Roles for Motifs of Cell Cycle Regulating Kinases
Beyond Substrate Selection of Individual Kinases
by
Jes Alexander
B.S., Biology and Chemistry
Stanford University, 2000
Submitted to the Department of Biology
in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy in Biology
at the
Massachusetts Institute of Technology
May 2008
C 2008 Jes Alexander. All rights reserved.
The author hereby grants to MIT permission to reproduce and to distribute publicly
paper and electronic copies of this thesis Aocuirent in whole or in part.
..
Department of Biology
May 14, 2008
C ertified by..........................................
Michael B. Yaffe
Associate Professor of Biology
Thesis Supervisor
Accepted by ...........
.........................
Stephen P. Bell
Professor of Biology
Chairman, Graduate Student Committee
MASSA
SI
OFTE
CINOOGy
AUG 1.2 2008
LIBRARIES
ARCHMES
Roles for Motifs of Cell Cycle Regulating Kinases
Beyond Substrate Selection of Individual Kinases
By
Jes Alexander
Submitted to the Department of Biology on May 14, 2008 in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy in Biology
ABSTRACT
Errors in the cell cycle, particularly during mitosis, have recently been implicated in
tumorigenesis and cancer formation. Several protein kinases, including the major mitotic
kinases Cdkl, Aurora A, Aurora B, Nek2, and Plkl, and the Polo-like kinase (Plk)
family, Plkl, Plk2, Plk3, and Plk4, are known to play important roles in the regulation of
mitosis and other phases of the cell cycle to prevent such errors. However, the
mechanisms underlying many of the roles of these kinases are unknown because the
appropriate substrates have not been identified. To aid in substrate identification and to
gain insight into the role of substrate specificity in the regulation of these kinases, we
have determined the motifs of the major mitotic kinases and the Plk family of kinases by
Positional Scanning Oriented Peptide Library Screening (PS-OPLS), which gives
complete, unbiased motifs. The PS-OPLS motif of Plkl revealed a previously unreported
specificity of Plkl leading to the identification of new candidate substrates, including
p31/Comet, which is involved in the silencing of the spindle assembly checkpoint for
anaphase onset. Additionally, the motifs and the localizations of the major mitotic
kinases lead us to put forward a model potentially explaining how the major mitotic
kinases, despite having overlapping localizations and access to the same sites on
substrates, phosphorylate distinct sites. This model, if true, would suggest that the motifs
and localizations of all of the major mitotic kinases coordinately direct the group of
kinases to the appropriate substrates. The motifs of the Plk family reveal that Plk2 and
Plk3 phosphorylate sites previously primed by phosphorylation by a tyrosine kinase,
suggesting that these kinases may integrate tyrosine kinase and serine/threonine kinase
signaling in a novel way. Finally, we find that Plk3 has a phosphorylation dependent
toggle switch which changes the motif of the kinase. The result suggests that the motif of
a kinase is not necessarily static, as was previously thought, and that the motif of certain
kinases may be modulated for different roles. These results give us deeper insight into
the function of kinases at the level of motifs, substrates, and the regulation of groups of
kinases.
Thesis Supervisor: Michael B. Yaffe
Title: Associate Professor of Biology
Table of Contents
Abstract
2
Acknowledgments
4
Chapter One
The Structural Basis for Kinase Substrate Specificity and the Known Roles of the
5
Major Mitotic Kinases and Polo-like Kinase Family in the Cell Cycle
Kinase Substrate Specificity
The Major Mitotic Kinases
The Polo-like Kinase Family
Summary and Preview
References
7
23
34
37
39
Chapter Two
The Determination of the Motifs of the Major Mitotic Kinases: Motifs and
Localizations of the Major Mitotic Kinases May Coordinately Regulate the
Kinases as a Group
Abstract
Introduction
Experimental Procedures
Results
Discussion
References
55
56
57
60
65
78
83
Chapter Three
Determination of the Substrate Specificities of the Polo-like Kinase Family
Reveals that Plk3 has a Motif that Changes Depending on the Phosphorylation
Status of the Activation Loop
90
Abstract
Introduction
Experimental Procedures
Results
Discussion
References
91
92
94
100
112
115
Chapter Four
Conclusions and Future Directions
118
Analysis of Motifs of Kinases Related by Function Versus those Related
by Sequence
119
Summary and Future Directions
121
References
130
Acknowledgments
I would like to thank all of the members of the Yaffe lab for their technical and
intellectual support, for generally making the lab a good place to work, and for putting up
with me over the years. I would like to particularly thank the other graduate students in
the lab, past and present, Sarah Bissonette, Drew Lowery, Duaa Mohammad, and Andrea
Tentner, who I have had the good fortune to get to know as we suffered through graduate
school together. Their technical, intellectual, and emotional support and friendship have
been immensely helpful -- Sarah Bissonette for doing and saying what needed to be done
and said, Drew Lowery for knowing every detail published about the cell cycle and for
sharing that information with me during discussions about my work, and Duaa for being
a good friend, discussing my work with me, providing technical help, and for entertaining
me with stories about her life. I would also like to thank all the postdocs in the lab, past
and present, Chris Ellson, Scott Floyd, Alexandra Gardino, Jeong-Ho Hong, Kazu Kishi,
Dan Lim, Rune Linding, Coky Nguyen, Gerry Ostheimer, Michael Pacold, Christian
Reinhardt, Erik Wilker, and Marcel van Vugt, for their time and effort in teaching me and
creating an environment conducive to learning -- in particular, Dan Lim for intellectual
and technical support and putting up with my ranting, Christian Reinhardt for working
with me on the phospho-dependent binding domain screen and his extra patience and
effort in teaching me, Michael Pacold and Scott Floyd for career advice, Chris Ellson for
being a good friend, discussing my work with me, telling the truth when it needed to be
said, giving technical assistance (also known as cloning), and being generally supportive.
Special thanks goes to Mary Stewart who keeps the lab running and without whom I
would never have been able to complete a single experiment. Finally, I would like to
thank Mike for hiring some good people, procuring money to fund my experiments, and
for teaching me what to do and what not to do if I am ever as fortunate as he has been and
have the opportunity to head my own lab.
Outside the lab, I would like to thank Ben Turk for helping us get the PS-OPLS
technique running in the Yaffe lab, Karl Clauser in Steve Carr's lab for his valiant efforts
in attempting to find Plk3 substrates, and Eric Weiss for a fruitful collaboration. I would
also like to thank my prelim and thesis committee including Steve Bell, Paul Chang, Ray
Erikson, Frank Gertler, Terry Orr-Weaver, and Peter Sorger for their time and effort,
helpful suggestions, and for pushing me to become a better scientist through their
insightful and needed critique. Additionally, my undergraduate research advisor Doug
Brutlag and my research advisor at the NCI, John Weinstein, were so encouraging. I
would like to thank them for their role in getting me to this point.
Most importantly, I would like to thank all the people outside of the biology
community who have been so supportive and have made me happy despite all my painful
experiences in graduate school. I would like to thank the University of Kansas men's
basketball team who won the NCAA National Championship this year (2008). Following
them for the last twenty years through the failures and disappointments to their run
through the tournament and to the championship this year has not only made me happy,
but has been a huge inspiration to me. More importantly, my parents and sister have been
extremely understanding and have supported me in every way they could, as they have in
all my endeavors. And last, but far from least, my biggest debt of gratitude is to my wife,
Alexis, who, despite being a medical resident, has found time to support me through the
difficult times in graduate school and in life.
Chapter One
The Structural Basis for Kinase Substrate Specificity and the Known
Roles of the Major Mitotic Kinases and Polo-like Kinase Family in the
Cell Cycle
Contributions:
I wrote all sections. I created Figures 1.1, 1.6, and 1.7. Other figures were adapted from
the sources cited.
The goal of the eukaryotic cell cycle is the replication of the genetic material and
division of the genetic material and other contents of the cell to two daughter cells. The
cell cycle is divided into four phases, G 1, S, G2, and M. M phase includes two separate,
but coordinated processes, mitosis and cytokinesis. M-phase is particularly complicated
as a series of events during mitosis, including centrosome separation and maturation,
spindle assembly, chromosome segregation, and spindle disassembly, and during
cytokinesis must occur in rapid succession without any errors. Errors in M-phase result
in genetic instability and chromosomal aberrations, which are hallmarks of cancer and
may be initiating events in tumorigenesis (1, 2).
To prevent errors, the events within each phase of the cell cycle are regulated by,
among other things, the carefully timed expression and destruction of proteins, formation
of specific protein-protein interactions, and transduction
of signals.
Protein
phosphorylation, performed by protein kinases, is the most common mechanism
employed by the cell for signal transduction and the cell cycle is not an exception. There
are an estimated 700,000 potentially phosphorylatable sites at any given time in a
eukaryotic cell (3). Each protein kinase chooses specific sites on specific substrates.
While the actual number of sites may be fewer than the estimated 700,000 since some of
these sites may not be surface accessible, each kinase must select which sites to
phosphorylate from a large number of inappropriate sites.
Clearly, the substrate
specificity of each kinase is important for the fidelity of processes such as the cell cycle.
Understanding the regulation and substrate specificity of kinases involved in the
cell cycle and, in particular mitosis, should give insight into the process of tumorigenesis
and may lead to modes of prevention or treatments of cancer (1, 2). This thesis examines
the substrates specificities of two sets of kinases (Figure 1.1). The major mitotic kinases,
Cdkl, Aurora A, Aurora B, Nek2, and Plkl, are a functional family with all members
playing significant roles regulating the processes of mitosis (4). The other set of kinases,
the Polo-like kinase (Plk) family, which includes Plkl, Plk2, Plk3, and Plk4, are related
by sequence, but have functions associated with different phases of the cell cycle (5).
Knowledge of the substrate specificity of these kinases will aid in the discovery of their
substrates, allowing a deeper understanding of the roles of these kinases in the cell cycle.
Additionally, the substrate specificities of these sets of kinases may give us other insights
Centrosome
Separation
SMajor
Chromatin
Condensation
Mitotic Kinases:
Cdkl 0
Aurora A 0
B
(2 0
k1 0
o-like
ases:
kl •
k2
k3 0
k4 0
Figure 1.1: Roles of the Major Mitotic Kinases and Polo-like Kinases in the cell
cycle. Each kinase is represented by a colored dot as indicated on the right side of the
figure. Schematics of a cell at each stage of the cell cycle are represented and the
phases are color coded, orange for M-phase and red for Gl-, S-, and G2-phases.
Spindle Assembly Checkpoint is abbreviated SAC and microtubule is abbreviated MT.
into mechanisms of substrate specificity involved in the regulation of the processes in
which these kinases function and how substrate specificity differs between kinases
related by function and those related by sequence.
Kinase Substrate Specificity
There are 518 protein kinases in the human genome (6). These kinases can be
divided into seven major groups based on sequence similarity with many kinases not
falling into any of these groups. The major mitotic kinases, which cooperate functionally
B.
Figure 1.2: Phylogeny of human kinases based on sequence similarity of the kinase
domains with the seven major groups of kinases labeled in small lettering. A. the
Major Mitotic Kinases, Cdkl, Aurora A, Aurora B, Nek2, and Plki are labeled in large
lettering and indicated by red circles and B. The Polo-like kinase family, Plkl, Plk2,
Plk3, and Plk4, are indicated by the red oval. This figure was adapted from (6).
to regulate mitosis, show significant sequence diversity and do not cluster to a single
phylogenetic class. Cdkl is a member of the CMGC kinases, Aurora A and Aurora B are
very similar to AGC kinases, Nek2 is most similar to the CAMK group, and Plk1 is most
similar to the AGC kinases (Figure 1.2A). In contrast to the major mitotic kinases,
members of the Polo-like family of kinases, which seem to have different roles in distinct
parts of the cell cycle, have very high sequence similarity and cluster in the phylogeny
near the AGC kinases (Figure 1.2B). Despite the clustering of kinases into different
groups based on sequence, all protein kinases catalyze the same reaction: transfer of the
y-phosphate from a molecule of ATP to the hydroxyl group of a serine, threonine, or
tyrosine residue of the substrate protein. For phosphorylation of a substrate to occur, the
site to be phosphorylated on the substrate must bind in the active site cleft of the kinase
while the kinase is active. For most kinases, the phosphorylation reaction is regulated at
a minimum of three levels: kinase activation, recognition of the sequence surrounding the
phosphorylatable residue by the active site cleft, and co-localization of the kinase with
the substrate. These aspects are discussed herein.
Kinase Domain Structure and the Structural Basis of Kinase Activation
The first X-ray crystal structure of a kinase domain was that of the AGC kinase
group member cAMP-Dependent Protein Kinase also known as Protein Kinase A (PKA)
(7). Subsequent crystal structures of other kinase domains have shown that the protein
kinase domain fold is highly conserved (7-9). Kinase domains are bi-lobed with a
smaller NH2-terminal lobe (N-lobe) and a larger COOH-terminal lobe (C-lobe) (Figure
1.3). The N-lobe is primarily responsible for positioning ATP for y-phosphate transfer
and consists of a five-stranded P-sheet (composed of strands p1 to P5) and two a-helices
including aC. The loop between pl and P2 is the phosphate binding loop (P-loop). The
C-lobe is primarily responsible for peptide substrate binding and phospho-transfer. It is
mostly a-helical containing six a-helices including aEF and aF and four small P-strands
including P6 through 09. The loop between 36 and 37 is the catalytic loop directly
involved in phosphate transfer and the loop between P8 and 09 is the Mg 2+ binding loop.
Between the N-lobe and C-lobe is the active site cleft, where peptide and ATP are bound
and catalysis occurs.
N lobe
*Il~ 1·
02,
Lys 72
;J
C4
97
Asp 14
n!
Pt
Inhibitor
C lobe
Figure 1.3: Ribbon representation of the X-ray crystal structure of activated PKA in
complex with the PKI pseudo-substrate. The activation segment is in red, the catalytic
loop is in green, the P-loop is in orange, the aC is in purple, and PKI is in yellow. The
invariant catalytic residues, the activation loop phosphoresidue, and the lobes are
labeled. This figure is adapted from (16).
Several residues are invariant because of their important roles in catalysis (Figure
1.3 and 1.4). Using the amino acid numbering scheme for PKA, the ATP phosphates are
positioned by a glycine rich region in the P-loop; a Mg 2+ ion, which bridges the 0- and yphosphates and is stabilized by D184 in the Mg 2+ loop; and K72 in 13, which forms an
hydrogen bond with 0- and y-phosphates and is itself localized by a salt-bridge
interaction with E91 on aC. The Mg 2+ ion interacts with the P- and y-phosphates of the
bound ATP molecule and is required for phospho-transfer. D166 is the residue directly
involved in phosphate transfer. D166 was originally thought to act as a base to remove
the proton from the hydroxyl group on the substrate serine or threonine or from the
phenol group on the substrate tyrosine allowing the resulting alkoxide or phenoxide anion
to perform a nucleophilic attack on the ATP y-phosphate resulting in phospho-transfer
(10, 11). However, the current thinking is that D166 is not only a base catalyst, but
Figure 1.4: Schematic representation of the invariant catalytic residues of PKA, also
showing the RD arginine R165. This figure is from (11).
instead may play a major role in positioning the substrate serine or threonine hydroxyl or
tyrosine phenol for attack on the ATP y-phosphate (12). D166 itself is positioned by
hydrogen bonding interactions from N171 also from the catalytic loop.
Most kinases can attain at least two conformational states, a minimally active
state and a maximally active state. Conformational changes in the activation segment
cause the kinase to switch between these states by positioning the catalytic residues
11
Catalytic
Loop
6. 1~e
R
Figure 1.5: Schematic representation of the activation segment showing the parts of
the segment including the DFG and APE tripeptides, P+1 loop, and activation loop.
This figure is adapted from (15).
appropriately for phospho-transfer (7, 13, 14). The activation segment runs from a highly
conserved tripeptide, 184-DFG-186, in the middle of the Mg 2 + binding loop to another
conserved tripeptide, 206-APE-208, which is at the junction of the N-terminus of ucEF
and the loop preceding it, the P+l loop (Figure 1.5). Within the activation segment, in
order from N-terminal to C-terminal, are the Mg 2+ binding loop, 19, the activation loop,
and the P+1 loop, named for its interaction with the amino acid one position C-terminal
to the phospho-acceptor residue, although it interacts with other positions as well.
In many kinases, activation requires phosphorylation on serine, threonine or
tyrosine residues on the activation loop (15).
autophosphorylation
or by
phosphorylation
This phosphorylation can occur through
by
a
different
kinase.
In the
unphosphorylated state, the activation segment often collapses into the active site cleft
blocking access to ATP and the peptide substrate. Phosphorylation of the activation loop
can occur at one or more residues depending on the kinase, but in most cases one of the
sites plays a dominant role
(T197 in PKA)
in the
conformational
change.
Phosphorylation on the activation loop causes refolding of the activation segment by
formation of interactions between the phosphoresidue and a group of amino acids
composing a basic pocket (7, 13, 14). In PKA, pT197 interacts with H87 on ctC, R165
Figure 1.6: Structure of Protein Kinase A (PKA) (PDB ID: 1ATP) showing the
interaction between the phosphoresidue on the activation segment and the basic
pocket. Basic pocket residues (green) and the activation segement phosphothreonine
(yellow) are shown in stick representation. The activation segment (red) and the rest
of the protein (white) are shown in the backbone representation. In this figure, the Nlobe as at the top and the C-lobe is at the bottom and the structure, as compared to
Figure 1.3, has been rotated 1800 around an axis running from N-lobe to C-lobe.
13
on the catalytic loop, and K189 on f39 (Figure 1.6).
All kinases that require
phosphorylation for activation have an arginine (R165) just preceding the invariant
catalytic aspartate (D166) (Figure 1.3, 1.4, 1.6) (11). Kinases with this RD sequence are
known as RD kinases. Interaction between pT197 and R165 seems to have little effect on
the positioning of the catalytic loop, but is likely to be important for positioning of the
activation segment and in particular the DFG tripeptide (15). Interaction with K189 also
seems to be involved in positioning the DFG tripeptide (7, 13, 14). Interaction with H87
plays a role in positioning the N- and C-lobes with respect to each other through
positioning of the aC. In the inactive state of the kinase, electrostatic repulsion between
the amino acids of the basic pocket is thought to prevent the activation segment from
achieving the active state conformation (15).
Neutralization of the basic pocket by
interaction with the dianionic pT197 allows the activation segment to take its active state
conformation and facilitates activation of the kinase.
Upon phosphorylation, the conformational changes associated with the interaction
of the phosphoresidue with the basic pocket are propagated through the kinase resulting
in the activation segment moving out of the active site cleft, repositioning of the DFG
tripeptide sequence, and positioning of the catalytic residues allowing efficient phosphotransfer. Repositioning of the 184-DFG-186 tripeptide localizes D184 for chelation of
the Mg2+ ion which is required for positioning of the phosphates of ATP for catalysis (7,
11, 13, 14, 16). Additionally, F185 forms hydrophobic interactions with amino acids on
aC, which are reinforced by interactions between the pT197 and H87 on aC. These
interactions with aC are important for positioning of the invariant catalytic E91, which is
on aC and forms an ionic interaction with the invariant catalytic K72. As mentioned
above this interaction is required to localize K72 for its role in positioning the phosphates
of ATP. In serine/threonine kinases, phosphorylation on the activation loop also results
in a hydrogen bonding interaction between the P+1 loop conserved serine or threonine
(T165) and residues in the catalytic loop. Though not part of the activation loop, the loop
between aEF and aF seems to play a role in stabilization of the conformation of the
activation loop in the active state of the kinase. Together these conformational changes
activate the kinase.
Residues other than the RD arginine of the basic pocket are not as highly
conserved, even in kinases that require phosphorylation for activation (15). For example,
in Aurora A, another kinase activated by phosphorylation on the activation loop, the
equivalent residue to K189 is not basic (17). Instead, a lysine on the activation loop
interacts with the activation loop phosphorylation sites. The basic pocket itself, however,
is highly conserved, even in kinases not requiring phosphorylation for activation.
Glycogen Synthase Kinase 3 beta (Gsk33) and Casein Kinase 1 (CK1) are examples (1820). The amino acid in the equivalent position to T197 is a valine in Gsk3P and an
asparagine in CK1, making phosphorylation on the equivalent site impossible. GSK3P
has a basic pocket composed of the RD arginine, a residue from atC, and a residue from
P9; CK1 has a basic pocket composed of the RD arginine and a residue from P9. The
crystal structures of both kinases have been solved and in both cases the basic pocket was
found to be occupied by an anion from the crystallization solution. Crystal structures of
Gsk3P have also been solved in which the basic pocket is empty (21). In these structures,
both the activation segment of Gsk3P and the basic pocket are in the active conformation
confirming that Gsk3P is active without phosphorylation on the T197 equivalent site or
an anion in the basic pocket. The structure also implies that the basic pocket is open to
be filled by an anion. Both of these kinases phosphorylate substrates that have previously
been phosphorylated (primed) by another kinase on a nearby residue.
Gsk33
phosphorylates a serine or threonine four amino acids N-terminal to a phosphoserine
(22).
CK1 phosphorylates a serine or threonine three amino acids C-terminal to a
glutamate, aspartate, phosphoserine or phosphothreonine (23). Taken together, these data
suggest that a phosphoresidue on the substrate can fit into the basic pocket, implying that
the basic pocket is involved in substrate selection rather than kinase activation in these
kinases. It should be noted that no kinase has been described for which the required
priming phosphorylation is catalyzed by a tyrosine kinase.
Structural Basis for Kinase Substrate Specificity
The original studies of kinase substrate specificity were performed on PKA in the
mid-to-late 1970's and showed that primary structure of the substrate was primarily
responsible for the choice of sites phosphorylated (24). At each position near the
phosphoacceptor residue, certain amino acids increase the likelihood that a kinase will
phosphorylate the site while others reduce the likelihood. The consensus sequence is a
way of representing which sites a kinase is likely to phosphorylate. There are two types
of consensus sequences, minimal consensus sequences and optimal consensus sequences.
The minimal consensus sequence delineates which amino acids in which positions are
required for phosphorylation. The optimal consensus sequence delineates which amino
acids in which positions result in a site that is highly favored by the kinase. In either
case, each position of the sequence is represented as a list of amino acids separated by
slashes ("/"). The amino acids in each list are considered to have similar effects on the
likelihood that the kinase will phosphorylate the site. An X is used when there is no
significant difference among the amino acids in their effect on the likelihood of
phosphorylation.
A motif is a broader concept including consensus sequence, but
expanding it to include all residues in each position affecting the likelihood of
phosphorylation, both positively and negatively. The PKA optimal consensus sequence
was determined to be R-R-X-S/T-(p, where X is any amino acid and (pis a hydrophobic
amino acid (24). By convention, the phosphoacceptor site (S/T) is position 0, positions
N-terminal to this are numbered with more negative numbers going away from position 0
(e.g. X is -1) and positions C-terminal to the phosphoacceptor are numbered with more
positive numbers going away from position 0 (e.g. 9 is +1).
An understanding for how specificity is mediated at the molecular level was not
gained until kinetic studies on PKA alanine scanning mutants were performed and the Xray crystal structure of PKA in complex with the PKI inhibitor peptide was solved in the
early 1990's (25-27) This structure has served as a common reference point for the
analysis of all subsequent kinase structures. The sequence of the inhibitor in the region
of interaction with the active site cleft of the kinase, with the exception of an alanine
instead of a serine or threonine at what would be the phosphoacceptor site, is the same as
the optimal sequence. The crystal structure of the complex shows that the substrate binds
in an extended conformation in the active site cleft, explaining the importance of primary
structure (Figure 1.2, Figure 1.7), rather than secondary or tertiary structure, in
specificity. Additionally, there are clear interactions between amino acids composing the
surface of the active site cleft and the amino acids of the inhibitor peptide, particularly the
Figure 1.7: Structure of PKA in complex with the PKI peptide (PDB ID: 1ATP)
showing the amino acids in the PKA active site cleft that interact with the amino acids
in the bound pseudo-substrate peptide.
PKA is shown in the space filling
representation. Amino acids E127, E170, and E230, which interact with the arginines
in the -3 and -2 positions on the peptide are shown in green, as are L198 and L205,
which interact with the isoleucine in the +1 position of the peptide. The rest of the
kinase is shown in white.
The peptide (orange) is shown in the backbone
representation, except for the two arginines and the isoleucine mentioned above,
which are shown in stick representation.
17
arginines at positions -3 and -2 and the hydrophobic amino acid at position +1. The R in
the -3 position forms electrostatic interactions with E127 and E170 on the surface of the
active cleft formed by the N-lobe and the R in the -2 position forms an electrostatic
interaction with E230 in aF in the C-lobe.
The hydrophobic amino acid in the +1
interacts with a hydrophobic pocket formed by L198 and L205 in the P+l loop of the
activation segment.
Interestingly, Cyclin Dependent Kinases (CDKs) and Mitogen Activated Protein
Kinases (MAPKs), members of the CMGC group, have a very different specificity in the
+1 position.
These kinases only phosphorylate sequences with a proline in the +1
position. This is in contrast to kinases of the AGC and CAMK groups, Aurora A kinase,
as well as probably others, which will not phosphorylate any sequence with proline in the
+1 position. The crystal structure of activated Cdk2 in complex with Cyclin A3 and a
substrate peptide has been solved and seems to explain the rare specificity of prolinedirected kinases (8). Proline is unique in that its backbone nitrogen cannot hydrogen
bond due to the cyclic nature of the amino acid. The cyclic nature of the amino acid also
results in kink in any peptide chain containing proline. The authors suggested that any
other amino acid in the +1 position would result in an unsatisfied hydrogen bond due to a
lack of a hydrogen bonding partner from the amino acids of the kinase in the vicinity of
the +1 position.
Such an unsatisfied hydrogen bond would make any amino acid other
than proline much less favorable in the +1 position. However, this selectivity may also
be due to the kink in the peptide chain caused uniquely by proline.
Subsequent crystal structures of other kinases with selectivity in the -3, -2, and +1
positions have demonstrated the role of residues homologous to the ones described above
in the selectivities of these kinases (28). The structures have also shown that for most
kinases, only residues within four or five positions on either side of the phosphoacceptor
site on the substrate interact with the active site cleft, which explains why the residues
involved in kinase-substrate recognition are commonly found within four of five
positions of the phosphoacceptor (3, 8, 27). However, selectivity at positions farther out
can occur through interactions with the outer surface of the kinase domain (27). Other
determinants of specificity also exist for specific kinases.
As mentioned previously,
priming phosphorylation can interact with the basic pocket in the active site cleft of
certain kinases.
In general, the active site cleft is complementary to the substrate
sequence allowing interactions based on charge, hydrogen bonding, or hydrophobicity,
and this complementarity defines the kinase specificity.
Techniques for Kinase Motif Determination
The original approach used to determine the specificity of PKA employed kinetics
studies on a series of peptides based on a site found to be phosphorylated by PKA in
pyruvate kinase (24, 29). The peptides were designed such that each had a different
single amino acid mutation.
This technique employs the comparison of affinity and
reaction velocity for each of the different peptides, from which the effect on specificity of
different amino acids in each position can be inferred. This technique is particularly
useful for determining a minimal consensus sequence, as this just requires alanine
scanning. It is less useful if a substrate site is not known as a site must be determined
prior to employing this technique. It is also less useful if the goal is to determine an
optimal consensus sequence or a motif, as to do this correctly requires kinetic studies on
peptides with single mutations where the mutations scan through each amino acid at each
position of the sequence. This can amount to hundreds or thousands of experiments. If
all amino acids are not tested in each position, the resulting optimal consensus sequence
or motif may be incomplete and, therefore, biased.
To overcome these problems the technique of oriented peptide library screening
(OPLS) was developed in the mid 1990's (30). In this technique, an in vitro kinase assay
is performed on a degenerate library of peptides instead of a peptide with a single
sequence.
The library consists of peptides of the same length each, with a fixed
phosphoacceptor residue in the center, which orients the library.
All other positions in
the library are composed of a mixture of amino acids. There are usually four to seven
such degenerate positions on each side of the phosphoacceptor and each usually includes
an equal mixture of all 20 amino acids except for serine, threonine, and tyrosine, as these
would disorient the library, and cysteine, as this might result in disulfide cross-links
between peptides.
After the reaction with the kinase, those peptides that were
phosphorylated are separated from the unphosphorylated peptides and sequenced by
Edman Degradation.
The percentage of each amino acid at each position of the
phosphorylated peptides is determined from these results. The ratio of the percentage of
a given amino acid in a given position to the percentage of that amino acid in that
position from the original library gives the relative enrichment of that amino acid and is a
measure of the selectivity of the kinase for that amino acid in that position. To further
examine specificity, the same technique can be repeated with a library containing, in
addition to the fixed phosphoacceptor residue, a second fixed position corresponding to
the highest selectivity determined with the previous library. This amplifies the selectivity
ratios of less highly selected positions to better distinguish between signal and noise.
OPLS advantages include not requiring a known substrate site, requiring much
less work to achieve an unbiased motif, and determination of selectivity at each position
independently of all other positions. This technique has been successfully employed on
many kinases (30-36). However, it has a few important drawbacks.
The process of
sequencing is error prone because of preview and lag, which are caused by excessive or
incomplete peptide cleavage during Edman Degradation, as compared with the bulk of
the peptides. When this occurs around a position of high selectivity for a given amino
acid, the positions just prior to (preview) and just after (lag) will often show high
selectivity for the same amino acid. As a result some subjectivity is required to decide
which positions are showing real selectivity and which are showing artifacts of the
sequencing reaction. This problem extends to positions in which certain amino acids are
strongly selected against. Preview and lag will have the effect of covering up the extent
of negative selection, making interpretation of residues mediating negative selection
difficult and somewhat unreliable. Finally, this technique does not allow for testing the
role of priming phosphorylation in specificity. To do this would require the addition of
phosphoamino acids to the peptide library. However, phosphoamino acids would make it
impossible to distinguish between the peptides phosphorylated by the kinase and those
not phosphorylated by the kinase since all peptides would have phosphoresidues.
A more advanced technique based on OPLS, positional scanning oriented peptide
library screening (PS-OPLS), which was published in 2004, addresses the deficiencies in
OPLS (37). In this technique instead of a single in vitro kinase assay on single oriented
peptide library, in vitro kinase assays on 198 separate oriented peptide libraries are
performed in parallel (Figure 2.2). The stock peptide libraries are arranged in an array
with 9 rows and 22 columns.
Each peptide library is biotinylated and has two fixed
positions, the phosphoacceptor position and a position for testing kinase specificity.
Scanning across the columns puts all 20 amino acids plus phosphotyrosine and
phosphothreonine in the second fixed position.
Scanning down the rows moves the
second fixed position from -5, relative to the phosphoacceptor S/T, to +4.
At the
intersection of a column and a row is a library with the column's designated amino acid
fixed at the row's designated position and with all other positions degenerate except the
S/T phosphoacceptor site. The degenerate positions contain a mixture of all 20 amino
acids except cysteine. The kinase reactions are also arranged in 9 rows and 22 columns
and the libraries are added to the reactions in parallel maintaining their positions in the
array. The kinase assays are performed in the presence of y- 32P-ATP. After the assays
are performed, a portion of each reaction is transferred in parallel to a streptavidin
membrane to which the peptides bind. The membrane is washed to remove the unreacted
y-32P-ATP and then visualized by autoradiogram or phosphorimager. The darker the spot
at the intersection of a given column and given row, the more the substrate binding cleft
selects for the amino acid designated by the column at the position designated by the row.
The resulting data can be quantitated by densitometry.
PS-OPLS has been used successfully on several kinases (31, 37-44). A drawback
of the technique is that it requires significantly more kinase protein than is required for
OPLS since 198 reactions are being performed instead of just one. However, PS-OPLS
does not have the problems associated with the sequencing artifacts of OPLS. Therefore,
no subjectivity is required in determining which amino acids in which positions are
selected, and, importantly, this technique also allows the reliable determination of
negatively selected amino acids. In addition, since peptides phosphorylated in the kinase
reactions
do not need to be separated
from the unphosphorylated
peptides,
phosphoresidues can be included in the libraries. This allows testing for a role of priming
phosphorylation in specificity. Based on these advantages, we chose to use PS-OPLS for
the motifs determined in this work.
Other Determinants of Substrate Specificity
Whereas the motif of a kinase decides which sites on a substrate can be
phosphorylated and, therefore, makes the final determination as to whether a site will be
phosphorylated, phosphorylation requires that the kinase co-localize with the substrate
while the kinase is active.
This co-localization requirement is particularly important
since the binding of peptides in the kinase active site cleft has dissociation constants often
in the hundreds of micromolar to millimolar range (45). To compensate for low affinity,
either high local concentrations of the kinase in the vicinity of the substrate or secondary
binding sites are necessary to enable phosphorylation. To increase local concentrations,
determinants of substrate specificity other than motifs are often employed.
Different kinases use different determinants.
Some kinases recognize substrate
docking motifs, which are separate from the site to be phosphorylated, through kinase
domain interaction surfaces, which are separate from the active site cleft.
MAPKs
recognizing D domains on their substrates are examples of this (46). Some kinases have
targeting domains. Non-receptor tyrosine kinases often contain Src Homology 2 (SH2)
and/or Src Homology 3 (SH3) domains, which bind phosphotyrosine containing
sequences and proline-rich sequences in substrates, respectively (47). Polo-like kinase 1
(Plkl) has a phospho-dependent binding domain, the Polo-Box domain (PBD), that binds
S-pS/pT motifs (48). Two models have been proposed for the role of the PBD in Plkl
substrate interactions (49).
In the Processive model, the PBD directly binds Plkl
substrates; in the Distributive model, the PBD binds a protein which situates Plkl in close
proximity to other proteins which are substrates. Evidence for both modes of substrate
recognition exist (49). Some kinases have targeting subunits. The targeting subunits of
Cdks are Cyclins, with which they form stable heterodimers, as their name suggests.
Cyclin A directs Cdk2 to many of its substrates through interactions between a
hydrophobic patch on the Cyclin and an RXL motif on these substrates (50).
Some
kinases are localized to particular subcellular compartments or structures. Interaction
through an RXL motif seems to be dispensable for substrate recognition by Cdkl/Cyclin
B (51).
However, Cyclins are also responsible for Cyclin/Cdk complex localization.
Cdkl/Cyclin B1 is excluded from the nucleus prior to mitosis by a nuclear export
sequence (NES) and cytoplasmic retention sequence (CRS) on Cyclin B (52). During
early mitosis, phosphorylation on the CRS inactivates it causing rapid translocation of
Cdkl/Cyclin BI into the nucleus where its substrates are (53-55). In all cases, the local
concentration of kinase in the vicinity of substrate is increased, satisfying the spatial colocalization requirement and elevating the likelihood of phosphorylation (3).
The Major Mitotic Kinases
Mitosis is the part of the cell cycle in which the nuclear contents, particularly the
DNA which was replicated in S-phase, is divided between the emerging daughter cells.
This process relies on two major structures, the centrosome and the mitotic spindle. The
centrosome is the microtubule organizing complex (MTOC) and during the cell cycle it
undergoes replication in preparation for its role in mitosis. Prior to mitosis, cells contain
a duplicated centrosome consisting of two mother-daughter pairs of centrioles tethered
together and surrounded by pericentriolar material (PCM), which nucleates microtubules.
At the G2/M transition, the tether is severed and the centrosomes begin to move apart to
opposite ends of the cell. During this process, the centrosomes mature, meaning that
proteins such as y-tubulin and other regulators of microtubule assembly are recruited to
the PCM to enable the formation of the mitotic spindle. In the first phase of mitosis,
prophase, the chromosomes condense.
Prometaphase begins with breakdown of the
nuclear envelope allowing attachment of the chromosomes to the newly formed mitotic
spindle. Microtubules of the mitotic spindle attach to kinetochores, protein complexes
which bind to the centromeric DNA regions on the chromosomes. The kinetochore on
each chromosome must attach to microtubules from each centrosome to form bivalent or
amphitelic attachments.
Until this occurs, the spindle assembly checkpoint prevents
chromosome segregation.
When all kinetochores have formed attachments to
microtubules from each centrosome, the chromosomes align on the metaphase plate and
the cell has reached metaphase. At this point, the E3 ubiquitin ligase anaphase promoting
complex/cyclosome (APC/C) ubiquitinates Securin, which holds the sister chromatids
together, and Cyclin B causing their destruction and allowing anaphase onset. During
anaphase, the sister chromatids move toward the centrosomes as the centrosomes move
apart. In telophase, the chromosomes reach the opposite ends of the cell and decondense
and the nuclear envelope reassembles forming the two new nuclei of the daughter cells.
The major mitotic kinases, Cdkl, Aurora A, Aurora B, Nek2, and Plkl regulate
these events through carefully timed phosphorylation of specific sites on substrates. Each
of these events is regulated by more than one of these kinases and these kinases have
overlapping localizations, often even interacting with and phosphorylating the same
substrate proteins on different residues. What follows are brief descriptions of what is
known about each of these kinases including how they are activated, where they localize,
what their substrate specificity is, and what their known roles are with emphasis on
known substrates.
Cdkl
Cyclin-Dependent Kinases (Cdks) are dependent on cyclins for their activation
and targeting, as their name suggests. Cyclin-Cdk complexes are the master regulators of
the cell cycle with different cyclin-Cdk complexes controlling different parts of the cell
cycle. This regulation is based on carefully timed expression, activation, and destruction
of the cyclins and other regulators of these complexes. Vertebrates have the Gi-phase
cyclin D, the GI/S-transition cyclin E, the S-phase cyclin A, and the M-phase cyclin B.
Cdkl, 2, 4, and 6 have the same kinase motif, S/T-P-X-K/R, but K/R in the +2 seems not
to be necessary, particularly for Cdkl (8, 30, 36). Unlike other Cyclin-Cdk complexes
that also rely on an interaction between a hydrophobic patch on the cyclin and an RXL
motif on the substrate, Cdkl/Cyclin B seems not to require this for substrate selection
(51). The Cdkl/Cyclin B complex is the master regulator of mitosis and seems to play a
role in almost every aspect of mitosis.
Activation of Cdkl/Cyclin B, which occurs at the centrosome, marks entry into
mitosis (56).
Cyclin B expression begins in S-phase and protein levels peak at the
beginning of mitosis (57). Cyclin B binds to Cdkl on the aC helix which contains the
PSTAIRE sequence recognized by cyclins for binding.
This binding results in a
conformational change to a partially active state. Phosphorylation on the activation loop
at T160 by the Cdk activating kinase (CAK) is also required for full activation of the
kinase (8, 58).
However, prior to mitosis, Cdkl is inhibited by a second set of
phosphorylations on T14 and Y15 of the P-loop generated by Mytl and Weel kinases
(59, 60).
These phosphorylations inhibit the kinase by blocking access to the ATP
binding pocket through electrostatic repulsion with the phosphates of ATP. At the onset
of mitosis, these phosphorylations are removed by the phosphatases Cdc25B and Cdc25C
(61).
A small amount of active Cdkl/Cyclin B phosphorylates Weel, Cdc25B and
Cdc25C decreasing the activity of Weel, increasing the activities of Cdc25B and
Cdc25C, and generating a positive feedback loop which rapidly activates the rest of the
pool of Cdkl/Cyclin B (62, 63). As discussed later, Plkl also contributes to mitotic entry
through a role in generating this positive feedback loop.
Once active, Cdkl/Cyclin B activity plays a role in centrosome separation and the
resultant bipolar spindle.
At least one substrate of Cdkl/Cyclin B involved in this
process is known. HsEg5 is a kinesin-related motor protein that associates with the
centosome in mitosis, but not in interphase. Its localization to the centrosome at mitosis
is required for centrosome separation. Disruption of its localization results in monastral
spindles. Cdkl/Cyclin B triggers localization to the centrosome by phosphorylation on a
conserved site in the C-terminus (64).
Relocalization of Cdkl/Cyclin B allows it to regulate other mitotic processes.
There are three B-type cyclins. Cyclin B1 and B2 are ubiquitously expressed, whereas
Cyclin B3 seems to be restricted to the testis (65, 66). Cdkl binds Cyclin B 1 and Cyclin
B2 to form two separate complexes that perform different roles in mitosis based on their
different localizations (67). Prior to mitosis, both Cdkl/Cyclin BI and Cdkl/Cyclin B2
are sequestered in the cytoplasm by a cytoplasmic retention sequence (CRS) and a
nuclear export sequence (NES) (52). Cdkl/Cyclin B2 remains outside the nucleus first
localizing to the golgi where it phosphorylates GM130 for mitotic golgi fragmentation
(67, 68). Later in prophase, Cdkl/Cyclin B2 localizes diffusely in the cytoplasm (69).
Cdkl/Cyclin BI becomes phosphorylated on its NES and CRS causing abrogation of
nuclear export and enhanced nuclear import resulting in translocation of Cdkl/Cyclin B
into the nucleus (53-55). Once inside the nucleus, Cdkl/Cyclin Bl has several different
roles. It phosphorylates lamins to cause nuclear envelope breakdown (70, 71).
It also
localizes to chromatin, spindle microtubules, and unattached kinetochores (72).
Cdkl/Cyclin Bl is thought to play a role in chromatin condensation (73).
It
phosphorylates histone HI1, which was thought to mediate its role in chromatin
condensation, but the true significance of this phosphorylation has turned out to be less
clear (4). Cdkl/Cyclin B1 phosphorylates MAP4 and Stathmin/Opl8 to regulate spindle
microtubule dynamics (74-76). Additional substrates likely exist at the kinetochore, but
their identities are not known; Cdkl/Cyclin B1 localization specifically to unattached
kinetochores, however, suggests it may have a role in amphitelic attachment of
chromosomes to the spindle (72).
Another role of Cdkl/Cyclin B is to prepare cells for anaphase and exit from
mitosis through its own destruction.
The E3 ubiquitin ligase, Anaphase Promoting
complex/Cyclosome (APC/C), ubiquitinates mitotic proteins for destruction by the
proteasome. The APC/C can bind two WD40-repeat containing proteins that change its
substrate specificity, Cdc20 and Cdhl (77).
Cdc20 targets the APC/C to proteins
containing a sequence known as the destruction box (D-box). Cdhl targets the APC/C to
proteins containing a D-box or proteins containing a sequence known as the KEN box.
Cyclin-Cdk phosphorylation of Cdhl prevents it from interacting with the APC/C (78).
At the beginning of mitosis, Cdkl/Cyclin B is fully active and functions with Plkl to
phosphorylate subunits of APC/C to activate it (79). Degradation of D-box containing
proteins then begins with Cyclin A shortly after nuclear envelope breakdown in prophase
(80). Destruction of Cyclin B and Securin, which also contain D-boxes, does not occur
until after all chromosomes are aligned on the metaphase plate with amphitelic
attachments and the spindle assembly checkpoint has been silenced (81, 82). Securin is
an inhibitor of sister chromatid separation. Destruction of Cyclin B and Securin allows
chromosome segregation in anaphase. The destruction of Cyclin B inactivates Cdkl,
allowing Cdhl to bind the APC/C, thus, changing the APC/C substrate specificity to
allow other mitotic proteins to be destroyed. Geminin, an inhibitor of DNA replication,
is ubiquitinated by the APC/C Cdh l to allow relicensing of origins of replication (83).
Inactivation of Cdkl is required for mitotic exit, cytokinesis, chromatin decondensation,
and reformation of the nuclear envelope (4, 84).
Aurora A
Aurora Kinases are conserved from budding yeast to humans.
Budding yeast
have one Aurora, Ipll, whereas mammals have three, Aurora A, B, and C. Aurora B is
the ortholog of Ipll and will be discussed below. Aurora C is normally expressed only in
the testis, playing a role in meiosis (85).
Aurora A is ubiquitously expressed beginning in S-phase, through G2-phase, and
into mitosis with protein levels and activity peaking shortly after mitotic entry (86). It
localizes to the centrosome in S- and G2- phases and to both the centrosome and to a
portion of the spindle microtubules proximal to the centrosome in mitosis (87). There is
rapid exchange of localized Aurora A at both centrosomes and microtubules with a
cytoplasmic
pool of the kinase (88).
Aurora A is partially
activated by
autophosphorylation on its activation loop at T288 (equivalent to PKA T197) (17).
However, pT288 does not completely interact with the basic pocket resulting in the
activation segment still partially blocking the active site cleft. Additionally, since pT288
is not protected by the basic pocket, it is subject to dephosphorylation by Protein
Phosphatase 1 (PP 1), which regulates Aurora A activity (89). TPX2, which binds Aurora
A after nuclear envelope breakdown, fully activates Aurora A. The binding of TPX2 to
Aurora A causes the activation segment to move out of the active site cleft and buries
pT288 resulting in full activation of the kinase and preventing its dephosphorylation (17).
The motif of Aurora A has been determined to be R-X-S/T-(p, where
p is any
hydrophobic amino acid (90, 91). Based on loss of function studies, Aurora A plays roles
in centrosome maturation, mitotic entry, and centrosome separation (92-94).
The centrosome is the microtubule organizing center in metazoans and consists of
a pair of centrioles surrounded by pericentriolar material (PCM). The pericentriolar
material is the site of microtubule nucleation. Centrosome maturation, which occurs in
G2 and early mitosis, involves the recruitment of proteins to the pericentriolar material
such as y-tubulin ring complex and y-tubulin, which are involved in microtubule
nucleation. Aurora A is localized to centrosomes through Plkl activity, and Aurora A
activity at the centrosome is required for centrosome maturation including the
recruitment of y-tubulin, although how this is mediated is unclear (95).
One role of
Aurora A in the process of maturation is the phosphorylation of NDEL1, which is
required for TACC3 recruitment to the centrosome (93).
TACC3, which is also a
substrate of Aurora A, interacts with XMAP215, which stabilizes microtubules (96, 97).
PAK, Ajuba, and TPX2 interact individually with Aurora A at the centrosome and
activate it. The kinase PAK plays a role in focal adhesion dynamics as does Hefl,
another protein that interacts with Aurora A at the centrosome, suggesting Aurora A may
coordinate mitosis with substratum adhesion (98, 99). Another known substrate of Aurora
A at the centrosome is the kinase Lats2. Phosphorylation of Lats2 by Aurora A may play
a role in localizing Lats2 to the centrosome where it is required for centrosome
maturation (100,
101).
TPX2, which fully activates Aurora A kinase and is
phosphorylated by it, is responsible for re-localization of part of the Aurora A pool from
the centrosome to the spindle microtubules (102).
The localization of TPX2 and,
therefore, that of Aurora A on the spindle microtubules is also regulated by Plkl (95).
The role of Aurora A on microtubules is unknown.
Aurora A plays at least two roles in mitotic entry.
It is involved in the
localization of Cdkl/Cyclin B to the centrosome where the activation of Cdkl/Cyclin B
begins (94). Aurora A also phosphorylates Cdc25B to activate it (103). As discussed
previously, Cdc25B activates Cdkl/Cyclin B by removing inhibiting phosphorylations,
which results in a positive feedback loop resulting in activation of the pool of
Cdkl/Cyclin B and mitotic entry. Cdkl/Cyclin B also phosphorylates PP1, reducing its
activity.
This results in increased Aurora A activation and furthering the positive
feedback loop for Cdkl/Cyclin B activation and generating a positive feedback loop for
Aurora A activation (104).
In Drosophila, the mutation of the Aurora A homolog results in monopolar
spindles suggesting a role for Aurora A in centrosome separation (105). This phenotype
has also been seen in human cell culture treated with Aurora A inhibitors or siRNA (92,
106). However, the role of Aurora A in the separation or the maintenance of separation
of centrosomes is far from clear. The Xenopus ortholog of Aurora A has been shown to
phosphorylate the Xenopus ortholog of Eg5, which as discussed previously, is
phosphorylated by Cdkl/Cyclin B to localize it to the centrosome for its role in
centrosome separation (107). However, the role of the phosphorylation by Aurora A is
unknown.
Aurora A is destroyed by the proteasome beginning in anaphase, but its levels do
not become undetectable until Gl-phase (84, 108). Aurora A contains a D-box which is
only active in concert with the D-box activating domain (DAD) in Aurora A (109).
Phosphorylation in the DAD protects Aurora A from ubiquitination and subsequent
degradation (110). Although the D-box should be recognized by both Cdc20 and Cdhl,
ubiquitination of Aurora A is mediated by the Cdhl form of the APC/C (108, 111).
Aurora B
Aurora B is expressed from S- to M-phase with protein levels peaking in mitosis
(112). It is part of a complex including the non-enzymatic proteins INCENP, survivin,
and borealin, known as the chromosomal passenger complex (CPC) (113). As part of this
complex, Aurora B travels to several localizations during mitosis (114). In prophase, the
complex can be found diffusely over chromosomes, on both the arms and on the
centromeres.
By metaphase, the complex is concentrated on the centromeres, and in
anaphase and telophase, the complex moves to the spindle midzone and midbody,
respectively.
There is exchange of the complex with a cytoplasmic pool at the
centromeres through metaphase, but turnover is greatly reduced in anaphase at the
spindle midzone (113). The binding of INCENP to Aurora B activates Aurora B. Initial
binding of INCENP changes the position of the aC helix causing a partially active
conformation of the activation segment (115). Phosphorylation of INCENP by Aurora B
and autophosphorylation on its activation loop moves the aC and the activation segment
to the fully active conformation. In addition to its role in activating Aurora B, INCENP
is thought to be involved in localization of the complex, as are survivin and borealin
(116-118).
The Ipll motif, R/K-X-S/T-I/L/V, determined based on comparison of
substrate sites has been assumed to be the motif of Aurora B (110, 119).
Aurora B may plays roles in chromosome condensation, spindle assembly,
kinetochore-microtubule attachments, and spindle assembly checkpoint.
Aurora B
phosphorylates Histone H3 on S10 and S28, which correlates with chromosome
condensation during prophase, but the causal link, if one exists, is not known (120).
At least two modes of spindle assembly exist: a mode driven by the centrosomes,
in which centrosomes nucleate microtubules and kinetochore capture stabilizes them and
a mode in which microtubules that assemble in the cytoplasm are stabilized in the vicinity
of the chromatin. Aurora B seems to play a role in the latter form of spindle assembly at
least in Xenopus egg extracts as inhibition of Aurora B or depletion of the CPC inhibits
spindle assembly.
This may occur through phosphorylation and inhibition of
Stathmin/Opl8, which destabilizes microtubules or through phosphorylation and relocalization of MCAK, which also destabilizes microtubules (121, 122).
Aurora B affects kinetochore-microtubule attachments through a role in formation
of the trilaminar kinetochore (123). It also plays a role in fixing incorrect kinetochoremicrotubule attachments in response to a lack of tension on the kinetochore. Release of
kinetochore-microtubule attachments seems to involve phosphorylation of Hec 1, which is
part of a complex required for the kinetochore to bind microtubules (124). In budding
yeast, lack of tension seems to be sensed by the INCENP-survivin complex that binds at
the kinetochore-microtubule interface (125). This complex may recruit Aurora B when
there is a lack of tension, but not when there is appropriate tension. Aurora B seems to
recruit MCAK at merotelic attachments in a kinase activity dependent manner, where
MCAK can depolymerize inappropriately attached microtubules (126). This suggests a
mechanism in which Aurora B causes the detachment of inappropriately attached
microtubules, either syntelic or merotelic, by phosphorylating Hecl and can then cause
depolymerization
of the previously merotelically attached microtubules through
recruitment of MCAK. Aurora B is also involved in maintaining activation of the spindle
checkpoint until all chromosomes have aligned with appropriate microtubule attachments
on the metaphase plate (127).
Phosphorylation by Cdkl/Cyclin B and then Plkl on
INCENP regulates Aurora B in anaphase (128). Finally, Aurora B activity is required for
cytokinesis (113).
Nek2
Nek2 is a member of a family of kinases consisting of eleven members in human.
The kinase domain of Nek2 is the most similar of this family to the founding member of
the family in Aspergillus nidulans, Never In Mitosis A (NIMA) (129).
NIMA was
discovered in a screen for mutants that do not undergo mitosis, as its name suggests
(129). Nek2 protein levels and activity peak in S- and G2-phases (130). Nek2 contains a
leucine zipper motif C-terminal to its kinase domain which it uses for homodimerization,
which allows autophosphorylation for activation of the kinase (131). The activation of
Nek2 is also regulated by PP1 which dephosphorylates Nek2 to deactivate it (132). PP1
is inhibited by Nek2 phosphorylation, as well as Cdkl/Cyclin B phosphorylation and the
binding of Inhibitor-2, a protein that specifically inhibits PP1 (132, 133). This suggests a
positive feedback loop in activation of Nek2.
The optimal phosphorylation motif of
Nek2 on substrates is not known.
In S-phase and G2, Nek2 localizes to the core of the centrosome at the proximal
ends of the centrioles between the two pairs of centrioles, where it is both structural and
required for centrosome separation (134). The proteins Rootletin and Centrosomal Nek2associated protein 1 (C-Napl) are also at the proximal ends of the centrioles interacting
with each other to tether the two pairs of centrioles together (134, 135).
Nek2
phosphorylates both Rootletin and C-Napl causing them to be displaced from the
centrosome at the G2/M transition. Additionally, p-Catenin interacts with Rootletin and
when Rootletin is displaced by Nek2, p-Catenin relocates to other parts of the centrosome
(136).
The displacement of C-Napl, Rootletin and p-Catenin allow centrosome
separation. p-Catenin was also shown to be a substrate of Nek2, but the significance of
this is not clear.
Nek2 has also been implicated in regulating microtubule organization at the
centrosome. Ninein-like protein (Nlp) interacts with gamma-tubulin ring complex at the
centrosome in interphase cells, but does not localize to the centrosome in mitosis
suggesting that it has a role in microtubule nucleation in interphase, but not mitosis.
Nek2 and Plkl phosphorylate Nlp to cause its displacement from the centrosome (137).
How this process is important for mitosis is unclear.
There are at least three splice variants of Nek2: Nek2A, Nek2B, and Nek2C.
Nek2A and Nek2C both have a KEN-box and a D-box, but Nek2B has neither. Although
Cdc20 is required for Nek2A and Nek2C ubiquitination and degradation, APC/C
recognizes Nek2A and Nek2C directly, not through Cdc20 (138, 139).
As a result,
Nek2A and Nek2C are degraded shortly after nuclear envelope breakdown in a spindle
assembly checkpoint independent manner.
Nek2B levels remain constant through
mitosis, but then decrease in Gl-phase (140). How this decrease in protein levels occurs
is unknown.
Plk1
Polo-like kinase 1 (Plkl) is a member of the Polo-like kinase family, which is
named after the founding member discovered in Drosophila,Polo (141). Polo was named
for its mutant phenotype of abnormal spindle poles (141).
Like Drosophila,
Saccharomyces only has a single Polo family member, Cdc5. However, humans have 4
Polo kinases, Plkl, Plk2, Plk3, and Plk4. All four members play roles in the cell cycle,
but the roles of Plk2, Plk3, and Plk4 are less well understood than that of Plkl, which is
the ortholog of Cdc5 (142). Plk2, Plk3, and Plk4 will be discussed below.
Plkl is expressed from late S-phase into M-phase with protein levels peaking in
mitosis (143). It localizes to the centrosome in S- and G2-phases, to the centrosome and
kinetochores in prophase, to the centrosome, kinetochores, and spindle microtubules in
metaphase, and to the centrosomes, kinetochores, and spindle midzone in anaphase (143).
Plkl localized to the centrosomes and kinetochores exchanges rapidly with a cytoplasmic
pool of the protein (Kazuhiro Kishi, Personal Communication). In addition to an Nterminal kinase domain, Plkl has two C-terminal Polo-box motifs, which have recently
been found to fold together to form a Polo-box domain (PBD) (144). The Polo-box
domain of Plkl was found to be a phospho-dependent binding domain that binds
sequences with the consensus sequence S-pS/pT (48).
Kinase activation occurs
concurrently with expression and involves at least two possibly related mechanisms.
Kinase activity is inhibited by an interaction between the kinase domain and the PBD.
This autoinhibition is relieved by the binding of the PBD to phospho-peptides (144). It
has also been reported phosphomimic mutation of the activation loop reduces the
interaction between the kinase domain and PBD, suggesting that phosphorylation on the
activation loop, aside from activating the kinase by changing the activation loop
conformation to the active state, may also relieve the autoinhibition (145). The upstream
kinase that phosphorylates and Plkl for activation is not known. The consensus sequence
for Plkl phosphorylation has been determined by kinetics on peptides based on a known
site in Cdc25C. The resulting motif, D/E-X-S/T-(p-X-D/E, where ( is any hydrophobic
amino acid, is biased (146).
Plkl is involved in centrosome separation and centrosome maturation. The
original study of Polo mutants demonstrated monopolar spindles and subsequent work
has confirmed this (141, 147, 148). However, the role of Plkl in centrosome separation
is still unknown, possibly because its roles in separation and maturation are not clearly
separable. Several roles for Plkl in centrosome maturation are at least partially known.
Plk1 binds and recruits y-tubulin to the centrosome for microtubule nucleation (148, 149).
Plkl plays a role in inhibiting the microtubule destabilizing protein Stathmin/Opl8 (150).
As previously mentioned, it also regulates Nlp in conjunction with Nek2 (137).
Plkl
phosphorylates TCTP likely regulating its association with microtubules as it binds
microtubules to stabilize them until metaphase (151).
Plkl also phosphorylates Asp
which is involved in microtubule organization at the centrosome (152).
Plk1 plays two roles in mitotic entry. Plkl phosphorylates Cdc25C as part of the
activation
of Cdc25C
(153).
As
mentioned
previously
Cdc25C
removes
phosphorylations by Weel and Mytl that inhibit Cdkl/Cyclin B. Further contributing to
Cdkl/Cyclin B activation, Plkl phosphorylates Weel and Mytl resulting in their
inactivation (154, 155). Cdkl/Cyclin B generates PBD binding sites on Cdc25C and
Mytl that enable Plkl phosphorylation of these proteins (153, 155).
Plkl may play at least two roles associated with the kinetochore. It is required for
the association of Hecl, which is required for kinetochore-microtubule attachments,
Mad2, which is involved in the spindle checkpoint, and other proteins involved in
kinetochore structure (156).
It may also play a direct role in the spindle assembly
checkpoint by generating the 3F3/2 phosphoepitope on kinetochores lacking tension
(156, 157).
Plkl also regulates the metaphase-anaphase transition. Plkl activates the APC/C
in two ways. At the beginning of mitosis, Plkl and Cdkl/Cyclin B phosphorylate the
APC/C to activate it (79). Early mitotic inhibitor 1 (Emil) inhibits the APC/C in Sphase, but is degraded in prometaphase allowing activation of the APC/C.
Plkl,
stimulated by Cdkl/Cyclin B, phosphorylates Emil resulting in its destruction (158, 159).
The APC/C degrades Securin, which inhibits the protease Separase from degrading
cohesin and allowing sister chromatid separation (160).
In budding yeast, Cdc5
phosphorylates the cohesin Sccl to increase its affinity for separase, thus further
contributing the metaphase-anaphase transition. Although Plkl has roles in cytokinesis,
it begins to be ubiquitinated by the APC/C and destroyed in anaphase and is gone by Giphase (84, 161).
The Polo-like Kinase Family
As mentioned above, the Polo-like kinase family in humans has four members,
Plkl, Plk2, Plk3, and Plk4. This family of kinases is defined by two conserved features,
an N-terminal kinase domain and at least one Polo-box motif in the C-terminus. The
kinase domains are highly conserved with the highest sequence similarity between Plk2
and Plk3. Plkl is the next most similar and Plk4 is the least. Plkl, Plk2, and Plk3 have
two Polo-box motifs, which fold together to form a Polo-box Domain (PBD) (162). The
Polo-box domain of Plkl was shown to be a phospho-dependent binding domain and the
binding consensus sequences for the PBDs of Plkl, Plk2, and Plk3 were published to be
essentially the same, S-pS/pT. Plk4 only has a single Polo-box, the function of which
seems to be different from that of the other Plks, as described below. All members of this
family seem to play non-redundant roles in the cell cycle as the localizations and patterns
of expression are not the same. However, the roles of Plk2, Plk3, and Plk4 are poorly
understood and almost no substrates for these kinases are known.
Plk2
Plk2 was discovered as a serum inducible immediate early gene in NIH 3T3 cells
(163). It is expressed in Gl and protein levels and activity peak and fall off in Gl (163,
164). Phosphomimic mutation on the activation loop was found to increase the activity
of the kinase 10-fold suggesting that the kinase may need to be phosphorylated on the
activation loop to be fully activated (164).
Plk2 is a non-essential gene in mouse, but embryos showed slight growth
retardation and MEFs from these mice showed delayed S-phase entry, suggesting a role
for Plk2 in the cell cycle (165).
At least one cell cycle role may be in centriole
duplication as Plk2 localizes to the centrosome and both overexpression and knockdown
results in abnormal numbers of centrioles (166).
In addition to being expressed in
response to serum, it is expressed after X-ray and UV irradiation and other DNA
damaging agents due to p53 transactivation, implying a role in response to cellular
stresses (167, 168). Plk2 is also expressed in postmitotic cells, where it has been reported
to interact with calcium and integrin binding protein (CIB), which inhibits its activity and
spine associated RapGAP (SPAR) which is involved in remodeling dendritic spines (164,
165, 169, 170). Association with both of these proteins is through its PBD. However,
the phospho-dependency of the binding has not been checked. SPAR seems to be a
substrate of Plk2, but the site of phosphorylation was not identified (170). Indeed, no
phosphorylation sites have been mapped and the kinase motif has not been determined.
Plk3
Much of data on Plk3 is conflicting and/or surprisingly similar to that of Plkl or
Plk2. Like Plk2, Plk3 was originally discovered as an immediate early gene transcribed
in response to fibroblast growth factor (FGF) treatment of NIH 3T3 cells (171). Like
Plk2, Plk3 mRNA is expressed only in G1 (171, 172).
Reports of protein expression
have been inconsistent. Plk3 protein levels were originally reported to remain constant
throughout the cell cycle (173, 174). However, recently Plk3 protein levels have been
reported to track with the mRNA levels: increasing, peaking, and falling off in G1 (175).
The localization of the kinase is also controversial. The kinase was originally described
to localize to the cellular cortex (176, 177).
Later it was suggested to localize to
centrosomes, mitotic spindle, and the midbody during mitosis, which is very similar to
the distribution of Plkl (178, 179). Subsequent work suggests that the kinase does not
localize to any of these structures, but instead to the nucleolus (175). The mechanism of
activation of the kinase has not been examined at a biochemical level and the motif has
not been determined.
Plk3 seems to have the same function as Plk2 in neurons as it also interacts with
CIB and SPAR (169, 170). However, its role in the cell cycle seems to be very different
from that of Plk2.
Surprisingly, Plk3 has been reported to both inhibit as well as
promote cell cycle progression.
Plk3 seems to become transcribed and activated after oxidative stress and DNA
damage to mediate a stress response (174, 180, 181). This activation is at least partially
mediated by phosphorylation by Chk2 (181). Plk3 then phosphorylates Chk2, resulting
in further activation of Chk2 (182). It has also been reported to phosphorylate S20 of p53
(181). Phosphorylation at S20 of p53 is important for the cell cycle arrest mediated by
p53 and was previously reported to be phosphorylated by Chkl and Chk2 after DNA
damage (183, 184). Chkl and Chk2 have motifs requiring an R in the -3 position (185).
The S20 site of p53 has an E in the -3 position making it a more likely target of Plk3 than
Chkl or Chk2, based on the sequence similarity between Plkl and Plk3 kinase domains
and what is known about the Plkl motif (146).
Plk3 has recently been reported to
generate a priming phosphorylation for Glycogen Synthase Kinase 3 beta (Gsk31) on
Cdc25A resulting in phosphorylation of Cdc25A by Gsk3P and the degradation of
Cdc25A (186). Since Cdc25A promotes cell cycle progression, Plk3 is implicated in
inhibition of the cell cycle by two different pathways, one through stabilization of p53
and one through destruction of Cdc25A.
Plk3 has also been reported to phosphorylate Cdc25C resulting in the
translocation of Cdc25C into the nucleus as part of mitotic entry (187). It also may be
involved in promoting golgi fragmentation which is required for proper mitosis (188).
Recently, Plk3 has been shown to be required for entry into S-phase (175). Finally
overexpression of Plk3 results in cell cycle arrest and apoptosis, which was explained at
least partially as the outcome of defects in cytokinesis (177). Based on these reports,
Plk3 seems to also have a complex role in promoting the cell cycle. The role of Plk3 in
the cell cycle remains mysterious, but determination of the motif and additional
substrates should help to clear up the controversy.
Plk4
Plk4, like the other Plk family members, has an N-terminal kinase domain, but
unlike the other family members, it only has a single polo-box, which does not function
as a phospho-dependent binding domain. Instead, the polo-box seems to function in
homodimerization (189).
A second region that mediates homodimerization
independently of the polo-box also exists between the kinase domain and polo-box (189).
Plk4 is expressed from late Gl-phase through mitosis with mRNA levels falling in early
G1 (190). The cell cycle regulation of protein and kinase activity levels has not been
determined. Plk4 localizes to the nucleolus in G2 and to the centrosome in G2 and M
(191). The centrosome localization seems to be mediated by the polo-box (189). The
motif of Plk4 has been reported to be X-X-S/T-(p-(p-X/P, where
Xis a charged residue and 9p is a large hydrophobic residue (192).
Plk4 seems to be required for mitosis as the Plk4 knockout mouse is embryonic
lethal with the embryo dying with large numbers of cells in mitosis with high Cyclin B
levels (191). MEFs from mice heterozygous for Plk4 showed centrosome amplification,
multipolar spindles, and aneuploidy (193).
Plk4 has been shown to be required for
centriole duplication, likely explaining the phenotypes seen in the Plk4 heterozygote
MEFs (194, 195). Substrates of the kinase have yet to be determined.
Summary and Preview
Protein kinases play key roles in regulation of the cell cycle. Activation of a
kinase frequently involves phosphorylation on the activation loop. The phosphoresidue
on the activation loop interacts with a basic pocket in the active site cleft causing a
conformational change that aligns the catalytic residues for optimal phospho-transfer.
The motif of a kinase is determined by the residues in the active site cleft which select
sequences that the kinase can phosphorylate based on their ability to make favorable
electrostatic, hydrogen-bonding, and/or hydrophobic interactions. In the case of kinases
that do not require phosphorylation for activation, the basic pocket can contribute to the
specificity of the kinase by binding sequences that have been previously phosphorylated
upstream or downstream of the site to be phosphorylated. Determinants of localization of
a kinase play an important role in substrate specificity by increasing the local
concentration of kinase in the area of the substrate. Activation, motifs, and localizations
play key roles in the substrate specificity of the major mitotic kinases Cdkl, Aurora A,
Aurora B, Nek2, and Plkl. Each of the events in mitosis is regulated by more than one of
these kinases and these kinases have overlapping localizations often even interacting with
and phosphorylating the same substrate proteins on different residues. The motifs of this
functional family have not been completely defined. Very little is known about the
activation, motifs, and localizations of the rest of the Polo-like kinase family, Plk2, Plk3,
and Plk4, or their roles in the cell cycle. Very few substrates of these kinases have been
identified.
In this thesis, we set out to determine complete, unbiased motifs of two sets of
kinases involved in the regulation of the cell cycle. The motifs of the functionally
related, but sequence dissimilar major mitotic kinases were determined (Chapter Two) as
well of those of the sequence related, but functionally distinct Polo-like kinase family
(Chapter Three). The motif of Plkl led to the discovery of new substrates. Additionally,
the motifs of the major mitotic kinases along with their localizations lead us to put
forward a hypothesis to explain how despite overlapping localizations and overlapping
motifs, these kinases are able to phosphorylate distinct substrates sites, suggesting that
motifs may be involved in the coordinated regulation of these kinases as a group not just
the kinases individually. The motifs of the Polo-like kinase family result in the exciting
novel motif of Plk3 involving specificity for phosphotyrosine and the even more exciting
result that the motif of a kinase is not necessarily 'static' as was previously thought. In
Chapter 4, the motifs of these two sets of kinases are compared and contrasted with
respect to sequence and resulting function and follow-up experiments are proposed.
References
1.Kops GJ, Weaver BA, Cleveland DW. On the road to cancer: Aneuploidy and the
mitotic checkpoint. Nat Rev Cancer. 2005 Oct;5(10):773-85.
2. Jallepalli PV, Lengauer C. Chromosome segregation and cancer: Cutting through the
mystery. Nat Rev Cancer. 2001 Nov;1(2): 109-17.
3. Ubersax JA, Ferrell JE,Jr. Mechanisms of specificity in protein phosphorylation. Nat
Rev Mol Cell Biol. 2007 Jul;8(7):530-41.
4. Nigg EA. Mitotic kinases as regulators of cell division and its checkpoints. Nat Rev
Mol Cell Biol. 2001 Jan;2(1):21-32.
5. Dai W. Polo-like kinases, an introduction. Oncogene. 2005 Jan 10;24(2):214-6.
6. Manning G, Whyte DB, Martinez R, Hunter T, Sudarsanam S. The protein kinase
complement of the human genome. Science. 2002 Dec 6;298(5600):1912-34.
7. Knighton DR, Zheng JH, Ten Eyck LF, Ashford VA, Xuong NH, Taylor SS, Sowadski
JM. Crystal structure of the catalytic subunit of cyclic adenosine monophosphatedependent protein kinase. Science. 1991 Jul 26;253(5018):407-14.
8. Brown NR, Noble ME, Endicott JA, Johnson LN. The structural basis for specificity of
substrate and recruitment peptides for cyclin-dependent kinases. Nat Cell Biol. 1999
Nov; 1(7):438-43.
9. Canagarajah BJ, Khokhlatchev A, Cobb MH, Goldsmith EJ. Activation mechanism of
the MAP kinase ERK2 by dual phosphorylation. Cell. 1997 Sep 5;90(5):859-69.
10. Madhusudan, Trafny EA, Xuong NH, Adams JA, Ten Eyck LF, Taylor SS, Sowadski
JM. cAMP-dependent protein kinase: Crystallographic insights into substrate recognition
and phosphotransfer. Protein Sci. 1994 Feb;3(2):176-87.
11. Johnson LN, Noble ME, Owen DJ. Active and inactive protein kinases: Structural
basis for regulation. Cell. 1996 Apr 19;85(2):149-58.
12. Adams JA. Kinetic and catalytic mechanisms of protein kinases. Chem Rev. 2001
Aug; 101(8):2271-90.
13. De Bondt HL, Rosenblatt J, Jancarik J, Jones HD, Morgan DO, Kim SH. Crystal
structure of cyclin-dependent kinase 2. Nature. 1993 Jun 17;363(6430):595-602.
14. Zhang F, Strand A, Robbins D, Cobb MH, Goldsmith EJ. Atomic structure of the
MAP kinase ERK2 at 2.3 A resolution. Nature. 1994 Feb 24;367(6465):704-11.
15. Nolen B, Taylor S, Ghosh G. Regulation of protein kinases; controlling activity
through activation segment conformation. Mol Cell. 2004 Sep 10;15(5):661-75.
16. Huse M, Kuriyan J. The conformational plasticity of protein kinases. Cell. 2002 May
3;109(3):275-82.
17. Bayliss R, Sardon T, Vernos I, Conti E. Structural basis of aurora-A activation by
TPX2 at the mitotic spindle. Mol Cell. 2003 Oct;12(4):851-62.
18. Dajani R, Fraser E, Roe SM, Young N, Good V, Dale TC, Pearl LH. Crystal structure
of glycogen synthase kinase 3 beta: Structural basis for phosphate-primed substrate
specificity and autoinhibition. Cell. 2001 Jun 15;105(6):721-32.
19. ter Haar E, Coll JT, Austen DA, Hsiao HM, Swenson L, Jain J. Structure of
GSK3beta reveals a primed phosphorylation mechanism. Nat Struct Biol. 2001
Jul;8(7):593-6.
20. Longenecker KL, Roach PJ, Hurley TD. Three-dimensional structure of mammalian
casein kinase I: Molecular basis for phosphate recognition. J Mol Biol. 1996 Apr
5;257(3):618-31.
21. Bax B, Carter PS, Lewis C, Guy AR, Bridges A, Tanner R, Pettman G, Mannix C,
Culbert AA, Brown MJ, Smith DG, Reith AD. The structure of phosphorylated GSK3beta complexed with a peptide, FRATtide, that inhibits beta-catenin phosphorylation.
Structure. 2001 Dec;9(12):1143-52.
22. Fiol CJ, Wang A, Roeske RW, Roach PJ. Ordered multisite protein phosphorylation.
analysis of glycogen synthase kinase 3 action using model peptide substrates. J Biol
Chem. 1990 Apr 15;265(11):6061-5.
23. Flotow H, Graves PR, Wang AQ, Fiol CJ, Roeske RW, Roach PJ. Phosphate groups
as substrate determinants for casein kinase I action. J Biol Chem. 1990 Aug
25;265(24):14264-9.
24. Kemp BE, Graves DJ, Benjamini E, Krebs EG. Role of multiple basic residues in
determining the substrate specificity of cyclic AMP-dependent protein kinase. J Biol
Chem. 1977 Jul 25;252(14):4888-94.
25. Gibbs CS, Zoller MJ. Rational scanning mutagenesis of a protein kinase identifies
functional regions involved in catalysis and substrate interactions. J Biol Chem. 1991
May 15;266(14):8923-31.
26. Gibbs CS, Zoller MJ. Identification of electrostatic interactions that determine the
phosphorylation site specificity of the cAMP-dependent protein kinase. Biochemistry.
1991 Jun 4;30(22):5329-34.
27. Knighton DR, Zheng JH, Ten Eyck LF, Xuong NH, Taylor SS, Sowadski JM.
Structure of a peptide inhibitor bound to the catalytic subunit of cyclic adenosine
monophosphate-dependent protein kinase. Science. 1991 Jul 26;253(5018):414-20.
28. Pinna LA, Ruzzene M. How do protein kinases recognize their substrates? Biochim
Biophys Acta. 1996 Dec 12;1314(3):191-225.
29. Kemp BE, Bylund DB, Huang TS, Krebs EG. Substrate specificity of the cyclic
AMP-dependent protein kinase. Proc Natl Acad Sci U S A. 1975 Sep;72(9):3448-52.
30. Songyang Z, Blechner S, Hoagland N, Hoekstra MF, Piwnica-Worms H, Cantley LC.
Use of an oriented peptide library to determine the optimal substrates of protein kinases.
Curr Biol. 1994 Nov 1;4(11):973-82.
31. Manke IA, Nguyen A, Lim D, Stewart MQ, Elia AE, Yaffe MB. MAPKAP kinase-2
is a cell cycle checkpoint kinase that regulates the G2/M transition and S phase
progression in response to UV irradiation. Mol Cell. 2005 Jan 7;17(1):37-48.
32. Nair SA, Kim MH, Warren SD, Choi S, Songyang Z, Cantley LC, Hangauer DG.
Identification of efficient pentapeptide substrates for the tyrosine kinase pp60c-src. J Med
Chem. 1995 Oct 13;38(21):4276-83.
33. Nishikawa K, Toker A, Johannes FJ, Songyang Z, Cantley LC. Determination of the
specific substrate sequence motifs of protein kinase C isozymes. J Biol Chem. 1997 Jan
10;272(2):952-60.
34. Obata T, Yaffe MB, Leparc GG, Piro ET, Maegawa H, Kashiwagi A, Kikkawa R,
Cantley LC. Peptide and protein library screening defines optimal substrate motifs for
AKT/PKB. J Biol Chem. 2000 Nov 17;275(46):36108-15.
35. Ruzzene M, Songyang Z, Marin O, Donella-Deana A, Brunati AM, Guerra B,
Agostinis P, Cantley LC, Pinna LA. Sequence specificity of C-terminal src kinase (CSK)-a comparison with src-related kinases c-fgr and lyn. Eur J Biochem. 1997 Jun
1;246(2):433-9.
36. Songyang Z, Lu KP, Kwon YT, Tsai LH, Filhol O, Cochet C, Brickey DA, Soderling
TR, Bartleson C, Graves DJ, DeMaggio AJ, Hoekstra MF, Blenis J, Hunter T, Cantley
LC. A structural basis for substrate specificities of protein Ser/Thr kinases: Primary
sequence preference of casein kinases I and II, NIMA, phosphorylase kinase, calmodulindependent kinase II, CDK5, and Erkl. Mol Cell Biol. 1996 Nov;16(11):6486-93.
37. Hutti JE, Jarrell ET, Chang JD, Abbott DW, Storz P, Toker A, Cantley LC, Turk BE.
A rapid method for determining protein kinase phosphorylation specificity. Nat Methods.
2004 Oct; 1(1):27-9.
38. Bullock AN, Debreczeni J, Amos AL, Knapp S, Turk BE. Structure and substrate
specificity of the pim-1 kinase. J Biol Chem. 2005 Dec 16;280(50):41675-82.
39. Bunkoczi G, Salah E, Filippakopoulos P, Fedorov O, Muller S, Sobott F, Parker SA,
Zhang H, Min W, Turk BE, Knapp S. Structural and functional characterization of the
human protein kinase ASKi. Structure. 2007 Oct;15(10):1215-26.
40. Eswaran J, Bernad A, Ligos JM, Guinea B, Debreczeni JE, Sobott F, Parker SA,
Najmanovich R, Turk BE, Knapp S. Structure of the human protein kinase MPSK1
reveals an atypical activation loop architecture. Structure. 2008 Jan;16(1): 115-24.
41. Fu Z, Larson KA, Chitta RK, Parker SA, Turk BE, Lawrence MW, Kaldis P,
Galaktionov K, Cohn SM, Shabanowitz J, Hunt DF, Sturgill TW. Identification of yinyang regulators and a phosphorylation consensus for male germ cell-associated kinase
(MAK)-related kinase. Mol Cell Biol. 2006 Nov;26(22):8639-54.
42. Hutti JE, Turk BE, Asara JM, Ma A, Cantley LC, Abbott DW. IkappaB kinase beta
phosphorylates the K63 deubiquitinase A20 to cause feedback inhibition of the NFkappaB pathway. Mol Cell Biol. 2007 Nov;27(21):7451-61.
43. Pike AC, Rellos P, Niesen FH, Turnbull A, Oliver AW, Parker SA, Turk BE, Pearl
LH, Knapp S. Activation segment dimerization: A mechanism for kinase
autophosphorylation of non-consensus sites. EMBO J. 2008 Feb 20;27(4):704-14.
44. Rennefahrt UE, Deacon SW, Parker SA, Devarajan K, Beeser A, Chernoff J, Knapp
S, Turk BE, Peterson JR. Specificity profiling of pak kinases allows identification of
novel phosphorylation sites. J Biol Chem. 2007 May 25;282(21):15667-78.
45. Taylor S, Ghosh G. Protein kinases: Catalysis and regulation. Current Opinion in
Structural Biology. 2006;16(6):665.
46. Tanoue T, Adachi M, Moriguchi T, Nishida E. A conserved docking motif in MAP
kinases common to substrates, activators and regulators. Nat Cell Biol. 2000
Feb;2(2): 110-6.
47. Hubbard SR, Till JH. Protein tyrosine kinase structure and function. Annu Rev
Biochem. 2000;69:373-98.
48. Elia AE, Cantley LC, Yaffe MB. Proteomic screen finds pSer/pThr-binding domain
localizing Plkl to mitotic substrates. Science. 2003 Feb 21;299(5610):1228-31.
49. Lowery DM, Lim D, Yaffe MB. Structure and function of polo-like kinases.
Oncogene. 2005 Jan 10;24(2):248-59.
50. Schulman BA, Lindstrom DL, Harlow E. Substrate recruitment to cyclin-dependent
kinase 2 by a multipurpose docking site on cyclin A. Proc Natl Acad Sci U S A. 1998 Sep
1;95(18):10453-8.
51. Loog M, Morgan DO. Cyclin specificity in the phosphorylation of cyclin-dependent
kinase substrates. Nature. 2005 Mar 3;434(7029): 104-8.
52. Pines J, Hunter T. The differential localization of human cyclins A and B is due to a
cytoplasmic retention signal in cyclin B. EMBO J. 1994 Aug 15;13(16):3772-81.
53. Li J, Meyer AN, Donoghue DJ. Nuclear localization of cyclin B mediates its
biological activity and is regulated by phosphorylation. Proc Natl Acad Sci U S A. 1997
Jan 21;94(2):502-7.
54. Hagting A, Jackman M, Simpson K, Pines J. Translocation of cyclin B1 to the
nucleus at prophase requires a phosphorylation-dependent nuclear import signal. Curr
Biol. 1999 Jul 1;9(13):680-9.
55. Yang J, Song H, Walsh S, Bardes ES, Kornbluth S. Combinatorial control of cyclin
B 1 nuclear trafficking through phosphorylation at multiple sites. J Biol Chem. 2001 Feb
2;276(5):3604-9.
56. Jackman M, Lindon C, Nigg EA, Pines J. Active cyclin B1-Cdkl first appears on
centrosomes in prophase. Nat Cell Biol. 2003 Feb;5(2):143-8.
57. Pines J, Hunter T. Human cyclin A is adenovirus E1A-associated protein p60 and
behaves differently from cyclin B. Nature. 1990 Aug 23;346(6286):760-3.
58. Fisher RP, Morgan DO. A novel cyclin associates with MO 15/CDK7 to form the
CDK-activating kinase. Cell. 1994 Aug 26;78(4):713-24.
59. Parker LL, Piwnica-Worms H. Inactivation of the p34cdc2-cyclin B complex by the
human WEE1 tyrosine kinase. Science. 1992 Sep 25;257(5078):1955-7.
60. Liu F, Stanton JJ, Wu Z, Piwnica-Worms H. The human Myt1 kinase preferentially
phosphorylates Cdc2 on threonine 14 and localizes to the endoplasmic reticulum and
golgi complex. Mol Cell Biol. 1997 Feb;17(2):571-83.
61. Kramer A, Mailand N, Lukas C, Syljuasen RG, Wilkinson CJ, Nigg EA, Bartek J,
Lukas J. Centrosome-associated Chkl prevents premature activation of cyclin-B-Cdkl
kinase. Nat Cell Biol. 2004 Sep;6(9):884-91.
62. Mueller PR, Coleman TR, Dunphy WG. Cell cycle regulation of a xenopus Wee llike kinase. Mol Biol Cell. 1995 Jan;6(1):119-34.
63. Hoffmann I, Clarke PR, Marcote MJ, Karsenti E, Draetta G. Phosphorylation and
activation of human cdc25-C by cdc2--cyclin B and its involvement in the selfamplification of MPF at mitosis. EMBO J. 1993 Jan;12(1):53-63.
64. Blangy A, Lane HA, d'Herin P, Harper M, Kress M, Nigg EA. Phosphorylation by
p34cdc2 regulates spindle association of human Eg5, a kinesin-related motor essential for
bipolar spindle formation in vivo. Cell. 1995 Dec 29;83(7):1159-69.
65. Jackman M, Firth M, Pines J. Human cyclins B and B2 are localized to strikingly
different structures: B 1Ito microtubules, B2 primarily to the golgi apparatus. EMBO J.
1995 Apr 18;14(8):1646-54.
66. Nguyen TB, Manova K, Capodieci P, Lindon C, Bottega S, Wang XY, Refik-Rogers
J, Pines J, Wolgemuth DJ, Koff A. Characterization and expression of mammalian cyclin
b3, a prepachytene meiotic cyclin. J Biol Chem. 2002 Nov 1;277(44):41960-9.
67. Draviam VM, Orrechia S, Lowe M, Pardi R, Pines J. The localization of human
cyclins B 1 and B2 determines CDK1 substrate specificity and neither enzyme requires
MEK to disassemble the golgi apparatus. J Cell Biol. 2001 Mar 5;152(5):945-58.
68. Lowe M, Rabouille C, Nakamura N, Watson R, Jackman M, Jamsa E, Rahman D,
Pappin DJ, Warren G. Cdc2 kinase directly phosphorylates the cis-golgi matrix protein
GM130 and is required for golgi fragmentation in mitosis. Cell. 1998 Sep 18;94(6):78393.
69. Jackman M, Firth M, Pines J. Human cyclins Bi and B2 are localized to strikingly
different structures: B to microtubules, B2 primarily to the golgi apparatus. EMBO J.
1995 Apr 18;14(8):1646-54.
70. Heald R, McKeon F. Mutations of phosphorylation sites in lamin A that prevent
nuclear lamina disassembly in mitosis. Cell. 1990 May 18;61(4):579-89.
71. Peter M, Nakagawa J, Doree M, Labbe JC, Nigg EA. In vitro disassembly of the
nuclear lamina and M phase-specific phosphorylation of lamins by cdc2 kinase. Cell.
1990 May 18;61(4):591-602.
72. Bentley AM, Normand G, Hoyt J, King RW. Distinct sequence elements of cyclin B
promote localization to chromatin, centrosomes, and kinetochores during mitosis. Mol
Biol Cell. 2007 Dec; 18(12):4847-58.
73. Lamb NJ, Fernandez A, Watrin A, Labbe JC, Cavadore JC. Microinjection of
p34cdc2 kinase induces marked changes in cell shape, cytoskeletal organization, and
chromatin structure in mammalian fibroblasts. Cell. 1990 Jan 12;60(1): 151-65.
74. Verde F, Dogterom M, Stelzer E, Karsenti E, Leibler S. Control of microtubule
dynamics and length by cyclin A- and cyclin B-dependent kinases in xenopus egg
extracts. J Cell Biol. 1992 Sep; 118(5):1097-108.
75. Ookata K, Hisanaga S, Bulinski JC, Murofushi H, Aizawa H, Itoh TJ, Hotani H,
Okumura E, Tachibana K, Kishimoto T. Cyclin B interaction with microtubuleassociated protein 4 (MAP4) targets p34cdc2 kinase to microtubules and is a potential
regulator of M-phase microtubule dynamics. J Cell Biol. 1995 Mar;128(5):849-62.
76. Larsson N, Marklund U, Gradin HM, Brattsand G, Gullberg M. Control of
microtubule dynamics by oncoprotein 18: Dissection of the regulatory role of multisite
phosphorylation during mitosis. Mol Cell Biol. 1997 Sep; 17(9):5530-9.
77. Zur A, Brandeis M. Timing of APC/C substrate degradation is determined by fzy/fzr
specificity of destruction boxes. EMBO J. 2002 Sep 2;21(17):4500-10.
78. Lukas C, Sorensen CS, Kramer E, Santoni-Rugiu E, Lindeneg C, Peters JM, Bartek J,
Lukas J. Accumulation of cyclin B 1 requires E2F and cyclin-A-dependent rearrangement
of the anaphase-promoting complex. Nature. 1999 Oct 21;401(6755):815-8.
79. Kraft C, Herzog F, Gieffers C, Mechtler K, Hagting A, Pines J, Peters JM. Mitotic
regulation of the human anaphase-promoting complex by phosphorylation. EMBO J.
2003 Dec 15;22(24):6598-609.
80. den Elzen N, Pines J. Cyclin A is destroyed in prometaphase and can delay
chromosome alignment and anaphase. J Cell Biol. 2001 Apr 2;153(1):121-36.
81. Hagting A, Den Elzen N, Vodermaier HC, Waizenegger IC, Peters JM, Pines J.
Human securin proteolysis is controlled by the spindle checkpoint and reveals when the
APC/C switches from activation by Cdc20 to Cdhl. J Cell Biol. 2002 Jun
24;157(7):1125-37.
82. Clute P, Pines J. Temporal and spatial control of cyclin B 1 destruction in metaphase.
Nat Cell Biol. 1999 Jun;1(2):82-7.
83. Li A, Blow JJ. Non-proteolytic inactivation of geminin requires CDK-dependent
ubiquitination. Nat Cell Biol. 2004 Mar;6(3):260-7.
84. Pines J. Mitosis: A matter of getting rid of the right protein at the right time. Trends
Cell Biol. 2006 Jan;16(1):55-63.
85. Kimura M, Matsuda Y, Yoshioka T, Okano Y. Cell cycle-dependent expression and
centrosome localization of a third human aurora/Ipll-related protein kinase, AIK3. J Biol
Chem. 1999 Mar 12;274(11):7334-40.
86. Kimura M, Kotani S, Hattori T, Sumi N, Yoshioka T, Todokoro K, Okano Y. Cell
cycle-dependent expression and spindle pole localization of a novel human protein
kinase, aik, related to aurora of drosophila and yeast Ipll. J Biol Chem. 1997 May
23;272(21):13766-71.
87. Barr AR, Gergely F. Aurora-A: The maker and breaker of spindle poles. J Cell Sci.
2007 Sep 1;120(Pt 17):2987-96.
88. Stenoien DL, Sen S, Mancini MA, Brinkley BR. Dynamic association of a tumor
amplified kinase, aurora-A, with the centrosome and mitotic spindle. Cell Motil
Cytoskeleton. 2003 Jun;55(2):134-46.
89. Katayama H, Zhou H, Li Q, Tatsuka M, Sen S. Interaction and feedback regulation
between STK 5/BTAK/Aurora-A kinase and protein phosphatase 1 through mitotic cell
division cycle. J Biol Chem. 2001 Dec 7;276(49):46219-24.
90. Ohashi S, Sakashita G, Ban R, Nagasawa M, Matsuzaki H, Murata Y, Taniguchi H,
Shima H, Furukawa K, Urano T. Phospho-regulation of human protein kinase aurora-A:
Analysis using anti-phospho-Thr288 monoclonal antibodies. Oncogene. 2006 Dec
14;25(59):7691-702.
91. Ferrari S, Marin O, Pagano MA, Meggio F, Hess D, El-Shemerly M, Krystyniak A,
Pinna LA. Aurora-A site specificity: A study with synthetic peptide substrates. Biochem
J. 2005 Aug 15;390(Pt 1):293-302.
92. Marumoto T, Honda S, Hara T, Nitta M, Hirota T, Kohmura E, Saya H. Aurora-A
kinase maintains the fidelity of early and late mitotic events in HeLa cells. J Biol Chem.
2003 Dec 19;278(51):51786-95.
93. Mori D, Yano Y, Toyo-oka K, Yoshida N, Yamada M, Muramatsu M, Zhang D, Saya
H, Toyoshima YY, Kinoshita K, Wynshaw-Boris A, Hirotsune S. NDEL1
phosphorylation by aurora-A kinase is essential for centrosomal maturation, separation,
and TACC3 recruitment. Mol Cell Biol. 2007 Jan;27(1):352-67.
94. Hirota T, Kunitoku N, Sasayama T, Marumoto T, Zhang D, Nitta M, Hatakeyama K,
Saya H. Aurora-A and an interacting activator, the LIM protein ajuba, are required for
mitotic commitment in human cells. Cell. 2003 Sep 5;114(5):585-98.
95. De Luca M, Lavia P, Guarguaglini G. A functional interplay between aurora-A, Plkl
and TPX2 at spindle poles: Plkl controls centrosomal localization of aurora-A and TPX2
spindle association. Cell Cycle. 2006 Feb;5(3):296-303.
96. Gergely F, Draviam VM, Raff JW. The ch-TOG/XMAP215 protein is essential for
spindle pole organization in human somatic cells. Genes Dev. 2003 Feb 1;17(3):336-41.
97. Kinoshita K, Noetzel TL, Pelletier L, Mechtler K, Drechsel DN, Schwager A, Lee M,
Raff JW, Hyman AA. Aurora A phosphorylation of TACC3/maskin is required for
centrosome-dependent microtubule assembly in mitosis. J Cell Biol. 2005 Sep
26;170(7):1047-55.
98. Zhao ZS, Lim JP, Ng YW, Lim L, Manser E. The GIT-associated kinase PAK targets
to the centrosome and regulates aurora-A. Mol Cell. 2005 Oct 28;20(2):237-49.
99. Pugacheva EN, Golemis EA. The focal adhesion scaffolding protein HEF1 regulates
activation of the aurora-A and Nek2 kinases at the centrosome. Nat Cell Biol. 2005
Oct;7(10):937-46.
100. Abe Y, Ohsugi M, Haraguchi K, Fujimoto J, Yamamoto T. LATS2-ajuba complex
regulates gamma-tubulin recruitment to centrosomes and spindle organization during
mitosis. FEBS Lett. 2006 Feb 6;580(3):782-8.
101. Toji S, Yabuta N, Hosomi T, Nishihara S, Kobayashi T, Suzuki S, Tamai K, Nojima
H. The centrosomal protein Lats2 is a phosphorylation target of aurora-A kinase. Genes
Cells. 2004 May;9(5):383-97.
102. Kufer TA, Sillje HH, Komer R, Gruss OJ, Meraldi P, Nigg EA. Human TPX2 is
required for targeting aurora-A kinase to the spindle. J Cell Biol. 2002 Aug
19;158(4):617-23.
103. Dutertre S, Cazales M, Quaranta M, Froment C, Trabut V, Dozier C, Mirey G,
Bouche JP, Theis-Febvre N, Schmitt E, Monsarrat B, Prigent C, Ducommun B.
Phosphorylation of CDC25B by aurora-A at the centrosome contributes to the G2-M
transition. J Cell Sci. 2004 May 15;117(Pt 12):2523-31.
104. Marumoto T, Hirota T, Morisaki T, Kunitoku N, Zhang D, Ichikawa Y, Sasayama T,
Kuninaka S, Mimori T, Tamaki N, Kimura M, Okano Y, Saya H. Roles of aurora-A
kinase in mitotic entry and G2 checkpoint in mammalian cells. Genes Cells. 2002
Nov;7(11):1173-82.
105. Glover DM, Leibowitz MH, McLean DA, Parry H. Mutations in aurora prevent
centrosome separation leading to the formation of monopolar spindles. Cell. 1995 Apr
7;81(1):95-105.
106. Girdler F, Gascoigne KE, Eyers PA, Hartmuth S, Crafter C, Foote KM, Keen NJ,
Taylor SS. Validating aurora B as an anti-cancer drug target. J Cell Sci. 2006 Sep
1;119(Pt 17):3664-75.
107. Giet R, Uzbekov R, Cubizolles F, Le Guellec K, Prigent C. The xenopus laevis
aurora-related protein kinase pEg2 associates with and phosphorylates the kinesin-related
protein X1Eg5. J Biol Chem. 1999 May 21;274(21):15005-13.
108. Castro A, Arlot-Bonnemains Y, Vigneron S, Labbe JC, Prigent C, Lorca T.
APC/Fizzy-related targets aurora-A kinase for proteolysis. EMBO Rep. 2002
May;3(5):457-62.
109. Castro A, Vigneron S, Bemis C, Labbe JC, Prigent C, Lorca T. The D-boxactivating domain (DAD) is a new proteolysis signal that stimulates the silent D-box
sequence of aurora-A. EMBO Rep. 2002 Dec;3(12): 1209-14.
110. Carmena M, Earnshaw WC. The cellular geography of aurora kinases. Nat Rev Mol
Cell Biol. 2003 Nov;4(11):842-54.
111. Honda K, Mihara H, Kato Y, Yamaguchi A, Tanaka H, Yasuda H, Furukawa K,
Urano T. Degradation of human Aurora2 protein kinase by the anaphase-promoting
complex-ubiquitin-proteasome pathway. Oncogene. 2000 Jun 1;19(24):2812-9.
112. Bischoff JR, Anderson L, Zhu Y, Mossie K, Ng L, Souza B, Schryver B, Flanagan
P, Clairvoyant F, Ginther C, Chan CS, Novotny M, Slamon DJ, Plowman GD. A
homologue of drosophila aurora kinase is oncogenic and amplified in human colorectal
cancers. EMBO J. 1998 Jun 1;17(11):3052-65.
113. Ruchaud S, Carmena M, Earnshaw WC. Chromosomal passengers: Conducting cell
division. Nat Rev Mol Cell Biol. 2007 Oct;8(10):798-812.
114. Crosio C, Fimia GM, Loury R, Kimura M, Okano Y, Zhou H, Sen S, Allis CD,
Sassone-Corsi P. Mitotic phosphorylation of histone H3: Spatio-temporal regulation by
mammalian aurora kinases. Mol Cell Biol. 2002 Feb;22(3):874-85.
115. Sessa F, Mapelli M, Ciferri C, Tarricone C, Areces LB, Schneider TR, Stukenberg
PT, Musacchio A. Mechanism of aurora B activation by INCENP and inhibition by
hesperadin. Mol Cell. 2005 Apr 29;18(3):379-91.
116. Vader G, Kauw JJ, Medema RH, Lens SM. Survivin mediates targeting of the
chromosomal passenger complex to the centromere and midbody. EMBO Rep. 2006
Jan;7(1):85-92.
117. Gassmann R, Carvalho A, Henzing AJ, Ruchaud S, Hudson DF, Honda R, Nigg EA,
Gerloff DL, Earnshaw WC. Borealin: A novel chromosomal passenger required for
stability of the bipolar mitotic spindle. J Cell Biol. 2004 Jul 19;166(2):179-91.
118. Honda R, Korner R, Nigg EA. Exploring the functional interactions between aurora
B, INCENP, and survivin in mitosis. Mol Biol Cell. 2003 Aug;14(8):3325-41.
119. Cheeseman IM, Anderson S, Jwa M, Green EM, Kang J, Yates JR,3rd, Chan CS,
Drubin DG, Barnes G. Phospho-regulation of kinetochore-microtubule attachments by
the aurora kinase Ipllp. Cell. 2002 Oct 18;111(2):163-72.
120. Goto H, Yasui Y, Nigg EA, Inagaki M. Aurora-B phosphorylates histone H3 at
serine28 with regard to the mitotic chromosome condensation. Genes Cells. 2002
Jan;7(1):11-7.
121. Ohi R, Sapra T, Howard J, Mitchison TJ. Differentiation of cytoplasmic and meiotic
spindle assembly MCAK functions by aurora B-dependent phosphorylation. Mol Biol
Cell. 2004 Jun;15(6):2895-906.
122. Gadea BB, Ruderman JV. Aurora B is required for mitotic chromatin-induced
phosphorylation of Opl 8/Stathmin. Proc Natl Acad Sci U S A. 2006 Mar
21;103(12):4493-8.
123. Liu ST, Rattner JB, Jablonski SA, Yen TJ. Mapping the assembly pathways that
specify formation of the trilaminar kinetochore plates in human cells. J Cell Biol. 2006
Oct 9;175(1):41-53.
124. DeLuca JG, Gall WE, Ciferri C, Cimini D, Musacchio A, Salmon ED. Kinetochore
microtubule dynamics and attachment stability are regulated by Hec . Cell. 2006 Dec
1;127(5):969-82.
125. Sandall S, Severin F, McLeod IX, Yates JR,3rd, Oegema K, Hyman A, Desai A. A
Birl-Slil5 complex connects centromeres to microtubules and is required to sense
kinetochore tension. Cell. 2006 Dec 15;127(6):1179-91.
126. Knowlton AL, Lan W, Stukenberg PT. Aurora B is enriched at merotelic attachment
sites, where it regulates MCAK. Curr Biol. 2006 Sep 5;16(17):1705-10.
127. Morrow CJ, Tighe A, Johnson VL, Scott MI, Ditchfield C, Taylor SS. Bubl and
aurora B cooperate to maintain BubR1-mediated inhibition of APC/CCdc20. J Cell Sci.
2005 Aug 15;118(Pt 16):3639-52.
128. Goto H, Kiyono T, Tomono Y, Kawajiri A, Urano T, Furukawa K, Nigg EA,
Inagaki M. Complex formation of Plkl and INCENP required for metaphase-anaphase
transition. Nat Cell Biol. 2006 Feb;8(2):180-7.
129. O'regan L, Blot J, Fry AM. Mitotic regulation by NIMA-related kinases. Cell Div.
2007 Aug 29;2:25.
130. Fry AM, Schultz SJ, Bartek J, Nigg EA. Substrate specificity and cell cycle
regulation of the Nek2 protein kinase, a potential human homolog of the mitotic regulator
NIMA of aspergillus nidulans. J Biol Chem. 1995 May 26;270(21):12899-905.
131. Fry AM, Arnaud L, Nigg EA. Activity of the human centrosomal kinase, Nek2,
depends on an unusual leucine zipper dimerization motif. J Biol Chem. 1999 Jun
4;274(23):16304-10.
132. Helps NR, Luo X, Barker HM, Cohen PT. NIMA-related kinase 2 (Nek2), a cellcycle-regulated protein kinase localized to centrosomes, is complexed to protein
phosphatase 1. Biochem J. 2000 Jul 15;349(Pt 2):509-18.
133. Eto M, Elliott E, Prickett TD, Brautigan DL. Inhibitor-2 regulates protein
phosphatase-1 complexed with NimA-related kinase to induce centrosome separation. J
Biol Chem. 2002 Nov 15;277(46):44013-20.
134. Fry AM, Mayor T, Meraldi P, Stierhof YD, Tanaka K, Nigg EA. C-Nap 1, a novel
centrosomal coiled-coil protein and candidate substrate of the cell cycle-regulated protein
kinase Nek2. J Cell Biol. 1998 Jun 29;141(7):1563-74.
135. Bahe S, Stierhof YD, Wilkinson CJ, Leiss F, Nigg EA. Rootletin forms centrioleassociated filaments and functions in centrosome cohesion. J Cell Biol. 2005 Oct
10;171(1):27-33.
136. Bahmanyar S, Kaplan DD, Deluca JG, Giddings TH,Jr, O'Toole ET, Winey M,
Salmon ED, Casey PJ, Nelson WJ, Barth AL. Beta-catenin is a Nek2 substrate involved in
centrosome separation. Genes Dev. 2008 Jan 1;22(1):91-105.
137. Rapley J, Baxter JE, Blot J, Wattam SL, Casenghi M, Meraldi P, Nigg EA, Fry AM.
Coordinate regulation of the mother centriole component nlp by nek2 and plkl protein
kinases. Mol Cell Biol. 2005 Feb;25(4):1309-24.
138. Hayes MJ, Kimata Y, Wattam SL, Lindon C, Mao G, Yamano H, Fry AM. Early
mitotic degradation of Nek2A depends on Cdc20-independent interaction with the
APC/C. Nat Cell Biol. 2006 Jun;8(6):607-14.
139. Wu W, Baxter JE, Wattam SL, Hayward DG, Fardilha M, Knebel A, Ford EM, da
Cruz e Silva,E.F., Fry AM. Alternative splicing controls nuclear translocation of the cell
cycle-regulated Nek2 kinase. J Biol Chem. 2007 Sep 7;282(36):26431-40.
140. Hayward DG, Fry AM. Nek2 kinase in chromosome instability and cancer. Cancer
Lett. 2006 Jun 18;237(2):155-66.
141. Sunkel CE, Glover DM. Polo, a mitotic mutant of drosophila displaying abnormal
spindle poles. J Cell Sci. 1988 Jan;89 ( Pt 1)(Pt 1):25-38.
142. Lee KS, Grenfell TZ, Yarm FR, Erikson RL. Mutation of the polo-box disrupts
localization and mitotic functions of the mammalian polo kinase plk. Proc Natl Acad Sci
U S A. 1998 Aug 4;95(16):9301-6.
143. van Vugt MA, Medema RH. Getting in and out of mitosis with polo-like kinase-1.
Oncogene. 2005 Apr 18;24(17):2844-59.
144. Elia AE, Rellos P, Haire LF, Chao JW, Ivins FJ, Hoepker K, Mohammad D, Cantley
LC, Smerdon SJ, Yaffe MB. The molecular basis for phosphodependent substrate
targeting and regulation of plks by the polo-box domain. Cell. 2003 Oct 3;1 15(1):83-95.
145. Jang YJ, Lin CY, Ma S, Erikson RL. Functional studies on the role of the C-terminal
domain of mammalian polo-like kinase. Proc Natl Acad Sci U S A. 2002 Feb
19;99(4):1984-9.
146. Nakajima H, Toyoshima-Morimoto F, Taniguchi E, Nishida E. Identification of a
consensus motif for plk (polo-like kinase) phosphorylation reveals Mytl as a Plk1
substrate. J Biol Chem. 2003 Jul 11;278(28):25277-80.
147. van Vugt MA, van de Weerdt BC, Vader G, Janssen H, Calafat J, Klompmaker R,
Wolthuis RM, Medema RH. Polo-like kinase-1 is required for bipolar spindle formation
but is dispensable for anaphase promoting complex/Cdc20 activation and initiation of
cytokinesis. J Biol Chem. 2004 Aug 27;279(35):36841-54.
148. Lane HA, Nigg EA. Antibody microinjection reveals an essential role for human
polo-like kinase 1 (Plkl) in the functional maturation of mitotic centrosomes. J Cell Biol.
1996 Dec;135(6 Pt 2):1701-13.
149. Feng Y, Hodge DR, Palmieri G, Chase DL, Longo DL, Ferris DK. Association of
polo-like kinase with alpha-, beta- and gamma-tubulins in a stable complex. Biochem J.
1999 Apr 15;339 ( Pt 2)(Pt 2):435-42.
150. Budde PP, Kumagai A, Dunphy WG, Heald R. Regulation of Op 18 during spindle
assembly in xenopus egg extracts. J Cell Biol. 2001 Apr 2;153(1):149-58.
151. Yarm FR. Plk phosphorylation regulates the microtubule-stabilizing protein TCTP.
Mol Cell Biol. 2002 Sep;22(17):6209-21.
152. do Carmo Avides M, Tavares A, Glover DM. Polo kinase and asp are needed to
promote the mitotic organizing activity of centrosomes. Nat Cell Biol. 2001
Apr;3(4):421-4.
153. Elia AE, Cantley LC, Yaffe MB. Proteomic screen finds pSer/pThr-binding domain
localizing Plk1 to mitotic substrates. Science. 2003 Feb 21;299(5610):1228-31.
154. Watanabe N, Arai H, Iwasaki J, Shiina M, Ogata K, Hunter T, Osada H. Cyclindependent kinase (CDK) phosphorylation destabilizes somatic Weel via multiple
pathways. Proc Natl Acad Sci U S A. 2005 Aug 16;102(33):11663-8.
155. Inoue D, Sagata N. The polo-like kinase Plxl interacts with and inhibits Mytl after
fertilization of xenopus eggs. EMBO J. 2005 Mar 9;24(5): 1057-67.
156. Ahonen LJ, Kallio MJ, Daum JR, Bolton M, Manke IA, Yaffe MB, Stukenberg PT,
Gorbsky GJ. Polo-like kinase 1 creates the tension-sensing 3F3/2 phosphoepitope and
modulates the association of spindle-checkpoint proteins at kinetochores. Curr Biol. 2005
Jun 21;15(12):1078-89.
157. Wong OK, Fang G. Plxl is the 3F3/2 kinase responsible for targeting spindle
checkpoint proteins to kinetochores. J Cell Biol. 2005 Aug 29;170(5):709-19.
158. Moshe Y, Boulaire J, Pagano M, Hershko A. Role of polo-like kinase in the
degradation of early mitotic inhibitor 1, a regulator of the anaphase promoting
complex/cyclosome. Proc Natl Acad Sci U S A. 2004 May 25;101(21):7937-42.
159. Hansen DV, Loktev AV, Ban KH, Jackson PK. Plkl regulates activation of the
anaphase promoting complex by phosphorylating and triggering SCFbetaTrCP-dependent
destruction of the APC inhibitor Emil. Mol Biol Cell. 2004 Dec;15(12):5623-34.
160. Musacchio A, Salmon ED. The spindle-assembly checkpoint in space and time. Nat
Rev Mol Cell Biol. 2007 May;8(5):379-93.
161. Seong YS, Kamijo K, Lee JS, Fernandez E, Kuriyama R, Miki T, Lee KS. A spindle
checkpoint arrest and a cytokinesis failure by the dominant-negative polo-box domain of
Plkl in U-2 OS cells. J Biol Chem. 2002 Aug 30;277(35):32282-93.
162. Elia AE, Rellos P, Haire LF, Chao JW, Ivins FJ, Hoepker K, Mohammad D, Cantley
LC, Smerdon SJ, Yaffe MB. The molecular basis for phosphodependent substrate
targeting and regulation of plks by the polo-box domain. Cell. 2003 Oct 3;1 15(1):83-95.
163. Simmons DL, Neel BG, Stevens R, Evett G, Erikson RL. Identification of an earlygrowth-response gene encoding a novel putative protein kinase. Mol Cell Biol. 1992
Sep; 12(9):4164-9.
164. Ma S, Liu MA, Yuan YL, Erikson RL. The serum-inducible protein kinase snk is a
Gi phase polo-like kinase that is inhibited by the calcium- and integrin-binding protein
CIB. Mol Cancer Res. 2003 Mar;1(5):376-84.
165. Ma S, Charron J, Erikson RL. Role of Plk2 (snk) in mouse development and cell
proliferation. Mol Cell Biol. 2003 Oct;23(19):6936-43.
166. Warnke S, Kemmler S, Hames RS, Tsai HL, Hoffmann-Rohrer U, Fry AM,
Hoffinann I. Polo-like kinase-2 is required for centriole duplication in mammalian cells.
Curr Biol. 2004 Jul 13;14(13):1200-7.
167. Bums TF, Fei P, Scata KA, Dicker DT, El-Deiry WS. Silencing of the novel p53
target gene Snk/Plk2 leads to mitotic catastrophe in paclitaxel (taxol)-exposed cells. Mol
Cell Biol. 2003 Aug;23(16):5556-71.
168. Shimizu-Yoshida Y, Sugiyama K, Rogounovitch T, Ohtsuru A, Namba H, Saenko
V, Yamashita S. Radiation-inducible hSNK gene is transcriptionally regulated by p53
binding homology element in human thyroid cells. Biochem Biophys Res Commun. 2001
Nov 30;289(2):491-8.
169. Kauselmann G, Weiler M, WulffP, Jessberger S, Konietzko U, Scafidi J, Staubli U,
Bereiter-Hahn J, Strebhardt K, Kuhl D. The polo-like protein kinases fnk and snk
associate with a ca(2+)- and integrin-binding protein and are regulated dynamically with
synaptic plasticity. EMBO J. 1999 Oct 15;18(20):5528-39.
170. Pak DT, Sheng M. Targeted protein degradation and synapse remodeling by an
inducible protein kinase. Science. 2003 Nov 21;302(5649):1368-73.
171. Donohue PJ, Alberts GF, Guo Y, Winkles JA. Identification by targeted differential
display of an immediate early gene encoding a putative serine/threonine kinase. J Biol
Chem. 1995 Apr 28;270(17):10351-7.
172. Li B, Ouyang B, Pan H, Reissmann PT, Slamon DJ, Arceci R, Lu L, Dai W. Prk, a
cytokine-inducible human protein serine/threonine kinase whose expression appears to be
down-regulated in lung carcinomas. J Biol Chem. 1996 Aug 9;271(32):19402-8.
173. Chase D, Feng Y, Hanshew B, Winkles JA, Longo DL, Ferris DK. Expression and
phosphorylation of fibroblast-growth-factor-inducible kinase (fnk) during cell-cycle
progression. Biochem J. 1998 Aug 1;333 ( Pt 3)(Pt 3):655-60.
174. Bahassi el M, Conn CW, Myer DL, Hennigan RF, McGowan CH, Sanchez Y,
Stambrook PJ. Mammalian polo-like kinase 3 (Plk3) is a multifunctional protein involved
in stress response pathways. Oncogene. 2002 Sep 26;21(43):6633-40.
175. Zimmerman WC, Erikson RL. Polo-like kinase 3 is required for entry into S phase.
Proc Natl Acad Sci U S A. 2007 Feb 6;104(6):1847-52.
176. Holtrich U, Wolf G, Yuan J, Bereiter-Hahn J, Kam T, Weiler M, Kauselmann G,
Rehli M, Andreesen R, Kaufmann M, Kuhl D, Strebhardt K. Adhesion induced
expression of the serine/threonine kinase fnk in human macrophages. Oncogene. 2000
Oct 5;19(42):4832-9.
177. Conn CW, Hennigan RF, Dai W, Sanchez Y, Stambrook PJ. Incomplete cytokinesis
and induction of apoptosis by overexpression of the mammalian polo-like kinase, Plk3.
Cancer Res. 2000 Dec 15;60(24):6826-31.
178. Jiang N, Wang X, Jhanwar-Uniyal M, Darzynkiewicz Z, Dai W. Polo box domain of
Plk3 functions as a centrosome localization signal, overexpression of which causes
mitotic arrest, cytokinesis defects, and apoptosis. J Biol Chem. 2006 Apr
14;281(15):10577-82.
179. Wang Q, Xie S, Chen J, Fukasawa K, Naik U, Traganos F, Darzynkiewicz Z,
Jhanwar-Uniyal M, Dai W. Cell cycle arrest and apoptosis induced by human polo-like
kinase 3 is mediated through perturbation of microtubule integrity. Mol Cell Biol. 2002
May;22(10):3450-9.
180. Li Z, Niu J, Uwagawa T, Peng B, Chiao PJ. Function of polo-like kinase 3 in NFkappaB-mediated proapoptotic response. J Biol Chem. 2005 Apr 29;280(17): 16843-50.
181. Xie S, Wu H, Wang Q, Kunicki J, Thomas RO, Hollingsworth RE, Cogswell J, Dai
W. Genotoxic stress-induced activation of Plk3 is partly mediated by Chk2. Cell Cycle.
2002 Nov-Dec;1(6):424-9.
182. Bahassi el M, Myer DL, McKenney RJ, Hennigan RF, Stambrook PJ. Priming
phosphorylation of Chk2 by polo-like kinase 3 (Plk3) mediates its full activation by ATM
and a downstream checkpoint in response to DNA damage. Mutat Res. 2006 Apr
11;596(1-2):166-76.
183. Hirao A, Kong YY, Matsuoka S, Wakeham A, Ruland J, Yoshida H, Liu D, Elledge
SJ, Mak TW. DNA damage-induced activation of p53 by the checkpoint kinase Chk2.
Science. 2000 Mar 10;287(5459): 1824-7.
184. Shieh SY, Ahn J, Tamai K, Taya Y, Prives C. The human homologs of checkpoint
kinases Chkl and Cdsl (Chk2) phosphorylate p53 at multiple DNA damage-inducible
sites. Genes Dev. 2000 Feb 1;14(3):289-300.
185. O'Neill T, Giarratani L, Chen P, Iyer L, Lee CH, Bobiak M, Kanai F, Zhou BB,
Chung JH, Rathbun GA. Determination of substrate motifs for human Chkl and
hCdsl/Chk2 by the oriented peptide library approach. J Biol Chem. 2002 May
3;277(18):16102-15.
186. Kang T, Wei Y, Honaker Y, Yamaguchi H, Appella E, Hung MC, Piwnica-Worms
H. GSK-3 beta targets Cdc25A for ubiquitin-mediated proteolysis, and GSK-3 beta
inactivation correlates with Cdc25A overproduction in human cancers. Cancer Cell. 2008
Jan;13(1):36-47.
187. Bahassi el M, Hennigan RF, Myer DL, Stambrook PJ. Cdc25C phosphorylation on
serine 191 by Plk3 promotes its nuclear translocation. Oncogene. 2004 Apr
8;23(15):2658-63.
188. Xie S, Wang Q,Ruan Q, Liu T, Jhanwar-Uniyal M, Guan K, Dai W. MEKl-induced
golgi dynamics during cell cycle progression is partly mediated by polo-like kinase-3.
Oncogene. 2004 May 6;23(21):3822-9.
189. Leung GC, Hudson JW, Kozarova A, Davidson A, Dennis JW, Sicheri F. The sak
polo-box comprises a structural domain sufficient for mitotic subcellular localization. Nat
Struct Biol. 2002 Oct;9(10):719-24.
190. Fode C, Binkert C, Dennis JW. Constitutive expression of murine sak-a suppresses
cell growth and induces multinucleation. Mol Cell Biol. 1996 Sep;16(9):4665-72.
191. Hudson JW, Kozarova A, Cheung P, Macmillan JC, Swallow CJ, Cross JC, Dennis
JW. Late mitotic failure in mice lacking sak, a polo-like kinase. Curr Biol. 2001 Mar
20;11(6):441-6.
192. Leung GC, Ho CS, Blasutig IM, Murphy JM, Sicheri F. Determination of the
Plk4/Sak consensus phosphorylation motif using peptide spots arrays. FEBS Lett. 2007
Jan 9;581(1):77-83.
193. Ko MA, Rosario CO, Hudson JW, Kulkami S, Pollett A, Dennis JW, Swallow CJ.
Plk4 haploinsufficiency causes mitotic infidelity and carcinogenesis. Nat Genet. 2005
Aug;37(8):883-8.
194. Habedanck R, StierhofYD, Wilkinson CJ, Nigg EA. The polo kinase Plk4 functions
in centriole duplication. Nat Cell Biol. 2005 Nov;7(11):1140-6.
195. Kleylein-Sohn J, Westendorf J, Le Clech M, Habedanck R, Stierhof YD, Nigg EA.
Plk4-induced centriole biogenesis in human cells. Dev Cell. 2007 Aug;13(2):190-202.
Chapter Two
The Determination of the Motifs of the Major Mitotic Kinases: Motifs
and Localizations of the Major Mitotic Kinases May Coordinately
Regulate the Kinases as a Group
Jes Alexander, Dan Lim, Brian Joughin, Fabio Sessa, Frank Ivins, Andrew Fry, P. Todd
Stuckenberg, Andrea Mussachio, Steve Smerdon, Jan-Michael Peters, Michael Yaffe
Contributions:
Dan and I determined the motif of Plkl together (Figure 2.2A).
Bioinformatics
localization enrichment data (Figure 2.6A) and methods are from Brian. Data in Figure
2.3C is from Jan-Michael.
Figure 2.3D and 2.4D were created by Dan. Figure 2.5D
was created by Steve. Aurora A protein was from Todd, Aurora B protein was from
Fabio and Andrea, Nek2 protein was from Frank, Andrew, and Steve, and PlkI protein
was from Dan. I performed all the other experiments, wrote the rest of the sections, and
created the other figures.
Abstract
Mitosis is highly regulated, as errors in this process result in aneuploidy and
genetic instability. Regulation of these processes relies heavily on precise protein
phosphorylation of specific substrate sites and the major mitotic kinases, Cdkl/Cyclin B,
Aurora A, Aurora B, Nek2, and Plkl, play key roles in this regulation. Some of the
substrates of these kinases are known, but many remain to be identified as many of the
roles of these kinases in mitosis have yet to be understood at the molecular level. These
kinases have overlapping localizations and, therefore, have access to the same substrates
and sites on substrates.
Yet, each kinase phosphorylates distinct sites from those
phosphorylated by each of the other kinases. How this specificity is maintained is not
clear. To gain insight into this specificity and to aid in the identification of new
substrates of these kinases, we have determined the motifs of the major mitotic kinases
using Positional Scanning Oriented Peptide Library Screening (PS-OPLS) which gives
complete and unbiased motifs. We verified the motifs using kinetic studies on peptides
with single amino acid changes from the optimal peptide determined by PS-OPLS. The
motifs match known substrates of these kinases. Using structural models of the kinase
domains with the optimal peptides docked in, we were able to rationalize the substrate
specificity of these kinases. Plkl was found to have selectivity for N in the -2 position,
which was not known previously, and this led to the identification of new
phosphorylation sites in Sccl, Bubl, and p31/Comet. Based on the motifs and the
previously known localizations we put forward a hypothesis that may explain how these
kinases maintain distinct phosphorylation sites and, if true, would suggest that the
individual motifs and localizations of the major mitotic kinases cooperatively regulate the
substrate site selection of these kinases in a coordinated manner.
Introduction
Mitosis is the process by which the contents of the nucleus are partitioned to
daughter cells during cell division. This process includes centrosome separation and
maturation, chromatin condensation, nuclear envelope breakdown, spindle formation,
amphitelic chromosome attachment, and chromosome segregation.
The timing and
mechanics of each of these processes is carefully regulated as problems in these
processes result in aneuploidy, genetic instability, and cancer (1, 2). Regulation of these
processes relies heavily on precise protein phosphorylation of specific substrate sites and
the major mitotic kinases, Cdkl/Cyclin B, Aurora A, Aurora B, Nek2, and Plkl, play key
roles in this regulation (3). Although some of the substrates of these kinases are known,
many have yet to be identified (3). Additionally, how the specificity of these kinases is
established and maintained remains to be understood.
Cdkl/Cyclin B is the master regulator of mitosis and its activation marks entry
into mitosis. It has roles in chromatin condensation, nuclear envelope breakdown, and
spindle formation (4-10). Aurora A is involved in centrosome separation and maturation,
while Aurora B is involved in chromosome condensation and assuring amphitelic
chromosome attachment (11-16). Nek2 is required for centrosome separation and plays a
role in centrosome maturation (17, 18). Plkl has been shown to regulate Cdkl/Cyclin B
activation, centrosome maturation, spindle formation, and amphitelic chromosome
attachment (19-24).
Not surprisingly, since these kinases cooperate to regulate processes within
mitosis, they have overlapping localizations (Figure 2.1). At and just prior to the G2/M
transition, Aurora A and Plkl are localized to the pericentriolar material, while Nek2
localizes to the proximal centriole at the core of the centrosome and Aurora B localizes to
the centromeres of the decondensed, but duplicated chromosomes (17, 25-27). At this
time, Cdkl/CyclinB activity is low but increasing. At metaphase, Cdkl/Cyclin B is
maximally active and diffusely localized throughout the cell, but intense activity appears
to be present at the centrosomes, spindle microtubules, and kinetochores (10, 28-30).
Aurora A and Plkl co-localize with Cdkl/Cyclin B to the pericentriolar material and the
Figure 2.1:
The major mitotic kinases have overlapping localizations in mitosis.
Localizations of the major mitotic kinases, Cdkl/Cyclin B (red), Aurora A (green),
Aurora B (purple), Plkl (blue), and Nek2 (yellow), at the G2/M transition (left) and at
metaphase (right) are shown.
spindle microtubules (25, 31). Plkl also co-localizes with Aurora B and Cdkl/Cyclin B
at the kinetochores (10, 25, 27).
Although mitosis has been studied extensively, there is still much to understand
about the process at the molecular level. Some substrates of these kinases are known, but
there seem to be more substrates to be identified, as the known substrates do not to
explain all the phenotypic changes these kinases are thought to be involved in.
Additionally, despite overlapping subcellular localizations, and therefore, access to
overlapping sets of potential phosphorylation sites on substrates, these kinases have
distinct and apparently mutually exclusive sets of phosphorylation sites in vivo, as seems
to be required for the proper regulation of mitosis. How the activities of, and substrate
targeting by, these kinases are coordinated so that each phosphorylates its own sites, but
not those of the other kinases is not clear. Since consensus phosphorylation motifs of
individual kinases play important roles in substrate selection for many well-studied
kinases, we suspected that determining the optimal motifs for these mitotic kinases might
shed additional light on the problem of substrate selection and co-localization, as well as
help in the identification of new substrates of these kinases.
We, therefore, have determined the motifs of Cdkl/Cyclin B, Aurora A, Aurora
B, Nek2, and Plkl, using Positional Scanning Oriented Peptide Library Screening (Figure
2.2A), which gives an unbiased and complete motif. In particular, this method allows
each amino acid in each position in a sequence motif to be tested independently of the
rest of the sequence, and defines not only which amino acids are positively selected in
each position, but also which are discriminated against.
Subsequent kinetic studies
verified the motifs determined by PS-OPLS, and the motifs closely matched several
previously mapped sites on known substrates. Additionally, we were able to rationalize
the basis for motif selection using models of the optimal peptide docked into the X-ray
crystal structures of these kinases. In addition, the new expanded motif of Plkl allowed
the identification of new substrates for Plkl. Finally, based on the motifs of these kinases
and their localizations, we propose the hypothesis that in mitosis, major mitotic kinases
with overlapping localization do not have overlapping motifs and major mitotic kinases
with overlapping motifs do not have overlapping localizations. Such regulation would
imply that all the individual motifs and localizations of the major mitotic kinases together
cooperatively regulate the substrate site selection of these kinases in a coordinated
manner, giving us the beginnings of a systems-level understanding of the regulation of
mitosis by protein phosphorylation.
Experimental Procedures
Kinase Protein Production and Purification
Recombinant full-length human wildtype C-terminally hexa-His-tagged Aurora
A, X. Laevis wildtype Aurora B(60-361):INCENP(790-847) complex, and full-length
human T175A Nek2A proteins were produced as described previously (32-34).
To
produce T210D Plkl kinase domain, nucleotide sequence encoding the kinase domain of
human Plkl (aa. 38-346) was cloned into NdeI and XhoI restriction sites in the pET28a
bacterial expression vector (Novagen).
The resulting fusion protein contains an N-
terminal hexa-His-tag and a thrombin protease cleavage site between hexa-His-tag and
the kinase domain. A Thr-210 mutation to Asp was introduced using the QuikChange
Site-Directed Mutagenesis Kit (Stratagene) and verified by sequencing. Protein was
expressed in E. coli Rosetta (Novagen) cells. A starter culture was grown overnight in
LB. 6L of TB culture was initiated by addition of the starter culture at a 1:100 dilution.
After 4 hours of growth, the 6L of TB culture were cooled to 40C on ice, IPTG was added
to a final concentration of 1 mM and the cultures were incubated at room temperature for
16 hours. Cells were pelleted by centrifugation at 4000 X g for 10 min at 40 C, the pellet
from each liter of culture was resuspended in -40 mL of lysis buffer (50 mM Tris, 500
mM NaCl, 14 mM P-mercaptoethanol), transferred to separate 50 mL Falcon tubes, and
lysed by sonication using a Branford sonicator with a probe tip (power level 8, duty cycle
50%, 2:33 min:sec, 2 times) in an ethanol-ice bath.
The lysate was clarified by
ultracentrifugation at 40,000 RPM in a Beckman 45Ti rotor for 30 min at 40C. Initial
purification utilized Ni-NTA affinity chromatography. Following ultracentrifugation, the
supernatant was mixed with 25 mL of packed Ni-NTA beads by end-over-end rotation at
40C for 2 hours. The bead slurry was loaded onto a column for FPLC with the aid of a
peristaltic pump. The column was transferred to the FPLC, washed with 280 mL of wash
buffer (500 mM NaC1, 14 mM 3-mercaptoethanol, 10 mM imidazole, 10 mM Phosphate,
pH 7), and the protein was eluted using buffer of the same composition except containing
300 mM imidazole. Fractions were analyzed by on-line UV absorbance and SDS/PAGE.
Fractions containing Plkl kinase domain protein were pooled, 100 units of thrombin
protease were added to cleave off the hexa-His-MBP tag, and the mixture was incubated
at 40C overnight. To remove the cleaved tag, the protein was re-passaged through NiNTA using the same conditions as described above. The resulting -10 mL of material
was applied to a Superose 12 column (16 mm X 30 cm) and eluted using gel filtration
buffer (500 mM NaC1, 2 mM DTT, 10 mM Tris, pH 8). Fractions (2 mL) were analyzed
for Plkl content by on-line UV absorbance and SDS-PAGE.
Phosphorylation Motif Determination by Peptide Library Array Screening
Positional Scanning Oriented Peptide Library Screening (PS-OPLS) was
performed following Hutti et al. (35). Briefly, solution-phase kinase reactions were
performed in parallel on 198 separate biotinylated, partially degenerate oriented peptide
libraries (Anaspec, Inc) arrayed in a 384 well microtiter plate in a 22 row X 9 column
format. Each peptide library contains an N-terminal biotin tag, a 50:50 mix of serine and
threonine at the orienting phospho-acceptor residue, a single second fixed amino acid
located between the -5 and +4 position, and a mixture of amino acids at all other
positions. Individual libraries contain any of the 20 natural amino acids as well as
phosphothreonine and phosphotyrosine in the second fixed position, corresponding to the
22 rows. Scanning down the rows in the array moves the position of the fixed amino acid
from -5 to +4 relative to the fixed phospho-acceptor residue, while scanning across the
columns changes the identity of the second fixed amino acid. As an example, the peptide
library with Lys fixed in the -4 position has the following sequence: Y-A-X-K-X-X-XS/T-X-X-X-X-A-G-K-K-biotin, where amino acids are represented in 1-letter code, and
X is an equal mixture of all 17 natural amino acids excluding Cys, Ser, and Thr to
prevent oxidation effects and eliminate secondary phosphorylation events. S/T denotes a
50:50 mix of Ser and Thr. Kinase reactions were performed at 300C in a total volume of
16 [LI containing 31.25 jlM peptide library, 100 jlM ATP, and 200 loCi of [32 P]-y-ATP, in
150 mM NaCl (500 mM NaCl for Nek2 kinase reactions), 10 mM MgC12, 1 mM DTT,
0.1% Tween 20, and 50 mM Tris, pH 7.5. For Aurora A and Aurora B, reactions were
performed for 6 hours with 0.25 and 0.792 plg of protein per reaction, respectively. For
Nek2, reactions were performed for 8 hours with 2.4 jlg of protein per reaction.
Reactions for the Plkl motif were performed for 3 hours with 0.5 jg of protein per
reaction. For Cdkl/Cyclin B, reactions were performed for 4 hours with 0.08 jtg of
protein complex (Millipore) per reaction. Following incubation, 2 j1 of each reaction
were simultaneously transferred to a streptavidin-coated membrane (Promega SAM 2
biotin capture membrane) using a 384 slot pin replicator (VP Scientific). The membrane
was washed three times with 140 mM NaC1, 0.1% SDS, 10 mM Tris, pH 7.4, three times
with 2 M NaC1, twice with 2 M NaCl containing 1% H3PO4, and once with water. The
extent of peptide library phosphorylation was determined by imaging the membrane with
a phosphorimager (Molecular Dynamics).
In Vitro Kinase Assays
Kinase assays for kinetic parameter determination were performed at 300 C in 90
tL of kinase reaction buffer (50 mM Tris, pH 7.5, 150 mM NaCl (500 mM NaCl for
Nek2 assays), 10 mM MgC12, 100 jpM ATP, 9 ICX [32 P]-y-ATP, and 1 mM DTT). Each
reaction contained 0.003 jtg of Aurora B:INCENP complex, 3.6 jlg of T175A Nek2
protein, or 0.054 gg of T210D Plkl kinase domain. The sequences of the optimal
peptides
for
Aurora
WFRMSIRGGYKKKK,
B,
Nek2
and
Plkl
GHDTSFYWAAYKKKK,
were
ARRHSMGWAYKKKK,
respectively.
These optimal
peptides were determined by taking the most highly selected amino acid from each
position, -4 to +3, of the PS-OPLS blot and using S as the phospho-acceptor residue, 0.
A C-terminal tyrosine was added to each peptide to allow determination of concentration
of peptide solutions by UV spectrometry, and four C-terminal lysines were appended to
increase solubility and electrostatic interaction with phosphocellulose paper. Additional
peptides with single amino acid changes from the optimal peptides were as indicated.
Concentrations of peptides were as indicated. 5 jiL of each reaction were spotted on
phosphocellulose at 0, 3, 6, 9, 12, and 15 minutes in duplicate. The phosphocellulose
paper was washed 4 times with 0.5% phosphoric acid, transferred vials containing 3 mL
of scintillation cocktail, and scintillation counted. From these kinase assays, Kin, Vmax,
and Vmax/Km values were determined by curve fitting assuming classic Michaelis-Menten
kinetics. For each concentration of peptide, care was taken to insure that less than 5% of
the total substrate was phosphorylated and that the reaction rate was linear with respect to
time.
Kinase assays for comparison of activity of each kinase against the optimal
peptides of each kinase were performed at 300C in 25 pLL of kinase reaction buffer. Each
reaction contained 0.008 jtg of Aurora B:INCENP complex, 1 jig of T175A Nek2
protein, or 0.046 jig of T210D Plkl kinase domain. Amounts of each kinase were chosen
so that reactions of each kinase with its optimal peptide would yield approximately equal
amounts of phosphorylated peptide. Concentrations of optimal peptide were set to the
Km of the kinase in the reaction. Reactions were performed in triplicate and 5 jtL of each
reaction was spotted on phosphocellulose at 0 minutes and 1 hour. The phosphocellulose
paper was washed 4 times with 0.5% phosphoric acid, added to vials containing
scintillation fluid, and counted.
Structural Modeling
The X-ray crystal structures of Plkl (PDB ID: 20WB), Aurora B (PDB ID:
2BFX), and Nek2 (PDB ID: 2JAV) were used as base models. Peptides were manually
docked into the substrate binding cleft based on threading the structures of the kinases
onto the structure of the PKA:PKI complex (38). Molecular surface representations of the
Aurora B, Nek2, and Plkl active sites were created with GRASP (33, 34, 36, 37) and
shaded by electrostatic potential using surface projections of charge calculated with
DelPhi (37).
Centrosome and Spindle Enrichment Informatics
A list of proteins associated with the centrosome was obtained from mass
spectrometry data of Andersen et al. (39). A list of spindle proteins was obtained from
mass spectrometry data of Sauer et al. (40). Sequences for these proteins were
downloaded directly from NCBI Entrez Protein using NCBI Entrez Utilities
(http://www.ncbi.nlm.nih.gov/entrez/query/static/eutils_help.html). The human proteome
was downloaded as version 3.23 of the file ipi.HUMAN.fasta from the International
Protein Index (http://www.ebi.ac.uk/IPI/), and a list was generated of the amino acid
sequences surrounding each serine or threonine amino acid residue within the sequences
of each protein in each of these three data sources. The human proteome yielded
1850231 sites from 66619 proteins. The centrosome data and spindle data yielded 35425
sites from 524 proteins and 34909 sites from 277 proteins, respectively. For each motif
of interest, statistical significance of that motif in the centrosomal or spindle proteomes
was calculated using the hypergeometric distribution:
P=
where N is the number of S/T sites in the human proteome, n is the number of S/T sites in
the centrosomal or spindle proteome, m is the number of motif sites in the human
proteome, and k is the number of motif sites in the centrosomal or spindle proteome.
This corresponds to the probability of seeing as many instances or more of the motif as
are seen in the centrosomal or spindle proteome by chance if drawing a dataset the same
size as the centrosomal or spindle proteome at random from the human proteome.
Results
Determination of the Optimal Consensus Motif of Plk1
To address our question of how major mitotic kinases phosphorylate distinct sites
and to aid in substrate identification, we determined the optimal consensus motif of Plkl
by Positional Scanning Oriented Peptide Library Screening (Figure 2.2A). Plkl showed
strong selection for D, N, or E in the -2 position relative to the S/T phospho-acceptor
residue, as evidenced by darker spots on the autoradiogram of the spotted streptavidin
membrane. Plkl also showed strong selection for hydrophobic amino acids in the +1
position, particularly F, Y, I, and M. Interestingly, P in the +1 was strongly discriminated
against as indicated by the arrowhead in the right panel of Figure 2.2A.
To validate the PS-OPLS results, we verified the motif revealed by this technique
by determining kinetic parameters for Plkl-dependent phosphorylation of the optimal
peptide (GHDTSFYWAAYKKKK) as well as peptides containing single amino acid
mutations in optimal peptide (Figure 2.2B).
Vmx/Km ratios of these variant peptides
followed the same trends as displayed in the PS-OPLS. The optimal peptide had neither
the lowest Km nor the highest Vmax, but had the highest Vmx/Km ratio which is consistent
with the peptide library screening approach determining optimal substrate sequences
based on maximum Vmax/Km ratio (i.e. maximal turnover). All variations of the optimal
peptide tested resulted in decreased Vma/Km ratios suggesting that the peptide chosen as
the optimal peptide is correct and verifying that the optimal peptide can be determined by
PS-OPLS by choosing the most highly selected amino acid in each position. Particularly,
changing the D to an A in the -2 resulted in a greater than 20-fold drop in the Vmx/Km
ratio and activity of Plkl against the peptide with P in the +1 was so low that it was not
possible to fit the data to a Michaelis-Menten curve. The lack of activity against this
peptide recapitulated the lack of activity against P in the +1 in the PS-OPLS (Figure
2.2A). Based on these data the optimal consensus sequence is D/N/E-X-S/T-4, where X
is any amino acid and D is hydrophobic amino acids except P. (Figure 2.2A).
To investigate the structural basis for the optimal motif we performed molecular
modeling studies using a previously published structure of the Plkl kinase domain and a
-5
0
YAIXXXXS/TXXXXAGKK-Biotin
A.
I
... 19 ...
-5o5o
S-400
o
[384 well plate: FIll 198
wells with in vitro
o:s,)
Buffer
2. Wash
Kinase
MgCI2
32
y- P -A
ATP
Buffer
Minimal Consensus Motif D/NIE-X-S/T-0Q(not P)
iduein Fied Ption
W
\
1. Transfer to Streptavidin Membrane
2. Wash
.
3. Image on Phosphorimager
8.
120g
E 8D
1D
>
40
repiae
Optimal
D-2N tide
D-2R tide
D-2A tide
F+1Atide
F+1E tide
545
273
5125
3478
10792
12765
1253
234
1821
371
204
114
2.3
0.86
0.36
0.11
0.02
0.01
F+1P tide
-
-
-
% (,pM) V,
l(pmouminipg)
V
lI
m
Optimal: GHDTSFYWAAKKKK
20
0
1000
2000
3000
[peptide] (pM)
C:ST
o:Kr
Figure 2.2: Plkl Selects D, N, or E in the -2 relative to the phospho-acceptor S/T and
hydrophobic amino acids in the +1, but strongly discriminates against P in the +1 in
opposition to Cdkl. A. Experimental flow of Positional Scanning Oriented Peptide
Library Screening (PS-OPLS) technique for determination of kinase substrate
specificity (left) and PS-OPLS blot of Plkl T210D kinase domain (a.a. 38-346). B.
Determination of kinetic parameters (Km, Vmax, and Vmax/Km ratio) for reactions of
Plkl T210D kinase domain with the optimal substrate peptide and peptides with single
amino acid changes from the optimal. Graph of reaction data fitted to the MichaelMenten equation for selected peptides as indicated (left) and table of kinetic
parameters (right). C. PS-OPLS blot of Cdkl/Cyclin B.
model of the PKA:substrate complex as base models (Figure 2.3C) (36). The optimal
peptide sequence, HDTSFYWA, was modeled into the active site in an extended
conformation. Selection for D and E in the -2 seems to be based on an electrostatic on
electrostatic and/or hydrogen-bonding interaction with K178.
This conclusion is
supported by the observation that mutation of the -2 residue to one incapable of forming
either type of interaction (i.e. A) reduced the Km by 6-fold, while replacement with a
positively charged residue (i.e. R) reduced the Km nearly 10-fold.
F and other
hydrophobic amino acids selected in the +1 appear to fit into a hydrophobic pocket
formed by L211, P215, and 1218 in the activation loop.
Substitution of small or
negatively charged amino acids in this position reduced the Km by a factor of 20 or
greater.
The motif we determined is consistent with mapped Plkl phosphorylation sites, of
which several are known (Figure 2.3A). Most Plkl phosphorylation sites contain a D or
E in the -2 and a hydrophobic amino acid in the +1 position. This motif is also consistent
with a previously published motif for Plkl: D/E-X-S/T-(p-X-D/E, where (p is any
hydrophobic amino acid (41). However, the new motif differs from the previous motif
since it expands the -2 position selection to include N, which has a similar level of
selectivity to that of D or E in that position and demonstrates no selection for D or E in
the +3 position (Figure 2.2A).
This expanded motif suggested that there may be novel substrates containing Plkl
phosphorylation sites with N in the -2 relative to the phospho-acceptor. To identify such
substrates we performed a phospho-proteomic
screen (Figure 2.3B).
Plkl
phosphorylation was inhibited using the Plkl specific inhibitor, BI 2536, in nocodozole
arrested HeLa cells. Phosphorylations on specific mitotic complexes were mapped by
mass spectrometry. In parallel, phosphorylations were mapped on these same mitotic
complexes from nocodozole arrested, but non-Plkl-inhibited HeLa cells.
Phosphorylations found in the non-Plkl-inhibited data set, but not in the inhibited data set
indicate Plkl dependent phosphorylation.
Many Plkl phosphorylation sites were
determined (data not shown) including 3 with N in the -2 position relative to the
phosphorylation (Figure 2.3C). Two of these proteins, Bubl and p31/Comet, are
involved in spindle assembly checkpoint and the third, Sccl/Rad21, is involved in sister
A.
Substrate
Protein
Brca2
BubR1
Cdc25C
Kizuna
MKLP2
Myti
Npml
TCTP
Weel
Mapped Plkl
Phosphorvlation Site(s)
DMS'39 W, DES231L
ELT•"•V, EAT 1'"V
EFS'R"L
DLT1'91
EHS' 2 L
49
6
L,DDS 435L, DLS69*D, EDT 1L
DSS"2
4
EDS M
DDS4L, TESuT
EDS 53A
References
2536
(57)
(58)
(59,60)
(61)
(62)
(41,63)
(64)
(65)
(66)
C,
Substrate
Protein
Sccl/Rad21
p31/Comet
Bubl
Mapped Plkl
Phosphorylation Site
NQS'~R
NAS'TE
NKS4CT
oteins
Figure 2.3: The substrate specificity motif of Plkl matches previously mapped PlkI
phosphorylation sites and aids in the identification of new N in the -2 sites. A.
Previously published mapped Plkl phosphorylation sites contain D or E in the -2 and
hydrophopbic amino acids in the +1. B. Experimental strategy for the identification
of new N in the -2 Plkl substrate sites. C.
Table of mapped N in the -2 Plkl
phosphorylation sites identified as in B. D. Molecular surface representation of the
Plkl active site with the optimal substrate peptide, HDTSFYWA, shown in stick
representation.
(positive).
The electrostatic potentials are colored red (negative) and blue
chromatid cohesion in mitosis (7, 42, 43). Regulation of the spindle assembly checkpoint
and sister chromatid cohesion are known roles of Plkl (44, 45).
The strong discrimination of Plkl against phosphorylating sites with P in the +1
suggests a possible mechanism by which Cdkl/Cyclin B and Plkl could phosphorylate
mutually exclusive sites. Cdkl is a proline-directed kinase with the consensus
phosphorylation motif, S/T-P-X-K/R (46). A P in the +1 is essential for Cdkl to
phosphorylate a site, which is indicated by the lack of selection of other amino acids in
the +1 position of the PS-OPLS blot (Figure 2.2C.) Since P in the +1 is essential Cdklmediated for phosphorylation and Plkl seems incapable of phosphorylating peptides with
P in the +1, Cdkl and Plkl should not be able to phosphorylate the same sites on
proteins. Aurora A has also been shown previously to strongly discriminate against sites
with P in the +1 suggesting that Aurora A and Cdkl also phosphorylate mutual exclusive
sites (47).
Determination of the Optimal Consensus Motif of Aurora A and Aurora B
To verify the discrimination against P in the +1 by Aurora A and to determine if
this is a more general mode of regulation between Cdkl and the major mitotic kinases,
we determined the motifs of Aurora A and Aurora B by PS-OPLS (Figure 2.4A). Both
Aurora A and Aurora B showed extremely strong selection for R in the -2 and strong
discrimination against all other amino acids in the -2. Surprisingly, even K in the -2 is a
poor substitute for R. Additionally, there is a much smaller selectivity for R in the -3
position. In the +1 position, although both kinases selected hydrophobic amino acids, the
specific set of amino acids selected differed slightly between Aurora A and B. Aurora B
selected I and M most strongly, while Aurora A selected F and I in addition to L and M.
Importantly, both Aurora A and Aurora B showed strong discrimination against P in the
+1. The similarity in overall phosphorylation motifs between Aurora A and B is not
surprising as these kinases have a high level of sequence similarity.
This result verified the previous report of Aurora A being largely incapable of
phosphorylating sites with P in the +1. To independently verify the PS-OPLS results for
Aurora B, we determined kinetic parameters for Aurora B-dependent phosphorylation of
the optimal peptide (ARRHSMGWAYKKK) and peptides with single amino acid
A.
K I H GFEDCAVTSRQPNML:YIpTYW
-
-5:X
-4:X
-3:X
-5:X
-4:X
-3:X
K I HGFEDCAVTSRQPNMLpYpTYW
-2:X
-2:X
,~-,,,&- X
0:
+2:X
+2:X
+3:X
+4:X
+3:X
+4:X
4000
3001
Peptide
Km (pM)
Vm, (pmol/min/pg) Vmx / K
Optimal
R-2D tide
247
1238
5425
2085
22
1.7
M+1P tide
--
--
--
Optimal: ARRHSMGWAYKKKK
e
2001
e
*
100(
Ce
0
Substrate
Protein
AurkA
Cdc25B8
HURP
Lats2
MBD3
p53
RalA
400
200
[peptide] (pM)
Mapped Aurora A
Phosphorylation Site(s)
RTT-"L
RRS V
RMS:"2L
RYSmL'L
RRS24G
RHS"IV
RKS'"L
References
(47)
(67)
(68)
(69)
(70)
(71)
(72)
Substrate
Protein
AurKB
CENP-A
Desmin
Histone H3
INCENP
MCAK
MgcRacGAP
Vimentin
References
Mapped Aurora B
Phosphorylation Site(s)
(73)
RKTm2 M
(74)
RRS7R
(75)
RTS %G
(13, 76)
RKS10T,RKS-"A
(50. 77)
RTS111S
(16, 78)
RKS195C
RLS'"T RRS"*T, RST'"S, RISIIG (79, 80)
(73, 81)
RSS5"V
Figure 2.4: Aurora A and Aurora B strongly select R in the -2 and hydrophobic
amino acids in the +1, but strongly discriminate against P in the +1. A. PS-OPLS
blots of Aurora A (left) and Aurora B:INCENP complex (right). B. Determination of
kinetic parameters (Ki, Vmax, and Vmax/Km ratio) for reactions of Aurora B:INCENP
with the optimal substrate peptide and peptides with single amino acid changes from
the optimal. Graph of reaction data fitted to the Michael-Menten equation for peptides
as indicated (left) and table of kinetic parameters (right). C. Previously published
mapped Aurora A (left) and Aurora B (right) phosphorylation sites contain R in the -2
and hydrophopbic amino acids in the +1. D. Molecular surface representation of the
Aurora B active site with the optimal substrate peptide, RRHSMGW, shown in stick
representation.
The electrostatic potentials are colored red (negative) and blue
(positive).
changes in which the R in the -2 was changed to a D and the M in the +1 was changed to
a P (Figure 2.4B). Changing R to D resulted in a 13-fold drop in Vm/Km as compared to
the optimal peptide. As with Plkl, the activity of Aurora B against the P in the +1
peptide so low that it was not possible to fit the data to a Michaelis-Menten equation.
Based on these data the minimal optimal consensus sequence for both Aurora A and
Aurora B is R-X-S/T-(, where X is any amino acid and Q is any hydrophobic amino acid
except P.
To investigate the structural basis for the optimal motif we performed molecular
modeling studies using the previously published structure of the Aurora B kinase domain
(Figure 2.4C) (33). The optimal peptide sequence, RRHSMGW, was modeled into the
active site in an extended conformation. Selection for R in the -3 seems to be based on
an electrostatic interaction with E177. Selection for R in the -2 is likely due to an
electrostatic interaction with E220 and E281. Consistent with this model, replacement of
the -2 R by a D resulted in a 5-fold decrease in Km. M and other hydrophobic amino
acids in the +1 position appear to fit into a hydrophobic pocket formed by W237 and
M249 on the activation loop and L256 of the active site cleft.
The motifs of both Aurora A and Aurora B are consistent with mapped Aurora A
and Aurora B phosphorylation sites (Figure 2.4C) as most Aurora phosphorylation sites
have an R in the -2 and a hydrophobic in the +1. The Aurora A motif presented here is
consistent with previously published consensus motifs (47, 48). The Aurora B motif, RX-S/T-D, corrects a previously deduced motif, R/K-R/K-X-S/T- c( (49, 50), which
claimed that K is an acceptable substitute for R in the -2 and allowed P in the +1. These
data suggest that Aurora A and Cdkl phosphorylate mutually exclusive sites, as do
Aurora B and Cdkl.
Determination of the Optimal Consensus Motif of Nek2
To determine if the other major mitotic kinase, Nek2, also follows the emerging
pattern of not phosphorylating sites with P in the +1, we performed PS-OPLS Nek2
(Figure 2.5A). Nek2 showed strong discrimination against non-hydrophobic amino acids
in the -3 position with particularly strong positive selection for F, L, and M. In the -2
position, Nek2 showed no strong preference, although R was slightly preferred to other
amino acids and P was disfavored. In the +1 position, Nek2 favored hydrophobic amino
acids and showed extremely strong discrimination against P in the +1. Nek2 also showed
discrimination against D and E in the + 1.
To verify the PS-OPLS data we measured kinetic parameters for Nek2-dependent
phosphorylation of its optimal peptide (WFRMSIRGGYKKK) and peptides containing
single amino acid changes from the optimal. Changing the F in the -3 of the optimal to a
V resulted in an approximately 8-fold decrease in the Vmax/Km as compared to the optimal
peptide, primarily due to a reduction in Km. V was chosen because it was neither favored
nor disfavored in the -3 position.
Non-hydrophobic amino acids, which were
discriminated against in the -3 position, would be expected to result in even larger
decreases Vmax/Km. As with Plkl and Aurora B, the activity of Nek2 against the P in the
+1 peptide was so low that it was not possible to fit the data to a Michaelis-Menten curve.
Based on these data the minimal optimal consensus motif for Nek2 is p-R/X-X-S/T-AR/H, where qp is any hydrophobic amino acid, X is any amino acid, and (D is any
hydrophobic amino acid except P.
To investigate the structural basis for the optimal motif we performed molecular
modeling studies using the previously published structure of the Nek2 kinase domain,
and since in this structure the activation loop is disordered, a model of the PKA:substrate
I I"1L•I-I"UL,/• V
-51
-4:)
-3:)
-2:)
I OK
•41"1•11VlLplpI ! vv
.~
I I, ~ r r ~rr r \, r c a r\ a ttl rr I rV~TV\~I
R
K I HGIF
I L A V I O
1.4r IN IVI L
i
r VV
04
Y~j
0:S/T>+-7)
+2:)
+3:)
+4:)
B.
120
100
E 80
E
Peptide
Optimal
F-3V tide
R-2D tide
K, (pM)
554
4597
13477
V, (pmol/min/pg) V,, / K,
143
0.26
153
0,033
677
0.050
I+1P tide
--
--
Optimal: WFRMSIRGGYKKKK
60
8 40
ptimal
3V tide
2D tide
1P tide
20
0
0
400
800
1200
[peptidel (pM)
1600
C,
Substrate
Protein
Nek2
Hecl
Dý
Mapped Nek2
References
Phosphorylation Site(s)
FAKT"-F
(34)
FALS "K
(82)
Figure 2.5: Nek2 strongly selects hydrophobic amino acids in the -3 and the +1, but
strongly discriminates against P in the +1. A. PS-OPLS blot of Nek2. B.
Determination of kinetic parameters (Km, Vmax, and Vmax/Km ratio) for reactions of
Nek2 with its optimal substrate peptide and peptides with single amino acid changes
from the optimal. Graph of reaction data fitted to the Michael-Menten equation for
peptides as indicated (left) and table of kinetic parameters (right). C. Previously
published mapped Nek2 phosphorylation sites contain hydrophobic amino acids in the
-3 and the +1. D. Molecular surface representation of the Nek2 active site with the
optimal substrate peptide, FRASIR, shown in stick representation. The electrostatic
potentials are colored red (negative) and blue (positive).
complex as base models (Figure 2.5C) (34). The optimal peptide sequence, FRASIR,
was modeled into the active site in an extended conformation. Selection for F and other
hydrophobic amino acids in the -3 seems to be based on a surface formed by two alanines
that are F and E, respectively, in PKA. Selection for R in the -2 may be due to an
electrostatic interaction with E208. I and other hydrophobic amino acids in the +1 of the
peptide likely fit into a hydrophobic pocket formed by F176 in the activation loop, and
P180 and M183 of the active site. The R and H in the +2 position of the peptide may
electrostatically interact with E48, which could also form hydrogen-bonds to hydroxyl
groups of S and T.
There are only two mapped Nek2 phosphorylation sites including the activation
loop autophosphorylation site, T175 (Figure 2.5D). The motif of Nek2 is consistent with
these mapped sites, which both contain F in the -3 position. No motif for Nek2 has been
published previously. Importantly, this motif confirms our hypothesis that Nek2, just like
Plkl, Aurora A, and Aurora B will not phosphorylate sites targeted by Cdkl and vice
versa.
Examination of the Role of Localization and Motifs on the Regulation of the Major
Mitotic Kinases
The motif data suggests a mode of regulation of mitotic kinase signaling based on
the presence or absence of a P in the +1 as a kind of switch allowing phosphorylation of a
site by either Plkl, Aurora A, Aurora B, or Nek2 if a residue other than proline is present
in the +1 or Cdkl if a proline is present in the +1. This regulation prevents Cdkl from
phosphorylating Plkl, Aurora A, Aurora B, or Nek2 sites and Plkl, Aurora A, Aurora B
and Nek2 from phosphorylating Cdkl sites. Previously, this proline specificity has been
suggested as a mode of regulation between Cdkl and Aurora A and between the CMGC
proline-directed kinase group, of which Cdkl is a member, and the basophilic CAMK
and AGC kinase groups, to which Aurora A and Aurora B are closely related but are not
members (47, 51). The data presented here expands this concept from just Aurora A to
include the other major mitotic kinases, Aurora B, Nek2, and Plkl.
To attempt to understand how the major mitotic kinases aside from Cdkl
phosphorylate mutually exclusive sites despite having overlapping localizations, we re-
examined the role of kinase localization on substrate selection performing an unbiased
bioinformatics analysis of potential sites for Plkl, Nek2, and Aurora kinases at the
centrosome and on the spindle apparatus compared to the proteome as a whole (Figure
2.6A). The Plkl motif was found to be enriched at the centrosome and spindle apparatus.
The Nek2 motif was enriched at the centrosome, but not on the spindle as might be
expected for a centrosomal protein. The Aurora motif was not significantly enriched at
the centrosome or the spindle. The lack of enrichment of Aurora kinase sites is, perhaps,
not surprising as there are many basophilic kinases that control processes outside of
mitosis. The enrichment of potential substrates at subcellular structures where Plkl and
Nek2 localize supports the intuitive idea that these kinases are likely to phosphorylate
proteins that co-localize to the areas of the cell with the largest abundance of these
kinases rather than to areas of low abundance of the kinase, and suggests that proteins at
these structures may have co-evolved to make use of the appropriate kinase for phosphodependent regulation.
This, in turn, suggests that localization may play a role in
maintaining mutually exclusivity of major mitotic kinase phosphorylation sites.
To further explore this emerging potential relationship between localization and
motifs in the regulation of major mitotic kinases we more closely examined the ability of
these kinases to phosphorylate the same sequences.
Although we have already
determined kinetic parameters for Plkl phosphorylating a D-2R variant of its optimal
peptide (Figure 2.2B) and Aurora B and Nek2 phosphorylating R-2D variants of their
optimal peptides (Figure 2.4B, 2.5B) and seen that these changes away from the optimal
resulted in significant reductions in the Vmax/Km ratio compared to the optimal, we
wanted to directly compare these kinases on a common set of peptides. The optimal
peptide of each kinase, Aurora B, Nek2, and Plkl, was reacted with each kinase in this
set (Figure 2.6B). As expected, each kinase phosphorylated its optimal peptide most
strongly. The amount of phosphorylation of the Plkl optimal peptide by Nek2 and
Aurora B was about 8% and less than 1% of the phosphorylation by Plkl, respectively.
The amount of Aurora B and Plkl phosphorylation of the Nek2 optimal peptide was
about 23% and less than 1% of the phosphorylation by Nek2, respectively. The extent of
Nek2 and Plkl phosphorylation of the Aurora B optimal peptide was about 1% of that of
Aurora B. These data suggest that kinases with non-overlapping localizations such
Enrichment Significance
Score
Kinase
Plkl
Nek2
Motif
[D/NIE]-X-[S/T]-[IILIMN/F/VV/Y
[IiL/M/V]-R-X[T-[S/T]-IIMN]
Centrosome
2,50e-11
2.32e-4
Spindle
2 83e-4
0.136
B.
100
--
n
Piki
Optimal
Nek2
Optimal
AurB
Optimal
Peptide
0 Plkl
O AurB
Aurora B
Aurora A - Aurora B Overlay
Ai nrmn A
[0 Nek2
F'
1
E
L
M.
U
L
-
-
ý -
- -
T|
Figure 2.6: The motif of Nek2 overlaps with that of the Aurora kinases and Plkl, but
the motifs of Plkl and the Aurora kinases are non-overlapping.
A.
Table of
significance scores for the enrichment of consensus motifs in centrosome and mitotic
spindle proteins.
B. Comparison of ability of each kinase (Plkl in black, Nek2 in
gray, and Aurora B in white) to phosphorylate each of the other kinases' optimal
substrate peptides as a percentage of the amount of phosphorylation of its own optimal
substrate peptide.
C. Pseudo-color PS-OPLS blots of Aurora A (yellow), Aurora B
(purple), Nek2 (cyan), and Plkl (red) were superimposed to show similarities and
differences between the motifs. Equal selectivity in the Aurora A-Aurora B overlay is
cyan, Aurora A-Nek2 overlay is red, Aurora A-Plkl overlay is purple, Nek2-Plkl
overlay is maroon, and Aurora B-Plkl overlay is green.
as Plkl or Aurora kinases and Nek2 may be able to phosphorylate the same sequences,
whereas kinases with overlapping subcellular localizations, such as Plkl and Aurora
kinases, are not able to phosphorylate the same sequences as a strict consequence of nonoverlapping phosphorylation motifs.
Discussion
Mitosis is an intricate and highly regulated process. Several kinases including the
master regulator Cdkl and the other major mitotic kinases, Aurora A, Aurora B, Nek2,
and Plkl, phosphorylate specific sites on specific proteins to orchestrate the process of
mitosis. To better understand the mechanisms underlying the roles these kinases play in
mitosis, we have determined unbiased, complete motifs of these kinases. The motif of
Cdkl has been published previously to be S/T-P-X-R/K, and we have verified this motif
(46). The motif of Nek2 is novel. The motifs of Plkl, Aurora A, and Aurora B have
been determined by PS-OPLS, with results that differed significantly from previously
published, but biased and incomplete motifs (41, 49, 50). We further verified the PSOPLS motifs by kinetic studies on individual peptides, testing the effects of altering
amino acids away from the optimal for individual positions. The motifs match known
substrates of these kinases and we were able to rationalize the structural basis for the
motifs using structural models of the kinase domains with the optimal peptide docked in.
As part of this study we have identified new Plkl substrates.
Motifs are
particularly useful when combined with phosphoproteomic mass spectrometry screens,
where they can help to identify the kinases that created the discovered phosphosites. The
previously reported motif of Plkl suggested that Plkl only phosphorylated sites with D or
E in the -2 position (41). PS-OPLS revealed the complete motif showing additional
specificity for N in the -2. The new motif was applied to a phosphoproteomic mass
spectrometry screen in which Plkl phosphorylation sites were determined by comparing
nocodozole arrested mitotic HeLa cells treated with the specific Plkl inhibitor BI 2536 to
HeLa cells just arrested in mitosis with nocodozole (52). Sites in Sccl/Rad21, Bubl, and
p31/Comet had N in the -2 matching the new Plkl consensus sequence and are Plkl
activity dependent suggesting that they are novel Plkl sites. Sccl is a subunit of the
cohesin complex that prevents chromosome segregation until it is cleaved by separase at
the metaphase-anaphase transition (53). Sccl is known to be phosphorylated at multiple
sites by Plkl enhancing the cleavage of Sccl by separase, although all of the sites have
not been mapped (53). The location of the phosphorylation site discovered in this study
is near one of the separase cleavage sites suggesting it may also function in increasing the
cleavability of Sccl. Bubl has previously been shown to be phosphorylated by Cdkl
recruiting Plkl to kinetochores, and Plkl has been shown to phosphorylate Bubl
although the site had not been determined (54). P31/Comet has recently been shown to
bind Mad2, causing Mad2 to release Cdc20 (55).
The Mad2-Cdc20 interaction is
required for maintenance of the spindle checkpoint until all kinetochores have obtained
amphitelic attachments (56). Plkl is known to be required for the recruitment of Mad2 to
the kinetochore for the spindle assembly checkpoint, but this phosphorylation in
p31/Comet suggests it may have other roles also (23).
Of the major mitotic kinases, Nek2 stands alone in its localization to the proximal
ends of the centrioles at the core of the centrosome. The other major mitotic kinases have
overlapping localizations with Aurora A, Plkl, and Cdkl localizing to the pericentriolar
material and spindle poles and Aurora B, Plkl, and Cdkl localizing to the kinetochores
and centromeres. As these kinases are all Serine/Threonine kinases, these overlapping
localizations suggest that these kinases will have access to the same proteins and,
therefore, the same potential phosphorylation sites. Since the mapped phosphorylation
sites of each kinase do not contain very many, if any, common sites with any of the other
kinases in the set, we wanted to understand how these kinases maintain their specificity.
The motifs in this study shed light on this problem. Figure 2.6C shows each motif
in pseudo-color and pairs of motifs superimposed. The overlay of the Aurora A motif in
green and Aurora B motif in purple shows how similar the motifs are, as the majority of
spots are neither purple nor green but bluish to cyan -- the R in the -2, in particular, which
dominates the motifs. As mentioned above, Aurora A and Aurora B have mutually
exclusive localizations during mitosis, which suggests that these kinases, despite having
essentially identical motifs, also phosphorylate mutually exclusive sites as they do not
have access to the same sites unless the protein being phosphorylated has a localization
overlapping with the localizations of both Aurora A and B. This is supported by the
known substrates for these kinases (Figure 2.4C). Additional proteins known to form
complexes with Aurora-A and -B, such as Bora and Survivin and INCENP, respectively,
may also play a role in substrate selection.
The Nek2-Plkl and Aurora A-Plkl motif overlays show the selectivity in the -2 of
the Plkl and Aurora A motifs and the selectivity in the -3 of the Nek2 motif. The optimal
peptide of Nek2 contains an R in the -2 which suggests that optimal Nek2 sequences
might be phosphorylatable by the Aurora kinases, which we have verified on individual
peptides (Figure 2.6B). Nek2 may also be able to phosphorylate the same sequences as
Plkl if such sequences had hydrophobic amino acids in the -3 to satisfy the Nek2 motif,
which we have verified on individual peptides (Figure 2.6B). However, Nek2 has been
shown to localize to the proximal centrioles at the core of the centrosome, whereas
Aurora A localizes to the periphery of the pericentriolar material, where Plkl is also
thought to localize (26). Nek2 and Aurora A or Plkl should, therefore, not be able to
phosphorylate the same sites unless the substrate localizes to both the core of the
centrosome and to the periphery of the pericentriolar material.
In the Aurora A-Plkl motif overlay, with Aurora A in green and Plkl in cyan, the
R in the -2 is green and the D, N, and E in the -2 are cyan. Similarly, in the Aurora-BPlkl motif overlay with Aurora B in purple and Plkl in cyan, the R in the -2 is purple and
the D, N, and E in the -2 are cyan. This difference in selection in the -2 suggests that
Plkl and Aurora kinases, despite having overlapping localizations, will not phosphorylate
the same sequences, which we verified using peptides (Figure 2.6B).
Integrating all of this data suggested to us a mode of regulation to enforce
specificity of phosphorylation by these kinases. Plkl and the Aurora kinases do not
phosphorylate the same sites despite overlapping localizations because their motifs are
non-overlapping. Nek2 and Plkl and the Aurora kinases do not phosphorylate the same
sites despite overlapping motifs because of non-overlapping localizations.
Thus we
hypothesize that the major mitotic kinases with overlapping localizations do not have
overlapping motifs and major mitotic kinases with overlapping motifs do not have
overlapping localizations.
Another way to visualize this hypothesis is shown in the Venn diagrams depicted
in Figure 2.7. Kinases exist in two functionally orthogonal spaces: 'localization space'
and 'motif space'. 'Localization space' reflects all of the subcellular locales where a
protein can reside. 'Motif space' contains the optimal sequence motifs of all S/T kinases.
In 'localization space', the red circle represents the localization of Cdkl, which overlaps
Motif Space
Localization Soace
m
m
I
ho
Figure 7: An alternate view of the proposed hypothesis in which kinase functionality
is represented by Venn diagrams of 'localization space' and 'motif space'.
In
'localization space', each circle represents the total subcellular locales available to the
kinase.
In 'motif space', each circle represents the sequences that can be
phosphorylated by the kinase.
In this representation, a major mitotic kinase can
overlap with every other kinase in at most one of these two spaces.
with all of the other major mitotic kinase.
Similarly, Plkl, whose localization is
represented by the blue circle, overlaps with the localizations of Aurora A (green) and
Aurora B (purple) localizations, but not Nek2 (yellow). The 'motif space' panel is
interpreted similarly. In the context of these two spaces an equivalent statement of the
hypothesis becomes evident: Each major mitotic kinase can overlap every other kinase
in, at most, one of localization or motif space.
As examples, Cdkl overlaps all of the
other kinases in 'localization space', but not in 'motif space' and Nek2 overlaps Plkl and
Aurora Kinases in 'motif space', but not in localization spaces'. This representation begs
the question of how temporal aspects of 'regulation fit into this model. Although, prior
events undoubtedly play roles in which substrates and which sites in these substrates are
phosphorylated by the major mitotic kinases, this aspect of regulation is beyond the scope
of this study.
This hypothesis, if true, implies a coordinated regulation of these kinases. It has
been known for some time that motifs and localization direct individual kinases to their
substrates.
However, what has been suggested by our data is that the motifs and
localizations of individual kinases may be cooperatively directing a group of kinases to
their substrates and that the major mitotic kinases may be under a coordinated selective
pressure to maintain their motifs and localizations, suggesting a systems-level regulation
and evolution of the major mitotic kinases and their substrates.
References
1. Kops GJ, Weaver BA, Cleveland DW. On the road to cancer: Aneuploidy and the
mitotic checkpoint. Nat Rev Cancer. 2005 Oct;5(10):773-85.
2. Jallepalli PV, Lengauer C. Chromosome segregation and cancer: Cutting through the
mystery. Nat Rev Cancer. 2001 Nov;1(2):109-17.
3. Nigg EA. Mitotic kinases as regulators of cell division and its checkpoints. Nat Rev
Mol Cell Biol. 2001 Jan;2(1):21-32.
4. Lamb NJ, Fernandez A, Watrin A, Labbe JC, Cavadore JC. Microinjection of p34cdc2
kinase induces marked changes in cell shape, cytoskeletal organization, and chromatin
structure in mammalian fibroblasts. Cell. 1990 Jan 12;60(1):151-65.
5. Heald R, McKeon F. Mutations of phosphorylation sites in lamin A that prevent
nuclear lamina disassembly in mitosis. Cell. 1990 May 18;61(4):579-89.
6. Peter M, Nakagawa J, Doree M, Labbe JC, Nigg EA. In vitro disassembly of the
nuclear lamina and M phase-specific phosphorylation of lamins by cdc2 kinase. Cell.
1990 May 18;61(4):591-602.
7. Morrow CJ, Tighe A, Johnson VL, Scott MI, Ditchfield C, Taylor SS. Bubl and aurora
B cooperate to maintain BubR1-mediated inhibition of APC/CCdc20. J Cell Sci. 2005
Aug 15;118(Pt 16):3639-52.
8. Ookata K, Hisanaga S, Bulinski JC, Murofushi H, Aizawa H, Itoh TJ, Hotani H,
Okumura E, Tachibana K, Kishimoto T. Cyclin B interaction with microtubuleassociated protein 4 (MAP4) targets p34cdc2 kinase to microtubules and is a potential
regulator of M-phase microtubule dynamics. J Cell Biol. 1995 Mar;128(5):849-62.
9. Larsson N, Marklund U, Gradin HM, Brattsand G, Gullberg M. Control of microtubule
dynamics by oncoprotein 18: Dissection of the regulatory role of multisite
phosphorylation during mitosis. Mol Cell Biol. 1997 Sep;17(9):5530-9.
10. Bentley AM, Normand G, Hoyt J, King RW. Distinct sequence elements of cyclin B 1
promote localization to chromatin, centrosomes, and kinetochores during mitosis. Mol
Biol Cell. 2007 Dec;18(12):4847-58.
11. Marumoto T, Honda S, Hara T, Nitta M, Hirota T, Kohmura E, Saya H. Aurora-A
kinase maintains the fidelity of early and late mitotic events in HeLa cells. J Biol Chem.
2003 Dec 19;278(51):51786-95.
12. Mori D, Yano Y, Toyo-oka K, Yoshida N, Yamada M, Muramatsu M, Zhang D, Saya
H, Toyoshima YY, Kinoshita K, Wynshaw-Boris A, Hirotsune S. NDEL1
phosphorylation by aurora-A kinase is essential for centrosomal maturation, separation,
and TACC3 recruitment. Mol Cell Biol. 2007 Jan;27(1):352-67.
13. Goto H, Yasui Y, Nigg EA, Inagaki M. Aurora-B phosphorylates histone H3 at
serine28 with regard to the mitotic chromosome condensation. Genes Cells. 2002
Jan;7(1):11-7.
14. DeLuca JG, Gall WE, Ciferri C, Cimini D, Musacchio A, Salmon ED. Kinetochore
microtubule dynamics and attachment stability are regulated by Hecl. Cell. 2006 Dec
1;127(5):969-82.
15. Sandall S, Severin F, McLeod IX, Yates JR,3rd, Oegema K, Hyman A, Desai A. A
Birl-Sli 15 complex connects centromeres to microtubules and is required to sense
kinetochore tension. Cell. 2006 Dec 15;127(6): 1179-91.
16. Knowlton AL, Lan W, Stukenberg PT. Aurora B is enriched at merotelic attachment
sites, where it regulates MCAK. Curr Biol. 2006 Sep 5;16(17):1705-10.
17. Fry AM, Mayor T, Meraldi P, Stierhof YD, Tanaka K, Nigg EA. C-Napl, a novel
centrosomal coiled-coil protein and candidate substrate of the cell cycle-regulated protein
kinase Nek2. J Cell Biol. 1998 Jun 29;141(7):1563-74.
18. Fu Z, Larson KA, Chitta RK, Parker SA, Turk BE, Lawrence MW, Kaldis P,
Galaktionov K, Cohn SM, Shabanowitz J, Hunt DF, Sturgill TW. Identification of yinyang regulators and a phosphorylation consensus for male germ cell-associated kinase
(MAK)-related kinase. Mol Cell Biol. 2006 Nov;26(22):8639-54.
19. Sunkel CE, Glover DM. Polo, a mitotic mutant of drosophila displaying abnormal
spindle poles. J Cell Sci. 1988 Jan;89 ( Pt 1)(Pt 1):25-38.
20. van Vugt MA, van de Weerdt BC, Vader G, Janssen H, Calafat J, Klompmaker R,
Wolthuis RM, Medema RH. Polo-like kinase-1 is required for bipolar spindle formation
but is dispensable for anaphase promoting complex/Cdc20 activation and initiation of
cytokinesis. J Biol Chem. 2004 Aug 27;279(35):36841-54.
21. Lane HA, Nigg EA. Antibody microinjection reveals an essential role for human
polo-like kinase 1 (Plkl) in the functional maturation of mitotic centrosomes. J Cell Biol.
1996 Dec;135(6 Pt 2):1701-13.
22. Elia AE, Cantley LC, Yaffe MB. Proteomic screen finds pSer/pThr-binding domain
localizing Plkl to mitotic substrates. Science. 2003 Feb 21 ;299(5610): 1228-31.
23. Ahonen LJ, Kallio MJ, Daum JR, Bolton M, Manke IA, Yaffe MB, Stukenberg PT,
Gorbsky GJ. Polo-like kinase I creates the tension-sensing 3F3/2 phosphoepitope and
modulates the association of spindle-checkpoint proteins at kinetochores. Curr Biol. 2005
Jun 21;15(12):1078-89.
24. Wong OK, Fang G. Plxl is the 3F3/2 kinase responsible for targeting spindle
checkpoint proteins to kinetochores. J Cell Biol. 2005 Aug 29;170(5):709-19.
25. van Vugt MA, Medema RH. Getting in and out of mitosis with polo-like kinase-1.
Oncogene. 2005 Apr 18;24(17):2844-59.
26. Roghi C, Giet R, Uzbekov R, Morin N, Chartrain I, Le Guellec R, Couturier A, Doree
M, Philippe M, Prigent C. The xenopus protein kinase pEg2 associates with the
centrosome in a cell cycle-dependent manner, binds to the spindle microtubules and is
involved in bipolar mitotic spindle assembly. J Cell Sci. 1998 Mar;111 ( Pt 5)(Pt 5):55772.
27. Crosio C, Fimia GM, Loury R, Kimura M, Okano Y, Zhou H, Sen S, Allis CD,
Sassone-Corsi P. Mitotic phosphorylation of histone H3: Spatio-temporal regulation by
mammalian aurora kinases. Mol Cell Biol. 2002 Feb;22(3):874-85.
28. Jackman M, Firth M, Pines J. Human cyclins B 1 and B2 are localized to strikingly
different structures: B to microtubules, B2 primarily to the golgi apparatus. EMBO J.
1995 Apr 18;14(8):1646-54.
29. Lowe M, Rabouille C, Nakamura N, Watson R, Jackman M, Jamsa E, Rahman D,
Pappin DJ, Warren G. Cdc2 kinase directly phosphorylates the cis-golgi matrix protein
GM130 and is required for golgi fragmentation in mitosis. Cell. 1998 Sep 18;94(6):78393.
30. Draviam VM, Orrechia S, Lowe M, Pardi R, Pines J. The localization of human
cyclins Bl and B2 determines CDK1 substrate specificity and neither enzyme requires
MEK to disassemble the golgi apparatus. J Cell Biol. 2001 Mar 5;152(5):945-58.
31. Kufer TA, Sillje HH, Korner R, Gruss OJ, Meraldi P, Nigg EA. Human TPX2 is
required for targeting aurora-A kinase to the spindle. J Cell Biol. 2002 Aug
19;158(4):617-23.
32. Satinover DL, Leach CA, Stukenberg PT, Brautigan DL. Activation of aurora-A
kinase by protein phosphatase inhibitor-2, a bifunctional signaling protein. Proc Natl
Acad Sci U S A. 2004 Jun 8;101(23):8625-30.
33. Sessa F, Mapelli M, Ciferri C, Tarricone C, Areces LB, Schneider TR, Stukenberg
PT, Musacchio A. Mechanism of aurora B activation by INCENP and inhibition by
hesperadin. Mol Cell. 2005 Apr 29;18(3):379-91.
34. Rellos P, Ivins FJ, Baxter JE, Pike A, Nott TJ, Parkinson DM, Das S, Howell S,
Fedorov O, Shen QY, Fry AM, Knapp S, Smerdon SJ. Structure and regulation of the
human Nek2 centrosomal kinase. J Biol Chem. 2007 Mar 2;282(9):6833-42.
35. Hutti JE, Jarrell ET, Chang JD, Abbott DW, Storz P, Toker A, Cantley LC, Turk BE.
A rapid method for determining protein kinase phosphorylation specificity. Nat Methods.
2004 Oct;1(1):27-9.
36. Kothe M, Kohls D, Low S, Coli R, Cheng AC, Jacques SL, Johnson TL, Lewis C,
Loh C, Nonomiya J, Sheils AL, Verdries KA, Wynn TA, Kuhn C, Ding YH. Structure of
the catalytic domain of human polo-like kinase 1.Biochemistry. 2007 May
22;46(20):5960-71.
37. Nicholls A, Bharadwaj R, Honig B. Grasp: Graphical representation and analysis of
surface properties. Biophys J. 1993;64:A166.
38. Knighton DR, Zheng JH, Ten Eyck LF, Xuong NH, Taylor SS, Sowadski JM.
Structure of a peptide inhibitor bound to the catalytic subunit of cyclic adenosine
monophosphate-dependent protein kinase. Science. 1991 Jul 26;253(5018):414-20.
39. Andersen JS, Wilkinson CJ, Mayor T, Mortensen P, Nigg EA, Mann M. Proteomic
characterization of the human centrosome by protein correlation profiling. Nature. 2003
Dec 4;426(6966):570-4.
40. Sauer G, Korner R, Hanisch A, Ries A, Nigg EA, Sillje HH. Proteome analysis of the
human mitotic spindle. Mol Cell Proteomics. 2005 Jan;4(1):35-43.
41. Nakajima H, Toyoshima-Morimoto F, Taniguchi E, Nishida E. Identification of a
consensus motif for plk (polo-like kinase) phosphorylation reveals Mytl as a Plkl
substrate. J Biol Chem. 2003 Jul 11;278(28):25277-80.
42. Yang M, Li B, Tomchick DR, Machius M, Rizo J, Yu H, Luo X. p3lcomet blocks
Mad2 activation through structural mimicry. Cell. 2007 Nov 16;131(4):744-55.
43. Vagnarelli P, Morrison C, Dodson H, Sonoda E, Takeda S, Earnshaw WC. Analysis
of Scc I-deficient cells defines a key metaphase role of vertebrate cohesin in linking sister
kinetochores. EMBO Rep. 2004 Feb;5(2):167-71.
44. Lenart P, Petronczki M, Steegmaier M, Di Fiore B, Lipp JJ, Hoffmann M, Rettig WJ,
Kraut N, Peters JM. The small-molecule inhibitor BI 2536 reveals novel insights into
mitotic roles of polo-like kinase 1. Curr Biol. 2007 Feb 20;17(4):304-15.
45. Alexandru G, Uhlmann F, Mechtler K, Poupart MA, Nasmyth K. Phosphorylation of
the cohesin subunit Sccl by Polo/Cdc5 kinase regulates sister chromatid separation in
yeast. Cell. 2001 May 18;105(4):459-72.
46. Songyang Z, Blechner S, Hoagland N, Hoekstra MF, Piwnica-Worms H, Cantley LC.
Use of an oriented peptide library to determine the optimal substrates of protein kinases.
Curr Biol. 1994 Nov 1;4(11):973-82.
47. Ferrari S, Marin O, Pagano MA, Meggio F, Hess D, El-Shemerly M, Krystyniak A,
Pinna LA. Aurora-A site specificity: A study with synthetic peptide substrates. Biochem
J. 2005 Aug 15;390(Pt 1):293-302.
48. Ohashi S, Sakashita G, Ban R, Nagasawa M, Matsuzaki H, Murata Y, Taniguchi H,
Shima H, Furukawa K, Urano T. Phospho-regulation of human protein kinase aurora-A:
Analysis using anti-phospho-Thr288 monoclonal antibodies. Oncogene. 2006 Dec
14;25(59):7691-702.
49. Cheeseman IM, Anderson S, Jwa M, Green EM, Kang J, Yates JR,3rd, Chan CS,
Drubin DG, Barnes G. Phospho-regulation of kinetochore-microtubule attachments by
the aurora kinase Ipllp. Cell. 2002 Oct 18;111(2):163-72.
50. Honda R, Korner R, Nigg EA. Exploring the functional interactions between aurora
B, INCENP, and survivin in mitosis. Mol Biol Cell. 2003 Aug;14(8):3325-41.
51. Zhu G, Fujii K, Belkina N, Liu Y, James M, Herrero J, Shaw S. Exceptional disfavor
for proline at the P + 1 position among AGC and CAMK kinases establishes reciprocal
specificity between them and the proline-directed kinases. J Biol Chem. 2005 Mar
18;280(11):10743-8.
52. Steegmaier M, Hoffmann M, Baum A, Lenart P, Petronczki M, Krssak M, Gurtler U,
Garin-Chesa P, Lieb S, Quant J, Grauert M, Adolf GR, Kraut N, Peters JM, Rettig WJ. BI
2536, a potent and selective inhibitor of polo-like kinase 1, inhibits tumor growth in vivo.
Curr Biol. 2007 Feb 20;17(4):316-22.
53. Hauf S, Roitinger E, Koch B, Dittrich CM, Mechtler K, Peters JM. Dissociation of
cohesin from chromosome arms and loss of arm cohesion during early mitosis depends
on phosphorylation of SA2. PLoS Biol. 2005 Mar;3(3):e69.
54. Qi W, Tang Z, Yu H. Phosphorylation- and polo-box-dependent binding of Plkl to
Bubl is required for the kinetochore localization of Plkl. Mol Biol Cell. 2006
Aug; 17(8):3705-16.
55. Xia G, Luo X, Habu T, Rizo J, Matsumoto T, Yu H. Conformation-specific binding
of p3l(comet) antagonizes the function of Mad2 in the spindle checkpoint. EMBO J.
2004 Aug 4;23(15):3133-43.
56. Fang G, Yu H, Kirschner MW. The checkpoint protein MAD2 and the mitotic
regulator CDC20 form a ternary complex with the anaphase-promoting complex to
control anaphase initiation. Genes Dev. 1998 Jun 15;12(12):1871-83.
57. Lin HR, Ting NS, Qin J, Lee WH. M phase-specific phosphorylation of BRCA2 by
polo-like kinase 1 correlates with the dissociation of the BRCA2-P/CAF complex. J Biol
Chem. 2003 Sep 19;278(38):35979-87.
58. Matsumura S, Toyoshima F, Nishida E. Polo-like kinase 1 facilitates chromosome
alignment during prometaphase through BubR1. J Biol Chem. 2007 May
18;282(20):15217-27.
59. Elia AE, Cantley LC, Yaffe MB. Proteomic screen finds pSer/pThr-binding domain
localizing Plkl to mitotic substrates. Science. 2003 Feb 21;299(5610):1228-31.
60. Toyoshima-Morimoto F, Taniguchi E, Nishida E. Plkl promotes nuclear translocation
of human Cdc25C during prophase. EMBO Rep. 2002 Apr;3(4):341-8.
61. Oshimori N, Ohsugi M, Yamamoto T. The Plk1 target kizuna stabilizes mitotic
centrosomes to ensure spindle bipolarity. Nat Cell Biol. 2006 Oct;8(10):1095-101.
62. Neef R, Preisinger C, Sutcliffe J, Kopajtich R, Nigg EA, Mayer TU, Barr FA.
Phosphorylation of mitotic kinesin-like protein 2 by polo-like kinase 1 is required for
cytokinesis. J Cell Biol. 2003 Sep 1;162(5):863-75.
63. Inoue D, Sagata N. The polo-like kinase Plxl interacts with and inhibits Mytl after
fertilization of xenopus eggs. EMBO J. 2005 Mar 9;24(5):1057-67.
64. Zhang H, Shi X, Paddon H, Hampong M, Dai W, Pelech S. B23/nucleophosmin
serine 4 phosphorylation mediates mitotic functions of polo-like kinase 1. J Biol Chem.
2004 Aug 20;279(34):35726-34.
65. Yarm FR. Plk phosphorylation regulates the microtubule-stabilizing protein TCTP.
Mol Cell Biol. 2002 Sep;22(17):6209-21.
66. Watanabe N, Arai H, Iwasaki J, Shiina M, Ogata K, Hunter T, Osada H. Cyclindependent kinase (CDK) phosphorylation destabilizes somatic Weel via multiple
pathways. Proc Natl Acad Sci U S A. 2005 Aug 16;102(33):11663-8.
67. Dutertre S, Cazales M, Quaranta M, Froment C, Trabut V, Dozier C, Mirey G,
Bouche JP, Theis-Febvre N, Schmitt E, Monsarrat B, Prigent C, Ducommun B.
Phosphorylation of CDC25B by aurora-A at the centrosome contributes to the G2-M
transition. J Cell Sci. 2004 May 15;117(Pt 12):2523-31.
68. Yu CT, Hsu JM, Lee YC, Tsou AP, Chou CK, Huang CY. Phosphorylation and
stabilization of HURP by aurora-A: Implication of HURP as a transforming target of
aurora-A. Mol Cell Biol. 2005 Jul;25(14):5789-800.
69. Toji S, Yabuta N, Hosomi T, Nishihara S, Kobayashi T, Suzuki S, Tamai K, Nojima
H. The centrosomal protein Lats2 is a phosphorylation target of aurora-A kinase. Genes
Cells. 2004 May;9(5):383-97.
70. Sakai H, Urano T, Ookata K, Kim MH, Hirai Y, Saito M, Nojima Y, Ishikawa F.
MBD3 and HDAC 1, two components of the NuRD complex, are localized at aurora-Apositive centrosomes in M phase. J Biol Chem. 2002 Dec 13;277(50):48714-23.
71. Liu Q,Kaneko S, Yang L, Feldman RI, Nicosia SV, Chen J, Cheng JQ. Aurora-A
abrogation of p53 DNA binding and transactivation activity by phosphorylation of serine
215. J Biol Chem. 2004 Dec 10;279(50):52175-82.
72. Wu JC, Chen TY, Yu CT, Tsai SJ, Hsu JM, Tang MJ, Chou CK, Lin WJ, Yuan CJ,
Huang CY. Identification of V23RalA-Ser194 as a critical mediator for aurora-A-induced
cellular motility and transformation by small pool expression screening. J Biol Chem.
2005 Mar 11;280(10):9013-22.
73. Yasui Y, Urano T, Kawajiri A, Nagata K, Tatsuka M, Saya H, Furukawa K,
Takahashi T, Izawa I, Inagaki M. Autophosphorylation of a newly identified site of
aurora-B is indispensable for cytokinesis. J Biol Chem. 2004 Mar 26;279(13):129973003.
74. Zeitlin SG, Shelby RD, Sullivan KF. CENP-A is phosphorylated by aurora B kinase
and plays an unexpected role in completion of cytokinesis. J Cell Biol. 2001 Dec
24;155(7):1147-57.
75. Kawajiri A, Yasui Y, Goto H, Tatsuka M, Takahashi M, Nagata K, Inagaki M.
Functional significance of the specific sites phosphorylated in desmin at cleavage furrow:
Aurora-B may phosphorylate and regulate type III intermediate filaments during
cytokinesis coordinatedly with rho-kinase. Mol Biol Cell. 2003 Apr;14(4):1489-500.
76. Sugiyama K, Sugiura K, Hara T, Sugimoto K, Shima H, Honda K, Furukawa K,
Yamashita S, Urano T. Aurora-B associated protein phosphatases as negative regulators
of kinase activation. Oncogene. 2002 May 9;21(20):3103-11.
77. Bishop JD, Schumacher JM. Phosphorylation of the carboxyl terminus of inner
centromere protein (INCENP) by the aurora B kinase stimulates aurora B kinase activity.
J Biol Chem. 2002 Aug 2;277(31):27577-80.
78. Ohi R, Sapra T, Howard J, Mitchison TJ. Differentiation of cytoplasmic and meiotic
spindle assembly MCAK functions by aurora B-dependent phosphorylation. Mol Biol
Cell. 2004 Jun;15(6):2895-906.
79. Ban R, Irino Y, Fukami K, Tanaka H. Human mitotic spindle-associated protein
PRC1 inhibits MgcRacGAP activity toward Cdc42 during the metaphase. J Biol Chem.
2004 Apr 16;279(16):16394-402.
80. Minoshima Y, Kawashima T, Hirose K, Tonozuka Y, Kawajiri A, Bao YC, Deng X,
Tatsuka M, Narumiya S, May WS,Jr, Nosaka T, Semba K, Inoue T, Satoh T, Inagaki M,
Kitamura T. Phosphorylation by aurora B converts MgcRacGAP to a RhoGAP during
cytokinesis. Dev Cell. 2003 Apr;4(4):549-60.
81. Goto H, Yasui Y, Kawajiri A, Nigg EA, Terada Y, Tatsuka M, Nagata K, Inagaki M.
Aurora-B regulates the cleavage furrow-specific vimentin phosphorylation in the
cytokinetic process. J Biol Chem. 2003 Mar 7;278(10):8526-30.
82. Chen Y, Riley DJ, Zheng L, Chen PL, Lee WH. Phosphorylation of the mitotic
regulator protein Hec 1 by Nek2 kinase is essential for faithful chromosome segregation. J
Biol Chem. 2002 Dec 20;277(51):49408-16.
Chapter Three
Determination of the Substrate Specificities of the Polo-like Kinase
Family Reveals that Plk3 has a Motif that Changes Depending on the
Phosphorylation Status of the Activation Loop
Jes Alexander, Dan Lim, Michael Yaffe
Contributions:
Dan and I determined the motif of Plki T210D together (Figure 3.4 and 3.5).
performed all the other experiments. I wrote all the sections and created all the figures.
I
Abstract
The Polo-like kinase (Plk) family in humans consists of four members, Plkl,
Plk2, Plk3, and Plk4. These kinases are non-redundant regulators of the cell cycle.
Although the role of Plkl in mitosis has been extensively explored, the roles of the other
Plk family members are just beginning to be understood and few substrates have been
identified. The complete, unbiased motifs of Plk2, Plk3, and Plk4 have not been
reported.
Using Positional Scanning Oriented Peptide Library Screening, we have
determined the motifs of all four wildtype Plk family members as well as the motifs of
activation loop phosphomimic mutants of these kinases. The four kinases have similar,
but distinct acidophilic motifs. Therefore, these kinases are likely to phosphorylate some
of the same sites, but will also phosphorylate different sites on proteins. The motifs of
Plk2 and Plk3 include specific selectivity for phosphotyrosine, suggesting that
phosphorylation of potential substrates by a tyrosine kinase may prime these substrates
for phosphorylation by Plk2 or Plk3. The motif of wildtype unphosphorylated Plk3 is
dramatically different from the motif of the Plk3 T219D phosphomimic mutant,
suggesting that phosphorylation on the activation loop of Plk3 may act as a switch to
change the motif of the kinase. In vitro phosphorylation of the kinase verified the
phosphorylation-based motif switch. This is the first report of kinase that has a motif that
is non-static. Finally, we examined the ability of the Plk3 Polo-box domain to bind
phosphopeptides and find that it binds unphosphorylated peptides more strongly that
phosphorylated peptides, suggesting that it does not function as a phospho-dependent
binding domain in substrate selection.
Introduction
The Polo-like kinase (Plk) family in humans consists of four members, Plkl,
Plk2, Plk3, and Plk4. These kinases are non-redundant regulators of the cell cycle. Each
member of the Plk family has an N-terminal kinase domain and at least one Polo-box
motif in the C-terminus. The family is named after the founding member in Drosophila,
Polo. Polo, itself, was named after the phenotype of abnormal spindles poles, caused by
mutation of the gene (1).
The ortholog of Polo in higher eukaryotes is Plkl, the sequence of which is
conserved from yeast (where the kinase is named Cdc5) to humans.
Plkl is a key
regulator of M-phase, involved in centrosome separation and maturation, mitotic entry,
spindle assembly, the spindle assembly checkpoint, and cytokinesis (2). An incomplete
motif has been published as D/E-X-S/T-p-D/E, where
(9is
any hydrophobic amino acid
(3). We have determined an unbiased, complete motif of Plkl previously to be D/N/E-XS/T-1, where I is any hydrophobic amino acid except proline (Figure. 2.2). Many
substrates of Plkl are known (Chapter 1, Figure 2.3). The two polo-box motifs in Plkl
were found to fold together into a phospho-dependent binding domain, the Polo-box
domain (PBD), which binds the motif S-pS/pT (4). In contrast to Plkl, very little is
known about the other Plk family members.
Complete, unbiased kinase motifs of Plk2 and Plk3 have not been reported, but
the binding motifs of the Plk2 and Plk3 PBDs, determined by phosphopeptide library
screening, were reported to be the same as that of the Plkl PBD, S-pS/pT (5). Plk2 and
Plk3 are known to have post-mitotic roles in neurons where Plk2 is known to
phosphorylate spine associated RapGAP (SPAR), but the phosphorylation site(s) were
not identified (6-8). Within the cell cycle, Plk2 seems to play a role in centriole
duplication, but no substrates have been identified (9).
Plk3's role in the cell cycle is controversial. Plk3 has been reported to play a role
in stress response pathways activated by oxidative stress and DNA damage (10, 11). As
part of its role, it has been reported to phosphorylate p53 S20 and Cdc25C S216 (11, 12).
Both sites are involved in arresting the cell cycle, although the amino acid sequences
flanking the phosphorylated residues are quite different. Surprisingly, Plk3 has also been
reported to phosphorylate Cdc25C at S191 to enable mitotic entry (13). Plk3 has also
been reported to have other pro-cell cycle roles being required for entry into S-phase and
playing roles in golgi fragmentation in mitosis, and in cytokinesis (14-16).
Plk4 may be required for mitosis, as the Plk4 knockout mouse is inviable and the
embryo dies with large numbers of cells in mitosis with high Cyclin B levels (17). Plk4
has also been shown to be required for centriole duplication explaining the presence of
abnormal spindles, which are seen in both overexpression and loss of function studies of
Plk4 (18-20). No substrates of Plk4 have been identified. Unlike the other Plk family
members, Plk4 only has a single polo-box motif, which may be involved in
homodimerization rather than phospho-dependent binding of substrates (21).
Identification of substrates should elucidate the roles of Plk2 and Plk4 in centriole
duplication and clarify the conflicting literature on Plk3. Importantly, many kinases
show strong selection for residues 3 amino acids N-terminal to the phospho-acceptor.
For the Ser-20 phospho-acceptor in p53, this residue is a Glu, while for the Ser-216 site
in Cdc25C, this residue is an Arg, making it unlikely that both sites are truly
phosphorylated directly by Plk3. To further clarify optimal phosphorylation motifs and
aid in substrate identification, we have determined the motifs of wildtype Plkl, Plk2,
Plk3, and Plk4 and the motifs of activation loop phosphomimic mutants of these kinases
using Positional Scanning Oriented Library Screening (PS-OPLS). Plk family members
have similar, but distinct acidophilic motifs. The motif of Plk3 is particularly interesting.
Specific selection for phosphotyrosine suggests that phosphorylation of the substrate by a
tyrosine kinase may prime for Plk3 phosphorylation of the substrate. Additionally,
Phosphorylation on the activation loop of Plk3 acts as a switch, dramatically changing the
motif of the kinase. This is the first report of a kinase with a motif that changes as a
function of activation loop phosphorylation.
Experimental Procedures
Kinase Protein Production and Purification
To produce wildtype and T239D Plk2 kinase domain protein, wildtype, K106A,
R184A, R209A, K217A, T219D, and T219A Plk3 kinase domain protein, and wildtype
and T170D Plk4 kinase domain protein, nucleotide sequences encoding the kinase
domains of human Plkl (aa. 38-346), Plk2 (aa.53-359), Plk3 (aa.49-345), and Plk4 (aa.1290) were cloned separately into the NdeI and XhoI restriction sites in a modified
pET28a expression vector (Novagen) (pET28A-His-MBP, courtesy of Dan Lim).
Mutations were introduced using the QuikChange Site-Directed Mutagenesis Kit
(Stratagene) and verified by sequencing. The resulting fusion proteins contain an Nterminal hexa-His-tag followed by MBP and a Tev protease cleavage site between MBP
and the kinase domain. Protein was expressed in E. coli Rosetta (Novagen) cells.
For Wildtype and T239D Plk2 kinase domain, wildtype, K106A, R184A, R209A,
K217A, T219D, and T219A Plk3 kinase domain, and T170D Plk4 kinase domain fusion
proteins, starter cultures were grown overnight in LB. 50 mL of TB culture was initiated
by addition of the starter culture at a 1:100 dilution. After 4 hours of growth, the 50 mL
of TB culture was cooled to 40 C on ice, IPTG was added to a final concentration of 1
mM, and the cultures were incubated at room temperature for 16 hours. Cells were
pelleted by centrifugation at 4000 X g for 10 min at 40 C, resuspended in 5 mL of lysis
buffer (50 mM Tris, 500 mM NaC1, 1 mM DTT), and lysed by sonication using a
Branford sonicator with a probe tip (power level 4, duty cycle 50%, 2:33 min:sec) in an
ethanol-ice bath. The lysate was clarified by centrifugation at 18,000 RPM in a Sorvall
SL-50T rotor in a Sorvall Super T21 table top centrifuge for 30 min at 40C. The
supernatant was added to 0.5 mL of packed amylose beads and end-over-end rotated a
40C overnight. Beads were washed 5 times rapidly with lysis buffer and protein was
eluted by addition of 1 mL of elution buffer (50 mM Tris, 500 mM NaC1, 1 mM DTT,
100 mM maltose) and end-over-end rotation for 2 hours at 40C.
To produce wildtype and T219A PKA phosphorylated Plk3 kinase domain
protein, the wildtype and T219A Plk3 kinase domain plasmid constructs described above
were used. Nucleotide sequence encoding the catalytic domain of PKA (aa.11-352) was
cloned into the pCDFDuet-1 expression vector (Novagen). Protein was expressed in E.
coli Rosetta (Novagen) cells containing either the Plk3 wildtype or Plk3 T219A and the
PKA expression vectors. Expression and purification was as described for other Plk3
constructs above.
Wildtype Plk4 kinase domain protein was expressed in E. coli Rosetta (Novagen)
cells. A starter culture was grown overnight in LB. 2L of TB culture was initiated by
addition of the starter culture at a 1:100 dilution. After 4 hours of growth, the 2L of TB
culture were cooled to 40C on ice, IPTG was added to a final concentration of 1 mM and
the cultures were incubated at room temperature for 16 hours. Cells were pelleted by
centrifugation at 4000 X g for 10 min at 40C, the pellet from each liter of culture was
resuspended in -40 mL of buffer (50 mM Tris, 500 mM NaCl, 14 mM 3mercaptoethanol), transferred to separate 50 mL Falcon tubes, and lysed by sonication
using a Branford sonicator with a probe tip (power level 8, duty cycle 50%, 2:33 min:sec,
2 times) in an ethanol-ice bath. The lysate was clarified by ultracentrifugation at 40,000
RPM in a Beckman 45Ti rotor for 30 min at 40C. Initial purification utilized Ni-NTA
affinity chromatography. Following ultracentrifugation, the supernatant was mixed with
25 mL of packed Ni-NTA beads by end-over-end rotation at 40C for 2 hours. The bead
slurry was loaded onto a column for FPLC with the aid of a peristaltic pump. The
column was transferred to the FPLC, washed with 280 mL of wash buffer A (500 mM
NaC1, 14 mM 1-mercaptoethanol, 10 mM imidazole, 50 mM Phosphate, pH 7), and the
protein was eluted in buffer with the same composition except containing 300 mM
imidazole.
The kinase domain was further purified by affinity chromatography on
amylose beads.
Fractions containing Plk4 protein from the Ni-NTA column, as
determined by on-line UV absorbance and SDS-PAGE, were loaded onto a column
containing 25 mL of packed amylose beads, washed with 100 mL of wash buffer B (50
mM Tris, 500 mM NaC1, 14 mM P-mercaptoethanol, 10 mM maltose), and the protein
was eluted in buffer with the same composition except containing 100 mM imidazole.
Fractions containing Plk4 protein were pooled, 5 mg of hexa-his tagged Tev protease was
added to cleave off the hexa-His-MBP tag, and the mixture was incubated at 40C
overnight. To remove the cleaved tag and Tev protease, the protein was re-passaged
through Ni-NTA using the same conditions as described above. The recovered material
was applied to a Superose 12 column (16 mm X 30 cm) and eluted using gel filtration
buffer (500 mM NaC1, 2 mM DTT, 10 mM Tris, pH 8). Fractions (2 mL) were analyzed
for Plk4 protein by on-line UV absorbance and SDS-PAGE.
To produce wildtype and T210D Plkl kinase domain, nucleotide sequence
encoding the kinase domain of human Plkl (aa. 38-346) was cloned into NdeI and XhoI
restriction sites in the pET28a bacterial expression vector (Novagen). The resulting
fusion protein contains an N-terminal hexa-His-tag and a thrombin protease cleavage site
between hexa-His-tag and the kinase domain. A Thr-210 mutation to Asp was introduced
using the QuikChange Site-Directed Mutagenesis Kit (Stratagene) and verified by
sequencing. Protein was expressed in E. coli Rosetta (Novagen) cells. A starter culture
was grown overnight in LB. 6L of TB culture was initiated by addition of the starter
culture at a 1:100 dilution. After 4 hours of growth, the 6L of TB culture were cooled to
40 C on ice, IPTG was added to a final concentration of 1 mM and the cultures were
incubated at room temperature for 16 hours. Cells were pelleted by centrifugation at 4000
X g for 10 min at 40C, the pellet from each liter of culture was resuspended in -40 mL of
lysis buffer (50 mM Tris, 500 mM NaCi, 14 mM P-mercaptoethanol), transferred to
separate 50 mL Falcon tubes, and lysed by sonication using a Branford sonicator with a
probe tip (power level 8, duty cycle 50%, 2:33 min:sec, 2 times) in an ethanol-ice bath.
The lysate was clarified by ultracentrifugation at 40,000 RPM in a Beckman 45Ti rotor
for 30 min at 40C.
Initial purification utilized Ni-NTA affinity chromatography.
Following ultracentrifugation, the supernatant was mixed with 25 mL of packed Ni-NTA
beads by end-over-end rotation at 40C for 2 hours. The bead slurry was loaded onto a
column for FPLC with the aid of a peristaltic pump. The column was transferred to the
FPLC, washed with 280 mL of wash buffer (500 mM NaC1, 14 mM P-mercaptoethanol,
10 mM imidazole, 10 mM Phosphate, pH 7), and the protein was eluted using buffer of
the same composition except containing 300 mM imidazole. Fractions were analyzed by
on-line UV absorbance and SDS/PAGE.
Fractions containing Plkl kinase domain
protein were pooled, 100 units of thrombin protease were added to cleave off the hexaHis-MBP tag, and the mixture was incubated at 40 C overnight. To remove the cleaved
tag, the protein was re-passaged through Ni-NTA using the same conditions as described
above. The resulting -10 mL of material was applied to a Superose 12 column (16 mm X
30 cm) and eluted using gel filtration buffer (500 mM NaC1, 2 mM DTT, 10 mM Tris, pH
8). Fractions (2 mL) were analyzed for Plkl content by on-line UV absorbance and SDSPAGE.
Phosphorylation Motif Determination by Peptide Library Array Screening
Positional Scanning Oriented Peptide Library Screening (PS-OPLS) was
performed following Hutti et al. (22). Briefly, solution-phase kinase reactions were
performed in parallel on 198 separate biotinylated, partially degenerate oriented peptide
libraries (Anaspec, Inc) arrayed in a 384 well microtiter plate in a 22 row X 9 column
format. Each peptide library contains an N-terminal biotin tag, a 50:50 mix of serine and
threonine at the orienting phospho-acceptor residue, a single second fixed amino acid
located between the -5 and +4 position, and a mixture of amino acids at all other
positions. Individual libraries contain any of the 20 natural amino acids as well as
phosphothreonine and phosphotyrosine in the second fixed position, corresponding to the
22 rows. Scanning down the rows in the array moves the position of the fixed amino acid
from -5 to +4 relative to the fixed phospho-acceptor residue, while scanning across the
columns changes the identity of the second fixed amino acid. As an example, the peptide
library with Lys fixed in the -4 position has the following sequence: Y-A-X-K-X-X-XS/T-X-X-X-X-A-G-K-K-biotin, where amino acids are represented in 1-letter code, and
X is an equal mixture of all 17 natural amino acids excluding Cys, Ser, and Thr to
prevent oxidation effects and eliminate secondary phosphorylation events, respectively.
S/T denotes a 50:50 mix of Ser and Thr. Kinase reactions were performed at 300 C in a
total volume of 16 til containing 31.25 jiM peptide library, 100 pM ATP, and 200 piCi of
[32P]-y-ATP, in 150 mM NaCl, 10 mM MgCl 2, 1 mM DTT, 0.1% Tween 20, and 50 mM
Tris, pH 7.5. Reactions for the Plkl T210D and wildtype motifs were performed for 3
hours with 0.5 jg of protein per reaction and 3 hours with 5 jg of protein per reaction,
respectively. Reactions for the Plk2 T239D motif and Plk4 wildtype motif were
performed for 6 hours with 0.5
jpg
of protein per reaction and 2 hours with 0.11pg of
protein per reaction. Reactions for Plk3 wildtype and T219D were performed for 6 hours
with 1 and 1 jig of protein per reaction, respectively. Reactions for Plk2 wildtype motif
and Plk4 T170D motif were performed for 6 hours with 0.5 jig of protein per reaction and
2 hours with 10 jtg of protein per reaction, respectively. Following incubation, 2 jil of
each reaction were simultaneously transferred to a streptavidin-coated membrane
(Promega SAM 2 biotin capture membrane) using a 384 slot pin replicator (VP Scientific).
The membrane was washed three times with 140 mM NaCl, 0.1% SDS, 10 mM Tris, pH
7.4, three times with 2 M NaC1, twice with 2 M NaCI containing 1%H3PO 4 , and once
with water. The extent of peptide library phosphorylation was determined by imaging the
membrane with a phosphorimager (Molecular Dynamics).
In vitro Kinase Assays
All kinase assays were performed at 300C for 45 minutes in a total volume of 40
tl containing kinase reaction buffer (50 mM Tris, pH 7.5, 150 mM NaC1, 10 mM MgCl 2 ,
100 jtM ATP, 9 tC [32 P]-y-ATP, and 1 mM DTT) and the amounts of kinase domain
protein and peptide substrate indicated below. For Plk3, 0.6 jig of the indicated wildtype
or mutant kinase domain protein were reacted against 0.5 mM pY+2tide, with sequence
GEDTSFpYWAAYKKKK, or Y+2tide, with sequence GEDTSFYWAAYKKKK. For
Plkl, 0.006 jtg of wildtype or T210D kinase domain protein were reacted with 0.5 mM
Plkltide, with sequence GHDTSFYWAAYKKKK. For Plk2, 0.12 jg of wildtype or
T219D kinase domain protein were reacted with 0.5 mM Y+2tide or pY+2tide. For Plk4,
0.019 jig of wildtype or T170D kinase domain protein were reacted with 0.5 mM
Plk4tide, with sequence GHETSFYWAAKKKK. Reactions were performed in triplicate
and 5 jtL of each reaction was spotted on phosphocellulose at 0 minutes and 45 minutes.
The phosphocellulose paper was washed 4 times with 0.5% phosphoric acid, added to
scintillation fluid, and scintillation counted.
IVT pulldown assays of Plk3 PBD
Six degenerate peptide libraries were used, which each had either a fixed phosphoresidue
or the corresponding non-phosphoresidue, which was flanked on either side with four
positions containing degenerate mixtures of amino acids. Additionally, each library was
biotinylated at its N-terminus. The phosphoresidue was one of pT, pS, or pY. As an
example, the pT peptide library had the following sequences: biotin-Z-G-Z-G-G-X-X-XX-pT-X-X-X-X-Ala-K-K, where Z indicates aminohexanoic acid, and X denotes all
amino acids except C. Streptavidin beads (Pierce, 75pmol /pL gel) were incubated with a
five-fold molar excess of each biotinylated library in 20 mM Tris/HCl (pH7.5), 125 mM
NaCi, 0.5% NP-40, 1 mM EDTA and washed four times with the same buffer to remove
unbound peptide. The bead immobilized libraries (30gtL gel) were added to 6 jtL of an in
vitro translated 35S-labeled phospho-binding domain construct in 200 tL binding buffer
(20 mM Tris/HCl (pH7.5), 125 mM NaC1,0.5% NP-40, 1 mM EDTA, 1 mM DTT, 4
jtg/mL pepstatin, 4 jpg/mL aprotinin, 4 glg/mL leupeptin, 200 itM Na 3VO 4, 50 mM NaF).
Plk3 PBD, Plkl PBD, and Src SH2 domain constructs included the amino acids indicated
from nucleotides sequences cloned into pCDNA3.1+. After incubation at 40C for 2-3
hours, the beads were rapidly washed 3 times with binding buffer prior to SDS-PAGE
(12%) and autoradiography. Determination of binding was based on comparison of the
resulting bands from the pulldown with phosphopeptide library and the nonphosphopeptide library.
Results
Kinase Motifs of wildtype Polo-like Kinase family members
Activation loop phosphorylation is a common mechanism of activation of most
kinases. However, most kinases also have a residual level of activity even without
phosphorylation on the activation loop.
It has previously been shown that Plkl is
phosphorylated at T210 on the activation loop and that this phosphorylation results in
activation (23, 24).
The phosphomimic T210D mutation has been shown to at least
partially replicate this activation. This T is conserved in the other three human Polo-like
kinase family members. An activation loop phosphomimic mutation has been published
to activate Plk2 also (7). For Plk3 or Plk4, neither the phosphorylation status of the
activation loop T nor the effect of activation loop phosphorylation or phosphomimic
mutations on kinase activity has been reported. To the best of our knowledge there have
been no reports of a difference in the specificity of a kinase when in the less active nonphosphorylated state as compared to the more active phosphorylated state. As we had no
reason to suspect that the wildtype, unphosphorylated Polo-like kinase family members
would behave any differently than other known kinases, we first determined the kinase
motifs of the wildtype kinases by Positional Scanning Oriented Peptide Library
Screening (Figure. 3.1).
In the Plkl PS-OPLS blot, the darkest spots, indicating strongly selected amino
acids in the optimal phosphorylation motif, correspond to D, N, and E in the -2 and
hydrophobic amino acids, particularly F, in the +1 (Figure 3.1). The lightest spot on the
blot, indicating a residue that is specifically discriminated against in the phosphorylation
motif, corresponds to P in the +1. The resulting optimal consensus motif of Plkl is
D/N/E-X-S/T-O, where X is any amino acid, S/T is the phosphoacceptor residue and D is
any hydrophobic amino acid except P. This motif is the same as the Plkl T210D motif
(Figure 2.2 and Figure 3.1). The optimal consensus motif of Plk4 T170D is D/N/E-XS/T-(p-Y/W, where cp is any hydrophobic amino acid. This motif is similar to but not
exactly the same as a motif of Plk4 that was published while this study was underway.
That approach employed phosphorylation of arrays of peptides synthesized on a
membrane, in which each position of an individual reference peptide was replaced with
100
Plki wildtvDe
PIk2 wildtvme
C
LAi
&.
"0
PIk3 wildtvne
LL
PIk4 wildtvoe
Fixed Residue
Figure 3.1: Positional Scanning Oriented Peptide Library Screening (PS-OPLS) blots
of wildtype Polo-like kinase (Plk) family kinase domains. The blot shows relative
amounts of radioactive phosphate incorporation in in vitro kinase assays into
degenerate peptide libraries with the indicated amino acid fixed at the indicated
position relative to the phosphoacceptor S/T. The darker the spot, the more highly
selected the amino acid indicated by the columns is at the position indicated by the
row.
each of the 20 amino acids while keeping the other amino acids of the peptide fixed (25).
In contrast to this approach, PS-OPLS systematically tests each amino acid in each
position independently of all others positions resulting in unbiased motifs and each kinase
reaction in this technique occurs in solution which results in fewer artifacts than methods
involving substrate attached to solid supports (22). The motifs of Plk2 and Plk3 are very
101
similar, but not the same. In both motif blots, the pY in the +2 is the strongest spot.
However, for Plk2, the E in the -3 and the F in the +1 are almost as strong as the pY in
the +2. For Plk3, the selectivity for pY in the +2 dominates the blots as no other spot is
nearly as dark. The optimal consensus motif of Plk2 is E-X-X-S/T-F-pY and of Plk3 is
S/T-X-pY, where pY is phosphotyrosine. This is an intriguing result as no other kinase is
known to select specifically for pY in its motif, and it suggests that tyrosine kinase
phosphorylation primes substrates for Plk3 phosphorylation and, perhaps for Plk2
phosphorylation as well.
To verify this unexpected motif, we compared the activity of Plk3 against an
optimal peptide containing pY in the +2 (pY+2tide) with its activity against a peptide
with the same sequence except with Y substituted for pY in the +2 (Y+2tide) (Figure
3.2A, 3.2B). We chose to focus on Plk3 since the pY in the +2 was dominant in the Plk3
motif, but less so in the Plk2 motif. The activity of Plk3 against the pY+2tide was
approximately three times the activity against the Y+2tide confirming the selectivity for
pY in the Plk3 motif revealed by PS-OPLS.
The pY in the motif of Plk3 interacts with the basic pocket that in most kinases
interacts with the activation loop phosphorylation
The motif of Plk3 is reminiscent of the unusual Gsk330 motif, which is S/T-X-XX-pS/pT, where pS is phosphoserine and pT is phosphothreonine. As is indicated by its
motif, a priming S or T phosphorylation 4 amino acids C-terminal to the phosphoacceptor S or T on the substrate is required to allow Gsk3p to phosphorylate the phosphoacceptor residue on that substrate (26). Gsk33 is also unusual in that it does not require
phosphorylation on its activation loop to attain an active confirmation (27). In most
kinases including MAP kinases, Cyclin-dependent kinases, PKA, Akt, and Aurora
kinases, phosphorylation on the activation loop allows the kinase to attain an active state
through conformational changes caused by interaction of the phosphate with a basic
pocket composed of positively-charged residues on the catalytic loop and N-terminal lobe
of the kinase (27, 28). In contrast, the X-ray crystal structure of Gsk33 suggests that the
pS/pT priming phosphorylation on the substrate fits into this basic pocket instead, which
accounts for the pS/pT-containing motif of Gsk3p3 (27).
102
120,000
100,000!80,000
60,000
40,000
20,000
pY+2tide
m Plk3 Wildtype
Y+2tide
m PIk3 T2190
35
2.0
NWr
1.5
III
10
0J
-0.5
ST219A
m T219D
m WT (PKA Phosphorylated)
m T219A (PKA Phosphorylated)
*
*
[
[
*
*
K106A (Basophilic Pocket)
R209A (Basophilic Pocket)
K217A (Basophilic Pocket)
T219H (Electrostatic Repulsion)
T219K (Electrostatic Repulsion)
T219R (Electrostatic Repulsion)
-1.0
Figure 3.2: In vitro kinase assays comparing activity of Plk3 against a peptide with
pY in the +2 position (pY+2tide) with a peptide with the same sequence except with Y
in the +2 (Y+2tide) show that A. overall Plk3 activity is unaffected by activation loop
phosphomimic mutation (T219D), but that the activity of the kinase against a peptide
with pY in the +2 is reduced and B. that this reduction in activity against the pY+2tide
is recapitulated by phosphorylation at T219 and may be due to an interaction between
K106 and the pY on the substrate peptide. Data in B is shown as the logarithm base
two of the ratio of the counts per minute (CPM) of the pY+2tide to the Y+2tide for
ease of visualization.
103
Since the motifs of Plk3 and Gsk3p are similar, we hypothesized that like in the
case of Gsk33, the priming phosphorylation on the substrate might be interacting with the
basic pocket in Plk3. To test this hypothesis we measured the activity of mutants, in
which residues potentially forming the basic pocket were mutated to alanines, against
pY+2tide and Y+2tide (Figure 3.2B.). As the structure of Plk3 has not been solved, but
the Polo-like kinase family is most similar to AGC kinases, we chose potential Plk3 basic
pocket residues based on sequence alignment of Plk3 with PKA and Aurora A. for which
the basic pocket is known based on solved X-ray crystal structures of these kinases. The
amino acids in PKA that form the basic pocket are H87 on aC, R165 which is the RD
arginine on the catalytic loop, and K189 on 39. The equivalent residues in Plk3 are
K106, R184, and A208, respectively. Since A208 is not a basic residue, but R209 is, we
decided to include this residue as a potential basic pocket residue instead of A208.
Aurora A also does not have a basic residue equivalent to K189. Instead it has a lysine
two amino acids N-terminal to the phosphoacceptor T on the activation loop that is part
of the basic pocket. We included the equivalent residue, K217 in Plk3, in the group of
potential basic pocket residues. For easier visual comparison, the ratios of activity
against the pY+2tide to the activity against the Y+2tide in Figure 3.2B are shown as log
ratios. The ratio of the activity of the unphosphorylated, wildtype kinase against the
pY+2tide to the activity against the Y+2tide, as mentioned previously, was approximately
3 resulting in a log ratio of 1.6. The log ratios of the activities of the R209A and K217A
mutants were nearly the same as that of the wildtype at 1.5, but the log ratio for the
K106A mutant was approximately -0.75, meaning that the activity of the K106A mutant
against the Y+2tide was greater than against the pY+2tide. This suggests that K106
interacts with the pY on the substrate. The R184A mutation eliminated kinase activity
making assessment of its role in binding the pY on the substrate impossible.
Another approach to interrogating the role of the basic pocket in pY selectivity is
to test the activity of Plk3 constructs that have an intact basic pocket, but that prevent the
pY on the substrate from interacting with the pocket.
Since in most kinases, the
activating phosphorylation on the activation loop fits into the basic pocket as do
activation loop phosphomimic mutations, we would expect a phosphorylation or
phosphomimic mutation on the activation loop of Plk3 to fit into the pocket and exclude
104
the pY on the substrate. We examined the effect of phosphorylation on T219 and the
effect of the T219D mutation on activity against the pY+2tide and Y+2tide (Figure 3.2B).
Plk3 was co-expressed in bacteria with PKA, which has a motif, R-X-X-S/T- cp, that
matches the sequence surrounding T219 on Plk3, 216-RKKT 219I-220. Plk3 protein was
then purified away from PKA protein by affinity chromatography using the His-MBP tag.
To control for phosphorylation of sites other than T219, we also co-expressed Plk3
T219A with PKA.
The activities of PKA phosphorylated wildtype and PKA
phosphorylated T219A Plk3 were determined. The log ratio of PKA phosphorylated
T219A Plk3 was slightly higher than unphosphorylated wildtype, while the log ratio of
PKA phosphorylated wildtype was near zero, suggesting that phosphorylation on the
activation loop dramatically reduced the specificity of the kinase for the pY in the +2.
Recapitulating the results with PKA phosphorylated Plk3 and the basic pocket mutations,
activity of T219D Plk3 against the pY+2tide was brought down to the same level as the
activity against the Y+2tide resulting in a log ratio near zero.
Since negative charge on the activation loop could electrostatically repulse the pY
on the substrate irrespective of any effect on the basic pocket, we also tested the activities
of T219H, T219K, and T219R mutants, which will have increased positive charge on the
activation loop. The positive charge would be expected to electrostatically attract the pY
on the substrate and, consequently, result in an increased pY+2tide to Y+2tide activity
ratio as compared to the wild-type, unphosphorylated form, if the mechanism for the
reduced ratio in T219D and pT219 Plk3 is only based on electrostatic repulsion, rather
than via the basic pocket. The log ratios of these three mutants were all less than that of
the wildtype, unphosphorylated form. It therefore, seems likely that the pY on the
substrate fits into the basic pocket, which would explain the remarkable pY-directed
motif.
Activation Loop Phosphomimic mutation does not increase the activity of all
members of the Polo-like kinase family
The activities of wildtype, unphosphorylated and T219D Plk3 against the Y+2tide
were approximately the same, despite higher activity of the wildtype, unphosphorylated
form of Plk3 against the pY+2tide as compared to T219D Plk3 (Figure 3.2A, 3.2B). This
105
0
B.
A.
40.000
40,00
80,000
35,000
30,000
60,000ooo-
25,000
a-
0 20,000
40,000
15,000
10,000
20,000.
5,000.
0
0
pY+2tide
I PIk2 Wildtype
Y+2tide
0 PIk2 T239D
C.
1 n nonn
80,000,
2 60,000
40,000
20,000
0
PIk4 Wildtype
PIk4 T1700
Figure 3.3: In vitro kinase assays comparing the activity wildtype A. Plkl, B. Plk2,
and C. Plk4 with their activation loop phosphomimic mutants showing that PlkI and
Plk2 are activated by the phosphomimc mutant, but that Plk4 T170D has significantly
reduced activity.
suggests that the overall activity of the kinase does not change with the activation loop
phosphomimic mutation. This is unexpected as the sequence homology between Plk3
and Plkl is quite high and Plkl activity has been shown to increase with the activation
loop phosphomimic mutation (23).
106
This result led us to wonder whether Plk4 was activated by an activation loop
phosphomimic mutation like Plkl or unaffected like Plk3. In addition, we wanted to
know the effect of the phosphomimic mutation on the activity of Plk2 against the
pY+2tide as compared to the Y+2tide. We measured the activity of Plkl wildtype and
T210D against the optimal peptide based on the Plk1 motif, Plk2 wildtype and T239D
against the pY+2tide and Y+2tide, and Plk4 wildtype and T170D against a peptide based
on the Plk4 motif (Figure 3.1). The phosphomimic mutation increased the activity of
Plkl approximately 10-fold, agreeing with the previously published data (Figure. 3.3A)
(24). The activity of both wildtype and T239D Plk2 against the pY+2tide was greater
than against the Y+2tide.
The activity of Plk2 T239D against the pY+2tide was
approximately 7-fold higher than the wildtype against the pY+2tide and activity of Plk2
T239D against the Y+2tide was more than 40-fold greater than the wildtype against the
same peptide (Figure 3.3B). Thus unlike Plk3, the phosphomimic mutation in Plk2
significantly increases its activity and this increase in activity was not specific to the
sequence tested. The activity of Plk4 wildtype was more than 200-fold higher than Plk4
T170D (Figure 3.3C).
Kinase motifs of activation loop phosphomimic mutants of Polo-like Kinase family
members
That activation loop phosphorylation or phosphomimic mutation does not change
the overall activity of Plk3, but does reduce the activity of this kinases against pY in the
+2 containing sequences, suggests that these modifications to the kinase may simply be
altering the specificity of this kinase from that of the wildtype, unphosphorylated form.
To examine this possibility, we determined the kinase motif of Plk3 T219D as well as the
motifs of Plkl T210D, Plk2 T239D, and Plk4 T170D as controls to determine if this is a
general property of Polo-like kinase family members or specific to Plk3 (Figure 3.4). The
optimal consensus motif of Plkl T210D and Plk4 T170D are D/N/E-X-S/T-( and
D/N/E-X-S/T-(p-Y/W, respectively (Figure 3.4). These motifs are the same as the motifs
of the wildtype, unphosphorylated kinases (Figure 3.4). The optimal consensus motif of
Plk2 T239D is E-D/E-pY-S/T-4-pY/Y/W. The motif of Plk3 T219D is E-D/E-pY/D/ES/T-1.
This motif is distinctly different from the motif of the wildtype,
107
Plki T210D
Plkl wildtvpe
·T·v-
'"~
r~~~~
-3
#'
*9~I
9,
*2
*
*1
*
*
+4
K I HGFE
V
CAV TSR QPN MLpYpTYW
K IH GF ED C AV T SR
PIk2 wildtype
PN MLpYpTYW
n
i r
r c
u
• v
I
-.r
r
.
I
I
PIk2 T239D
C
0o.
I
HGFEOCAVTSR
o
I HG F EOC AV TS RQPN MLpYpTY W
PNMLpYpTYW
PIk3 witdtype
0-
PIk3 T219D
-4
I-
S3
+2
+4
I HGFEDCAV TSRQPNMLpYpTYW
K I HGFE DCAV TSR QPN MLpYoTY
Plk4 wildtype
PIk4 T170D
3
-2... .
*e
424 1
+3
-
-5
+
1
.4
-4
-3
.2
QC
41 9
'AM
K I H 73F EOC A V TSR 0 PNM LpYpTYW
-
;2J
I
R
E
M
K I H GF E DCAV TSR 0 PNM LoypTYW
Fixed Residue
Figure 3.4: Comparison of the motifs of the wildtype Plk family members with the
motifs of the activation loop phophomimic mutants. In the red-green overlay in the
third column, the motifs of the wildtype kinases are green and the motifs of the
phosphomimic mutants are red. Equal contribution from both motifs results in a
yellow spot.
unphosphorylated Plk3, S/T-X-pY (Figure 3.4). Clearly, the phosphomimic mutation
changes the motif. This is a striking result.
Phosphorylation on the activation loop of
Plk3 acts as a toggle to switch the kinase from a pY-directed kinase to an acidophilic
kinase with a motif of E-D/E-pY/D/E-S/T-Q. To our knowledge, this is the first report of
a kinase that has a motif that switches.
Comparison of the motifs of the most active form of each of the Polo-like kinase
family members
Although most closely related to the AGC group of kinases, the Polo-like kinase
family does not belong to any of the seven major groups of kinases (TK, TKL, STE, CK,
108
PIk2 T2390
Plk1 T2100D - PIk2 T239D
Plkl T210D
-5
-4
-3
-2
-1
.1
,4-Z
*3
+4
m
U.
4
3
Figure 3.5: Pseudo-color PS-OPLS blots of the more active form (wildtype or
activation loop phosphomimic mutant) of Plkl (purple), Plk2 (red), Plk3 (yellow), and
Plk4 (cyan) were superimposed to show similarities and differences between the
motifs. Equal selectivity in the Plkl-Plk2 overlay is dark blue, Plkl-Plk3 overlay is
blue, Plkl-Plk4 overlay is yellow, Plk2-Plk3 overlay is green, and Plk3-Plk4 overlay
is magenta.
AGC, CAMK, and CMGC). The Polo-like kinase family may be thought of as a group in
and of itself and, therefore, studying the motifs of this family may give us insight into the
diversity of motifs within any of the major groups of kinases without having to determine
the motifs of tens of kinases. We compared the motifs of Plkl T210D, Plk2 T239D,
wildtype, unphosphorylated Plk3, and wildtype, unphosphorylated Plk4 as these motifs
are of the more active form. To more easily visualize the similarities and differences
between the motifs, Figure 3.5. shows pairs of motifs in pseudocolor superimposed. The
motifs of Plk2 T239D and Plk3 are very similar. Selection for pY in the +2 is clear, but
other weak selection in Plk3 is evident in the Plk2 motif. This similarity is not surprising
as their sequence homology is higher than that of any other pair of Polo-like kinase
family members. The E in the -3 position distinguishes the motifs of these two kinases
from that of Plkl T210D or Plk4. The D/N/E in the -2 is selected by all members of the
family, with weaker selection by Plk2 T239D, Plk3, and Plk4, but strong selection by
Plkl T210D. In the +1 position, all members of the family, except Plk3, show strong
109
selection for F as well as a general selection for hydrophobic amino acids, which are
weak selections for Plk3. We have previously shown that Plkl will not phosphorylate
sequences with P in the +1 and the motif of Plk2 T239D shows similar discrimination
against P in the +1 as that of Plkl (Figure 2.2). This is in contrast to the motif of Plk4
which tolerates P in the +1. Selection for Y and W in the +2 is strongest in the Plk4
motif and to a lesser extent that of Plkl T210D, but is also present in the motif of Plk2
T239D. Overall, the members of the family have kinase motifs that are similar, but the
motifs of Plkl T210D, Plk2 T239D Plk3, and Plk4 all contain aspects that make them
distinct from each other. This suggests that these kinases may phosphorylate some of the
same sites, but will also phosphorylate different sites on proteins.
The Polo-Box Domain of Plk3, unlike that of Plkl, is not a phospho-dependent
binding domain
To further explore the biochemistry of Plk3, we re-examined the ability of the
Plk3 PBD to function as a phospho-dependent binding domain. It was previously
published that the Plkl PBD is a phospho-dependent binding domain that binds
sequences matching the motif, S-pS/pT-X (4). Based on the high sequence homology
including the critical residues, H538 and K540 in Plkl and H590 and K592 in Plk3, that
interact with the phosphate it was suggested that Plk3 PBD was also a phosphodependent binding domain (5). Three constructs were tested for their ability to bind to
pS, pT, or pY oriented peptide libraries as compared to non-phosphorylated control
peptide libraries. The three constructs were full-length Plk3 (aa. 1-646), a construct
including a C-terminal portion of the kinase domain through the PBD (aa. 312-646), and
a construct starting just C-terminal to the kinase domain and including the PBD (aa. 335646) (Figure 3.6). The construct composed of amino acids 312-646 has been published to
localize to the centrosome and the construct including amino acids 335-646 is the
equivalent of the Plkl 326-603 sequence that was originally discovered to be a phosphodependent binding domain (4, 29). All three constructs showed stronger binding to the
non-phosphorylated control peptide libraries than the phospho-peptide libraries, whereas
the Plkl PBD bound more strongly to the pS and pT oriented peptide libraries than the
non-phosphorylated controls and the Src SH2 domain bound more strongly to the pY
110
I
Nj pSS S
PIk3 (aa 1-646)
-
NZ pT
N'Spy y
•\°pTT
•l•S
•i pyy
PIk3 (aa 312-646)
PIk3 (aa 335-646)
Piki (aa 326-603)
Src (aa 77-262)
4
Figure 3.6: In vitro transcription/translation pulldown assays of Plk3 Polo-box
domain (PBD) constructs with phosphorylated or unphosphorylated peptide libraries
shows that the the Plk3 PBD binds more strongly to non-phosphopeptides than to
phosphopeptides suggesting it is not a phospho-dependent binding domain. Three
constructs of the Plk3 PBD were tested with two controls, the Plkl PBD and the Src
SH2 domain.
oriented library than the non-phosphorylated control (Figure 3.6). These data suggest
that the Plk3 PBD is not a phospho-dependent binding domain, but may actually bind
non-phosphorylated sequences. Unfortunately, it is not possible to determine from this
technique what sequences are preferred by this domain.
111
Discussion
In this study, we determined the motifs of the Polo-like kinase family members,
Plkl, Plk2, Plk3, and Plk4, as both the wildtype and with the T210D, T239D, T219D, and
T170D phosphomimic mutations, respectively, on the activation loops.
The
phosphomimic mutation motifs of Plkl and Plk4 were essentially the same as the
wildtype motifs of these kinases. The Plkl motif is D/N/E-X-S/T-0 and the Plk4 motif is
D/N/E-X-S/T-(p-Y/W, where (Dis any hydrophobic amino acid except proline and (p is
any hydrophobic amino acid. Unexpectedly, the wildtype motifs of Plk2 and Plk3 were
different from the activation loop phosphomimic mutation motifs of these kinases. The
wildtype motif of E-X-X-S/T-F-pY and of Plk3 is S/T-X-pY. The Plk2 T239D motif is
E-D/E-pY-S/T-A-pY/Y/W and the Plk3 T219D motif is E-D/E-pY/D/E-S/T-4.
The
motifs of Plk3 make it clear that Plk3 is an acidophilic kinase making it unlikely that if
phosphorylates the S216 site in Cdc25C.
Phosphomimic mutation on the activation loop seems to have different effects on
different members of the Polo-like kinase family, suggesting that phosphorylation on the
activation loops would have similarly diverse effects.
We reconfirmed that
phosphomimic mutation on the activation loop of Plkl stimulates kinase activity.
Phosphomimic mutation on the activation loop of Plk2 and Plk3 seems to switch the
motif. The motif switch for Plk2 is less dramatic than that of Plk3 as the strongest
determinants of specificity in both the wildtype, unphosphorylated and Plk2 T239D
motifs are the same. Additionally, the phosphomimic mutation increased the activity of
the kinase by more than 40-fold compared to the wildtype, suggesting that the kinase may
need to be phosphorylated in vivo for activity. In Plk3, the phosphomimic mutation does
not seem to affect catalytic competency of the kinase, but seems only to switch the motif.
The motif switch is quite dramatic, changing the motif from being dominated by
selectivity for pY in the +2 to having specificity for acidic residues in the positions Nterminal to the phosphoacceptor and almost reducing the pY in the +2 specificity to
background levels. The motif switching is not an artifact of the phosphomimic mutation
since phosphorylation in the activation loop also switches the motif. This is an extremely
important result as it shows that kinases do not necessarily have static motifs, as was
112
previously assumed in the field. The pY in the +2 specificity of the wildtype kinase is at
least in part mediated by K106, which may be part of a basic pocket involved in both
phospho-dependent kinase activation as well as substrate selection in the absence of
activation loop phosphorylation. Since the structure of Plk3 has not been solved, it is
impossible to know for sure. The phosphomimic mutation of Plk4 resulted in a greater
than 200-fold drop in activity of the kinase. Although, based on sequence alignment,
T170 of Plk4 seems to be the equivalent residue to Plkl T210, which has been shown to
be phosphorylated for activation of Plkl, this can only be verified with crystal structures.
It may be that Y169 in Plk4 is actually the residue that is phosphorylated for activation of
Plk4. Nevertheless, the results lead to the hypothesis that phosphorylation on the
activation loop may function as a way of shutting off Plk4 kinase activity. Although
clearly possible, whether or not phosphorylation is actually employed in vivo to swap
Plk3 motifs or shut off kinase activity of Plk4 remains to be tested.
Intriguingly, the Plk2 and Plk3 motifs show specific selection for pY in the +2
position. In other kinase motifs when pY is selected, either other negatively charged
residues or large hydrophobic residues are also selected in the same position, as in the
case of CKII and PAK kinases, suggesting that what is important for these kinases is
charge or hydrophobicity, not the presence of phosphotyrosine itself (22, 30). Two
kinases are known to require priming phosphorylations within a few amino acids of the
site to be phosphorylated, GSK313, as mentioned previously, and CKI, which has the
motif pS/pT/D/E-X-X-S/T (31). In both cases, the priming phosphorylation is on a pS or
pT residue.
No kinase has previously been described that requires priming
phosphorylation on a pY residue. This is particularly interesting in a serine/threonine
kinase as it suggests that Plk2 and Plk3 may integrate tyrosine kinase signaling with
serine/threonine kinase signaling in a novel way.
One other serine/threonine kinase integrates tyrosine kinase signaling and
serine/threonine kinase signaling in a similar, but distinct manner. Protein Kinase C delta
(PKC8) contains a Conserved Domain 2 (C2) domain that has been found to be a
phosphotyrosine binding domain (32).
Thus, the C2 domain can recognize a
phosphotyrosine motif on a substrate and the PKC kinase domain could then
phosphorylate the substrate at another site. This is similar to the PBD of Plkl, which
113
binds a phosphoserine/phosphothreonine motif on a substrate and allows Plkl to
phosphorylate that protein or other proteins in the vicinity (33). We re-examined the
ability of the Plk3 PBD to function as a phospho-dependent binding domain and found
that it binds more strongly to non-phosphorylated sequences suggesting that it is not a
phospho-dependent binding domain. Both Plkl and Plk3 seem to function as molecular
logic gates, integrating the activity of another kinase with its own and outputting a
phosphorylation on the substrate. Plkl functions as an AND gate. Its two inputs are
phosphorylation status of the activation loop of the kinase domain, which decides if the
kinase is active or not, and phosphorylation status of the substrate or a protein nearby,
which is recognized by the PBD. Both inputs should be in the phosphorylated state for
output phosphorylation of the substrate by Plkl.
PKC6 functions similarly.
Plk3
functions as an Exclusive OR (XOR) gate. Its two inputs are phosphorylation status of
the activation loop, which decides which motif the kinase will phosphorylate and the
phosphorylation status of the substrate, which is recognized by the active site cleft. For
Plk3 exactly one input should be in the phosphorylated state for output phosphorylation
on the substrate. The phosphopeptide binding function of the PBD, which is conserved in
orthologs of Plkl from yeast to humans, seems in Plk3 to have been subsumed by the
Plk3 kinase domain, leaving the Plk3 PBD able to handle other as yet undefined
functions (5). Since the kinase domain is responsible for recognition of the priming
phosphorylation, phosphorylation by Plk3 occurs on more carefully chosen sites than in
the case of Plkl because the priming phosphorylation must be in the same protein exactly
two amino acids away from the residue to be phosphorylated.
Kinase motifs become more important and useful as the number of
phosphoproteomic mass spectrometry screens performed increases. The motifs combined
with phosphoproteomic data and bioinformatics approaches can help to identify the
kinases creating the phosphorylation sites and thereby predict sites phosphorylated by
these kinases (34). The substrates of Plk2, Plk3, and Plk4 are almost unknown and their
roles are only beginning to be understood. In this study we have determined the motifs of
these kinases, which should aid in the identification of substrates and help to increase our
standing of the roles of these kinases in regulation of both the cell cycle and other postmitotic functions.
114
References
1. Sunkel CE, Glover DM. Polo, a mitotic mutant of drosophila displaying abnormal
spindle poles. J Cell Sci. 1988 Jan;89 ( Pt 1)(Pt 1):25-38.
2. van de Weerdt BC, Medema RH. Polo-like kinases: A team in control of the division.
Cell Cycle. 2006 Apr;5(8):853-64.
3. Nakajima H, Toyoshima-Morimoto F, Taniguchi E, Nishida E. Identification of a
consensus motif for plk (polo-like kinase) phosphorylation reveals Mytl as a Plkl
substrate. J Biol Chem. 2003 Jul 11 ;278(28):25277-80.
4. Elia AE, Cantley LC, Yaffe MB. Proteomic screen finds pSer/pThr-binding domain
localizing Plkl to mitotic substrates. Science. 2003 Feb 21;299(5610):1228-31.
5. Elia AE, Rellos P, Haire LF, Chao JW, Ivins FJ, Hoepker K, Mohammad D, Cantley
LC, Smerdon SJ, Yaffe MB. The molecular basis for phosphodependent substrate
targeting and regulation of plks by the polo-box domain. Cell. 2003 Oct 3;1 15(1):83-95.
6. Pak DT, Sheng M. Targeted protein degradation and synapse remodeling by an
inducible protein kinase. Science. 2003 Nov 21;302(5649):1368-73.
7. Ma S, Liu MA, Yuan YL, Erikson RL. The serum-inducible protein kinase snk is a Gl
phase polo-like kinase that is inhibited by the calcium- and integrin-binding protein CIB.
Mol Cancer Res. 2003 Mar;1(5):376-84.
8. Kauselmann G, Weiler M, WulffP, Jessberger S, Konietzko U, Scafidi J, Staubli U,
Bereiter-Hahn J, Strebhardt K, Kuhl D. The polo-like protein kinases fnk and snk
associate with a ca(2+)- and integrin-binding protein and are regulated dynamically with
synaptic plasticity. EMBO J. 1999 Oct 15;18(20):5528-39.
9. Warnke S, Kemmler S, Hames RS, Tsai HL, Hoffmann-Rohrer U, Fry AM, Hoffinann
I. Polo-like kinase-2 is required for centriole duplication in mammalian cells. Curr Biol.
2004 Jul 13;14(13):1200-7.
10. Bahassi el M, Conn CW, Myer DL, Hennigan RF, McGowan CH, Sanchez Y,
Stambrook PJ. Mammalian polo-like kinase 3 (Plk3) is a multifunctional protein involved
in stress response pathways. Oncogene. 2002 Sep 26;21(43):6633-40.
11. Xie S, Wu H, Wang Q, Kunicki J, Thomas RO, Hollingsworth RE, Cogswell J, Dai
W. Genotoxic stress-induced activation of Plk3 is partly mediated by Chk2. Cell Cycle.
2002 Nov-Dec; 1(6):424-9.
12. Ouyang B, Li W, Pan H, Meadows J, Hoffmann I, Dai W. The physical association
and phosphorylation of Cdc25C protein phosphatase by prk. Oncogene. 1999 Oct
28; 18(44):6029-36.
115
13. Bahassi el M, Hennigan RF, Myer DL, Stambrook PJ. Cdc25C phosphorylation on
serine 191 by Plk3 promotes its nuclear translocation. Oncogene. 2004 Apr
8;23(15):2658-63.
14. Zimmerman WC, Erikson RL. Polo-like kinase 3 is required for entry into S phase.
Proc Natl Acad Sci U S A. 2007 Feb 6;104(6):1847-52.
15. Xie S, Wang Q,Ruan Q, Liu T, Jhanwar-Uniyal M, Guan K, Dai W. MEKl-induced
golgi dynamics during cell cycle progression is partly mediated by polo-like kinase-3.
Oncogene. 2004 May 6;23(21):3822-9.
16. Jiang N, Wang X, Jhanwar-Uniyal M, Darzynkiewicz Z, Dai W. Polo box domain of
Plk3 functions as a centrosome localization signal, overexpression of which causes
mitotic arrest, cytokinesis defects, and apoptosis. J Biol Chem. 2006 Apr
14;281(15):10577-82.
17. Hudson JW, Kozarova A, Cheung P, Macmillan JC, Swallow CJ, Cross JC, Dennis
JW. Late mitotic failure in mice lacking sak, a polo-like kinase. Curr Biol. 2001 Mar
20;11(6):441-6.
18. Habedanck R, Stierhof YD, Wilkinson CJ, Nigg EA. The polo kinase Plk4 functions
in centriole duplication. Nat Cell Biol. 2005 Nov;7(11):1140-6.
19. Ko MA, Rosario CO, Hudson JW, Kulkami S, Pollett A, Dennis JW, Swallow CJ.
Plk4 haploinsufficiency causes mitotic infidelity and carcinogenesis. Nat Genet. 2005
Aug;37(8):883-8.
20. Fode C, Binkert C, Dennis JW. Constitutive expression of murine sak-a suppresses
cell growth and induces multinucleation. Mol Cell Biol. 1996 Sep;16(9):4665-72.
21. Leung GC, Hudson JW, Kozarova A, Davidson A, Dennis JW, Sicheri F. The sak
polo-box comprises a structural domain sufficient for mitotic subcellular localization. Nat
Struct Biol. 2002 Oct;9(10):719-24.
22. Hutti JE, Jarrell ET, Chang JD, Abbott DW, Storz P, Toker A, Cantley LC, Turk BE.
A rapid method for determining protein kinase phosphorylation specificity. Nat Methods.
2004 Oct;1(1):27-9.
23. Jang YJ, Ma S, Terada Y, Erikson RL. Phosphorylation of threonine 210 and the role
of serine 137 in the regulation of mammalian polo-like kinase. J Biol Chem. 2002 Nov
15;277(46):44115-20.
24. Lee KS, Erikson RL. Plk is a functional homolog of saccharomyces cerevisiae Cdc5,
and elevated plk activity induces multiple septation structures. Mol Cell Biol. 1997
Jun; 17(6):3408-17.
116
25. Leung GC, Ho CS, Blasutig IM, Murphy JM, Sicheri F. Determination of the
Plk4/Sak consensus phosphorylation motif using peptide spots arrays. FEBS Lett. 2007
Jan 9;581(1):77-83.
26. Fiol CJ, Mahrenholz AM, Wang Y, Roeske RW, Roach PJ. Formation of protein
kinase recognition sites by covalent modification of the substrate. molecular mechanism
for the synergistic action of casein kinase II and glycogen synthase kinase 3. J Biol
Chem. 1987 Oct 15;262(29):14042-8.
27. Dajani R, Fraser E, Roe SM, Young N, Good V, Dale TC, Pearl LH. Crystal structure
of glycogen synthase kinase 3 beta: Structural basis for phosphate-primed substrate
specificity and autoinhibition. Cell. 2001 Jun 15;105(6):721-32.
28. Nolen B, Taylor S, Ghosh G. Regulation of protein kinases; controlling activity
through activation segment conformation. Mol Cell. 2004 Sep 10;15(5):661-75.
29. Jiang N, Wang X, Jhanwar-Uniyal M, Darzynkiewicz Z, Dai W. Polo box domain of
Plk3 functions as a centrosome localization signal, overexpression of which causes
mitotic arrest, cytokinesis defects, and apoptosis. J Biol Chem. 2006 Apr
14;281(15):10577-82.
30. Rennefahrt UE, Deacon SW, Parker SA, Devarajan K, Beeser A, Chernoff J, Knapp
S, Turk BE, Peterson JR. Specificity profiling of pak kinases allows identification of
novel phosphorylation sites. J Biol Chem. 2007 May 25;282(21): 15667-78.
31. Flotow H, Graves PR, Wang AQ, Fiol CJ, Roeske RW, Roach PJ. Phosphate groups
as substrate determinants for casein kinase I action. J Biol Chem. 1990 Aug
25;265(24):14264-9.
32. Benes CH, Wu N, Elia AE, Dharia T, Cantley LC, Soltoff SP. The C2 domain of
PKCdelta is a phosphotyrosine binding domain. Cell. 2005 Apr 22;121(2):271-80.
33. Lowery DM, Lim D, Yaffe MB. Structure and function of polo-like kinases.
Oncogene. 2005 Jan 10;24(2):248-59.
34. Linding R, Jensen LJ, Ostheimer GJ, van Vugt MA, Jorgensen C, Miron IM, Diella F,
Colwill K, Taylor L, Elder K, Metalnikov P, Nguyen V, Pasculescu A, Jin J, Park JG,
Samson LD, Woodgett JR, Russell RB, Bork P, Yaffe MB, Pawson T. Systematic
discovery of in vivo phosphorylation networks. Cell. 2007 Jun 29;129(7):1415-26.
117
Chapter Four
Conclusions and Future Directions
Contributions:
I wrote all sections.
118
Analysis of Motifs of Kinases Related by Function Versus those Related by
Sequence
The Polo-like kinase family has high sequence similarity, but each member of the
family seems to have a different role in cell cycle regulation. The motifs of the Polo-like
kinase family were found to be similar, but distinct. In particular, Plk3 T219D and the
more active forms of Plkl, Plk2, and Plk4 all showed selection for D and N in the -2
position and a hydrophobic amino acid in the +1 position. This similarity suggests that
given access to the same proteins, these kinases might phosphorylate the same sites on
these proteins. In addition to these common specificities, each kinase selected for other
amino acids in other positions that differentiated their motifs, like selection for E in the -3
by Plk2 and the strong selection for W and Y in the + 2 by Plk4, suggesting that each
member of the family will also phosphorylate distinct sites on substrates. As noted in
Chapter 1, these kinases have common localizations and could have common substrates,
particularly at the centrosome, where each has been reported to localize.
The expression patterns of the kinases likely contribute to their distinct roles,
which are likely mediated by phosphorylation of distinct substrates. Plk4 protein is
expressed from late Gl through the cell cycle to early in the next Gl, when it becomes
undetectable (1). Plk2 is expressed in Gl with protein and activity levels peaking in midG1 and then falling off before S-phase (2). Plkl protein is expressed from late S-phase
until it is destroyed in late M-phase and early Gl (3). Plk2 may be at the centrioles
during the period in which Plk4 protein is undetectable preparing centrioles for
duplication prior to Plk4's role in centriole duplication (4). Plk2 protein levels probably
fall off at approximately the same time as Plk4 levels are rising again. Plkl and Plk4
may be at the centrosome at the same time in mitosis, but Plkl seems to be associated
with the pericentriolar material whereas Plk4 is associated specifically with the centrioles
(5), suggesting that they may have non-overlapping localizations at the centrosome
possibly preventing phosphorylation of the same substrates. How Plk3 fits into this
picture is unclear, since when Plk3 protein is expressed during the cell cycle or where it
localizes is controversial. However, the motif of wildtype, unphosphorylated Plk3 is
distinct from the motifs of the other kinases and so it likely has distinct substrates. It
119
seems likely that timing of expression and localization play an important role in substrate
specificity of the Plk family.
The substrate specificity of the Protein Kinase C (PKC) family, which has high
sequence similarity and highly similar motifs, is regulated similarly to the above
proposed regulation of the Plk family. However, the PKC family is different from the
Plk family, since PKC kinases have some redundant roles (6). All nine members of the
family select for an R in the -3 and hydrophobic amino acids in the +1 position (7). Each
has slightly different specificities at other positions of the motif. Regulation of PKC
substrate specificity seems to involve differential expression as well as localization, like
the Plk family (8). The PKCs are different from the Plks in that instead of temporal
regulation of expression, PKCs are expressed in different cell types (9). Like the Plk
family, the PKCs have different localizations. Localization is mediated by receptors for
activated C-kinase (RACKs) (8). RACKs likely limit the substrates with which each
kinase can interact in addition to allosterically activating the bound kinase.
The
similarities in regulation of substrate specificity of the PKCs to the proposed regulation
of the Plk family lend some credence to the proposed model of regulation.
Unlike the Polo-like kinases, which function in different aspects of the cell cycle
and are expressed during different phases of the cell cycle, the major mitotic kinases must
all be expressed during mitosis. The major mitotic kinases, which, except for Aurora A
and Aurora B, do not have very high sequence similarity, but do cooperate in function,
have very different motifs. Cdkl will not phosphorylate any sites phosphorylated by any
of the other major mitotic kinases and vice versa despite overlapping localizations; Plkl
will not phosphorylate any sites phosphorylated by the Aurora kinases and vice versa
despite overlapping localizations. However, the motifs of Aurora A and Aurora B are
almost identical. We have proposed that the localization of these kinases is a key factor
in generating their specificity as Aurora A localizes to the centrosome and spindle
microtubules whereas Aurora B localizes to the DNA and kinetochores (10, 11). It might
be that for kinases that function together, like the major mitotic kinases, localization and
motifs play larger roles in generating specificity than for kinases which do not, like the
PKC or Plk family, which can rely on differential expression to generate some of the
specificity.
120
Summary and Future Directions
Although motifs have been known for some time to direct individual kinases to
their substrates, the motifs presented in this study suggest that motifs can play larger
roles. In Chapter 2, we put forward a hypothesis that, if true, would imply that the motifs
and localizations of individual kinases cooperatively direct a group of kinases to their
substrates in a coordinated manner. Although motifs function at the level of individual
kinases they may also function at the level of groups of functionally related kinases. In
Chapter 3, we showed that the motifs of Plk2 and Plk3 select for phosphotyrosine,
suggesting that phosphorylation by a tyrosine kinase on the substrate primes for
phosphorylation of the substrate by Plk2 and Plk3.
Motifs are not just involved in
passing along signals, but can actually integrate signals. In the case of Plk2 and Plk3,
motifs may integrate tyrosine kinase signaling with serine/threonine kinase signaling in a
novel way. Also in Chapter 3, we showed that Plk3 has a phosphorylation based toggle
switch that changes the motif of the kinase. This result suggests that the motif of a kinase
is not necessarily static and for some kinases it might be modulated to cause the kinase to
phosphorylate different sites under different conditions.
Thus, the motif of certain
kinases may actually change the functional roles of those kinases.
Coordinated Regulation of the Major Mitotic Kinases by Motifs and Localizations
In Chapter 2, based on the results of biochemical characterization of the motifs of
the major mitotic kinases we put forward a hypothesis potentially explaining how the
major mitotic kinases are regulated to phosphorylate distinct sites in proteins despite
having access to the same sites: In mitosis, major mitotic kinases with overlapping
motifs do not have overlapping localizations and major mitotic kinases with overlapping
localizations do not have overlapping motifs. If this hypothesis accurately describes the
regulation theses kinase are under, then mislocalization of a kinase A so that it colocalizes with a kinase B with an overlapping motif should allow the sites phosphorylated
by the kinase B to be phosphorylated by kinase A. Additionally, changing the motif of a
kinase C that is co-localized with a kinase D so that both kinases have the same motif
should allow the kinase C to phosphorylate sites that had originally only been
121
phosphorylated by the kinase D. Testing either of these predictions of the model directly
is difficult because all of the major mitotic kinases are pleiotropic and modified function
by either mislocalization or changing of the motif would result in overlaid, difficult to
interpret phenotypes. In particular, the localizations of substrates and kinases would be
altered. A better way to test the hypothesis would be to manipulate a substrate rather
than a kinase itself. Either mislocalization of a substrate from the localization of one
kinase to that of another kinase with an overlapping motif or mutation of a site known to
be phosphorylated by one kinase to match the motif of second kinase with the same
localization, but a different motif would make interpretation of the results of the
experiment easier, since an appropriately chosen substrate will not be pleiotropic.
One such convenient substrate is the budding yeast, Scc 1. Sccee is a component of
the cohesin complex which holds sister chromatids together during mitosis to prevent
premature chromosome segregation (12).
When all chromosome have obtained
amphitelic attachments and the spindle checkpoint has been silenced, the APC/C
ubiquitinates securin resulting in the destruction of securin (13). Securin inhibits the
protease separase from degrading Sccl (14). In budding yeast, degradation of Sccl is
essential for anaphase onset (15). Separase cleaves Sccl at two sites, of which the Cterminal site is preferred by the endopeptidase. A few amino acids N-terminal to each of
the cleavage sites is a site matching the Plkl consensus motif (173-DTSL-176 and 261DNSV-264).
Alexandru et al. showed that the yeast ortholog of Plkl, Cdc5,
phosphorylates Sccl at these sites to make it a better substrate for separase in a securin
deletion strain (16). Mutation of the serines to non-phosphorylatable residues resulted in
delay of Sccl degradation by about 30 minutes and increased the number of cells with
undivided nuclei undergoing cytokinesis, known as the "cut" (chromosomes untimely
torn) phenotype. Phosphorylation of See 1 was assayed by electrophoretic mobility shift.
Mutation of the sites phosphorylated by Cdc5 to match the motif of Ipll (D173R
and D261R), the budding yeast ortholog of Aurora B, or Cdkl (L176P and V264P) would
allow the hypothesis to be tested as both Ipll and Cdkl have appropriate localizations
(Angelika Amon, Personal Communication). The yeast system would allow this to be
done with relative ease, and wildtype and non-phosphorylatable Sccl could be used as
controls. The same assays used by Alexandru et al. could be used. Phosphorylation and
122
degradation of Sccl can be monitored over a time-course by electrophoretic mobility
shift and the appearance of cleavage products, respectively.
The number of yeast
displaying the "cut" phenotype would also be monitored. If the hypothesis is true, the
motif mutant Sccl should be phosphorylated and degraded with kinetics more similar to
the wildtype Scc l than the non-phosphorylatable mutant. Also there should be fewer
cells displaying the "cut" phenotype for the motif mutant than for the nonphosphorylatable mutant. The proposed experiments are currently in progress.
It is also possible that the motif mutant Sccl will result in a different phenotype.
Both Ipll and Cdkl are under different regulation than Cdc5. The motif mutant Sccl
might be phosphorylated earlier or later by Ipll or Cdkl than it would be by Cdc5,
resulting in premature or delayed separase cleavage of the motif mutant Sccl. While a
delayed separase cleavage may just result in a delayed, but otherwise normal, anaphase, a
premature cleavage prior to all kinetochores attaining amphitelic attachments would
cause a lack of tension, an extended spindle checkpoint, and, probably, premature
cytokinesis resulting in aneuploid cells (12).
A phenotype for the Sccl motif mutant different than the wildtype would have
important implications for the regulation of mitosis. It would suggest that it matters
which kinase phosphorylates a given site.
In essence, each site must not only be
phosphorylated, but must be phosphorylated in the appropriate setting, which would be
arranged by the differential upstream regulation of the appropriate major mitotic kinase.
Such a result would explain why the regulation suggested by the hypothesis would be
required and why more than one kinase is necessary for proper regulation of mitosis. As
mitosis is an intricate process, this differential upstream regulation must be important in
fine tuning when events occur in mitosis. Although, temporal regulation is not explicitly
stated in the hypothesis, it would implicitly be part of the hypothesis. Whether or not
there is a different phenotype for the Sccl motif mutant as compared to the wildtype, if
the hypothesis is true, there is a coordinated regulation of the major mitotic kinases based
on the individual kinases' motifs and localizations.
123
Regulation of p31/Comet Function by Plkl
p31/Comet was shown to be a candidate substrate of Plkl in nocodozole arrested
cells.
Previously, p31/Comet was shown to be involved in silencing the spindle
checkpoint by binding Mad2 (17). Mad2 has at least two conformations, open-Mad2 and
closed-Mad2 (18). In the "Mad2 Template" Model, closed-Mad2 binds Madl at the
kinetochores for its localization to the kinetochore (19). Open-Mad2 binds closed-Mad2
on kinetochores without amphitelic attachments where it can also bind Cdc20. Upon
binding of Cdc20, open-Mad2 changes conformation to closed-Mad2 and the Mad2Cdc20 complex dissociates from the kinetochore. The interaction of Mad2 with Cdc20
prevents interaction of Cdc20 with the APC/C preventing anaphase onset. After a
kinetochore has obtained an amphitelic attachment, p31/Comet binds the closed-Mad2 on
the kinetochore preventing open-Mad2 from binding closed-Mad2 and preventing the
formation of Mad2-Cdc20 complexes (20).
How p31/Comet is prevented from binding Mad2 on kinetochores still lacking
amphitelic attachments has not been determined, although we hypothesize that
phosphorylation by Plkl may play a role. The role of Plkl in the spindle assembly
checkpoint is unclear. Plkl is required for recruitment of proteins involved in the spindle
assembly checkpoint such as Madl and Mad2, and it also has been reported to generate
the 3F3/2 phosphoepitope which appears on kinetochores lacking amphitelic attachments
to the spindle (21-23). The X-ray crystal structure of the Mad2-p3 I/Comet complex has
been solved (24). However, the candidate Plkl phosphorylation site on P31/Comet is not
near the interface suggesting it does not directly inhibit interaction of the Mad2p31/Comet complex. However, the phosphorylation may still be a signal resulting in the
prevention of p31/Comet from binding Mad2 prior to amphitelic attachment. To test this
hypothesis, a non-phosphorylatable mutant p31/Comet could be expressed in p31/Comet
knockdown cells. Appropriate controls would be expression of wildtype p31/Comet and
a phosphomimic mutant in the p3 1/Comet knockdown cells. If the hypothesis is true, the
spindle checkpoint would be silenced early resulting in premature chromosome
segregation and increased aneuploidy as compared to the wildtype control.
The
phosphomimic might show delayed anaphase onset.
124
Promotion and Inhibition of the Cell Cycle by Plk3
As discussed previously, Plk3 has been reported to have roles both in promoting
the cell cycle and promoting cell cycle arrest (25-31). There are at least two possible
explanations for how Plk3 could mediate this conflicting behavior, one involving the
PBD and the other involving the kinase domain.
In Chapter 3, the PBD of Plk3 was shown to not be a phospho-dependent binding
domain, and this has been corroborated by an independent study published recently (32).
However, the PBD has previously been shown to mediate protein-protein interactions
with CIB and SPAR as part of its non-cell cycle related roles in neurons (33, 34). The
interaction with CIB seems to occur in a phospho-independent manner as only one of the
Polo-box motifs of the pair of Polo-box motifs comprising the full Polo-box domain is
required.
GFP-tagged Plk3 PBD has been reported to localize to centrosomes also
suggesting a binding function (35). When cell lysates from untreated cells or cells
subjected to cellular stress were fractionated on a size-exclusion column, Plk3 was
reported to associate with complexes of different sizes (27). The later data is somewhat
suspect as it relies on an antibody, which we have found to immunoprecipitate the heat
shock protein, Hsp70, rather than Plk3 based on mass spectrometry (data not shown) and
recognizes a band by Western that is not fainter in lysates from Plk3 knockdown cells as
compared to untreated cells (36). Nevertheless, the data suggest the hypothesis that the
Plk3 PBD may function as context dependent binding domain forming complexes with
different proteins after cellular stress than during a normal cell cycle.
The switch
deciding which proteins it interacts with may be on the PBD itself or may be a property
of the interacting proteins.
More excitingly, the kinase domain may contain the switch which decides
whether Plk3 contributes to cell cycle progression or cell cycle arrest. In Chapter 3,
phosphorylation or phosphomimic mutation of Plk3 at T219, which is on the activation
loop, was found to act as a toggle switch to swap between a motif almost entirely
consisting of specificity for pY in the +2 position and the generally acidophilic motif, ED/E-pY/D/E-S/T-f,
where (D, is any hydrophobic amino acid. Although this result
suggests that motif swapping is possible for kinases, which is an important result in itself,
it does not address whether or not this toggle is employed in vivo on Plk3 and to what
125
end. Phosphorylation by Plk3 on p53 and Cdc25A has been reported to occur on
acidophilic sites (ETFS20D in p53 and EST8 0D in Cdc25A) to cause cell cycle arrest and
Plk3 has been reported to be phosphorylated by Chk2 following cellular stress, although
the site was not identified (26, 27, 37). The activation loop T219 is part of a sequence
(EQRKKT 219I) that matches a minimal Chk2 consensus sequence of R-X-X-S/T (38).
Together these data suggest the hypothesis that Plk3 may be phosphorylated on T219 of
the activation loop to swap its motif from being dominated by pY specificity for the roles
of Plk3 in promoting the cell cycle to being acidophilic for the roles of Plk3 in inhibiting
the cell cycle.
Although there is an antibody that recognizes the phosphorylated activation loop
of Plkl, no such antibody exists for Plk3 (39). The sequences of Plkl and Plk3 on the
activation loop are sufficiently similar, however, that the antibody might recognize Plk3
phosphorylated on the activation loop. To test the antibody, recombinant phosphorylated
and unphosphorylated Plk3 could be used in a Western blot. If the antibody does
recognize T219 phosphorylated Plk3, an easy experiment presents itself that could test
this hypothesis using cell lines. Plk3 could be immunoprecipitated from lysates from
cells exposed to ionizing radiation to cause DNA damage and the phosphospecific
antibody used in a Western blot to determine if Plk3 is phosphorylated. An equal amount
of lysate from undamaged cells would be used as a control. If the antibody does not
recognize T219 phosphorylated Plk3, a mass spectrometric approach would be required.
One possible experiment to clarify the relevance of motif switching might be to
make Plk3 knock-down cells, or analogue-specific inhibited forms of Plk3, and ask
whether T219A or T219D mutants of Plk3 were equivalent to wild-type Plk3 in their
ability to reverse a knock-down or chemically inhibited phenotype. In our preliminary
experiments, RNAi-mediated knock-down of Plk3 appeared to induce cell lethality,
although the absence of a good antibody against Plk3 limits our ability to make strict
conclusions.
Modulation of Src Homology 2 (SH2) Domain Binding by Plk3
Tyrosine kinase signaling is involved in, among other processes, cellular
substratum adhesion (40). We and others have shown that GFP-tagged Plk3 localizes to
126
the cellular cortex where tyrosine kinase signaling regulating substratum adhesion is
concentrated (data not shown), (40, 41). Additionally, overexpression of Plk3 results in
cellular rounding and release from the substratum, suggesting Plk3 may play a role in
regulating substratum adhesion (data not shown), (41). The relevant substrates are
unknown. SH2 domains, which are found in many proteins involved in tyrosine kinase
signaling, bind tyrosine phosphorylated sequences in proteins (42, 43). The motifs of
many SH2 domains have been determined and were found to select tyrosine
phosphorylated sequences based on the amino acids C-terminal to the phosphotyrosine
residue (44, 45). However, recently a more extensive study of SH2 domain binding
motifs has revealed that many SH2 domains also have specificity for amino acids one and
two positions N-terminal to the phosphotyrosine residue (46).
Specificity for
phosphoserine or phosphothreonine in any position was not examined in any of these
studies (44-46). The motif of wildtype, unphosphorylated Plk3 is S/T-X-pY, suggesting
that Plk3 selects sequences for phosphorylation that are similar to sequences bound by
SH2 domains. This similarity in specificity suggests the hypothesis that Plk3 could
phosphorylate phosphotyrosine containing sequences to either increase or decrease their
affinity for SH2 domains. Phosphorylation by Plk3 would place two negative charges on
the amino acid two positions N-terminal to the phosphotyrosine, which could affect SH2
binding as some domains specifically select negatively charged amino acids in this
position while some others select serine or threonine (46). Such modulation of SH2
domain binding and consequently tyrosine kinase signaling may mediate Plk3's role in
substratum adhesion.
Shutting off Plk4 Activity
We showed that the wildtype Plk4 kinase domain was approximately 200-fold
more active the activation loop phosphomimic mutant, Plk4 T170D. This suggests that
phosphorylation on the activation loop of Plk4 might act as a toggle to turn on or off
kinase activity when it is no longer needed either locally or globally. Plk4 is involved in
centriole biogenesis and overexpression results in multiple centrioles (5). Since Plk4 is
expressed from G1 to the next Gl, but centriole duplication takes place in S-phase, it may
127
be that Plk4 kinase activity at the centrioles needs to be inhibited, but activity elsewhere
is still required since Plk4 also seems to be required for mitosis (1, 47).
No kinase has been reported for which phosphorylation on the activation loop has
resulted in inactivation. So, this would be a novel mode of regulation of kinase activity.
To determine if Plk4 is phosphorylated on the activation loop will require a mass
spectrometric approach as an antibody is not currently available. One way to test if
phosphorylation on the activation loop plays a role in regulating Plk4 activity with
respect to centriole duplication would be to express a non-phosphorylatable Plk4 mutant
at near endogenous levels in HeLa cells in the setting of a Plk4 knockdown and an Sphase
arrest
by
aphidicolin
and
look
for
centriole
overduplication
by
immunofluorescence using C-Napl as a marker (5). Expression of wildtype Plk4 and
catalytically inactive Plk4 in the same setting would be appropriate controls.
If
inactivation of Plk4 by phosphorylation on the activation loop is required, then centrioles
would overduplicate in cells expressing the non-phosphorylatable mutant, but not in cells
expressing the wildtype or catalytically inactive forms.
Enhancing the PS-OPLS Technique
Although phosphorylation of proteins is the most widespread post-translational
modification employed by the cell for signal transduction, ubiquitination, sumoylation,
acetylation, and methylation, among others, are also known to play roles in signal
transduction (8, 48). Currently, it is not known if post-translational modifications other
than phosphorylation are part of the motif of any kinases. Ubiquitin and SUMO are too
large to fit in the active site cleft of a kinase, but acetylations and methylations are small
modifications that could play a role in kinase specificity. On proteins such as histones
and p53, phosphorylation, acetylation, and methylation occur on residues in close enough
proximity to each other that one post-translationally modified residue could act as a
specificity determinant
for the enzymes creating subsequent post-translational
modifications (49, 50).
Although, there are no clear examples of acetylated or
methylated residues acting as kinase specificity determinants, phosphorylation of p53 at
S371 inhibits methylation on the adjacent K372 suggesting that the methyltransferase
affinity for the sequence is reduced by the phosphoresidue (50). If phosphoresidues can
128
affect methyltransferase affinity, it does not seem unreasonable that methylated or
acetylated residues could be determinants of kinase specificity. Interestingly, Aurora B
phosphorylates Histone H3 next to trimethylated K9 at S10 during mitosis (51, 52).
Although the unmethylated site matches the motif of Aurora B, the effect of methyllysine on the specificity of the kinase for the site is unknown. It may be that Aurora B
has a higher affinity for a site with methyl-lysine in the -1 position. The peptide libraries
used in the current PS-OPLS technique allow for the testing of specificity towards
phosphothreonine and phosphotyrosine in addition to the twenty natural amino acids, but
the libraries could be expanded to include other post-translational modifications such as
methyl-lysine, dimethyl-lysine, trimethyl-lysine, methyl-arginine, and acetyl-lysine to
allow us to determine if these post-translational modifications affect the specificity of
kinases.
Additionally, this technique could also be modified to enable the determination of
motifs of other enzymes that post-translationally modify proteins. A peptide spot array
has been used to determine the specificity of one methyltransferase (53).
In this
approach, peptides containing single amino acids changes from the sequence of a known
site were synthesized on a membrane in separate spots. Each amino acid in each position
was represented in a different spot. The PS-OPLS technique has the advantages that the
reaction occurs in solution, which reduces artifacts associated with the solid support, and
that each amino acid is tested in each position independently of all others resulting in an
unbiased motif. Methyltransferases and acetyltransferases might be amenable to this
approach. As acetyltransferases and methyltransferase modify not only histones, but
transcription factors, and other proteins, the motifs of these enzymes might help to
determine substrates, which would help to define the roles of these enzymes (50, 54).
The peptide libraries used for the determination of the motifs of these enzymes would
have to be oriented on lysines.
S-adenosyl-methionine with a radioactively labeled
methyl group would be need for the methyltransferases; Acetyl CoA with a radioactively
labeled acetyl group would be needed for the acetyltransferases. The rest of the
experimental set up, however, could be the same.
129
References
1.Fode C, Binkert C, Dennis JW. Constitutive expression of murine sak-a suppresses cell
growth and induces multinucleation. Mol Cell Biol. 1996 Sep;16(9):4665-72.
2. Ma S, Liu MA, Yuan YL, Erikson RL. The serum-inducible protein kinase snk is a G1
phase polo-like kinase that is inhibited by the calcium- and integrin-binding protein CIB.
Mol Cancer Res. 2003 Mar;1(5):376-84.
3. van Vugt MA, Medema RH. Getting in and out of mitosis with polo-like kinase-1.
Oncogene. 2005 Apr 18;24(17):2844-59.
4. Warnke S, Kemmler S, Hames RS, Tsai HL, Hoffmann-Rohrer U, Fry AM, Hoffmann
I. Polo-like kinase-2 is required for centriole duplication in mammalian cells. Curr Biol.
2004 Jul 13;14(13):1200-7.
5. Habedanck R, Stierhof YD, Wilkinson CJ, Nigg EA. The polo kinase Plk4 functions in
centriole duplication. Nat Cell Biol. 2005 Nov;7(11): 1140-6.
6. Lessmann E, Leitges M, Huber M. A redundant role for PKC-epsilon in mast cell
signaling and effector function. Int Immunol. 2006 May;18(5):767-73.
7. Nishikawa K, Toker A, Johannes FJ, Songyang Z, Cantley LC. Determination of the
specific substrate sequence motifs of protein kinase C isozymes. J Biol Chem. 1997 Jan
10;272(2):952-60.
8. Ubersax JA, Ferrell JE,Jr. Mechanisms of specificity in protein phosphorylation. Nat
Rev Mol Cell Biol. 2007 Jul;8(7):530-41.
9. Salamanca DA, Khalil RA. Protein kinase C isoforms as specific targets for
modulation of vascular smooth muscle function in hypertension. Biochem Pharmacol.
2005 Nov 25;70(11):1537-47.
10. Barr AR, Gergely F. Aurora-A: The maker and breaker of spindle poles. J Cell Sci.
2007 Sep 1;120(Pt 17):2987-96.
11. Crosio C, Fimia GM, Loury R, Kimura M, Okano Y, Zhou H, Sen S, Allis CD,
Sassone-Corsi P. Mitotic phosphorylation of histone H3: Spatio-temporal regulation by
mammalian aurora kinases. Mol Cell Biol. 2002 Feb;22(3):874-85.
12. Sonoda E, Matsusaka T, Morrison C, Vagnarelli P, Hoshi O, Ushiki T, Nojima K,
Fukagawa T, Waizenegger IC, Peters JM, Earnshaw WC, Takeda S. Sccl/Rad21/Mcdl is
required for sister chromatid cohesion and kinetochore function in vertebrate cells. Dev
Cell. 2001 Dec;1(6):759-70.
13. Cohen-Fix O, Peters JM, Kirschner MW, Koshland D. Anaphase initiation in
saccharomyces cerevisiae is controlled by the APC-dependent degradation of the
anaphase inhibitor Pdslp. Genes Dev. 1996 Dec 15;10(24):3081-93.
130
14. Ciosk R, Zachariae W, Michaelis C, Shevchenko A, Mann M, Nasmyth K. An
ESP 1/PDS 1 complex regulates loss of sister chromatid cohesion at the metaphase to
anaphase transition in yeast. Cell. 1998 Jun 12;93(6):1067-76.
15. Uhlmann F, Lottspeich F, Nasmyth K. Sister-chromatid separation at anaphase onset
is promoted by cleavage of the cohesin subunit Sccl. Nature. 1999 Jul 1;400(6739):3742.
16. Alexandru G, Uhlmann F, Mechtler K, Poupart MA, Nasmyth K. Phosphorylation of
the cohesin subunit Scel by Polo/Cdc5 kinase regulates sister chromatid separation in
yeast. Cell. 2001 May 18;105(4):459-72.
17. Xia G, Luo X, Habu T, Rizo J, Matsumoto T, Yu H. Conformation-specific binding
of p3 l(comet) antagonizes the function of Mad2 in the spindle checkpoint. EMBO J.
2004 Aug 4;23(15):3133-43.
18. Luo X, Tang Z, Xia G, Wassmann K, Matsumoto T, Rizo J, Yu H. The Mad2 spindle
checkpoint protein has two distinct natively folded states. Nat Struct Mol Biol. 2004
Apr;11(4):338-45.
19. De Antoni A, Pearson CG, Cimini D, Canman JC, Sala V, Nezi L, Mapelli M, Sironi
L, Faretta M, Salmon ED, Musacchio A. The Madl/Mad2 complex as a template for
Mad2 activation in the spindle assembly checkpoint. Curr Biol. 2005 Feb 8; 15(3):214-25.
20. Peters JM. Checkpoint control: The journey continues. Curr Biol. 2008 Feb
26;18(4):R170-2.
21. Wong OK, Fang G. Loading of the 3F3/2 antigen onto kinetochores is dependent on
the ordered assembly of the spindle checkpoint proteins. Mol Biol Cell. 2006
Oct; 17(10):4390-9.
22. Ahonen LJ, Kallio MJ, Daum JR, Bolton M, Manke IA, Yaffe MB, Stukenberg PT,
Gorbsky GJ. Polo-like kinase 1 creates the tension-sensing 3F3/2 phosphoepitope and
modulates the association of spindle-checkpoint proteins at kinetochores. Curr Biol. 2005
Jun 21;15(12):1078-89.
23. Wong OK, Fang G. Plxl is the 3F3/2 kinase responsible for targeting spindle
checkpoint proteins to kinetochores. J Cell Biol. 2005 Aug 29;170(5):709-19.
24. Yang M, Li B, Tomchick DR, Machius M, Rizo J, Yu H, Luo X. p3 Icomet blocks
Mad2 activation.through structural mimicry. Cell. 2007 Nov 16;131(4):744-55.
25. Li Z, Niu J, Uwagawa T, Peng B, Chiao PJ. Function of polo-like kinase 3 in NFkappaB-mediated proapoptotic response. J Biol Chem. 2005 Apr 29;280(17):16843-50.
26. Xie S, Wu H, Wang Q, Kunicki J, Thomas RO, Hollingsworth RE, Cogswell J, Dai
W. Genotoxic stress-induced activation of Plk3 is partly mediated by Chk2. Cell Cycle.
2002 Nov-Dec; 1(6):424-9.
131
27. Bahassi el M, Conn CW, Myer DL, Hennigan RF, McGowan CH, Sanchez Y,
Stambrook PJ. Mammalian polo-like kinase 3 (Plk3) is a multifunctional protein involved
in stress response pathways. Oncogene. 2002 Sep 26;21(43):6633-40.
28. Bahassi el M, Hennigan RF, Myer DL, Stambrook PJ. Cdc25C phosphorylation on
serine 191 by Plk3 promotes its nuclear translocation. Oncogene. 2004 Apr
8;23(15):2658-63.
29. Xie S, Wang Q,Ruan Q,Liu T, Jhanwar-Uniyal M, Guan K, Dai W. MEKI-induced
golgi dynamics during cell cycle progression is partly mediated by polo-like kinase-3.
Oncogene. 2004 May 6;23(21):3822-9.
30. Zimmerman WC, Erikson RL. Polo-like kinase 3 is required for entry into S phase.
Proc Natl Acad Sci U S A. 2007 Feb 6;104(6):1847-52.
31. Conn CW, Hennigan RF, Dai W, Sanchez Y, Stambrook PJ. Incomplete cytokinesis
and induction of apoptosis by overexpression of the mammalian polo-like kinase, Plk3.
Cancer Res. 2000 Dec 15;60(24):6826-31.
32. van de Weerdt BC, Littler DR, Klompmaker R, Huseinovic A, Fish A, Perrakis A,
Medema RH. Polo-box domains confer target specificity to the polo-like kinase family.
Biochim Biophys Acta. 2008 Mar 5
33. Kauselmann G, Weiler M, WulffP, Jessberger S, Konietzko U, Scafidi J, Staubli U,
Bereiter-Hahn J, Strebhardt K, Kuhl D. The polo-like protein kinases fnk and snk
associate with a ca(2+)- and integrin-binding protein and are regulated dynamically with
synaptic plasticity. EMBO J. 1999 Oct 15;18(20):5528-39.
34. Pak DT, Sheng M. Targeted protein degradation and synapse remodeling by an
inducible protein kinase. Science. 2003 Nov 21;302(5649): 1368-73.
35. Jiang N, Wang X, Jhanwar-Uniyal M, Darzynkiewicz Z, Dai W. Polo box domain of
Plk3 functions as a centrosome localization signal, overexpression of which causes
mitotic arrest, cytokinesis defects, and apoptosis. J Biol Chem. 2006 Apr
14;281(15):10577-82.
36. Zimmerman WC, Erikson RL. Finding Plk3. Cell Cycle. 2007 Jun 1;6(11):1314-8.
37. Kang T, Wei Y, Honaker Y, Yamaguchi H, Appella E, Hung MC, Piwnica-Worms H.
GSK-3 beta targets Cdc25A for ubiquitin-mediated proteolysis, and GSK-3 beta
inactivation correlates with Cdc25A overproduction in human cancers. Cancer Cell. 2008
Jan;13(1):36-47.
38. O'Neill T, Giarratani L, Chen P, Iyer L, Lee CH, Bobiak M, Kanai F, Zhou BB,
Chung JH, Rathbun GA. Determination of substrate motifs for human Chkl and
hCdsl/Chk2 by the oriented peptide library approach. J Biol Chem. 2002 May
3;277(18):16102-15.
132
39. van de Weerdt BC, van Vugt MA, Lindon C, Kauw JJ, Rozendaal MJ, Klompmaker
R, Wolthuis RM, Medema RH. Uncoupling anaphase-promoting complex/cyclosome
activity from spindle assembly checkpoint control by deregulating polo-like kinase 1.
Mol Cell Biol. 2005 Mar;25(5):2031-44.
40. Frame MC, Fincham VJ, Carragher NO, Wyke JA. v-src's hold over actin and cell
adhesions. Nat Rev Mol Cell Biol. 2002 Apr;3(4):233-45.
41. Holtrich U, Wolf G, Yuan J, Bereiter-Hahn J, Kam T, Weiler M, Kauselmann G,
Rehli M, Andreesen R, Kaufmann M, Kuhl D, Strebhardt K. Adhesion induced
expression of the serine/threonine kinase fnk in human macrophages. Oncogene. 2000
Oct 5;19(42):4832-9.
42. Pawson T, Gish GD, Nash P. SH2 domains, interaction modules and cellular wiring.
Trends Cell Biol. 2001 Dec;1 1(12):504-11.
43. Schlessinger J, Lemmon MA. SH2 and PTB domains in tyrosine kinase signaling. Sci
STKE. 2003 Jul 15;2003(191):RE12.
44. Songyang Z, Shoelson SE, Chaudhuri M, Gish G, Pawson T, Haser WG, King F,
Roberts T, Ratnofsky S, Lechleider RJ. SH2 domains recognize specific phosphopeptide
sequences. Cell. 1993 Mar 12;72(5):767-78.
45. Songyang Z, Shoelson SE, McGlade J, Olivier P, Pawson T, Bustelo XR, Barbacid
M, Sabe H, Hanafusa H, Yi T. Specific motifs recognized by the SH2 domains of csk,
3BP2, fps/fes, GRB-2, HCP, SHC, syk, and vav. Mol Cell Biol. 1994 Apr;14(4):2777-85.
46. Huang H, Li L, Wu C, Schibli D, Colwill K, Ma S, Li C, Roy P, Ho K, Songyang Z,
Pawson T, Gao Y, Li SS. Defining the specificity space of the human SRC homology 2
domain. Mol Cell Proteomics. 2008 Apr;7(4):768-84.
47. Hudson JW, Kozarova A, Cheung P, Macmillan JC, Swallow CJ, Cross JC, Dennis
JW. Late mitotic failure in mice lacking sak, a polo-like kinase. Curr Biol. 2001 Mar
20;11(6):441-6.
48. Seet BT, Dikic I, Zhou MM, Pawson T. Reading protein modifications with
interaction domains. Nat Rev Mol Cell Biol. 2006 Jul;7(7):473-83.
49. Yang XJ. Multisite protein modification and intramolecular signaling. Oncogene.
2005 Mar 3;24(10):1653-62.
50. Morgunkova A, Barlev NA. Lysine methylation goes global. Cell Cycle. 2006
Jun;5(12):1308-12.
51. Fischle W, Tseng BS, Dormann HL, Ueberheide BM, Garcia BA, Shabanowitz J,
Hunt DF, Funabiki H, Allis CD. Regulation of HP1-chromatin binding by histone H3
methylation and phosphorylation. Nature. 2005 Dec 22;438(7071):1116-22.
133
52. Hirota T, Lipp JJ, Toh BH, Peters JM. Histone H3 serine 10 phosphorylation by
aurora B causes HP1 dissociation from heterochromatin. Nature. 2005 Dec
22;438(7071):1176-80.
53. Rathert P, Zhang X, Freund C, Cheng X, Jeltsch A. Analysis of the substrate
specificity of the dim-5 histone lysine methyltransferase using peptide arrays. Chem Biol.
2008 Jan;15(1):5-11.
54. Sterner DE, Berger SL. Acetylation of histones and transcription-related factors.
Microbiol Mol Biol Rev. 2000 Jun;64(2):435-59.
134