zmm mutants show decreased levels of crossovers at both 23 and 33°C, with a more severe defect at 33°C [16]. Nevertheless, msh4 and msh5 mutants showed improved gamete viability at 33°C (Figure 1A and B). This was not the case for the remaining zmm mutants (Tables 1 and 2). At 23°C, all of the zmm mutant strains showed the characteristic increase in the two- and zero-viable spore classes, consistent with meiosis I non-disjunction (Table 1). The Y55 msh5 and the msh4 msh5 mutants displayed a more severe decrease in the proportion of four-viable spored tetrads compared to the msh4 strain at this lower temperature. Since Msh5 has previously been suggested to be important for chromosome segregation in crossover-defective mms4 mlh1 strains [25], it is possible that Msh5 has a role independently of Msh4. If so, this is strain-specific as no differences in gamete viability between the msh4 and msh5 strains was observed in SK1 (Table 2). Regardless of this, the temperature-dependent improvement in gamete viability was observed in all mutants defective for MSH4, MSH5 or both, irrespective of strain background. Thus, our data suggest two temperature-dependent processes: Msh5 may have a strain-specific chromosome segregation function at 23°C, whereas elevated temperature (33°C) positively impacts chromosome segregation in the absence of Msh4 or Msh5. We focus on the latter phenotype.

To verify that the improved spore viability at 33°C was due to improved chromosome segregation, we assessed the segregation of chromosome III in tetrads. To this end, we generated diploid strains where LacO (Y55, BR) or TetO (SK1) repeats had been inserted near the centromere of chromosome III. These strains also express the recombinant LacI-GFP or TetR-GFP that recognize the LacO and TetO repeats, respectively. Upon the completion of meiosis, each of the four spores in the ascus should contain a single GFP focus, provided meiosis I segregation occurred normally. In contrast, a meiosis I non-disjunction event will generate an ascus where two of the spores contain two GFP signals and two spores lack GFP altogether. In all of the three strains, meiosis I non-disjunction frequencies were decreased ∼3-fold at 33°C compared to 23°C (Figure 1C). Thus, improved chromosome segregation is the most likely explanation for the enhanced spore viability in the msh4 and msh5 strains at 33°C.

msh4 mutants are defective in crossing over [12], [16], [26]. Therefore, we considered whether crossover formation was different at 33°C in the msh4 mutant. Classical genetic analysis of crossing over on chromosome III (Figure 1D) in the four-viable spored tetrads revealed similar crossover levels in the msh4 mutant at 23°C and 33°C (Table 3). These observations suggest that total levels of crossovers within the population of viable spores are unaffected by temperature in the msh4 mutant.

To determine unambiguously that msh4 mutants have improved chromosome segregation independently of enhanced crossover frequencies, we assessed the segregation of a homeologous chromosome pair. In this diploid strain of Y55, one of the chromosome IIIs had been replaced by the co-linear chromosome from the sibling species, S. paradoxus. The resulting sequence divergence (10–20%) [27], [28] suppresses crossing over in ∼90% of wild-type meioses [24]. We failed to observe a single crossover in the LEU2-MAT interval in the inspected tetrads at 23°C or 33°C in the msh4 strain (data not shown), suggesting that crossing over is not increased. The homeologous chromosome pair missegregates in 7.3% and 1.7% of meioses at 23°C and at 33°C respectively (Table 4). These observations support the hypothesis that temperature regulates segregation independently of modulating crossover levels on chromosome III.

It is unclear whether temperature improves the segregation of the homeologous chromosome pairs in the presence of MSH4, since the homeolog pair did not display any detectable temperature-dependent decrease in its non-disjunction frequency at 33°C. In the wild-type strain, the homeolog pair missegregated in 8.2% of meioses at 23°C, similar to previous observations [24], and 6.9% at 33°C (Table 4). These observations raise the possibility that the temperature-dependent modulation of chromosome segregation might be specific to msh4 mutants.

Progression to meiosis I in zmm mutants in the SK1 background is more severely abrogated at 33°C than at 23°C. This is accompanied by a more severe defect in crossing over at the HIS4LEU2 hotspot at 33°C (15% of wild type) compared to 23°C (40–50% of wild type) [16]. Nevertheless, both the msh4 and msh5 mutants in SK1 displayed relatively good viability, particularly at 33°C (Table 2). It is therefore possible that preferential progression of the meiotic cells with relatively high accuracy of chromosome segregation may be selected. Although this could explain the phenotype in SK1, in Y55 and BR, sporulation frequencies were similar at the two temperatures in the msh4 strains (>45%, Table 1 and data not shown). Importantly, further decreasing the sporulation frequency of msh4 in Y55 by increasing temperature to 37 or 39°C, does not influence viability or crossover frequencies (Tables 1 and 3).

Homolog pairing was also not affected by temperature. Using fluorescent in situ hybridization (FISH) with probes against chromosome III as well as V in the msh4 mutant (EY137), we observed homolog pairing frequencies that were similarly high at both 23 and 33°C (>95%, data not shown), consistent with previous observations that deletion of MSH4 or MSH5 does not influence the juxtapositioning required for recombination in the Spo11-independent Cre/Lox system [29]. Nevertheless, homolog recognition mediated by meiotic recombination is presumably still required for the improved segregation in msh4 since spo11 msh4 double mutants displayed <2% spore viability (irrespective of temperature), similar to a spo11 mutant (Table 5).

Finally, since other zmm mutants did not display the improvement in chromosome segregation at 33°C, we carried out epistasis analysis. ZIP1, which encodes the transverse element of the synaptonemal complex, is required for the temperature-mediated improvement in chromosome segregation (Table 5). The msh4 zip1 mutant displayed similar spore viabilities at 23 and 33°C.

All strains are listed in Supplementary Table S1 and were verified by PCR and/or Southern blot as well as meiotic segregation. Initial screening of gamete viability in zmm mutants was carried out in SK1 on 2% KAC-COM-agar 2% KAC, 0.22% yeast extract, 0.05% D-glucose, 0.087% complete drop-out mixture (Abdullah and Borts 2001), 2% agar, pH 7.0. This medium was also used for Y55. For BR, sporulation on solid medium was carried out on 2% KAC, 0.2% yeast extract, 0.1% D-glucose, 0.1% complete amino acid mix (Bio101), and 2% agar, pH 7.0. Spore viability was determined by standard tetrad dissection.

Meiotic spreads and FISH analysis was carried out, as described [33]. GFP was visualised using a Deltavision IX70 system (Applied Precision) using the softWoRx software, and an Olympus Plan Apo 100×1.4 numerical aperture objective lens.

Map distances were calculated according to Perkins [34]. We used the Fisher exact test, t-test for proportions, or G-test for homogeneity, adjusting the P-values according to Dunn-Sidak or Tukey-Kramer methods when multiple comparisions where made.