The Arabidopsis thaliana ASY1 gene is essential for homologous chromosome synapsis. Antibodies specific to Asy1 protein and its homologue BoAsy1 from the related crop species Brassica oleracea have been used to investigate the temporal expression and localization of the protein in both species. Asy1 is initially detected in pollen mother cells during meiotic interphase as numerous punctate foci distributed over the chromatin. As leptotene progresses the signal appears to be increasingly continuous and is closely associated with the axial elements but not to the extended chromatin loops associated with them. By the end of zygotene the signal extends almost the entire length of the synapsed homologues, although not to the telomeres. The protein begins to disappear as the homologues desynapse, until by late diplotene it is no longer associated with the chromosomes. Immunogold labelling in conjunction with electron microscopy established that Asy1 localizes to regions of chromatin that associate with the axial/lateral elements of meiotic chromosomes rather than being a component of the synaptonemal complex itself. These data together with the previously observed asynaptic phenotype of the asy1 mutant suggest that Asy1 is required for morphogenesis of the synaptonemal complex, possibly by defining regions of chromatin that associate with the developing synaptonemal complex structure.
Meiosis occupies a central role in the life cycles of all sexually reproducing eukaryotes, and it is through this process that eukaryotes achieve chromosome reduction, ordered chromosome segregation into the haploid products and recombination of entire chromosomes and chromosome segments. The biochemical processes that result in meiotic homologous recombination are set in a context of orchestrated chromatin rearrangements that are essential for the completion of the meiotic sequence. Following premeiotic S phase, the homologous chromosomes condense, and pairing is initiated. In some species at least, preliminary pairing events are detectable in meiotic interphase that have been defined as homologue cognition and alignment ( Schwarzacher, 1997), although these are apparently not detected in some other species ( Dawe et al., 1994; Scherthan et al., 1998). Ultimately, in all species, homologous chromosomes are juxtaposed in an intimate synaptic association along their entire length. Running between the synapsed chromosomes is a proteinaceous structure, the synaptonemal complex (SC) ( Moses, 1968; Von Wettstein et al., 1984), whose structure appears to be evolutionarily highly conserved. Detailed microscope studies in a range of higher eukaryotes have revealed that during early prophase I, prior to SC polymerization, each homologous chromosome, comprising two sister chromatids, is organized into a stacked parallel array of loops ( Zickler and Kleckner, 1999) around an axial element or core. Subsequently, during the zygotene stage of prophase I, the central region of the SC assembles between the axes, now called lateral elements, resulting in synapsis of the homologues. The central region is composed of a series of numerous transverse filaments that lie perpendicular to the two lateral elements and a central element that runs between and parallel to the lateral elements ( Heyting, 1996; Zickler and Kleckner, 1999).
Despite a range of proposals, the real function of the SC is as yet unresolved. Nevertheless its ubiquitous occurrence, its remarkable evolutionary structural conservation and its temporal and spatial association with chromosome synapsis and at least some events of homologous recombination highlight the central importance of the SC in the meiotic process. One approach to dissecting its function (or functions) is to analyze the composition and interactions of SC proteins, and also SC-associated proteins, and to investigate the structure and regulation of the genes that encode them. Another important goal is to understand how these proteins interact with DNA and chromatin.
In recent years studies in yeast, and mammals in particular, have produced considerable progress in the identification of genes encoding SC and SC-associated proteins (reviewed in Heyting, 1996; Roeder, 1997; Zickler and Kleckner, 1999). However in the case of higher plants, little progress has been made towards identification of SC genes. A major barrier is that despite the availability of the Arabidopsis thaliana genome sequence and the apparent structural conservation of the SC, the SC proteins themselves and those associated with them exhibit a high degree of sequence divergence. For example Zip1 from yeast and Scp1 from rats are structurally similar transverse filament proteins, yet they exhibit no sequence homology other than that expected for two coiled-coil proteins ( Heyting, 1996). We have therefore used a different route to identify candidate genes encoding SC proteins. Arabidopsis T-DNA insertion lines were screened for lines exhibiting reduced fertility. Meiotic mutants were then confirmed by cytological analysis of pollen mother cells ( Ross et al., 1997, Ross et al., 1997). One such mutant, asy1, exhibits an asynaptic phenotype and dramatically reduced chiasma formation. Subsequently the ASY1 gene was cloned and found to encode a 596 amino-acid polypeptide with some similarity to the N-terminus of the yeast axial core-associated protein Hop1 ( Caryl et al., 2000). The two proteins exhibit 28% identity and 51% similarity over the first 250 amino acids. This corresponds to a HORMA domain (Hop1, Rev7 and MAD2) ( Aravind and Koonin, 1998), a sequence that is found in a number of proteins that interact with chromatin, including Him-3, a Caenorhabditis elegans protein that also encodes a component of the meiotic chromosome core ( Zetka et al., 1999).
In this study we have used an antibody raised against recombinant Asy1 to carry out a detailed analysis of spatial and temporal expression of Asy1 throughout meiosis. These studies have been carried out in both the model plant species Arabidopsis and the related species Brassica oleracea, which is not only an important corp but is technically more amenable to detailed cytological analyses ( Armstrong et al., 1998). This has enabled us to establish that ASY1 encodes a protein that interacts with chromatin associated with the chromosome axes. We also explore the extent to which the predicted HORMA domain in Asy1 is indicative of a functional similarity with other meiotic genes known to possess this structural feature.
Materials and Methods
Plants used in this study were Arabidopsis thaliana accessions Columbia and Wassilewskija and Brassica oleracea var. alboglabra A12DHd [a genetically homozygous line derived from B. oleracea var. alboglabra by spore culture ( Bohoun et al., 1996)]. Plants were grown in multipurpose compost in a greenhouse under supplementary light (16 hours light, 8 hours dark).
Nucleic acid isolation
Genomic DNA was isolated using the Nucleon Phytopure (Amersham Pharmacia Biotech). Total RNA was isolated using the Rneasy Plant Mini Kit (Qiagen).
Nucleic acid hybridization
Capillary blotting was used to transfer 5 μg of digested genomic DNA onto Hybond N+ membrane (Amersham Pharmacia Biotech) and fixed by baking at 80°C for 2 hours. Southern blots were hybridized using Rapid-hyb (Amersham Pharmacia Biotech) and probed with radiolabelled DNA ( Feinberg and Vogelstein, 1983; Feinberg and Vogelstein, 1984). The blots were then washed at high stringency (0.2× SSC, 0.1% SDS 65°C). Radioactivity was detected by autoradiography. The probe used was the 2145 by Asy1 cDNA (GenBank accession AF157576) minus the first 61 bp.
PCR and RT-PCR analysis
The Omniscript RT kit (Qiagen) was used to synthesize cDNA from B. oleracea meiotic bud RNA using an Oligo-dT(16) primer. Cloned Pfu DNA polymerase (Stratagene) was then used to amplify part of the BoASY1 gene using primers FP2 and R1 (designed to the Arabidopsis ASY1 gene). The PCR products were cloned into the pZErO-2 vector (Invitrogen) and sequenced. This sequence data was used to design BoASY1-specific primers for RACE PCR. The cDNA for the RACE PCR was prepared using Superscript II reverse transcriptase (Invitrogen), and the PCR amplifications were carried out with Super Taq DNA polymerase (HT Biotechnology Ltd). 5′ RACE was carried out using AUAP (Life Technologies) with BoASY1-pecific nested primers R2 and R3. The cDNA was tailed with dCTP using Terminal Transferase (Roche). The 3′ RACE used R4 together with BoASY1 specific nested primers F2 and F3. PCR products were cloned into the pCR2.1 vector (Invitrogen) and sequenced.
R3: TTCACCCTTTCTAGCTG; AUAP: GGCCACGCGTCGACTAGTAC;
F2: CTTTTCTGACTTCATTTAGCTT; F3 CTGACAGCCAAGATGTCATG.
Nucleic acid sequencing
Nucleic acid sequencing was carried out using the BigDye Terminator Sequencing v2.0 Ready Reaction kit (PE Biosystems). Sequencing reactions were run out on an ABI 3700 DNA analyzer by the Functional Genomics Unit, University of Birmingham, UK.
The coding region of the Arabidopsis ASY1 gene was cloned into the protein expression vector pGEX-6P-1 (Amersham Pharmacia Biotech) as an N-terminal fusion to glutathione S-transferase. Upon induction, the GST-Asy1 fusion protein accumulated as insoluble inclusion bodies in E. coli SK1592. Purified, refolded recombinant protein was prepared as described previously ( Kakeda et al., 1998). Rabbit polyclonal antiserum was produced against the GST-Asy1 fusion protein (ISL, Poole, UK). The antibody was found to be specific for the Asy1 component of the GST-Asy1 fusion protein and did not cross-react with the GST component.
Plant proteins were extracted in 1% SDS, 5% β-mercaptoethanol and 50 mM Tris-HCl, pH 7.5, containing 1 Complete Mini protease inhibitor cocktail tablet (Roche) in 10 ml extraction buffer), and insoluble material was removed by centrifugation. Protein samples were separated by SDS-PAGE and electroblotted onto Hybond C extra nitrocellulose membrane (Amersham Pharmacia Biotech). Western blots were incubated with anti-Asy1 antiserum (1:2000 dilution) followed by anti-rabbit antibodies conjugated to horseradish peroxidase (1:5000 dilution) (Sigma-Aldrich). Protein bands were visualized using ECL reagents (Amersham Pharmacia Biotech) and detected by autoradiography.
Preparation of spreads and sections
Individual flower buds were dissected out from one inflorescence at a time, of either Brassica or Arabidopsis, and ordered by size in a Petri dish containing damp filter paper. Only the smallest buds were included, and any buds containing yellowish anthers, indicating the presence of pollen, were discarded. The relationship of meiotic stage to floral morphology has been described earlier ( Armstrong and Jones, 2001; Armstrong et al., 2001). Single anthers from individual buds were assessed for their meiotic stage after staining and lightly squashing the PMCs in acetic orcein. The remaining five anthers, from buds at appropriate meiotic stages, were placed in a cavity slide, and a small volume (20 μl for Brassica, 5 μl for Arabidopsis) of enzyme digestion mixture was added. The digestion mixture includes 0.1 g cytohelicase, 0.0375 g sucrose, 0.25 g polyvinylpyrollidone MW 40,000 (all Sigma) ( Albini et al., 1984). Anthers were incubated in the digestion mixture for 8 minutes (Brassica) or 2 minutes (Arabidopsis) at 37°C in a moisture chamber. After this time the anthers were gently tapped out in the digestion mixture using a brass rod to release PMCs, and anther wall debris was removed from the suspension thus created.
For immunofluorescence LM analysis, subsequent steps were performed using good quality pre-cleaned glass slides. For immunogold EM analysis, the slides were pre-coated with a plastic film and glow-discharged to make them hydrophilic. Slides were dipped one at a time into a solution of 0.75% plastic Petri dish (Greiner) dissolved in chloroform and left in an upright position to dry. The slides were then glow-discharged by exposure to UV light at 10-1 torr.
To prepare the spreads, 10 μl of Lipsol spreading medium (1% Lipsol detergent in freshly distilled water buffered to pH 9.0 with borate buffer) was drawn up with a pipette. The liquid was then expelled such that the drop was retained at the end of the pipette tip. 1-2 μl of the cell suspension was then drawn up using a separate pipette, and this was combined with the Lipsol drop. The mixture was expelled onto a treated slide. The cells were then monitored by phase contrast microscopy. When the chromatin started to spill out from PMCs, 10 μl of fixative (4% paraformaldehyde pH 8.0) was added. The slide was then allowed to dry in a fume-hood. Wax-embedded sections of anthers were prepared as described previously ( Armstrong and Jones, 2001).
Slides were first immersed in wash (PBS + 0.1% Triton) for 2×5 minutes. To block non-specific antibody binding (optional depending on background levels), 100 μl of blocking buffer (1% BSA in PBS) was applied directly to the slides, covered with parafilm and incubated at room temperature for 45 minutes. 100 μl of primary (anti-Asy1) antibody, diluted 1:500 in blocking buffer (PBS + 0.1% Triton + 1% BSA) was applied directly to the slides, covered with parafilm and incubated overnight at 4°C in a moisture chamber. The slides were then washed (2×5 minutes) before adding 100 μl of the secondary antibody (anti-rabbit FITC, Sigma, 1:50 in blocking buffer) and incubating for 90 minutes at room temperature. Finally the slides were washed for 2×5 minutes, mounted in DAPI (10 μg/ml) in Vectashield antifade mounting medium. Slides were examined by fluorescence microscopy using a Nikon Eclipse T300 microscope. Capture and analysis of images was achieved using an image analysis system (Applied Imaging).
Immunogold (electron microscopy) localization
Suitably spread PMC nuclei on plastic-coated slides were located by phase contrast microscopy and their positions marked. The unstained spreads on plastic film were then transferred to nickel electron microscope (EM) grids by flotation on a clean water surface. After picking up the grids, and allowing them to dry, the locations of spreads on the grids was once more checked by phase contrast microscopy. The following immunological and staining steps were all carried out by placing the grids on drops of reagent on a hydrophobic plastic surface. The grids were first taken through two changes (2×5 minutes) of PBS. Grids were next transferred to primary (anti-Asy1) antibody diluted 1:500 in blocking buffer (as for LM) and incubated overnight at 4°C, then washed (3×5 minutes). The grids were then incubated with a secondary antibody for EM (goat anti-rabbit IgG [BB International] conjugated to 5 nm gold particles 1:50 dilution) for 90 minutes at room temperature, followed by three washes (3×5 minutes) and 2×5 minutes washes in distilled water. Finally the grids were stained using the most appropriate method of the following three different procedures: (1) Uranyl acetate (30%); incubation for 7 minutes at room temperature followed by washing in methanol and then distilled water; (2) osmium tetroxide (0.1% in water) - incubation for 30 minutes at room temperature, followed by washing in distilled water; (3) phosphotungstic acid (1% in absolute ethanol), incubation for 10 minutes at room temperature. Overall uranyl acetate and phosphotungstic acid resulted in a higher level of background chromatin staining, but despite this, phosphotungstic acid was preferable for staining the SC. Osmium tetroxide did not produce as much background staining, but provided only poor definition of the SC. The stained grids were examined using a Jeol 1200Ex electron microscope.
Identification and cloning of BoASY1, a Brassica oleracea orthologue of ASY1
Cytological procedures have been developed that have improved the analysis of meiotic chromosomes in Arabidopsis (Ross et al., 1996). Nevertheless such analyses remain technically challenging, primarily because of the small size of the floral structures and to a lesser extent the small physical size of the Arabidopsis chromosomes. As a member of the Cruciferae family, Brassica oleracea is closely related to Arabidopsis, and mapping and sequencing studies have confirmed that there exists a high degree of synteny and sequence conservation between the two species ( O'Neill and Bancroft, 2000). It is therefore of interest to determine if B. oleracea possesses an ASY1 orthologue and if so, the level of conservation. The underlying rationale of this approach is that if the B. oleracea gene exhibits a high degree of homology to ASY1 then it should be feasible to use an anti-Asy1 antibody to investigate spatial and temporal expression of the protein in both species. This capitalizes on the fact that it is far easier to isolate anthers at specific meiotic stages from B. oleracea than it is from Arabidopsis and that it also possesses a significantly larger genome.
Initially, a Southern blot of B. oleracea genomic DNA was prepared and probed at high stringency with the ASY1 coding region. This confirmed the presence of closely related, low copy number sequences within the Brassica genome (data not shown). A PCR-based approach was then used to clone the putative B. oleracea orthologue of ASY1. Arabidopsis gene-specific primers were used to amplify the coding region of BoASY1 from B. oleracea flower bud cDNA. Brassica gene-specific primers were then designed on the basis of this sequence and used in RACE PCR to identify the 5′ and 3′ ends of the gene. Finally, the sequences were assembled to produce a full-length BoASY1 cDNA sequence of 2182 bp (Accession: AF410429) ( Fig. 1A).
A comparison of BoASY1 and ASY1 showed that the genes were 87% identical across the full length of the cDNA-coding region (position 102 to 1904). Sequencing of B. oleracea genomic DNA corresponding to this region revealed that BoASY1 is composed of at least 21 exons. This is the same as the coding region of ASY1 (which has 22 exons in total, one comprising the 5′ UTR). The similarity is reflected when the BoAsy1 and Asy1 proteins are examined. Both are almost identical in length (599 and 596 amino acids respectively) with a similar pI (5.2 and 5.04 respectively) and molecular weight (67.98 kDa and 67.3 kDa respectively). The proteins share 83% identity and 89% similarity across the full length of the sequences ( Fig. 1B), with conservation highest at the N-terminus.
This region includes a HORMA domain ( Aravind and Koonin, 1998) extending across the first 250 amino acids of the protein. There are two potential in-frame translation start codons separated by 3 bp, of which only the first has the Kozak consensus sequence ( Kozak, 1995) necessary for efficient translation initiation.
Asy1/BoAsy1 is expressed in early meiotic cells in both Arabidopsis and Brassica
To begin the analysis of Asy1 function, an antibody was raised against the full-length recombinant protein expressed in E. coli. The expression of Asy1 and BoAsy1 was then determined in protein extracts from Arabidopsis flower buds and B. oleracea anthers collected at different stages of meiosis. We have previously reported that there is a clear relationship between meiotic progression in pollen mother cells and flower bud length in Arabidopsis ( Armstrong et al., 2001). Protein extracts were therefore prepared from bud size series taken from a single inflorescence representing pre-meiosis through to the tetrad stage. In the case of Brassica, protein extracts were prepared from meiotically staged anthers.
Western analysis of Arabidopsis bud extracts revealed expression of a single protein with a molecular mass of slightly below 69 kDa, which is in agreement with a predicted size for Asy1 of 67.3 kDa. Expression was first detected in buds at a size corresponding to early male meiosis, the signal then gradually decreased in older buds as they approached the tetrad stage ( Fig. 2A). Interestingly there was a slight increase in band intensity in the extract prepared from buds at the tetrad stage. We believe that this reflects the asynchrony between male and female meiosis in Arabidopsis that we have previously reported ( Armstrong and Jones, 2001). Male meiosis initiates at an earlier stage of bud development so that by the time the PMCs have reached the tetrad stage, the embryo sac mother cells are still only in prophase I.
Western blots also revealed that BoAsy1 is expressed during early meiosis in B. oleracea, and because cytologically staged Brassica anthers were analyzed, expression could be more directly related to the meiotic stage of the material ( Fig. 2B). This analysis revealed that the Asy1 protein begins to accumulate during meiotic interphase. It increases, peaking at leptotene before gradually decreasing towards the later stages of meiosis. We noticed some slight variation in the rate of disappearance of the protein. Generally it was no longer detectable after diplotene/diakinesis, but in a few cases it apparently persisted to later stages of meiosis. This is probably because of some meiotic asynchrony of the anthers within a single bud, so that the stage indicated by cytological analysis of a single anther may not always correspond precisely with the stages present in the remaining anthers. Expression of BoAsy1 was not detected in the vegetative tissues of the stem and leaf.
Asy1 protein is located in pollen mother cells undergoing meiosis
The anti-Asy1 antibody was next used to immunolocalize Asy1 in Arabidopsis (wild-type) anther locule sections prepared at prophase I and at the tetrad stage after meiosis is complete ( Fig. 3). Binding of FITC-labelled anti-Asy1 Ab to the PMCs that are located in the central region of the anther locule was clearly detected at prophase I ( Fig. 3A) but was no longer detectable in PMCs at the tetrad stage ( Fig. 3B). Antibody binding was not found in any of the locule cells that surround the PMCs at either time point. When pre-immune serum was used on comparable sections in place of the anti-Asy1 antibody, no signal was detected.
Immunolocalization of Asy1 and BoAsy1 to meiotic nuclei and chromosomes
Expression in both species was first detected as punctate foci at meiotic interphase ( Fig. 4A, Fig. 5A). During leptotene, when the axial element first becomes visible, punctate signals are still present but mixed with some stretches of more continuous signal ( Fig. 5B). As the homologous chromosomes continue to condense and synapse through zygotene, the signal becomes associated with the entire length of the lateral elements, although not with the nucleolar region ( Fig. 4B, Fig. 5C). One possibility is that the initially punctate signals gradually extend as meiosis proceeds to give a more continuous signal. Alternatively the apparent continuity of the signal may be caused by juxtaposition of individual foci resulting from condensation of the homologues as meiosis proceeds. Either way it is apparent that the signal intensity is not absolutely uniform, and interestingly the intensity of the signal mirrors that of the DAPI signal along the axes. Continuity of the signal is maintained throughout pachytene ( Fig. 4C, Fig. 5D) but begins to disappear as the homologues desynapse during diplotene ( Fig. 5E). During these stages it is evident that the antibody is localized to the chromosome/bivalent axes and not to the DAPI-positive extended chromatin loops surrounding these structures. Association of BoAsy1 with the desynapsing lateral elements of the SC is maintained at early diplotene where antibody binding to both homologues is clearly visible. However, by late diplotene the signal is no longer associated with the chromosomes, although a few larger foci appear away from the chromosomes, suggesting that the protein accumulates as aggregates prior to degradation ( Fig. 5F). By diakinesis and at the tetrad stage the protein is no longer detectable ( Fig. 5G,H).
BoAsy1 does not localize to the telomere regions
Fluorescence in situ hybridization (FISH) of telomere repeat sequences used in conjunction with silver-stained surface spreads of pachytene chromosomes has revealed that the telomeres appear to abut onto the ends of SCs ( Cuñado and Santos, 1998). Given the close association of Asy1/BoAsy1 with the lateral elements, it was of interest to investigate the distribution of the protein relative to the telomere structures. Combined immunolocalization and FISH using a telomerespecific probe was performed on pachytene chromosome spreads of B. oleracea ( Fig. 6). This revealed that BoAsy1 extended to the telomeres but did not extend into the telomere region as there was little or no merging of the two signals ( Fig. 6A). At a higher resolution it appears that there may be a short region of chromatin not associated with the protein interposed between the signal derived from the fluorescent antibody and that from the telomere repeat, and the Asy1 signal does not extend to the end of the bivalent ( Fig. 6B-E).
EM reveals BoAsy1 localizes to axis-associated chromatin
To resolve the subcellular localization of Bo Asy1 in detail, an immunogold-labelled antibody was applied to spreads prepared from Brassica PMCs at pre-leptotene, early zygotene and pachytene ( Fig. 7). This revealed that in meiotic interphase, prior to the appearance of axial elements, BoAsy1 is distributed over the diffuse chromatin with the gold particles lying in short linear arrays ( Fig. 7A). In leptotene nuclei, as the axial elements appear, it is clear that the gold particles are associated with the chromatin rather than the axial elements ( Fig. 7B). The same pattern of localization to axis-associated chromatin is apparent in zygotene and in pachytene nuclei containing tripartite SC ( Fig. 7C,D). Thus these observations taken in conjunction with the fluorescence microscopy indicate that Asy1 is localized to regions of chromatin that are closely associated with the chromosome axes rather than the axes themselves.
ASY1 is conserved in Brassica oleracea
The development of Arabidopsis as a model plant system for molecular studies has provided a route to the identification of meiotic genes in higher plants. This has enabled us to identify ASY1, a gene required for synapsis of homologous chromosomes during prophase I of meiosis ( Caryl et al., 2000). We have now begun to investigate the role of the Asy1 in chromosome synapsis using both Arabidopsis and the closely related crop species B. oleracea. On the basis of the high degree of sequence homology between these two species, it was predicted that Brassica would possess an ASY1 orthologue with sufficient similarity to the Arabidopsis gene so that an antibody raised against Arabidopsis Asy1 would recognize the protein encoded by the Brassica gene. Cloning and sequence analysis of BoASY1 from B. oleracea confirmed that there is indeed a high degree of similarity and identity between the Arabidopsis ASY1 gene and its Brassica ortholog.
The observation of several bands when Brassica genomic DNA was probed with ASY1 was not unexpected and suggests the presence of at least one other ASY1-related sequence in the Brassica genome. We have previously reported the identification of ASY2, an Arabidopsis gene that encodes a predicted 58.5 kDa protein. Asy2 exhibits 57% amino-acid identity with Asy1 over the N-terminal half of the protein that includes the HORMA domain region, although there is no significant homology over the remainder of the protein. The function of Asy2 is currently unknown. The possibility that it is a functional homologue of Asy1 may be excluded on the basis that it is expressed in an asy1 mutant line. It seems likely that ASY1/2 arose as a result of an ancient duplication of the Arabidopsis genome that is known to have occurred during the evolutionary past of the species and that subsequently there has been a divergence in the functions of the two genes ( Arabidopsis Genome Initiative, 2000). It is therefore likely that homologues of the two genes are also present in the B. oleracea genome. Moreover it is quite possible that additional related copies exist, since it is evident that the Brassica genome has undergone triplication and has expanded about 1.5 times compared with Arabidopsis ( O'Neill and Bancroft, 2000).
Asy1 expression is specific to meiosis
A feature of ASY1 is that gene transcripts are detectable in both reproductive and vegetative tissues of Arabidopsis ( Caryl et al., 2000). This is not restricted to ASY1, as vegetative expression of other meiotic genes in Arabidopsis has also been reported ( Doutriaux et al., 1998; Bai et al., 1999). In other species, transcription of meiotic genes appears to be more strictly confined to reproductive cells, although this is not without exception; for example, the rat SCP1 gene is expressed, albeit at a low level, in the brain ( Kerr et al., 1996). In the case of ASY1, we suggest that expression might be subject to some degree of regulation at the level of translation since the phenotype of asy1 is clearly meiotic with no detectable effect on vegetative growth ( Ross et al., 1997, Ross et al., 1997; Caryl et al., 2000). This suggestion is supported by the results of the western analysis and immunolocalization studies on Arabidopsis anther locule sections, which clearly demonstrate that expression of Asy1 protein is restricted to cells undergoing meiosis ( Fig. 3). Expression of the protein has not been detected in any of the vegetative cells tested to date.
Asy1 is associated with chromatin in close proximity to the axial elements/lateral elements of homologous chromosomes
Mutant asy1 plants exhibit an asynaptic phenotype with a substantially reduced chiasma frequency ( Ross et al., 1997, Ross et al., 1997). The genetic lesion becomes apparent early in prophase I. Telomere pairing at the interphase/leptotene transition that precedes synapsis in wild-type plants is unaffected in asy1, and development of the chromosome axes appears to be substantially normal at the light microscope level ( Armstrong et al., 2001). However, synapsis fails, and the normal events of zygotene and pachytene are not seen. This failure of synapsis suggests that asy1 plants might be defective in a component of the SC or in a protein required for establishment of the SC. Light microscopy studies using FITC-labelled anti-Asy1 initially led us to believe that the protein may be an integral component of the chromosome axes. The signal is clearly associated with both the axial elements prior to synapsis and the lateral elements of the fully synapsed homologues. Moreover, whereas the chromatin not immediately adjacent to the axes was detectable using DAPI, there was no evidence that it interacted with the anti-Asy1 Ab (for example see Figs 5B-E and Fig. 6D-E). However close inspection of the FITC signal revealed that the signal was never truly uniform along the chromosome cores, and some variation in intensity was evident. The regions of increased intensity corresponded to regions of chromatin adjacent to the chromosome axes that exhibited an increased staining intensity with DAPI.
The application of electron microscopy in conjunction with immunogold labelling confirmed that Asy1 is a chromatin-associated protein rather than an integral component of the axial/lateral elements, such as the rat SC proteins SCP2 and SCP3 ( Schalk et al., 1998). However, it is apparently confined to chromatin immediately adjacent to these structures. Thus it belongs to the general category of axis-associated proteins, similar to, for example, TOPO II ( Moens and Earnshaw, 1989). It seems likely that the HORMA domain plays a role in the interaction with chromatin. Whether this structural motif in conjunction with a DNA target is sufficient, or other proteins are required, remains to be resolved. Either way, there is clearly selectivity in the regions of chromatin to which Asy1 associates. This point was further emphasized by the observation that the localization of BoAsy1 along the lateral elements did not extend into the telomere region. It will be of interest to determine if the distribution of the SC structure also follows this pattern relative to the telomere region.
The function of Asy1 is currently a matter for speculation. To our knowledge a protein with the pattern of distribution described above (see below for further comments) has not previously been described. An obvious consequence of mutation of ASY1 is an inability to establish the SC, but its precise role in the establishment of the SC is unclear. Asy1 protein is detected as a chromatin-associated signal before the chromosome axes appear. As the axes become apparent during leptotene, the protein localizes to the chromatin close to the axial elements, and this persists throughout zygotene and pachytene. As meiosis progresses, the paired homologous chromosomes desynapse and the SC breaks down; eventually Asy1 is no longer associated with the axes but some residual protein is detected as protein complexes. It seems possible that these correspond to so-called polycomplexes that become apparent in a wide range of species, including higher plants during diplotene ( Zickler and Kleckner, 1999). They are composed of SC material and their appearance coincides with SC degradation. Although these observations indicate a link between Asy1 and the SC, the nature of this remains to be resolved. On the other hand, the diplotene observations may simply represent extrachromosomal protein aggregates not associated with polycomplexes.
As Asy1 is clearly detectable at early leptotene, a substantial period before synapsis commences, Asy1 may have an role in the juxtaposition of the homologues that occurs in early prophase I. This is supported by our recent observation that in an asy1 mutant homologue, synapsis does not progress beyond the pairing of telomeres in leptotene ( Armstrong et al., 2001). In this case the failure to establish the SC may be a consequence of an inability to pair rather than the protein directly mediating SC assembly. However, we cannot rule out the possibility that in the asy1 mutant transient/unstable pairing occurs but has not been detected by the methodology we have applied. If so, then the protein may act at the interface between homologue pairing and synapsis rather than at an earlier stage, possibly recruiting the bases of the chromatin loops to the developing axial/lateral elements. Such a role may be reflected in the observation that the protein remains associated with the homologues throughout pachytene and its degradation coincides with desynapsis and SC disassembly.
The relationship of Asy1 to SC-associated proteins from other organisms
Given the fundamental nature of meiosis it might seem reasonable to assume that the genes involved would be highly conserved in different species. Although this assumption appears to hold true for genes that encode proteins with a defined enzymatic activity, notably recombination proteins, the position is less clear for other meiotic proteins such as those involved in chromosome synapsis. Thus, even in the case of the SC, which is an evolutionarily conserved structure, identification of homologues to known SC components has proved elusive, suggesting that functional conservation is not necessarily reflected in sequence homology ( Heyting, 1996).
When ASY1 was originally isolated it was noted that it possessed some limited similarity to the yeast HOP1 gene ( Hollingsworth et al., 1990). The predicted proteins are virtually the same size and a BLAST database search revealed that both possess a HORMA domain located towards their N-terminal region. However set against this was a lack of any identifiable homology in the C-terminal half of each protein. In particular, a zinc finger DNA-binding domain and sequences at the C-terminus that are required for Hop1 function are entirely absent in Asy1 ( Hollingsworth et al., 1990; Caryl et al., 2000). A HOP1 homologue has also been reported in the yeast Kluyveromyces lactis ( Smith and Roeder, 2000). The protein KlHop1 has an overall identity of 40% with Hop1 and although this is indicative of the divergence that is apparent amongst SC-associated proteins, it should be noted that regions essential for Hop1 function are conserved in the K. lactis protein. A limited degree of homology has also been noted between Hop1 and the Him3 protein from C. elegans ( Zetka et al., 1999). At 239 amino acids, Him3 is less than half the size of Hop1 (and Asy1) but it contains a HORMA domain that encompasses virtually the entire protein. However in a pairwise comparison it exhibits only 16% identity and 31% similarity to the Hop1 domain ( Zetka et al., 1999). This is somewhat lower than the homology shared between the Hop1 and Asy1 HORMA domains ( Caryl et al., 2000). Also, when Him3 was compared with the Asy1 HORMA domain we failed to detect any significant sequence homology. Thus on balance, the overall high degree of sequence divergence and an absence of any defined biological role for the HORMA domain provides little firm evidence to suggest any functional similarity between the three proteins beyond sharing the potential to interact with chromatin.
We anticipated that determining the localization of Asy1 (and BoAsy1) would provide further insight into to the functional relationship, or lack of, between the three proteins. A comparison of the immunolocalization studies using antibodies raised against the proteins reveals differences in the distribution of each protein during meiosis, although some caution in reaching firm conclusions is required since, to date, localization of Hop1 and Him3 has been based on light microscopy only.
Yeast Hop1 associates with chromosomes during leptotene as a series of foci ( Smith and Roeder, 1997). This is dependent upon Red1, a protein that is required for axial element assembly; Red1 also shows a similar distribution. Evidence suggests that the two proteins work both in concert and possess independent activities ( de los Santos and Hollingsworth, 1999; Hollingsworth and Ponte, 1997; Woltering et al., 2000). As meiosis progresses into zygotene, Hop1 becomes more broadly distributed over each chromosome, although less so than for Red1, and is absent from the nucleolar region. By late pachytene, Hop1 has disassociated from the chromosomes, and it has been suggested that this is coincident with synapsis ( Smith and Roeder, 1997). This pattern of events would suggest that Hop1 is not an integral structural component of the axial cores. Thus, although in this respect the yeast protein seems similar to Asy1/BoAsy1, there is some discrepancy in their temporal distribution. In many respects the distribution of Asy1/BoAsy1 is more reminiscent of Red1, which localizes to both unsynapsed and synapsed chromosomes and persists during pachytene ( Smith and Roeder, 1997). Moreover immunofluorescence using anti-Red1 antibody reveals that the protein is initially present as individual punctate signals that develop into a semi-continuous signal. However, in contrast to Red1, but like Hop1, the plant proteins do not localize to the nucleolus ( Smith and Roeder, 1997). In addition, there is no similarity at the sequence level between Asy1 and Red1, indeed none of the predicted proteins in the Arabidopsis proteome exhibits homology to the yeast protein.
The distribution of Him3 is, however, rather different from both Hop1 and Asy1/BoAsy1, being reminiscent of Cor1, a mammalian axial element protein ( Zetka et al., 1999; Dobson et al., 1994). Initially the protein localizes as numerous foci during early meiosis. As the chromosome axes develop, the punctate appearance is replaced by more continuous staining of the axial elements. This pattern persists throughout prophase I and, importantly, is retained on the chromosome cores following desynapsis of the homologues up to the metaphase I -anaphase I transition. On the basis of this finding it is currently postulated that Him3 is an integral component of the chromosome core that may play a role in sister chromatid cohesion ( Zetka et al., 1999). Thus although the spatial distribution of Asy1/BoAsy1 appears similar to that of Him3, it is quite obvious that there are significant temporal differences as Asy1 clearly disassociates from the chromosome cores before the end of prophase I. Overall there is no evidence to suggest that Asy1/BoAsy1 is involved in sister chromatid cohesion and hence is functionally distinct from Him3.
In summary, it appears that Asy1/BoAsy1 may perform a role similar to that of Hop1/Red1; however further clarification will require identification of the meiotic proteins with which it interacts. The exact role of the protein remains to be established. In the absence of Asy1/BoAsy1, SC morphogenesis cannot take place, suggesting that the protein performs a crucial function, possibly at the interface of the axis-associated chromatin and the nascent SC structure.
We are grateful to Stephen Price, Karen Staples and the EM unit for technical support and to Nancy Kleckner for helpful comments regarding the work. The work was supported by the Biotechnology and Biology Research Council, UK.
↵ * These authors contributed equally to this work
- Accepted July 8, 2002.
- © The Company of Biologists Limited 2002