Distinct functions of S. pombeRec12 (Spo11) protein and Rec12-dependent crossover recombination (chiasmata) in meiosis I; and a requirement for Rec12 in meiosis II

Structure of rec12⁺ gene and complementation by rec12⁺cDNA

While rec12⁺ was reported to encode a protein of 139 amio acids [30], conceptual translation of the surrounding region suggested that rec12⁺ encoded a larger protein with homology to Spo11 of S. cerevisiae and Top6A of Sulfolobus shibatae [17, 21]. Alignments and splicing sequence profiles [31] were used to identify four putative introns (Figure 1A). Oligonucleotide primer pairs flanking each of the putative introns were used for PCR and RT-PCR to amplify genomic DNA and meiotic mRNA, respectively. In each case, the product from mRNA was shorter than that from genomic DNA, confirming the presence of four introns (Figure 1A).

Presence of four introns suggested that the rec12⁺ cDNA encodes a protein of 345 amino acids with high homology to eukaryotic Spo11 proteins (Figure 1B). To test this hypothesis, we introduced a full-length genomic clone and a sequence-confirmed cDNA into inducible expression vectors [32] in such a way that the first methionine in predicted exon 1 would be used for translation (Figure 1C). Each of these constructs restored wild-type levels of recombination to rec12-117 strains (Figure 2A, 2B). We conclude that the cDNA encodes a functional Rec12 protein of 345 amino acids in length. This conclusion was confirmed by Western blotting of an epitope-tagged Rec12 protein expressed from the endogenous rec12⁺ locus: a meiotically-induced protein band of the expected size was observed (W.D. Sharif and W.P. Wahls, unpublished observations).

Rec12 (Spo11) protein shares homology with the catalytic (Top6A) subunit of Topoisomerase VI, which functions as a heterotetramer with Top6B [16, 17]. However, no Top6B homolog is present in most eukaryotes and the Rec12 (Spo11) enzyme does not need to affect strand passage in the manner of canonical type II topoisomerases [11]. We therefore tested whether overexpression of rec12⁺ would lead to a hyper-recombination phenotype, thereby implicating function as a homodimer. Very low levels of ectopic rec12⁺ expression restored wild-type recombination levels to the rec12-117 mutant (Figure 2B), confirming that functional Rec12 protein was produced. However, upon overexpression of rec12⁺ the frequency of recombinants did not increase significantly above wild-type levels (Figure 2C, 2D). Thus, Rec12 protein in excess of wild-type levels is biologically inactive; some other factor becomes rate-limiting under these circumstances.

The rec12-117 mutant was isolated in a genetic screen; the rec12-117 mutation reportedly confers a 100-fold or greater reduction of intragenic recombination and a 6-fold reduction of intergenic recombination [33]. To define more precisely the requirement for Rec12 in meiotic chromosome dynamics, we used gene replacement to generate a null (rec12-D15) mutation and a mutation in the putative active site tyrosine (rec12-Y98F). We then determined the effects of these mutations upon intergenic (crossover) recombination within intervals of chromosomes I, II, and III. We also employed a combination of cytological and genetic approaches to reveal the nature of meiotic chromosome segregation defects for each chromosome in each of the three mutants.

Rec12 and its active site tyrosine are required for chiasmatic meioses

Models for recombination posit that gene conversion and reciprocal exchange (crossover recombination) are alternative outcomes of a common intermediate [11]. Intragenic recombination at the ade6 locus, a measure of gene conversion, was reduced more than 100-fold in rec12-117 mutants [33], and we have obtained similar results with the rec12-117, rec12-D15, and rec12-Y98F mutants (Figure 2; W.D. Sharif and W.P. Wahls, unpublished data). Because reciprocal exchange accompanies more than half of all gene conversions at the ade6 locus [34, 35], the rec12 mutants must also have a reduced number of reciprocal exchanges in the vicinity of ade6.

If Rec12 protein is required for all (or most) meiotic recombination, then rec12 mutants might fail to produce a sufficient number of reciprocal exchanges to ensure that each chromosome pair receives at least one crossover. To address this issue, we determined the number of crossovers present in regions of chromosomes I, II, and III in the rec12-D15 (deletion) and rec12-Y98F (active site) mutants. Multiple genetic intervals encompassing approximately 20% of the genome (444 cM) were analyzed.

Rec12 is required for proper meiotic chromosome segregation

The rec12 mutant meioses are achiasmatic and would thus be expected to suffer chromosome segregation errors during meiosis I. To gain insight into this possibility, asci from meiotic cultures were stained with a DNA specific fluorescent dye (DAPI) and visualized by differential interference contrast (DIC) and fluorescence microscopy. The majority of asci from wild-type cells contained four well-rounded spores, each with a single DAPI-staining body of equal intensity (Figure 3A). The rec12 mutants were proficient at meiosis and underwent two meiotic divisions as revealed by ascus development and the distribution of chromosomes. However, each of the rec12 mutants exhibited a high frequency of chromosome segregation errors that were sometimes accompanied by defects in spore formation and/or ascus development (Figure 3B). The data from a large number of asci from each mutant are presented schematically in Figure 3C.

The rec12-117, rec12-D15, and rec12-Y98F mutants exhibited qualitatively and quantitatively similar cytological phenotypes (Figure 3C). They each produced between 60% and 66% of asci with gross defects in DNA distribution. These mutant phenotypes were not uniform; different asci from the same cross were differentially affected (Figure 3). This heterogeneous phenotype is characteristic of a variety of mutations that directly affect meiotic chromosome dynamics [37–39]. Aberrant (or random) assortment of the three chromosome pairs (or twelve chromatids) leads to heterogeneous DNA distributions within different asci.

The rec12 mutants also exhibited high frequencies of additional developmental defects such as abnormal spore placement, reduced spore numbers, and abnormal spore formation (Figure 3C). It has been suggested previously [38] that these phenotypes are characteristic of a defect in meiosis II because spore formation in S. pombe is controlled by the spindle pole body of the MII spindle [40–42].

S. pombe produces ordered tetrads. If chromosomes segregated aberrantly during the MI reductional division, and then segregated appropriately during the equational division of MII, one would expect sister spores to receive identical complements of chromosomes. However, among rec12 mutant asci with four spores and an aberrant DNA distribution (Figure 3C, classes b, c, d, e), only 26% or less exhibited a pattern characteristic of simple MI nondisjunction (Figure 3C, classes c, e). The remaining 74% or more had DNA distributions characteristic of precocious sister chromatid separation and/or MII segregation errors – DNA was unequally distributed between two sister spores (Figure 3C, classes b, d). In some cases a defect in chromosome segregation during MII was visualized more directly. For example, in many cases DNA masses failed to separate properly into two sister spores during the second meiotic division (Figure 3B; Figure 3C, classes f, g, j, k, m). Thus, Rec12 protein is required for some function of MII, as well as for recombination during MI.

rec12mutants produce a high frequency aneuploid meiotic products

Since S. pombe contains only three chromosome pairs, aberrant assortment during one or both of the meiotic divisions can produce, by chance, a significant fraction of meiotic products that receive at least one copy of each chromosome and are hence viable [38, 43]. Aberrant segregation can also produce nullisomic aneuploids (missing one or more chromosomes), polysomic aneuploids (having additional copies of individual chromosomes), and diploids [38, 39, 44]. In S. pombe the haploids and diploids are viable, nullisomic aneuploids are inviable, and disomic aneuploids tend to lose the extra chromosome or chromosomes and become haploid [45]. We therefore determined the frequencies of spore viability and meiotic diploidy as genetic measures of chromosome segregation errors.

Approximately 50% of the meiotic products from the rec12-117 and rec12-D15 (null) mutants were inviable (Figure 4A), indicating that about 50% of the products were nullisomic for one or more chromosomes. This value is close to the frequency (58%) of nullisomics that one would expect if chromosome segregation were completely random in either one of the two meiotic divisions. Interestingly, the rec12-Y98F mutant produced a significantly higher frequency of inviable spores than the rec12-117 and rec12-D15 mutants (Figure 4A), suggesting that rec12-Y98F is a separation of function mutation.

As a second measure of aberrant segregation, we determined the frequency of diploid spore colonies in the mutants. Each of the rec12 mutants behaved similarly and produced approximately 10% diploid meiotic products, more than 30-fold higher than the frequency produced by wild-type cells (Figure 4B). If chromosome segregation were completely random in either one of the two meiotic divisions, one would expect approximately 4% of the viable products to be disomic for all three chromosomes. The value obtained in the rec12 mutants was 2.5-fold higher than the predicted random value, indicating that the segregation errors are not simply the consequence of random chromosome assortment (nondisjunction) occurring exclusively during one of the two meiotic divisions.

Rec12 is required for chiasmatic and achiasmatic chromosome segregation during MI; and its active site tyrosine is required for chromosome segregation during MII

Because each of the rec12 mutants had aberrant meiotic chromosome segregation (Figure 3), produced a high frequency of viable meiotic products (Figure 4A), and produced a high frequency of meiotic diploids (Figure 4B), we were able to monitor directly the segregation patterns of all three chromosome pairs in each of the mutants. Meiotic diploids were genotyped for heteroallelic, centromere-linked markers on each of the three chromosomes. Those data were compared to the inheritance patterns predicted for each type of chromosome missegregation (Figure 5).

Approximately half of the diploid products from the rec12-117 and rec12-D15 (null) meioses were heterozygous for centromere-linked markers on chromosomes I, II, and III (Figure 5E). Thus, those meioses did not exhibit simple MI nondisjunction, because 100% of diploids arising from MI nondisjunction would be heterozygous at the centromeres (Figure 5B). Precocious separation of sister chromatids during MI [44, 46] can also produce diploid meiotic products, 67% of which would be heterozygous for centromere markers (Figure 5D) [38]. However, the average heterozygosities in the rec12-117 and rec12-D15 mutants were significantly below this value, indicating that these mutants must have suffered some proportion of MII segregation errors. By applying mapping functions for aneuploid meiotic products [47], we can set a lower limit of 22% and 29% MII nondisjunction for the rec12-117 and rec12-D15 meioses analyzed, respectively.

Surprisingly, the rec12-Y98F (active site) mutants produced meiotic diploids that were almost exclusively homozygous for centromere-linked markers (Figure 5E). The average heterozygosity for loci on all three chromosomes was 4.7 ± 4.1%. This indicates that about 95% of those rec12-Y98F mutant cells that produced diploid spores must have successfully completed the reductional segregation of chromosomes (MI) and failed to properly segregate sister chromatids during the equational division (MII).

Conserved structure and function of Rec12 (Spo11) to initiate recombination

Identification of four introns (Figure 1) and complete complementation of the rec12-117 hyporecombination defect by the full-length cDNA (Figure 2) confirmed that Rec12 protein is homologous to Spo11 of other eukaryotes [11]. Both Rec12 and its active site tyrosine are required for intragenic and intergenic recombination at multiple locations on all three chromosomes (Figure 2, Tables 1, 2, data not shown). We observed an average, 256-fold and 263-fold reduction in crossover recombination in the rec12-D15 and rec12-Y98F mutants for intervals encompassing ~20% of the genome. Since wild-type S. pombe has 44 crossovers per meiosis (between 11 and 18 per chromosome) [36], the mutants statistically have less than 0.1 crossovers per chromosome per meiosis and are thus achiasmatic.

It was reported previously that rec12-117 mutants exhibit only a 6-fold reduction in crossover recombination [33], which is at odds with our value of more than a 250-fold reduction in the rec12-D15 (null) and rec12-Y98F (active site) mutants (Table 1). This discrepancy is not due to the nature of the rec12-117 allele, as we have also observed a 200-fold or greater reduction in crossovers in rec12-117 mutants (W.D. Sharif and W.P. Wahls, unpublished data). Our observation that rec12-117 mutants produce a high frequency of viable diploid meiotic products, about half of which are heterozygous for centromere linked markers (Figures 4, 5, Table 2), provides a straightforward reason for the discrepancy: The previous study failed to exclude diploid spore colonies and scored only for double prototrophic recombinants (the double auxotrophic, reciprocal recombinants were not analyzed). Presumably the bulk of the colonies reported previously as "recombinants" were actually diploids (or disomics) exhibiting intergenic complementation.

We used dosage studies to see if Rec12 functions in a manner predicted by homology to Spo11/Top6A. Upon overexpression of Rec12 protein, the frequency of recombination did not increase beyond wild-type levels (Figure 2). The simplest interpretation is that Rec12 functions in a multisubunit protein complex; a complex in which other components are required either for efficient catalysis of dsDNA breaks or for correct subcellular localization of Rec12. In budding yeast the active site mutation exhibits allele-specific semidominance when heterozygous with wild-type [51], and we have made similar observations in fission yeast (W.D. Sharif and W.P. Wahls, unpublished data). This provides evidence that a dimer of Rec12/Spo11 protein, harboring two functional active sites, is required for catalysis of dsDNA breaks in each organism. Thus, with regard to primary protein sequence (Figure 1), inferred subunit structure (Figure 2 and data not shown), strict requirement for meiotic recombination (Figure 2, Table 1), and dependence upon its active site tyrosine (Table 1), Rec12 protein is a structurally and functionally conserved member of the Spo11 family of enzymes that catalyze meiotic dsDNA cleavage to initiate recombination [11].

Rec12 has dual roles in chiasmatic and achiasmatic (distributive) chromosome segregation during Meiosis I

In the absence of crossover recombination (chiasmata) and a backup distributive (achiasmatic) system for chromosome segregation, MI nondisjunction should occur and 100% of any resulting diploid spore colonies would be heterozygous for centromere linked markers (Figure 5B). However, only about 50% of diploid meiotic products from rec12-D15 (null) mutant meioses were heterozygous for centromere linked markers (Figure 5E), indicating that some fraction of the products were derived from precocious separation of sister chromatids (PSS) and/or MII nondisjunction (Figure 5C, 5D). Since PSS does not occur in rec12 null mutants [52], those centromere-homozygous chromosomes must have experienced a proper reductional segregation and a failed equational segregation (discussed below).

The rec12-D15 (null) mutants are achiasmatic (Table 1) and, as anticipated, exhibit a high frequency of MI nondisjunction (Figures 3, 5, Table 2). However, the frequency of MI nondisjunction observed is only half of that expected, suggesting the presence of a distributive (achiasmatic) system of chromosome segregation. There is precedence for a distributive MI segregation system in several organisms. [7, 53]. In rec7 hyporecombination mutants of S. pombe the fidelity of segregation of homologs during MI exceeds the expected random level [10]. Similarly, MI segregation of GFP-tagged centromeres in rec12 deletion mutants occurs at non-random levels (L. Davis and G.R. Smith, personal communication), in good agreement with our genetic results. In each case identified, this distributive system of fission yeast can partially circumvent the MI segregation errors of achiasmatic chromosomes; approximately one-quarter to one-half of those homologs expected to nondisjoin (i.e., move to the same pole) end up segregating distributively (i.e., move to opposite poles).

While the rec12-D15 (null) mutants frequently suffer MI nondisjunction, the rec12-Y98F (active site) mutants produce meiotic diploids with nearly normal MI segregation patterns (Figures 3, 5, Table 2). This difference is unrelated to crossover frequency, because the active site and null mutants have identical deficiencies in crossover recombination (Table 1). Thus, Rec12 protein has one or more functions in MI that are distinct from its functions to catalyze recombination and mediate chiasmatic segregation. We propose that the Rec12-Y98F protein facilitates distributive segregation of achiasmatic chromosomes during MI.

A meiotic checkpoint could provide additional time for missegregating chromosomes to correct their erroneous wanderings during MI [10]. We propose two likely mechanisms by which the Rec12 protein improves the MI segregation of achiasmatic chromosomes. First, Rec12 may be required for an "MI nondisjunction checkpoint" that delays meiotic cells in which proper MI disjunction has failed to occur. If Rec12 were required for such a checkpoint, the null mutants would be achiasmatic, would fail to delay for the time period required to establish distributive segregation, and would consequently undergo high rates of MI nondisjunction. In the presence of Rec12-Y98F protein the checkpoint would be active, thereby allowing the achiasmatic chromosomes sufficient time to reorient and undergo distributive segregation. Meiotic progression in spo11 (rec12) null mutants of budding yeast is accelerated [29], supporting our hypothesis that Rec12 may participate in an "MI nondisjunction checkpoint."

As an alternative (but not mutually exclusive) mechanism, Rec12 protein may participate directly in the machinery of distributive segregation. In female meioses of D. melanogaster, heterochromatic pairing of homologs can substitute for chiasmata to ensure bipolar spindle attachment of bivalents and subsequent distributive segregation [4–6, 8]. Rec12 may serve a similar role in maintaining a paired (or pseudo-paired) state between achiasmatic chromosomes prior to the first meiotic division. In support of this possibility, Spo11 of mice colocalizes extensively along homologously synapsed chromosomes during pachytene and dissociates when the homologs segregate from one-another during anaphase of MI [25].

Rec12 has a role in equational chromosome segregation during Meiosis II that is dependent upon its active site tyrosine

Cytological and genetic data each demonstrated that the rec12-D15 (null) and rec12-Y98F (active site) mutants frequently suffer chromosome segregation errors during MII (Figures 3, 5, Table 2). This role for Rec12 in equational chromosome segregation may be conserved, as spo11 mutants of Sordaria macrospora also exhibit frequent MII nondisjunction (S. Tesse, A. Storlazzi, S. Gargano, and D. Zickler, personal communication). Irregular partitioning of chromosomes during the second meiotic division has also been reported in fission yeast rec7 mutants [10], so MII missegregation may be a relatively common fate for chromosomes in hyporecombination mutants. As described previously in this paper and elsewhere [38], the biology of fission yeast facilitates the identification of MII segregation defects that are difficult to detect in other organisms that terminate meiotic progression or produce inviable meiotic products as a consequence of defects in meiotic chromosome dynamics.

Although they exhibit a striking difference in their apparent MI nondisjunction frequencies, the rec12-D15 (null) and rec12-Y98F (active site) mutants manifest at most a two-fold difference in the absolute frequency of meiotic diploids exhibiting MII missegregation patterns (Figures 4, 5, Table 2). Catalytically-inactive Rec12-Y98F protein apparently cannot enhance the fidelity appropriate MII segregation, relative to MII missegregation levels in cells lacking Rec12 entirely. We therefore propose that Rec12 protein mediates a signal, strictly dependent upon its active site tyrosine, that is required for the equational division of MII.

There are two main hypotheses as to the nature of the signal: First, Rec12 protein itself may participate in the signal transduction. In pat1-114 ^ts meioses Rec12 appears to persist until about the time of the second meiotic division (N. Herrera and S. Forsburg, personal communication), so Rec12 is available at the right time. Alternatively, some aspect of the crossovers or chiasmata may transduce the signal. This could involve structural components that persist on chromosomes at the positions where recombination and chiasmata had previously occurred, or could involve a diffusible component generated by recombination events.

At present, two hypothetical mechanisms by which a Rec12-dependent signal affects MII are proposed: First, Rec12-dependent (recombination-dependent) signals might trigger a block after MI, thereby allowing cells to skip a frequently irregular MII division [39]. However, the majority of the rec12 mutant meioses show evidence of two meiotic divisions (Figure 3) and many of the meiotic diploids exhibit mixed MI and MII missegregation patterns for centromere linked markers on the three different chromosomes (Table 2). In the absence of PSS, which does not occur at significant levels in rec12 mutants [52], such mixed segregation patterns can only be generated by a combination of errors in both reductional and equational segregation. A second hypothetical mechanism takes into account the coordinate, two-step release of sister chromatid cohesion that is essential for the two meiotic divisions [38]. Rec12-dependent (or recombination-dependent) signals may be required for the proper maintenance and/or dissolution of centromere-proximal sister chromatid cohesion in MII. These and other hypotheses regarding the unambiguous requirement for Rec12 in MII remain to be tested.

S. pombeculture

Intron mapping, cDNA cloning, and DNA sequence analysis

Cells harboring the pat1-114 ^ts allele were induced to enter meiosis [56] and were collected after 3 hours of meiotic induction. Genomic DNA and total RNA were prepared [57] and subject to PCR and RT-PCR [58] using primers flanking each putative intron.

Full-length rec12⁺ cDNA was obtained using RT-PCR. The 5' and 3' primers contained NdeI and BamHI restriction sites, respectively, and were designed to amplify the rec12⁺ cDNA from the first ATG in exon 1 to a position +142 bp downstream of the stop codon in exon 5. RT-PCR products were cloned by blunt-end ligation into pCR-Blunt (Invitrogen Corp.) and both strands were subject to DNA sequencing (GenBank accession no. AF195027).

Alignment of Rec12 with other Spo11 family members

Protein sequences homologous to that of S. pombe Rec12 were identified with a NCBI Blast search [59] using matrix BLOSUM62. Representative homologous sequences (see Figure 1 legend) were aligned using T-COFFEE [60], which optimizes alignment of sequences with low levels of homology. Output was prepared using a 50% threshold for identical and conserved residues.

Inducible pREP vector constructions

Genomic rec12⁺ DNA and rec12⁺ cDNA were cloned between the NdeI and BamHI restriction sites of the inducible vectors pREP2, pREP42, and pREP82 [32]. Induction of gene expression and northern blot analysis of rec12⁺ mRNA levels were as described [32, 57].

Genomic DNA was liberated from pYL78 [30] by partial EcoRI digestion and complete BamHI digestion, which cleaves within the polylinker site approximately 700 bp downstream of the rec12⁺ stop codon in exon 5. The DNA fragment with single-site cleavage by EcoRI at bp position +3, relative to the ATG start codon, was gel purified. A dsDNA oligonucleotide linker, containing cohesive NdeI and EcoRI ends, was constructed (5'-TATGAGCCATCATCATCATCATCATAGCATG-3' and 5'-AATTCATGCTATGATGATGATGATGATGGCTAC-3'). This linker encoded in frame the amino acids MSHHHHHHSM, with the last residue corresponding to the first methionine encoded in rec12⁺ exon 1. Trimolecular ligation was used to create pREP2-rec12⁺, pREP42-rec12⁺, and pREP82-rec12⁺ clones encoding Rec12 protein with 6X histidine tags. An NdeI-BamHI fragment of a cDNA clone with wild-type rec12⁺ sequence (see above) was sub-cloned into the pREP42 vector to generate pREP42-rec12⁺cDNA. The integrity of all clones was determined by a combination of restriction mapping, PCR, and DNA sequencing.

Construction of rec12-D15::ura4⁺ and rec12-Y98Falleles

A PCR based gene targeting approach [61, 62] was used to delete the rec12⁺ coding region and replace it with the ura4⁺ gene. Candidates were screened with a combination of PCR analysis, restriction digestion, and DNA sequencing to identify those with successful allele replacement.

Replacement of the endogenous rec12⁺ locus with the rec12-Y98F allele was achieved by transformation and a pop-in, pop-out approach [61, 63]. A 1.2 kbp XbaI-XbaI fragment of rec12⁺ was subcloned into pBluescript SKII(+) and subject to oligonucleotide directed mutagenesis [64] to simultaneously introduce a silent mutation generating an EcoRV site and to mutate the Y98 codon (GAG ATA TTT ATT ACA → GAG ATA T CT ATT TCA). After confirmation by DNA sequencing, the XbaI-XbaI fragment was subcloned into pURA4. pURA4-rec12-Y98F was linearized by digestion with NcoI. Transformation, forward selection for Uracil prototrophy, and reverse selection for FOA resistance were as described [61, 63]. Candidates were screened as described above.

Recombinant frequency determination

Mating, meiosis, and preparation of free spores were as described [54, 55, 65]. Intergenic and intragenic recombinant frequencies were determined as described [38, 65]. Because diploid spores could contain complementing markers and be mistaken for recombinants, we tested all spore colonies for diploidy by replica plating on complete media containing Phloxin-B (YEA-B). On these plates, diploid cells produce dark pink colonies, whereas haploid cells produce light pink colonies [54]. Diploid meiotic products were excluded from recombinant frequency determinations.

Diploid spore isolation and haploidization analysis

Identification of diploid spore colonies and their haploidization with m-flourophenylalanine (m-FPA) [55] were as described [38]. Fifty haploidized colonies derived from each diploid spore colony were replica plated to differentially supplemented minimal media to genotype the lys1 and ade6 loci. The tps13 alleles were scored by replica plating colonies onto rich (YEA) media and testing for growth at permissive (22°C) and restrictive (35°C) temperatures.

Microscopy

Asci from meiotic cultures were fixed with 70% ethanol at -20°C for at least 15 minutes, washed with H₂O, and stained with 4,6-diamidino-2-phenylindole (DAPI) at a final concentration of 1 μg/ml. Cells were examined by differential interference contrast (DIC) and fluorescence (DAPI) microscopy with a Zeiss axiophot (Carl Zeiss, Thornwood, NY). Images were analyzed using the MetaMorph software package (Universal Imaging, West Chester, PA).