Regulation of mitotic spindle biogenesis in budding yeast

Therefore, the assembly of the mitotic spindle is critically dependent on centrosome duplication, separation of sister centrosomes and microtubule dynamics.. Chromosome segregation, the

REGULATION OF MITOTIC SPINDLE BIOGENESIS

INSTITUTE OF MOLECULAR AND CELL BIOLOGY

NATIONAL UNIVERSITY OF SINGAPORE

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude and appreciation to A/P Uttam Surana, to whom I

am much indebted for his guidance, insightful and influential conversations, valuable advice and the freedom to explore my curiosities My sincere thanks also go out to members of my PhD

Supervisory Committee, A/P Mohan Balasubramanian and A/P Yang Xiaohang, for their constructive comments and encouragement

Special Thanks to my extended family, my labmates – Hong Hwa, Chee Seng, Wei Chun,

Jonathon, Joan, Zhang Tao, Saurabh, Jenn Hui, San Ling, Wee Kheng and Vaidehi for

sharing in the thrill along the path to discovery, help in various ways and for making lab life fun!

A big Thank You to all in CMJ lab for being such wonderful neighbours! I am also grateful to

Suniti Naqvi, Mithilesh Mishra and all in Mohan’s lab at TLL for their time in teaching me

fission yeast techniques I wish to thank Prof Mark Winey for providing technical expertise in

electron microscopy for various projects Thanks also to the EM Unit, NUS for my training and

Chee Peng for help despite his busy schedule

I am grateful to Drs Mark Winey, David Morgan, Wolfgang Zachariae, John Kilmartin,

Matthias Peter, David Pellman, Chris Hardy, Kyung Lee, Jiri Lukas and Michel Bouvier for

providing me with valuable reagents and Dr Mark Hall for helpful advice

Special Thanks to Ram, Trich, Jaya, Xianwen, Lee Thean, Kar Lai, Rida, Foong May, Suniti,

Vani, Srini, Shal, Nee, Indra for their friendship, the fun times and encouragement Thanks to all

at Opus Dei for their friendship and prayers, and for bringing out the best in me in my daily work

Most importantly, this thesis is dedicated to my loving FAMILY especially my parents for

having always encouraged me to aim high, for much-valued support, understanding, advice, sacrifices, prayers and constant cheer that made this journey of scientific discovery possible

Thank You Daddy, Mummy, Sharon and Renita!

Karen Crasta, June 2007

Abbreviations

Ab Antibody

1NM-PP1 4-amino-1-tert-butyl-3-(1-naphthylmethyl) pyrazolo [3, 4-d] pyrimidine

BSA bovine serum albumin

ECL Enhanced chemiluminescence

EDTA ethylenediamine tetraacetic acid

PCR polymerase chain reaction

PEG polyethylene glycol

PMSF phenylmethylsulfonylfluoride

Raff Raffinose

RNA ribonucleic acid

SDS sodium dodecyl sulfate

SSC saline sodium citrate

TE Tris-EDTA buffer

ts temperature-sensitive

YEP yeast extract-peptone

TABLE OF CONTENTS

Acknowledgements………i

Table of Contents……….ii

Summary……… vi

List of Tables……….viii

List of Figures……… ix

Abbreviations……… xi

Chapter 1 Introduction………1

1.1 Introductory Remarks………1

1.2 Overview of Budding Yeast Cell Cycle ……… 2

1.2.1 Saccharomyces cerevisiae cell cycle and cyclin-dependent kinase Cdc28……… 2

1.2.1.1 Inhibitory Phosphorylation on Cdc28-Tyr 19 ………4

1.2.1.2 Structural basis for Cdk activation ……….………5

1.2.1.3 Conditional cdc28 mutants……… 7

1.2.2 Coordination of cell cycle events and checkpoints……… 9

1.3 Protein Degradation in cell cycle control ……… 10

1.3.1 Ubiquitin-dependent proteolysis ……… 10

1.3.2 SCF……….11

1.3.3 APC………12

1.3.3.1 Selective substrate recognition by APC………12

1.3.3.2 APC-Cdc20 at metaphase-to-anaphase transition……….15

1.3.3.3 APC-Cdh1 at the end of mitosis………15

1.4 The bipolar mitotic spindle……… 17

1.5 The centrosome cycle……… 20

1.6 The spindle pole body (SPB) cycle……….25

Chapter 2 Materials and Methods………29

2.1 Materials……… 30

2.2 Methods……… 35

2.2.1 Strains and Culture Conditions……… 35

2.2.2 Cell Synchronization Procedures……… 36

2.2.3 GAL-HO induction for construction of polyploidy strains………36

2.2.4 Yeast Transformation……….36

2.2.5 Isolation of plasmid DNA from yeast cells………37

2.2.6 High-copy Suppression Screen……… 37

2.2.7 Preparation of Yeast Chromosomal DNA……… 38

2.2.8 Southern Blot Analysis……… 39

2.2.9 Northern Blot Analysis……… 39

2.2.10 Immunofluorescent Staining……… 40

2.2.11 Visualization of Fluorescent Protein Signals……….41

2.2.12 Flow Cytometric Analysis……… 42

2.2.13 Transmission Electron Microscopic Analysis………42

2.2.13.1 Chemical Fixation and Embedding of Yeast Cells……… 42

2.2.13.2 Microtome Sectioning, Staining and Viewing under TEM……… 43

2.2.14 Bioluminescence Resonance Energy Transfer (BRET2) assay……….44

2.2.15 Preparation of Cell Extracts for Protein Analysis……… 44

2.2.15.1 Cellular lysis using acid-washed glass beads……… 44

2.2.15.2 Protein Precipitation using Tri-Chloroacetic Acid (TCA)……… 45

2.2.16 Western Blot Analysis………45

2.2.17 Pulse-chase experiments……….46

2.2.18 Immunoprecipitation of HA and cmyc-tagged proteins……… 46

2.2.19 Kinase Assays………47

2.2.20 Detection of Cdc28 tyrosine phosphorylation………48

2.2.21 Detection of ubiquitin conjugates in vivo……… 48

2.2.22 Coomasie Blue Staining……….49

2.2.23 Silver Staining………49

2.2.24 Expression and Purification of GST-tagged proteins……….50

Chapter 3 The Regulatory Role of Cdc28 in SPB Separation………52

3.1 Background……….52

3.2 Results……….54

3.2.1 cdc28Y19E and cdc28-as1 cells are unable to separate SPBs………54

3.2.2 cdc28 mutants defective in SPB Separation Do Not Activate the Spindle Checkpoint……… ………… 56

3.2.3 Genetic Screen to Identify Downstream Targets of Cdc28 in SPB Separation….57 3.2.4 Ectopic Expression of Microtubule-Associated Proteins Induces Spindle Formation……….……… ……59

3.2.5 SPB separation does not require Cdc28-mediated phosphorylation of microtubule-associated proteins……… 63

3.2.6 Low Endogenous Levels of Cin8, Kip1 and Ase1 in cdc28Y19E and cdc28-as1 63

3.2.7 Defect in SPB separation is due to proteasomal degradation of microtubule associated proteins……… 68

3.2.8 APCCdh1, but not APCCdc20, Prevents SPB Separation………72

3.2.9 Cdh1 phosphorylation and Cin8 ubiquitylation in cdc28 mutants defective in spindle assembly………76

3.2.10 Cdc28/Clb activity controls Cdh1 subcellular localization and spindle assembly.77 3.2.11 Cdc28-phosphorylation sites in Cdh1 and stability of Cin8 and Clb2………… 79

Separation……… 80

3.2.13 Tyrosine dephosphorylation of Cdc28 temporally precedes spindle assembly during a normal cell cycle……… 83

3.3 Discussion……… 83

Chapter 4 Inactivation of Cdh1 by synergistic action of Cdc28 and Cdc5 is essential for spindle assembly……….94

4.1 Background……….94

4.2 Results……….94

4.2.1 Effects of ectopic expression of Cdc5 in cdc28-as1 cells……… 94

4.2.2 Phosphorylation of Cdh1 by Cdc5 requires priming by Cdc28……….99

4.2.3 Cdc5 has a role in bipolar spindle assembly in budding yeast……….101

4.2.4 Cdc5 degradation in cdc28-as1 cells………110

4.3 Discussion……….112

Chapter 5 Matters Arising……… 119

5.1 Synergistic action of Cdk1 and Plk1 on Mammalian Cdh1 ……….119

5.2 The Paradox: Active Cdh1 Is Degraded………121

5.3 Role of Cdc20 in SPB separation……… 124

5.4 Temporal regulation of satellite formation by inactivation of mitotic kinase………… 128

Chapter 6 Conclusion and Perspectives…… ……… 133

References……….138 Appendices

List Of Figures

Figure 2 Cdc28 function is regulated by inhibitory phosphorylation by Swe1 and

dephosphorylation by Mih1……….6

Figure 3 Schematic representation of a centrosome and spindle pole body (SPB)……… 21

Figure 6 The G2/M arrest phenotype of cdc28-as1 and cdc28Y19E cells is not due to

activation of the spindle checkpoint……… 58

Figure 7 Ectopic expression of microtubule-associated proteins induces spindle

Figure 10 Proteasomal degradation of microtubule-associated proteins in cdc28 mutants…70

Figure 11 APCCdh1-mediated degradation of microtubule-associated proteins prevents

spindle assembly………74

Figure 12 Phosphorylation status of Cdh1 determines its subcellular localization and Cin8

ubiquitylation……….78

Figure 13 Phosphorylation sites in Cdh1 essential for spindle formation……… 81

Figure 14 Cin8-bundling activity, not its motor activity, is required for SPB separation… 82

Figure 15 Tyrosine dephosphorylation correlates with the timing of SPB separation during

normal cell cycle………84

Figure 16 Model depicting role of activated Cdc28 (Cdk1) in SPB separation in budding

yeast………93

Figure 17 Ectopic expression of Cdc5 causes hyperphosphorylation of Cdh1 and SPB

separation……… 97

Figure 20 Inhibition of APCCdh1 by Cdc5 and Acm1, and involvement of Cdc5 in nuclear

export of Cdh1……… 109

Figure 22 A scheme for the regulation of SPB separation involving Cdc28, Cdc5, Cdh1,

Acm1 and microtubule-binding proteins Cin8 and Kip1……….111

Figure 24 Two-step kinase reaction of bacterially-expressed human Cdh1 with mammalian

Cdk1 and Plk1……… 123

Figure 27 Temporal regulation of half-bridge elongation and satellite deposition by

inactivation of mitotic kinase……… 131

List Of Tables

Table 1 List of yeast strains used in this study………30 Table 2 List of oligonucleotides used in this study……….33

SUMMARY

Partitioning of chromosomes equally to progeny cells is the central function of mitosis For chromosome segregation to proceed accurately, the sister chromatids must be attached to the highly-organized microtubule-based structure, the mitotic spindle Centrosomes (spindle pole body

or SPB in yeast) constitute the two opposite poles of the bipolar spindle and serve as microtubule organizing centers (MTOC) Microtubules radiating out of centrosomes eventually connect to sister chromatids and mediate their equal partitioning to the progeny cells during mitosis At the time of their birth, progeny cells inherit only one centrosome from the progenitor cell Therefore, the assembly of the mitotic spindle is critically dependent on centrosome duplication, separation of sister centrosomes and microtubule dynamics The duplication of the centrosome and SPB has been extensively investigated over the last two decades, but little is known about the processes and regulators underlying centrosome and SPB separation Since many aspects of mammalian centrosome and yeast SPB duplication are similar despite their overt structural differences, this study focuses on the regulation and mechanism of SPB separation in the budding yeast

Saccharomyces cerevisiae

Activation of Cdc28 (Cdk1) by tyrosine-19 dephosphorylation is known to be essential for

SPB separation, the first step in bipolar spindle formation (Lim et al., 1996) However the exact

role that Cdc28 plays in the process is not known In the first part of this study, we have investigated the regulatory role of Cdc28 in SPB separation We show that the ubiquitin ligase APCCdh1 acts as a potent inhibitor of spindle formation by promoting degradation of microtubule-associated proteins Cin8, Kip1 and Ase1 which are essential for SPB separation Activated Cdc28 kinase causes inactivation of APCCdh1 during S phase, resulting in the accumulation of these SPB-separation promoting proteins Since ectopic expression of Cin8, Kip1 or Ase1 is sufficient for SPB separation even in the absence of Cdc28-Clb activity, we propose that stabilization of these mechanical force-generating proteins is highly likely to be the predominant role of Cdc28-Clb in

SPB separation Interestingly, our results also indicate that SPB separation is dependent on the microtubule-bundling activity of Cin8 (a plus-end motor protein belonging to the conserved BimC family of spindle motors) and not on its motor function

While the first part of this study examines the consequences of Cdc28-mediated phosphorylation of Cdh1, the next part reveals that this phosphorylation is necessary to create a

phosphopeptide domain that acts as a docking site for the Polo kinase Cdc5 (Elia et al., 2003)

Once bound, Cdc5 further phosphorylates Cdh1 at distinct sites Hence, synergistic action of Cdc28 and Cdc5 is required for complete inactivation of Cdh1 and is thus important for spindle assembly Our results also show that Cdc5 becomes essential for SPB separation in the absence of Acm1, a negative regulator of APCCdh1 that is independent of the status of Cdh1 phosphorylation Although Polo kinase is a well-known regulator of centrosome separation (its exact role remains unknown), Cdc5 has never been implicated in SPB separation in budding yeast prior to this study Hence we have identified a novel role for Polo kinase in SPB separation, together with Cdc28 and Acm1

Finally, this study also deals with a few issues that arose during the course of this work It attempts to rationalize the results obtained and raises a number of questions that may be subjects for future investigations to better understand the intricacies involved in the regulation of spindle biogenesis

Chapter 1 Introduction

In the time it takes you to read this dissertation, several billions of cells in your body will have successfully completed the highly regulated process of mitosis, during which a mother cell containing duplicated chromosomes divides into two daughter cells, each with identical ploidy to the mother cell Chromosome segregation, the physical partitioning of chromosomes between progeny cells, is the central function of mitosis during a cell division cycle Accurate segregation

of sister chromatids is orchestrated by one of Nature’s most beautiful cellular structures, the mitotic spindle This strictly bipolar, microtubule-based apparatus pulls sister chromatids apart before mitosis is completed to ensure that duplicated chromosomes are distributed equally between daughter cells Therefore, without a spindle, efficient chromosome segregation is impossible and would lead to genomic instability and aneuploidy, often associated with cancers (reviewed in Jallepalli and Lengauer, 2001) Faithful chromosome segregation is thus critically dependent upon the formation of a bipolar mitotic spindle Hence, understanding the regulation of mitotic spindle biogenesis, the subject of this dissertation, is crucial for gaining insights into the chromosome segregation process that is central to cellular reproduction

The budding yeast Saccharomyces cerevisiae was used as an experimental system in this

study because its amenability to genetic manipulation has long established it as a model organism for investigations into most fundamental cellular control circuitry Well-known examples include

the study of cell division cycle (cdc) mutants which identified regulators involved in specific

functions during the cell cycle, and checkpoints, surveillance mechanisms that arrest cells at specific stages of the cell cycle (discussed in greater detail later) Since spindle biogenesis is coordinately regulated with other cell cycle events both in time and space, it is appropriate to

briefly describe the general nature of the cell division cycle in Saccharomyces cerevisiae

1.2 Overview of Budding Yeast Cell Cycle

1.2.1 Saccharomyces cerevisiae cell cycle and cyclin-dependent kinase Cdc28

The cell cycle of the budding yeast Saccharomyces cerevisiae is currently the best understood of

all eukaryotes Cell division in budding yeast is accomplished by the coordinated control of the cell cycle clock consisting of four distinct phases (G1, S, G2, M) with the G2 phase being extremely short (Fig 1) These phases are a temporally organized series of interlocking processes rather than a series of time intervals No group of regulatory proteins is as intimately connected to cell cycle progression as the cyclin-dependent kinases, Cdks In budding yeast the sole essential Cdk, Cdc28 (highly homologous to the prototype Cdc2 or Cdk1 and sometimes referred to as cdc2/Cdk1) is considered to be the master regulator of the cell cycle (reviewed in Mendenhall and Hodge, 1998) Catalytically inactive Cdc28 requires association with its regulatory partners, the cyclins, for activity Different Cdc28/cyclin complexes are active periodically during the cell cycle and are responsible for driving the cell from one phase to the next (Figure 1) In late G1, passage through START (START marks the commitment of the cell to a new cycle of division), requires activation of Cdc28 by the G1 cyclins Cln1, Cln2 and Cln3, while at S phase, Cdc28 associates with S-phase cyclins Clb5 and Clb6 to initiate chromosome duplication and at mitosis (M phase) with the mitotic cyclins, Clb1, Clb2, Clb3 and Clb4, to initiate M phase

Cdc28 exerts its effects by phosphorylating presumably a large number of proteins While the identity of many of these substrates remain unknown, it is known that Cdc28 phosphorylates its substrates at serine or threonine residues in a specific sequence context that is recognized by its active site The typical phosphorylation sequence for Cdks is [S/T*]PX[K/R], where S/T* indicates the phosphorylated serine or threonine, X represents any amino acid and K/R represents the basic amino acid lysine (K) or arginine (R) In most cases, the target serine (S) or threonine (T) residue is followed by a proline (P); it is also highly favourable for the target residue to have a

basic amino acid two positions after the target residue (Brown et al., 1999)

Activation of the distinct Cdc28-cyclin complexes initiates a number of cellular and morphological events START, which requires Cdc28-Cln complexes, is not a singular event but signifies integration of various intracellular and external cues after which cells are irreversibly committed to the division program This commitment triggers a number of cellular events among which emergence of a bud (also referred to as a daughter cell) and duplication of spindle-pole body (yeast equivalent of mammalian centrosome essential for mitotic spindle formation) are morphologically most prominent Cdc28-Clb5/Clb6 complex is required for onset of S phase and the initiation of DNA replication As cells progress towards late S phase, the Cdc28-Clb1/Clb2/Clb3/Clb4 complexes accumulate to initiate the entry into M phase and biogenesis of the mitotic spindle Once a spindle is assembled, it grows in length and forges an amphitellic (bipolar) attachment to duplicated chromosomes Upon arrival of the appropriate cues, the spindle helps to transmit one set of chromosomes to each of the progeny cells One of the most central and spectacular event of the cell cycle, chromosome segregation is highly coordinated to ensure accurate transmission of chromosomes At the end of M phase, cyclins undergo proteolytic destruction resulting in a rapid decline in Cdc28 activity which allows cells to exit mitosis Mitotic cyclin destruction is the most prominent hallmark of the end of M phase which re-sets the cell cycle clock and permits cells to undergo cytokinesis (cell separation) and proceed to the next cycle of division

1.2.1.1 Inhibitory Phosphorylation on Cdc28-Tyr 19

Cyclin binding alone is not sufficient to fully activate Cdc28 Its complete activation also requires

a combination of phosphorylation and dephosphorylation of evolutionary-conserved residues

within its sequence Phosphorylation of a conserved threonine-169 (equivalent to Thr-167 in S

pombe and Thr-161 in human cells) in the T-loop adjacent to the kinase active site is catalyzed by

the Cdk-activating kinase (CAK) (Gould et al., 1991; Kaldis et al., 1996) In mammalian cells,

yeast, phosphorylation precedes cyclin binding (Kaldis et al., 1998; Ross et al., 2000) In both

cases however, cyclin binding and not phosphorylation is the highly regulated, rate-limiting step in Cdk activation

Two inhibitory phosphorylations at threonine-14 and tyrosine-15 (equivalent to Tyr-19 in

S cerevisiae) within the ATP-binding domain also have important functions in the regulation of

Cdk activity The phosphorylation state of these residues, first described in fission yeast

Schizosaccharomyces pombe, is controlled by a balance of opposing kinase and phosphatase

activities acting at these sites which influence initiation of mitosis Tyr-15 on Cdc2 is phosphorylated by Wee1 kinase, which impedes Cdc2 activity and prevents entry into mitosis, and

is dephosphorylated by Cdc25 phosphatase, which reverses this phosphorylation and activates Cdc2 (Fig 2) These critical steps at mitotic entry appear to be largely conserved throughout

evolution (Russell et al., 1989; Nurse, 1990) Substituting Tyr-15 with phenylalanine, which

mimics constitutively dephosphorylated tyrosine, results in premature onset of mitosis (Gould and Nurse, 1989) Thus, the state of Tyr-15 phosphorylation is central in the decision to enter mitosis

In budding yeast, Cdc28, like Cdc2, also shows a marked loss of Tyr-19 phosphorylation in cells

arrested in mitosis (Amon et al., 1992) and is phosphorylated by Swe1 (Wee1 homologue) and

dephosphorylated by Mih1 (Cdc25 homologue) Surprisingly however, substitution of Tyr-19 by

phenylalanine does not lead to precocious mitosis (Amon et al., 1992), indicating that Tyr-19

dephosphorylation is not the rate-limiting step for initiation of mitosis Thus, other pathways whose activation in conjunction with Cdc28 kinase is necessary for triggering mitosis could exist

in budding yeast

1.2.1.2 Structural basis for Cdk activation

From structural studies of Cdk2 and Cdk2-cyclin A complex (De Bondt et al., 1993; Jeffrey et al., 1995; Russo et al., 1996; Brown et al., 1999), it has been deciphered that in the absence of cyclin,

comprising also of a lysine residue and an aspartic acid residue) on the highly conserved PSTAIRE helix of Cdk1 is situated outside its catalytic cleft In addition, the T-loop that contains Thr-161 sits in front of the cleft and blocks access of substrates to ATP Cyclin binding induces a conformational change such that it both moves the T loop away, allowing access to this ATP-binding site, as well as places the glutamic acid residue inside the cleft to realign the active site residues which can then coordinate a magnesium ion and ATP phosphate atoms for catalytic activity Phosphorylation of Thr-161 in the T-loop by CAK causes additional conformational changes that further modify the substrate-binding surface, greatly increasing its affinity for protein substrates Phosphorylation of Thr-161 flattens the T-loop which then moves closer to the cyclin; this region serves as a key component of the binding site for protein substrates containing the S/T-P-X-K/R motif mentioned previously The proline residue in this motif interacts with the backbone

of the T-loop while the positively-charged lysine or arginine residue interacts with the negatively charged phosphate on Thr-161 The Thr-14 and Tyr-15 residues of Cdc2 are located within the ATP-binding region Phosphorylation of these residues exerts an inhibitory effect by hindering the phosphate transfer from ATP to substrates due to electrostatic repulsion between the phosphates

on these residues and the phosphates of ATP These phosphorylations inhibit kinase activity even when Cdc2 is bound by cyclin A and Thr-161 is phosphorylated Hence, dephosphorylation of Thr-14 and Tyr-15 constitutes a crucial final step in the activation of Cdc2

1.2.1.3 Conditional cdc28 mutants

Many fundamental insights into cell cycle control have been gained through the isolation of

temperature-sensitive (ts) cell division cycle (cdc) mutants Lee Hartwell, who was granted the

Nobel Prize in 2001, pioneered cell cycle genetics in budding yeast with his discovery that upon

shifting to the restrictive temperature, many ts cdc mutants arrest their cell population with the same morphology (Hartwell et al., 1970), suggesting that they arrest at a specific point at which

are essential for cell cycle progression, cdc mutations are conditional i.e the gene product is functional only when subjected to certain conditions Because most cdc mutants are temperature-

non-sensitive, their gene products are functional at the permissive temperature (typically 24˚C) and are

inactivated at the restrictive temperature (typically 37˚C) One of the earliest analyses of cdc mutants led to the discovery of the CDC28 gene (Hartwell et al., 1974; Beach et al., 1982)

A requirement for Cdc28 at the G1/S and G/2M transitions was identified from phenotypic

analyses of various temperature-sensitive alleles of CDC28 For example, the cdc28-4 mutant cells (histidine at position 128 replaced by tyrosine) arrested in G1 with unbudded cells,

unreplicated DNA and no spindles, indicating that these cells were defective in undergoing

passage through START (Reed, 1980) The cdc28-1N mutant cells (proline at position 250

replaced by leucine), on the other hand, arrested in G2/M with large hyperpolarized buds,

replicated DNA and a short spindle (Piggott et al., 1982; Surana et al., 1991) These phenotypes

clearly suggested that Cdc28 function was required at both G1/S and G2/M transitions While it was known that Tyr19 dephosphorylation is necessary for the onset of mitosis, the clue to it being

required for a specific mitotic event came from the analysis of cdc28-Y19E allele (tyrosine at position 19 replaced by glutamic acid) Cells driven by cdc28-Y19E arrest with large

hyperpolarized buds and replicate their DNA, but fail to assemble a bipolar spindle, suggesting a

role for Cdc28 in spindle formation (Lim et al., 1996)

Besides “traditional” genetics, a “chemical genetics” approach has also been utilized in the analysis of Cdc28 function in the cell This strategy involves making a mutation in the ATP-binding site that does not affect the activity of the kinase but creates an extra pocket to fit an ATP analogue bearing an accessory group This sensitizes the mutant kinase to inhibition by the ATP

analogue that is too bulky to fit wild-type kinases and consequently, cannot inhibit them The analogue-sensitive cdc28-as1 allele carries a phenylalanine to glycine substitution at amino acid

position 88 (F88G) which alters the ATP-binding pocket and confers sensitivity to a bulky ATP

[3, 4-d] pyrimidine) (Bishop et al., 2000) These cells progress normally through the cell cycle;

however, chemical inhibition with 500 nM of 1NM-PP1 induced these cells to arrest at G2/M with

large hyperpolarized buds, replicated DNA and without spindles (Bishop et al., 2000) A higher

concentration of 5000 nM induces a uniform G1 arrest in which cells failed to form buds, replicate DNA or form spindles

1.2.2 Coordination of cell cycle events and checkpoints

Efficient and accurate chromosome segregation depends on the precise temporal coordination of cellular events which ensure orderly progression through the cell cycle Although intrinsic timing mechanisms can initiate cell cycle events in the correct sequence, the order of events also depends

on surveillance mechanisms known as “checkpoints” Checkpoints ensure that if a certain event is interrupted or executed erroneously, the subsequent phase of the cell cycle is not initiated (Hartwell and Weinert, 1989)

In budding yeast, four major checkpoint controls have been described: Morphogenetic checkpoint, DNA replication checkpoint, DNA damage checkpoint and Spindle checkpoint While the morphogenetic checkpoint delays cell cycle progression in response to perturbations of cell polarity that prevent bud formation, the DNA replication checkpoint prevents entry into mitosis in response to the inhibition of DNA replication caused by drugs such as hydroxyurea that

result in stalled replication forks (Osborn et al., 2002) DNA damage checkpoint is activated in

response to genotoxic stresses and prevents segregation of damaged chromosomes (Melo and Toczyski, 2002) Spindle checkpoint is triggered when perturbation in various aspects of spindle dynamics are detected Following DNA replication in S phase, duplicated chromosomes are held

together by the cohesin complex till anaphase (reviewed in Nasmyth et al., 2002) Sister chromatid

cohesion ensures that the two sister chromatid kinetochores attach stably to microtubules

emanating from opposite spindle poles during bi-orientation (Tanaka et al., 2000) Proper

sister kinetochores to microtubules Amphitelic attachment in turn generates tension on sister kinetochores as the poleward force exerted on chromosomes by microtubules is counteracted by cohesion between sister chromatids The separation of sister chromatids at the metaphase to anaphase transition is then triggered by proteolytic cleavage of the cohesion subunit Scc1

(Uhlmann et al., 2000) In the absence of proper bipolar attachment or sufficient tension across

sister kinetochores, the spindle checkpoint is activated to inhibit onset of anaphase The spindle checkpoint is also activated when cells encounter errors in spindle assembly (reviewed in

Musacchio and Hardwick, 2002; Tan et al., 2005) Another checkpoint called the spindle

positioning checkpoint, delays mitotic exit and cytokinesis when spindles are misaligned with respect to the mother-bud axis (Lew and Burke, 2003)

1.3 Protein Degradation in cell cycle control

As described earlier, activation of the cyclin-dependent kinase controls the ordered sequence of cell division events Likewise, destruction mechanisms also play an important role in cell cycle progression and are ideally suited to generate unidirectionality in the cell cycle since they are irreversible processes Proteolysis is particularly critical for sister chromatid separation at the metaphase-to-anaphase transition and for cyclin destruction to mark the end of mitosis

1.3.1 Ubiquitin-dependent proteolysis

The ubiquitin system drives the cell cycle by timely destruction of numerous regulatory proteins Aaron Ciechanover, Avram Hershko and Irwin Rose were granted the 2004 Nobel Prize for their pioneering work that led to the discovery that proteins targeted for destruction are attached to multiple copies of a conserved 76-residue ubiquitin protein in a process known as ubiquitylation (Hochstrasser, 1996; Hershko and Ciechanover, 1998) Polyubiquitinated proteins are then recognized and degraded by the proteasome Ubiquitylation of a substrate requires the activity of a

ATP-dependent activation of the C-terminus of ubiquitin, forming a covalent thioester bond between the terminal glycine of ubiquitin (Gly76) with a cysteine in the active-site of E1 In the second step, an ubiquitin-conjugating enzyme (E2) transiently receives the activated ubiquitin from E1, again on a conserved cysteine residue Finally, a ubiquitin ligase (E3) transfers ubiquitin from E2 to a lysine side-chain on the target protein E3 ubiquitin ligases, either alone or in conjunction with E2, are responsible for target recognition Polyubiquitination can occur by isopeptide bond formation between the C-terminal Gly76 of one ubiquitin and any of the seven

lysine residues of the next ubiquitin in the polyubiquitin chain (Peng et al., 2003) Polyubiquitin

chains linked through Lys48 are the most common and recruit proteins to the proteasome for degradation In addition, chains linked through Lys29 or Lys63 have also been reported in yeast cells but these do not target proteins for proteolysis (reviewed in Hershko and Ciechanover, 1998) Biochemical studies suggest that a polyubiquitin chain containing at least four ubiquitin molecules (tetraubiquitin) is required to label protein substrates to render them polyubiquitinated for

recognition and subsequent degradation by the 26S proteasome (Thrower et al., 2000)

In the cell cycle context, two multi-subunit E3 ligases, namely SCF (Skp1, Cullin, F box protein complex) and APC (anaphase promoting complex), play central roles in G1/S, G2/M and M/G1 transitions (also see Fig 1)

1.3.2 SCF

The multisubunit E3 ligase SCF obtained its name from its three components – Skp1, Cullin and the F-box (Cardozo and Pagano, 2004) and is required at the G1/S transition There are several different F-box proteins which recruit specific target proteins to the SCF Hence, the F-box protein determines substrate specificity of the SCF A well-known F-box protein in budding yeast is Cdc4

Sic1, a Cdc28-Clb inhibitor that prevents entry into S-phase (Schwob et al., 1994), is

phosphorylated before its ubiquitylation by the SCF complex that includes Cdc34 (E2 ubiquitin–

1997) Cdc28-Cln1/Cln2/Cln3 kinase complexes phosphorylate Sic1, targeting it for destruction

by the SCF, allowing the activation of the S-phase cyclins Clb5 and Clb6 Thus, SCF plays an

important role in triggering entry into S phase

1.3.3 APC

The anaphase-promoting complex (APC) is a multimeric E3 ubiquitin ligase consisting of at least

a dozen subunits (Zachariae and Nasmyth, 1999; Peters, 2006) The APC ubiquitylates substrates with the help of three co-factors, the ubiquitin-activating enzyme (E1), a ubiquitin-conjugating enzyme (E2) and an activator protein It has been proposed that polyubiquitination reactions occur

in two ways (Carroll and Morgan, 2002): either via a distributive manner, whereby the addition of each new ubiquitin involves a separate interaction between the substrate and the APC or in a processive manner, where repeated cycles of ubiquitination occur while the substrate remains bound to the APC It has been shown that the kinetics of polyubiquitination is tightly correlated

with relative timing of degradation of APC substrates in the cell cycle (Rape et al., 2006) The

more processive the polyubiquitination of an APC substrate, the earlier it is degraded The more distributive the polyubiquitination of an APC substrate, the more likely it may be converted to the basal state via deubiquitination carried out by deubiquitinating enzymes upon dissociation from APC, thus taking a longer time to achieve the tetraubiquitinated state

1.3.3.1 Selective substrate recognition by APC

How are substrates recognized by the APC? As mentioned earlier, substrate ubiquitylation

by the APC requires activator proteins, either Cdc20 or Cdh1 (also known as Hct1) in mitotically dividing cells (reviewed in Peters, 2006) Cdc20 binds and activates the APC at the metaphase-to-anaphase transition, whereas Cdh1 maintains APC activity during late mitosis through G1

(Schwab et al., 1997; Visintin et al., 1997, Fang et al., 1998; Zachariae et al., 1998) Cdc20 and

an IR tail (consensus IR at extreme C-terminus) These mediate APC binding at its tetratrico

peptide repeat (TPR) subunits (Schwab et al., 2001; Passmore et al., 2003; Vodermaier et al.,

2003) Both activators also contain a C-terminal WD40-repeat domain that is predicted to fold into

a seven-bladed β-propeller structure Regulated association of Cdc20 and Cdh1 with the APC can

confer different substrate specificities on the APC (Schwab et al., 1997; Visintin et al., 1997) For

example, Cdh1 binds to Clb2 but is unable to bind to another APC substrate Pds1 However, Pds1

is able to associate with Cdc20 Ubiquitylation of APC substrates depends on specific degradation motifs The best characterized are the destruction box (D-box) with the consensus amino acid sequence RXXLXXXN and the KEN-box with the consensus amino acid sequence KEN (Glotzer

et al., 1991; Pfelger and Kirschner, 2000) The D-box is recognized by both APCCdc20 and APCCdh1

while the KEN box is preferentially recognized by APCCdh1 Both Cdc20 and Cdh1 bind substrates directly, and these interactions depend on an intact D-box and/or KEN-box motif (Burton and

Solomon, 2001; Hilioti et al., 2001; Pfleger et al., 2001) Additional motifs have also been identified, including the A-box (Littlepage and Ruderman, 2002) and the GxEN motif (Castro et

al., 2003) The substrate specificity of Cdh1 led to the “substrate-adaptor” hypothesis where

activator proteins bind substrates through conserved motifs, for example the D and KEN boxes just mentioned, and recruit them to the APC for ubiquitylation Consistent with this, it has been

shown that substrates can only associate with APC in the presence of Cdc20 or Cdh1 (Passmore et

al., 2003) However, this hypothesis has been challenged by findings from Tim Hunt’s group that

APC from mitotic Xenopus egg extracts can bind to the D-box peptide in the complete absence of Cdc20 and Cdh1 (Yamano et al., 2004) Prompted by their findings, the authors proposed that the

activator protein could act in a substoichiometric manner to induce conformational changes in APC to provide a form capable of binding to substrate

Recent findings (Burton et al., 2005; Kraft et al., 2005) demonstrate the direct recognition

of D and KEN boxes (by means of synthetic peptides) by APC activators Kraft et al (2005)

interaction is required for processive substrate ubiquitylation They also observed that the activator interaction stimulates the activator-substrate interaction since more D-box peptide

APC-crosslinked to Cdc20 with APC present than when it was depleted Burton et al (2005), on the

other hand, concluded that substrate-activator interaction stimulates the activator-APC interaction They showed that an intact D-box domain within a substrate was required to stimulate the association between the Cdh1-substrate complex and the APC Together, these findings suggest that APC-activator and activator-substrate interactions can influence one another through D-box dependent mechanisms Indeed, this is consistent with a model whereby both APC and activators contribute recognition sites for substrates, providing the first direct evidence for a ternary complex

between the APC, the activator and the substrate (Passmore et al., 2005) The role of APC in

substrate recognition is more mysterious Surprisingly, not only are the interactions between

substrates and activator D-box dependent, but also those between substrates and APC (Yamano et

al., 2004; Eytan et al., 2006) The identity of the APC subunit that mediate these interactions is

unknown but the Doc1 protein is an attractive candidate since APC lacking Doc1 exhibits a processivity defect (Carroll and Morgan, 2002) Furthermore, Doc1 has been shown to play a

critical role in D-box recognition by APC (Passmore et al., 2003; Carroll et al., 2005) Clearly,

structural analysis of Cdh1 or Cdc20 bound to substrates containing one or both degradation motifs will be required to address the various hypotheses regarding interactions that lead to substrate ubiquitylation The capability of the APC to selectively recognize its substrates at the correct time is essential for several key events in mitosis: the initiation of anaphase, exit from mitosis and the preparation for the next round of DNA replication Importantly, Cdc20 and Cdh1 activate APC in a sequential manner in mitosis, with Cdc20 functioning at anaphase onset and Cdh1 at the end of mitosis through G1

1.3.3.2 APC-Cdc20 at metaphase-to-anaphase transition

The most well known function of APC, one that lent “APC” its name, is its role at the onset of anaphase As described earlier, after S phase, duplicated sister chromatids are held together by a cohesin complex which prevents premature segregation of sister chromatids (Nasmyth, 2002) Just prior to anaphase, when sister chromatids are properly bioriented on the spindle, APC is activated

by Cdc20 (Peters, 2006) At the onset of anaphase, APCCdc20 targets securin (Pds1) for mediated destruction, which bind to the separase Esp1 prior to anaphase and causes its inhibition APC-mediated destruction of Pds1 liberates separase (Esp1) allowing it to cleave the cohesin subunit Scc1 and dissolve sister-chromatid cohesion to permit partitioning of chromosomes by the

ubiquitin-mitotic spindle (Uhlmann et al., 2000; Nasmyth, 2002) Cdc20 itself is also regulated during the

cell cycle Cdc20 is degraded during exit from mitosis in an APCCdh1-dependent manner It is therefore resynthesized during S phase and G2 when APCCdh1 is inactive (Fang et al., 1998; Prinz

et al., 1998; Shirayama et al., 1998; Kramer et al., 2000) In addition, the binding of Cdc20 to

APC is stimulated by cyclin-dependent kinases through phosphorylation of APC core subunits (Rudner and Murray, 2000) In contrast, Cdh1 is inhibited by Cdc28

1.3.3.3 APC-Cdh1 at the end of mitosis

APCCdh1 is fully active during G1 phase to ensure continued degradation of mitotic cyclins during G1 Phosphorylation of Cdh1 during subsequent S phase prevents its association with the

APC (Jaspersen et al., 1999; Zachariae et al., 1998) Thus, APCCdh1 is inactivated upon phosphorylation by early-expressed Cdc28/Clb3, Clb4, Clb5 kinase complexes during S phase, allowing cells to progressively accumulate mitotic cyclin Clb2 critical for the onset of mitosis

(Huang et al., 2001; Yeong et al., 2001) Cdh1 phosphorylation is also required for export of Cdh1 into the cytoplasm contributing to its inactivation (Jaquenoud et al., 2002) It has recently been

shown that Cdh1 forms a complex with three other proteins Acm1, Bmh1 and Bmh2 during S

conserved 14-3-3 family comprising of phosphoserine-binding proteins (van Heusden et al.,

1995) Acm1 acts as a negative regulator of APCCdh1 when complexed with Bmh1 and Bmh2 and its binding is independent of the status of Cdh1 phosphorylation

Cdh1 phosphorylation persists till the end of mitosis when the protein phosphatase Cdc14 reverses Cdc28-mediated phosphorylation and promotes Cdh1 binding to APC, thereby activating APCCdh1 (Visintin et al., 1998; Jaspersen et al., 1999) The activity of Cdc14 itself is under the

control of the Mitotic Exit Network (MEN) pathway proteins which include Tem1, Net1, polo kinase Cdc5 and Cdc15 (Zachariae, 1999) Cdc14 not only activates APCCdh1, but also dephosphorylates the Cdc28 inhibitor Sic1 (causing its stabilization and further inhibition of the Cdc28-Clb complexes) and the Sic1 transcription factor Swi5 (thereby enhancing Sic1 production), and thus brings about the inactivation of the mitotic kinase by three complementary mechanisms (Peters, 2002)

APCCdh1 is required for degradation of the mitotic cyclin Clb2, resulting in inactivation of

the mitotic kinase (Schwab et al., 1997, Visintin et al., 1997) In addition, it has been shown that

APCCdc20 plays a role in partial proteolysis of Clb2 (Baumer et al., 2000; Yeong et al., 2000;

Wasch and Cross, 2002) At the metaphase-to-anaphase transition, Clb2 is partially degraded in a D-box dependent manner by APCCdc20, partially reducing the activity of mitotic kinase Cdc28-Clb2 It was estimated that approximately 50% of Clb2 normally present in metaphase cells gets degraded in a Cdc20-dependent manner by the time cells reach telophase Interestingly, a second fraction of Clb2 persists during anaphase that was protected from APCCdc20-mediated destruction This partial Cdc28 inactivation results in the initiation of Cdc14-mediated activation of Cdh1 The second phase of proteolysis begins at the end of mitosis and is dependent on APCCdh1 At this stage, APCCdh1 also degrades other proteins such as the microtubule-associated proteins Cin8 and

Ase1 and polo kinase Cdc5 (Juang et al., 1997; Charles et al., 1998; Hildebrandt and Hoyt, 2001; Castro et al., 2003)

Thus far we have described the events involving APC complex that lead to chromosome segregation and cyclin destruction In the next section, we encounter the elegant macromolecular structure that captures, aligns, and eventually separates the sister chromatids during anaphase, the bipolar mitotic spindle

1.4 The bipolar mitotic spindle

As mentioned earlier, partitioning of chromosomes equally to progeny cells is the central function

of mitosis This process must be extremely precise To achieve this high-fidelity of chromosome segregation, eukaryotic cells have developed a marvellous self-organizing molecular apparatus called the mitotic spindle

The spindle is one of the most extensively studied cellular structure in cell cycle research and its history can be traced as far back as 1882 in the detailed illustrations of the famous German anatomist Walther Flemming, who had coined the term “mitosis” from the Greek word “thread”

(mitos to reflect the shape of the paired sister chromatids; Flemming, 1882) Detailed analysis of

spindle morphology and progression through mitosis was limited to fixed preparations but Flemming’s drawings clearly showed the shape of the mitotic spindles along with evidence for filamentous organization In the 1950s, development of polarization microscopy with camera attachments allowed time-lapse analysis and confirmed that spindles are made up of filaments running parallel to chromosome movement (reviewed in Rieder and Khodjakov, 2003) The introduction of electron microscopy (EM) allowed the first observation of the microtubule

cytoskeleton of S cerevisiae, described as a “fiber apparatus” present in the nucleus in 1966

(Robinow and Marak, 1966) Early EM studies also showed that the spindle was made up of parallel, straw-shaped structures termed “microtubules” (reviewed in Rieder and Khodjakov, 2003) At this point, molecular cell biologists were interested in solving how microtubules (MTs) reorganized from a radial array during interphase into a spindle during mitosis However, this

breakthrough came in 1968 when a heterodimer of α- and β-tubulin was identified as the protein

subunit of MTs (Weisenberg et al., 1968) In 1972, polymerization of brain tubulin in vitro was described, which marked the beginnings of the study of MT dynamics (Weisenberg, 1972)

Concurrent studies revealed that these MTs were initiated by the spindle poles (the organizing center (MTOC) of a spindle) during the polymerization process, and that each MT is polarized, with a slow-growing minus-end embedded in the MTOC and a fast-growing plus-end that extends distally to it

microtubule-Tubulin biochemistry soon led to the concept of polymerization dynamics driven by GTP hydrolysis known as “dynamic instability” (Mitchison and Kirschner, 1984), where microtubule ends undergo stochastic changes from a polymerizing to depolymerizing state β-tubulin subunits are exposed at the faster polymerizing plus-end and α-tubulin subunits at the slower-polymerizing minus-end of the MT Hydrolysis of the GTP-bound β-tubulin occurs soon after addition of the tubulin heterodimer The bulk of the MT polymer thus consists of GDP-associated tubulin, while the newly added tubulin heterodimers are still bound to GTP This is called the GTP cap The energy from GTP hydrolysis fuels dynamic instability, the rapid oscillation between extended periods of growth and shrinkage at the MT end When the rate of addition of GTP-tubulin is greater than the rate of GTP hydrolysis, the GTP cap is maintained and the MT will continue to grow rapidly When GDP-tubulin is less tightly associated with the MT and rapidly dissociates, there is a switch from rapid growth to rapid shrinkage The transition from growth to shrinkage is termed “catastrophe” while the transition from shrinkage back to growth is termed “rescue”

The cell uses the inherent dynamic properties of MTs for assembly and function of the mitotic spindle In a typical metaphase spindle, the two spindle poles are in a “face-to-face” configuration, separated by a set of overlapping microtubules emanating from each spindle pole towards the other (pole-to-pole microtubules) A second set of MTs termed astral MTs radiate from each spindle pole towards the cell cortex A third set, kinetochore MTs, emanate from the

achieved by a “search-and-capture” mechanism to find kinetochores, first proposed by Kirschner and Mitchison (1986), based on dynamic instability where MT plus ends extend and retract till a kinetochore captures a MT end, forming a tension-ridden bipolar spindle As mentioned previously, it is now known that each sister kinetochore is linked to one pole that is directly facing

it while the other is attached to the opposite pole, thus setting up what is known as bi-orientation

(Tanaka et al., 2000) By the time cells reach metaphase, all the chromosomes are bi-oriented and

aligned at the equatorial plate

Control of microtubule dynamics alone is not sufficient to organize MTs Antagonistic forces exerted by plus and minus-end directed kinesins also generate a bipolar spindle The BimC family of plus-end kinesin motor proteins (members show strong sequence similarity in an amino-

terminal motor domain) like Eg5 in mammalian cells and Cin8 and Kip1 in S cerevisiae, crosslink

and slide apart antiparallel microtubules to form bipolar spindles by binding and bundling MTs

(Kashina et al., 1997) This plus-end directed force pushing spindle poles apart appears to be counterbalanced by a force pulling them inwardly The kinesin-related Kar3 in S cerevisiae is a minus-end directed motor with its motor domain at the carboxyl-terminus (Endow et al., 1994)

When both Cin8 and Kip1 activities were eliminated, preformed spindles collapsed with separated poles rapidly moving back together This collapse effect was suppressed by reducing the activity

of the minus-end directed motor Kar3 (Saunders et al., 1997) Deficiency in Cin8/Kip1 activity or overexpression of KAR3 led to shorter spindles Overproduction of Cin8, on the other hand,

increased spindle length Therefore, Cin8 and Kip1 required for an outwardly-directed force and Kar3 required for an inward force, produce balancing forces that determine proper spindle length

Due to the dynamic nature of microtubules and the presence of microtubule-associated proteins (MAPs), the spindle poles tend to move away from each other, causing the sister chromatids to be pulled in the opposite direction In a metaphase-spindle, this tendency is counteracted by cohesins that tether the sister-chromatids together Such opposing forces within

All spindles are bipolar, but the structure of the spindle pole differs in different organisms The spindle pole is pivotal to the biogenesis of the mitotic spindle In mammalian cells, the spindle poles are termed centrosomes (Fig 3, left panel) In yeast, the poles are called spindle pole bodies (SPBs) (Fig.3, right panel) Higher plants and the oocytes of many vertebrates do not contain centrioles and depend on self-organizing properties of microtubules and MAPs to generate the spindle poles In many animals, assembly of the first spindle in the fertilized egg is dependent on the MTOC (in the form of a basal body centriole) contributed by the sperm since sometime during oogenesis, oocyte centrosome degenerates Hence, during fertilization, the sperm contributes not

only DNA but also MTOC for construction of a spindle (Simerly et al., 1995) The incoming

centriole then recruits maternal components that constitute the pericentriolar material (PCM) (Holy and Schatten, 1991) During a mitotic division cycle, a cell inherits only one centrosome from its progenitor and has to build a complex spindle structure starting from this centrosome In each cell cycle, centrosomes are duplicated and separated precisely to serve as two poles of the mitotic spindle, each acting as the MTOC for nuclear and astral microtubules, and this is generally referred to as the centrosome cycle

1.5 The centrosome cycle

The centrosome was first discovered by Walther Flemming in 1875, and subsequently named by Theodor Boveri in 1887 (Flemming, 1875; Boveri; 1887) Named for its location near the cell centre, Boveri described the centrosome as a “pair of centrioles surrounded by differentiated cytoplasm” In animal cells, each centrosome is composed of a pair of centrioles and the surrounding dense fibrillar mass known as the pericentriolar material (Fig 3, left panel) The centrioles in the pair are referred to as mother and daughter centrioles where the mother centriole can be distinguished by the presence of distal and sub-distal appendages The centrioles themselves are cylindrical structures, each built from nine microtubule-triplets (doublet or singlets

that their long axes are perpendicular to each other (also known as orthogonal arrangement) (Fig

3, left panel) Incidentally, centrioles are very similar in structure to basal bodies, the organelle located at the base of cilia which are themselves conserved structures that have diverse and essential roles in development (Johnson and Rosenbaum, 1992)

In order to build two centrosomes from one, the pair of centrioles undergoes a duplication cycle (Fig 4, top panel) During G1 phase of the cell cycle, the centrioles lose their orthogonal arrangement in that the daughter centriole separates slightly from the mother centriole but remains tethered by a flexible connection As cells enter S phase, a precursor centriole (procentriole) appears perpendicular to the proximal end of each of the parental centriole and continues to elongate through S phase to attain the same length as the parental ones At G2/M, the pairs of centrioles disconnect completely, along with the divided pericentriolar material, to form two separate (mother and daughter) centrosomes This is followed by maturation of the immature daughter centriole by full acquisition of appendages that contain protein such as cenexin, ninein,

CEP110 or ε-tubulin (Lange and Gull, 1995; Ou et al., 2002; Chang et al., 2003) It is interesting

to note that like DNA replication, the duplication of centrioles is a semi-conservative process in that each pair of centrioles consists of one old and one new member That centriole duplication is semi-conservative but nucleus-independent has given support to the idea that centrosomes contain their own nucleic acids, like mitochondria and chloroplasts do Indeed, presence of centrosome-

associated RNA in surf clam oocyte has been reported (Alliegro et al., 2006); however the role it

might play in centriole duplication is far from clear

Centrosomal defects were originally proposed to lead to aneuploidy and cancer in 1914 by Boveri (reviewed in Brinkley and Goepfert, 1998) He proposed that cancer cells arise by the aberrant replication and activity of centrosomes leading to increased centrosome number that cause aneuploidy during mitosis Indeed, many cancer cells show a high incidence of centrosome amplification; for example, nearly 80% of invasive breast tumour cells have amplified centrosome

separase has been implicated in a licensing mechanism that restricts centrosome duplication so that

it occurs only once per cell cycle In addition, excess polo kinase Plk4 activity has been shown to trigger extra procentriole formation, suggesting that regulation of Plk4 is involved in preventing the occurrence of multiple centrosomes (reviewed in Nigg, 2007)

While centrosome duplication has been fairly well-documented, the separation of duplicated centrosomes is poorly understood In the case of vertebrate cells, the mother and daughter centrioles (and the duplicated centrosomes) appear to be tethered since in centrosome preparations, centrioles remain paired Electron microscopy studies show an electron dense material between mother and daughter centrioles in isolated centrosomes However, the molecular nature of this linker (intercentriolar linkage) remains unclear (It should be noted that the linkage between mother and daughter centrioles is what after duplication becomes the linkage between the

two pairs of centrioles i.e centrosomes) While in vivo existence of a linker is not yet

unequivocally proven, some of the proteins that regulate centrosome separation (or centriole disjunction) have been identified Nek2A, a NIMA related kinase (Fry, 2002), forms a complex with the catalytic subunit of phosphatase PP1 and C-Nap1, a large coiled-coil protein (280 kDa) thought to provide a docking site for the linker Nek2A is capable of phosphorylating both itself and C-Nap1, whereas PP1 can dephosphorylate both Nek2A and C-Nap1 It has been reported that inhibition of C-Nap1 activity or over-expression of Nek2A leads to premature separation of mother and daughter centrioles Moreover, while co-expression of PP1 and Nek2A can prevent centriole separation, inhibition of PP1 activity promotes the separation These observations suggest that mutual regulation of the C-Nap1, PP1 and Nek2A is important for the control of centrosome cohesion Recently, a C-Nap1 interacting protein, called Rootletin, has been shown to

participate in centriolar cohesion (Bahe et al., 2005) It is also phosphorylated by Nek2 and is

removed from centrosomes at the onset of mitosis As in the case of C-Nap1, inhibition of Rootletin results in centrosome splitting

Motor proteins are known to play an essential role in bipolar spindle assembly (Kashina et

al., 1997) Eg5, a plus-end directed homotetrameric motor protein belonging to BimC (Kinesin-5)

family is of particular importance in this context because its properties and functions have been

studied in some detail (Cochran et al., 2004; Kwok et al., 2004) It can both push apart pole microtubules as well as recruit microtubules into bundles (Kapitein et al., 2005) Inhibition

pole-to-of Eg5 by a small molecule inhibitor, monastrol, prevents centrosome separation and leads to monopolar spindles suggesting that its functions are essential for biogenesis of a bipolar spindle

(Mayer et al., 1999) The association of Eg5 with spindle apparatus is regulated via phosphorylation by Cdk1 in human cells (Blangy et al., 1995), suggesting a direct involvement of

Cdk1 in centrosome separation In addition to Cdk1, polo and aurora-A kinases are also

implicated in centrosome separation In both Xenopus and Drosophila, loss of aurora-A kinase activity results in a failure to separate centrosomes (Glover et al., 1995; Giet et al., 1999) In both

Drosophila and human cells, centrosomes fail to separate in the absence of polo kinase (Sunkel

and Glover, 1988; Lane and Nigg, 1996) At least in vertebrate cells, it is possible that polo and aurora-A kinases may contribute to complete removal of C-Nap1 from centrosomes (Faragher and Fry, 2003)

1.6 The spindle pole body (SPB) cycle

Some of the paradigms of centrosome duplication have been based on the exquisite work done on spindle pole bodies (SPB), the centrosome-equivalent in yeast (Adams and Kilmartin, 1999) At the structural level, SPBs do not bear much resemblance to centrosomes In budding yeast

Saccharomyces cerevisiae, the SPB is a cylindrical structure embedded in the nuclear envelope

and comprises three distinct layers (Fig 3, right panel) The outer plaque faces the cytoplasm, contains proteins such as Tub4, Spc98, Spc97 and Spc72, and nucleates astral microtubules (Jaspersen and Winey, 2004) The inner plaque faces the nucleoplasm and extends the nuclear microtubules that consist of both pole-pole microtubules and the kinetochore microtubules The

central plaque is embedded in the nuclear envelope and anchors both outer and inner plaques Proteins such as Spc29, Spc42 and Cmd1 have been localized to this layer The central plaque also contains an electron-dense structure known as the ‘half-bridge’ The half-bridge is an important appendage because its distal end mediates the assembly of a new SPB A number of proteins including Cdc31 (centrin homologue), Kar1 and Sfi1 (similar to human Sfi1) and Mps3 have been localized to this structure

The duplication of SPBs, like centrosomes, is an intricate affair At the time of its birth the yeast daughter inherits from its mother one SPB bearing a half-bridge During G1, the precursor for a new SPB, called satellite, is assembled on the cytoplasmic side of the half-bridge’s distal tip (Fig 4, bottom panel) This consists of Spc42, Spc29, Nud1 and Cnm67, with Spc42

being the core crystal of the SPB (Bullitt et al., 1997) As cells traverse START, the satellite

expands to form a duplication plaque whose structure resembles the cytoplasmic side of the mature SPB This step requires Spc42 self-assembly where trimerization of Spc42 dimers is

required to form the plaque (Bullitt et al., 1997) This is facilitated by phosphorylation of Spc42

by Cdc28 and Mps1 kinases (Donaldson and Kilmartin, 1996; Jaspersen et al., 2004b) During

this time, the half-bridge (about 90 nm in length) also extends under the duplication plaque and forms a complete bridge (about 150 nm in length) by fusion of its cytoplasmic and nuclear fronts Kilmartin and colleagues recently showed that Cdc31 binds Sfi1 on multiple conserved repeats and this association forms a filament that is precisely the length of the half-bridge with the amino-

terminus of Sfi1 located at the SPB core and the carboxyl-terminus at the distal tip (Li et al.,

2006) They proposed a model in which half-bridge elongation occurs by addition of the Cdc31 complex to the carboxyl-terminus of Sfi1-Cdc31 already present in the existing half-bridge After a complete bridge is formed, the bridge then retracts to some extent allowing insertion of the duplication plaque into the nuclear membrane and the assembly of the nucleoplasmic side Finally, the old and the new SPBs lie in a side-by-side configuration, interconnected by the bridge