A clean start: degradation of maternal proteins at the oocyte-to-embryo transition.

Abstract

In many organisms, the transition from oocyte to embryo occurs in the absence of mRNA transcription. Therefore, early developmental programs rely on maternal mRNAs and proteins that are synthesized during oogenesis. The regulated translation of maternal RNAs is essential for the proper deployment of regulatory factors during early embryogenesis. Recent studies suggest that the degradation of maternal proteins by the ubiquitin-proteasome pathway is also crucial for the oocyte-to-embryo transition. In this article, we explore the hypothesis that the coordinated degradation of germline proteins is essential for remodeling the oocyte into a totipotent zygote that is capable of somatic development.

Full text

Page 1

A clean start: degradation of maternal
proteins at the oocyte-to-embryo
transition
Cynthia DeRenzo and Geraldine Seydoux
Department of Molecular Biology and Genetics, Johns Hopkins School of Medicine, 725 North Wolfe Street, 515 PCTB, Baltimore,
MD 21205-2185, USA
In many organisms, the transition from oocyte to
embryo occurs in the absence of mRNA transcription.
Therefore, early developmental programs rely on
maternal mRNAs and proteins that are synthesized
during oogenesis. The regulated translation of maternal
RNAs is essential for the proper deployment of regulat-
ory factors during early embryogenesis. Recent studies
suggest thatthe degradation ofmaternal proteins bythe
ubiquitin–proteasome pathway is also crucial for the
oocyte-to-embryo transition. In this article, we explore
the hypothesis that the coordinated degradation of
germline proteins is essential for remodeling the oocyte
into a totipotent zygote that is capable of somatic
development.
Multicellular organisms begin their development with a
remarkable transition. At fertilization, the quiescent egg
is transformed into a dynamic embryo that is ready to
differentiate into many cell types. This transition begins
with oocyte maturation when a hormonal or sperm-
dependent cue signals to the egg to resume meiosis and
prepare for fertilization. In turn, fertilization triggers,
probably through Ca2Coscillations, several cellular and
developmental changes that eventually remodel the egg
into a totipotent zygote [1]. These changes include the
switch from meiosis to mitosis and the switch from a germ
cell to a zygote that can generate both somatic and germ
cells. In many organisms, these changes occur before the
onset of mRNA transcription and, therefore, depend on
modulation of the maternal dowry of RNAs and proteins
that are present in the oocyte. The recruitment of
maternal RNAs for translation has long been recognized
as being a widespread mechanism to generate new
proteins in maturing oocytes and fertilized eggs
(for review, see Ref. [2]). Conversely, RNAs that are no
longer needed are actively degraded in the early embryo
(for review, see Ref. [3]). In this article, we explore the
hypothesis that protein degradation functions as a
complementary mechanism to erase the oogenic program
and drive eggs into embryogenesis. We describe recent
published examples of oocyte-protein degradation in
Caenorhabditis elegans and consider the possibility that
similar mechanisms operate in other organisms (Table 1).
Transitioning from meiosis to mitosis: degradation of
MEI-1 and MEI-2
Before fertilization, oocytes arrest during meiosis, typi-
cally in prophase of meiosis I or after maturation in
Table 1. Oocyte proteins targeted for degradation in different organismsa
OrganismTargetDegron E3 ligase Regulator promotes (P) or
inhibits (I) degradation
Refs
Caenorhabditis
elegans
MEI-1 PESTMEL-26–CUL-3 MBK-2 kinase (P) [5,8,11,13–15,17,21,51,52]
OMA-1PEST MBK-2 kinase (P) [18–21]
PIE-1
POS-1
MEX-1
MEX-5
MEX-6
CCCH fingerZIF-1–CUL-2MEX-5 and MEX-6 (P)
PAR-1 kinase (I)
MBK-2 kinase(P)
[21,32,33]
XenopusCPEBPESTCDC2 kinase (P)[23–27]
DrosophilaOskarPar-1 kinase (I)[42]
VasaFat facets DUB (I)
Gustavus
[45,46]
ZebrafishVasa[50]
aAbbreviations: DUB, deubiquitinating enzyme; ZIF-1, zinc-finger-interacting factor 1.
Corresponding authors: Cynthia DeRenzo (cderenz@jhmi.edu), Geraldine Seydoux (gseydoux@jhmi.edu).
www.sciencedirect.com0962-8924/$ - see front matter Q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.tcb.2004.07.005
Review
TRENDS in Cell BiologyVol.14 No.8 August 2004

Page 2

meiosis II. Fertilization triggers the completion of
meiosis and the transition to mitosis. The meiotic and
mitotic spindles differ in their behavior and morphology
(Figure 1) but, remarkably, these structures form in the
same cytoplasm, often within minutes of each another
[4]. In C. elegans, the transition from the small meiotic
spindle to the large mitotic spindle requires down-
regulation of the microtubule-severing complex katanin
[4,5]. The katanin complex is a p60–p80 heterodimeric
complex that is capable of disassembling microtubules
[6]. In C. elegans, the katanin subunits are encoded by
the mei-1 and mei-2 genes and are required specifically
for meiosis: loss-of-function mutations in either mei-1 or
mei-2 disrupt meiosis without affecting mitosis [7].
MEI-1 and MEI-2 proteins accumulate on the meiotic
spindle and are rapidly turned over after meiosis [4,8].
Mutations that block the turnover of MEI-1 and MEI-2
cause the proteins to accumulate on the mitotic spindle
and interfere with spindle elongation and rotation [4,5].
Two classes of mutation were initially found to block
MEI-1 and MEI-2 turnover: a mutation in mei-1 itself [9]
and mutations in a second locus, mel-26 [5]. The mei-1
mutation disrupts a PEST motif [9], which is an amino
acid sequence that is known to promote the rapid
turnover of proteins [10]. These observations suggested
that MEI-1 might be actively targeted for degradation
after meiosis.
The first clue that the ubiquitination pathway (Box 1)
might be involved in the degradation of MEI-1 and MEI-2
came with the discovery of rfl-1, another locus required for
MEI-1 degradation [11]. RFL-1 is the homolog of yeast
Uba3, a member of the neddylation pathway that
activates cullins by modifying them with the ubiquitin-
like protein Nedd8 (for review, see Ref. [12]). RNA
interference (RNAi) screening of the five C. elegans cullins
identified CUL-3 as the one that is crucial for MEI-1
degradation [11]. These data suggested that MEI-1 and
MEI-2 might be targeted by an E3-ubiquitin-ligase-
containing CUL-3 for ubiquitination and degradation.
Recent work focusing on MEL-26 has confirmed this
hypothesis.
MEL-26 is a Broad-Complex, Tramtrack and Bric-a-
brac (BTB)-domain protein that binds to CUL-3
in vitro and in vivo [13–15]. Remarkably, MEL-26
also binds to MEI-1 [13,14]. Mutations in the PEST
sequence of MEI-1 that stabilize MEI-1 during mitosis
block the MEI-1–MEL-26 interaction. A CUL-3–MEL-26
complex that is immunoprecipitated from 293T cells
promotes MEI-1 ubiquitination in vitro [15]. Together,
these data suggest that MEL-26 links MEI-1 to CUL-3,
thus stimulating the ubiquitination of MEI-1. In other
E3 ligases that have been described to date, substrates
that are to be ubiquitinated are linked to a cullin by
two intermediate proteins: a substrate-recruitment
factor that binds to the substrate and a BTB-domain
protein (e.g. SKP-1 and Elongin C) that binds to the
substrate-recruitment factor and the cullin (Box 1).
MEL-26 is the first BTB protein that has been shown
to merge the properties of the cullin-binding protein
and the substrate-recruitment factor into a single
polypeptide but it is likely that others will follow.
CUL-3 binds to several BTB proteins in C. elegans
and Schizosaccharomyces pombe [14,16]. In C. elegans,
cul-3-mutant zygotes have more-complex phenotypes
than mel-26 zygotes have [5,11,17], which is consistent
with CUL-3 functioning with other BTB partners to
degrade substrates other than MEI-1.
Degradation of other PEST proteins
The other CUL-3 substrates are not yet known but
another example of a PEST protein that is degraded in
C. elegans zygotes is OMA-1. OMA-1 is a putative RNA-
binding protein that functions in oocyte maturation
[18,19]. The levels of OMA-1 protein are highest in
maturing oocytes and they decrease rapidly after fertiliza-
tion. Like MEI-1, OMA-1 contains a PEST motif,
mutations in which stabilize OMA-1 in embryos [20].
The PEST mutation does not interfere with the function of
OMA-1 during oocyte maturation but it disrupts cell-fate
specification in embryos, which leads to embryonic
lethality. How stabilized OMA-1 interferes with cell fates
is not yet known but it seems to be linked to the ability
of OMA-1 to stabilize other maternal proteins. The E3 that
is required for OMA-1 degradation has not yet been
reported but it seems to be distinct from the E3 that
targets MEI-1 [21].
An example in another organism of a PEST-domain
protein that is degraded at the egg-to-embryo transition
is cytoplasmic polyadenylation-element-binding protein
(CPEB). CPEB is an RNA-binding protein, best charac-
terized in Xenopus, that regulates the translation of
oocyte RNAs during maturation (for review, see
Ref. [22]). CPEB levels remain high midway through
oocyte maturation and drop suddenly before the first
meiotic division [23–25]. Inappropriate stabilization of
CPEB interferes with meiosis, which indicates that
CPEB degradation is required for the completion of
egg maturation [26]. Several lines of evidence implicate
(a)(b)
(c)(d)
Figure 1. MEI-1 degradation at the meiosis-to-mitosis transition. (a–b) Caenorhab-
ditis elegans zygotes stained for microtubules (green) and DNA (blue). The meiotic
spindle is small and located close to the cortex (a). By contrast, the mitotic spindle
(b) is much larger and is located more centrally. (c–d) MEI-1 (red) accumulates on
the meiotic spindle and throughout the cytoplasm during meiosis (c), and has been
mostly degraded by the onset of mitosis (d). Images courtesy of Thimo Kurz and
Bruce Bowerman.
Review
TRENDS in Cell BiologyVol.14 No.8 August 2004421
www.sciencedirect.com

Page 3

the ubiquitin–proteasome system in this process. CPEB
is ubiquitinated when it is incubated in egg extracts [26]
and CPEB degradation can be inhibited either by
deleting the PEST domain or by exposing oocytes to
proteasome inhibitors [25]. CPEB degradation is activated
upon phosphorylation by CDC2 kinase [26,27] but the E3
involved has not yet been reported.
Transition from the ‘germline-only’ egg to the
‘soma-and-germline’ embryo: localized degradation of
germ-plasm proteins
The destruction of CPEB occurs primarily in the vegetal
pole of the Xenopus egg. In embryos, CPEB persists in the
animal pole, in which it is required for cell division [28].
There is another class of maternal protein that continues
to function in embryos, albeit only in certain cells. These
are the proteins that reside in the germ plasm, a
specialized cytoplasm essential for the formation of the
embryonic germline. Germ-plasm proteins are not
required in somatic cells but they must be maintained in
the embryonic germline. Recent evidence from C. elegans,
Drosophila and zebrafish suggests that germ-plasm
proteins are actively degraded outside of the germline
and are protected from degradation in germ cells.
C. elegans germ-plasm proteins
Among the germ-plasm proteins in C. elegans, the
behavior of three CCCH-finger proteins – PIE-1, POS-1
and MEX-1 – is understood best. These putative RNA-
binding proteins are restricted to the germline by at least
two mechanisms (Figures 2,3). First, at each asymmetric
division that gives rise to a somatic blastomere and a
germline blastomere, PIE-1, POS-1 and MEX-1 are segre-
gated preferentially to the germline daughter [29–31].
Second, they are actively degraded in somatic blastomeres
[32,33].Thisdegradationdependsonzinc-finger-interacting
factor(ZIF)-1,whichisaproteinthatbindstoCCCHfingers
[33].Twolines ofevidencesuggestthatZIF-1isasubstrate-
recruitment factor for a multisubunit E3 ubiquitin ligase.
First, ZIF-1 contains a suppressor of cytokine signaling
(SOCS) box, which is a motif that is typical of substrate-
recruitment factors that complex with Elongin C (the BTB
partner for CUL-2) [34]. ZIF-1 binds to Elongin C in
yeast two-hybrid assays and this interaction requires
the SOCS box [33]. Second, RNAi knockdowns of ZIF-1,
Elongin C, CUL-2 or the E2 ubiquitin-conjugating
enzyme UBC5 (also known as LET-70) stabilize
CCCH proteins in somatic cells. These observations
are consistent with ZIF-1 functioning in the context of
an E3 ubiquitin ligase that targets CCCH proteins for
degradation in somatic cells (Box 1). RNAi knockdowns
of Elongin C, CUL-2 and UBC5 cause cell-cycle defects
[33,35], in addition to blocking CCCH-finger degra-
dation, which suggests that these core E3 components
alsoregulatethedegradationofotherproteins.Bycontrast,
zif-1(RNAi) embryos seem to have normal early divisions,
which is consistent with the specificity of ZIF-1 for CCCH
proteins [33]. zif-1(RNAi) embryos die during late embry-
ogenesis,possiblybecausetheperduranceofCCCHproteins
interferes with somatic development.
Whereas the CCCH proteins are unstable in somatic
cells, they are stable in germline blastomeres. How is
degradation restricted to somatic cells? This question is
particularly perplexing because CUL-2 is present in all
Box 1. Ubiquitin and protein degradation
The prevalent mode of regulated protein degradation in eukaryotic cells
is ubiquitination (the covalent attachment of a polyubiquitin chain to a
protein), followed by degradation of the protein by the proteasome (for
review, see Ref. [60]). Ubiquitination is initiated by a ubiquitin-activating
enzyme (E1) that activates ubiquitin for nucleophilic attack. Activated
ubiquitin is passed on to a ubiquitin-conjugating enzyme (E2) that, in
conjunction with a ubiquitin-ligating protein (E3), transfers the ubiquitin
tothesubstratethatistobedegraded.Mostorganismshaveonlyasmall
number of highly conserved E1 and E2 enzymes that participate in all
ubiquitination reactions.By contrast, E3 ligases are numerousand often
multimeric, which reflects their unique role in target recognition.
Typically, multimeric E3 ligases are composed of a cullin-containing
core (e.g. CUL-1, 2 or 3) and a variable substrate-recognition factor. The
latter provides substrate specificity to the ubiquitination reaction by
binding to one or a small number of targets. For example, some targets
arerecruitedbyafamilyofsubstrate-recognitionfactorsknownasF-box
proteins (Figure Ia). F-box proteins link the target to the CUL-1–RBX-1
core of an E3 ligase through the SKP-1 bridging protein. Similarly, a
second group of substrate-recognition factors, suppressor of cytokine
signaling(SOCS)-boxproteins,links targets toaCUL-2–RBX-1coreofan
E3 ligase through an Elongin-B–Elongin-C bridging complex (Figure Ib).
Athirdtypeofcomplexhasbeendescribedrecently[13–16],inwhichthe
substrate-recruitment factor and the bridging protein are merged into
one protein (Figure Ic). In this complex, targets are recruited to CUL-3
directly by a Broad-Complex, Tramtrack and Bric-a-brac (BTB)-domain
protein such as MEL-26.
RBX-1
RBX-1
TRENDS in Cell Biology
CUL-3
MEL-26
Ub
(BTB)
MEI-1
(PEST)
RBX-1
CUL-1
Ub
Target
F box
SKP-1
(BTB)
CUL-2
Ub
PIE-1
(CCCH)
ZIF-1
(SOCS)
El
C
(BTB)
E2
E2
El
B
E2
(a)(b)(c)
Figure I. Examples of multisubunit E3 ubiquitin ligases. Abbreviations: BTB, Broad-Complex, Tramtrack and Bric-a-brac; SOCS, suppressor of cytokine signaling; Ub,
ubiquitin; ZIF-1, zinc-finger-interacting factor 1.
Review
TRENDS in Cell BiologyVol.14 No.8 August 2004422
www.sciencedirect.com

Page 4

blastomeres and the same is probably true for the other
core E3-ligase components, all of which seem to have
essential cell-cycle functions [33,35]. The localization of
endogenous ZIF-1 has not yet been reported. Like other
substrate-recruitment factors, ZIF-1 seems to be
unstable: a green fluorescent protein (GFP)–ZIF-1 fusion
is visible only in embryos with depleted levels of CUL-2
and, under these conditions, it is ubiquitous [33].
What restricts the activity of the ZIF-1–Elongin-C–
CUL-2 ligase to somatic cells? The answer seems to be two
CCCH-finger proteins: MEX-5 and MEX-6. Unlike the
other CCCH proteins, MEX-5 and MEX-6 segregate
preferentially to the somatic daughters at each asym-
metric division [36] (Figures 2,3). Genetic experiments
suggest that MEX-5 and MEX-6 function redundantly
[36] and that, together, they are required to activate
ZIF-1-dependent degradation in somatic cells [33]. In
wild-type embryos, MEX-5 and MEX-6 levels are high
in somatic blastomeres [36]. The levels are kept low in
germline blastomeres by PAR-1, a serine/threonine
kinase enriched in the germ lineage and required for
overall anterior–posterior polarity [37]. In embryos that
lack PAR-1, MEX-5 and MEX-6 levels are high in all cells
[36] and ZIF-1-dependent degradation is activated in all
cells [33] (Figure 2). By contrast, the RNAi knockdown of
MEX-5 and MEX-6 stabilizes CCCH-finger proteins in all
cells in both wild-type and par-1-mutant embryos. These
results suggest that MEX-5 and MEX-6 activate ZIF-1-
dependent degradation and that PAR-1 opposes this
activity, perhaps by maintaining low levels of MEX-5
and MEX-6 in germline cells.
How do MEX-5, MEX-6 and PAR-1 function? These
proteins regulate many anterior–posterior differences in
the embryo [36,37], which raises the possibility that they
affect ZIF-1-dependent degradation indirectly. Like other
CCCH-finger proteins, however, MEX-5 and MEX-6 bind
to ZIF-1 and are subject to ZIF-1-dependent degradation
themselves [33]. It has been suggested that high levels of
substrate proteins stabilize substrate-recruitment factors
in other systems [38]. This raises the possibility that
MEX-5 and MEX-6 activate ZIF-1-dependent degradation
by stabilizing ZIF-1 specifically in somatic blastomeres.
It will be important to determine whether the binding of
MEX-5 and MEX-6 to ZIF-1 is essential for activating
ZIF-1 and, furthermore, how PAR-1 interferes with
this process.
Drosophila germ-plasm proteins
Studies using Drosophila support the view that germ-
line proteins are unstable outside of the germline and
are stabilized in germ cells by PAR-1. In Drosophila,
RNAs that encode proteins that are important for
germline development (e.g. oskar) are transported on a
microtubule network to the posterior pole of the oocyte
during oogenesis [39]. The PAR-1 kinase also localizes
to the posterior pole and is required to polarize the
oocyte microtubule network [40,41]. Using a combin-
ation of hypomorphic par-1 alleles with sufficient
activity to polarize the oocyte, Reichman et al.
determined that PAR-1 is also required to stabilize
the germ-plasm protein Oskar [42]. Oocytes with
reduced PAR-1 activity localize oskar RNA as in wild-
type oocytes but they accumulate lower levels of Oskar
protein in the posterior. Conversely, the overexpression
of PAR-1 kinase increases Oskar levels and Oskar phos-
phorylation, which suggests a direct interaction between
these proteins. Consistent with this hypothesis, purified
PAR-1 phosphorylates Oskar protein in vitro and this
increases Oskar stability in oocyte extracts. The activity
that is responsible for Oskar degradation is not yet known
but it is likely to involve the ubiquitin pathway. When
incubated in Xenopus extracts, Oskar is ubiquitinated and
degraded in a proteasome-dependent manner, which
suggests that Oskar might be targeted by an E3 ubiquitin
ligase in the same way as described for CCCH proteins in
C. elegans.
The analysis of Vasa in Drosophila further impli-
cates the ubiquitin pathway in the localization of
germ-plasm proteins. Vasa is another essential regu-
lator of germ-cell fate that is recruited to the posterior
pole of the oocyte by Oskar [43,44]. Vasa accumulation
in the posterior requires not only Oskar but also the
TRENDS in Cell Biology
Zygote
AB
EMS
C
Soma
Germline
P1
P2
P3
(a)
(b)
MEX-5 and MEX-6
PIE-1
PAR-1
ZIF-1–CUL-2-
dependent
degradation

?
Somatic
blastomeres
Germline
blastomeres
Figure 2. Soma-specific degradation of CCCH-finger proteins in Caenorhabditis
elegans embryos. (a) Early C. elegans lineage. A series of asymmetric divisions
gives rise to somatic (AB, EMS, C, D) and germline (P1–P4) blastomeres. At each
asymmetric division, MEX-5 and MEX-6 (blue) are segregated preferentially to the
somatic daughter, whereas PIE-1, MEX-1, and POS-1 (red) are segregated
preferentially to the germline daughter. After division, all CCCH-finger proteins
persist in the germline blastomeres but are degraded (broken lines) in somatic
blastomeres. The kinase PAR-1 (green) segregates specifically with the germline
blastomeres. (b) High levels of MEX-5 and MEX-6 activate zinc-finger-interacting
factor (ZIF)-1-dependent degradation in somatic blastomeres. PAR-1 prevents this
degradation in germline blastomeres by restricting MEX-5 and MEX-6 to the
anterior. Experiments in Drosophila suggest that PAR-1 might also have a direct
protective function in the germline.
Review
TRENDS in Cell BiologyVol.14 No.8 August 2004423
www.sciencedirect.com

Page 5

deubiquitinating enzyme Fat facets [45] and the
SOCS-box protein Gustavus [46]. Several lines of
evidence support a direct role for Fat facets in
promoting Vasa stability in the posterior pole. Vasa
and Fat facets interact physically in vivo, as demon-
strated by co-immunoprecipitation [45]. In fat facets
mutants, Vasa is polyubiquitinated and accumulates at
lower levels in the posterior compared with wild type,
resulting in fewer germ cells during embryogenesis.
Together, these observations suggest that Fat facets
stabilizes Vasa in the posterior by reversing the
ubiquitination of Vasa and that this activity is
essential for accumulating enough Vasa for efficient
germ-cell formation.
The role of the SOCS-box protein Gustavus is less
clear. As described for ZIF-1, SOCS-box proteins
typically function as substrate-recruitment factors for
E3 ubiquitin ligases [34] – they bind to specific targets
and stimulate their ubiquitination and degradation.
However, like Fat facets, Gustavus is required for Vasa
accumulation in the oocyte posterior pole [46], which is
not easily reconciled with a potential role in stimulat-
ing Vasa turnover. Gustavus binds to a domain in Vasa
that is essential for Vasa localization in vivo, which
suggests that the Gustavus–Vasa interaction is,
instead, essential for Vasa translocation to the pos-
terior. It will be important to determine whether
Gustavus functions in the context of an E3 ligase or
in another capacity.
Zebrafish Vasa
Evidence from studies in zebrafish also supports the
view that Vasa localization depends, at least in part,
on localized protein degradation. vasa RNA localizes to
four discrete regions on cleavage furrows in the early
embryo [47–49]. The localization of vasa RNA depends
on sequences in the 30-untranslated region (UTR) [50].
When these sequences are removed, vasa RNA is found
throughout the embryo but, surprisingly, Vasa protein
is detected only in primordial germ cells. Conversely,
when GFP-coding sequences are fused to the vasa
30-UTR, GFP is found throughout the embryo, even
though the RNA is maintained only in germ cells.
These observations suggest that Vasa protein is
actively targeted for degradation outside of the germ-
line. Although the ubiquitin–proteasome pathway has
not yet been implicated in this process, the parallels
with C. elegans and Drosophila suggest that zebrafish
Vasa is under similar regulation.
Coordinated protein degradation?
The translational recruitment of maternal RNAs is
regulated in a coordinated manner. CPEB co-regulates
the translation of many RNAs during oocyte maturation
(for review, see Ref. [22]). Is maternal-protein degradation
also under the coordinated control of a single gene? Recent
evidence from C. elegans supports this possibility and
points to the dual specificity and Yak1-regulated kinase
(DYRK) kinase MBK-2 as being a candidate master
WT WT
GFP–MEX-5 GFP–PIE-1
zif-1(RNAi)zif-1(RNAi)
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
(l)
(m)
(n)
(o)
(p)
Figure 3. Zinc-finger-interacting factor (ZIF)-1-dependent degradation of PIE-1 and MEX-5 in somatic cells. Images from time-lapse recordings of live wild-type (WT) embryos
(a–d, i–l) and embryos in which ZIF-1 has been depleted by RNA interference (RNAi) (e–h, m–p). At each asymmetric division, green fluorescent protein (GFP)–MEX-5 (a–h)
segregates with the somatic daughter (left-hand side of each image), whereas GFP–PIE-1 (i–p) segregates with the germline daughter (right-hand side of each image). This
segregation is not absolute and small amounts of protein remain in the other daughter. GFP–PIE-1 and GFP–MEX-5 are both turned over in somatic cells but, possibly, at
different rates [compare (k) with (c)]. Eventually, MEX-5 is eliminated completely from somatic cells (at time points beyond those shown). In zif-1(RNAi) embryos, asymmetric
segregation occurs as in wild-type embryos [compare (b) with (f), and (j) with (n)] but degradation in somatic daughters fails [compare (d) with (h), and (l) with (p)].
Review
TRENDS in Cell Biology Vol.14 No.8 August 2004424
www.sciencedirect.com