719
Development 103, 719-724 (1988)
Printed in Great Britain © The Company of Biologists Limited 1988
DNA methylation in the developing marsupial embryo
MARY E. STEVENS1, PETA M. MAIDENS2, EDWARD S. ROBINSON34, JOHN L. VANDEBERG3,
ROGER A. PEDERSEN5 and MARILYN MONK2
1
Gladstone Foundation Laboratories for Cardiovascular Disease, PO Box 40608, San Francisco, CA 94140, USA
MRC Mammalian Development Unit, University College London, Wolfson House, 4 Stephenson Way, London NW1 2HE, UK
^Department of Genetics, Southwest Foundation for Biomedical Research, San Antonio, TX 78285, USA
^School of Biological Sciences, Macquarie University, Sydney 2109 NSW, Australia
^Department of Anatomy and Laboratory of Radiobiology and Environmental Health, University of California, San Francisco, CA
94143, USA
2
Summary
Marsupial development differs from early development of placental mammals in that the blastocyst is
unilaminar, so that both embryonic and extraembryonic cells are derived from a single layer of cells
(protoderm) which faces the blastocyst cavity. Also, all
cells in female marsupial conceptuses so far examined
show preferential paternal X-inactivation. To test for
a possible correlation between cell position, paternal
X-inactivation and DNA hypomethylation, marsupial
DNA preparations from three regions, embryo, vascular yolk sac and avascular yolk sac, were digested with
methyl-specific restriction endonucleases, separated
on agarose gels and end-labelled with 32P-dCTP. The
size distribution of the fragments obtained indicated
three levels of methylation: high methylation of embryonic DNA, intermediate levels of methylation of
vascular yolk sac DNA and hypomethylation of avascular yolk sac DNA. The degree of methylation of
repeat sequences, observed as discrete bands in endlabelled HpaU digests, was correlated with the overall
methylation of tissue DNA. Thus, the difference in
methylation in embryonic and extraembryonic DNA
was similar to that described for the mouse conceptus,
and the outside cell position of marsupial fetal precursor cells did not correlate with hypomethylation.
Hpall tiny fragments, which indicate the presence of
CpG-rich islands of DNA, were evident in the marsupial digests. In the mouse DNA, these islands are
associated with gene transcription and provide one
route to cloning of unique gene sequences.
Introduction
undermethylated in all derivatives of two extraembryonic cell lineages in early mouse embryos,
whereas cells of the embryonic lineage are highly
methylated (Razin et al. 19846; Chapman et al. 1984;
see Rossant, 1986 for review). This tissue-specific
methylation pattern correlates with the type of Xinactivation in these lineages in female conceptuses
(Takagi & Sasaki, 1975; West et al. 1977; Harper etal.
1982): derivatives of the extraembryonic trophectoderm and primitive endoderm undergo paternal Xinactivation, whereas random X-inactivation occurs
in the pluripotent primitive ectoderm.
A possible correlate to paternal X-inactivation and
hypomethylation of DNA is that cells of the two
extraembryonic mouse lineages occupy an 'outside'
position in early development and are in contact with
Recent evidence suggests that methylation of cytosine residues in vertebrate DNA has a regulatory
role; methylation at specific sites in the vicinity of
certain genes suppresses transcription of those genes
(reviewed by Doerfler, 1983; Razin et al. 1984a).
Although information concerning the establishment
or alteration of methylation patterns is incomplete,
variation in overall DNA methylation clearly occurs
during normal development at different times and in
different tissues. Monk et al. (1987) describe an initial
loss of genomic methylation during preimplantation
development, followed by de novo methylation occurring to different degrees in the embryonic, extraembryonic and germ cell lineages. Other studies show
that repetitive DNA sequences are substantially
Key words: methylation, marsupial embryos, mammalian
embryogenesis, blastocyst, DNA.
720
M. E. Stevens and others
Table 1. Summary of end-labelling experiments
No. of
isolates
No. of
end-label
analyses
1
2
4
3
4
4
5
10
5
9
DNA
Stage
15
21
25
26
27
the blastocyst cavity. This suggests a possible role for
cell position in determining methylation patterns. To
test this hypothesis, marsupial embryos (Monodelphis domestica, the gray short-tailed opossum) were
used as a model system. Marsupial development
differs from that of the mouse in that all cells of the
embryo and extraembryonic membranes undergo
preferential paternal X-inactivation (Johnston &
Robinson, 1987; VandeBerg et al. 1987). Furthermore, in marsupials, no inner cell mass is ever formed
at the blastocyst stage as it is in the embryo of
placental mammals (see Wimsatt, 1975; Selwood,
1986, for review). Instead, all embryonic and extraembryonic cell lineages are derived from early blastocyst cells arranged as a single layer called the protoderm. These protoderm cells can be regarded as
'outside' in position and are in contact with the
blastocyst cavity. This study was designed to determine whether marsupial embryonic DNA also shows
hypomethylation, or whether the embryonic DNA is
highly methylated, as is DNA of mouse somatic
tissues.
Materials and methods
Marsupial embryonic tissue and DNA preparation
All embryos were recovered in Dulbecco's phosphatebuffered saline and staged following the numbering sequence of McGrady (1938) for Didelphis virginiana. Ten
stage-15 embryos were dissected into embryonic and extraembryonic regions, and the samples were pooled. Similarly,
these same regions were obtained from two sets of five
pooled stage-21 embryos. Older embryos (stages 25-27)
were dissected into three regions - embryonic, vascular
yolk sac and avascular yolk sac - and analysed individually.
All samples were promptly frozen and stored at —80°C
until use.
Cells were lysed in 200 fi\ STE (100 mM-NaCl, 50 mM-Tris,
pH7-5, lOmM-EDTA) containing 1% sodium dodecyl
sulphate and 200 ^g of proteinase KmP 1 and incubated
overnight at 51 °C. DNA was isolated using standard
phenol/chloroform extraction followed by ethanol precipitation. The DNA pellet was redissolved in TE (100 mMTris, lOmM-EDTA) and treated with 25 ^g of ribonucleaseml" 1 .
End-labelling
Table 1 summarizes the stages and the number of repeti-
tions of end-label experiments for each sample. DNA
preparations were digested with Mspl and Hpall and endlabelled with a(32P)-dCTP using the Klenow fragment of
DNA polymerase I (Maniatis et al. 1982). A HindlU digest
of 5 ng of A DNA was end-labelled to serve as markers.
Following electrophoresis (16 h, 20 V) on a 0-7% agarose
gel, Southern blot (Southern, 1975) of the labelled fragments and exposure of the filters to X-ray film, the
fragment-size distribution was observed. Efficient transfer
of high molecular weight DNA from the gels to the filter
was verified by absence of counts remaining in the gel.
Retention of low molecular weight fragments {HpaU tiny
fragments) by the filters was variable, but this did not affect
the interpretation of the results, which was based on the
distribution of fragment sizes towards the top of the gel.
Uncut DNA samples were always analysed (data not
shown) to ensure the high molecular weight quality of the
DNA preparations. Densitometry was performed with a
Sigma FTR20 scanning densitometer. Degree of methylation was correlated with the distribution of labelled
fragments at the top of the lanes, as well as the relative
number of fragments in each gel region. Repeated analysis
of replicate isolates of DNA from a particular tissue at a
particular stage of development (Table 1) showed reproducible results, thus confirming differences between tissues
and stages.
Results
Differences in global DNA methylation between
embryonic and extraembryonic regions
Marsupial development differs from that of eutherians in that no inner cell mass (ICM) is formed
during the blastocyst stage (see Fig. 1A) Thus, the
embryo proper in marsupials does not originate from
the inner cell mass, but from cells of the protoderm.
In this study, marsupial conceptuses from several
different stages were separated into three regions embryo, vascular yolk sac and avascular yolk sac and the methylation patterns of the total genomic
DNA from each region were compared. The embryonic stages used for this study are illustrated in
Fig. IB.
Monk et al. (1987) used a highly sensitive endlabelling technique for determining DNA methylation levels in extremely small embryonic samples.
DNA preparations were digested with two isoschizomeric restriction endonucleases: Mspl, which will
cleave at the DNA sequence CCGG regardless of the
methylation state of the internal cytosine; and Hpall,
which cuts at the same CCGG sequence, but only if
no methylation is present. The resulting fragments
were end-labelled (see Materials and methods),
loaded directly onto an agarose gel, blotted and
autoradiographed. Comparison of the Mspl and
Hpall fragment sizes gives an indication of the extent
of methylation of the total genomic DNA.
Methylation in marsupial embryos
721
Inner cell
mass
• Blastocoel •
Mouse
blastocyst
^
Marsupial
blastocyst
• Embryo
Embryonic
region
- Extraembryonic
region
Stage 25
Stage 21
Stage 15
emb
vas
Avascular
yolk
sac
Fig. 1. (A) Diagrammatic
comparison of a preimplantation
eutherian embryo, the 3-5 days
gestation mouse blastocyst (left)
and marsupial embryo (right). The
marsupial blastocyst lacks an inner
cell mass so that all cells are in
contact with the blastocyst cavity.
(B) Marsupial conceptuses from
three developmental stages showing
the embryonic, vascular yolk sac
and avascular yolk sac regions.
Bar, lmm.
avas
Fig. 2. Hpall digests of DNA from the embryonic (emb),
vascular yolk sac (vas) and avascular yolk sac (avas) of
stage-27 marsupial embryos. Note the skewing of the
label toward the higher molecular weight DNA fragments
(top of lane) in the embryonic sample.
An indication of the degree of methylation of the
total DNA of a particular tissue may be observed by a
comparison of Mspl and Hpall digests at the tops of
the lanes. Methylated DNA is indicated by a distribution of the label skewed towards the higher molecular weight in the HpaU lanes. A typical exper-
Fig. 3. Densitometry tracing of Hpall digest DNA from
embryonic (A), vascular yolk sac (B) and avascular yolk
sac (C) regions of a stage-27 marsupial conceptus.
iment is shown in Fig. 2. The level of methylation in
the various regions was as follows: embryonic (emb)
tissue > vascular yolk sac (vas) > avascular yolk sac
(avas). A high degree of skewing of the label toward
the high molecular weight methylated fragments was
722
M. E. Stevens and others
A Msp Hpa
seen in embryonic tissue. Vascular yolk sac showed
less skewing, whereas the avascular yolk sac DNA
produced patterns similar to the Msp lane (not
shown), indicating hypomethylation of the DNA.
This pattern of methylation (emb > vas > avas) was
always observed in stage-25, -26 and -27 conceptuses
and is confirmed by a comparison of densitometry
tracings of the lanes (Fig. 3). Stage-21 embryos
showed a similar trend, but stage-15 embryos showed
a somewhat different pattern (Fig. 4). At developmental stage 15, global methylation levels were quite
similar between embryonic and extraembryonic
DNA. Due to the paucity of embryonic tissue,
however, it was not possible to examine earlier
stages. No stages between stages 15 and 21 were
examined.
Repetitive bands were observed in the DNA of all
samples and at all stages. A prominent l-4kb band
was seen in both the Mspl and the Hpall tracks
(arrow, Fig. 5), whereas a ladder of higher molecular
weight bands was more prominent in the. Hpall lanes.
Restriction Msp
enzyme
Hpa
Hpa
emb emb
Fig. 5. Mspll and Hpall digests of stage-26 embryonic
DNA. The large arrow denotes the prominent 1-4 kb
band seen in all samples. Small arrows indicate the ladder
of higher molecular weight bands which are more
prominent in the Hpall lanes.
The identity of these DNAs is unknown. In the
mouse, repetitive sequences may occur in tandemly
repeated fashion in centromeric DNA or be distributed in dispersed fashion (e.g. MLF, mouse interspersed family) (Bennet et al. 1984).
Fragment-size distribution in all samples showed
that a large number of fragment ends are represented
in the small molecular weight region of the gel. This
distribution is similar to that observed in the mouse,
where a high proportion of small fragments are
derived from the clustered CpG sequences that occur
upstream from many genes (Bird et al. 1985).
Discussion
Stage 1 5
region emb
emb
15
vas
avas
Fig. 4. End-labelling of DNA from stage-15 conceptuses.
Embryonic DNA was digested with Mspl or Hpall; only
the HpaU digest of the vascular and avascular yolk sac is
shown.
We examined global methylation patterns in DNA
from several tissues of the marsupial conceptus at
different stages of development in order to test the
correlation between paternal X-inactivation, cell position and hypomethylation. The earliest stage of
marsupial embryo development that we could examine was stage 15 (Fig. IB), when the conceptus exists
as a late bilaminar blastocyst (McCrady, 1938). The
Methylation in marsupial embryos
primitive streak has not yet appeared. Due to the
small size of both stage-15 and stage-21 embryos, we
could not clearly separate the vascular and avascular
yolk sac regions, but could compare embryonic versus
combined vascular and avascular tissues. It was
necessary to pool several embryos (see Materials and
methods) in order to obtain sufficient tissues for
analysis. Embryonic and extraembryonic DNA methylation patterns were more similar to each other at
stage 15 than later in development. This pattern
changed by stage 21, by which time the mesodermal
tissue was separated, and the endoderm, ectoderm
and vascular yolk sac regions of the extraembryonic
tissue were beginning to form adjacent to the embryo. The methylation levels of embryonic and extraembryonic tissues of stage-21 and later-stage conceptuses were strikingly different. In all cases, global
methylation of DNA from embryonic regions was
much higher than that from extraembryonic DNA.
Thus, despite the differences in embryonic cell
position and paternal X-inactivation between mouse
and marsupial development, global DNA methylation differences between embryonic and extraembryonic regions are comparable.
There are other similarities between marsupial and
mouse DNA. For example, repetitive sequences
occur in DNA from all samples from each developmental stage (Fig. 5). The nature of these sequences
in marsupials, other than their molecular size, is
unknown. They resemble the dispersed repetitive
sequences, the mouse interspersed family (MIF),
seen in mouse embryos (Sanford et al. 1984).
End-labelling of marsupial DNA also produced
heavy labelling at the bottom of the tracks, which was
indicative of the high proportion of fragments derived
from short fragments similar to those derived from
islands of clustered CpG sequences. These sequences
occur upstream of many active genes (Bird et al. 1985)
and are unmethylated both in marsupial DNA and
mouse DNA.
The pattern of DNA methylation of the marsupial
genome during development thus is similar to that of
the most extensively studied placental mammal, the
mouse. The general order of global methylation was
the same for both species, with embryonic DNA
being more highly methylated than extraembryonic
DNA, despite the occurrence of paternal X-inactivation and the outside cell position of marsupial
embryonic cells. Both the marsupial and the mouse
have unmethylated repetitive fragments as well as
similar Hpa tiny fragments in end-labelled gels. These
developmental aspects of methylation may thus have
evolved before the divergence of metatherians and
eutherians, or evolved independently.
This work was supported by NATO Grant no. RG86/
723
0168 and by funds provided by the International Cancer
Research Data Bank Programme of the National Cancer
Institute and the International Union Against Cancer. The
marsupial collections were funded by NIH Grant RR01602.
We thank Mary Jo Aivaliotis and Janice MacRossin for
their patience and expertise in identifying timed pregnant
females and Debbie Coller for manuscript preparation.
References
BENNET, K. L., HILL, R. E., PIETRAS, D. F.,
WOODWORTH-SUTAI, M., KANE-HAAS, C , HOUSTON, J.
M., HEATH, J. K. & HASTIE, N. D. (1984). Most highly
repeated dispersed DNA families in the mouse
genome. Mol. cell. Biol. 4, 1561-1571.
BIRD, A., TAGGART, M., FROMMER, M., MILLER, O. J. &
MACLEOD, D. (1985). A fraction of the mouse genome
that is derived from islands of non-methylated CpGrich DNA. Cell 40, 91-99.
CHAPMAN, V., FORRESTER, L., SANFORD, J., HASTIE, N. &
ROSSANT, J. (1984). Cell lineage specific
undermethylation of mouse repetitive DNA. Nature,
Lond. 307, 284-286.
DOERFLER, W. (1983). DNA methylation and gene
activity. A. Rev. Biochem. 52, 93-124.
HARPER, M. I., FOSTEN, M. & MONK, M. (1982).
Preferential paternal X inactivation in extraembryonic
tissues of early mouse embryos. J. Embryol. exp.
Morph. 67, 127-135.
JOHNSTON, P. G. & ROBINSON, E. S. (1987). X
chromosome inactivation in female embryos of a
marsupial mouse (Antechinus stuartii). Chromosoma
95, 419-423.
MANIATIS, T., FRTTSCH, E. F. & SAMBROOK, J. (1982).
Molecular Cloning - A Laboratory Manual. Cold
Spring Harbor, New York: Cold Spring Harbor
Laboratory.
MCCRADY, E. (1938). The Embryology of the Opossum.
Am. Anat. Memoirs No. 16, Philadelphia: Wistar
Institute of Anatomy and Biology.
MONK, M., BOUBELIK, M. & LEHNERT, S. (1987).
Temporal and regional changes in DNA methylation in
the embryonic, extra-embryonic and germ cell lineage
during mouse embryo development. Development 99,
371-382.
RAZIN, A., CEDAR, H. & RIGGS, A. (1984a). DNA
Methylation. Biochemistry and Biological Significance.
New York: Springer-Verlag.
RAZIN, A., WEBB, C , SZYF, M., YISRAELI, J.,
ROSENTHAL, A., NAVEY-MANY, T., SCIAKY-GALLILI, N.
& CEDAR, H. (1984ft). Variations in DNA methylation
during mouse cell differentiation in vivo and in vitro.
Proc. natn. Acad. Sci. U.S.A. 81, 2275-2279.
ROSSANT, J. (1986). Development of extraembryonic cell
lineages in the mouse embryo. In Experimental
Approaches to Mammalian Embryonic Development,
pp. 97-120. Cambridge: Cambridge University Press.
SANFORD, J. P., FORRESTER, L., CHAPMAN, V., CHANDLEY,
A. & HASTIE, N. (1984). Methylation patterns of
724
M. E. Stevens and others
repetitive DNA sequences in germ cells of Mus
musculus. Nucleic Acids Res. 12, 2822-2836.
SELWOOD, L. (1986). Cleavage in vitro following
destruction of some blastomeres in the marsupial
Antechinus stuartii (Macleay). J. Embryol. exp. Morph.
92, 71-84.
SOUTHERN, E. M. (1975). Detection of specific sequences
among DNA fragments separated by gel
electrophoresis. J. molec. Biol. 98, 503-517.
TAKAGI, N. & SASAKI, M. (1975). Preferential inactivity of
the paternally derived X chromosome in the
extraembryonic membranes of the mouse. Nature,
Lond. 256, 640-642.
VANDEBERG, J. L., ROBINSON, E. S., SAMOLLOW, P. B. &
ROBINSON, P. G. (1987). X-linked gene expression and
X-chromosome inactivation: Marsupials, mouse, and
man compared. In lsozymes: Current Topics in
Biological and Medical Research 15, 225-253.
WEST, J. D., FRELS, W. I., CHAPMAN, V. M. &
PAPAIOANNOU, V. E. (1977). Preferential expression
of
the maternally derived X chromosome in the mouse
yolk sac. Cell 12, 873-882.
WIMSATT, W. A. (1975). Some comparative aspects of
implantation. Biol. Reprod. 12, 1—40.
{Accepted 11 May 1988)