An Analysis of Univalent Segregation in Meiotic Mutants of Arabidopsis thaliana: A Possible Role for Synaptonemal Complex

Abstract

During first meiotic prophase, homologous chromosomes are normally kept together by both crossovers and synaptonemal complexes (SC). In most eukaryotes, the SC disassembles at diplotene, leaving chromosomes joined by chiasmata. The correct co-orientation of bivalents at metaphase I and the reductional segregation at anaphase I are facilitated by chiasmata and sister-chromatid cohesion. In the absence of meiotic reciprocal recombination, homologs are expected to segregate randomly at anaphase I. Here, we have analyzed the segregation of homologous chromosomes at anaphase I in four meiotic mutants of Arabidopsis thaliana, spo11-1-3, dsy1, mpa1, and asy1, which show a high frequency of univalents at diplotene. The segregation pattern of chromosomes 2, 4, and 5 was different in each mutant. Homologous univalents segregated randomly in spo11-1-3, whereas they did not in dsy1 and mpa1. An intermediate situation was observed in asy1. Also, we have found a parallelism between this behavior and the synaptic pattern displayed by each mutant. Thus, whereas spo11-1-3 and asy1 showed low amounts of SC stretches, dsy1 and mpa1 showed full synapsis. These findings suggest that in Arabidopsis there is a system, depending on the SC formation, that would facilitate regular disjunction of homologous univalents to opposite poles at anaphase I.

MEIOSIS is a specialized form of cell division that makes haploid gametes from diploid cells. Chromosome number is reduced because there is a single round of DNA replication followed by two consecutive rounds of chromosome segregation. In most organisms, alignment of homologs, genetic recombination, and synaptonemal complex (SC) assembly are the processes that link homologous chromosomes together during the first meiotic prophase (Roeder 1997). SC usually disassembles at diplotene and homologous chromosomes remain joined by chiasmata, stable connections between homologs formed at the sites of crossovers, which, in conjunction with sister-chromatid cohesion, promote the correct co-orientation of bivalents at metaphase I (MI) and the reductional segregation of homologs at anaphase I (AI). In the absence of meiotic reciprocal recombination, homologs are expected to segregate randomly at anaphase I. However, in certain organisms nonexchange chromosomes segregate properly in most meioses (Wolf 1994). Among them, those systems displayed by Drosophila melanogaster males and females, Saccharomyces cerevisiae, and Schizosaccharomyces pombe are well documented (Dawson et al. 1986; Guacci and Kaback 1991; Hawley and Theurkauf 1993; Molnar et al. 2001a,b; Sharif et al. 2002; McKim et al. 2002; Davis and Smith 2003).

In Drosophila females, two different achiasmate segregation systems are involved in the accurate segregation of chromosomes (Hawley 1989; Hawley and Theurkauf 1993; Hawley et al. 1993a,b). In the first segregation system, pairing and segregation of nonexchange chromosomes require both homology between centromere adjacent heterochromatin (Hawley et al. 1993a; Dernburg et al. 1996; Karpen et al. 1996) and the presence of the Nod protein (Carpenter 1973; Zhang et al. 1990; Theurkauf and Hawley 1992; Cui and Hawley 2005). In the second one, two achiasmate heterologous chromosomes also segregate from each other with high fidelity (Dernburg et al. 1996). The existence of these systems is relevant because during Drosophila female meiosis the small fourth chromosomes never recombine and the X chromosomes fail to exchange in 5% of meioses (Theurkauf and Hawley 1992). In Drosophila males, all chromosomes are achiasmate and the stable pairing requires a specific region of heterochromatin on the sex chromosomes and of euchromatin on the autosomes (see, for review, McKee 1998).

In S. cerevisiae, two nonexchange artificial chromosomes (Dawson et al. 1986) or two monosomic nonhomologous chromosomes (Guacci and Kaback 1991) segregate properly in ∼90% of meioses. On the other hand, in rec7, rec14, and rec15 mutants of S. pombe, which are devoid of meiotic recombination, each chromosome segregates properly >75% of the time (Molnar et al. 2001a,b). Furthermore, Davis and Smith (2003) verified that in rec12 mutants, in which DNA double-strand breaks are essentially eliminated, homologous achiasmate chromosomes segregate to opposite poles at anaphase I in 63% of cells. In these mutants, Meu13, a homolog of the S. cerevisiae Hop2 protein, contributes to the recombination-independent chromosome pairing observed (Nabeshima et al. 2001). Recently, Davis and Smith (2005) have demonstrated that Dlc1, a member of dynein light-chain family, is involved in the correct segregation of achiasmate chromosomes in S. pombe.

To address whether the plant model species Arabidopsis thaliana (2n = 10) possesses an achiasmate segregation system similar to those described above, we analyzed the segregation of chromosomes 2, 4, and 5 in four different meiotic mutants. All mutants can be characterized by a high frequency of univalents at diplotene metaphase I, namely: mpa1 (meiotic prophase aminopeptidase), dsy1 (desynaptic), asy1 (asynaptic), and spo11-1-3 (sporulation defective). The MPA1 gene encodes a puromycin-sensitive aminopeptidase that is essential for meiosis in Arabidopsis (Sánchez-Morán et al. 2004). The DSY1 gene has not been identified yet but synapsis appears normal in the mutant (Ross et al. 1997). On the contrary, ASY1, homologous to the HOP1 gene of S. cerevisiae, is known to be essential for normal synapsis (Caryl et al. 2000). Finally, SPO11 encodes a type II topoisomerase, which is likely required for initiation of meiotic recombination in all species (Keeney et al. 1997; Grelon et al. 2001).

MATERIALS AND METHODS

Plant material:

Four different meiotic mutants of A. thaliana (L.) (2n = 10), produced by insertion mutagenesis with a portion of the Ti plasmid of Agrobacterium tumefaciens (T-DNA), were used in this study: dsy1, mpa1, asy1, and spo11-1-3. The spo11-1-3 mutant, line Salk_009440, was obtained from the collection of T-DNA mutants of the Salk Institute Genome Analysis Laboratory (Col-0 accession) (SiGnAL, http://signal.salk.edu/cgi-bin/tdnaexpress) (Alonso et al. 2003) and provided by the Nottingham Arabidopsis Stock Centre (http://nasc.nott.ac.uk/). The dsy1 and asy1 mutants (Ws accession), previously described by Ross et al. (1997), were kindly provided by Gareth Jones (School of Biological Sciences, University of Birmingham, Birmingham, UK). The mpa1 mutant was isolated in our laboratory from Spanish T-DNA transformed lines. Plants were cultivated in greenhouses under a 16-hr-day/8-hr-night photoperiod at 20° and 70% humidity.

Cytology:

Fixation:

Flower buds in stages 9/10 and 10/11 of floral development (Smyth et al. 1990) were used to analyze metaphase I and second-division stages in pollen mother cells (PMCs) and metaphase I in embryo-sac mother cells (EMCs). They were fixed in Carnoy's fixative (absolute ethanol–chloroform–glacial acetic acid: 6:3:1) overnight at room temperature. The fixative was replenished and buds were stored at −20° until required.

Slide preparations:

Air-dried spreads of PMCs and EMCs were carried out according to Fransz et al. (1998) and Armstrong et al. (1998), with minor modifications as described by Sánchez-Morán et al. (2001). Fixed buds were transferred to a watch-glass containing absolute ethanol–glacial acetic acid fixative (3:1) that was changed twice to remove the original Carnoy's fixative. Buds were then washed twice in 10 mm citrate buffer, pH 4.5, at room temperature. Buffer was removed before incubation with an enzyme mixture of 0.3% w/v pectolyase, 0.3% w/v cytohelicase, and 0.3% w/v cellulose (Sigma, St. Louis) in citrate buffer for 2 hr at 37°. The enzyme mixture was replaced by ice-cold citrate buffer to stop the reaction. Single buds, immersed in a small volume of buffer, were transferred to slides with a Pasteur pipette and macerated with a needle. Afterward, 10 μl of 60% acetic acid was added to each slide, which was placed on a hot plate at 45° for 1 min, and another 10 μl when the slide was not on the hot plate. Preparations were refixed with 200 μl of cold 3:1 fixative. Once the fixative was removed, the slides were dried and stained with the fluorochrome 4′,6-diamidino-2-phenylindole (DAPI, 10 μg/ml) in Vectashield antifade mounting medium (Vector Laboratories) to choose those containing metaphase II (MII) cells.

Fluorescence in situ hybridization:

The fluorescence in situ hybridization (FISH) technique was performed according to Sánchez-Morán et al. (2001). The following DNA probes were used:

• Clone pTa71 (Gerlach and Bedbrook 1979): Containing a 9-kb EcoRI fragment of Triticum aestivum consisting of the 18S-5.8S-25S rRNA genes and the spacer regions. Digoxigenin dUTP was incorporated by nick translation following the manufacturer's instructions (Roche) and it was detected by FITC antibodies.

• Plasmid pCT4.2: Containing the 5S rRNA gene from A. thaliana as a 500-bp insert cloned in pBlu. Biotin dUTP was also incorporated by nick translation and detected by avidin-Cy3 antibodies.

Preparations were counterstained with DAPI in Vectashield antifade mounting medium. Chromosomes were examined by using epifluorescence microscopy (Olympus BX60) and recorded with an Olympus DP70 camera. Figures were composed using PHOTOSHOP 7.0 software (Adobe Systems, San Jose, CA).

Synaptonemal complex preparations:

Surface spreads of meiocytes were performed according to Albini (1994) with minor modifications. Briefly, fresh unfixed anthers in stage 9 of floral development (Smyth et al. 1990) were squashed onto a cavity slide with a drop of digestion medium [0.01 g snail gut enzyme cytohelicase (Sigma) 0.125 g polyvinyl pyrrolidone and 0.19 g sucrose in 12.5 ml sterile distilled water]. After ∼5 min of enzymatic treatment, a single drop of meiocyte suspension was transferred into 0.05% Triton X-100 detergent on slides. The meiocytes were allowed to spread for ∼5 min and then about five drops of the fixative solution (4% paraformaldehyde and 1.7% sucrose in distilled water, adjusted to pH 8.9 with borate buffer) were added. Slides were dried overnight on a warm plate at 30°, rinsed in distilled water for 1 min at room temperature, and air dried. For silver impregnation, a few drops of aqueous AgNO₃ (30%) were placed on the preparations, which were then covered with a patch of nylon cloth at 30°–40° until they turned into a yellow color (∼45 min). Once the mesh was removed, the slides were rinsed in distilled water and dried. Prophase I nuclei were examined under an electron microscope (Jeol 1010).

RESULTS

FISH identification of Arabidopsis meiotic chromosomes:

In this work, we have applied FISH, using 5S and 45S rDNA probes, to distinguish all of the five chromosome pairs of the Arabidopsis complement (Fransz et al. 1998; Sánchez-Morán et al. 2001, 2002). The acrocentric chromosomes 2 and 4 bear 45S sequences on their short arms (corresponding to the location of the nucleolus organizing regions), whereas 5S sequences are located proximally on the short arms of chromosomes 4 and 5. Furthermore, in the Col-0 accession there is a 5S locus in the short arm of chromosome 3 (Figure 1, A and B). Although chromosomes 1 and 3 differ in size, this difference is not appreciable at metaphase II cells because chromosomes are extremely condensed. On these grounds, this research will be focused on the meiotic segregation patterns of chromosomes 2, 4, and 5 in four meiotic mutants. It must be noted that this analysis has been carried out, except for the mutant spo-11-1-3, in meiocytes from flower buds of the same dsy1, asy1, and mpa1 plants employed in previous studies (Sánchez-Morán et al. 2001, 2004).

Figure 1.—

FISH images of metaphase I cells from wild-type Arabidopsis plants, accessions Col-0 (A) and Ws (B). 45S rDNA loci and 5S rDNA loci are identified by green and pink signals, respectively. Bars, 10 μm.

Segregation analyses:

The spo11-1-3 mutant:

This meiotic mutant, denominated spo11-1-3, arose as consequence of a T-DNA insertion in the SPO11 gene and corresponds to the line Salk_009440. Figure 2 shows the location of the T-DNA insertions in spo11-1-1 (Grelon et al. 2001) and spo11-1-3 mutants and the location of the point mutation produced by treatment with ethyl methanesulfonate (EMS) in spo11-1-2 (Grelon et al. 2001).

Figure 2.—

Genomic organization of the AtSPO11-1 locus and localization of spo11-1 mutations. Solid boxes represent exons and shaded lines represent 5′ and 3′ untranslated regions. The locations of the T-DNA insertions in spo11-1-1 (Grelon et al. 2001) and spo11-1-3 (Salk_009440) and the point mutation produced by EMS in spo11-1-2 (Grelon et al. 2001) are indicated by vertical arrows.

SC formation was very limited in the 20 zygotene–pachytene nuclei analyzed. Indeed, the overall amount of synapsis never exceeded 10% of the total axial element length (Figure 3A). Consequently, a total absence of bivalents at MI was the main feature of both male and female meiosis (200 and 20 meiocytes analyzed, respectively). Univalents always moved entirely to one of the anaphase I poles (reductional segregation). However, some chromosome segregation abnormalities must occur at the first division because abnormal metaphase II cells with three (13.87%: 19/137) and four (3.65%: 5/137) separate groups of chromosomes were observed (supplemental Figure 1S at http://www.genetics.org/supplemental/). Canonical meiosis II PMCs, mostly metaphase II pairs, were classified in two groups depending on whether homologous chromosomes migrated to different poles at anaphase I (1:1) or not (2:0), and this was performed separately for each pair of the chromosomes analyzed (see Figure 4,A and B, as examples of regular disjunction and nondisjunction, respectively). Assuming random chromosome segregation, we would expect 50% MII cell pairs of each type. Observed and expected values are shown in Table 1. Chi-square analyses did not reveal significant differences for any of the chromosomes studied.

Figure 3.—

Electron micrographs of silver-stained zygotene–pachytene nuclei. (A) spo11-1-3. (B) dsy1. (C) mpa1. (D) asy1. Arrows in A and D indicate short SC stretches that are enlarged. Bars, 5 μm.

Figure 4.—

Examples of regular disjunction (A, C, E, and G) and nondisjunction (B, D, F, and H) of chromosome pairs 2, 4, and 5 in meiosis II cells. (A and B) spo11-1-3. (C and D) dsy1. (E and F) mpa1. (G and H) asy1. Individual chromosomes bearing 45S and 5S rDNA sequences are identified. (D) Telophase I. (A) Prophase II. (B, C, E, F, and H) Metaphase II. (G) Anaphase II. Bars, 10 μm.

TABLE 1.

Mean cell bivalent frequencies and segregation patterns displayed by chromosomes 2, 4, and 5 in the four meiotic mutants analyzed

	Chromosome no.	Mean cell bivalent frequency	Regular disjunction 1:1		Nondisjunction 2:0		No. MII cell pairs
Mutant					Observed	Expected		Observed	Expected		χ21
spo11-1-3	2	∼0	50	57.00	64	57.00	114	1.72
	4	∼0	56	57.00	58	57.00	114	0.04
	5	∼0	56	57.00	58	57.00	114	0.04
dsy1	2	0.06	59	42.93	22	38.07	81	12.80*
	4	0.10	55	37.40	13	30.60	68	18.40*
	5	0.12	66	47.04	18	36.96	84	17.37*
mpa1	2	0.18	44	33.63	13	23.37	57	7.80*
	4	0.07	39	28.35	14	24.65	53	8.60*
	5	0.15	43	30.47	10	22.53	53	12.11*
asy1	2	0.36	75	63.92	19	30.08	94	6.00*
	4	0.24	66	58.90	29	36.10	95	2.25
	5	0.25	60	55.63	29	33.37	89	0.92

Observed and expected values of regular disjunction (1:1) and nondisjunction (2:0) for each chromosome pair in MII cells. Expected values were calculated assuming random chromosome segregation at anaphase I.

^*Significant differences at 5% level.

The dsy1 mutant:

Synapsis appeared to proceed normally until completion at pachytene (25 nuclei analyzed; Figure 3B), but at diplotene most homologs were not associated by chiasmata. The mean cell bivalent frequency at metaphase I was 0.51 with an overall mean cell chiasma frequency of 0.57.

Those MII PMCs containing chromatids (0, 1.45, and 1.18% for chromosomes 2, 4, and 5, respectively) were excluded from the segregation analysis. This finding, in addition to the fact that some chromosome pair(s) could not be accurately identified in some cells, explains the different number of MII cells analyzed for each chromosome (Table 1). In the segregation analysis, we have taken into account the mean cell bivalent frequency of each chromosome at MI since bivalent formation implies that homologs will migrate to opposite poles at AI with a very high probability. This is the reason because the expected values under the random homologous segregation hypothesis do not correspond with 50% of MII cells analyzed. Chi-square analyses revealed that univalents for chromosomes 2, 4, and 5 segregate to opposite poles at AI with a frequency significantly higher than that expected by random (Table 1; Figure 4, C and D).

The mpa1 mutant:

Although five apparent normal SCs were always observed in each of the 20 pachytene nuclei analyzed (Figure 3C), univalents were predominant at diplotene. The mean cell bivalent frequency at MI was 0.75, with an overall mean cell chiasma frequency of 0.9.

The segregation analysis revealed results similar to those obtained in the dsy1 mutant; i.e., chromosomes 2, 4, and 5 displayed significantly higher frequencies of regular disjunction than those expected by random (Table 1; Figure 4, E and F).

The asy1 mutant:

At zygotene–pachytene, axial/lateral elements showed gaps, indicating a certain fragmentation throughout their length. A small amount of SC stretches, never exceeding the 30% of the total axial element length, was observed in some of the 20 nuclei analyzed (Figure 3D). The mean cell bivalent frequency was 1.27, with an overall mean cell chiasma frequency of 1.39.

Those MII PMCs containing chromatids as result of univalent equational segregation at AI (10.48, 12.04, and 11.00% for chromosomes 2, 4, and 5, respectively) were excluded from the segregation analysis. The results of such analysis, bearing in mind the mean cell bivalent frequency for each chromosome, are shown in Table 1. Examples of regular homologous disjunction and nondisjunction are showed in Figure 4, G and H. Whereas univalents for chromosome pair 2 segregated to opposite poles with a frequency significantly higher than that expected by random, those univalents for chromosomes 4 and 5 did not.

DISCUSSION

Although reciprocal recombination is generally required for proper homologous segregation at AI, some organisms possess systems capable of partitioning one or more pairs of homologs when they appear as univalents at MI. This is the case with Drosophila males and females in which different mechanisms are called upon to segregate achiasmate chromosomes in every meiosis (Hawley and Theurkauf 1993; McKee 1998). Similar systems, although not so efficient, have been described in S. cerevisiae and S. pombe (Dawson et al. 1986; Guacci and Kaback 1991; Davis and Smith 2003). It must be noted, however, that mutations that abolish meiotic recombination in Drosophila females display significantly increased levels of AI nondisjunction, perhaps because the system cannot accommodate the whole chromosome complement (Hawley 1989). Likewise, S. cerevisiae is unable to promote nonrandom segregation when meiotic recombination is abolished throughout the genome. It is known that chromosome V segregates at random in spo11 mutant meioses (Klein et al. 1999). On the contrary, in recombination-defective mutants of S. pombe, reductional segregation of chromosomes at AI is nonrandom (Molnar et al. 2001a,b; Davis and Smith 2003).

The results obtained in the univalent segregation analysis carried out in the Arabidopsis mutants are intriguing:

1. There is not an apparent relationship between the segregation behavior of univalents and their synaptic history. Thus, reductional segregation is the rule in asynaptic univalents of spo11-1-3 and desynaptic univalents of dsy1. On the other hand, in asy1, each pair of asynaptic univalents displayed ∼11% of equational segregation. The fragmentation of the axial elements observed in zygotene–pachytene nuclei of asy1 (Figure 3D) concurs with the hypothesis that Asy1 is required for morphogenesis of the SC, possibly by defining regions of chromatin that associate with the developing SC structure (Armstrong et al. 2002). Perhaps this protein is also involved in some aspect(s) of chromosome reorganization that predispose chromosomes to unipolar orientation.

2. Whereas chromosome pairs 2, 4, and 5 segregated randomly at AI in spo11-1-3, they did not in dsy1 and mpa1. An intermediate situation was observed in asy1, in which chromosome pairs 4 and 5 segregated at random and those of chromosome 2 migrated preferentially to opposite poles (Table 1, Figure 4). These differences in chromosome behavior among mutants are also accompanied by differences in SC formation (Figure 3), in such a way that a relationship between presence of complete SC at pachytene and regular disjunction of homologs at AI can be established.

The existence of a low amount of SC formation in spo11-1-3 is perhaps surprising because in Arabidopsis recombination is a prerequisite for synapsis. Therefore, one would expect that total absence of recombination would lead to complete absence of synapsis (Grelon et al. 2001). Whether this synapsis is between nonhomologous chromosomes or whether it is independent of DNA double-strand break (DSB) formation remains to be established. Likewise, nuclei with full synapsis have been observed in spo11 mutants of budding yeast (Bhuiyan and Schmekel 2004). Random chromosome segregation has also been claimed to exist in the almost sterile dmc1 mutant of A. thaliana (Couteau et al. 1999); however, the authors stated that they did not detect proper metaphase II cells in the DAPI-stained slides and also did not identify any of the univalent chromosome pairs. Unfortunately, no data about SC formation were provided. The different behavior of univalents in asy1 might be explained, assuming that chromosome 2's are more actively involved in SC formation than chromosomes 4 and 5 (Figure 3D) and that this would facilitate their significant nonrandom regular disjunction at AI. Indeed, in this mutant, chromosome 2 exhibits the highest bivalent and chiasma frequencies of the set, the reverse situation of that found in dsy1 and in wild type (Sánchez-Morán et al. 2001).

It is probable that spindle checkpoint genes (Cheslock et al. 2005) and SC components play a role in the segregation of the univalents at the first meiotic division. From pachytene onward, some SC components might secure homologous centromeres with the active faces of their kinetochores facing away from one another in the meiotic spindle. In fact, in the desynaptic mutants, mpa1 and dsy1, homologous univalents maintain a conspicuous parallel disposition throughout diplotene diakinesis that could facilitate the adequate interactions between kinetochores and microtubules (supplemental Figure 2S at http://www.genetics.org/supplemental/). In budding yeast, SC does not seem critical for nonexchange segregation because the segregation of achiasmate homologs was not randomized by the absence of SC in zip1 mutants (Kemp et al. 2004). However, the SC protein Zip1 may play a role in coupling yeast centromeres that would facilitate homologous pairing and subsequent synapsis initiation (Tsubouchi and Roeder 2005). Meiotic centromere pairing of nonexchange chromosomes, actually of the heterochromatic adjacent regions, seems to be sufficient to establish an initial bipolar attachment of microtubules following segregation of chromosomes to opposite poles (Kemp et al. 2004; Stewart and Dawson 2004). In Arabidopsis, initial localization of Zyp1 is dependent upon DSB formation and occurs when recombination is at an early stage, prior to extensive strand invasion. However, subsequent Zyp1 polymerization to bring about homologous chromosome synapsis cannot occur if recombination is blocked at an early step (Higgins et al. 2005). On the other hand, in some meiotic systems the segregation of achiasmate chromosomes is mediated by modified SC structures that persist after SC breakdown (Rasmussen 1977; Wolf 1994). In Drosophila females, the Mtrm protein (matrimony) might act by modifying some component of the SC that originally connected the heterochromatic regions of chromosomes during meiotic prophase. These modifications could play an important role in the achiasmate chromosome segregation by holding homologs together long enough to ensure their orientation to opposite poles (Harris et al. 2003). In mouse, a meiosis-specific structural protein, Sycp3, which appears in axial and lateral elements of the SC, is involved in the monopolar attachment of sister kinetochores during the first meiotic division (Parra et al. 2004). Recently, Page et al. (2003, 2005) have reported that in marsupial males the pairing of the achiasmatic X and Y chromosomes during meiotic prophase is maintained by a dense plate developed from their axial elements.

At present, the real role of the SC remains enigmatic. However, its ubiquitous presence and its structural conservation throughout different organisms reflect the importance of this proteinaceous structure in the meiotic process. This study suggests that in A. thaliana there is a SC-dependent system that promotes the accurate segregation of univalents to opposite poles at AI. It is less efficient than those described in Drosophila males and females and does not seem to specifically ensure the segregation of nonexchange chromosomes, since the chromosomes of A. thaliana do not generally fail to recombine and form five bivalents at meiosis. The further characterization of SC proteins and spindle checkpoint genes in plants may lead to a better understanding of this segregation system.