Shugoshin promotes efficient activation of spindle assembly checkpoint and timely spindle disassembly

Shugoshin proteins are evolutionary conserved across eukaryotes with some species-specific cellular functions ensuring the fidelity of chromosome segregation. Shugoshin being present at various subcellular locales, acts as an adaptor to mediate various protein-protein interactions in a spatio-temporal manner. Here, we characterize shugoshin (Sgo1) in the human fungal pathogen, Candida albicans. Interestingly, we discover a novel in vivo localization of Sgo1 along the length of the mitotic spindle. Further, Sgo1 performs a hitherto unknown function of facilitating timely disassembly of spindle in this organism. We observe that Sgo1 retains its centromeric localization and performs its conserved functions that include regulating the centromeric condensin localization, chromosome passenger complex (CPC) maintenance and sister chromatid biorientation. We identify novel roles of Sgo1 as a spindle assembly checkpoint (SAC) component with functions in maintaining the SAC proteins, Mad2 and Bub1, at the kinetochores, in response to faulty kinetochore-microtubule attachments. These findings provide an excellent evidence of the functional rewiring of shugoshin in maintaining genomic stability.


Introduction 1
The bipolar attachment of sister chromatids to opposite spindle poles is mediated by dynamic 2 interactions between kinetochores and microtubules during mitosis. Such biorientation keeps 3 the sister chromatids under tension. by counter-balancing opposing pulling forces. The force 4 towards the spindle poles generated by microtubule depolymerization is opposed by the  In mitosis, the function of shugoshin as a sensor of tension between the sister 11 chromatids is believed to be driven by the enrichment of this protein at the centromeres that chromosome segregation appear to be conserved across the organisms with some species 5 specific modulation, we wished to examine its roles in C. albicans for the first time to 6 investigate whether and how aneuploidy can be generated due to alteration of its function. 7 We observed that unlike several organisms but like S. cerevisiae, C. albicans harbors a single www.candidagenome.org) and of the SGO genes from various other eukaryotes. Using 23 conventional CLUSTALW, GBLOCKS (Castresana 2000) and BLAST programs, we 24 independently identified the conserved N-terminal domain (residues 26-103) and the SGO 1 motif (residues 326-364) in ORF 19.3550 to assign the same as Sgo1 in C. albicans (CaSgo1) 2 (Suppl. Fig. S1A). The alignment of Sgo proteins from various organisms depicts the relative 3 positions of these conserved regions (Suppl. Fig. S1B). We constructed a phylogenetic tree of 4 shugoshin, using CLUSTAL Omega and Interactive Tree of Life (iTOL) (Ciccarelli et al 5 2006, Letunic et al 2006, to determine the evolutionary position of CaSgo1 with respect to 6 other ascomycetes species (Suppl. Fig. S1C). The tree suggests that CaSgo1 is evolutionarily 7 closer to ScSgo1 as there is a recent common ancestor which becomes evident when we 8 compared the sequences of CaSgo1, ScSgo1 and SpSgo1 (Suppl. Fig. S1D). In several  (Fig. 1A). Since, many organisms including the 22 unicellular fission yeast S. pombe and multicellular eukaryotes like mammals, contain two 23 functionally distinct paralogs of shugoshin, namely SGO1 and SGO2, we searched for an 24 8 SGO2 ortholog in CGD but failed to identify one. We thus report that like S. cerevisiae, C. 1 albicans has a single shugoshin homolog SGO1. 2 3 Sgo1 is not essential for viability in C. albicans 4 To test essentiality of Sgo1 in C. albicans during vegetative growth, we constructed a 5 conditional mutant of SGO1 using gene replacement and conditional expression (GRACE) 6 strategy, wherein, one copy of SGO1 was deleted and the remaining copy was placed under a 7 regulatable promoter (P MET3 ). MET3 promoter is repressed in the presence of cysteine and 8 methionine (CM) and is de-repressed in their absence. Whereas the cells with depletion of an 9 essential protein CaSth1 (Prasad et al 2019) did not grow, we observed that Sgo1-depleted 10 cells (sgo1/P MET3 SGO1) did not lose viability (Suppl. Fig. S2A). This suggests that Sgo1, like 11 in S. cerevisiae, is non-essential for mitotic growth in C. albicans. We then constructed a 12 strain deleted for both the copies of SGO1 (sgo1Δ/Δ) and verified the deletions by Southern 13 hybridization (Suppl. Fig. S2B). We did not observe any phenotype of sgo1Δ/Δ cells under 14 normal growth condition where the cells grew like wild type without any defect in 15 chromosome missegregation or cell cycle arrest (Suppl. Fig. S2C). We observed a lower 16 survival rate of sgo1Δ/Δ cells, when treated with anti-microtubule drug nocodazole (Suppl.  To assess the functions of Sgo1 in the cell cycle of C. albicans we looked at the sub-cellular 3 localization of Sgo1. For this, we fused one copy of SGO1 with two tandem copies of 4 Candida-optimized GFP (Gerami-Nejad et al 2001) at the carboxy (C)-terminus in the wild 5 type strain SN148 to generate SGO1/SGO1-2GFP and on further marking of α tubulin with 6 RFP to generate SGO1/SGO1-2GFP::TUB1/TUB1-RFP strains. We also constructed 7 SGO1/SGO1-2GFP::TUB4/TUB4-MCHERRY strain expressing Sgo1-2GFP along with 8 Tub4-mCherry that marks the spindle poles. We verified that Sgo1-2GFP was functional as 9 sgo1Δ/SGO1-2GFP::TUB4/TUB4-MCHERRY strain with Sgo1-2GFP as sole source of Sgo1, 10 did not show any sensitivity to nocodazole as compared to increased sensitivity of sgo1Δ/Δ 11 mutant strain (sgo1Δ/sgo1Δ::TUB4/TUB4-MCHERRY) (Suppl. Fig. S2E).

12
Live cell analysis revealed that Sgo1-2GFP demonstrated dynamic localization 13 patterns throughout the cell cycle except in telophase (Fig. 1B). In the unbudded G1/S cells, 14 Sgo1-2GFP appeared as a dot-like signal close to MTOC indicative of kinetochore 15 localization, as the C. albicans kinetochores earlier showed to remain clustered near MTOC 16 (Thakur andSanyal 2011, 2012). In C. albicans metaphase cells, Sgo1-2GFP could be seen 17 as one dot or two dots near duplicated SPBs. Interestingly, with the advent of anaphase, Sgo1 18 shows a spindle-like localization, perhaps due to localization at the mitotic spindle, and 19 persists there till late anaphase (telophase) when the signal disappears, presumably due to the 20 spindle dis-assembly. The disappearance of Sgo1 signal at telophase could possibly be due to 21 its dissociation and/or degradation (Fig. 1B). This persistent, dynamic and spindle 22 localization pattern of Sgo1 is in contrast with what has been observed in other yeast species 23 including S. cerevisiae and S. pombe. 24 Disappearance of Sgo1 with the spindle disassembly would suggest that its spindle 1 localization might require intact MTs. To investigate that, we followed the Sgo1-2GFP 2 staining with (-NOC) or without (+NOC) the MTs in the cells deleted for a SAC component 3 MAD2 to avoid SAC-mediated cell cycle arrest upon disruption of MTs (Fig. 1C). Although 4 SPB separation was poor in the absence of MTs in mad2Δ/Δ mutant cells, we failed to 5 observe a spindle-like localization of Sgo1 but observed either no or weak and diffused 6 localization in nocodazole treated anaphase cells (Fig. 1C, +NOC) as compared to the cells in 7 the presence of MTs (Fig. 1C, -NOC) suggesting that Sgo1 is indeed localized along the 8 spindle MTs. To examine whether this localization is mediated by an association between the 9 MTs and its MT binding motif that we identified through in-silico analysis (Fig. 1A), we 10 expressed a mutant version of Sgo1 expressing Sgo1-MTΔ lacking the MT binding motif 11 (from 258 to 358 residues) fused with GFP at the C-terminus (Suppl. Fig. S2F) as the sole 12 source of Sgo1. However, we failed to detect any lack of spindle association of Sgo1 MTΔ-

13
GFP as compared to Sgo1-GFP (Suppl. Fig. S2G). It is possible that Sgo1 uses some other 14 domain or its interaction with some other MT binding protein(s) is required for its association 15 with the MTs.

17
Sgo1 has a role in timely spindle disassembly 18 Since we observed that Sgo1 associates with the mitotic spindle (Fig. 1B), we wished to 19 investigate its significance. We failed to observe any gross abnormalities in spindle (marked 20 by Tub1-GFP) morphology in the wild type and sgo1Δ/Δ strains (Suppl. Fig. S3A) negating 21 any function of Sgo1 on the spindle assembly. However, when we counted those cells with 22 SPB-SPB (Tub4-Tub4) distance > 6 µm, we observed 47% more cells in the sgo1Δ/Δ mutant 23 than the wild type showed intact spindle suggesting a possible defect in the spindle 24 disassembly in the mutant ( Fig. 2A). Additionally, in support of the defect in spindle 25 disassembly, we could observe that sgo1Δ/Δ mutant exhibited longer anaphase spindles as 1 compared to the wild type cells. We used mad2Δ/Δ mutant, that showed wild type spindle 2 lengths, to conclude that this spindle defect of sgo1Δ/Δ was distinct from its SAC related 3 defect (see later results) at the centromeres (Fig. 2B). In HeLa cells, depletion of Sgo1 causes 4 an increase in metaphase spindle length due to absence of tension between the sister 5 chromatids (Salic et al 2004). To assess if the longer anaphase spindles in C. albicans 6 sgo1Δ/Δ cells is a consequence of longer metaphase spindles, we analyzed the metaphase 7 spindle lengths in the wild type and sgo1Δ/Δ cells but observed no significant difference 8 between them (Suppl. Fig. S3B).

9
Next, we sought to address how loss of Sgo1 might affect spindle disassembly. Since 10 in S. cerevisiae, Ipl1 is required for spindle disassembly during late anaphase (Woodruff et al 11   2009), we argued that lack of Sgo1 in C. albicans may hinder normal timing of spindle 12 disassembly by misregulating Ipl1 localization and/or its activity at the spindle. But we failed 13 to observe any significant difference in gross intensity of Ipl1-2GFP along the spindle in 14 anaphase cells (segregated SPBs, intact spindle) between wild type and sgo1Δ/Δ mutant 15 (Suppl. Fig. S3C). However, similar to the observed intact long anaphase spindle ( Fig. 2A), 16 in the mutant 42% more cells over the wild type showed Ipl1 staining persistent along the 17 spindle when SPB-SPB distance was more than 6 µm (Fig. 2C). In order to correlate the 18 temporal dissociation of Ipl1 with spindle disassembly, we simultaneously monitored spindle 19 (Tub1-RFP) and Ipl1. We scored cells harboring more than 6 µm spindle and assessed the 20 presence or absence of Ipl1 on the spindle. The percentage of cells with Ipl1 staining on the 21 spindle in the sgo1Δ/Δ mutant was comparable with the wild type (Fig. 2D, type I). We could 22 not observe any early dissociation of Ipl1 from the spindle in the mutant cells. The percentage 23 of cells with intact spindles but no Ipl1 on them was similar in the mutant and the wild type 24 (Fig. 2D, type II). This suggests that in the mutant although Ipl1 resides on the longer spindle 25 like wild type, it cannot promote spindle disassembly due to absence of Sgo1. These results 1 indicate that the presence of Sgo1 at the spindle might be required for activity of Ipl1 at the 2 spindle for its timely disassembly. We conclude that in C. albicans, Sgo1 displays a spindle 3 disassembly function reminiscent of S. cerevisiae Ipl1, albeit does not perform it at least by 4 mis-localizing Ipl1.  Since an absence of tension causes SAC localization at the centromeres, we sought to 23 examine whether in the absence of tension, Sgo1 is enriched at the centromeres in a SAC-24 dependent manner. Sgo1-TAP ChIP in mad2Δ/Δ mutant strain (mad2Δ/Δ::SGO1/SGO1-25 13 TAP), in the presence or absence of microtubules, was performed. A significant drop in Sgo1 1 localization at the centromeres in the absence of Mad2 as compared to its presence ( Fig. 3A) 2 confirmed that Sgo1 enrichment at the centromeres is Mad2-dependent at metaphase.

3
However, Sgo1 localization, albeit at lesser extent, at the centromeres in the cycling cells is 4 Mad2 independent (compare wild type and mad2Δ/Δ, -NOC, Fig. 3A). Further, Sgo1-2GFP 5 intensity in the wild type and mad2Δ/Δ metaphase stage cells in the presence (-NOC) or 6 absence (+NOC) of microtubules ( Fig. 3B) was found to be consistent with the ChIP data 7 suggesting that the alterations in Sgo1 level is not due to fixation of the cells with 8 formaldehyde. Based on these observations, we conclude that Sgo1 is associated with 9 centromeres under normal condition in a Mad-independent manner but its recruitment at the   We observed a sharp decrease in viability from 60 to 0% after 2 to 8 h of nocodazole  Since, in C. albicans, Sgo1 exhibits SAC function (Fig. 4) where Sgo1 and Mad2 may 5 function in tandem to ensure efficient checkpoint arrest, we reasoned that these two proteins 6 may interact directly or indirectly. To investigate if localization of Mad2 at the unattached 7 kinetochore depends on Sgo1, we tagged Mad2 with 3HA at the C-terminus in wild type and 8 sgo1Δ/Δ strains, harboring Cse4-GFP as the centromere marker. Absence of Sgo1 did not 9 affect the expression of Mad2 (Suppl. Fig. S6A). We then performed the chromatin spread 10 assay after treating these strains with nocodazole for increasing duration. Following 2 h of 11 treatment, there was no difference in Mad2 co-localization with Cse4-GFP between wild type 12 and sgo1Δ/Δ strains (Fig. 5A). However, when we prolonged the duration of nocodazole-

24
To examine if Sgo1 has a similar role in sister chromatids bi-orientation in C. albicans, the 25 wild type and the mutant (sgo1Δ/Δ) strains harboring an array of tet operators integrated 1 immediately adjacent to CEN7 on one copy of chromosome 7 and expressing TetR-GFP 2 fusion protein (Burrack et al, 2013) were used. The spindle poles (SPBs) were marked with 3 Tub4-mcherry in these strains. For biorientation assay, we treated both the wild type and the 4 mutant strains with nocodazole to disassemble the spindle, and then released them into 5 nocodazole-free media to allow for de novo reassembly of the mitotic spindle (Suppl. Fig.   6 S7A). Biorientation was scored as two GFP dots each proximal to two separated SPBs we examined the chromosome segregation in these cells following full spindle recovery. We 18 observed that in wild type strain around 86% of the cells showed proper chromosome 19 disjunction, whereas in sgo1Δ/Δ strain the same dropped to 43% (Suppl. Fig S7B). This 20 further confirmed that during establishment of de novo KT-MT attachment Sgo1 is required 21 to promote sister chromatin biorientation to ensure accurate chromosome segregation.  Additionally, across sexually reproducing organisms, Sgo1 protects centromeric cohesin 8 during meiosis I (reviewed in Marston 2015). Therefore, we wished to test whether Sgo1 has 9 any role in sister chromatid cohesion during mitosis in C. albicans. We performed the sister 10 chromatid cohesion (SCC) assay using the strains used for the biorientation assay. Both the 11 wild type and sgo1Δ/Δ strains were treated with 20 µg/ml nocodazole to remove the 12 microtubule-based pulling force. Under this tension-less condition, in the cells with 13 undivided DAPI mass, the cohered sister-CEN7-GFP dots will appear as a single dot ( Fig. 14 7A, +NOC, type I) due to diffraction limit, whereas non-cohered sisters will appear as two

23
Single shugoshin protein in C. albicans 24 Since shugoshin was first identified in flies as MEI-S332 (Kerrebrock et al 1992), this protein 1 exists in nearly all eukaryotes as a single protein or as two paralogues. While in flies and 2 budding yeast, a single shugoshin is present, in fission yeast and in vertebrates two 3 paralogues are present. Through in-silico analysis, in C. albicans we found a single shugoshin 4 protein (Suppl. Fig. 1). In organisms with two paralogues, each protein has separate and 5 overlapping functions suggesting that the separation of functions occurred late during 6 evolution. In fact, it is believed that the centromeric cohesion protection function of 7 shugoshin has been acquired later, from its function in sister chromatid bi-orientation  Although the function of shugoshin as a sensor for the lack of tension appears to be 2 conserved, the interaction of shugoshin with SAC for the purpose of onward signaling of the 3 response varies in a species-specific manner. In human, mouse and frog, shugoshin (Sgo2 or 4 Sgo) has been shown to physically interact with SAC protein, Mad2 in meiosis but not in 3 albicans we show that, the SAC-mediated metaphase arrest was compromised in CaSgo1-null 4 cells harboring defects in KT-MT attachments (Fig. 4A, Suppl. Fig. S5B). Possibly, the 5 observed drop in cell viability of the Casgo1 mutant following exposure to anti-microtubule 6 drug (Fig. 4D, Suppl. Fig. S5A) might be, akin to S. cerevisiae, due to failure in de novo 7 sister chromatid bi-orientation during spindle recovery from the drug. Nevertheless, the 8 generation of multibudded cells in presence of the drug (Suppl. Fig. S5B), argues for 9 attenuation of SAC and bypass of the mitotic arrest in absence of Sgo1 in C. albicans. These 10 results indicate that in C. albicans Sgo1 plays a role for the activation of SAC at the 11 centromeres. Consistent to this we observed that the maintenance of SAC, at least Mad2 and 12 Bub1, is compromised in Sgo1-null cells (Fig. 5). Interestingly, we also observed a reduction Overall, in this study, we characterize shugoshin for the first time in C. albicans. 3 Consistent to the conserved functions of shugoshin across eukaryotes, we observed several 4 mitotic functions which are similar to other organisms (Fig. 8). Notably, we report here two 5 important adapter functions along the cell cycle that have not been yet reported in any other 6 organisms including fungi (Fig. 8). First, shugoshin in C. albicans may play additional roles   Table S1, Table S2 and Table S3, respectively in the supplementary material.  For generating conditional mutant, the native promoter of the second copy of SGO1 in 11 strain SGY8001 was shuffled with MET3 promoter by linearizing (using BstBI site in SGO1) 12 and integrating pCaDIS-SGO1, which was constructed by cloning upstream sequence (1283 13 bp) of SGO1 as BamH1-Pst1 fragment in-frame with MET3 promoter sequence in pCaDIS. For SGO1-2GFP tagging, we first constructed plasmid pBS-SGO1-GFP-HIS1 by cloning 6 SGO1 in pBS-GFP-HIS1 plasmid. We next cloned GFP-linker in the above plasmid to 7 generate pBS-SGO1-2GFP-HIS1. This was then linearized within SGO1 by partial restriction 8 digestion using BstB1 and was integrated into strains SN148 and J110 to generate strains 9 SGY8156 and SGY8243, respectively.

16
For SGO1-TAP tagging, a Candida optimized tandem affinity purification (TAP) tag 17 cassette was amplified from plasmids pFA-TAP-ARG4 and pFA-TAP-HIS1 using primers 18 SA6 and SA7 and strains SN148 and J110 were transformed with this cassette to replace one 19 copy of SGO1 with SGO1-TAP, generating strains SGY8038 and SGY8130, respectively. For 20 tagging the second copy of SGO1 in strain SGY8038, same strategy was followed using 21 plasmid pFA-TAP-HIS1.

4
For IPL1-2GFP and TUB1-RFP, one copy of IPL1 and TUB1 was tagged with 2GFP 5 and RFP respectively, by linearizing (XbaI) and integrating plasmids pBS-IPL1-2GFP-HIS1 6 and pBS-TUB1-RFP, respectively to generate the indicated strains (Table S1). CIp10-TUB4-MCHERRY-LEU2 to construct the indicated strains (Table S1)     Live cell DAPI staining for sister chromatid cohesion (SCC) assay 20 DAPI (Invitrogen, D1306) was added at the final concentration of 7 µg/ml to the early log 21 culture and the cells were allowed to grow for one more generation. Subsequently, the cells 22 were grown for 2 h in the absence (-NOC) or presence (+NOC) of 20 µg/ml nocodazole.

23
Cells were then harvested, washed and imaged.

24
Live cell microscopy and image processing 1 All live cell microscopy images, except for bi-orientation assay, were captured using confocal 2 laser scanning microscope LSM 780. Images were processed using Zeiss Zen 2012 software.

3
All intensity measurements of fluorescence signals were done using ImageJ software and 4 intensity values (a.u) reported after correction with the background signal. Imaging for bi-5 orientation assay was done using AxioObserver Z1 Zeiss inverted microscope and processed 6 using AxioVs40 V 4.8.2.0 software. (1 µg/ml) was used for nuclear staining.              were treated with 20 µg/ml nocodazole for 90 min and were released into drug-free media for 8 15 min before they were imaged. Representative images of the cells showing bioriented (type 9 I) or monooriented (type II) sister chromatids. In both these strains, SPBs were marked with 10 Tub4-mcherry and one copy of chromosome 7 was marked with TetO-GFP-TetR (see text).

11
Scale bar ~5 µm. Histogram plot depicts distribution of type I and II categories among the 12 cell population of the indicated strains; n ≥ 186.