Fission yeast Ase1PRC1 is required for the G2-microtubule damage response.

Schizosaccharomyces pombe delays entry into mitosis following G2 microtubule damage. This pathway is dependent on Rad26ATRIP, the regulatory subunit of the Rad26ATRIP/Rad3ATR DNA damage response (DDR) complex. However, this G2 microtubule damage response pathway acts independently of the G2 DNA damage checkpoint pathway. To identify other proteins in this G2 microtubule damage pathway, we previously screened a cDNA overexpression library for genes that rescued the sensitivity of rad26Δ cells to the microtubule poison thiabendazole. A partial cDNA fragment encoding only the C-terminal regulatory region of the microtubule bundling protein Ase1 PRC1 was isolated. This fragment lacks the Ase1PRC1 dimerization and microtubule binding domains and retains the conserved C-terminal unstructured regulatory region. Here, we report that ase1Δ cells fail to delay entry into mitosis following G2 microtubule damage. Microscopy revealed that Rad26ATRIP foci localized alongside Ase1PRC1 filaments, although we suggest that this is related to microtubule-dependent double strand break mobility that facilitates homologous recombination events. Indeed, we report that the DNA repair protein Rad52 co-localizes with Rad26ATRIP at these foci, and that localization of Rad26ATRIP to these foci depends on a Rad26ATRIP N-terminal region containing a checkpoint recruitment domain. To our knowledge, this is the first report implicating Ase1PRC1 in regulation of the G2/M transition.

INTRODUCTION

A diverse array of intrinsic and extrinsic insults threatens genomic integrity, and those that lead to double strand DNA breaks (DSBs) are particularly toxic. In humans, repair of DSBs occurs primarily via non-homologous end-joining (NHEJ) and homologous dependent repair (HDR) [1, 2]. NHEJ and HDR are stimulated by a trinity of synergistic phosphatidylinositol 3-kinase-related kinases (PIKKs): DNA-dependent protein kinase catalytic subunit (DNA-PKcs); ataxia-telangiectasia mutated (ATM); and ataxia-telangiectasia and Rad3-related (ATR) that initiate a number of other DDR responses as well, including cell cycle checkpoint control over G1/S, G2/M and S-phase [3-6]. ATM and ATR have also been implicated in pathways that respond to interphase microtubule (MT) damage. For example, MT toxins cause G1 and G2 cell cycle arrest in cultured cells by mechanisms that target downstream components of ATM and ATR signaling, including CHK1, CDC25C and CDC2 [7-11]. Exactly why eukaryotic cells keep the integrity of interphase microtubules under surveillance, and how they detect compromised interphase MTs, are areas for exploration.

Yeast model systems provide a framework for dissecting human DDR pathways [12]. In fission yeast, the conserved Rad26ATRIP/Rad3ATR complex directs DDRs similar to those in humans. The Rad26ATRIP/Rad3ATR complex also participates in a pathway that delays entry into mitosis following G2-MT damage [13]. Rad26ATRIP plays a cytoplasmic role during this response, as disruption of its C-terminal nuclear export signal prevents both cytoplasmic accumulation of Rad26ATRIP and the G2/M delay that follows treatment with MT poisons. Notably, this nuclear export allele of rad26ATRIP does not compromise the DDR. Therefore, the checkpoint responses to DNA and G2-MT damage are genetically separate pathways dependent upon Rad26ATRIP. Also using fission yeast, Balestra and Jimenez [14] demonstrated that G2-MTdamage delays entry into mitosis through stabilization of Wee1WEE1, a negative regulator of mitotic cyclin-dependent kinase (CDK) activity and target of the DDR [15-17]. We therefore suspect that the Rad26ATRIP/Rad3ATR complex initiates stabilization of Wee1WEE1 when G2-MTs are damaged. Here, we present evidence that the microtubule bundling protein Ase1PRC1 is another component of this G2-MT damage response in fission yeast.

MATERIALS AND METHODS

PCR cloning of and into GFP-tagging expression vectors: PCR amplification of ase1 and rad26 cDNAs from a cDNA library (gift of M. Yamamoto) were performed using high-fidelity Phusion Master Mix (MO531S; New England Biolabs). Primer sequences appear in Table 1. Amplicons were cloned using the pJET blunt-end cloning kit (K1231; Thermo Scientific), transformed into DH5-Alpha Mix & Go! competent cells (T3007; ZYMO Research), and recombinant vectors were purified using a midi prep kit (12143; QIAGEN). cDNA inserts were then directionally cloned into the thiamine-repressible, amino-terminal pRep41 eGFP and LEU2 vector [18] as NdeI-BamHI fragments and transformed into yeast using the yeast transformation protocol described by Bähler et al [19].

Table 1

Primers used in this study

Primer Set Name	Sequence	Resulting protein Size, aa
Full Ase1 gene	Forward: P-5’ CAT ATG CAA ACA GTA ATG ATG 3’Reverse: P-5’ GGA TCC TTA AAA GCC TTC TTC 3’	731
Ase1-C-term	Forward: P-5’ CAT ATG GCT ATG ACG AGT CCA 3’Reverse: P-5’ GGA TCC TTA AAA GCC TTC TTC 3’	218
1Full Rad26 gene	Forward: P-5’ CAT ATG ATG ATG GCT GAT GAA AGT 3’Reverse: P-5’ GGA TCC CTA AAA ATT AGT GTA CAA 3’	615
2Rad26Δ1–103aa	Forward: P-5’ CAT ATG TCC GAA GCT AAT AAT GCC 3’Reverse: P-5’ GGA TCC CTA AAA ATT AGT GTA CAA 3’	512
3Rad26Δ310–615aa	Forward: P-5’ AT ATG ATG ATG GCT GAT GAA AGT 3’Reverse: P-5’ ACA TCA TCG TCG ATT TAG GAT CC 3’	309
4Rad26Δ109–615aa	Forward: P-5’ AT ATG ATG ATG GCT GAT GAA AGT 3’Reverse: P-5’ GGA TCC CTA TGA GGC TTG TGA GTT TAC GG 3’	108
5Rad26Δ1–25aa–Δ147–615aa	Forward: P-5’ CAT ATG GAA TTA GAG CAA CAA GCT CAA ACC 3’Reverse: P-5’ GGA TCC CTA CTC ATG AAA TAG GGA TTT CGT 3’	121
6Rad26Δ1–40aa–Δ147–615aa	Forward: P-5’ CAT ATG GTA GTT GTA CCG AGT GAA AAG CAA 3’Reverse: P-5’ GGA TCC CTA CTC ATG AAA TAG GGA TTT CGT 3’	106
7Rad26Δ113–123aa	Forward: P-5’ CAA CGG AAA TTA GAG GAG CTT AAA AAA GAA 3’Reverse: P-5’ TAA CGA TTC TAG GGC GGC ATT ATT AGC TTC 3’	604

Physiological methods: Spindle-pole-body (SPB) separation was monitored in different strains containing cdc25.22 and the SPB marker cut12-egfp [20]. Cells of each strain were cultured in YE5S liquid medium to optical density (OD) 0.3 at 30°C, then shifted to 37°C for 3 hours 5 minutes before 16 mg/ml Carbendzim (MBC) was added. Cultures were maintained in the presence of MBC at 37°C for another 25 minutes before downshifting to 20°C and releasing cells from the cdc25.22 block. The percentage of cells containing two Cut12-EGFP foci was determined every 20 minutes following this downshift. Three trials were performed, and 200 cells were scored at every 20-minute time point. Strains used in this study are outlined in Table 2.

Table 2

Fission yeast strains used in this study

Strain	Genotype	Origin
TE236	leu1-32 ura 4-d18 h−	Kostrub et al., (1998)
TE257	rad26::ura4+ ade6-704 leu1-32 ura4-D18 h−	Al-Khodairy et al., (1994)
TW1197	rad26-GFP::kanR leu1-32 ura 4-d18 h−	Baschal et al., (2006)
MCI728	z:adh15:mcherry-atb2:natMX6 leu1-32 ura-D18 h+	gift of Meredith Betterton
FY20056	ase1-GFP::kanR leu1 h−	Yeast Resource Center, yeast.nig.ac.jp
ST754	ase1D::kanR leu1-32 ura4-D18 h+	gift of Thibault Courtheoux
TW1341	ase1D::kanR cdc25.22 cut12-gfp	this study
TW1345	ase1D::kanR rad26::ura+ cdc25.22 cut12-GFP	this study
TW1300	rad26::ura4+ cdc25.22 cut12-egfp:ura4+	Herring et al., (2010)
MKSP765	rad26-mcherry::kanR leu1-32 ura4-D18 h+	gift of Megan King
MKSP2074	rad52-mcherry::kanr ura4-D18 leu1-32 his3-D1	gift of Megan King
FY20720	sad1-mcherry::KanR leu1 his2 h+	Yeast Resource Center, yeast.nig.ac.jp

Microscopy: To visualize EGFP fusion proteins, 1 ml aliquots from cultures grown in liquid YE5S to OD 0.3 were centrifuged and resuspended in cold methanol for one minute, washed twice in 100 ml SlowFade Component C (SlowFade Antifade Kit, Molecular Probes) and air dried on coverglass. Once dried, 4.5 μl SlowFade Component A was dropped on the coverglass, which was then placed onto a slide. Achieving yeast monolayers that adhered tightly to the coverslips was crucial to observing fluorescence signals. To help ensure that such layers formed, the coverglass was soaked in acetone for one day, scrubbed with dishwashing soap, wiped with 70% ethanol and air dried prior to use. Images were acquired using a Leica DM5000 equipped with a Leica DFC350FX R2 digital camera and Leica FW4000 software.

RESULTS AND DISCUSSION

Previously, we screened a cDNA library for genes that, when overexpressed (OE), permit rad26Δ cells to grow on medium containing thiabendazole (TBZ) [21]. Of 10,000 transformants, four cDNAs representing Rad24, SPCC70.01, DASH complex subunit Dam1, and the C-terminal 218 amino acids of Ase1 (aa 513–731; Fig. 1A) were identified. Here, we investigate the role of Ase1 in the rad26-dependent G2-MT damage response more closely, since both ase1+ and rad26+ are required for proper cell morphology and minichromosome stability [13, 22]. Ase1 is a conserved MT-bundling protein containing a dimerization domain, a MT-binding domain, and an unstructured C-term [22-26]. Below, we confirm that OE of the unstructured Ase1-C-term rescues the growth of rad26Δ cells on media containing a MT toxin, and report that Ase1 is required to delay mitotic entry following insult to G2-MTs.

Figure 1

Ase1 is required to delay SPB separation following G 2 -microtubule damage. A. Map of the dimerization domain, MT-binding domain and C-terminal region of Ase1. B. Overexpression of the C terminal (513–731) amino acids of Ase1 rescues the MBC sensitivity of rad26Δ cells. Thiamine (thia) represses transcription of vector inserts. In this experiment, a medium strength nmt thiamine-repressible promoter was fused to full length Ase1-GFP, Ase1-C-term-GFP, or empty GFP vector, and transformed into rad26Δ cells under repressive conditions (EMM + thiamine). Two transformants representing each condition were streaked to repressive conditions (EMM + thiamine) for two days before replica plating to media with inducing conditions and a pink vital dye. C. Map of the genotypic distributions on the EMM-thia replica plates shown in (B). D. Overexpressed full length Ase1-GFP assembled into bright linear structures within nonviable cells. Overexpressed Ase1-C-term was diffusely localized throughout viable cells. Bar = 5 μm. E. Ase1 is required to delay SPB separation following MBC treatment. Strains containing the cdc25.22 and cut12-gfp alleles were arrested in G2 before releasing into medium with or without 16 mg/ml MBC. At each time point, the percentage of cells containing two Cut12-GFP foci was calculated from 150 cells. WT refers to rad26+ ase1+ cells with cdc25.22 cut12-gfp alleles in the background

We observed nmt-promoter driven OE of full length Ase1-GFP to be toxic, presumably due to MT-hyper-bundling (Fig. 1B, C; [22]). Within these nonviable cells, we observed that OE Ase1-GFP assembled into bright linear structures (Fig. 1D), consistent with MT-hyper-bundling at overlapping microtubules [22]. Overexpression of the 218aa unstructured C-term of Ase1 was not toxic, and rescued rad26Δ growth on carbendazim (MBC; Fig. 1B, C). This C-term fragment localized diffusely throughout cells (Fig. 1D) and did not appear to integrate within the yeast MT cytoskeleton. Therefore, OE Ase1-C-term may rescue rad26Δ growth on MBC in a regulatory manner, as opposed to a structural one.

Next, we tested if ase1+ is required for the rad26+-dependent checkpoint response to G2-MT damage (Fig. 1E). The temperature-sensitive cdc25.22 allele was used to reversibly block cells in G2, and the spindle-pole-body (SPB) marker Cut12-EGFP was used to monitor SPB separation and mitotic entry [20,27,28]. After wildtype cells were released from the G2 block into nutrient rich YE5S medium containing MBC, SPB separation was prevented for greater than three hours (Fig. 1E). This result is consistent with those of Akera, Sato and Yamamoto [29], who observed that SPBs fail to separate during MBC treatment. Following release of rad26Δ, ase1Δ, and rad26Δ ase1Δ cells from the G2 block into MBC media, precocious SPB separation began within the first hour in all three strains (Fig. 1E). That the kinetics of SPB separation were similar among the three strains suggests that Ase1 and Rad26 operate in the same pathway. We also tested if OE Ase1-C-term restored the G2-delay to rad26Δ cells during treatment with MT-toxin. However, the nmt-promoter requires growth in nutrient-deficient, synthetic media to drive OE, and we observed that cells grown under these nutrient-limiting conditions fail to delay G2/M following treatment with MT poisons (data not shown). This may be due to activation of a stress MAP kinase response that overrides this G2-MT damage response in nutrient-depleted conditions [30]. In summary, we conclude that Ase1 participates in the Rad26-dependent G2-MT damage response pathway.

Ase1 bundles MTs, and its loss results in an increased number of interphase MT bundles, as well as short mitotic spindles that are prone to collapse [22,31]. In agreement, we observed that the number of MT bundles in ase1Δ cells was significantly greater than the numbers in WT and rad26Δ cells, which were similar (Fig. 2A, B). In addition, the length of mitotic spindles decorated with either Atb2-GFP or Ase1-GFP was similar in WT and rad26Δ cells (Fig. 2C, D). Therefore, rad26+ does not appear to share ase1+-dependent functions related to MT bundling or spindle stability.

Figure 2

Rad26 does not influence MT structure, although it does tangentially localize with Ase1 and MTs. All strains were grown at 30ºC in liquid YE5S to optical density 0.3 and fixed with cold methanol. The number of MT- bundles per cell (A, B) and spindle length (C, D) were normal in rad26Δ cells. (E, F and G) Rad26 tangentially localizes with Ase1 and MTs, but not the SPB. Strains expressing mcherry-Atb2 (MTs), Rad26-GFP, Rad26-mcherry, Ase1-GFP and Sad1-mcherry were used in these experiments (false colors are sometimes shown); DAPI (blue) was used to visualize nuclei in (D). All bars =5 μm

Next, we tested if Rad26ATRIP and Ase1PRC1 colocalize with one another. Cycling fission yeast have approximately six faint nuclear Rad26 foci and, occasionally, one or two somewhat brighter foci [32]. We observed that these bright Rad26 foci tangentially localize with Ase1 (Fig. 2E) and MTs (Fig. 2F), but not to the spindle pole body (Fig. 2G). However, we believe that this association between Rad26 and Ase1 is related to the DDR, and not the G2-MT damage response, because these bright Rad26 foci co-localized with Rad52 (Fig. 3A). Rad52 is a DNA recombination-promoting protein that displaces RPA-coated ssDNA to facilitate HDR [33]. At sites of DNA damage, Rad52 foci co-localize with DNA checkpoint proteins [34,35] and comigrate with the nuclear envelope-spanning linker of nucleoskeleton and cytoskeleton (LINC) complex [36]. Comigration of Rad52-foci with LINC is a microtubule- and Rad3-dependent process that directs HDR of damaged DNA. Therefore, we suggest that Ase1 tangential localization with Rad26/Rad52-foci is likely playing a role in this microtubule-dependent DDR pathway as opposed to the G2-MT damage response.

Figure 3

Rad26 and Rad52 colocalize. A. Colocalization of Rad26 and Rad52 was observed in cells grown at 30ºC in liquid YE5S to optical density 0.3 and fixed with cold methanol. B. Overexpressed Rad26-GFP colocalizes with Rad52-mcherry. Cells were grown in promoter de-repressing liquid conditions (EMM − thiamine) for 20 hours and fixed with methanol. Bar=5 μm

To identify the region of Rad26ATRIP that mediates co-localization with Rad52, we cloned the rad26+ cDNA in front of GFP in a medium strength thiamine-repressible expression vector. Rad26ATRIP is 615 amino acids long and contains a coiled-coil motif in the N-terminus ([37]; Fig. 4A). Overexpression of the full length Rad26-GFP fusion protein resulted in production of bright nuclear foci that also co-localized with Rad52 (Fig. 3B). A large C-terminal truncation of Rad26 ATRIP did not affect localization to foci (Fig. 4A, construct 3), demonstrating that the N-terminal half of the protein is sufficient for focus formation. A region near the N-terminal border of the coiled-coil was critical for focus formation, since neither 1–108aa (Fig. 4A, construct 4) or 104-615aa (Fig. 4A, construct 2) localized to foci. This led to identification of an N-terminal region within 26-146aa that directed focus formation (Fig. 4A, construct 5). The intact coiled-coil region of Rad26 was not required for focus formation, since an 11 amino acid deletion within the coiled-coil permitted focus formation (Fig. 4A, construct 7).

Figure 4

Rad26 contains an N-terminal Checkpoint Recruitment Domain Required for Focus Formation. A. Summary of Rad26-GFP truncations. CC = coiled-coil region; ∙∙∙ = DNA removed; ✓ = positive for foci. Bar = 5 μm. B. An area that resembles the Checkpoint Recruitment Domain (CRD) of ATRIP [39] lies within the N-terminal region of Rad26 that directs focus formation (25–146aa).

Within ATRIP, an N-terminal region between 43–108aa contains a major RPA-ssDNA interacting domain [38]. An acidic checkpoint recruitment domain (CRD) exists within this N-terminal region (54–68aa) that mediates ATRIP focus formation and interaction with a basic cleft of RPA subunit RPA70N [39,40]. This CRD is conserved in budding yeast Ddc2ATRIP where it also mediates focus formation and interaction with RPA. While destruction of this motif impairs ATRIP localization to nuclear foci, it does not completely impair ATR signaling. Here, we identify a region resembling the CRD that exists within the 25–146aa region of Rad26 that directed focus formation (Fig. 4A, construct 5; Fig. 4B). Experiments to determine if it is required for Rad3ATR DNA checkpoint signaling in S. pombe are planned.

To our knowledge, this is the first report implicating Ase1PRC1 in regulation of the G2/M transition. During G2, Ase1PRC1 bundles and stabilizes interphase MTs [31]. During the transition from G2 to M, Ase1PRC1 localizes to the duplicated SPBs as they separate, yet Ase1PRC1 is not required for this event in wildtype cells [41-43]. Later, Ase1PRC1 is required to stabilize the pre-anaphase spindle and then the spindle midzone, where it influences the future site of division [22,23,31,44-47]. Ase1PRC1 plays two regulatory roles following entry into mitosis. Prior to metaphase, Cdc2CDK1 phosphorylates target sites in the Ase1-C-term to prevent release of the kinesin-like motor Klp9 from Ase1PRC1 [48]. After metaphase, the Clp1/Flp1CDC14 phosphatase removes these inhibitory phosphate groups to permit release of Klp9 from Ase1PRC1 in order to initiate anaphase B microtubule sliding. Ase1PRC1 is also required for the Clp1/Flp1CDC14- and septation initiation network (SIN)-dependent cytokinesis checkpoint that inhibits successive nuclear divisions following perturbation of actomyosin ring components [22,49-52]. In summary, Ase1PRC1 has well established structural and regulatory roles during later mitotic events, and this report here appears to be the first to identify an earlier G2/M role for Ase1PRC1.

Conflict of Interest

We declare there is no conflict of interest to report.