspring
of Various
Noncar-
Genomic DNA was digested with the methylation-sensitive
restriction
endonuclease
Hpall, and Southern
analysis
was performed
as
detailed in Experimental
Procedures.
The probe used is illustrated in
Figure 1 and consists of a fragment from the a constant region of the
immunoglobulin
gene. The pattern observed in offspring from noncarrier crosses
is shown in lane 1. The result for offspring carriers with
maternally derived transgenes
is shown in lane 2; that for offspring carriers with paternally derived transgenes,
in lane 3. The pattern in presumed homozygotic
carrier offspring is illustrated in lane 4. In animals
inheriting the transgene
from the female parent (lane 2) large DNA
fragments
(12 and 8 kb), as well as a 2.4 kb fragment, are detected.
In animals inheriting the transgene from the male parent (lane 3) the
large fragments are not present and smaller fragments (4.3, 3,2.4, and
1.4 kb) appear. The results indicate that the transgene
inherited from
the male parent is relatively undermethylated
compared with the transgene inherited from a female parent. Animals inheriting a transgenic
allele from each parent display a combined pattern.
Relationship
of Methylation
Pattern to
Transgene Expression
Because expression requires passage of the transgene
through the male parent and is observed in the adult as
well as the fetal myocardium (data not shown), it is not
likely to be the result of prezygotic RNA synthesis (paternal RNA). Furthermore, since information required for expression must be acquired during gametogenesis
in the
male or shortly after fertilization and since the degree of
methylation of cytosine nucleotides appears to correlate
with the transcriptional activity of certain genes (reviewed
by Razin et al., 1984) and is altered during spermatogenesis (Groudine and Conklin, 1985) methylation seemed a
likely mechanism by which to convey this regulatory information. Predictions of this model are easily tested
Genomic DNAs extracted from tail samples of carrier
mice were digested with Hpall, a restriction endonuclease
that cleaves CCGG sequences only when neither cytosine residue is methylated. The resulting transgenic restriction fragment pattern was detected using the standard probe derived from the immunoglobulin
portion of
the transgene (Figure 1). Three patterns, one of which represented the sum of the other two, could be discerned in
mice inheriting the transgene (Figure 4). The first pattern,
referred to as methylated, was observed in all carrier offspring of female carriers mated with noncarrier males. It
consisted of a 12 kb fragment and less intense 8 and 2.4
kb fragments (Figure 4, lane 2). A second pattern, referred
to as undermethylated,
was observed in all carrier offspring of carrier males mated to noncarrier females. The
12 and 8 kb fragments were absent and had been
replaced by three smaller fragments of 4.3,3.0, and 1.4 kb
and a greatly increased signal from the 2.4 kb fragment
(Figure 4, lane 3). A third pattern, referred to as combined,
was observed in some of the carrier offspring of matings
between heterozygote carriers. It consisted of the sum of
fragments seen in the other two patterns (Figure 4, lane
4). Offspring from heterozygote matings that display the
combined pattern are presumably homozygotic for the
transgene, inheriting it from both male and female parents. A summation of the analysis for 193 transgene carriers is shown in Table 2. While the numbers of offspring
from heterozygote crosses are small, fewer offspring with
undermethylated
patterns were observed than would
have been expected. These results raise the possibility of
transmission distortion, suggesting a deleterious effect of
the expressed gene (see below).
When DNAs from the transgene carriers were digested
with the restriction endonuclease
Mspl, an isoschizomer
of Hpall that is insensitive to the methylation status of
CCGG sequences, only one fragment 1.4 kb in length was
observed regardless of which parent donated the transgene (data not shown). These data indicate that, while certain Hpall sites in the transgene are methylated regardless of paternal inheritance, others are methylated only
when inherited through the female parent. Furthermore,
the methylation pattern observed in DNA derived from the
tail was observed in every other somatic organ analyzed,
including the heart (see below). It is evidently a global
methylation pattern, indifferent to whether or not the gene
is actively expressed in a particular organ.
The relationship of the methylation pattern to expres-
Parental-Specific
723
Table 3. Correlation
with Its Myocardial
Gene
Expression
of Methytation
Status of the Transgene
Expression
in Individual Transgenic
Mice
Status
of Transgene
Expression
-12
Kb-
Methylation
Status
of Transgene
Expressed
Not Expressed
-&OKb-
Undermethylated
Methylated
23
0
0
36
-43
Kb-
-30 Kb-2AKb-
sion of the transgene was examined in 61 animals (Table
3). In the 23 transgenic carriers exhibiting the undermethylated pattern, all express the transgene in the heart.
In contrast, none of the 36 transgenic carriers exhibiting
the methylated pattern express the transgene. No exceptions to these patterns were detected. Expression of the
transgene is strictly correlated with its degree of methylation. We have examined nine additional strains of transgenie mice that carry various fusions of the mouse mammary tumor virus LTR and c-myc gene (Stewart et al.,
1964); none exhibit parental related expression or methylation patterns of the c-tnyc gene.
The methylation pattern of the transgene in any particular animal is not dependent upon the sex of that animal,
but rather upon the sex of the animal’s transgenic parent.
To demonstrate this further, the methylation pattern of the
transgene was examined as inherited from the four possible paternal permutations of sex and methylation genotype (Figure 5). The results demonstrate that both female
and male carrier offspring of transgenic males exhibit an
undermethylated
pattern of the transgene irrespective of
the methylation pattern of the male parent (Figures 6A and
58). Likewise, both female and male carrier offspring of
transgenic females exhibit a methylated pattern for the
transgene irrespective of the methylation pattern of the fe
male parent (Figures 5C and 50).
Yethylation
Pattern Imprinted during Gametogenesis
Since the methylation pattern of the transgene is observed in all tissues in the adult animal, it is likely to be
imprinted either during gametogenesis in the parent or at
an early stage of embryogenesis
in the offspring. These
possibilities can easily be distinguished
in the male parent since about 30% of the testicular cells of adult males
are in some stage of gametogenesis (Hecht, 1967). A male
animal that inherits the transgene from its female parent
would be expected to carry a methylated version of the
transgene in every somatic organ. If the maternally inherited pattern were to be altered in the germ line, an undermethylated pattern would begin to emerge in the DNA
derived from testicular tissue. Conversely, a male animal
inheriting the undermethylated
pattern from his father
should maintain that pattern in all his somatic organs and
transmit it unaltered during gametogenesis.
The methylation patterns observed below conform precisely to these
predictions.
In a male inheriting the transgene from the female parent, all the somatic organs exhibit a methylation pattern
(Figure 6). In the testes, however, a fragment of 4.3 kb appears, corresponding
to one of the fragments present in
animals exhibiting the undermethylated
pattern. These
-1.4 Kb-
12 Kb
8.0 Kb
4.3 Kb
3.0 Kb
2.4 Kb
.!I
,. ‘4
1.4 Kb
Figure 5. Southern Analysis Demonstrating
Alteration
ation Status of the Transgene
through Inheritance
of the Methyl-
Male carriers
displaying
either the methylated
(A) or the undermethylated (B) pattern of the transgene
donate an undermethylated
transgene to both their male and female carrier offspring.
In contrast,
female carriers with a methylated (C)or an undermethylated
(D) transgene donate a methylated transgene to both their female and male carrier offspring.
P
34
12 kb-
w
8.0 kb-
4.3 kb-
Figure 6. Southern Analysis of Genomic DNA from Various Organs of
an Adult Male Tmnsgenic
Carrier Mouse That Inherited the More
Highly Methylated
Form of the Transgene from the Female Parent
The methylated form of the transgene
is present in
sessed by the presence of the 12 and 8 kb bands. In
sperm in all stages of development
are present, an
form of the transgene
is detected hy the appearance
all tissues as asthe testes, where
undermethylated
of a 4.3 kb band.
Cell
724
data indicate that the female parent-imprinted
methylation pattern is being eliminated in the male offspring’s
testis during gametogenesis.
The methylation state of the transgene in various organs of female carriers was also examined, although the
small number of germ cells in the ovary makes it difficult
to evaluate the role played by the ovary in imprinting. Females inheriting the transgene from their fathers display
the undermethylated
pattern in all somatic organs as well
as in the ovary (data not shown). Females inheriting the
gene from their mothers exhibit the methylated pattern in
all organs including the ovary (data not shown). Since the
number of mature eggs in the female is small compared
with the amount of tissue in the ovary and since the second meiotic division does not occur until after fertilization,
we feel that the sensitivity of this assay is not adequate to
detect a possible shift from the undermethylated
to the
methylated pattern during gametogenesis
in the female.
The question of germ-line versus zygotic imprinting in the
female therefore awaits resolution,
possibly by hormonally inducing superovulation
and collecting sufficient
oocytes for DNA analysis.
Discussion
A Molecular Explanation
for Imprinting
Phenomena
Observed in the Mouse
The parental-specific
expression of the transgene observed in this strain of mice supports and may explain the
molecular basis for previous data demonstrating
the
nonequivalence
of the maternal and paternal genomes.
Both the nuclear transfer experiments
(McGrath and
Solter, 1984a; McGrath and Solter, 1986; Mann and LovellBadge, 1984; Renard and Babinet, 1986; Surani et al.,
1984; Surani et al., 1986a; Surani et al., 1986b; Surani,
1987) and the classic genetic experiments involving intercrosses of animals with translocations
(Cattanach and
Kirk, 1985; Cattanach, 1986; Searle and Beechey, 1978;
Searle and Beechey, 1985; Solter, 1987; Surani, 1987)
have demonstrated that portions of the genome must be
inherited from each parent in order for development of the
zygote to proceed to term. Moreover, a specific mouse mutant, hairpin tail (Thp) (Johnson, 1974) has been identified
which supports the contention that specific regions of the
mouse genome exhibit parental specificity (McGrath and
Solter, 1984b). In this mutant, a deletion occurs in the
proximal portion of chromosome 17 (Bennett et al., 1975).
If the deletion is inherited from the male parent, the offspring are viable. If the deletion is inherited from the female parent, the fetus does not develop to term (Johnson,
1974). Thus it has been hypothesized that certain genes
located in the proximal portion of chromosome 17 are required for development and presumably these genes are
transcribed only when inherited from the female parent.
Additional evidence that parental imprinting influences
the expression of genes important in development comes
from the analysis of intercrosses of heterozygotes for
Robertsonian translocations of chromosomes 11 and 13.
Offspring that are parentally disomic or nullisomic for
regions on each chromosome are produced (Cattanach
and Kirk, 1985). Offspring maternally disomic for the proximal portions of chromosome 11 (and thus paternally nullisomic for the same region) are small compared with normal littermates. In contrast, offspring disomic for the
paternal contribution of the same region of chromosome
11 are large compared with their littermates. A similar observation has been made in the distal region of chromosome 2. Although animals either paternally or maternally
disomic for the distal portion of chromosome 2 live for only
a few days after birth, two distinct phenotypes have been
observed. Maternally disomic offspring are hypokinetic
and have flat bodies, while paternally disomic offspring
are hyperkinetic and have square bodies. These results
can be interpreted in terms of a parental effect on gene expression. The data presented here directly demonstrate
that parental imprinting of a specific gene completely
correlates with its subsequent expression.
Parental Imprinting
Effects a Transgene Methylation
Pattern Universal in all Tissues, yet Expression
Is Restricted to the Myocardium
The physical mechanism of the demonstrated parental imprinting is as yet unknown. Numerous studies are available to suggest a close correlation of gene expression
with methylation status (see Razin, 1984, for review), but
exceptions do exist, and thus the precise role of methylation is yet to be determined. 5Azacytidine has been used
to block methylation successfully and subsequently alter
gene expression, further suggesting that the expression
of certain genes is dependent, at least in part, upon the
degree of methylation of cytosine nucleotides (Jones,
1984; Jaenisch et al., 1985). Previous studies have
demonstrated that certain genes expressed constitutively
in the embryo are protected from de novo methylation during spermatogenesis,
while genes that are not expressed
undergo methylation during gametogenesis
in the male
(Groudine and Conklin, 1985).
The data in this report clearly demonstrate that the parental origin of a specific gene (i.e., the transgene)
governs the degree of methylation of that gene. The undermethylated form of the transgene is expressed, while
the more methylated form is not. Transgenic alleles inherited from the male parent were undermethylated
in
comparison to those inherited from the female parent. The
correlation between parental origin, methylation status,
and expression (as assessed by mRNA content) of the
transgene is absolute, suggesting that undermethylation
of the transgene plays a role in its subsequent expression.
Other physical properties of the genome, possibly depending on methylation, may be involved in parental
imprinting. These include the presence of DNAase hypersensitivity sites and the presence of specific conformations such as Z-DNA (Conklin and Groudine, 1984). These
properties could be governed in part by the methylation
states of the regulatory sequences of specific genes. On
the other hand, it is evident that methylation status alone
does not determine the expression of the transgene since
the male-imprinted
pattern of transgene methylation oc-
Parental-Specific
725
Gene
Expression
curs throughout the organism
in the heart. Undermethylation
sufficient, for the subsequent
but expression occurs only
may be necessary, but not
expression of the gene.
A Methyiation
Pattern is imprinted
in the Female
and Eliminated
in the Male Germ Line
Parental imprinting of the transgene occurs in the germ
cells of the male. This is documented by the presence of
undermethylated
forms of the transgene in the testes of
males exhibiting a more highly methylated form of the
transgene in all other tissues. Although not proven, it is
anticipated that imprinting of the methylation pattern also
occurs during gametogenesis
in the female. Our studies
suggest that four separate mechanisms must occur during gametogenesis
to yield the results observed in this
strain of transgenic mouse. In the oocyte the transgene either must be maintained in the methylated state or, if not
already methylated, must undergo de novo methylation
before fertilization. In the male germ cells the transgene
must be maintained in the undermethyiated
state or become undermethyiated
either through a passive mechanism (such as inhibition of a maintenance methylase) or
through the action of a specific demethylase. While de
novo and maintenance methyiases have been identified
in mammalian systems (see Razin, 1984, for a review), no
specific demethylase has as yet been identified.
is imprinting
Dependent upon Transgenic Sequences
or the Site of integration?
The question arises as to whether the parental-specific
expression observed in these mice is determined by the
site of integration of the transgene, by the specific sequences in the transgenic construct, or by a combination
of these elements. To determine whether this property is
conveyed with the inserted fragment or is a property of the
chromosomal site of integration, we tested another transgenie strain into which an identical DNA fragment had
been introduced. The results of these experiments, still
preliminary because of sample size, indicate that the degree of methylation of sequences within the transgene
does vary to a certain extent with parental origin. But in
contrast to the original strain of mice, exceptions to the established
pattern of parental imprinting
occur. Such
results suggest that the paternal imprinting effect may be
an easily damaged or “leaky” property of the transgene.
This property may be subject to additional influences
such as other control elements operating at the site of integration or alterations of transgene structure that might
occur as a consequence of the illegitimate recombination.
A Possible Explanation
for the Behavior
of Certain Human Alleles
The results of this study may help to explain the pattern
of inheritance noted for certain human diseases. For example, the genetic abnormality responsible for Huntington’s
disease has been localized to chromosome 4 (Gusella et
al., 1983). This disorder normally displays an autosomal
dominant mode of inheritance with high penetrance, but
an exception does exist. The juvenile form of Huntington’s
disease appears to be inherited preferentially (but not exclusively) from the male parent (Merritt et al., 1989). Diabetic alleles within the HLA complex appear to be preferentially transmitted by the father, a finding that might
account for the increased occurrence of diabetes in their
offspring (Vadheim et al., 1986). Another disease demonstrating parental influence of inheritance is the infantile
form of myotonic muscular dystrophy. Although the locus
for myotonic dystrophy has been localized to a region of
chromosome 19 (Bartlett et al., 1967) the congenital form
of the disease is almost exclusively inherited from the female parent (Roses et al., 1979). While these exceptions
might not be afielic with the major forms of these inherited
disorders, they might also represent human manifestations of parental-specific gene expression. The occurrence
of hydatidiform moles in humans is the clearest example
of the aberrant fate of human extraembryonic
tissue
whose genome is derived entirely from the male parent
(Wake et al., 1984).
Experlmentel
lkansgenic
Procedures
Mice and Vector
Constructlons
The plasmid used for the pronuclear
injection consists of a portion of
the 3’ LTR of the RSV fused to a DNA fragment from the mouse S107
plasmacytoma(Kirsch
etal., 1961). In thisplasmacytoma,
a 12;15chromosomal hanslocation
removed most of the first (noncoding)
c-mt+c
exon and the two normal c-rn~z promoters,
and juxtaposed
(5’ to 5’) the
truncated
c-m
gene to the switch sequence
of the immunoglobulin
heavy chain gene. A 17 kb EcoRl fragment of this rearranged
c-myc
gene that includes the region from the EcoRl site in the immunoglobulin a constant region to the EcoRl site 6 kb 3’ to the c-m
gene was
ligated into a derivative of the pRSVCAT plasmid described by Gorman
et al. (1962). The pRSVCAT plasmid containing
the RSV LTR sequences
was digested with EcoRl to delete a 3’ portion of the RSV
sequence
and all chloramphenicol
acetyttransferase
and SV40 sequences. The 17 kb EcoRl fragment from the S107 plasmacytoma
was
then ligated into the EcoRl site that is 3’ of the RSV sequence
in the
pRSVCAT derivative. The final construct
(termed RSV-SW)
consists
of a 2.2 kb EcoRI-Accl
fragment of pBR322, a 437 bp Pvull-EcoRI
fragment containing
a portion of the 3’ RSV LTR, and the 17 kb fragment containing the fused immunoglobulin
heavy chain gene and the
truncated
c-myc gene.
The RSV-S107 plasmid was linearized at a unique Kpnl site 3’to the
c-w
gene and injected into one of the pronuclei of fertilized one-cell
mouse eggs derived from a C57BII6.J
x CD-1 mating. The injected
eggs were transferred
into pseudopregnant
foster females and allowed
to develop to term. A male founder animal was identified through
Southern
blot analysis
of DNA extracted
from the tail. Over 700
descendants
of this founder were bred, largely to CD-1 mice, and gene
typed.
DNA and RNA Isolation
DNA was isolated from a short segment of the tail or from organs by
a modification
of the method of Davis et al. (1960). The resultant nucleic acid pellet was resuspended
in 300 ul of a 10 mM Tris (pH 7.4)
0.1 mM EDTA solution. RNA was isolated by homogenizing
tissue in
a Polytron homogenizer
using a buffer containing
4 M guanidine
isothiocyanate,
2.5 mM sodium citrate (pH 7) and 0.1 M 6-mercaptoethanol. Total RNA was isolated by the method of Chirgwin et al. (1979)
by centrifugation
through a 5.7 M cesium chloride, 25 mM sodium
acetate cushion.
DNA and RNA Analysis
DNAs were digested with the appropriate
restriction
endonuclease,
electrophoresed
on 0.7% agarose, transferred
to nitrocellulose
or nylon filters by the method of Southern (1975) and hybridized to a 1.4 kb
Cell
726
EcoRI-Xbal
fragment containing the a constant region from the RSVS107 plasmid. The relationship of the probe to the vector used for injection is shown in Figure 1. RNA was analyzed
for expression
of the
transgene by the RNAase protection assay (Melton et al., 1984). The
probe used consisted of a 1.9 kb BamHI-Pstl
fragment of the mouse
c-myc cDNA (containing
portions of the first and second exons of
c-myc) cloned into pGEM-3 (Promega, Madison, WI). The vector was linearized at a unique Nob site in the first exon of c-myc, and 32P-labeled
antisense transcripts
to portions of the first and second exon of c-rnyc
were synthesized
using SP6 polymerase
and [3zP]UTP by the previously described
method (Melton et al., 1984). Total cellular RNA was
hybridized to the radiolabeled antisense c-myc RNA probe for at least
6 hr at 50% in a reaction mixture containing 20 ug of total RNA, 10s
cpm antisense probe, 50% formamide,
and 4 mM PIPES buffer (pH
6.7). Single-stranded
RNA was digested with a solution containing
RNAase A and RNAase Tl, and the RNAases
were inactivated
with
proteinase K and SDS as described by Melton et al. (1984). The samples were extracted
with phenol, precipitated
with ethanol, resuspended in 90% formamide, and electrophoresed
on an 8 M urea, 6%
acrylamide
gel. The gels were dried on Whatman
3MM paper and
exposed to Kodak XAR-5 X-ray film with an image intensifier at -70%
for 24 to 72 hr.
Acknowledgments
We are most grateful to Racheal Wallace for her assistance
in many
of the animal experiments
and to Ann Kuo and Twila Jackson for their
expert technical assistance.
This work was supported
in part by grants
from E. I. DuPont de Nemours, Inc. (to P L.), American Business Foundation for Cancer Research,
Inc. (to P. L.). and a National Institutes of
Health Award HL26831 (to J. L. S.). J. L. S. is an Established Investigator of the American Heart Association.
P L. is a Senior Investigator
of
the Howard Hughes Medical Institute.
The costs of publication of this article were defrayed in part by the
payment
of page charges.
This article must therefore
be hereby
marked “advertisement”
in accordance
with 18 USC. Section 1734
solely to indicate this fact.
Received
July 10, 1987.
Bartlett, R. J., Pericak-Vance,
M. A., Yamaoka, L., Gilbert, J., Herbstreith, M.. Hung, W.-Y., Lee, J. E.. Mohandas,
T., Bruns, G., Laberge,
C., Thibault, M.-C., Ross, D.. and Roses, A. D. (1987). A new probe for
the diagnosis of myotonic muscular
dystrophy.
Science 235, 16481650.
Barton, S. C., Surani, M. A. H., and Norris, M. L. (1984).
nal and maternal
genomes
in mouse development.
374-376.
Role of paterNature 371,
Bennett, D., Dunn, L. C., Spiegelman,
M.. Artzt. K., Cookingham,
and Schermerhorn,
E. (1975). Observations
on a set of radiation
duced dominant
T-like mutations
in the mouse. Genet. Res.
95-l 08.
Cattanach.
B. M. (1986).
Exp. Morphol. 97(suppl.)
Parental
137-150.
origin
effects
in mice.
J.,
in26,
J. Embryol.
Cattanach,
B. M., and Kirk, M. (1985). Differential activity of maternally
and paternally
derived chromosome
regions in mice. Nature 375,
496-498.
Chirgwin, J. M., Przybyla, A. E., MacDonald,
R. J., and Rutter, W. J.
(1979). Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease.
Biochemistry
78, 5294-5299.
Conklin, K. F., and Groudine, M. (1984). Chromatin structure and gene
expression.
In DNA Methylation:
Biochemistry
and Biological Significance, A. Razin, H. Cidar, and A. Riggs, eds. (New York: SpringerVerlag), pp. 293-345.
Davis, R. W., Thomas, M., Cameron, J., St. John, T l?, Scherer, S.. and
Padgett, R. A. (1980). Rapid DNA isolations for enzymatic
and hybridization analysis. Meth. Enzymol. 65, 404-411.
Gorman,
Howard,
C. M., Merlino. G. T.. Willingham,
M. C., Pastan, I., and
B. H. (1982). The Rous sarcoma virus long terminal repeat is
a strong promoter when introduced into a variety of eukaryotic
cells by
DNA-mediated
transfection.
Proc. Natl. Acad. Sci. USA 79,6777-6781.
Groudine. M., and Conklin, K. F. (1985). Chromatin
structure
novo methylation
of sperm DNA: implications
for activation
paternal genome. Science 228, 1061-1068.
and de
of the
Gusella, J. F., Wexler, N. S., Conneally, l? M., Naylor, S. L., Anderson,
M. A., Tanzi. R. E., Watkins, P C., Ottina, K., Wallace, M. R., Sakaguchi, A. Y., Young, A. B., Shoulson, I., Bonilla, E., and Martin, J. B.
(1983). A polymorphic
DNA marker genetically
linked to Huntington’s
disease. Nature 306. 234-238.
Hecht, N. B. (1987). Regulation of gene expression
during mammalian
spermatogenesis.
In Experimental
Approaches
to Mammalian
Embryonic Development,
J. Rossant,
and R. A. Pedersen, eds. (Cambridge: Cambridge
University
Press), pp. 151-195.
Jaenisch.
R., Schnieke, A., and Harbers, K. (1985). Treatment of mice
with 5-azacytidine
efficiently
activates
silent retroviral genomes
rn
different tissues. Proc. Natl. Acad. Sci. USA 82, 1451-1455.
Johnson, D. R. (1974). Hairpin tail: a case of post-reductional
tion in the mouse egg? Genetics
76, 795-805.
gene ac-
Jones, P. A. (1984). Gene activation by 5-azacytidine.
In DNA Methylation: Biochemistry
and Biological Significance.
A. Razin, H. Cidar,
and A. Riggs, eds. (New York: Springer-Verlag),
pp. 165-219.
Kirsch, I. R.. Ravetch, J. V., Kwan, S. P.. Max, E. E., Ney, R. L., and
Leder, f? (1981). Multiple immunoglobulin
switch region homologies
outside the heavy chain constant region locus. Nature 293, 585-587
McGrath, J., and Salter, D. (1984a). Completion of mouse embryogenesis requires both the maternal and paternal genomes. Cell 37, 179-183.
McGrath, J., and Solter, D. (1984b). Maternal Thp lethality in the mouse
is a nuclear, not cytoplasmic,
defect. Nature 308, 550-551.
McGrath,
J.. and Salter, D. (1986). Nucleocytoplasmic
interactions
the mouse embryo. J. Embryol. Exp. Morphol. 97(suppl.) 277-289.
in
Mann, J. R., and Lovell-Badge,
R. H. (1984). Inviability of parthenogenones is determined
by pronuclei,
not egg cytoplasm.
Nature 370.
66-67.
Melton, D. A.,
Green, M. R.
RNA and RNA
riophage SP6
Krieg, P. A., Rebagliati, M. R., Maniatis, T., Zinn, K., and
(1984). Efficient in vitro synthesis
of biologically
active
hybridization
probes from plasmids containing a bactepromoter. Nucl. Acids Res. 72. 7035-7056.
Merritt, A. D., Conneally, F! M., Rahman, N. F., and Drew, A. L. (1969).
Juvenile Huntington’s chorea. In Progress in Nemo-Genetics,
A. Barbeau and T. R. Brunette, eds. (Amsterdam:
Excerpta Medica Foundation), pp. 645-650.
Overbeek,
I? A., Lai, S.-P, Van Quill, K. R., and Westphal,
H. (1986).
Tissue-specific
expression
in transgenic mice of a fused gene containing RSV terminal sequences.
Science 237, 1574-1577.
Razin, A. (1984). DNA methylation
patterns: formation and biological
functions.
In DNA Methylation:
Biochemistry
and Biological
Significance. A. Razin, H. Cidar, and A. Riggs, eds. (New York: SpringerVerlag), pp. 127-146.
Razin, A., Cidar, H., and Riggs, A. D. (1984). DNA Methylation:
Biochemistry
and Biological Significance
(New York: Springer-Verlag).
Renard, J. P. and Babinet, C. (1986). Identification
velopmental effect on the cytoplasm of one-cell-stage
Proc. Natl. Acad. Sci. USA 83, 6883-6886.
of a paternal demouse embryos.
Roses, A. D., Harper, P S., and Bossen, E. H. (1979). Myotonic muscular dystrophy.
In Handbook
of Clinical Neurology,
P J. Vinken and
G. W. Bruyn, eds. (New York: North-Holland),
pp. 485-531.
Searle, A. G., and Beechey, C. V. (1978). Complementation
studies
mouse translocations.
Cytogenet.
Cell Genet. 20, 282-303.
with
Searle, A. G., and Beechey,
C. V. (1985). Noncomplementation
phenomena
and their bearing on nondisjunctional
effects. In Aneuploidy: Etiology and Mechanisms,
V. L. Dellarco, P E. Voytek, and
A. Hollaender, eds. (New York: Plenum Press), pp. 363-376.
Salter, D. (1987). Inertia of the embryonic
genome in mammals, Trends
Genet. 3, 23-27.
Southern,
fragments
Stewart,
E. M. (1975). Detection of specific
separated by gel electrophoresis.
T. A., Pattengale,
P. K.. and Leder,
sequences
among DNA
J. Mol. Biol. 98, 503-517.
P (1984).
Spontaneous
Parental-Specific
727
Gene
Expression
mammary
adenocarcinomas
in transgenic
mice
press MTVlmyc fusion genes. Cell 38, 627-637.
that carry
and ex-
Surani, M. A. H. (1967). Evidence and consequencesof
differences
between maternal and paternal genomes during embryogenesis
in the
mouse. In Experimental
Approaches
to Mammalian Embryonic
Development, J. Rossant, and R. A. Pedersen, eds. (Cambridge:
Cambridge
University
Press), pp. 401-436.
Surani, M. A. H., Barton, S. C., and Norris, M. L. (1964). Development
of reconstituted
mouse eggs suggesting
imprinting of the genome during gametogenesis.
Nature 308, 546-550.
Surani, M. A. H.. Barton, S. C., and Norris, M. L. (1966a). Nuclear transplantation in the mouse: heritable differences
between parental genomes after activation of the embryonic
genome. Cell 45, 127-136.
Surani, M. A. H.. Reik, W., Norris,
fluence of germline modifications
mouse development.
J. Embryol.
M. L., and Barton, S. C. (196613). Inof homologous
chromosomes
on
Exp. Morphol. 97(suppl.) 123-136.
Vadheim, C. M., Rotter, J. I., MacLaren, N. K., Riley, W. J., and Anderson, C. E. (1986). Preferential transmission
of diabeticalleles
within the
HLA gene complex. N. Eng. J. Med. 375, 1314-1316
Wake, N., Seki, T., Fujita, H.. Okubo, H., Sakai, K., Okuyama,
K..
Hayashi, H., Shiina, Y., Sato, H., Kuroda, M., and Ichinoe. K. (1964).
Malignant potential of homozygous
and heterozygous
complete moles.
Cancer Res. 44. 1226-1230.
Note
Added
in Proof
Two very recent reports (Reik et al. 11967). Nature 328, 246-251;
Sapienza
et al. [1967). Nature 328, 251-254)
describe
transgenic
mouse lines in which, respectively,
1 out of 7 and 4 out of 5 transgenic
loci tend to display partially imprinted
methytation
patterns.
In the
former case, the transgene, an immunoglobulin/chloramphenicol
transacetylase
sequence,
is not expressed
regardless
of the transmitting
parent. In the latter case, quail troponin I transgenes
were evidently not
tested for expression.
In view of this frequency of legacy-related
methylation, it seems likely that the expressionlmethylation
we observe depends on our specific insert and on the locus into which it is inserted.
The locus is likely to imprint the methylation
pattern, but the inserted
fragment of DNA must be capable of responding
to it.