Download Control of DNA Synthesis and Mitosis by the Skp2-p27

Molecular Cell
414
a more regulated export pathway. By coordinating cargo
selection ahead of coat polymerization at the tER, the
cell can efficiently designate that this bud’s for the Golgi.
Paul LaPointe, Cemal Gurkan,
and William E. Balch
The Scripps Research Institute
Departments of Cell and Molecular Biology and
The Institute for Childhood and Neglected Disease
10550 N. Torrey Pines Road
La Jolla, California 92130
Selected Reading
Antonny, B., and Schekman, R. (2001). Curr. Opin. Cell Biol. 13,
438–443.
Antonny, B., Gounon, P., Schekman, R., and Orci, L. (2003). EMBO
Rep. 4, 419–424.
Aridor, M., Bannykh, S.I., Rowe, T., and Balch, W.E. (1999). J. Biol.
Chem. 274, 4389–4399.
Figure 1. Models for ER Export
The conventional model for COPII vesicle budding asserts that the
COPII coat machinery is solely responsible for cargo selection (Antonny and Schekman, 2001). Activation of Sar1 through the action
if its GEF, Sec12, leads to the recruitment of Sec23/24, which passively selects cargo for export via coat polymerization promoted
by Sec13/31. In contrast, a scaffolding model places COPII coat
recruitment downstream of a regulatory scaffold. In this new model,
Sec12 is activated after specific arrangements for vesicle formation
have been completed to integrate cargo packaging with coat recruitment to drive vesicle budding.
Aridor, M., Fish, K.N., Bannykh, S., Weissman, J., Roberts, T.H.,
Lippincott-Schwartz, J., and Balch, W.E. (2001). J. Cell Biol. 152,
213–229.
Bannykh, S.I., and Balch, W.E. (1997). J. Cell Biol. 138, 1–4.
Barlowe, C. (2003). Trends Cell Biol. 13, 295–300.
Bickford, L.C., Mossessova, E., and Goldberg, J. (2004). Curr. Opin.
Struct. Biol. 14, 147–153.
Espenshade, P., Gimeno, R.E., Holzmacher, E., Teung, P., and Kaiser, C.A. (1995). J. Cell Biol. 131, 311–324.
Gimeno, R.E., Espenshade, P., and Kaiser, C.A. (1995). J. Cell Biol.
131, 325–338.
Huang, M., Weissman, J.T., Beraud-Dufour, S., Luan, P., Wang, C.,
Chen, W., Aridor, M., Wilson, I.A., and Balch, W.E. (2001). J. Cell
Biol. 155, 937–948.
Nishimura, N., and Balch, W.E. (1997). Science 277, 556–558.
The data presented by Glick and colleagues (Soderholm et al., 2004) raise the possibility that COPII recruitment to the ER membrane is likely only a final step of
Palade, G. (1975). Science 189, 347–358.
Control of DNA Synthesis
and Mitosis by the
Skp2-p27-Cdk1/2 Axis
control of p27 degradation by Skp2 not only allows DNA
replication to proceed properly and on schedule but
also ensures the correct sequence of occurrence of S
phase and mitosis during the cell cycle (Nakayama et
al., 2004).
Skp1 and Skp2 were discovered as interactors of
cyclin A, hence the name S phase kinase-associated
protein 1 and 2 (Zhang et al., 1995). In vitro experiments
had shown that these proteins were unable to directly
modulate the activity of cyclin A-associated kinases, yet
inhibition of Skp2 in living cells prevented entry into S
phase. About 1 year later, it was found that Skp2 is a
member of the large eukaryotic family of F box proteins
(Bai et al., 1996). Work in yeast subsequently demonstrated that Skp1, together with another protein known
as Cul1 and an F box protein, are all subunits of SCF
ubiquitin ligases (reviewed by Reed, 2003). Each SCF
complex was found to contain a different F box protein
that provided specificity by recruiting different substrates for ubiquitinylation and subsequent proteolysis.
At the same time, a different line of research had
shown that the levels of the mammalian CDK inhibitor
A new study reveals a novel role for p27 in inhibiting
Cdk1 activity at G2/M and shows that p27 deficiency
almost completely rescues the aberrations observed
in Skp2⫺/⫺ mice, demonstrating that p27 is the principal downstream effector of the SCFSkp2 ubiquitin ligase.
In order to generate two daughter cells with the same
DNA content and one centrosome, the parent cell needs
to replicate its genome and duplicate the centrosome
only once per each cell division cycle. The molecular
machinery governing S phase is therefore interconnected with that controlling M phase, with a need for
inhibition of the latter while DNA replication is occurring
and vice versa. A new paper by Nakayama and colleagues in the May Developmental Cell shows that the
Soderholm, J., Bhattacharyya, D., Strongin, D., Markovitz, V., Connerly, P.L., Reinke, C.A., and Glick, B.S. (2004). Dev. Cell 6, 649–659.
Previews
415
p27 are mostly controlled by the ubiquitin system and
that its ubiquitinylation is promoted by the ubiquitinconjugating enzyme Ubc3 (Pagano et al., 1995). Subsequent work demonstrated that the ubiquitinylation of
p27 requires its phosphorylation on Thr-187. Since SCF
complexes were known to preferentially recognize substrates when phosphorylated and because Ubc3 works
in concert with SCF complexes, it was suggested that
p27 was a substrate of an SCF ligase. Soon after, Skp2
was shown to be the specific F box protein required for
the degradation of p27, explaining why Skp2 is a positive
regulator of the S phase and connecting the two converging lines of research (reviewed in Reed, 2003).
When the mouse Skp2 locus was deleted by Nakayama’s group, Skp2 function in controlling p27 cellular
abundance was confirmed (Nakayama et al., 2000).
Skp2-deficient mice are smaller than littermate controls
and show a generalized hypoplasia, exactly the opposite
phenotype from that observed a few years before in
p27⫺/⫺ mice. Accordingly, Skp2⫺/⫺ mouse embryonic
fibroblasts (MEFs) grew more slowly than controls. Histologic examination revealed that Skp2⫺/⫺ hepatocytes,
epithelia of bronchioles and proximal renal tubules, and
MEFs had nuclei larger than those of wild-type littermates. Further examination of these four cell types
showed that they all had more than two centrosomes
per cell. In addition, Skp2⫺/⫺ hepatocytes display polyploidy, proving that in certain cell lineages Skp2 is required to avoid DNA rereplications (note that ploidy and
centrosome number was analyzed only in the tissues in
which nuclear enlargement was originally observed). In
a subsequent paper, it was shown that Skp2⫺/⫺ mice
subjected to partial hepatectomy reestablished liver
mass by increasing hepatocyte size rather than hepatocyte number because of the inability to enter mitosis
(Minamishima et al., 2002). The fact that certain tissues
are more affected than others by Skp2 deficiency is
likely due to redundant pathways present in nonaffected
cell types. This is in agreement with the majority of cell
cycle knockouts, in which the effect of losing a particular
gene is almost invariably tissue specific.
No known F box protein targets only one substrate,
and indeed, in addition to p27, several studies suggested that Skp2 directs the degradation of as many
as ten other substrates (reviewed in Reed, 2003). To
understand which substrate was responsible for the
phenotype of Skp2⫺/⫺ mice, Nakayama’s team started
from the best-characterized one, namely p27, and generated a mutant mouse deficient in both Skp2 and p27.
These mice have a body size similar to the p27⫺/⫺ mice,
demonstrating that p27 accumulation is responsible for
the generalized hypoplasia of Skp2⫺/⫺ mice. Importantly, hepatocytes, bronchiolar epithelia cells, renal tubule cells, and MEFs of Skp2⫺/⫺;p27⫺/⫺ mice do not
show the nuclear enlargement and centrosome overduplication typical of Skp2⫺/⫺ cells. Finally, Skp2⫺/⫺;
p27⫺/⫺ hepatocytes do not show endoreduplication and,
when induced to proliferate, reacquired the ability to
enter mitosis.
At this point, the story started to look more and more
like another where mutations in the yeast Pop1 F box
protein gene, which is required for degradation of the
Cdk1 inhibitor Rum1, induces mitotic failure and endoreduplication (Kominami and Toda, 1997). Nakayama and
Figure 1. Mechanism of Activation of Cdk1 and Cdk2 Complexes
by Skp2
Skp2 is required for the ubiquitinylation of p27 at G1/S and during
S and G2 phases. Specifically, Skp2 targets for degradation only a
subpopulation of p27: that in complex with Cdk2 and, now we have
learned, with Cdk1 but not the fraction in complex with Cdk4 and
Cdk6. Furthermore, we still don’t know whether p27 in complex with
Cdk3 and other partners is targeted by Skp2 or different ubiquitin
ligase(s). The story is additionally complicated by the fact that in
somatic cells Cdk2 and Cdk1 each form different complexes with
at least three major cyclins (cyclin E1/cyclin E2/cyclin A1 and cyclin
A2/cyclin B1/cyclin B2, respectively; herein generically referred to
as cyclin E, cyclin A, and cyclin B). The details of which complex
promotes DNA replication, inhibits DNA rereplication, or induces
entry into mitosis are not well understood. When the Skp2 gene is
deleted in mice, p27 is stabilized, resulting in a generalized hypoplastic phenotype. In certain tissues (see text), a compensatory
increase in cyclin E levels occurs, such that cyclin E-Cdk2 activity
is not compromised by the accumulation of p27. In contrast, cyclin
A-Cdk2, cyclin A-Cdk1, and cyclin B-Cdk1 are inhibited, thus slowing
down the entry in mitosis. A delay in G2 with normal cyclin E-Cdk2
activity is the likely cause of centrosome overduplication. In addition, at least in hepatocytes, a more severe inhibition of Cdk1 activity
attenuates the prevention of rereplication, resulting in endoreduplication.
colleagues hypothesized that the similarity extended
also to the ability of p27 to inhibit not only cyclin E-Cdk2
and cyclin A-Cdk2, two kinases promoting DNA synthesis, but also Cdk1, a kinase promoting entry into mitosis.
Inhibition of Cdk1 (historically known as Cdc2) by p27
had already been shown in vitro by Tony Hunter’s group
10 years earlier (Toyoshima and Hunter, 1994), but it
was forgotten because of the overwhelming literature
showing p27 inhibition of the G1/S transition. Thus, it
was found that in Skp2⫺/⫺ MEFs there is an increased
association of p27 with both Cdk1 and Cdk2 with a consequent reduction in the kinase activity of cyclin A-Cdk2,
cyclin A-Cdk1, and cyclin B-Cdk1 (Figure 1). In contrast,
cyclin E-Cdk2 remains unaffected because of a compensatory increase in cyclin E levels. This probably does
the trick: in many cell lineages Cdk1 activity is too low
to allow efficient entry into mitosis, but Cdk2 activity is
Molecular Cell
416
high enough to promote DNA synthesis. Because
Skp2⫺/⫺ cells spend too much time in G2 and the centrosome cycle is dissociated from the cell division cycle,
centrosome overduplication occurs, becoming more evident with the increase in cell doubling. But how does
one explain the endoreduplication observed in hepatocytes? In these cells, Cdk1 inhibition must be more severe, since a mitotic block rather than a prolonged G2/
M transition is found. A CDK (perhaps cyclin A-Cdk1) is
necessary to disassemble the prereplication complex
after an origin has fired, thus preventing rereplication
from the same origin (Figure 1). So, in Skp2⫺/⫺ hepatocytes Cdk1 activity is strongly suppressed, blocking not
only the entry into mitosis but also the prevention of rereplication.
The cell can be considered an ensemble of networked
molecular machines. The ubiquitin system allows modular regulatory components of these machines to quickly
disappear thus contributing to the synchronization of
the cellular gears. The new findings from Nakayama et
al. provide new interesting insights into how, by regulating the activity of both Cdk2 and Cdk1 complexes, Skp2
controls the interrelationship between S and M phase.
Overexpression of Skp2 observed in many human tumors (Pagano and Benmaamar, 2003) may contribute
to the deregulated proliferation and genetic instability
typical of cancer cells, not only by increasing the activity
of CDKs but also by disturbing the delicate temporal
equilibrium of the CDK regulatory network.
Michele Pagano
Department of Pathology
New York University School of Medicine and
NYU Cancer Institute
550 First Avenue, MSB 599
New York, New York 10016
FoxO: Linking New Signaling
Pathways
and stress responses in mammalian cells. It has been
demonstrated that members of the FKHR (FoxO) family
of transcription factors are negatively regulated by nuclear exclusion promoted by PI(3)-kinase transduction
of growth factor-mediated signals. Biochemical analysis
of several FoxO proteins has demonstrated that this
negative regulation is mediated by phosphorylation of
conserved residues by the serine/threonine kinase Akt/
PKB (reviewed in Burgering and Kops, 2002). Results
of several studies have shown that members of the FoxO
family of transcription factors play a critical and evolutionarily conserved role in the downregulation of cellular
responses (e.g., proliferation or differentiation) normally
elicited by growth factors activating the PI(3) kinase
signal transduction pathway.
The roles of three of the FoxO family members, FoxO1,
FoxO3a, and FoxO4, in normal development and physiology have recently been elucidated through the disruption of each of the genes in mice (Castrillon et al., 2003,
Hosaka et al., 2004). Foxo1 null embryos died on embryonic day 10.5 (E10.5) as a consequence of incomplete
vascular development, suggesting a crucial role of this
transcription factor in vascular formation. Both Foxo3a
null and Foxo4 null mice were viable and grossly indistinguishable from their wild-type littermates. However,
Foxo3a null females showed age-dependent infertility
and had abnormal ovarian follicular development. In
contrast, histological analyses of Foxo4 null mice did
not identify any consistent abnormalities. Thus, the
physiological roles of Foxo genes are functionally diverse in mammals.
Two recent reports (Seoane et al., 2004; Hu et al., 2004)
reveal new roles for FoxO proteins in cell proliferation
and tumorigenesis. Seoane and colleagues show that
FoxO proteins play key roles in the TGF␤-dependent
activation of p21Cip1 by partnering with Smad3 and
Smad4. FoxG1, a protein from a distinct Fox subfamily,
binds FoxO/Smad complexes and blocks p21Cip1 expression. These interactions establish a relationship
between the PI3K pathway, FoxG1, and the TGF␤/
Smad pathways. The second report identifies I␬B kinase as a negative regulator of FoxO proteins, suggesting a mechanism for relieving negative regulation
of cell cycle and promoting tumor cell proliferation.
Transcription factors play a role in the differential expression of genes by binding to their DNA regulatory
elements. Such binding leads to the recruitment of additional factors to the initiation complex and results in
either the potentiation or repression of transcriptional
initiation. Members of the FoxO (Forkhead bOX-containing protein, O sub-family) family of transcription factors have been shown to play roles in a variety of cellular
processes from longevity, metabolism, and reproduction in C. elegans to regulation of gene transcription
downstream from insulin, cell cycle arrest, apoptosis,
Selected Reading
Bai, C., Sen, P., Hofman, K., Ma, L., Goebel, M., Harper, W., and
Elledge, S. (1996). Cell 86, 263–274.
Kominami, K., and Toda, T. (1997). Genes Dev. 11, 1548–1560.
Minamishima, Y.A., Nakayama, K., and Nakayama, K.I. (2002). Cancer Res. 62, 995–999.
Nakayama, K., Nagahama, H., Minamishima, Y., Matsumoto, M.,
Nakamichi, I., Kitagawa, K., Shirane, M., Tsunematsu, R., Tsukiyama,
T., Ishida, N., et al. (2000). EMBO J. 19, 2069–2081.
Nakayama, K., Nagahama, H., Minamishima, Y.A., Miyake, S., Ishida,
N., Hatakeyama, S., Iemura, S., Natsume, T., and Nakayama, K.I.
(2004). Dev. Cell 6, 661–672.
Pagano, M., and Benmaamar, R. (2003). Cancer Cell 4, 251–256.
Pagano, M., Tam, S., Theodoras, A., Beer, P., Delsal, S., Chau, I.,
Yew, R., Draetta, G., and Rolfe, M. (1995). Science 269, 682–685.
Reed, S. (2003). Nat. Rev. Mol. Cell Biol. 4, 855–864.
Toyoshima, H., and Hunter, T. (1994). Cell 78, 67–74.
Zhang, H., Kobayashi, R., Galaktionov, K., and Beach, D. (1995).
Cell 82, 915–925.