Supplemental Material

This article contains the following supporting material:

• Supplemental Table 1
• Supplemental Figure 1 - Lack of Sgo2 does not affect cohesion at centromeres in nda3KM311-arrested cells. nda3KM311 cen2-GFP, sgo2Δ nda3KM311 cen2-GFP, swi6Δ nda3-KM311 cen2-GFP and bub1Δ nda3-KM311 cen2-GFP cells were incubated at 18°C for 8 hours. Four categories of cells were distinguished: cen2-GFP was detected (1) as a single dot, (2) as a doublet, (3) as two spots separated roughly by 0.5μm and finally (4) as two spots on opposite sides of the nucleus. Situation (3) was found only in cells lacking Swi6 and likely indicates the loss of centromere cohesion (arm cohesion maintains the two centromere spots in proximity). The number of cells in each category was quantified (right panel, a minimum of 700 cells was counted in total in 4 separate experiments). Situation (4) was found predominantly in cells lacking Bub1 and indicates the total separation of the two sister chromatids due to the activation of separase (cells are spindle checkpoint-defective in the absence of Bub1, see (Bernard et al., 1998), (Vanoosthuyse et al., 2004)). We detected only a small but reproducible increase in the proportion of doublets (situation (2)) in cells lacking Sgo2. The same phenomenon has been reported previously in meiotic cells lacking Sgo2 (Rabitsch et al., 2004). Similarly lack of Sgo2 did not significantly affect centromeric cohesion in the presence of functional kinetochore-microtubule attachments in cells arrested in metaphase by over-expressing the checkpoint component Mad2 (He et al., 1997) (not shown).
• Supplemental Figure 2 - Wild-type and sgo2Δ cells release from an nda3KM311-dependent arrest with similar timing. nda3KM311 and nda3KM311 sgo2Δ cells were arrested for 8 hours at 18°C, filtered and released in pre-warmed media at 36°C. Cells were fixed at different time-points in cold methanol and processed for tubulin immuno-fluoresence. Cells were scored in three categories: (i) pro-metaphase cells, that did not yet reform a spindle and were still stuck in the arrest; (ii) cells with a spindle (metaphase, anaphase A or B) and finally (iii) interphase cells. A minimum of 200 cells were counted for each time-point. The data presented is from one representative experiment out of three repeats.
• Supplemental Figure 3 - Localisation of the Passenger proteins in the nucleolus and in metaphase (A) Pic1/INCENP fused to mCherry (Pic1-mCherry) co-localises with the nucleolar marker Gar2 fused to GFP (Gar2-GFP) in interphase cells. Note that all Passenger proteins co-localise throughout the cell-cycle (Figure 2 and see text for details). (B) Cells expressing GFP-Bir1/Survivin and the kinetochore marker Ndc80 were arrested in metaphase by over-expressing the spindle checkpoint component Mad2 (He et al., 1997). An explanatory diagram is presented on the right. SPB stands for Spindle Pole Body and OTR refers to “Outer Repeat” that designs the inner-centromere. (C) Aurora B and Ndc80 were imaged simultaneously in cycling cells. In interphase, Aurora B localised to the nucleolus and the SPB-centromere cluster, like the other two Passenger proteins Survivin and INCENP (see Figure 2). In metaphase Aurora B localised on centromeres, telomeres (see (D)) and on SPBs (see arrows and (E)). (D) In metaphase, Aurora B colocalised with the telomere marker Taz1. (E) Aurora B co-localised with the SPBs marker Sad1 in metaphase.
• Supplemental Figure 4 - Lack of Sgo2 does not affect the overall centromere structure. (A) Lack of Sgo2 does not affect the localisation on kinetochore (Ndc80-CFP) of the DASH complex component Ask1 in a normal metaphase (Ask1-GFP). (B) Lack of Sgo2 does not alleviate silencing of the ade6+ inserted in the centromeric heterochromatin (OTR::ade6+) and are therefore unable to grow on media lacking adenine. The rik1.304 mutation (Allshire et al., 1995) was used as positive control for alleviation of silencing. Cells were spotted on rich media or media lacking adenine and grown for three days at 28°C.
• Supplemental Figure 5 - Sgo2 localisation is not affected by the Bir1/Survivin mutation cut17.275. On all panels, an overnight culture of cut17.275 cells was shifted at the restrictive temperature of 36°C for 0-5 hours. The scale bar represents 2 μm. (A) cut17.275 cells expressing Ark1/Aurora B-GFP, Sgo2-mCherry and the kinetochore marker Ndc80-CFP were observed after two hours at 36°C. Both Ark1/Aurora B and Sgo2 still localised on centromeres in these cells. (B) Pic1/INCENP still localised on centromeres and on the spindle after 2 hours at the restrictive temperature. In anaphase cells, Sgo2 re-localised on the chromatin like in the wild-type (see Figure 2A). (C) Over time at 36°C, Ark1/Aurora B localisation on centromeres became less robust. Indeed after 4 hours at 36°C, Ark1/Aurora B was partially released in the nucleolus (arrows) whilst Sgo2 was maintained on centromeres. (D) Quantification of fluorescence intensities on centromeres at 1h versus 4h after the shift at the restrictive temperature. For each centromere (n=20 in each condition), the fluorescence intensities of Ark1 (GFP), Sgo2 (m-Cherry) and Ndc80 (CFP) were compared to each other as ratios. At 4 hours, Ark1 fluorescence intensities decreased by roughly 50% whilst Sgo2 remained unaffected. (E-G) The Bir1/Survivin mutation cut17.275 induced a mad2-dependent mitotic delay at the restrictive temperature. (E) Quantification over time of the percentage of cells where Ark1/Aurora B and Pic1/INCENP were found on centromeres. A minimum of 600 cells in three different experiments was counted for each time-point. That percentage peaked after 2 hours and decreased again, suggesting that cut17.275 cells underwent a delay in early mitosis at the restrictive temperature. A great number of cut17.275 cells show clear kinetochore-microtubule attachment defects (not shown and (Morishita et al., 2001; Huang et al., 2005)), suggesting that the delay is spindle checkpoint-dependent. (F) Consistent with this, the spindle checkpoint component Mad3 was also recruited to kinetochores in these cells with a similar timing. However in the absence of the spindle checkpoint component Mad2 the delay was abolished. (G) After four hours at the restrictive temperature, only a few cells maintained Sgo2, Pic1/INCENP or Ark1/Aurora B on centromeres and instead Sgo2, Pic1/INCENP and Ark1/Aurora B re-localised to their interphase localisation pattern (not shown). Consistent with these observations, the mitotic index measured by spindle staining increased after three hours at the restrictive temperature and then decreased again. Cells were shifted at 36°C, fixed in cold methanol at different time-points and processed for anti-tubulin immuno-fluorescence to identify mitotic cells. Results presented are the average of two independent experiments. Note that we found mitotic spindles harder to fix at the restrictive temperature.
• Supplemental Figure 6 - Lack of Sgo2 did not significantly affect the stability of either Bir1/Survivin or Pic1/INCENP in checkpoint-arrested cells. Extracts were prepared from cells arrested in mitosis at 18°C using the tubulin mutation nda3KM311 . Serial dilutions of extracts and immunoprecipitated complexes were loaded on the gel. The checkpoint component Mad1 was used as a loading control.
• Supplemental Figure 7 - nda3KM311-arrested cells show several types of kinetochore-SPBs connections. Upon shift to the restrictive temperature of 18°C, nda3KM311 cells arrest in mitosis with hyper-condensed chromosomes and an activated spindle checkpoint. Despite lack of microtubules, kinetochores and SPBs maintain connections. (A) nda3KM311 cells arrested for 8 hours at 18°C were shifted back to 36°C for one minute and fixed with paraformaldehyde. In these conditions an anti-tubulin antibody recognised SPBs. In these cells, kinetochores were visualised using an anti-GFP antibody recognising the kinetochore component Ndc80 tagged with –CFP (Ndc80-CFP). Two examples of arrested cells are shown. In both examples, a kinetochore pair has no connection with any SPBs (red circle). In the example N°1 (top), the duplicated SPBs are next to each other. Each kinetochore of a pair is connected to a different SPB (see enlargment on the right): the kinetochore-SPB connection is therefore amphitelic. In the example N°2, the duplicated SPBs have moved away from each other. Each SPB is connected to a pair of kinetochores. The kinetochore-SPB connection is therefore syntelic or monotelic. (B) nda3KM311 cells arrested for 8 hours at 18°C were fixed with paraformaldehyde and the nuclear envelope was visualised with the monoclonal antibody Mab414 that recognises nucleoporins (Covance Research Products). In the arrest, the nuclear envelope often loses its circular shape and ultimately forms several nuclear vesicles.