A study on premature segregation of unreplicated chromosomes 4

-4.2.3 Ectopic expression of Sic1 and Cdh1 prevent premature spindle elongation in Cdc6 depleted cells In Chapter 3, we provided evidence that premature spindle elongation in cells lack

Chapter4 Regulation of Spindle Elongation by

Cdc34 4.1 Background

In the previous chapter, we have explored the various reasons responsible for premature nuclear division in Cdc6 depleted cells It has long been thought that precocious segregation of chromosomes is due to premature onset of mitosis However, recent evidence suggests that premature chromosome segregation can occur without the onset of mitosis, APC activation, cohesion cleavage or biorientation of kinetochores

The DNA replication checkpoint has been reported to thwart untimely chromosome segregation not by inhibiting mitotic entry but by directly regulating spindle dynamics and by preventing replication fork collapse to allow duplication of centromeric DNA and hence kinetochore bi-orientation (Krishnan et al 2004) To be more precise, the critical effectors of DNA replication checkpoint, Mec1 and Rad53, (orthologues of human ATM/ATR- and Chk2-like kinases, respectively) prevent precocious chromosome segregation by suppressing the accumulation of spindle elongation effectors such as Cin8 and Stu2, thus precluding premature induction of spindle elongation during early S phase (Krishnan et al 2004) (Krishnan et al 2005)

A recent study in yeast suggests that concerted action by two prominent kinases Cdk1 and polo (Cdc5) are required to fully inactivate Cdh1, an activator of the E3 ubiquitin ligase APC (anaphase promoting complex) (Crasta et al 2008) APCCdh1 is essential for proteolytic destruction of microtubule associated proteins such as Cin8, Kip1 and Ase1 Therefore, Cdh1 inactivation is required to halt the

destruction of these proteins to facilitate spindle assembly and spindle elongation These findings helped uncover an additional mechanism by which DNA damage checkpoint prevents premature chromosome segregation (Zhang et al 2009) It was shown that activation of the DNA damage checkpoint leads to inactivation of Cdc5 polo kinase via Rad53-mediated phosphorylation (Zhang et al 2009) Thus, inactivation of both Cdk1 and Cdc5 by the checkpoint prevents Cdh1 inactivation, which in turn continues to mediate the destruction of spindle elongation-inducing proteins Cin8, Kip1 and Ase1 Thus these studies envisaged that the DNA damage checkpoint, in addition to suppressing cohesin cleavage, maintains APCCdh1 in an active state to restrain spindle extension until the damaged chromosomes are repaired Hence, both DNA replication and DNA damage checkpoints can prevent premature chromosome segregation by restricting the accumulation of microtubule associated proteins such as Cin8, Kip1, Ase1 and Stu2 The exact mechanism as to how DNA replication checkpoint promotes destabilization of the microtubule associated proteins remains unclear The activated DNA damage checkpoint leads to the perturbation of Cdh1/Cdk1/polo/MAPs control circuit; Cdh1 is maintained in an active state and spindle is restrained from prematurely elongating (Crasta et al 2008) Since cells lacking Cdc6 also encountered premature spindle elongation resulting from an accumulation of MAPs, it is possible that Cdc6 (putative G1-M checkpoint component) is able to directly regulate spindle elongation by targeting the Cdh1-Cdk1/polo/MAPs control circuit This notion would be consistent with Cdc6’s role in the inactivation of Cdk1 This chapter explores the relationship between Cdc6 and spindle dynamics

chromosome segregation is a characteristic associated specifically with cdc6 mutant

or it is a common characteristic of cells that can undergo START but are unable to initiate S phase To test this notion, we monitored spindle behaviour in cells lacking Cdc7 and Cdc45 Cdc7 is a serine/threonine kinase essential for DNA replication and requires Dbf4 for its activity (hence the term Ddk for Dbf-dependent kinase) (Bousset

et al 1998) (Jares et al 2000) Cdc45 is required for the initiation and elongation step

of DNA replication (Zou et al 1997) The CDC45 gene was shown to genetically

interact with components of replication factors such as MCM and ORC We

constructed cdc7Δ GAL-CDC7 (US5582) and cdc45Δ GAL-CDC45 (US5585) strains

for this study Both strains were first arrested in metaphase by growth in galactose medium containing nocodazole Cells were then transferred to YEPD for 30 minutes

to allow depletion of Cdc7 and Cdc45 before they were released into YEPD medium containing α-factor to synchronize them in G1 phase Once the cultures were uniformly arrested in G1, α-factor was removed and cells were released into YEPD

medium at 25˚C and samples taken every 20 minutes to score the spindle lengths As shown in Figure 15, cells lacking Cdc7 arrested in G1 with 1N DNA but proceeded to elongate the spindles from 80 minutes onwards with spindle length exceeding 4µm Similarly, cells lacking Cdc45 also arrested in G1 with a large bud and unreplicated chromosomes but proceeded to extend the spindle from 80 minutes onwards with spindle length exceeding 4µm (Figure 16) Thus, cells defective in Cdc7 or Cdc45 traverse START, assemble a bud and fail to initiate DNA replication, but prematurely

segregate the unreplicated chromosomes despite the presence of a functional CDC6

One interpretation of these results is that Cdc6, Cdc7 and Cdc45 are all involved in preventing untimely spindle elongation and segregation of unreplicated chromosomes Alternatively, it is not inconceivable that premature segregation is not specifically due

to the lack of Cdc6, Cdc7 or Cdc45 functions but is a common characteristic of cells that have traversed START but fail to initiate DNA replication In other words, the premature partitioning of the chromosomes is regulated by some unknown

mechanism and the mutations in genes such as CDC6, CDC7 and CDC45 only allow

manifestation of this underlying regulation because of their common inability to initiate DNA replication

Glu NOC Glua

Glu 60min Glu 120min Glu 180min

Nomarski DAPI Anti-tubulin

Time (mins) 0

>4 um

Figure 15 Premature segregation of unreplicated chromosomes in cells lacking

Cdc7

We constructed cdc7 7 strain for this study The strain was first arrested

in metaphase by growth in galactose medium containing nocodazole Cells were then

transferred to YEPD for 30 minutes to allow depletion of Cdc7 before they were

released into YEPD medium containing -factor to synchronize them in G1 phase

Once the cultures were uniformly arrested in G1, -factor was removed and cells were

released into YEPD medium at 25 C and samples taken every 20 minutes to score the

spindle lengths and for FACS analysis

short spindle (0-2 mm)

medium spindle (2-4 mm)

long spindle (>4 mm)

Spindle Length

Nomarski DAPI Anti-tubulin

0 20

>4 um

Glu NOC Glua

Glu 60min Glu 120min Glu 180min

Figure 16 Premature segregation of unreplicated chromosomes in cells

lack-ing Cdc45

strain was first arrested in metaphase by growth in galactose medium containing nocodazole Cells were then transferred to YEPD for 30 min-

utes to allow depletion of Cdc45 before they were released into YEPD medium

and for FACS analysis.

2N

short spindle (0-2 mm)

medium spindle (2-4 mm)

long spindle (>4 mm)

Time (mins)

Spindle Length

4.2.2 Depletion of Cdc6 in cdc34-1 cells fails to promote spindle

assembly or spindle elongation

To directly address the possibility of a role for CDC6 in spindle regulation, we utilize cdc34-1 cells which traverse START and duplicate their centrosomes but can neither

initiate S phase nor assemble a short spindle because the inter SPB bridge remains

unbroken If Cdc6 is involved in regulating spindle biogenesis and dynamics,

cdc34-1 cells lacking Cdc6 is expected to promote spindle assembly or spindle elongation For this experiment, we constructed cdc34-1 (US6005) and cdc34-1 cdc6Δ MET- CDC6 (US7012) strains expressing GFP-tagged spindle pole body component Spc42

to ascertain spindle length in both strains To ensure Cdc6 is degraded completely, both strains were arrested in nocodazole supplemented with methionine to repress

CDC6 transcription These cells were then released into methionine medium (+Met)

at non-permissive temperature of 36˚C Samples from both cdc34-1 and cdc34-1 cdc6Δ MET-CDC6 cells were collected at 180 minutes and analyzed by immunofluorescence microscopy As shown in Figure 17, 100% of both cdc34-1 and cdc34-1 cdc6Δ MET-CDC6 strains exhibit one Spc42-GFP dot indicating that the

duplicated SPBs, have not separated and no spindle was assembled In addition, immunofluorescence staining of 180 minutes sample showed no sign of spindle formation or spindle elongation These results imply that Cdc6 does not influence spindle biogenesis significantly

Nomarski DAPI Spc42GFP

Nomarski DAPI Anti-tubulin

-4.2.3 Ectopic expression of Sic1 and Cdh1 prevent premature

spindle elongation in Cdc6 depleted cells

In Chapter 3, we provided evidence that premature spindle elongation in cells lacking Cdc6 was due to the accumulation of microtubule associated proteins such as Cin8, Kip1 and Ase1 Since it has been shown previously that APCCdh1 is responsible for the ubiquitination and proteasome-mediated degradation of Cin8, Kip1 and Ase1, we therefore considered the possibility that Cdh1 inactivation may be relevant to Ase1, Cin8 and Kip1 stability in Cdc6 depleted G1 arrested cells If this is true, then over-expression of Cdh1 would be expected to promote degradation of spindle elongation effectors such as Cin8 in Cdc6 depleted cells To test this, we arrested two separate

cultures of cdc6Δ MET-CDC6 GAL-HA3-CDH1 CIN8-HA3 (US7022) in +Met+glucose medium (to repress transcription of CDC6) containing nocodazole

Subsequently, both cultures were released from metaphase arrest into +Met medium containing α-factor to ensure that the cells were arrested in G1 One culture was then released into +Met+glucose medium at room temperature to inhibit over-expression

of Cdh1 Another culture was released into medium containing Raffinose+Galactose

to drive over-expression of GAL-HA3-CDH1 As shown in Figure 18A, Cdh1

expression was apparent from 80 minutes after the release As Cdh1 over-expression peaked from 120 minutes onwards, Cin8 was concurrently degraded from 120 minutes The unstable Cin8 was not present in sufficient amounts to allow Cdc6 deficient cells to elongate their spindles, resulting in the phenotype observed: cells with short spindles (Figure 18A) In contrast, in the control culture where Cdh1 was not over-expressed, Cin8 remained stable The results of this experiment imply that Cdh1 may be inactive in Cdc6 depleted cells, leading to the accumulation of proteins such as Cin8 and, thus, premature spindle elongation

It is known that Sic1 is responsible for inhibition of Cdk1/Clb5, 6 kinases in G1 phase (prior to S phase onset) (Barberis et al 2005) Moreover, active Cdk1 is also responsible for inactivating Cdh1 via phosphorylation at multiple sites (Crasta et

al 2008) Therefore if Cdk1 activity is inhibited, Cdh1 will remain active To further verify that premature spindle elongation in Cdc6 depleted cells is a result of Cdh1 inactivation (with accumulation of microtubule associated proteins such as Cin8), we expressed non-degradable version of Sic1 in Cdc6 depleted cells We first arrested

cdc6Δ MET-CDC6 GAL-ndSIC1 (US7013) strain in nocodazole containing medium supplemented with methionine to repress CDC6 transcription Cells were then

released into α-factor containing medium to resynchronize cells in G1 The cells were subsequently released into medium containing Raffinose and Galactose to facilitate over-expression of non-degradable Sic1 Samples were taken at 160 minutes to monitor the presence of long spindles Almost 100% of the cells failed to elongate their spindles and arrested in G1 with large buds and short spindles (Figure 18B) This is similar to the previous experiment where Cdh1 over-expression in cells lacking Cdc6 led to arrest with short spindles Clearly, Cdh1 inactivation is mandatory for premature spindle elongation in Cdc6 deficient cells This also suggests that the Sic1 degradation step must be tightly regulated since untimely degradation of Sic1 promotes Cdk1-mediated inactivation of Cdh1, accumulation of Cin8 and Ase1 and hence premature spindle elongation If SCF-mediated degradation

of Sic1 is an essential step in determining the fate of spindle in Cdc6 depleted cells, then there is a strong possibility that SCF-component Cdc34 is important in the regulation of spindle elongation

Nomarski DAPI Anti-tubulin

Nomarski DAPI Anti-tubulin

Figure 18 Ectopic expression of Sic1 and Cdh1 prevent premature spindle elongation in

Cdc6 depleted cells

in +Met+glucose medium (to repress transcription of ) containing nocodazole

Subse-quently, both cultures were released from metaphase arrest into +Met medium containing

+Met+glucose medium at room temperature to inhibit over-expression of Cdh1 Another culture

was released into medium containing raffinose+galactose to drive over-expression of 3

Samples were collected for immunofluorescence and Western Blot analysis.

B strain was arrested in nocodazole containing medium

supplemented with methionine to repress transcription Cells were then released into

medium containing raffinose and galactose to facilitate over-expression of non-degradable Sic1

Samples were taken at 160 minutes for immunofluorescence analysis.

A

B

4.2.4 cdc34 and cdc34 cdc6 mutant cells can assemble short bipolar

spindles in the absence of Cdh1 or Sic1 but fail to elongate them

Once cells have traversed START, Cdc34 function becomes critical for G1/S transition due to its involvement in Sic1 degradation Cells deficient in Cdc34 function are unable to initiate S phase and arrest in G1 phase (post START) with a bud and duplicated SPBs but are unable to assemble a spindle in accordance with the regulatory scheme shown in figure below In contrast, cells lacking Cdc6 are also unable to undergo S phase and arrest in G1 phase (post START) but with a prematurely extended spindle This raises the possibility of Cdc34’s involvement in regulating spindle elongation since Cdc34 is functional in cells lacking Cdc6 Previous experimental data had shown that ectopic expression of Sic1 and Cdh1 are sufficient to prevent premature spindle elongation in cells lacking Cdc6 This is suggestive of a role for Cdc34 in Sic1 degradation With Sic1 degraded, Cdk1 can inactivate Cdh1; cells prematurely elongate the spindle mirroring that seen in Cdc6 deficient cells We tested this notion by depleting Sic1 or Cdh1 independently in cells deficient in Cdc34 and investigate if it is sufficient to permit spindle assembly and spindle elongation

We treated cdc34-1 cdh1Δ (US7015), cdc34-1 cdc6Δ MET-CDC6 cdh1Δ (US6016) and cdc6Δ MET-CDC6 sic1Δ (US5690) cells with nocodazole containing

+Met medium to deplete Cdc6 Subsequently, these cells were released into medium containing methionine at 36˚C to inactivate Cdc34 As expected these cells arrested with multiple buds in G1 phase in the absence of DNA replication Immunofluorescence staining of the spindles revealed that these cells are capable of assembling short spindles but are unable to extend them (Figure 19) This suggests that both Sic1 degradation and Cdh1 inactivation are critical steps in promoting

assembly of short spindles, however, neither event alone can promote spindle elongation in Cdc6 deficient cells These experiments also strengthen the notion that spindle elongation seen in Cdc6 deficient cells is not due to the lack of Cdc6 function per se The results also imply that Cdc34 (E2 enzyme), acting in synergy with SCF and Cdc4 (F-box protein) to degrade Sic1, not only regulates G1-S transition but also facilitates Cdh1 inactivation which leads to the stabilization of microtubule associated proteins and sets up the context for the assembly of a short spindle

Nomarski DAPI Anti-tubulin

Nomarski DAPI Anti-tubulin

Nomarski DAPI Anti-tubulin

36˚C 180min

36˚C 180min

36˚C 180min

Figure 19 cdc34 and cdc34 cdc6 mutant cells can assemble short bipolar

spin-dles in the absence of Cdh1 or Sic1 but fail to elongate them.

cells were arrested in +Met medium containing nocodazole to deplete Cdc6 inactivate Cdc34 Samples were collected at 180 minutes for immunofluorescence analysis

Sub-4.2.5 Sic1 degradation promotes Cdh1 inactivation and short spindle

assembly

Before returning to the relationship between Cdc34, microtubule associated proteins and spindle elongation in Cdc6 deficient cells, we take a short detour to explore the connection between Sic1 and Cdh1 in the context of the regulatory scheme outlined in Figure 20 below, since it is important for the regulation of spindle dynamics in Cdc6 mutant

The question we intended to address is whether Sic inactivation indeed leads to mediated inactivation of Cdh1 The active Cdk1 has been shown to phosphorylate Cdh1 on S16, S42, T157 and T173 residues creating polo-box binding domains recognized by Cdc5 (yeast Polo-like kinase) This leads to the recruitment of Cdc5 to Cdh1 and additional phosphorylation on S125 and S259 residues (Crasta et al 2008) Thus, multiple phosphorylations by these two kinases are required to inactivate Cdh1 completely To confirm the notion that Sic1 degradation promotes Cdk1-mediated

Cdk1-inactivation of Cdh1, we monitored the phosphorylation status of Cdh1 in cdc34-1 (US5677) and cdc34-1 sic1Δ (US5722) cells expressing HA3-CDH1 from its native

locus Cells were synchronized in G1 phase using α-factor and then released into YEPD medium at 36˚C As expected, Cdh1 immunoprecipitated from cells deficient

in Cdc34 showed no hyper-phosphorylation (i.e Sic 1 active, Cdk1 inactive, Cdh1 active), whereas it is highly phosphorylated in Cdc34 deficient cells lacking Sic1 (Figure 21) Treatment with calf intestinal alkaline phosphatase (CIP) causes higher molecular weight, slower mobility bands to disappear suggesting that their lower

mobility is indeed due to phosphorylation (Figure 21) As mentioned earlier, cdc34-1

cells arrest in G1 with large buds, unreplicated DNA and no bipolar spindle

However cdc34-1 sic1Δ cells arrested in G2 with large buds and short spindles (data

not shown) These observations support the notion that Cdh1 inactivation and short spindle assembly is tightly connected with Sic1 degradation mediated by Cdc34 In

the following section we make use of these observations and use SIC1 or CDH1 deletion to allow cdc34-1 mutant cells to assemble a short spindle and explore the

cellular requirements for spindle elongation

Cdh1 in cdc34-1 and cells expressing CDH1-HA 3 from

collected for immunoprecipitation and Western Blot analysis

4.2.6 Ectopic expression of microtubule associated proteins induces

spindle elongation in Cdc34 deficient cells devoid of Cdh1

The experiments described in the preceding sections suggest that premature spindle elongation in Cdc6 deficient cells is a consequence of the loss of regulation of microtubule associated proteins such as Ase1, Cin8 and Kip1 If so, then overexpression of Ase1, Cin8 and Kip1 would be expected to induce spindle

elongation in cdc34-1 cdh1Δ and cdc34-1 cdc6Δ MET-CDC6 cdh1Δ cells, given that they are able to assemble short spindles (Figure 19) To test this possibility, cdc34-1 cdc6Δ MET-CDC6 cdh1Δ cells carrying GAL-CIN8 (US7024) or GAL-KIP1 (US7026) and cdc34-1 cdh1Δ cells carrying GAL-ASE1 (US7028) were first

synchronized in G2-M by nocodazole treatment in raffinose medium supplemented

with methionine to fully repress CDC6 transcription These cells were then

subjected to second synchronization in the subsequent G1 phase by α-factor treatment in medium containing raffinose and methionine to ensure complete depletion of Cdc6 The cells were pre-induced with galactose for 1 hour to express Cin8, Kip1 or Ase1 and then released into medium containing raffinose and

galactose at 34˚C As a control, cdc34-1 (US1688) and cdc34-1 cdc6Δ MET-CDC6 cdh1Δ (US7016) cells without GAL constructs were treated identically As expected, cdc34-1 cells arrest with no spindle whereas cdc34-1 cdh1Δ and cdc34-1 cdc6Δ MET-CDC6 cdh1Δ cells arrest with short spindles, (Figure 19 and 22) confirming the

previous findings that Cdh1 inactivation promotes stability of microtubule associated proteins leading to short spindle assembly in Cdc34 deficient cells However, over-

expression of Ase1 in cdc34-1 cdh1Δ and Cin8 or Kip1 separately in cdc34-1 cdc6Δ MET-CDC6 cdh1Δ cells allowed ~60% of the cells to arrest with long spindles,

indicating that over-expression of these proteins could induce spindle elongation

(Figure 22 and 23) These findings suggest that microtubule associated proteins are

the limiting factors preventing spindle elongation in cdc34-1 cdh1Δ, cdc34-1 cdc6Δ MET-CDC6 cdh1Δ and cdc34-1 cdc6Δ MET-CDC6 sic1Δ cells (Figure 19)

Additionally, since only 60% of the cells are able to elongate their spindles, these results also suggest that the over-expressed microtubule associated proteins remain unstable in Cdc34 deficient cells despite the absence of active Cdh1 If so, then Cdc34 may have an additional novel role in regulating stability of microtubule associated proteins

To test this notion, we also treated cdc34-1 cells carrying GAL- ASE1 (US7027), GAL-CIN8 (US7023) and GAL-KIP1 (US7025) as described above We

found that overexpression of microtubule associated proteins can only induce approximately 30% of the cells to assemble short spindles but cannot elicit spindle elongation (Figure 22 and 23) These observations are consistent with our hypothesis

that microtubule associated proteins become more unstable in cdc34-1 cells when

Cdh1 is fully activated

Nomarski DAPI Anti-tubulin

Figure 22 Ectopic expression of microtubule associated proteins induces spindle elongation in Cdc34 deficient cells devoid of Cdh1.

28% short spindles 72% no spindle

58% long spindles 42% short spindles

60% long spindles 40% short spindles

Figure 23 Ectopic expression of microtubule associated proteins induces spindle elongation in Cdc34 deficient cells devoid of Cdh1

cdc34-1

-4.2.7 Cdh1 resistant microtubule associated proteins cannot induce

complete spindle elongation in cdc34-1 and cdc34-1 cdh1Δ cells

Since over-expression of Ase1, Cin8 and Kip1 cannot induce spindle elongation in

100% of the cdc34-1 cdh1Δ double mutant cells, it is possible that these proteins are unstable even in the absence of Cdh1 We test this possibility by treating cdc34-1 and cdc34-1 cdh1Δ cells carrying non-degradable (nd) Cin8 (US7018 and US7020) or

Ase1 (US7019 and US7021) with α-factor to synchronize them in G1 phase Subsequently, they were released into YEPD medium at 36˚C to inactivate Cdc34 function and samples were taken at 240 minutes for spindle length analysis As shown in Figure 24, expression of ndCin8 can only induce short spindle assembly in

16% of cdc34-1 cells and long spindle in 25% of cdc34-1 cdh1Δ cells Similarly, expression of ndAse1 also induces short spindle assembly in 15% of cdc34-1 cells and long spindle in 20% of cdc34-1 cdh1Δ cells (Figure 25) This strongly suggests

that Cdh1 resistant microtubule associated proteins such as ndAse1 and ndCin8 are still unstable in Cdc34 deficient cells, since only very small proportion of the cells can elongate their spindles Hence these results are consistent with our proposed novel role involving Cdc34 in regulating the stability of microtubule associated proteins in G1 phase

Nomarski DAPI Anti-tubulin

Nomarski DAPI Anti-tubulin

36˚C 240min

25% long spindles 75% short spindles

16% short spindles 84% no spindle

36˚C 240min

Figure 24 Cdh1 resistant microtubule associated proteins cannot induce

complete spindle elongation in cdc34-1 and cells

We treated cdc34-1 and cells carrying non-degradable