/ . Embryol. exp. Morph. Vol. 67, pp. 127-135, 1982
Printed in Great Britain © Company of Biologists Limited 1982
127
Preferential paternal X inactivation in extraembryonic tissues of early mouse embryos
By MARY 1. HARPER,1 MANDY FOSTEN2
AND MARILYN MONK2
From the MRC Mammalian Development Unit,
Wolfson House, University College London
SUMMARY
The preferential expression of the maternal X chromosome seen in certain extraembryonic
membranes of the mouse was studied by investigating the tissues from which these membranes
are derived during early development. The electrophoretic variant of the X-coded enzyme
PGK-1 (phosphoglycerate kinase) was used to distinguish the expression of the maternal
from the paternal X chromosome in heterozygous females.
Both the extraembryonic ectoderm and primary endoderm of 6^-day female egg cylinders
gave almost exclusive expression of the maternal form of the enzyme whereas the epiblast
gave near equal expression of the two parental alleles. No paternal PGK-1 band could be
detected in samples of pooled 3^-day blastocysts, but after 3 or 4 days of culture in vitro a
faint paternal band was seen in the resultant outgrowths. The activity of the maternal band
in these latter samples had increased greatly from that of the blastocysts, consistent with
preferential expression of the maternal Pgk-\ allele in the trophoblastic cells of the outgrowths, while both alleles are expressed in inner-cell-mass cells. The results strongly support
the idea that non-random X-chromosome expression is due to preferential paternal X inactivation in trophectoderm (from which extraembryonic ectoderm is derived) and in primary
endoderm, and not to cell selection.
INTRODUCTION
In female eutherian mammals one of the two X chromosomes is inactivated
early in development to produce adult somatic tissues that are mosaic with
respect to X-linked alleles (Lyon, 1961: see reviews by Lyon, 1972; Gartler &
Andina, 1976; Monk, 1978). Although X inactivation appears to be random in
mouse foetal tissues, certain extraembryonic membranes display preferential
expression of the maternally derived X chromosome. The two X chromosomes
may be marked either cytologically in female embryos heterozygous for an
X-chromosome translocation, or biochemically in females heterozygous for the
two alleles of PGK-1 (phosphoglycerate kinase-1, E.C.2.7.2.3).
Takagi & Sasaki (1975) used Cattanach's translocation to look at the
1
Author's present address: Cold Spring Harbor Laboratory, P.O. Box 100, Cold Spring
Harbor, New York 11724, U.S.A.
2
Authors' address: MRC Mammalian Development Unit, Wolfson House, 4 Stephenson
Way, London NW1 2HE, U.K.
5-2
128
M. I. HARPER, M. FOSTEN AND M. MONK
distribution of the differentially staining, inactive X chromosome in postimplantation embryos. At 8£ days gestation 50 % of the late-replicating X chromosomes were paternal in the embryonic portion of the conceptus (consistent with
random inactivation) whereas over 90 % of the inactive X chromosomes were
paternal in the yolk sac and chorion, and 70 % in the allantois.
Studies of the X-linked electrophoretic variant of PGK-1 (Nielsen & Chapman, 1977) have supported the cytological work of Takagi's group. West, Frels,
Chapman & Papaioannou (1977) found that the maternal Pgk-1 allele is
preferentially expressed in 13^-day female mouse yolk sacs. It appears that the
paternal allele is not expressed at all in the yolk-sac endoderm at this age thus
suggesting inactivation of the paternal X chromosome. Using control crosses
and embryo transfers these workers convincingly showed that the uneven expression of PGK-1 is not due to a selective pressure exerted by the phenotype
of the female genital tract. From work on other tissues (West, Papaioannou,
Frels & Chapman, 1978) it appears that all.derivatives of trophectoderm and
primary endoderm show preferential maternal X-chromosome expression.
Either cell selection or non-random inactivation could explain the uneven
expression of the two X chromosomes. If cell selection is the cause then it must
occur before 9\ days as Frels & Chapman (1980) have found preferential
maternal expression of PGK-1 in the yolk sac and mural trophectoderm at this
age with no evidence of the paternal form in the latter tissue. We therefore
decided to look at an earlier stage.
At 3£ days blastocysts contain low levels of PGK-1 that appears largely
maternal in origin, whereas by 6£ days the enzyme is embryonic and has risen
over 100 fold (Kozak & Quinn, 1975). We employed cellogel electrophoresis to
study the expression of PGK-1, which is a monomer, in dissected tissues from
heterozygous 6^-day females and in blastocysts and blastocyst outgrowths
cultured in vitro. The results suggest that the paternal Pgk-1 allele may never be
expressed in trophectoderm and primary endoderm and that the non-random
expression of the X chromosome seen in extraembryonic tissues is due to
preferential inactivation of the paternal X chromosome and not to cell selection.
MATERIALS AND METHODS
Mice carrying the Pgk-1 variant allele, Pgk-1&, were kindly provided by John
West, these mice being derived from feral Mus musculus musculus from Denmark
(Nielsen & Chapman, 1977). With the help of Dr Anne McLaren the Pgk-1*
allele was established on a C3H inbred background by repeated backcrossing.
The stock now consists of females homozygous and males hemizygous for the
Pgk-la allele, and is colony maintained. To produce heterozygous (Pgk-1&/
Pgk-P) female embryos, reciprocal matings were set up with random-bred
MF1 mice (Olac 1976 Ltd) as the source of the Pgk-lh allele. Fertilization is
Non-random X inactivation in early mouse embryos
129
assumed to occur at midnight and the day of detection of the vaginal plug is
designated day 0.
Preimplantation embryos were collected from females that had been superovulated by intraperitoneal injection of 5 i.u. of PMS (pregnant mare serum)
followed 48 h later by 5 i.u. of HCG (human chorionic gonadotrophin) and
immediately caged with males. 3^-day blastocysts were flushed from uteri of
pregnant females into PB1 medium (Whittingham & Wales, 1969). If to be
collected directly the embryos were passed through two washes of PBl/PVP
(PB1 medium containing 4 mg/ml of polyvinylpyrrolidone, Sigma, in place of
albumin) in watch glasses. Embryos were pooled and placed in 1 or 2/d of
medium in 10/4 Drummond microcaps. The ends of the microcaps were sealed
in a small flame, and the samples were stored at - 70 °C. The numbers in each
group are given in the results. Some blastocysts were cultured in vitro, as
described in Monk & Ansell (1976), for up to 4 days to produce trophoblastic
outgrowths and were harvested as above. 6|-day postimplantation embryos
were collected from normally mated females. The egg cylinders were dissected
individually in PBl/PVP as described by Monk & Harper (1979). The separated
epiblast, extraembryonic ectoderm and primary endoderm (Fig. \d) were
washed and collected individually (each in 2 JLL\ medium) as above.
Supernatant extracts were prepared by freeze thawing the sample three times,
followed by centrifugation at 2000 rev./min at 4 °C. The samples were then
analyzed for PGK-1 expression by running them on cellogel for \\ h at 200 V.
The electrophoresis and positive staining of PGK-1 were done according to
Biicheref al. (1981). Thymocyte extracts from hemizygousPGK-lAandPGK-lB
males were used as controls for PGK-1 A and PGK-1B migration respectively.
RESULTS
PGK expression in 6\-day embryonic tissues
To determine when the preferential expression of the maternal X chromosome
can first be detected, tissues of 6-^-day embryos fiom crosses $ Pgk-l^/Pgk-l01 X
<$Pgk-lb and qPgk-P/Pgk-l* X $Pgk-l»- were examined for PGK expression.
Individual samples of epiblast, extraembryonic ectoderm (e.e.e.) and primary
endoderm from a single egg cylinder (Fig. 1 a) were run on cellogel. Figure 1 (b-d)
shows representative gels from the reciprocal crosses and the results are recorded
in Fig. 2.
Of the 33 embryos from the two reciprocal crosses 20 expressed only one
PGK-1 band in the epiblast, and when 15 of these were tested for expression in
the endoderm and e.e.e. only the same PGK-1 band was detected. These 20
embryos were therefore presumed to be hemizygous males.
Thirteen embryos expressed a heterozygous PGK-1A/PGK-1B phenotype in
the epiblast and were therefore female. The expression of the two bands was
130
M. I. HARPER, M. FOSTEN AND M. MONK
(b)
(a)
©
PGK-1A
end
PGK-1B
epi
e.e.e.
A
Id)
(0
©
end
m
PGK-2
©
PGK-1A
PGK-1A
PGK-1B
PGK-1B
epi
end
epi
end
Fig. 1. PGK expression in 6^-day dissected tissues, (a) Diagram of 6i-day egg
cylinder, (b) Male embryo from cross Pgk-lb/Pgk-lb $xPgk-Ja <?. (c) Female
embryo from cross Pgk-lh[Pgk-P $xPgk-la $. id) Female embryo from cross
Pgk-l*/Pgk-la $xPgk-lb cJ. Gel lc includes a sample of testis from a Pgk-la male
to show the position of migration of the autosome-coded testis specific PGK-2
which runs more anodally than the X-coded PGK-1. Gel Id: other female embryos
from this cross showed more equivalent activities in the three tissues. Abbreviations:
epi - epiblast, end - primary endoderm, e.e.e. - extraembryonic ectoderm, A control sample PGK-IA, B - control sample PGK-IB, T - testis sample from
PGK-IA male.
fairly even although some female epiblasts showed slightly more PGK-IA (see
Fig. 1 c, d); others, slightly more PGK-IB. In most cases the e.e.e. and endoderm
regions showed approximately equal activity of PGK-1 to the epiblast, but in
these extraembryonic tissues all the activity was maternal. In four cases a very
pale paternal band was seen in extraembryonic tissues. In one sample of e.e.e.,
from the cross $Pgk-l&/Pgk-l& X $Pgk-lh, the PGK-IB band was uneven and
was almost certainly due to spillover from the standard PGK-IB sample run
Non-random X inactivation in early mouse embryos
(a) 9
15
131
ib) 6
8 10
epi
end
epi
(c) 9
epi
end
end
e.e.e.
(d)
epi
end
Fig. 2. PGK expression in dissected tissues of 33 embryos at 6i days, (a) Female
embryos from cross Pgk-la/Pgk-l* ? x Pgk-lb <J. (b) Male embryos from the same
cross as in (a), (c) Female embryos from cross Pgk-1*/Pgk-1* ? x Pgk-1* 6. id) Male
embryos from the same cross as in (c). Cross-hatched areas - tissues expressing
PGK-1A. Clear areas - tissues expressing PGK-1B.
next to it. The other three cases of paternal bands in endoderm or e.e.e. were
most likely due to contamination of the tissue with epiblast, as in two cases the
embryo dissections had not been recorded as 'clean'.
These results strongly suggest that cell selection is not the cause of the
preferential expression of X m (maternal X chromosome) seen in later derivatives
of e.e.e. and primary endoderm, unless selection could occur even earlier than
6£ days gestation.
PGK expression in blastocysts and blastocyst outgrowths in vitro
Embryo-coded Pgk-1 expression occurs after implantation (Kozak & Quinn,
1975). We investigated whether Pgk-1 expression occurs in vitro, together with
the preferential inactivation of the paternal X chromosome in extraembryonic
tissues as seen in vivo. Mixtures of heterozygous (female) and hemizygous (male)
blastocysts were collected and cultured for up to four days to produce trophoblastic outgrowths. No paternal PGK-1 expression was seen in large numbers
of 3^-day blastocysts (Fig. 3 a, b) or after one day of culture (data not shown).
After 3 and 4 days of culture there was considerable increase in PGK-1 activity
and the paternally derived PGK-1 band could be detected indicating expression
of the embryonic genome (Fig. 3 a, b).
In both crosses the maternally derived enzyme band in 3- and 4-day
132
M. I. HARPER, M. FOSTEN AND M. MONK
(a)
(b)
©
©
PGK-1A
m
^H
PGK-1B
bl
0G4
OG3
A
RGK-1A
A
bl
OGdil
OG3
PGK-1B
B
Fig. 3. PGK expression in blastocysts and in blastocyst outgrowths, (a) Embryos
from cross Pgk-lb/Pgk-lb $xPgk-la <5. Approximately 20 blastocysts and 20
outgrowths were applied to gel. (b) Embryos from crossPgk-la/Pgk-la $ x Pgk-lb <$.
Approximately 25 blastocysts and 12 outgrowths were applied to the gel. Abbreviations : bl - blastocysts, OG4 - 4-day outgrowths, OG3 - 3-day outgrowths,
OG dil - a 1 in 4 dilution of the outgrowth sample.
outgrowths is far in excess of the paternal, and the activity of the former has
increased considerably from that of the blastocysts as shown in Fig. 3 a, b.
Although males should make up 50 % of the embryo population their contribution alone is unlikely to account for the degree of imbalance between the bands.
The preferential expression of the maternal enzyme is almost certainly due to
synthesis by the trophoblastic cells (in line with the in vivo results), which
constitute a large proportion of the outgrowths.
In Fig. 3b a dilution of the outgrowth sample shows that detection of the
paternal PGK band could be diluted out and still leave more maternally derived
activity than seen in the blastocyst band (Fig. 3 b). Therefore even if blastocysts
expressed paternal PGK to the same proportion of total activity as in outgrowths, it would be undetectable due to the low PGK activity at this stage. A
limitation for all samples showing a single PGK-1 band is the argument that if
more activity were applied to the gel a very weak component might then be
detected. However the approximate two-fold difference in X-linked PGK
activity in blastocysts derived from XX and XO mothers (Kozak & Quinn,
1975) is a strong argument that the PGK-1 activity in blastocysts is predominantly maternally derived and not embryo-coded.
In Fig. 3 a there appears to be a third PGK-1 band migrating ahead of
PGK-1 A in the outgrowth samples, which was also visible in a repeat of the gel.
We do not know the explanation for this extra PGK-1 band. Figure 1 c shows
the position of the autosome-coded PGK-2 from the testes.
Non-random X inactivation in early mouse embryos
133
DISCUSSION
Various workers have found that female tissues derived from the trophoblast
and primary endoderm (cell lineages based on Gardner & Papaioannou, 1975)
express only the maternal X whereas those derived from the epiblast have equal,
or almost equal, expression of the two parental X chromosomes (Takagi &
Sasaki, 1975; West et al. 1977, 1978; Frels & Chapman, 1980). Three different
mechanisms (Takagi, 1976) could be proposed for this non-random expression
of the X chromosome in certain tissues:
(1) Random X inactivation with subsequent reversal to inactivate the paternal
X chromosome,
(2) Random X inactivation with subsequent cell selection in favour of the
maternal X chromosome,
(3) Preferential inactivation of the paternal X.
The first explanation appears unlikely, and Takagi (1976) has shown that
there is no reversal of the allocyclic X in 6£- and 7^-day embryos by a doublelabelling technique. It is known that the differentiated state of the X chromosome is very stable, such that clones can be produced with the same X active
through multiple generations (Hamerton et al. 1971; Chapman & Shows, 1976).
Even very rigorous pressures have only succeeded in reactivating the individual
selected gene on the inactive X (Kahan & DeMars, 1975).
If cell selection (2) does occur it is not through the phenotypic pressure of the
maternal reproductive tract (West et al. 1977; Frels & Chapman, 1980). There
does not appear to be anything intrinsically wrong with the paternal X in
trophectoderm and primary endoderm as it is capable of being expressed in the
membranes of XO embryos (Frels & Chapman, 1979). Frels & Chapman (1980)
have studied the PGK-1 expression at 9£ days and found only the maternal band
after electrophoresis of the mural trophoblast, limiting the possible paternal
expression to less than 0-5 %. Takagi's group (Takagi & Sasaki, 1975; Takagi,
Wake & Sasaki, 1978) investigated even earlier stages. In the extraembryonic
portion of the egg cylinder at 1\ days 90 % of the differentially staining X
chromosomes are paternal, and even at 6^- days and 3^- days over 85 % of the
allocyclic X chromosomes are derived from the father.
By (s\ days most, if not all, cells of the embryo have undergone X inactivation
(Takagi & Oshimura, 1973; Monk & Harper, 1979). Our results show that at
this stage both PGK-1 A and PGK-1B bands from heterozygous females appear
of approximately equal intensity in the epiblast, suggesting that X inactivation
is random in this tissue. Although Takagi & Sasaki (1975) found 65 % of
allocyclic X chromosomes were paternal in the embryonic portion of the 7^-day
egg cylinders they examined, this may be due to a skewing effect from primary
endoderm as the authors do not say whether or not this layer was included in
their cell spreads.
The work reported here shows that the extraembryonic ectoderm and the
134
M. I. HARPER, M. FOSTEN AND M. MONK
primary endoderm at d\ days display marked, if not exclusive, expression of the
maternal Pgk-1 allele, arguing against cell selection being the cause of nonrandom expression in certain older tissues. A faint paternal band is detected in
some of these tissues which may be due either to a low level of paternal Xchromosome expression or, more likely, to contamination from the epiblast.
Blastocyst attachment and outgrowth in vitro may mimic implantation in vivo
(Monk & Ansell, 1976; Monk & Petzoldt, 1977). The results presented show
that embryo-coded Pgk-1 expression also occurs in outgrowths. The predominance of the enzyme form coded by the maternally inherited X chromosome is
compatible with the expression of only the maternal allele in the trophoblast
cells. It appears that the X chromosomes of the trophectoderm differentiate at
the blastocyst stage (Monk & Kathuria, 1977; Kratzer & Gartler, 1978) and
those of primary endoderm by 6 days gestation (Monk & Harper, 1979). It is
possible that the paternal allele of Pgk-1 is never active in the trophectoderm or
the primary endoderm and X-chromosome differentiation may occur in these
tissues prior to the embryonic expression of PGK-1.
The results overall strongly support the third mechanism for non-random Xchromosome expression, namely preferential inactivation of the paternally
derived X chromosome in trophectoderm (from which the extraembryonic
ectoderm is derived) and in primary endoderm.
We wish to thank John West for the original PGK-1 A mice, Anne McLaren for help in
establishing our present PGK-1 A colony, and Andy McMahon and John West for helpful
discussion.
REFERENCES
T., BENDER, W., FUNDELE, R., HOFNER, H. & LINKE, I. (1981). Quantitative evaluation of electrophoretic allo and isozyme patterns. FEBS Lett. (In the Press).
CHAPMAN, V. M. & SHOWS, T. B. (1976). Somatic cell genetic evidence for X-chromosome
linkage of three enzymes in the mouse. Nature 259, 665-667.
FRELS, W. I. & CHAPMAN, V. M. (1979). Paternal X chromosome expression in extraembryonic membranes of XO mice. J. exp. Zool. 210, 553-560.
FRELS, W. I. & CHAPMAN, V. M. (1980). Expression of the maternally derived X chromosome
in the mural trophoblast of the mouse. /. Embryol. exp. Morph. 56, 179-190.
GARDNER, R. L. & PAPAIOANNOU, V. E. (1975). In The Early Development of Mammals (ed.
M. Balls & A. E. Wild), 2nd Symposium of the British Society for Developmental Biology,
pp. 107-132, Cambridge University Press.
GARTLER, S. M. & ANDINA, R. J. (1976). Mammalian X-chromosome inactivation. In
Advances in Human Genetics, Vol. 7 (ed. H. Harris & K. Hirschhorn), pp. 99-140. Plenum
Press, New York & London.
HAMERTON, J. L., RICHARDSON, B. J., GEE, P. A., ALLEN, W. R. & SHORT, R. V. (1971).
Non-random X chromosome expression in female mules & hinnies. Nature 232, 312-315.
KAHAN, B. & DEMARS, R. (1975). Localized derepression on the human inactive X chromosome in mouse-human cell hybrids. Proc. natn. Acad. Sci., U.S.A. 72, 1510-1514.
KOZAK, L. P. & QUINN, P. J. (1975). Evidence for dosage compensation of an X-linked gene
in the 6-day embryo of the mouse. Devi Biol. 45, 67-73.
KRATZER, P. G. & GARTLER, S. M. (1978). HGPRT activity changes in preimplantation
mouse embryos. Nature 274, 503-504.
LYON, M. F. (1961). Gene action in the X-chromosome of the mouse (Mus musculus L.).
Nature 190, 372-373.
BUCHER,
Non-random X inactivation in early mouse embryos
135
LYON, M. F. (1972). X-chromosome inactivation and developmental patterns in mammals.
Biol. Rev. 47, 1-35.
MONK, M. (1978). Biochemical studies on mammalian X-chromosome activity. In Development in Mammals, Vol. 3 (ed. M. H. Johnson), pp. 189-223. North-Holland, Amsterdam.
MONK, M. & ANSELL, J. (1976). Patterns of lactate dehydrogenase isozymes in mouse
embryos over the implantation period in vivo and in vitro. J. Embryol. exp. Morph. 36,
653-662.
MONK, M. & HARPER, M. I. (1979). Sequential X chromosome inactivation coupled with
cellular differentiation in early mouse embryos. Nature 281, 311-313.
MONK, M. & KATHURIA, H. (1977). Dosage compensation for an X-linked gene in preimplantation mouse embryos. Nature 270, 599-601.
MONK, M. & PETZOLDT, U. (1977). Control of inner cell mass development in cultured
mouse blastocysts. Nature 265, 338-339.
NIELSEN, J. T. & CHAPMAN, V. M. (1977). Electrophoretic variation for X-chromosome-linked
phosphoglycerate kinase (PGK-1) in the mouse. Genetics 87, 319-325.
TAKAGI, N. (1976). Stability of X chromosome differentiation in mouse embryos. Hum.
Genet. 34, 207-211.
TAKAGI, N. & OSHIMURA, M. (1973). Fluorescence and Giemsa banding studies of the
allocyclic X chromosome in embryonic and adult mouse cells. Expl Cell Res. 78, 127-135.
TAKAGI, N. & SASAKI, M. (1975). Preferential inactivation of the paternally derived X
chromosome in the extraembryonic membranes of the mouse. Nature 256, 640-642.
TAKAGI, N., WAKE, N. & SASAKI, M. (1978). Cytological evidence for preferential inactivation
of the paternally derived X chromosome in XX mouse blastocysts. Cytogenet. Cell Genet.
20, 240-248.
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.
WEST, J. D., PAPAIOANNOU, V. E., FRELS, W. I. & CHAPMAN, V. M. (1978). Preferential
expression of the maternally derived X chromosome in extraembryonic tissues of the
mouse. In Genetic Mosaics and Chimaeras in Mammals (ed. L. B. Russell), pp. 361-377.
Plenum Press, New York & London.
WHITTINGHAM, D. G. & WALES, R. G. (1969). Storage of two-cell mouse embryos in vitro.
Aust. J. biol. Sci. 22, 1065-1068.
(Received 29 June 1981, revised 15 July 1981)