Download Polo kinase and progression through M phase in Drosophila

Oncogene (2005) 24, 230–237
& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00
www.nature.com/onc
Polo kinase and progression through M phase in Drosophila: a perspective
from the spindle poles
David M Glover*,1
1
Cancer Research UK Cell Cycle Genetics Research Group, Department of Genetics, University of Cambridge, Downing Street,
Cambridge CB2 3EH, UK
Genes for the mitotic kinases Polo and Aurora A were
first identified in Drosophila through screens of maternal
effect lethal mutations for defects in spindle pole
behaviour. These enzymes have been shown to be highly
conserved and required for multiple functions in mitosis.
Polo is stabilized at the centrosome by association with
Hsp90. It is required for centrosome maturation on Mphase entry in order to recruit the c-tubulin ring complex
and activate the abnormal spindle protein, Asp. These
events facilitate the nucleation of minus ends of microtubules at the centrosome. The localization of Polo at the
kinetochore and the mid-zone of the central spindle
together with the phenotypes of polo mutants point to
functions at the metaphase to anaphase transition and in
cytokinesis. The latter are mediated, at least in part,
through the Pavarotti kinesin-like motor protein and its
conserved counterparts in other metazoans.
Oncogene (2005) 24, 230–237. doi:10.1038/sj.onc.1208279
Keywords: Polo kinase; centrosome; spindle; microtubule-organizing centre
Isolating cell cycle mutants in Drosophila
Genetic studies of mitotic progression in Drosophila
have given valuable insight into the regulation of this
process in metazoan cells. In the 1980s, my own
laboratory adopted two strategies to identify mitotic
mutants: a search through existing collections of
maternal effect lethal mutants that had been generated
with the primary objective of studying early embryonic
development, and the use of P-element transposonmediated mutagenesis to isolate mutants that showed
zygotic lethality. The rationale for such screens lay in
the biology of Drosophila development and in particular
the times during development at which the protein
machinery required for cell division might become
limiting. For its first two hours of development, the
embryo is a syncytium within which nuclei undertake
highly simplified cell cycles of alternating S- and Mphases at roughly 10-min intervals gradually lengthen*Correspondence: DM Glover; E-mail: [email protected]
ing to no more than 20 min to allow the onset of zygotic
transcription from cycle 10 onwards. Individual somatic
cells are not formed until the 14th cycle at which point a
G2 phase is introduced, the duration of which is
regulated by the accumulation and activation of String,
the fly cdc25 tyrosine phosphatase that activates cyclindependent kinase 1 (Cdk1). In order to achieve the
prodigious increase in nuclei from the single zygotic
complement to roughly 10 000 nuclei during syncytial
development, the egg is loaded with a dowry of
maternally provided cell cycle regulatory molecules.
Such proteins are synthesized in the nurse cells of the
female germ line during oogenesis. Thus, a class of
mutations was expected that when homozygous in the
mother would cause her to produce eggs deficient in
essential cell cycle regulatory proteins. In addition to
such maternal effect lethal mutations, another class of
mutation was also known to result in lethality when
homozygous in the zygote. In such cases, recessive lethal
mutations are carried out in each heterozygous parent.
The heterozygous mother is able to produce a dowry for
all her offspring, 25% of which will be homozygous for
the recessive mutation. Such homozygous mutants can
develop until a stage at which the essential maternally
supplied cell cycle protein becomes limiting. This is
usually either in the early cell cycles that follow
cellularization of the embryo or later in larval development when a burst of cell proliferation is required to
produce the adult tissues during metamorphosis. string
mutants, for example, arrest development during the G2
phase of cycle 14 (Edgar and O’Farrell, 1990). Mutants
of dd4, a gene now known to encode a component of the
g-tubulin ring complex (g-TuRC), show late larval
lethality (Gatti and Baker, 1989; Barbosa et al., 2000).
The lethal phase is in part a reflection of the degree to
which the protein is turned over during the mitotic
cycles.
Attracted to the poles
One of the first collections of maternal effect mutants
screened by my laboratory was kindly provided to us by
Christiani Nüsslein-Volhard. This proved to be a
treasure trove of cell cycle regulatory mutants. Some
of these identified genes encoded proteins strictly needed
Roles of Polo kinase in Drosophila
DM Glover
231
for the embryonic division cycles, such as gnu, a gene we
found to be required to turn off DNA replication upon
the completion of female meiosis (Freeman et al., 1986;
Freeman and Glover, 1987), and that is now known to
encode a component of a complex of molecules required
to regulate cyclin B levels (Lee et al., 2003; Renault et al.,
2003). Others were weak hypomorphic alleles of genes
whose function is required both in the syncytial
embryonic divisions and for cell division at later
developmental stages. This category included the
original alleles of both polo (Sunkel and Glover, 1988)
and aurora (Glover et al., 1989; Leibowitz, 1990). In
examining the defective syncytial embryos produced by
homozygous mutant mothers, we noticed several mutants that had defects at their spindle poles. This led us
to adopt a series of names for the corresponding genes
after the magnetic poles of the Earth or geo-magnetic
phenomena associated with them. Whereas polo and
aurora subsequently proved to be highly appropriate
names, in other cases our nomenclature was no so
predictive of function; lodestar, for example (Girdham
and Glover, 1991), encodes a protein actually not
directly related to spindle pole function but required
for chromatin structure and to release transcripts by
RNA polymerase II in an ATP-dependent manner! We
were later able to isolate stronger hypomorphic alleles
for both polo and aurora that showed zygotic lethality,
suggesting the requirement for the two gene products in
cell division throughout development (Glover et al.,
1989, 1995; Llamazares et al., 1991). Indeed, the second
allele of polo that we found originated in a collection of
P-element mutants established by Roger Karess. This
collection was also to liberate rough deal (Karess and
Glover, 1989) that we now know to encode a component
of the zest-white 10 checkpoint complex and abnormal
anaphase resolved (aar) that encodes a regulatory
subunit of protein phosphatase 1A, and that also
participates in regulating the metaphase–anaphase
transition (Mayer-Jaekel et al., 1993). The P-element
tag allowed us to identify the polo gene and show for the
first time that it encoded a putative protein kinase
(Llamazares et al., 1991). Once the gene was cloned,
antibodies that were able to immunoprecipitate active
kinase were raised against it. Studies of carefully staged
and individually selected syncytial embryos indicated
that the enzyme was activated in mitosis, its activity
peaking slightly later than Cdk1 (Fenton and Glover,
1993).
The phenotype of the original polo1 allele gave a
number of clues to the function of the gene product.
Multiple mitotic defects were evident in mutant
embryos, but characteristically the centrosomal antigen,
CP190 (at that time termed the Bx63 antigen), was not
recruited to discrete centrosomal bodies (Sunkel and
Glover, 1988; see also below). The study also showed
that, although homozygous mutants of polo1 could
develop to adulthood, they still showed mitotic defects
during the larval and pupal stages of development. Over
a decade later, a detailed cytological examination of the
phenotype of polo1 in the female germ line revealed
defects in the organization of the common pole shared
by the tandem spindles of the second meiotic division
(Riparbelli et al., 2000). Moreover, the polar body nuclei
resulting from the meiotic divisions failed to arrest their
development and went on to undertake some mitotic
divisions. This was associated with a failure in the
development of the sperm aster, a dramatic array of
microtubules (MTs) that normally facilitates migration
of the female pronucleus to its male counterpart. In the
absence of the sperm aster, the two pronuclei remained
apart and both underwent a limited number of
independent cycles of haploid mitoses. Thus, a series
of defects, most of which are related to an inability of
MT-organizing centres to function correctly, led to a
variety of developmental problems. In addition to the
above functions of polo in the female germ line, the gene
was also shown to function in the male germ line, where
it was required for correct chromosome segregation in
meiosis (Sunkel and Glover, 1988). This first study and
the subsequent work of Llamazares et al. (1991) showed
that defects in the spindle poles were a characteristic but
not exclusive aspect of the mutant phenotype.
Mitotic functions of Drosophila Polo and its conserved
counterparts
The first realization that Polo kinase was highly
conserved came when Clay et al. (1993) serendipitously
found a mouse homologue and when the budding yeast
CDC5 gene was also identified as a homologous protein
kinase (Kitada et al., 1993). The cdc5 mutant had been
identified in a part of Hartwell’s seminal work to
identify cell division cycle mutants that defined all of the
central concepts of cell cycle control. cdc5 fell into a
group of mutants that showed a late mitotic phenotype.
In such arrested cells, the bud did not separate from the
mother cell and grew so as to form a ‘dumb-bell’-shaped
pair of connected cells in which the spindle appeared to
persist. This group of genes is now known as the mitotic
exit network (MEN), a conserved cassette of genes that
function in late mitosis. Thus, there appeared to be a
paradox: on the one hand, the phenotype of polo in
Drosophila melanogaster seemed to suggest a function
early in the mitotic cycle in spindle formation, whereas
that of cdc5 in Saccharomyces cerevisiae, a late anaphase
function. This led us to ask what the phenotypes of a
Polo-like kinase might be in another genetically
tractable organism. This was the fission yeast Schizzosaccharomyces pombe that had been so valuable in
characterizing the mitotic roles of the p34cdc2 kinase that
we now know as Cdk1. Disruption of the fission yeast
Plk gene, plo1, led to two distinct phenotypes: the
formation of monopolar spindles and a failure of
septation (Ohkura et al., 1995). The former phenotype
was not dissimilar to aspects of the polo phenotype in
flies and the latter was already known from other fission
yeast mutants that define the septum-inducing network
(SIN). It had already become apparent at that time that
the budding yeast MEN and fission yeast SIN were
analogous regulatory cassettes. Thus, the fission yeast
study first clearly showed that a Polo-like kinase could
Oncogene
Roles of Polo kinase in Drosophila
DM Glover
232
function at two different mitotic stages and thereby
pointed to the possibility of reconciling the apparently
disparate observations from Drosophila and budding
yeast. Whether the MEN and SIN networks are truly
replicated in metazoan cells is still by no means certain.
However, as I will discuss below, a role for Polo in the
late stages of mitosis was subsequently demonstrated.
In the right place at the right time
The polo-like kinases have been shown to follow similar
dynamic patterns of localization during progression
through mitosis essentially in all metazoans examined.
Perhaps the most dramatic illustration of the dynamics
of the protein kinase has come from studies by Claudio
Sunkel’s laboratory (Moutinho-Santos et al., 1999).
These workers expressed a GFP-tagged form of Polo
from its own promoter in a transgenic fly line. The
tagged enzyme localized to centrosomes at mitotic entry
and then became associated with the nuclear membrane
until its breakdown. During prometaphase, Polo kinase
associated with kinetochores and prior to cytokinesis
with the mid-part of the central spindle, a characteristic
structure that forms in late anaphase and that is
essential for cytokinesis.
This dynamic pattern of localization led to an
expectation that there would be multiple substrates for
the kinase at its multiple cellular sites. This prediction is
being borne out, although to date still relatively few
substrates have been determined. A characteristic
feature of the polo-like kinases is the conserved Poloboxes in the C-terminal domain of the protein. Studies
from a variety of organisms have indicated that the Polo
boxes are required to target the protein to specific
cellular sites (for example, Lee et al., 1998; Reynolds and
Ohkura, 2003). Recently, it has been shown that the
Polo boxes bind to a phosphopeptide motif (Elia et al.,
2003a, b; Barr et al., 2004). It seems likely that this
docking site is phosphorylated by either cdk1, MAPK or
even plk itself to facilitate binding and perhaps
explaining some evidence for co-operativity of Polo
with other mitotic kinases.
The search for bona fide substrates with the Drosophila Polo was initiated by Alvaro Tavares, who
compared the proteins phosphorylated by exogenous,
bacculovirus-expressed Polo kinase added to extracts of
cytoplasm from polo1 mutant or wild-type embryos
(Tavares et al., 1996). Endogenous Polo phosphorylated
a similar series of proteins. These included b-tubulin and
two proteins of 85 kDa and approximately 200 kDa.
Subsequently, it appears that these last two proteins
are likely to be an 85 kDa MT-associated protein
released from MTs following phosphorylation by Polo
(Cambiazo et al., 2000) and Asp, the product of the
abnormal spindle gene (Saunders et al., 1997; Avides and
Glover, 1999; Avides et al., 2001; see below).
As I will detail below, some other Polo substrates
have been identified in Drosophila. However, without
doubt, there remain many more to be determined. One
characteristic of proteins that have been phosphorylated
Oncogene
by Polo is that they become recognizable by the
monoclonal antibody MPM2. This antibody raised
against a mitotic phospho-epitope by Potu Rao’s lab
in the early 1980s (Davis et al., 1983) recognizes scores
of proteins in Western blots of mitotic but not
interphase cells. It was shown by Kumagai and Dunphy
(1996) that the Xenopus plk1 could phosphorylate and
participate in activating cdc25 and in so doing generated
an MPM2 epitope. The Plks are not the only protein
kinases able to generate such epitopes, but in Drosophila
the level of cellular MPM2 immunoreactivity correlates
directly with the severity of polo mutant alleles,
suggesting that Polo kinase generates such epitopes on
a range of substrates (Logarinho and Sunkel, 1998).
Moreover, the sites of immunostaining of mitotic cells
by MPM2 broadly correspond to the known localization
patterns of plks. Thus, this antibody is likely to find
continued application in studying the cell biology of
Polo-mediated phosphorylation.
Centrosome maturation
Although much remains to be discovered, we now have
a skeletal outline of the role of Polo kinase in
centrosome maturation in Drosophila. The inability to
recruit the centrosomal antigen, now known as CP190,
seen in polo1-derived embryos appears to be a secondary
consequence of a failure to recruit the g-TuRC, a
complex known to cap the minus ends of MTs on
centrosomes and that is required for their nucleation.
CP190 has been shown to interact with the g-TuRC and
indeed, when this complex is disrupted in mutants of
dd4, a gene encoding a 91 kDa component of the
complex, CP190 then becomes dispersed around the cell
(Barbosa et al., 2000). Strikingly, neither g-tubulin nor
CP190 are recruited to the mitotic centrosome either in
strong hypomorphic polo mutants (Donaldson et al.,
2001; Figure 1) or following depletion of the protein
kinase by RNAi in cultured cells (Bettencourt-Dias et al.,
2004). A failure of centrosomes to recruit g-tubulin has
also been observed in HeLa cells following microinjection of antibodies against Polo-like kinase (Lane and
Nigg, 1996).
Some understanding of centrosome maturation in
Drosophila has been gained through the use of an in vitro
system developed to study MT nucleation (Figure 2a).
Partially purified centrosomes prepared from extracts of
Drosophila embryos are capable of nucleating asters of
MTs on a microscope slide. Such preparations lose their
nucleating ability if stripped with potassium iodide, but
this can be restored upon ‘add-back’ of a high-speed
supernatant of cytoplasmic extract. If the cytoplasmic
extract is depleted of the g-TuRC, then the cytoplasmic
extract is no longer able to restore MT-nucleating
activity (Moritz et al., 1998). A second MT-associated
protein (MAP) Asp, product of the abnormal spindle
gene, is also essential for the cytoplasmic extract to be
able to restore MT-nucleating activity (Avides and
Glover, 1999). A deficiency in the ability to restore
MT nucleation is thus seen with cytoplasmic extracts
Roles of Polo kinase in Drosophila
DM Glover
233
Wild-type
polo mutant
Centrosomin
γ-tubulin
Figure 1 Polo kinase is required for the recruitment of g-tubulin to the centrosome. Mitotic figures from wild-type and polo9 mutant
cells in which MTs are stained green and DNA, red. The upper panels are stained to reveal the core centrosomal protein, centrosomin
(blue). Note that this core antigen is present at the centrosomes of both wild-type and polo mutant cells. The lower panels are stained to
reveal g-tubulin (blue). Note that this antigen is present at the poles of the wild-type spindle but is not detectable at the centrosomes of
the polo mutant. Immunostaining was carried out by Mary Donaldson
depleted of Asp protein either through the experimental
use of antibodies or as a result of asp mutations. This
functional deficiency can be rescued by adding a purified
preparation of Asp protein to the mutant cytoplasm.
Typically, asp mutants have mitotic spindles with
highly unfocused poles and likely show a high mitotic
index. The finding that polo1 and aspE3 double mutants
showed a further synergistic increase in mitotic index
could be interpreted if Asp were a substrate of the Polo
kinase (Gonzalez et al., 1998). This was later shown
directly by Avides et al. (2001), who found that
phosphorylation of Asp by Polo in vitro generated a
mitotic phospho-epitope recognized by the monoclonal
antibody MPM2. Although Asp still localizes to the
minus ends of MTs in polo mutants, it appears that the
Asp protein has to be phosphorylated by Polo for it to
be active. Cytoplasmic extracts from polo mutants
would not rescue the nucleating activity of salt-stripped
centrosome preparations, but this could be restored by
adding back phosphorylated Asp protein, and further
enhanced by active Polo kinase (Avides et al., 2001).
Therefore, it seems that, in Drosophila, Polo kinase plays
at least two roles in the mitotic maturation of
centrosomes: a role both in recruiting the g-TuRCs
and in activating Asp protein that is recruited independently of Polo activity. Together, these two minus end
MT-associated proteins then enable the increase in MT
density that is seen at the centrosome upon the entry
into mitosis (Figure 2b). The finding in other organisms
of MT-associated protein substrates of the plks suggests
that these enzymes may have further roles in regulating
MT behaviour during mitosis (Budde et al., 2001; Yarm,
2002).
Many questions remain unanswered. What is the Polo
substrate whose phosphorylation facilitates recruitment
of the g-TuRC? What is the centrosomal docking site for
Polo kinase itself? Is there a specific docking molecule or
are its centrosomal substrates sufficient? How is Polo
kinase activated at the centrosome? Is there a functional
equivalent of the spindle pole body component Cut12 of
fission yeast that binds Plo1 and promotes its activity,
and are kinases of the NIMA family required for this
process (Grallert and Hagan, 2002; MacIver et al.,
2003)? The ability to apply a concerted biochemical and
genetic approach in Drosophila may well enable some of
these questions to be approached. We are beginning to
glimpse some aspects of the regulation of Polo kinase
activity at the centrosome. It appears in both Drosophila
and human cells that the Plks require the chaperone
protein Hsp90 to maintain their stability (de Carcer
et al., 2001). In Drosophila, this interaction is particularly important at the centrosome, where the two
proteins co-localize. In either Hsp90 mutants or in
cells treated with the Hsp90 inhibitor, geldanamycin,
Oncogene
Roles of Polo kinase in Drosophila
DM Glover
234
a
geldanamycin
treated
control
geldanamycin treated +
Polo kinase
b
hsp90
Polo
P
hsp90
Polo
P
hsp90
P
Polo
Figure 2 Studies of MT nucleation by Drosophila centrosomes in vitro have led to a model for Polo function in the mitotic maturation
of the centrosome. (a) Purified preparations of centrosomes lose their ability to nucleate asters of MTs following treatment with KI.
This can be restored by incubation in extracts of untreated cytoplasm (control), whereupon centrosomes (green–yellow in overlap) then
nucleate rhodamine-labelled tubulin (red). Treatment with geldanamycin destroys this ability of cytoplasmic extracts to rescue MT
nucleation (centre panel). Addition of exogenous active Polo kinase overcomes the effect of geldanaycin (right-hand panel). Similar
experimentation has shown that the abnormal spindle protein is required and that to be active is must be phosphorylated by Polo (see
text for detail). These are taken from experiments of Carmo Avides. (b) A model for the mitotic maturation of the Drosophila
centrosome. The dark grey cylinders represent the centrioles and the large light grey sphere, the pericentriolar material. Polo and
associated hsp90 are represented in green, Asp in red and the g-tubulin ring complex in yellow. Active Polo kinase is required to recruit
the g-tubulin ring complex and to activate Asp. These molecules then co-operate to nucleate MTs (red)
centrosome segregation and maturation are affected (de
Carcer et al., 2001). The ability of the cytoplasmic
extracts to restore the nucleating ability of salt-stripped
centrosomes is also lost following geldanamycin treatment, and this can be restored by addition of active Polo
kinase. This indicates the potential importance of such
chaperone proteins for processes in which complex
molecular structures need to be rapidly assembled and
disassembled during the mitotic cycle. Thus, it seems
likely that other chaperone protein complexes will also
be required at the centrosome.
Polo and the metaphase–anaphase transition
A number of indirect observations have implicated the
Plks as having a role in activating the anaphasepromoting complex or cyclosome (APC/C) in Xenopus
Oncogene
(Descombes and Nigg, 1998), budding yeast (Charles
et al., 1998; Shirayama et al., 1998) and mammalian cells
(Kotani et al., 1998, 1999). The phosphorylation
patterns of APC/C components have been carefully
studied by Golan et al. (2002), who have shown that
Plk1 and cdk1-cyclin B have additive effects in
phosphorylating and activating the APC/C, the former
preferentially phosphorylating Cdc16 and Cdc23 and
the latter, Cdc27. It was also recently shown that the
fission yeast Polo-like kinase, Plo1, physically interacts
with the Cut23 (Apc8) component of the APC/C (May
et al., 2002). These observations have prompted a study
of genetic interactions between alleles of polo and
makos, a gene encoding the Drosophila Cdc27 (Deak
et al., 2003). However, this did not cast any light on the
potential function of Polo in activating APC/C, but
instead led to the unexpected finding that the weakly
hypomorphic phenotype of polo1 in larval neuroblasts
Roles of Polo kinase in Drosophila
DM Glover
235
was enhanced by mks with respect to its effect upon
centrosome structure and function. The polo1 mutant
shows some limited recruitment of the g-TuRC to the
centrosome in mitosis. In an mks background, however,
the centrosomal phenotype of polo1 resembles the
stronger polo hypomorphs with no detectable g-TuRC,
suggesting that the APC/C is required either directly or
indirectly to fully activate Polo kinase function. Moreover, the mks1 polo1 double mutants also showed absence
of MPM2 reactivity at the centrosome consistent with
reduced Polo kinase activity. This has led to the
speculation that an early mitotic function of the APC/
C may be required for the full activation of Polo kinase
and the recruitment of the g-TuRC to the centrosome.
One possibility is that the APC/C is needed to degrade
cyclin A or some other critical regulatory molecule that
remains present in mks1polo1 cells.
An indication that Polo may facilitate aspects of
sister chromatid separation at the metaphase–anaphase
transition comes from a report that downregulation of
Polo kinase function will enhance the effects of
stabilized Securin. It is generally thought that Securin
is destroyed by the APC/C to release separase and
promote sister separation. In Drosophila, however, a
Securin stabilized by a D-box mutation alone is
insufficient to prevent separase activity, but the effects
of this mutation can be enhanced either by reduced Polo
kinase or by stabilized cyclin A (Leismann and Lehner,
2003). The mechanisms of this interaction are presently
a mystery. It seems likely to involve some interplay
between Polo kinase and the Cyclin A : cdk1, as this
cyclin has only been found in complexes with cdk1 in
Drosophila.
Although the above genetic interactions give some
tantalizing glimpses of Polo’s participation in regulatory
networks, they do not yet permit any mechanistic
insight. Thus, much remains to be done to clarify the
roles of Polo kinase at the metaphase–anaphase transition both in Drosophila and other organisms.
Roles for Polo in cytokinesis
As I have discussed above, studies in fission yeast
pointed to the major role played by Plo1 kinase in
directing septum formation. Indeed, when overexpressed, it could drive cells into the septation pathway
irrespective of whether cells were arrested in G1, G2 or
undergoing repeated rounds of S phase (Ohkura et al.,
1995). Indeed, it has been suggested that Plo1 acts
upstream of the SIN to induce septation (Tanaka et al.,
2001). This raised the question of whether Plks could
have similar functions late in mitosis in metazoan cells?
A hint of this possibility first came from an observation
of Lee et al. (1995) that the kinesin-like motor protein
MKLP1 was associated with immunoprecipitates of
murine Plk1. At that time, it was believed that the main
function of this motor was in anaphase B spindle
elongation (Nislow et al., 1992). However, when the
counterpart of MKLP1 was identified in Drosophila
(Adams et al., 1998), it was shown, through its mutant
phenotype, to have a role in organizing the central
spindle late in mitosis in order for the correct execution
of cytokinesis. The failure of cytokinesis in embryos
results in the formation of large cells, hence the name
assigned to the locus, pavarotti. We now know that this
family of motor proteins plays a conserved role in
organizing the central spindle in anticipation of cytokinesis and that one of the functions of the motor is to
deliver an essential activating protein for a small
GTPase (MgcRacGAP in mammalian cells; RacGAP50C in Drosophila) to the equator of the cell. Just
as in mammalian cells, Polo kinase and Pav-KLP coimmunoprecipitate, and furthermore they co-localize in
the central part of the spindle.
A functional role for Polo in cytokinesis in metazoan
cells was again first revealed through the phenotype
of polo1 mutants in spermatogenesis (Carmena et al.,
1998; Herrmann et al., 1998). This cell type has proved
to be extremely valuable in assessing whether proteins
that function early in mitosis or meiosis have later
roles in cytokinesis. This is because primary spermatocytes of Drosophila are unusual in having a very weak
spindle integrity checkpoint. Thus, mitotic defects that
would cause a significant metaphase arrest in somatic
cells only delay the progress of male meiosis by a matter
of some minutes. Consequently, polo mutant spermatocytes are able to undertake the metaphase–anaphase
transition and then begin to prepare for cytokinesis.
As polo1 is a weak hypomorph, some cells within the
cysts of primary spermatocytes are able to undertake
cytokinesis, but in many cells the central spindle
fails to form (Figure 3) and Pav-KLP is not recruited
to a ring-like structure at the equator as would
normally occur.
Together, these pointers suggested a role for the Plks
in directing the function of this Pavarotti-KLP family of
motor proteins. The molecular interactions between
Plk1 and MKLP1 have now been characterized in some
detail by Liu et al. (2004). However, the exact function
of the Plks in regulating the Pav-KLP class of motors is
still uncertain and in mammalian cells is further
complicated by the finding of another related motor
protein, MKLP2, also phosphorylated by Plk1 (Neef
et al., 2003).
The precise roles for Polo in the emerging network
that regulates cytokinesis are not yet clear. It is likely
that numerous substrates have to be phosphorylated by
Polo to ensure the correct execution of this process.
Moreover, it appears important to inactivate Plk1 in a
timely manner to complete cytokinesis correctly (Lindon
and Pines, 2004). It is curious as to how many proteins
play roles in organizing the centrosome and the early
mitotic spindle have roles in cytokinesis. Studies in
Drosophila have revealed that not only Polo but also
Asp (Wakefield et al., 2001; Riparbelli et al., 2002) and
the g-TuRC (Sampaio et al., 2001; Barbosa et al., 2003)
are required for the correct organization of the central
spindle and thus cytokinesis. Perhaps Polo kinase
ensures the phosphorylation of a set of familiar
substrates both at the onset and at the end of the
mitotic process.
Oncogene
Roles of Polo kinase in Drosophila
DM Glover
236
microtubules
DNA
metaphase
anaphase
late anaphase
- telophase
Figure 3 Polo is required to form the central spindle for cytokinesis. The main panels show meiotic figures from cysts of
spermatocytes from polo1 males. The insets represent single meiotic figures from equivalent wild-type cysts. The left-hand column
shows staining of MTs and the right-hand column, DNA. The spindle MTs of wild-type and polo mutant cells are virtually
indistinguishable at metaphase (upper panels). At anaphase the distinctive central spindle structure forms in anticipation of cytokinesis
in wild-type cells (arrowhead, left-hand central inset). This structure is lacking in the polo mutant cells at this stage (arrow, main lefthand central panel). At late anaphase/telophase, the central spindle has begun to condense in the early stages of mid-body formation in
the wild-type cell (arrowhead, left-hand lower inset). No such structure is present in the polo mutant cells at a similar stage (arrow,
main left-hand lower panel). Immunostaining carried out by Maria Giovanna Riparbelli and Giuliano Callaini
Concluding remarks
Further understanding of the regulation, both in
time and space, of the interplay between the plks
and their docking proteins and/or substrates will
be important to fully appreciate the intricacies of
mitotic control. It seems likely that Drosophila will
continue to play a role in unravelling these networks.
It is an ideal organism in which to apply proteomics
to study the interactions of Polo with other proteins.
Its fully sequenced genome and the ease of applying
RNAi in cultured cells and in the whole organism
provides the opportunity to test the functions of
proteins identified in this way. Moreover, it still offers
the opportunity of applying classical genetic approaches
Oncogene
to identify genes that participate in regulatory networks.
Knowledge gained in this way should be readily
applicable to the study of the polo-like kinases in
human cells. This is of particular interest in the
long term as these enzymes, their regulators and
substrates are attractive drug targets in the treatment
of human cancer.
Acknowledgements
I would like to thank Cancer Research UK, MRC, BBSRC
and the EU for research funding, and all past and present
members of my research group and collaborators for their
contributions to this work. I am particularly grateful to
Monica Bettencourt-Dias for her comments on the manuscript.
Roles of Polo kinase in Drosophila
DM Glover
237
References
Adams RR, Tavares AA, Salzberg A, Bellen HJ and Glover
DM. (1998). Genes Dev., 12, 1483–1494.
Avides MC and Glover DM. (1999). Science, 283, 1733–1735.
Avides MC, Tavares AA and Glover DM. (2001). Nat. Cell
Biol., 3, 421–424.
Barbosa V, Gatt M, Rebollo E, Gonzalez C and Glover DM.
(2003). J. Cell Sci., 116, 929–941.
Barbosa V, Yamamoto R, Carpenter ATC, Henderson DS
and Glover DM. (2000). Genes Dev., 14, 3126–3139.
Barr FA, Sillje HHW and Nigg EA. (2004). Nat. Rev. Mol.
Cell Biol., 5, 429–440.
Bettencourt-Dias M, Giet R, Sinka R, Mazumdar A, Lock
WG, Balloux F, Zafiropoulos PJ, Yamaguchi S, Winter S,
Carthew R, Cooper M, Jones D, Frenz L and Glover DM.
(2004). Nature, in press.
Budde PP, Kumagai A, Dunphy WG and Heald R. (2001).
J. Cell Biol., 153, 149–158.
Cambiazo V, Logarinho E, Pottstock H and Sunkel CE.
(2000). FEBS Lett., 483, 37–42.
Carmena M, Riparbelli M, Minestrini G, Tavares AA, Adams
RR, Callaini G and Glover DM. (1998). J. Cell Biol., 143,
659–671.
Charles J, Jaspersen S, Tinker-Kulberg R, Hwang L, Szidon A
and Morgan D. (1998). Curr. Biol., 8, 497–507.
Clay F, McEwen S, Bertoncello I, Wilks A and Dunn A.
(1993). Proc. Natl. Acad. Sci. USA, 90, 4882–4886.
Davis FM, Tsao TY, Flowler SK and Rao PN. (1983). Proc.
Natl. Acad. Sci. USA, 80, 2926–2930.
de Carcer G, Avides MD, Lallena MJ, Glover DM and
González C. (2001). EMBO J., 20, 2878–2884.
Deak P, Donaldson M and Glover DM. (2003). J. Cell Sci.,
116, 4147–4158.
Descombes P and Nigg E. (1998). EMBO J., 17, 1328–1335.
Donaldson MM, Tavares AA, Ohkura H, Deak P and Glover
DM. (2001). J. Cell Biol., 153, 663–677.
Edgar BA and O’Farrell PH. (1990). Cell, 62, 469–480.
Elia AEH, Cantley LC and Yaffe MB. (2003a). Science, 299,
1228–1231.
Elia AEH, Rellos P, Haire LF, Chao JW, Ivins FJ, Hoepker
K, Mohammad D, Cantley LC, Smerdon SJ and Yaffe MB.
(2003b). Cell, 115, 83–95.
Fenton B and Glover DM. (1993). Nature, 363, 637–639.
Freeman M and Glover DM. (1987). Genes Dev., 1, 924–931.
Freeman M, Nusslein-Volhard C and Glover DM. (1986).
Cell, 46, 457–468.
Gatti M and Baker BS. (1989). Genes Dev., 3, 438–453.
Girdham C and Glover DM. (1991). Genes Dev., 5, 1786–1799.
Glover DM, Alphey L, Axton JM, Cheshire A, Dalby B,
Freeman M, Girdham C, Gonzalez C, Karess R, Leibowitz
M, Llamazares S, Maldonado-Codina MG, Raff J, Saunders
RDC, Sunkel C and Whitfield WGF. (1989). J. Cell Sci., 12,
277–291.
Glover DM, Leibowitz MH, McLean D and Parry H. (1995).
Cell, 81, 95–105.
Golan A, Yudkovsky Y and Hershko A. (2002). J. Biol.
Chem., 277, 15552–15557.
Gonzalez C, Sunkel CE and Glover DM. (1998). Chromosoma,
107, 452–460.
Grallert A and Hagan IM. (2002). EMBO J., 21, 3096–3107.
Herrmann S, Amorim I and Sunkel CE. (1998). Chromosoma,
107, 440–451.
Karess RE and Glover DM. (1989). J. Cell Biol., 109,
2951–2961.
Kitada K, Johnson AL, Johnston LH and Sugino A. (1993).
Mol. Cell. Biol., 13, 4445–4457.
Kotani S, Tugendreich S, Fujii M, Jorgensen P-M, Watanabe
N, Hoog C, Hieter P and Todokoro K. (1998). Mol. Cell, 1,
371–380.
Kotani S, Tanaka H, Yasuda H and Todokoro K. (1999).
J. Cell Biol., 146, 791–800.
Kumagai A and Dunphy W. (1996). Science, 273, 1377–1380.
Lane H and Nigg E. (1996). J. Cell Biol., 135, 1701–1713.
Lee K, Yuan Y-L, Kuriyama R and Erikson R. (1995). Mol.
Cell. Biol., 15, 7143–7151.
Lee KS, Greenfell TZ, Yarm FR and Erikson RL. (1998).
Proc. Natl. Acad. Sci. USA, 95, 9301–9306.
Lee LA, Van Hoewyk D and Orr-Weaver TL. (2003). Genes
Dev., 17, 2979–2991.
Leibowitz MH. (1990). PhD thesis, University of London,
UK.
Leismann O and Lehner CF. (2003). J. Cell Sci., 116,
2453–2460.
Lindon C and Pines J. (2004). J. Cell Biol., 164, 233–241.
Liu X, Zhou T, Kuriyama R and Erikson RL. (2004). J. Cell
Sci., 117, 3233–3246.
Llamazares S, Moreira A, Tavares AA, Girdham C, Spruce
BA, Gonzalez C, Karess RE, Glover DM and Sunkel CE.
(1991). Genes Dev., 5, 2153–2165.
Logarinho E and Sunkel CE. (1998). J. Cell Sci., 111,
2897–2909.
MacIver FH, Tanaka K, Robertson AM and Hagan IM.
(2003). Genes Dev., 17, 1507–1523.
May KM, Reynolds N, Cullen CF, Yanagida M and Ohkura
H. (2002). J. Cell Biol., 156, 23–28.
Mayer-Jaekel RE, Ohkura H, Gomes R, Sunkel CE,
Baumgartner S, Hemmings BA and Glover DM. (1993).
Cell, 72, 621–634.
Moritz M, Zheng Y, Alberts BM and Oegema K. (1998).
J. Cell Biol., 142, 775–786.
Moutinho-Santos T, Sampaio P, Amorim I, Costa M and
Sunkel CE. (1999). Biol. Cell, 91, 585–596.
Neef R, Preisinger C, Sutcliffe J, Kopajtich R and Nigg EA.
(2003). J. Cell Biol., 162, 863–875.
Nislow C, Kuriyama R and McIntosh JR. (1992). Nature, 359,
543–547.
Ohkura H, Hagan I and Glover DM. (1995). Genes Dev., 9,
1059–1073.
Renault AD, Zhang X-H, Alphey LS, Frenz LM, Glover DM,
Saunders RDC and Axton JM. (2003). Development, 130,
2997–3005.
Reynolds N and Ohkura H. (2003). J. Cell Sci., 116,
1377–1387.
Riparbelli MG, Callaini G and Glover DM. (2000). J. Cell
Sci., 113, 3341–3350.
Riparbelli MG, Callaini G, Glover DM and Avides M do C.
(2002). J. Cell Sci., 115, 913–922.
Sampaio P, Rebollo E, Varmark H, Sunkel CE and Gonzalez
C. (2001). Curr. Biol., 11, 1788–1793.
Saunders RDC, Avides M do C, Howard T, Gonzalez C and
Glover DM. (1997). J. Cell Biol., 137, 881–890.
Sunkel C and Glover DM. (1988). J. Cell Sci., 89, 25–38.
Shirayama M, Zachariae W, Ciosk R and Nasmyth K. (1998).
EMBO J., 17, 1336–1349.
Tanaka K, Petersen J, MacIver F, Mulvihill DP, Glover DM
and Hagan IM. (2001). EMBO J., 20, 1259–1270.
Tavares AA, Glover DM and Sunkel CE. (1996). EMBO J.,
15, 4873–4883.
Wakefield JG, Bonaccorsi S and Gatti M. (2001). J Cell Biol.,
153, 637–648.
Yarm FR. (2002). Mol. Cell. Biol., 22, 6209–6221.
Oncogene