Nonproteolytic ubiquitylation regulates the APC/Cinhibitory function of XErp1 Dissertation zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften (Dr. rer. nat.) vorgelegt von Eva Beate Hörmanseder an der MathematischNaturwissenschaftliche Sektion Fachbereich Biologie Tag der mündlichen Prüfung: 16. Dezember 2011 1. Referent: Prof. Dr. Thomas U. Mayer 2. Referent: Prof. Dr. Martin Scheffner 3. Referent: Prof. Dr. Olaf Stemmann
Non-proteolytic ubiquitylation regulates the APC/C-inhibitory
function of XErp1function of XErp1
Doktors der Naturwissenschaften (Dr. rer.
nat.)
vorgelegt von
2. Referent: Prof. Dr. Martin
Scheffner
3. Referent: Prof. Dr. Olaf Stemmann
1.2. The APC/C counteracts the
activity of Cdk1 7
1.3. The “wait anaphase signal”: The
SAC inhibits the APC/C in
mitosis 9
1.4. Regulation of APC/CCdc20 activity
in meiosis 11
1.5. The postulation of MPF and
CSF 12
1.6. The discovery of Mos as a
CSF component 13
1.7. Identification of the CSF
component XErp1 14
1.8. XErp1 inactivation upon CSF
release 15
1.9. The molecular mechanism of
XErp1 mediated APC/C inhibition 16
1.10. Feedback loops controlling XErp1
activity during CSF arrest 18
1.11. Aim of this project 20
2. RESULTS 21
2.1. UbcX can suppress SAC activity
in Xenopus egg extract 21
2.2. UbcX can suppress CSF activity
in Xenopus egg extract 22
2.3. Elevated UbcX activity prevents
meiosis I - meiosis II
transition in
Xenopus oocytes 24
activity 25
2.5. Does USP44 counteract UbcX to
maintain CSF arrest? 26
2.6. An eight-fold increase in UbcX
activity is required for CSF
release. 27
2.7. UbcX levels increase during
oocyte maturation and remain constant
during CSF release and embryonic cell
cycles 28
2.8. UbcX dependent CSF release can
be suppressed by XErp1 29
2.9. UbcX mediated ubiquitylation disrupts
the APC/C - XErp1 complex 30
2
2.10. XErp1 is the main target
of UbcX mediated ubiquitylation in
CSF
extract 32
2.11. Ubiquitylation of XErp1 is
dependent on the APC/C and
independent
of SCFβ TRCP 33
2.12. Dissociation of XErp1 upon
Cdk1 phosphorylation does not require
ubiquitylation 35
2.13. Cdc20 degradation is not
required for CSF arrest maintenance
36
3. DISCUSSION 38
3.1. Regulation of spindle checkpoint
signaling by UbcH10/UbcX 39
3.1.1. The spindle assembly checkpoint
can be inactivated by UbcX in
Xenopus egg extract 39
3.1.2. Is an APC/C inhibitor
targeted for ubiquitylation during
SAC
signaling? 41
3.2. UbcX mediated ubiquitylation of
XErp1 regulates its APC/C inhibitory
activity 43
3.2.1. Cdc20 is not destabilized in
CSF arrested egg extract 43
3.2.2. UbcX mediated ubiquitylation of
XErp1 regulates its APC/C inhibitory
activity 44
3.2.3. Are ubiquitin hydrolases
counteracting the activity of UbcX
during CSF
arrest? 46
3.3. Is the regulation of UbcX
activity important during the meiotic
cell
cycle? 48
3.3.2. Could UbcX participate in the
inactivation of XErp1 upon
fertilization? 48
pathways regulating the activity of
XErp1 49
3.4. Could ubiquitylation of XErp1
be required for its APC/C
inhibitory
activity? 50
5.2. Plasmids 55
5.2.3. Cloning and Mutagenesis 57
5.3. Proteins 57
5.3.2. His-tagged protein expression in
SF9 cells 58
5.3.3. His-tagged protein purification
from bacteria and SF9 cells 58
5.3.4. Coupled in vitro
transcription/translation (IVT) 59
5.4. Antibodies 59
5.4.2. Affinity purification of antibodies
59
5.5. Gel electrophoresis and immunoblot
analysis 60
5.6. Xenopus egg extracts 61
5.6.1. Xenopus CSF egg extract
preparation 61
5.6.2. Extract manipulations 62
5.7. Xenopus oocyte injections 64
6. LITERATURE 65
7. APPENDIX 75
7.1. Summary 75
7.2. Zusammenfassung 75
7.3. Acknowledgements 76
1. INTRODUCTION
Most eukaryotes reproduce sexually,
where cells from two parents
fuse to
generate a single cell, the zygote,
which develops into a new
organism (Figure
1.1.). Since the combination of two
diploid cells would lead to the
duplication
of the chromosomal content at every
generation, sexual reproduction depends
on a process called meiosis.
Figure 1.1. The life cycle of
vertebrates. Cells in vertebrates
proliferate mitotically in the
diploid phase to form a
multicellular organism. Sexual reproduction
begins with meiosis to generate
haploid cells, which fuse upon
fertilization to form a new
organism.
1.1. Meiosis and meiotic maturation
Meiosis is a specialized form of
nuclear division that leads to
the generation of
cells containing half the normal
complement of chromosomes from diploid
oocytes (Figure 1.2. a, Alberts et
al., 2002). (Alberts et al.,
2002).
Before entering the meiotic program,
oocytes are diploid like somatic
cells and
contain two copies of each
chromosome, one of them inherited
from each
parent. Meiosis begins with an
S-phase (Petronczki et al., 2003)
in which
5
chromosomes are replicated to produce
sister chromatid pairs tightly linked
by
cohesion (Klein et al., 1999).
Next, the duplicated homologues pair
to form
tetrads and undergo homologues
recombination, a process important for
generating genetic variation and to
guarantee accurate segregation of
the
homologues at the following nuclear
division. Homologous recombination
starts with the introduction of
DNA double-strand breaks (DSB) at
almost
variable positions along the chromosome
(Sun et al., 1989). In
most of the
cases, DSBs are repaired without
rendering the DNA sequence of
the two
homologs. Sometimes however, the repair
leads to the formation of
a
continuous DNA strand between two
homologous chromatids, which can lead
to a reciprocal DNA exchange or
crossover (Allers and Lichten,
2001). The
result is a strong physical linkage
between the two homologous
chromosomes
as long as the sister chromatid
arms are held together by
cohesion. As a result,
the homologous chromosomes become
bioriented on the first meiotic
spindle
and after cohesin cleavage at the
chromosome arms at anaphase I,
exactly one
of the two homologous chromosomes
is segregated into each daughter
cell
(Buonomo et al., 2000). After the
completion of meiosis I, cells
enter directly
the next division cycle without
replicating the chromosomes. In meiosis
II,
similar to mitosis, sister chromatids
are divided into the two
daughter cells by
the cleavage of centromeric cohesion
upon anaphase II onset.
Together,
meiotic divisions result in the
production of four haploid cells,
which can be
differentiated into special reproductive
cells, i.e. the egg and the
sperm.
In animals, oocytes arrest before
the first meiotic division at
prophase I, and
these immature oocytes or stage VI
oocytes can stop at this point
for decades
(Hunt, 1989). The production of
a fertilizable egg from such an
immature
oocyte involves a process called
oocyte maturation (Figure 1.2. b).
Upon
hormonal induction, immature oocytes resume
meiosis I and undergo germinal
vesicle breakdown (GVBD) which is
visible on the surface of the
oocytes by the
appearance of a white dot. Meiosis
I is completed with the
extrusion of the
first polar body after which the
oocytes proceed directly through
meiosis II
6
where the second polar body is
extruded and haploid gametes are
produced.
In vertebrates like Xenopus laevis,
oocytes complete meiotic maturation
with
an arrest at metaphase of meiosis
II, in which they await
fertilization. From the
viewpoint of cell-cycle control, the
major questions are concerning
the
mechanisms underlying the induction and
regulation of oocyte maturation as
well as the arrest of mature
oocytes at metaphase of meiosis
II and its release
upon fertilization (Tunquist and Maller,
2003).
Figure 1.2. The meiotic program. (a)
In meiosis, after DNA replication,
two divisions generate haploid
gametes. For clarity, only one
chromosome is depicted. (b) Meiosis
in vertebrates is arrested at
two stages. After DNA synthesis,
the oocytes grow to their final
size and arrest at meiotic
prophase I. Progesterone induces meiotic
maturation and the production of
an egg arrested at meiotic
metaphase II. Fertilization triggers
the completion of Meiosis II
and a diploid zygote is
formed (Adapted from Morgan,
2007).(Morgan, 2007)
1.1. Cdk1/cyclin B drives the meiotic
cell cycle
The ordered progression of the
meiotic cell cycle, like the
mitotic cell cycle, is
mediated mainly by the activity
of cyclin dependent kinases (Cdks)
and
ubiquitin ligases (Murray, 2004). Cdks
are serine-threonine kinases that
are
activated by their regulatory subunit,
the cyclins. In mitotic G1,
low Cdk1
activity is important for the
resetting of the origins of DNA
replication. Rising
Cdk activity triggers the firing
of DNA replication origins and
as S-phase
progresses and DNA replication
continues, the activity of Cdk1/CylinB1
promotes entry into mitosis, which
is characterized by nuclear
envelope
7
condensation. After the successful division
of the replicated chromosomes into
two daughter cells, the cell needs
again low Cdk1 activity to exit
mitosis and to
enter G1. Therefore, low Cdk
activity followed by high activity
links DNA
replication to progression through
mitosis (Porter, 2008) – the
basis for the
mitotic cell cycle.
In Xenopus meiosis, the hormone
progesterone induces entry into
metaphase I
by the activation and amplification
of Cdk1/cyclin B by inducing
both the
dephosphorylation of inhibitory residues
on Cdk1 and the accumulation
of
cyclin B (Tunquist and Maller,
2003). Progression from metaphase I
to
anaphase I is accompanied by a
drop in cyclin B levels
and decreasing Cdk1
activity. But unlike in mitotic
cells, cyclin B is not completely
degraded upon
anaphase onset but appears to be
reduced to half (Furuno et
al., 1994;
Iwabuchi et al., 2000). While it
remains controversial whether this
drop in
cyclin B levels is required for
meiotic progression (Peter et al.,
2001; Taieb et
al., 2001), the inhibition of
complete cyclin B degradation is
essential for the
persistence of M-phase and the
inhibition of DNA replication (Ohe
et al., 2007).
Thus, the oocytes directly enter a
second M-phase, where the
stabilization of
cyclin B levels is important for
establishing the second meiotic
arrest. Upon
fertilization, cyclin B is degraded,
Cdk1 is inactivated and the
zygotes enter
mitotic cell cycles.
1.2. The APC/C counteracts the activity
of Cdk1
Anaphase onset requires the
inactivation of both Cdk1 kinase
and the
inactivation of the anaphase inhibitory
protein securin. Securin prevents
cohesin cleavage and thus the
irreversible step of sister chromatid
separation
by keeping the cohesin directed
protease separase inactive (Uhlmann et
al.,
1999; Uhlmann et al., 2000).
Both, Cdk1/cyclin B and securin
activity is
regulated by the E3 ubiquitin
ligase anaphase promoting complex/cyclosome
(APC/C). It mediates the specific
ubiquitylation of cyclin B and
securin (Sudakin
8
et al., 1995; Zou et al., 1999)
thereby targeting them for
destruction by the 26
S proteasome at anaphase onset.
The APC/C is an unusual large
E3 ubiquitin ligase that consists
of at least 13
subunits including proteins with cullin
and RING-finger domains (Zachariae
and
Nasmyth, 1999). In addition, the
APC/C associates with coactivator
proteins
called Cdc20 and Cdh1 (Pesin and
Orr-Weaver, 2008), which bind
transiently to
the APC/C core complex and are
thought to regulate both the
activity and
substrate specificity of the APC/C.
While in somatic mitotic cell
cycles, the
coactivator of the APC/C alternates
between Cdc20 and Cdh1, the
main
coactivator required for meiosis and
early embryonic cell cycles has
been
reported to be Cdc20 (Lorca et
al., 1998). The APC/C together
with its
coactivator is responsible for substrate
recognition and thus confers
specificity
to the ubiquitylation reaction (Peters,
2006). It functions at the
last step of a
cascade of enzymes that sequentially
act to transfer ubiquitin to
the target
protein (Hershko and Ciechanover,
1998). Free ubiquitin is first
covalently
attached to an ubiquitin-activating enzyme
E1 via a thioester bond. It
is then
transferred to an ubiquitin-conjugating
enzyme E2 where it forms a
thioester
bond with the active site
cystein. The main E2 enzyme
cooperating with the
APC/C has been identified in clam
as E2-C (Hershko et al., 1994)
and orthologs
were found in Xenopus named UbcX
(Yu et al., 1996), and in
humans named
UbcH10 (Townsley et al., 1997).
In Xenopus, UbcX is essential
for APC/C
activity, since a dominant negative
mutation in the active site
cystein (C114S)
inhibits APC/C dependent substrate
ubiquitylation (Townsley et al.,
1997), and
the depletion of UbcX inhibits APC/C
substrate degradation (data not
shown).
In the final step of APC/C
dependent ubiquitylation, the E2-bound
ubiquitin is
covalently attached to a lysine
residue in the target protein. In
this reaction,
the APC/C is thought to approximate
the substrate and the E2-ubiquitin
and to
position them for efficient ubiquitin
transfer (Peters, 2006). Recently,
it has
been shown that in human cells,
UbcH10 forms an E2-enzyme module
with
Ube2S, and both enzymes were shown
to be important for the
formation of
9
ubiquitin chains on APC/C substrates,
where UbcH10 conjugates the
first
ubiquitin to the lysine residue of
the substrate and Ube2S then
elongates the
ubiquitin chain (Garnett et al.,
2009; Williamson et al., 2009;
Wu et al., 2010).
As a consequence, ubiquitylation can
target proteins to the 26 S
proteasome, a
high molecular weight protease complex
that hydrolyses its substrates
into
short peptides and thus inactivates
them irreversibly. Alternatively,
ubiquitylation can act as a
reversible posttranslational modification of
a
protein to regulate its activity
(Hershko and Ciechanover, 1998).
1.3. The “wait anaphase signal”: The
SAC inhibits the APC/C in
mitosis
Mitotically and meiotically dividing
cells depend on ubiquitin-mediated
proteolysis of key cell-cycle
regulators at the correct time
(Pesin and Orr-
Weaver, 2008). In mitosis, a
conserved mechanism called the
spindle assembly
checkpoint (SAC) guarantees an equal
segregation of the chromosomes to
the
two nascent daughter cells (Musacchio
and Salmon, 2007). The SAC is
activated
by missattached or unattached kinetochores
(Nicklas et al., 1995; Rieder
et al.,
1995; Rieder et al., 1994) and
prevents the APC/C from
ubiquitylating cyclin B
and securin. Although it is not
yet completely understood how
the SAC
inactivates the APC/C, it is well
accepted that the primary target
of the SAC is
the APC/C coactivator Cdc20 (Hwang
et al., 1998; Kim et al.,
1998) and that
SAC activity is propagated by a
number of conserved proteins
including Mad1,
Mad2 and Bub3/BubR1 (Hoyt et
al., 1991; Li and Murray, 1991).
Current
models of SAC mediated APC/C
inactivation suggest that Mad2 binds
to Cdc20
in conjunction with BubR1 and Bub3
to form the “Mitotic Checkpoint
Complex”
(MCC), which binds to the APC/C
and renders it inactive (Sudakin
et al., 2001).
Once all kinetochores are properly
attached, it has been suggested
that the
inhibitory MCC complexes have to be
actively dissociated by APC/C
dependent,
non-proteolytic ubiquitylation of Cdc20 to
turn off the SAC. Specifically,
it has
been shown that addition of the
E2 ubiquitin conjugating enzyme
UbcH10 to
SAC-arrested cell extract triggers the
APC/C-dependent multi-ubiquitylation of
10
Cdc20, and possibly other components
of the APC/C–Cdc20-MCC complex,
resulting in the release of Mad2
and BubR1 from Cdc20 (Reddy et
al., 2007). In
checkpoint arrest conditions, this
ubiquitylation reaction is antagonized
by the
activity of the ubiquitin hydrolase
USP44 (Figure 1.3.), which removes
ubiquitin
from Cdc20 (Stegmeier et al.,
2007). As soon as the last
kinetochore is
attached, ubiquitylation of Cdc20 is
thought to exceed its
deubiquitylation,
Cdc20 is freed from the MCC
and the APC/C can be
rapidly activated in a
switch-like manner.
Figure 1.3. Dynamic ubiquitylation and
deubiquitylation regulate SAC activity.
During mitotic checkpoint arrest,
ubiquitylation of Cdc20 by UbcX,
which leads to the dissociation
of the APC/C inhibitors Mad2 and
BubR1, needs to be counteracted
by USP44 dependent deubiquitylation
of Cdc20 to maintain SAC
mediated APC/C inhibition.
A different model contradicts this
view of SAC arrest and instead
suggests that
in cells with an active SAC,
Cdc20 in complex with the
MCC proteins is
ubiquitylated and targeted for
destruction, and this degradation
is important
for inactivating the APC/C (Ge et
al., 2009; Nilsson et al.,
2008). Supporting this
model, experiments in budding yeast
and human cells have shown that
Cdc20
is ubiquitylated and degraded during
SAC arrest and overexpression of
Cdc20
could overcome the SAC mediated
inhibition of the APC/C (King et
al., 2007;
Pan and Chen, 2004). Importantly, a
non-ubiquitylatable form of Cdc20
where
every lysine was mutated to an
arginine was insensitive to the
checkpoint
arrest and activated the APC/C
(Nilsson et al., 2008). These
results contradict a
model where Cdc20 ubiquitylation causes
its activation and rather support
the
latter model where ubiquitylation
inactivates Cdc20.
11
The regulation of APC/C activity
is especially important during oocyte
maturation in vertebrates where meiosis
is arrested twice to coordinate
oocyte
development with the events of
meiosis (Figure 1.4.).
In prophase I, the APC/C has to
be inactive to maintain chromosome
cohesion
(Pesin and Orr-Weaver, 2008). When
oocytes mature, the APC/C needs
to
become active at the metaphase I
- anaphase I transition to
allow the
degradation of securin and the
separation of the homologous
chromosomes
(Buonomo et al., 2000; Siomos et
al., 2001). In contrast to all
organisms tested,
the requirement of the APC/C for
meiosis I - meiosis II
transition is
controversial in Xenopus. Although
microinjections of Xenopus oocytes with
inhibitory antibodies or antisense
oligonucleotides directed against the
APC/C
coactivator Cdc20 did not disrupt
progression through meiosis I
(Peter et al.,
2001; Taieb et al., 2001), it
is possible that these approaches
did not eliminate
APC/C activity completely. Nevertheless,
the complete degradation of cyclin
B
must be prevented also in Xenopus
to maintain M-phase and to
inhibit S-phase
(Ohe et al., 2007), suggesting
that the APC/C needs to be
regulated to
contribute to this modulation of
cyclin B levels.
Figure 1.4. Oocyte maturation on a
molecular level: Cdk1 and APC/C.
The cell cycle in meiosis is
driven by the activity of
Cdk1/cyclin B which is counteracted
by the APC/C, the relative
activities of which through the
maturation process are illustrated
(adapted from Wu and Kornbluth,
2008).
At the second meiotic arrest at
metaphase II, the APC/C needs
to be inhibited
to stabilize cyclin B and
securin to prevent premature anaphase
onset and
12
parthenogenetic activation of the egg.
Upon fertilization, APC/C activation
is
required to induce the exit from
the metaphase II arrest (Lorca
et al., 1998;
Peter et al., 2001) and thereby
allowing entry into early embryonic
cell cycles.
While the spindle checkpoint is
important for the metaphase arrest
and APC/C
inhibition in mitotic cells in
the presence of unattached
kinetochores, it is
unlikely that the SAC mediates
the metaphase arrest observed in
mature
vertebrate eggs. Evidence against such
a hypothesis includes the fact
that CSF
arrest is terminated by fertilization
and the following elevation in
cytoplasmic
calcium levels, but calcium addition
does not overcome SAC arrest
(Minshull et
al., 1994). Additionally, the SAC
requires kinetochores and microtubule
depolymerization, whereas neither is
required for meiotic metaphase II
arrest
(Tunquist and Maller, 2003). What
inhibits oocytes at metaphase of
Meiosis II?
1.5. The postulation of MPF and CSF
In 1971, Yoshio Masui and Clement
L. Markert performed experiments in
Rana
pipiens oocytes and embryos that
became fundamental for the
identification of
the mechanisms mediating the metaphase
II arrest in mature oocytes
(Masui
and Markert, 1971).
Specifically, they observed that injection
of immature oocytes with endoplasm
of mature oocytes induced meiotic
maturation. Therefore they postulated
that
maturation is induced by a
maturation promoting factor (MPF) which
is
released by hormonal induction and
remains active in the mature
egg (Figure
1.5.). To analyze whether the
same activity could accelerate cell
divisions in
embryonic cells, they injected endoplasm
of the mature egg into one
cell of a
two-cell stage embryo. Surprisingly,
they found that the injected
blastomere
arrested at the next mitosis,
prompting them to propose the
existence of a
cytostatic factor (CSF) present in
the mature egg that is
responsible for
inducing the metaphase II arrest
(Figure 1.5.). Additionally, this
activity is
13
inactivated upon fertilization, since
injection of blastomeres with
endoplasm of
fertilized embryos did not cause
cell-cycle arrest.
Figure 1.5. The discovery of MPF
and CSF. Illustration of the
oocyte- and blastomere-injection assays
originally performed by Masui and
Markert in 1971 that led to
the identification of the maturation
promoting factor MPF and the
cytostatic factor CSF.
While MPF was soon identified to
be the activity of cyclin
dependent kinase
Cdk1 together with its regulatory
subunit cyclin B (Gautier et
al., 1990; Gautier
et al., 1988; Lohka et al.,
1988; Murray et al., 1989),
the discovery of the
molecular identity of the CSF took
more than three decades.
1.6. The discovery of Mos as a
CSF component
To identify the CSF activity that
mediates the metaphase II arrest,
three criteria
were proposed for a protein or
an activity to be a CSF:
(1) The activity emerges
during oocyte maturation and peaks
in the metaphase II arrested
egg. (2)
Injection of blastomeres with the
activity induces mitotic arrest
and (3)
fertilization triggers the inactivation of
the factor (Masui and Markert,
1971).
The first protein identified meeting
these criteria was the kinase
Mos. Mos is
expressed during oocyte maturation (Sagata
et al., 1988); Figure 1.6.), it
could
induce mitotic arrest when injected
into blastomeres of a dividing
embryo
14
(Sagata et al., 1989) and it
was degraded upon fertilization
(Lorca et al., 1991).
To understand the detailed molecular
mechanism linking Mos to the
metaphase II arrest, the signaling
pathway of the kinase was
investigated.
Biochemical analysis revealed that Mos
can activate the mitogen
activated
protein kinase (MAPK) pathway (Posada
et al., 1993) resulting in the
activation
of the ribosomal S6 kinase (Rsk),
and functional analysis of the
members of this
pathway showed that they are
required for CSF arrest (Abrieu
et al., 1996;
Bhatt and Ferrell, 1999; Cross
and Smythe, 1998; Gotoh and
Nishida, 1995;
Gross et al., 1999; Haccard et
al., 1993; Kosako et al.,
1994a, b). Therefore, the
Mos activated MAPK-pathway was proposed to
be a molecular component of
the CSF. Since both, the Mos-MAP
kinase pathway and APC/C
inhibition are
responsible for CSF arrest, it
seemed possible that these two
pathways are
interconnected. However, it remained
unclear how Rsk as the terminal
kinase
in this cascade was communicating
with the cell-cycle machinery to
establish
the CSF arrest.
1.7. Identification of the CSF component
XErp1
Reportedly, polo-like kinase Plx1 is
required CSF inactivation and
APC/C
activation (Descombes and Nigg, 1998).
Specifically, it has been shown
that
Xenopus egg extracts depleted of
Plx1 fail to release the CSF
arrest upon
increasing cytoplasmic calcium levels.
Therefore, a yeast two-hybrid screen
was performed to identify proteins
that interacted with Plx1
(Schmidt et al.,
2005), and this approach led
finally to the identification of
the sought after
component of CSF, the XErp1
protein. XErp1 nicely satisfied the
Masui and
Markert criteria proposed for CSF.
First, XErp1 is synthesized during
Xenopus
oocyte maturation; it starts to
be detectable at the MI-MII
transition and it
accumulates as oocytes proceed through
meiosis II where it reaches
highest
levels at metaphase II (Figure
1.6.); second, exogenous introduction
of XErp1
into one blastomere of a two-cell
stage embryo promoted a cell-cycle
arrest
and third, XErp1 was degraded after
fertilization in a Plx1 dependent
manner.
15
Importantly, XErp1 is essential for
CSF arrest as Xenopus egg
extracts arrested
at metaphase II depleted of
XErp1 were unable to maintain
CSF arrest and
entered interphase.
Further characterization XErp1 revealed the
C-terminus of the protein, which
is
sufficient for CSF arrest maintenance,
shares high sequence similarity with
the
mitotic APC/C inhibitor Emi1 and
like Emi1, XErp1 was shown to
inhibit the
APC/C directly (Schmidt et al.,
2005). Therefore, XErp1 is a
CSF specific APC/C
inhibitor.
Figure 1.6. Oocyte maturation and CSF
on a molecular level. Oocyte
maturation is driven by the
activities of Cdk1/cyclin B, the
APC/C and CSF factors Mos and
XErp1, ad the relative
activities during oocyte maturation are
depicted on the left (adapted
from Kornbluth, 2008).
Since XErp1 was shown to be a
substrate of Rsk, the Mos-MAPK
pathway could
finally be linked to the
regulation of the APC/C. Rsk
phosphorylation was
shown to increase the inhibitory
activity of XErp1 in CSF
arrested eggs, which
will be described later.
1.8. XErp1 inactivation upon CSF release
As proposed by Masui and Markert,
fertilization causes the inactivation
of CSF.
The first response of an egg
to fertilization is an elevation
in cytoplasmic
calcium levels, which results in the
activation of calcium/calmodulin dependent
kinase II (CaMKII;(Lorca et al.,
1993). The identification of XErp1
as a CaMKII
16
substrate provided insights into how
fertilization is connected with
CSF
inactivation (Figure 1.7.;(Hansen et al.,
2006; Liu and Maller, 2005;
Rauh et al.,
2005).
Figure 1.7. Fertilization mediated CSF
inactivation. Fertilization (1) triggers
the activation of CaMKII (2)
which phosphorylates XErp1 (3)
creating a docking site for Plx1
(4). Plx1 in turn phosphorylates
XErp1 creating a phosphodegron
(5), which is recognized by the
ubiquitin ligase SCFβ
TRCP. XErp1 ubiquitylation targets it
for degradation (6) and thus
CSF inactivation, the APC/C becomes
active (7) and cells complete
meiosis II (adapted from Rauh
et al., 2005).
CaMKII mediated phosphorylation of XErp1
provides a docking site for
Plx1 on
XErp1. Through Plx1 mediated
phosphorylation of XErp1 a
phosphodegron is
created and XErp1 is recognized by
the SCFβ TRCP complex, an ubiquitin
E3 ligase
that ubiquitylates and targets XErp1
for degradation. Consequently, calcium
triggers CSF inactivation resulting in
APC/C activation and the fertilized
egg can
proceed with embryonic cell divisions.
1.9. The molecular mechanism of XErp1
mediated APC/C inhibition
In CSF arrested eggs, XErp1
maintains the metaphase II arrest
by directly
inhibiting the APC/C. The binding
of XErp1 to the APC/C is
essential for its
inhibitory activity as mutants
defective in APC/C binding are
inefficient in
17
inhibiting the APC/C (Wu et al.,
2007b). The well-conserved C-terminal
peptide
sequence of XErp1, termed the RL
tail, was reported to mediate
the
recruitment of XErp1 by serving
as a docking site to the
APC/C (Ohe et al.,
2010). Binding to the APC/C allows
and enhances the inhibitory
interactions of
two other sequence elements of XErp1,
the D-box and the ZBR-domain.
While
it is well established that all
three elements are critical for
APC/C inhibition,
the specific contribution of the
D-box and the ZBR domain to
the inhibition of
the APC/C by XErp1 remain elusive
(Nishiyama et al., 2007; Ohe
et al., 2010;
Tang et al., 2010).
Notably, all three elements are
conserved between XErp1 and Emi1,
a somatic
paralog of XErp1, whose APC/C
inhibitory activity is required to
prevent DNA
re-replication (Di Fiore and Pines,
2007; Machida and Dutta, 2007)
suggesting
that XErp1 and Emi1 share the
same mode of APC/C inhibition.
Emi1, when
bound to the APC/C together with
the E2 enzyme UbcH10, was
shown to
inhibit the correct engagement of the
substrate to the APC/C thereby
reducing
substrate ubiquitylation (Summers et
al., 2008). Further studies on
Emi1
suggested that it acts as an
APC/C pseudosubstrate and the
D-box mediates
APC/C binding, while its ZBR
mediates APC/C inhibition (Miller
et al., 2006).
Consistently, it has been shown that
Emi1 mutated in its ZBR does
not inhibit
the APC/C but rather is quickly
targeted for destruction by the
APC/C. Given
that XErp1 – like Emi1 –
contains a D-box and ZBR, it
is tempting to speculate
that XErp1 acts as well as
a pseudosubstrate. However, previous
studies
suggest that XErp1 does not
compete with substrates for APC/C
binding but
rather interferes with the transfer
of ubiquitin to substrate proteins
bound to
the APC/C (Tang et al., 2010).
Furthermore, our preliminary
experiments
revealed that in contrast to
Emi1, mutation of the ZBR of
XErp1 does not
convert it into an APC/C substrate
corroborating the idea that XErp1
inhibits
the APC/C by a mechanism distinct
to the one of Emi1.
18
Together, although it is established
that XErp1 needs to be
recruited to the
APC/C to exert its inhibitory
function, the exact molecular
mechanism of XErp1
mediated APC/C inhibition remains elusive.
1.10. Feedback loops controlling XErp1
activity during CSF arrest
During metaphase II arrest, the
Mos-MAPK pathway was shown to
activate
XErp1 by upregulating both the
stability and activity of XErp1
(Isoda et al.,
2011; Wu et al., 2007a; Wu et
al., 2007b). The Mos-MAPK pathway
activates
the kinase Rsk (Bhatt and
Ferrell, 1999; Gross et al.,
1999), which
phosphorylates XErp1 at residues in
the central region (Inoue et
al., 2007;
Nishiyama et al., 2007) leading to
the recruitment of the protein
phosphatase
PP2A containing the regulatory subunit
B56β or B56ε to XErp1 (Wu
et al.,
2007a). PP2A- B56β,ε antagonizes
N-terminal and C-terminal inhibitory
phosphorylations of XErp1 by Cdk1
(Isoda et al., 2011). Cdk1
phosphorylations
destabilize XErp1 and decrease its
affinity for the APC/C (Wu et
al., 2007a; Wu
et al., 2007b).
Figure 1.8. Oocyte maturation and CSF
on a molecular level. Oocyte
maturation is driven by the
activities of Cdk1/cyclin B, the
APC/C and CSF factors Mos and
XErp1, ad the relative
activities during oocyte maturation are
depicted on the left (adapted
from Kornbluth, 2008). On the
right, a simplified signaling
network controlling the activity of
XErp1 is illustrated (adapted
from Isoda et al., 2011).
19
Specifically, it has been shown that
multiple N-terminal Cdk1 phosphorylation
motifs bind cyclin B1-Cdk1 itself
as well as Plk1 and CK1 δ/ε
to inhibit XErp1
(Isoda et al., 2011). While Plk1
phosphorylation was shown to
partially
destabilize XErp1, Cdk1 and CK1δ/ε
phosphorylations are thought to
cooperatively inhibit XErp1 binding to
the APC/C (Figure 1.8.). Since
Cdk1 levels
are high during the Metaphase II
arrest, constant phosphorylation of
XErp1
would lead to gradual XErp1
inactivation and CSF release. By
recruiting PP2A-
B56β,ε to counteract the inhibitory
phosphorylations, the Mos MAPK-
pathway
keeps XErp1 active and therefore
maintains CSF arrest (Figure 1.8.).
At the
same time, this mechanism allows to
maintain Cdk1 activity at the
correct level
during CSF arrest (Figure 1.9.(Wu
and Kornbluth, 2008; Wu et
al., 2007b).
Continuous cyclin B synthesis during
CSF arrest leads to a temporal
increase in
Cdk1/cyclin B activity, which in turn
leads to an increase in the
phosphorylation
of XErp1, since the activity of
PP2A on XErp1 remain equal.
XErp1
phosphorylated by Cdk1 dissociates from
the APC/C leading to a
transient
APC/C activation and slow degradation
of cyclin B.
Figure 1.9. Cdk1/cyclin B2 and
PP2A regulate XErp1’s association with
the APC/C. Phosphorylation of
XErp1 by Cdk1/cyclin B2 leads
to the dissociation of XErp1
from the APC/C, which is
counteracted by PP2A, which
dephosphorylates XErp1 and promotes
XErp1 association with the APC/C.
Therefore, the continuous synthesis of
cyclin B induces a slow
degradation of
cyclin B during CSF arrest.
Otherwise, continuous synthesis would
create an
amount of cyclin B that cannot
be degraded by the APC/C
anymore in a short
time. This would result in a
slow and gradual rather than a
switch-like exit from
CSF arrest as observed upon
fertilization.
20
1.11. Aim of this project
XErp1 is an APC/C inhibitor
operating in CSF arrested oocytes.
However, the
exact molecular mechanism of APC/C
inhibition and its regulation is
unknown.
The D-box and the RL-tail of
XErp1 mediate the binding of
XErp1 to the APC/C,
most likely to position the ZBR
of XErp1 correctly to inactivate
the APC/C.
However, the interaction with the
APC/C needs to be dynamic to
allow slow
cyclin B degradation during CSF
arrest. Phosphorylation and
dephosphorylation
of XErp1 can regulate its
association with the APC/C, and
the Mos-MAPK
pathway was shown to promote XErp1
association. Intrigued by the
findings on
APC/C regulation by the spindle
checkpoint, we would like to
understand if a
dynamic balance of
ubiquitylation/deubiquitylation of Cdc20,
XErp1 and/or
other components of the APC/C is
also required for CSF arrest.
In addition, we
would like to test whether Cdc20
turnover is required for CSF
arrest and if
XErp1 regulates this potential
turnover. Thus, these studies will
provide a
deeper understanding of how the
XErp1-APC/CCdc20 interaction is regulated
and
2. RESULTS
In this study, we show that
non-proteolytic ubiquitylation of XErp1
regulates its
APC/C inhibitory function during CSF
arrest in Xenopus egg extracts.
This
section describes the experiments
demonstrating that ectopic UbcX, the
E2
enzyme of the APC/C, induces
release from SAC- and CSF arrest.
The release
from CSF arrest is APC/CCdc20
dependent and in the presence
of elevated UbcX
activity, XErp1 is ubiquitylated resulting
in the dissociation of XErp1
from the
APC/C. Hence, the APC/C inhibitory
activity of XErp1 in CSF
arrest can be
modulated in an UbcX-dependent manner.
Furthermore, evidence is provided
that in contrast to SAC arrested
somatic cells, Cdc20 is not
degraded during
meiotic CSF arrest suggesting that
CSF arrest is not mediated
by the
destabilization of Cdc20.
2.1. UbcX can suppress SAC activity
in Xenopus egg extract
The finding that in human
somatic cells, the APC/C can
liberate itself from
inhibition by the SAC (Reddy et
al., 2007) prompted us to
analyze whether a
similar mechanism operates in Xenopus
eggs or egg extracts to
regulate APC/C
activity during SAC and - more
interestingly - during CSF
arrest. In Xenopus
eggs, SAC activity was reported to
be absent but can be induced
by increasing
the ration of nucleus to cytoplasm
in the presence of spindle
poisons (Minshull
et al., 1994). To analyze the
effect of UbcX on SAC arrest
in Xenopus eggs, we
prepared CSF arrested egg extract and
triggered SAC arrest by the
microtubule
poison nocodazole in the presence
of high concentrations of sperm
nuclei
(Figure 2.1. a). Under these
conditions, calcium addition did not
result in APC/C
activation as in vitro translated
35S-securin remained stable (Figure
2.1. b,
panel 1). Westernblot (WB) analysis
revealed that XErp1 was
efficiently
22
inhibition was due to SAC- but
not CSF-activity. Addition of
recombinant wild
type UbcX (UbcXwt) to SAC arrested
extracts caused APC/C activation and
35S-
securin degradation (Figure 2.1. b,
panel 2). This effect was
dependent on the
catalytic activity of UbcX, as the
addition of a catalytic inactive
form of UbcX
(UbcXci) had no effect on
35S-securin stability (Figure 2.1.
b, panel 3). Therefore,
the mechanism of UbcX mediated
SAC inactivation is conserved between
humans and Xenopus.
Figure 2.1. Ectopic UbcXwt overrides
SAC-arrest in Xenopus egg
extract. (a) CSF-extracts containing
35S-securin was supplemented with
nocodazole and high concentrations of
sperm to activate the SAC. CSF
arrest was released by calcium
addition. (b) At the indicated
time points after the addition
of the specified reagents samples
were taken and 35S-securin was
detected by autoradiography and
XErp1 and α-tubulin by WB. CSF,
cytostatic factor; SAC, spindle
assembly checkpoint; 35S-securin, in
vitro translated, 35S-labeled securin;
wt, wild type; ci, catalytical
inactive.
2.2. UbcX can suppress CSF activity
in Xenopus egg extract
To analyze if an increase in
the activity of UbcX similarly
influences CSF
mediated APC/C inhibition, ectopic
UbcXwt was added to CSF arrested
egg
extract supplemented with a low
concentration of sperm nuclei and
35S-securin
(Figure 2.2. a). Interestingly, also
in these extracts ectopic UbcX
caused APC/C
activation and CSF release in the
absence of the calcium signal,
as indicated by
23
panel 2). However - unlike in
extracts treated with calcium -
XErp1 remained
stable and showed an increase in
its electrophoretic mobility following
exit
from meiosis (Figure 2.2. c, panel
1 and 2), suggesting that
UbcXwt causes CSF
inactivation by different means than
XErp1 degradation. The addition of
UbcXci
or dialysis buffer had no effect
on CSF arrest (Figure 2.2. b,
c, panel 3 and 4),
suggesting that the observed CSF
override is dependent on an
increase in the
catalytic activity of UbcX.
Additionally, the human homologue of
UbcX was equivalent in the
ability to
overcome CSF arrest in Xenopus egg
extract, as the addition of
catalytic active
UbcH10 triggered premature CSF release
(Figure 2.2. d, panel 3),
demonstrating that both UbcX and
UbcH10 are interchangeable in inducing
CSF release.
24
2.3. Elevated UbcX activity prevents
meiosis I - meiosis II
transition in
Xenopus oocytes
To collect evidence for UbcX
mediated regulation of CSF arrest
in vivo, we
injected recombinant UbcX into Xenopus
stage VI oocytes arrested at
prophase
of meiosis I. We induced oocyte
maturation by the addition of
progesterone
and followed the resumption of