Molecular Basis of Differential Gene Expression in Themouse Preimplantation Embryo

Table of contents:
- Overview of preimplantation development
- Alteration of transcription and remodeling of chromatin
in the early preimplantation mouse embryo
Maternal to embryonic transition of transcription
Remodelling of chromatin
- Signaling pathways in the early preimplantation embryo
Signaling pathways involved in the maternal to
embryonic transition
Mitosis promoting factor
Protein kinases C and A
Signaling pathways involved in compaction and
polarization
E-cadherin-catenin
Protein kinase C and E-cadherin
- Establishment of differential gene expression
- Conclusion
- Acknowledgements
- References
Overview of preimplantation development
During preimplantation development, a fertilized egg
develops into a blastocyst that is able to implant into
the uterine wall. Morphological changes during preimplantation development can be categorized into three
main stages: increase in cell number (cleavage), cellular flattening and polarization (compaction), and production of an embryonic cavity or blastocoel (blastulation). Cleavage occurs in all preimplantation stages;
however, an increase in cell number is more noticeable
in the first four stages, which are the “1-cell stage”
(zygote), “2-cell stage”, “4-cell stage”, and the “8-cell
stage”. During cleavage, transcription from the embry

onic genome begins, maternal mRNA degrades to a
large extent, and the control of development switches
gradually from maternal to embryonic (Schultz et al.,
1999). During the fourth cell cycle, the cells (blastomeres) of the 8-cell stage embryo start to polarize
and flatten against each other, to produce a “partially
compacted 8-cell stage” embryo, which with the
increase in membrane-membrane adhesion of adjacent
blastomeres converts into a ball-like structure called
the “fully compacted 8-cell stage” or “late 8-cell
stage” (Figure 1). Since the blastomeres at the end of
the fourth cell cycle undergo mitosis and cytokinesis,
the 16-cell stage embryo at the beginning of fifth cell
cycle appears “de-compacted”, but after cytokinesis, it
converts into a fully compacted 16-cell stage embryo
or “morula”. The polarized blastomeres of the 16-cell
stage embryo simply divide to give rise to the two different cell types of the sixth cell cycle (Rossant and
Vijh, 1980; Rossant and Tam, 2004); outer cells (future
trophoblast) and inner cell mass cells (future ICM). In
the 32-cell stage morula, epithelial-type junctional
complexes form between trophoblasts. When a blastocoel begins to form during the 32-cell stage, the
embryo is called a blastocyst. In “early blastocysts”,
the size of the blastocoel is about 1/3 of the size of the
inner cell mass (ICM). The “blastocyst stage” embryo
in the seventh cell cycle has approximately 64 cells
and has developed a blastocoel approximately equal in
size to the ICM. In addition, at this stage a differentiated layer of primitive endoderm (PE) or hypoblast has
developed in the vicinity of blastocoel (Rossant and
Tam, 2004). During the eighth cell cycle, in the
“expanded blastocyst” the blastocoel comes to fill
almost all of the internal space in the embryo.
Preimplantation development concludes at this time
with the release of the embryo from the zona pellucida
(hatching) and its implantation into the uterine wall
(Becker and Davies, 1995; Johnson, 1996).
After fertilization during preimplantation development, there are three major cellular transitions. These
are transition from the maternal to embryonic control
of development, blastomere polarization and compaction, and blastocoel formation. In this review, the
molecular basis of differential gene expression during
these transitions is discussed. More specifically, the
roles of chromatin remodeling and cell signaling pathways in the establishment of differential gene expression will be elaborated in detail.
Alteration of transcription and remodeling of
chromatin in the early preimplantation mouse
embryo
Maternal to embryonic transition of transcription:
During oocyte maturation and 12h before ovulation,
the germinal vesicle breaks down, and this signals the
beginning of degradation of much of the RNA that is
accumulated during oocyte growth. At the same time,
the rate of protein synthesis declines, due to degradation of RNA or to translational control. The time
course for decay of maternal transcripts varies between
genes (Gosden et al., 1997). In fact, there is a complicated network of regulatory mechanisms, where stored
RNAs are selectively polyadenylated for
translation/degradation rather than being affected
globally. Specific mRNAs are stored in the cytoplasm
as mRNA-protein complexes and are isolated from the
translational apparatus by masking proteins (Curtis et
al., 1995; Verrotti et al., 1996).
The maternal to embryonic transition is the switch
in control of development from products of the maternal genome to products of the embryonic genome
(Telford et al., 1990). While full control of development by embryonic transcripts takes at least until the
blastocyst stage, the “switch” is experimentally
defined as the time of the first burst of transcription
from the embryonic genome. This corresponds with
when development becomes sensitive to transcriptional inhibitors (Telford et al., 1990). In mice the experimentally defined switch is at the early 2-cell stage.
However, most proteins in the embryo will still be
maternally derived at this point (Figure 1). In addition,
recent works indicate that genes involved in ribosome
biogenesis and assembly, protein synthesis, RNA
metabolism and transcription are over-represented at
the two-cell stage, suggesting that genome activation
during the 2-cell stage may not be as global and
promiscuous as previously proposed (Zeng and
Schultz, 2005) and not all the necessary transcripts are
made by the embryonic genome.
Approximately 70-90% of the polyadenylated RNA
in unfertilized eggs is lost between fertilization and the
late 2-cell stage, although there is little, if any, difference in total RNA content (Piko and Clegg, 1982;
Hamatani et al., 2004). This indicates that much of the
oocyte mRNA was for the purposes of oogenesis and
the early stages of post-fertilization development. For
example, it is shown that the translation of maternal
RNA is required for the initiation of zygotic genome
activation (Hamatani et al., 2006). It also indicates
possible functions for non-mRNA forms of RNA
which for example may be involved in epigenetic regulations of the embryonic genome (Rassoulzadegan et
al., 2006). The pattern for mRNA levels of common
structural and housekeeping genes is U-shaped with a
nadir at the late 2-cell stage and rising concentrations
after the 2-cell stage. Although, most maternal messages are gone by the end of the 2-cell stage in mice,
depletion of maternal proteins occurs over the next few
cleavage divisions (Kidder, 1992a; Piko et al., 1984).
With the expression of the embryonic genome, the
maternal components that direct early development
begin to be replaced (Schultz, 2002), and new products
characteristic of preimplantation development appear.
This has been shown by changes in the patterns of
metabolically labelled proteins in high-resolution twodimensional gel electrophoresis, during different times
after fertilization. The most pronounced changes are
due to the synthesis of proteins at the mid 2-cell stage,
which can be inhibited by α-amanitin (Latham et al.,
1991). In mouse, there are other lines of evidence indicating that the one-cell embryo has potential for transcription: 1) The concentrations of the transcription
factor Sp1 and the TATA box-binding protein (TBP)
increase in pronuclei of one-cell embryos in a timedependent fashion (Worrad et al., 1994). 2) Functional
RNA polymerase I and III are present in one-cell
embryos (Nothias et al., 1996). 3) 5-bromouridine 5'-
triphosphate sodium (BrUTP) is incorporated into
pronuclei of one-cell embryos, and incorporation is
sensitive to α-amanitin and RNase treatments (Aoki et
al., 1997 and 2003). 4) Injection of a luciferase
reporter gene under the control of the SV40 early promoter, into the male pronucleus at the early S phase
results in detectable luciferase activity in G2 one-cell
embryos (Ram and Schultz, 1993).
Remodeling of chromatin: Replication of DNA in the
first cell cycle could facilitate the access of maternally

derived transcription factors to their cis-acting DNAbinding sequences prior to the formation of nucleolsomes (Davis and Schultz, 1997). At the one-cell
stage, the two haploid pronuclei enter the S phase as
they migrate toward each other. Indeed, between start
of DNA replication and assembly of nucleosomes,
transcription factors are able to bind to DNA and start
transcription (Schultz et al., 1999; Schultz, 2002).
Aphidicolin treatment, that inhibits entry into the S
phase and DNA synthesis, decreases transcription of
some transcripts e.g. eukaryotic initiation factor 1A
(eIF-1A) (Davis et al., 1996). This treatment also
decreases BrUTP incorporation in G2 of the first cell
cycle, but only by 35% (Aoki et al., 1997 and 2003).
This indicates that there are two classes of genes, those
whose transcription is independent of DNA replication, and those whose transcription is linked to DNA
replication (Schultz et al., 1999; Schultz, 2002). In
another words, chromatin organization is partly
responsible for a transcriptionally repressive state in
the 1-cell stage embryo. There also is a higher rate of
BrUTP incorporation in male pronuclei, where protamines are replaced by egg histones. Similar to DNA
replication, the process of histone replacement also
may provide un-wrapped DNA to which transcription
factors bind (McLay and Clarke, 1997). In this regard,
it has been shown that a subclass of the high-mobilitygroup (HMG) proteins (a family of abundant lowmolecular weight mammalian chromosomal proteins),
HMG-I/Y, translocates into pronuclei of one-cell
embryos during the first round of DNA synthesis, and
that this promotes transcription (Beaujean et al., 2000).
HMG-I/Y may help to replace a subtype of histone H1
with histone H1°, which accumulates in the oocyte during oogenesis (Clarke et al., 1992). HMG-I/Y has high
affinity for the AT-rich sequences found in scaffold or
matrix-associated regions (SARs/MARs) and is able to
displace histone H1 and increase chromatin accessibility (Thompson et al., 1994; Thompson, 1996). The
result would be an increase in the rate of transcription
which by microarray data are shown to be related to
3254 genes (Hamatani et al., 2004).
Other evidence for the involvement of chromatin in
induction of a transcriptionally repressive state, comes
from studies that have shown that the activities of promoters and replication origins from the late 1-cell
stage to the 4-cell stage is repressed, and that this
repression can be relieved by either sodium butyrate
(inhibitor of histone deactylase) or by enhancers
(Majumder et al., 1993b; Wiekowski et al., 1997).
Recently, it has been shown that brahma-related gene
1 (BRG1), the catalytic subunit of a chromatin remodeling complex, SWItch/Sucrose NonFermentable
(SWI/SNF), is essential for maternal to embryonic
transition and is derived from maternal protein stores
in the oocyte (Bultman et al., 2006).
The dramatic biochemical changes of chromatin
have been observed during the early stages of pronuclear formation in the early preimplantation embryo
and several findings point to the importance of its
remodeling (Figure 1). The findings are: 1) Paternal
pronuclei, in the process of replacing their sperm-specific histones with somatic histones, have higher levels
of transcription than female pronuclei (Perreault,
1992; Thompson, 1996; Adenot et al., 1997; Aoki et
al., 1997). 2) Acetylated forms of several histones
(H4.Ac5, 8, 12; H3.Ac9/18 and H2A.Ac5) are transiently enriched in the nuclear periphery at the two-cell
stage. This enrichment is less frequently observed in
one-cell embryos and not at all at the 4-cell stage
(Worrad et al., 1995; Stein et al., 1997; Adenot et al.,
1997). In contrast, H3.Ac14, H3.Ac23, H4.Ac16, and
acetylated H2B are uniformly distributed throughout
the nucleus (Stein et al., 1997). 3) Different forms of
methylated, phosphorylated, and acetylated hiostones
H3, H4, and H2A have also been shown to stably and
dynamically mark the genome of the early mouse
embryo (Sarmento et al., 2004). 4) Somatic histone H1
is first detectable at the 4-cell stage (Clarke et al.,
1992) and its increase corresponds with a decrease in
HMG-I/Y. 5) Promoters without enhancers of microinjected plasmid DNA are transcribed in 1-cell embryos,
but strongly repressed in 2-cell embryos. This repression can be relieved by the inhibition of histone
deacetylases using sodium butyrate (Wiekowski et al.,
1991 and 1997; Majumder et al., 1993a; Nothias et al.,
1995). And 6) nuclei of early two-cell mouse embryos
readily support normal embryonic development when
transplanted into enucleated one-cell embryos, whereas nuclei from more advanced embryos do not
(McGrath and Solter, 1984; Robl et al., 1986; Howlett
et al., 1987). Aside from the above biochemical
changes of chromatin, however, it remains to be
demonstrated whether chromatin bears any structural
changes (Dehghani et al., 2005a) during the preimplantation period of development.
Signaling pathways in early preimplantation
embryo
Signaling pathways involved in the maternal to
embryonic transition: Presence of proper extracellu

lar signals is necessary to induce cells with appropriate
developmental history to take a specific developmental route. Different signaling pathways have been
found to be functional in the preimplantation mouse
embryo including those that are related to protein
kinase C (PKC), Wnt and its intracellular partners,
bone morphogenetic protein (BMP) Notch (Wang et
al., 2004a), mitogen activated protein (MAP) kinase
activated by Ras (Natale et al., 2004; Paliga et al.,
2005; Maekawa et al., 2005; Wang et al., 2004b), protein kinase A (PKA), and receptor tyrosine kinase (Heo
and Han, 2006).
- Mitosis promoting factor: During fertilization the
metaphase-II related arrest of the oocyte is broken by
fertilization. The hormonal signal stimulates the maturation promoting factor (MPF), now called the mitosis
promoting factor activity, and the sperm-derived signal
destroys CSF (cytostatic factor) activity. These activities are, in essence, kinase activities that regulate the
meiotic cell cycle. MPF is a single protein kinase that
induces mitosis. It has been shown that it is activated
by dephosphorylation of tyrosine and threonine, and
phosphorylation of threonine 161. Sustained phosphorylation of threonine 14 and dephosphorylation of
tyrosines inactivates MPF (Whitaker, 1996).
Experiments have identified a protein-serine/threonine
kinase known as Mos as an essential component of
CSF (Dekel, 1996). Mos is specifically synthesized in
oocytes around the time of completion of meiosis I and
is then required both for the increase in MPF activity
during meiosis II and for the maintenance of MPF
activity during metaphase II arrest. The downstream
kinase of Mos is Rsk, which inhibits action of the
anaphase-promoting complex and arrests meiosis at
metaphase II. At fertilization, the increase in cytosolic
Ca2+ signals the completion of meiosis. The anaphasepromoting complex will be activated by increase in
Ca2+. The resultant inactivation of MPF leads to completion of the second meiotic division, with asymmetric cytokinesis (as in meiosis I) giving rise to a second
small polar body (Cooper, 2000).
- Protein kinases C and A: The role played by PKC
in events associated with fertilization is controversial.
A study shows that a PKC activator, 4β-phorbol 12-
myristate 13-acetate (PMA), induces Ca2+ oscillations
in mouse oocytes (Cuthbertson and Cobbold, 1985),
whereas in human oocytes, it stops the oscillations,
whether added before or after sperm (Sousa et al.,
1996a,b; Sousa et al., 1997). Calcium ion oscillations
are intracellular changes in the concentration of calcium, which is closely related to fertilization. Also, staurosporine (a PKC inhibitor) causes an increase in intracellular Ca2+ (Jones et al., 1995; Jones, 1998). These
findings indicate that PKC might be an important signaling molecule to control the levels of intracellular
calcium. In the case of cortical granule (CG) release
which blocks the simultaneous entry of several sperms
into the oocyte (poly-spermy), it has been shown that
12-O-tetradecanoyl phorbol 13-acetate (TPA) and 1-
oleyl-2-acetyl-sn-glycerol (OAG); a compound structurally similar to diacyl glycerol (DAG; one of the second messengers produced by the action of phospholipase C that leads to activation of PKC), caused CG
release (Colonna and Tatone, 1993). In another study,
the CG release induced by phorbol esters was blocked
by PKC inhibitors, but the same inhibitors failed to
have any effects on the extent of CG release caused by
spermatozoa (Ducibella and LeFevre, 1997). This indicates that there is a biochemical pathway in oocytes in
which PKC activation leads to CG release but that this
pathway is not used by the spermatozoon at fertilization.
The role of PKC in egg activation is also disputed.
One of the criteria that has been used to assess egg activation is second polar body formation. Gallicano et al.
(1993) reported that phorbol esters can induce extrusion
of a second polar body in hamster oocytes. Since up to
50% of these polar bodies resorb within 1h of addition
of PMA, Moore et al. (1995) has argued that this may
not be a bona fide polar body, because cytokinesis does
not take place and its formation may be due to PKC disruption of the metaphase spindle, or to disruption of
cytoskeletal structure. Gallicano et al. (1997) reported
that PKC activation causes extrusion of the second
polar body in mouse oocytes and that the polar body is
resorbed after a few hours, similar to the case in hamsters. However, other studies have shown that phorbol
esters do not induce second polar body formation in
mouse (Cuthbertson and Cobbold, 1985; Colonna et al.,
1989; Moore et al., 1995). In addition, Ducibella and
LeFevre (1997) examined a myristoylated pseudosequence over a similar dose range to the one used by
Gallicano et al. (1997) for inhibition of second polar
body formation, and found it to be highly toxic. To date,
the study of PKC at fertilization and egg activation has
been limited to pharmacological manipulation that
relies heavily on phorbol esters, and these agents are not
specific for binding to PKC (Ahmed et al., 1993;
Wilkinson and Hallam, 1994; Kazanietz et al., 1995).
Therefore, the conclusions remain controversial.

There are several studies that suggest a role of protein phosphorylation in embryonic gene activation. An
inhibitor of PKA, H8 (N-2-methylaminoethyl isoquinoline-5-sulfonamide dihydrochloride), prevents
synthesis of the transcription requiring complex (TRC)
(Poueymirou et al., 1989). This effect of H8 on TRC
synthesis is likely to be at the level of transcription,
since TRC is an embryonic product of the two-cell
stage embryos (Schultz, 1993). This protein is detected following in vitro translation of RNA obtained from
2-cell embryos, but not from one-cell or 2-cell
embryos cultured in the presence of α-amanitin.
Indeed, H8 and α-amanitin, have similar effects on
TRC synthesis. These two compounds also inhibit the
increase in heat shock protein (hsp) 70 mRNA between
the one- and 2-cell stages (Manejwala et al., 1991).
Culture of one-cell embryos in cycloheximide under
conditions that inhibit more than 95% of protein synthesis does not prevent the increase in hsp 70 mRNA,
indicating that PKA affects maternally derived proteins, which are involved in transcription at the onecell stage. Inhibitors of the calmodulin-dependent protein kinase and PKC, do not prevent embryonic gene
activation (Schultz, 1993). We have shown that all of
the isoforms of PKC are present between the 2-cell and
blastocyst stages of mouse preimplantation development, and that each has a distinct, dynamic pattern and
level of expression (Pauken and Capco, 2000;
Dehghani and Hahnel, 2005). A transient increase in
the nuclear concentration of PKC δ and ε during the
early 4-cell stage has been shown to affect transcription (Dehghani et al., 2005b).
Signaling pathways involved in compaction and
polarization:
- E-cadherin-catenin: During the 8-cell stage, polarization is accompanied by intercellular adhesion mediated by the E-cadherin-catenin system (Larue et al.,
1994; Huber et al., 1996). The adhesion results in compaction and formation of incomplete apicolateral junction complexes (Fleming et al., 2000). Further biogenesis of the junctions, transforms the proto-epithelial
phenotype of 8-cell blastomeres into the mature
epithelial phenotype of trophoblast cells. Tight junctions, adherent junctions, desmosomes, and gap junctions are involved in this transformation. By approximately the 30-cell stage, functional junctional complexes have formed between apices of the outer cells
of the morula. This coincides with commitment of the
outer cells to become trophoblast cells as discussed
above. Polarization and compaction also involve
changed distribution of cytoskeletal elements (e.g.
actin filaments, microtubules), cytoplasmic organelles
(e.g. endocytic vesicles), microvilli, and components
of the cell cortex (e.g. actin binding proteins) (Fleming
et al., 1993; Fleming et al., 1994). In E-cadherin null
embryos, compaction occurs due to maternally inherited E-cadherin, but proper blastocysts do not form
(Larue et al., 1994; Kan et al., 2007). Although null
embryos form desmosomes and tight junctions, they
cannot maintain a coherent epithelium and die at
implantation. This was confirmed by treating homozygous null embryos with an antibody that blocks E-cadherin interaction and through removal of Ca2+
(required for E-cadherin interaction) (Riethmacher et
al., 1995). Together, these experiments clearly demonstrated the importance of E-cadherin in both the formation and maintenance of a polarized epithelium in the
preimplantation embryo. It was hypothesized that Ecadherin induces cytocortical polarization and that this
leads secondarily to polarization within the cytoplasm
(Fleming et al., 2001). Clayton and colleagues (1995),
using inhibitors of protein and cytoskeletal assembly,
showed that adhesion via E-cadherin is independent of
its surface expression. This suggested that the intracellular component of E-cadherin signaling pathway was
required for adhesion (Sefton et al., 1996). Molecular
analysis of cadherin-mediated adhesion complexes in
a variety of epithelial tissues elucidated the central role
of β-catenin. It not only binds to the cytoplasmic
domain of E-cadherin, α-catenin and filamentous
actin, but is also involved in the activation of several
target genes as transcription factor (Gumbiner, 1995;
Nollet et al., 1999). During mouse preimplantation
development, both α-catenin and β-catenin are maternally provided as proteins and mRNAs, and their transcription from the embryonic genome begins at the late
2-cell stage (Huber et al., 1996). Embryos null for β-
Catenin form blastocysts, implant and develop until
the egg-cylinder-stage embryos (Haegel et al., 1995),
however, α-catenin null embryos fail to form a functional trophectoderm (Torres et al., 1997).
There are several lines of evidence that emphasize
the role of protein phosphorylation in post-translational modifications related to compaction and polarization. Transcription and translation of embryonic genes
required for compaction take place which are then
completed by the late 4-cell stage (Kidder and
McLachlin, 1985; Levy et al., 1986). Bloom and
McConnell (1990) showed that some phosphoproteins
were only found in compacted 8-cell embryos and not
in other stages, suggesting a link between post-transla

tional mechanisms and compaction. Sefton and colleagues (1992) showed that the onset of uvomorulin
phosphorylation coincides with compaction and
hypothesized that this event converts uvomorulin from
a non-adhesive to an adhesive form. However, cell
flattening and gap junction formation take place in the
absence of E-cadherin phosphorylation, and staurosporine, an inhibitor of protein kinase activity, causes premature intercellular flattening of blastomeres
(O’Sullivan et al., 1993). This has been further tested
by using 6-dimethylaminopurine (6-DMAP), a serinethreonine kinase inhibitor that is able to induce premature cell flattening and gap junction formation at the 4-
cell stage. Premature flattening was inhibited when the
embryos were cultured in the presence of an anti-Ecadherin antibody or without extracellular Ca2+,
demonstrating that 6-DMAP-stimulated compaction
requires functional E-cadherin. Although, the direct
relationship of protein kinase inhibition with E-cadherin is not clear, however, it is obvious that compaction is affected by phosphorylation.
- Protein kinase C and E-cadherin: Parallel experiments with PKC, suggested that this kinase might be
involved in compaction. Yamamura et al. (1989)
found that PKC activators increased adhesion of cells
in 2-, 4-, and un-compacted 8-cell embryos. Soon
after, it was shown that this increased adhesion can be
inhibited by a monoclonal antibody to E-cadherin.
Indeed, PKC activation causes a rapid shift in the
localization of E-cadherin molecules, indicating that
PKC plays a role in the initiation of compaction via
direct or indirect effects on E-cadherin (Winkel et al.,
1990). One study showed that β-catenin, a subunit of
the cadherin protein complex, becomes phosphorylated during compaction, on serine/threonine residues
and at the same time PKC α redistributes to contact
sites as compaction initiates (Pauken and Capco,
1999), suggesting that β-catenin might be phosphorylated by this isozyme. In contrast, another study
showed that β-catenin is a major tyrosine-phosphorylated protein in oocytes and early cleavage-stage
embryos, and that the relative amount of phosphorylated β-catenin is greatly reduced during the morula-blastocyst transition, suggesting that tyrosine phosphorylation of β-catenin may represent a molecular mechanism to prevent E-cadherin from becoming adhesive
(Ohsugi et al., 1999). Additionally, a role for the
myosin light-chain kinase in activation of compaction
has been proposed (Kabir et al., 1996). Hence, the
relationship between the spatial location of a single
isozyme and a temporal event of preimplantation
development can be investigated by activation/inhibition of each isozyme individually.
In conclusion, homophilic adhesion between E-cadherin molecules is a primary regulator of compaction
and trophoblast differentiation. Phosphorylation
/dephosphorylation reactions are important for assembly of the E-cadherin complex. Molecular partners of
E-cadherin and their order of interaction during compaction remain to be identified. Since all the activators/inhibitors of PKC that have been used to date,
affect the family of PKC isozymes, the role of individual isozymes in this event is obscure. Also, future
experiments should determine which other cell adhesion systems are involved in compaction of preimplantation mammalian embryos.
Establishment of differential Gene expression
It is believed that the formation of different populations of cells is established during polarization and
compaction of 8-cell stage embryo. The individual 1/2,
1/4, and early 1/8 blastomeres arise by approximately
equal, but asynchronous cleavage divisions. They are
roughly spherical, radially symmetric, have no consistent developmental fate, and are totipotent (Johnson,
1996). During the fourth cell cycle, the 1/8 blastomeres undergo a process of cellular flattening (compaction) and cellular polarization (Figure 1). These
processes are the initial steps in the formation of a
communicating polarized epithelium. At the end of the
fourth cell cycle, the polarized blastomeres, now called
polarblasts cleave to produce two-cell types in the 16-
cell embryo, polarized outer cells (polarblasts), and
non-polar inner cells (pluriblasts) (Johnson and
McConnell, 2004; Johnson, 1996; Johnson and
Selwood, 1996).
An important aspect of development is establishment of a “differential gene expression” program.
While the presence of morphological differences
between outer and inner cells is only noticeable for the
first time at the 16-cell stage, every stage of preimplantation development displays a unique pattern of
gene expression (stage-specific gene expression)
(Kidder, 1992a). Transcription of several early embryonic genes has been shown to be stage-specific. For
example embryonic alkaline phosphatase (EAP) , histone H3, γ-actin, and connexin-43 are transcribed at
the 2-cell stage. However, β-actin and transforming
growth factor α (TGF-α) are not transcribed until the

4-cell stage, and glucose transporter 2 (GLUT-2) and
epidermal growth factor receptor (EGF-R) until the 8-
cell stage (Kidder, 1992b; Kidder, 1993). Interestingly,
some genes are transiently expressed only at the 8-cell
stage, U2af binding protein-related sequence (U2afbprs) (Latham et al., 1995). U2 auxiliary factor (U2AF)
is a non-snRNP (small nuclear ribonucleoprotein) protein which is required for the binding of U2 snRNP to
the pre-mRNA branch site. Some genes that are transcribed during oogenesis are not transcribed during the
preimplantation period, e.g. connexin-32, however,
some genes are transcribed during both periods, e.g. Ecadherin. Expression of many genes that are transcribed during the cleavage stages becomes cell typespecific in the morula and blastocyst, i.e. zona occludens 1 (ZO-1; tight junction protein 1), plakoglobulin
(gamma-catenin, a component of desmosomes),
claudin (a component of desmosomes), EAP, Na+-K+-
ATPase-α, oct-4 (Octamer-4, a homeodomain transcription factor of the POU family), Mash-2 (mammalian achaete-scute homologous protein-2) (Hahnel
et al., 1990; Guillemot et al., 1994; MacPhee et al.,
1994; Collins and Fleming, 1995; Yeom et al., 1996;
Dehghani et al., 2000; Moriwaki et al., 2007). Other
genes are only transcribed and expressed by outer and
trophoblasts later in differentiation, i.e. Desmocolin-2,
and several integrins (Collins et al., 1995; Sutherland
and Calarco-Gillam, 1983).
Networks of epigenetic pathways directly or indirectly (through chromatin) regulate transcription. The
stable determination of cell fate requires factors that
initiate transcriptional patterns and mechanisms that
sustain these patterns over time and through multiple
cell divisions (Hagstrom and Schedl, 1997).
Chromatin organization plays a role in the cellular
memory that maintains stable states of transcription
(Jacobs and van Lohuizen, 1999). In Drosophila
Melanogaster, two groups of proteins, the Polycomb
and trithorax groups provide transcriptional memory
by “freezing” transcription states. The Polycomb
group of proteins repress the expression of homeotic
genes (which determine the identity of the different
body segments along the anterior-posterior axis),
whereas the trithorax group proteins sustain expression of these genes (Hagstrom and Schedl, 1997;
Jacobs and van Lohuizen, 1999). These proteins have
been shown to have the same responsibility in embryoderived stem cells (Boyer et al., 2006). In mammalian
embryos, silencing and propagation of the silenced
state of one of the two X chromosomes within a
diploid female nucleus provides an example of transcriptional cellular memory. Much attention has been
focused on differential DNA methylation as a marker
for imprinting and X inactivation. However, it is
unclear whether DNA methylation acts as a primary
determinant of differential gene activity or whether it
simply reflects changes in chromatin structure that
determine differential activity (Wolffe, 1996; Wolffe
and Pruss, 1996). It appears that both nucleoprotein
organization and acetylation patterns are important
factors in the maintenance of the differential gene
activity of active and inactive X chromosomes. It has
been proposed that a superabundance of chromosomal
proteins or transcription factors specific for large
domains of DNA or individual genes, could maintain
active and repressive chromatin structures during
DNA replication (Wolffe, 1994).
Conclusion
Acetylation, phosphorylation, methylation and other
kinds of histone modification, alteration of long-range
chromatin, and stable incorporation of chromosomal
proteins into the structure of DNA are the major epigenetic modifications that target DNA and program transcription. These modifications can be regulated by cell
signaling pathways (Zlatanova and van Holde, 1992;
Bestor et al., 1994; Owen-Hughes and Workman,
1994; Edmondson and Roth, 1996; Felsenfeld et al.,
1996; Felsenfeld, 1996; Patterton and Wolffe, 1996;
Dillon et al., 1997; Elgin and Jackson, 1997; Vermaak
and Wolffe, 1998; Kornberg, 1999; Schreiber and
Bernstein, 2002). Presence and activity of several cell
signaling components have been experimentally confirmed in the preimplantation embryo. However, it
remains to be identified how different signals are
orchestrated, and which signals directly create, organize, and induce the molecular program of differential
gene expression.
Acknowledgments
I would like to thank Professor Ann Hahnel, (University of
Guelph, Canada) and Professor David Bazett-Jones
(University of Toronto, Canada) for their exciting and insightful research and guidance that has inspired me to prepare
this review.