Meiotic Chromosome Morphology and Behavior in zip1 Mutants of Saccharomyces cerevisiae

Corresponding author: G. Shirleen Roeder, Department of Molecular, Cellular and Developmental Biology, Yale University, P.O. Box 208103, New Haven, CT 06520-8103, shirleen.roeder{at}yale.edu (E-mail).

Communicating editor: S. JINKS-ROBERTSON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The yeast Zip1 protein (Zip1p) is a component of the central region of the synaptonemal complex (SC). Zip1p is predicted to form a dimer consisting of a coiled-coil domain flanked by globular domains. To analyze the organization of Zip1p within the SC, in-frame deletions of ZIP1 were constructed and analyzed. The results demonstrate that the C terminus but not the N terminus of Zip1p is required for its localization to chromosomes. Deletions in the carboxy half of the predicted coiled-coil region cause decreases in the width of the SC. Based on these results, a model for the organization of Zip1p within the SC is proposed. zip1 deletion mutations were also examined for their effects on sporulation, spore viability, crossing over, and crossover interference. The results demonstrate that the extent of synapsis is positively correlated with the levels of spore viability, crossing over, and crossover interference. In contrast, the role of Zip1p in synapsis is separable from its role in meiotic cell cycle progression. zip1 mutants display interval-specific effects on crossing over.

MEIOSIS is essential for diploid organisms to generate haploid gametes. Meiotic cells undergo a single round of DNA replication followed by a lengthy prophase and two successive nuclear divisions. During prophase of meiosis I, homologous chromosomes pair and recombine to ensure their proper segregation at meiosis I (reviewed by ROEDER 1997 ).

Pairing of homologous chromosomes culminates in the formation of a proteinaceous scaffold called the synaptonemal complex (SC), which is morphologically conserved across a wide variety of species (reviewed by VON WETTSTEIN et al. 1984 ). The SC consists of two dense parallel structures called lateral elements, separated by a less dense central region. In organisms that are particularly favorable for cytological study, two distinct substructures can be observed within the central region (MOSES 1968 ; WETTSTEIN and SOTELO 1971 ; SCHMEKEL et al. 1993A , SCHMEKEL et al. 1993B ; reviewed by SCHMEKEL and DANEHOLT 1995 ). The central element runs longitudinally along each complex, equidistant between the lateral elements; transverse filaments cross the central region, lying perpendicular to the longitudinal axis of the SC. Many transverse filaments span the entire width of the SC from one lateral element to the other, while others terminate at the central element. The lateral elements of the SC are separated by ~100 nm throughout the entire length of each pair of homologs. Each lateral element corresponds to the protein backbone of one pair of sister chromatids and is referred to as an axial element before it becomes part of mature SC. In some organisms, full-length axial elements are formed before the initiation of synapsis (reviewed by VON WETTSTEIN et al. 1984 ). In wild-type Saccharomyces cerevisiae, synapsis initiates before axial element formation is complete (DRESSER and GIROUX 1988 ; ALANI et al. 1990 ; PADMORE et al. 1991 ). Throughout this paper, synapsis is defined as SC formation and is recognized cytologically by the appearance of two parallel lateral elements separated by a uniform distance.

Another prominent and essential feature of meiosis is a high level of genetic recombination. Crossing over results in the reciprocal exchange of genetic information and is essential for proper chromosome segregation at meiosis I. Meiotic recombination and SC formation are concurrent events in S. cerevisiae (reviewed by PETES et al. 1991 ; ROEDER 1995 , ROEDER 1997 ; KLECKNER 1996 ). In this organism, most or all meiotic recombination is initiated by double-strand breaks (DSBs). DSBs occur early in prophase before synapsis begins (PADMORE et al. 1991 ). DSBs are converted to double Holliday junctions around the time of SC formation (SCHWACHA and KLECKNER 1994 ). Mature crossover products are detected near the end of pachytene, as the SC disassembles (PADMORE et al. 1991 ; GOYON and LICHTEN 1993 ; NAG and PETES 1993 ). The SC has been postulated to play a role in crossover interference, which decreases the probability of crossing over in regions that have already undergone exchange (EGEL 1978 ; MAGUIRE 1988 ; KING and MORTIMER 1990 ).

In S. cerevisiae, the ZIP1 gene encodes a component of the central region of the SC (SYM et al. 1993 ; SYM and ROEDER 1995 ). The Zip1 protein (Zip1p) localizes to synapsed meiotic chromosomes but not to unsynapsed axial elements (SYM et al. 1993 ). In the zip1 null mutant, homologous chromosomes are paired but not synapsed, and each pair of full-length axial elements is closely associated at a few sites (SYM et al. 1993 ; NAG et al. 1995 ; ROCKMILL et al. 1995 ). Based on DNA sequence analysis, Zip1p is predicted to contain an -helical coiled coil and to form a dimer consisting of a long, rod-shaped domain flanked by globular domains. Mutations that increase the length of the Zip1p coiled coil increase the width of the SC, suggesting that Zip1p is a component of transverse filaments (SYM and ROEDER 1995 ).

The Scp1 protein in rats, mice, and humans, and the homologous Syn1 protein in hamsters are similar to Zip1p in that each contains a long coiled-coil region, and antibodies against these proteins localize specifically to synapsed regions of meiotic chromosomes (MEUWISSEN et al. 1992 , MEUWISSEN et al. 1997 ; DOBSON et al. 1994 ; SAGE et al. 1995 ). Epitope mapping of the hamster Syn1 and rat Scp1 proteins has shown that the C terminus of each molecule lies in the lateral element region, and that the molecules protrude from the lateral elements into the central region of the SC (DOBSON et al. 1994 ; LIU et al. 1996 ; SCHMEKEL et al. 1996 ). The N termini of Scp1 molecules from opposite lateral elements may overlap (SCHMEKEL et al. 1996 ).

The zip1 null mutation causes meiotic arrest in some, but not all, yeast strain backgrounds (SYM et al. 1993 ; SYM and ROEDER 1994 ). In zip1 strains that arrest, recombination initiates, but no mature recombinants are produced (SYM et al. 1993 ). In zip1 strains that sporulate, the frequency of crossing over is decreased two- to threefold, and Holliday junctions persist longer than in wild type (SYM and ROEDER 1994 ; STORLAZZI et al. 1996 ). Crossover interference is completely eliminated in zip1 strains, consistent with a role for the SC in mediating interference (SYM and ROEDER 1994 ).

To investigate the structure and function of Zip1p, we have constructed and analyzed a set of in-frame deletion mutations affecting different domains of the protein. Our results offer insight into the organization of Zip1p within the SC, and they provide information about the relationship between synapsis and the other functions that Zip1p performs.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains:
Yeast manipulations were performed and media were prepared using standard procedures (SHERMAN et al. 1986 ). Cells were grown and induced for meiosis at 30° unless otherwise indicated. Yeast transformations were carried out by the lithium acetate method (ITO et al. 1983 ). Yeast strains used for cytological and sporulation analyses are isogenic derivatives of BR2495 (ROCKMILL and ROEDER 1990 ), which has the following genotype:

In each zip1 deletion mutant, both copies of the chromosomal ZIP1 gene were replaced with the indicated zip1 deletion allele by the two-step transplacement method (ROTHSTEIN 1991 ). The zip1 null mutant (MY152) is homozygous for zip1::URA3 (SYM and ROEDER 1995 ). Strain MY187 is MY152 carrying zip1-M1 (zip1-H2) on a multicopy (2µ) plasmid (pMB164; SYM and ROEDER 1995 ).

Strains used for tetrad analysis are isogenic derivatives of MY261 (SYM and ROEDER 1994 ), which has the following genotype:

In all MY261 derivatives, one copy of the ZIP1 gene carries the zip1::LEU2 disruption (SYM et al. 1993 ); the other copy is either not changed (for wild type) or replaced with the indicated zip1 deletion allele (for deletion mutants) or with the zip1::LYS2 disruption (for the null mutant; SYM and ROEDER 1994 ).

In-frame deletions of ZIP1:
All zip1 in-frame deletions were derived from pMB96, which was described previously (SYM et al. 1993 ). The zip1-N1 mutation was constructed as follows: First, the 3' fusion site for the deletion was created by removing a 69-bp NlaIV-NlaIV fragment (nucleotides 420–488) and ligating the remaining blunt ends, which generates a BamHI site. For the 5' fusion site, a BlpI site (position 58) was cut and filled in with the Klenow fragment of DNA polymerase I, and a 12-bp BamHI linker was inserted. Digestion at the 5' and 3' fusion sites with BamHI, followed by ligation, resulted in the zip1-N1 deletion. The zip1-NM1 mutation was constructed by ligation between the blunt ends resulting from digestion at NlaIV (position 486) and HincII (position 726) sites. The zip1-NM2 mutation was constructed by cutting at a BlpI site (position 58), filling in with Klenow fragment, inserting a 12-bp ClaI linker, cutting at the inserted ClaI site and at a Psp1406I site (position 723), and ligating the ends. The zip1-M1 mutation, previously referred to zip1-H2, was described by SYM and ROEDER 1995 . The zip1-M2 mutation was constructed as follows: To create the 5' fusion site, an 8-bp BglII linker was inserted at an EcoRV site (position 1222). For the 3' fusion site, a 12-bp BamHI linker was inserted at a PvuII site (position 2099). Digestion at the inserted BglII and BamHI sites and ligation of the resulting ends created zip1-M2. The zip1-MC1 mutation was constructed by cutting at EcoRV (position 1222) and XmnI (position 2393) sites and ligating the resulting blunt ends. The zip1-MC2 mutation was constructed by cutting at the BamHI site inserted during construction of zip1-M2, filling in with Klenow fragment, and then ligating to the blunt end resulting from digestion at an XmnI site (position 2393). The zip1-C1 mutation was created by cutting at a Bsu36I site (position 2365), filling in the ends, and ligating to the blunt end resulting from digestion at an HpaI site (position 2469). The zip1-C2 mutation has a 10-bp XbaI linker (New England Biolabs, Beverly, MA) inserted at the HpaI site (position 2469). This insertion results in an in-frame UAG stop codon.

A SacI-SalI fragment containing the zip1-N1 mutation was subcloned into a YIp vector, pRS306 (SIKORSKI and HIETER 1989 ), at SacI-SalI sites to create pTP86 for integration. For the integration of zip1-NM1, zip-NM2, zip1-M1, zip1-M2, and zip1-MC2, SacI-XhoI fragments containing the mutations were subcloned into the SacI-XhoI sites of pRS306 to make pTP110, pTP106, pTP62, pTP97, and pTP105, respectively. The BamHI-SphI fragments containing the zip1-MC1 and zip1-C1 deletions were cloned into the BamHI-SphI sites of YIp5 (STRUHL et al. 1979 ) to make the integration plasmids pTP5 and pTP6, respectively. A SacI-SalI fragment containing zip1-C2 was subcloned into the SacI-SalI sites of a YCp vector, pRS316 (SIKORSKI and HIETER 1989 ), to generate pTP98. Plasmids were targeted for integration at ZIP1 by cutting with EcoNI (pTP86, pTP110, and pTP106), Bsu36I (pTP62), HpaI (pTP97 and pTP105), or XbaI (pTP5 and pTP6).

Coiled-coil analysis:
The predicted amino acid sequence of Zip1p was analyzed using the COILS program (LUPAS et al. 1991 ) to calculate the probability that the sequence will adopt a coiled-coil conformation. The COILS program used for the analysis is located at http://ulrec3.unil.ch/software/COILS_form.html. Using the myosins, tropomyosins, intermediate filaments, desmosomal proteins, and kinesins matrix with window = 21, four stretches of sequence within Zip1p are predicted to adopt a coiled-coil conformation. These are the segments from residues 180 to 218, from 233 to 321, from 395 to 433, and from 463 to 748. Assuming 0.1485 nm per residue in a coiled-coil conformation (STEINERT et al. 1993 ), these coiled-coil segments are predicted to be 5.79, 13.22, 5.79, and 42.47 nm in length. Thus, the total length of the coiled-coil region of a Zip1p dimer is predicted to be 67.27 nm. In the zip1-M2 and zip1-MC1 mutations, a linker region of 29 amino acids (aa) is included in the deletions; in the zip1-NM1 mutation, a linker of 14 aa is included. Because these linkers are very short, it is assumed that the deletion of the linker does not cause any significant change in the length of the Zip1p dimer beyond that caused by the deletion in the coiled-coil segments. Also, it is assumed that the 16 aa deleted from the N-terminal globular domain in zip1-NM1 and the 51 aa deleted from the C-terminal globular domain in zip1-MC1 and zip1-MC2 do not significantly affect the length of the Zip1p dimer.

Cytology:
Meiotic chromosome spreads were prepared according to the method of DRESSER and GIROUX 1988 . All mutants were examined in multiple, independent spread preparations with similar results. For electron microscopy, spreads were stained with silver nitrate as described by DRESSER and GIROUX 1988 . Immunofluorescence procedures were performed according to SYM et al. 1993 . Chromosomal DNA was visualized by staining with a DNA-binding dye, DAPI. The rabbit anti-Zip1p antibody (SYM et al. 1993 ) was used at a 1:100 dilution, and Cy3-conjugated secondary antibody against rabbit IgG (Jackson Laboratories, West Grove, PA) was used at a 1:200 dilution. A Leitz DMRB microscope (Wetzlar, Germany) equipped with fluorescence and a PL APO x100 objective was used to observe antibody-stained preparations. Images were captured using an Imagepoint CCD camera (Photometrics, Tucson, AZ) and processed with the IPLab Spectrum software (Vienna, Virginia).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In-frame deletions affecting different domains of Zip1p:
According to the predicted amino acid sequence of Zip1p, the internal -helical coiled-coil domain extends approximately from aa 180 to aa 748, with three non-coiled-coil linker interruptions (aa 219–232, 322–394, and 434–462). Residues 1–179 comprise the N-terminal globular domain, and aa 749–875 form the C-terminal globular domain. To investigate the relationship between Zip1p structure and function, strains carrying in-frame deletion mutations in different regions of the protein were constructed and analyzed. Mutations are given letter designations that indicate the domain(s) in which each deletion begins and ends (Figure 1): N for the N-terminal globular domain, M for the coiled-coil domain in the middle of Zip1p, and C for the C-terminal globular domain. For example, N1 is confined to the N-terminal domain, while NM1 begins in the N-terminal domain and ends in the coiled-coil domain. Together, the deletion mutations analyzed span the entire length of the ZIP1-coding region.

View larger version (22K): In this window In a new window Download PPT slide   Figure 1. —Deletions of Zip1p. Based on the predicted three-dimensional conformation, Zip1p is divided into three regions: the N-terminal globular domain (N), the coiled coil in the middle of the protein (M), and the C-terminal globular domain (C). The vertical lines represent the boundaries between domains. The portion deleted in each mutant is indicated by the gap; the amino acids deleted are indicated on the right.

Deletion mutations were analyzed for their effects on chromosome synapsis, Zip1p localization, sporulation efficiency, spore viability, crossing over, and crossover interference. To assess spore viability, crossing over, and crossover interference, tetrad analysis was performed in an SK1 strain background in which the zip1 null mutant sporulates (SYM and ROEDER 1994 ). Cytological analyses and sporulation assays were carried out in a BR2495 strain background (ROCKMILL and ROEDER 1990 ) in which the zip1 null mutant undergoes pachytene arrest (NAG et al. 1995 ; data not shown). For cytological studies, meiotic chromosomes at various stages of meiotic prophase were surface spread. To assess chromosome synapsis, spread chromosomes were stained with silver nitrate and observed in the electron microscope. To monitor Zip1p localization, spread chromosomes were stained with anti-Zip1p antibodies, followed by appropriate secondary antibodies. For each mutant, a time course analysis of Zip1p localization was performed to monitor the rate of synapsis.

The N-terminal globular domain of Zip1p is not essential for synapsis:
In Zip1-N1p, 80% of the N terminus (aa 21–163) is deleted. The zip1-NM1 deletion (aa 164–243) includes one end of the N terminus and a segment of the coiled-coil domain. zip1-NM2 (aa 21–242) is the sum of zip1-N1 and zip1-NM1.

Based on electron microscopy, homologous chromosomes in zip1-N1 undergo full synapsis just as they do in wild type (Figure 2A and Figure B, and Figure 3). Fluorescence microscopic analysis demonstrates that the mutant Zip1-N1p localizes along the lengths of pachytene chromosomes; however, Zip1p staining is not as continuous as it is in wild type, even at late time points (Figure 4A and B; Table 1). In most zip1-NM1 nuclei, homologous chromosomes are fully synapsed (Figure 2C and Figure 3), but short unsynapsed regions are occasionally observed (Figure 2D). Zip1p staining appears to be linear, and nuclei displaying linear staining accumulate as cells arrest in pachytene (Figure 4C; Table 1). In the zip1-NM2 mutant, the extent of synapsis varies among nuclei from only short stretches of SC to nearly complete synapsis, but nuclei with fully synapsed chromosomes have not been detected (Figure 2E and Figure F, and Figure 3). Zip1p staining is punctate, even at late time points (Figure 4D; Table 1). Anti-Zip1p antibodies were raised against the C-terminal 264 aa; thus, the less continuous Zip1p staining observed in the zip1-N1 and zip1-NM2 mutants cannot result from a loss of epitopes recognized by the antibody.

View larger version (119K): In this window In a new window Download PPT slide   Figure 2. —Electron micrographs of meiotic chromosomes from mutants affecting the N- and C-terminal globular domains of Zip1p. An unsynapsed region in D is indicated by an arrow. Stretches of SC in E and F are indicated by arrowheads. All spreads were prepared from cells sporulated at 30°. Bar, 1 µm.

View larger version (21K): In this window In a new window Download PPT slide   Figure 3. —Summary of phenotypes of zip1 deletion mutants. SC formation: each pair of thick lines indicates the lateral elements (axial elements if not synapsed) of a pair of homologs; short vertical lines represent Zip1p, and the gray vertical lines represent unstable synapsis in zip1-NM1. Sporulation and spore viability data are taken from Table 3; ts, temperature sensitive. Crossing over is the sum of map distances in the MAT-CEN3 and CEN3-HIS4 intervals; data are taken from Table 4. Symbols for interference: +, wild-type interference; ±, decreased interference; -, no interference.

View larger version (44K): In this window In a new window Download PPT slide   Figure 4. —Immunolocalization of Zip1p. Spread nuclei from wild-type and zip1-deletion mutants were stained with DAPI and anti-Zip1p antibodies. The two staining patterns are superimposed. For the purpose of better contrast, DAPI staining is shown in red, and anti-Zip1p staining is shown in green. Regions of overlap are yellow, and the intense green-staining bodies in F and L are polycomplexes. Large arrowheads point to the nucleolar region on chromosome XII, which does not undergo synapsis and does not display Zip1p staining. Arrows point to condensed chromosomes that have incomplete Zip1p staining. Sharp arrowheads in K and L point to regions of Zip1p staining that are paired, but separate on homologs. The inserts in K and L are shown at 1.5-fold higher magnification. All spreads were prepared from cells sporulated at 30°. Bar, 2 µm.

These data suggest that the N terminus of Zip1p is not essential for chromosome synapsis. However, this domain might play a role in stabilizing Zip1p or the structure of the SC.

The C-terminal domain is essential for Zip1p localization to chromosomes:
The zip1-MC2 deletion removes the last 48 amino acids of the coiled-coil region and the first 51 residues of the C terminus (aa 701–799). In the zip1-C1 deletion, only 34 aa (791–824) are eliminated from the C terminus of Zip1p. The zip1-C2 deletion begins at aa 825 and extends to the end of the ZIP1-coding region.

In electron microscopic and immunofluorescence analyses, the zip1-MC2 mutant undergoes wild-type SC formation (Figure 2G and Figure 3). Furthermore, Zip1p localization appears normal with respect to both pattern and timing (Figure 4E; Table 1). In the zip1-C1 mutant, full-length axial elements are assembled and paired, but no SC is detected in silver-stained preparations (Figure 2H and Figure 3). The Zip1-C1p does not localize to chromosomes; instead, it assembles into polycomplexes, which are aggregates of Zip1p unassociated with chromosomes (Figure 4F). Polycomplexes are observed when wild-type Zip1p is overproduced (SYM and ROEDER 1995 ) and in mutants defective in SC formation (ALANI et al. 1990 ; BISHOP et al. 1992 ; LOIDL et al. 1994 ). The zip1-C2 mutant is indistinguishable from the zip1 null mutant; neither Zip1p staining nor SC formation were observed in spread nuclei of cells carrying this mutation (data not shown). This result is consistent with a previous report indicating that the nuclear localization signal for Zip1p is located at the extreme C terminus (BURNS et al. 1994 ), which is deleted in this mutant.

These results demonstrate that aa 791–824 are important for Zip1p localization to chromosomes and thus for synapsis. This region may be essential for the interaction of Zip1p with some other component(s) of the SC.

Deletions in the Zip1p coiled coil can decrease the width of the SC:
Both the zip1-M2 (aa 409–700) and zip1-MC1 (aa 409–799) mutants have deletions in the carboxyl half of the coiled-coil region (Figure 1). Electron microscopic examination demonstrates that homologous chromosomes are synapsed in both mutants (Figure 3 and Figure 5A and Figure B). Immunofluorescence analysis indicates Zip1p localization along the lengths of chromosomes similar to wild type (Figure 4G and Figure H). However, the rate of synapsis in both mutants is slower than in wild type (Table 1). In electron microscopic analysis, chromosomes at certain intermediate stages of synapsis are frequently detected in the mutants, but they are rarely observed in wild type. At these intermediate stages, synapsed chromosomes appear to have short unsynapsed regions at their ends, and/or unsynapsed axial elements coexist with fully synapsed chromosomes in the same nucleus (Figure 5C and Figure D). Consistent with the electron microscopic observations, incomplete Zip1p staining on condensed chromosomes (based on DAPI staining) is detected in both mutants (Figure 4I and Figure J). At late time points, the number of nuclei at intermediate stages decreases, and all chromosomes become fully synapsed (Figure 4G and Figure H, and Figure 5A and B; Table 1).

View larger version (80K): In this window In a new window Download PPT slide   Figure 5. —Electron micrographs of meiotic chromosomes from Zip1p coiled-coil mutants. All spreads were prepared from cells sporulated at 30°. Bar, 1 µm.

The width of the SC is obviously decreased in the zip1-M2 and zip1-MC1 mutants (Figure 6). To investigate further the relationship between zip1 deletions and SC width, the width of the SC in wild-type and all zip1 deletion mutants that make complete SC was measured from electron micrographs of SCs in which the lateral elements could be clearly distinguished. The width of the SC in the zip1-N1 and zip1-NM1 mutants is approximately the same as in wild type (Table 2). On the other hand, in the zip1-MC2, zip1-MC1, and zip1-M2 mutants, the width of the SC is decreased by 14, 66, and 52 nM, respectively (Table 2). These data show that the width of the SC can be decreased by deletions in the coiled-coil domain of Zip1p.

View larger version (64K): In this window In a new window Download PPT slide   Figure 6. —Regions of SC from coiled-coil deletion mutants of Zip1p. Bar, 200 nm.

Zip1p antibodies stain unsynapsed chromosomes in the zip1-M1 mutant:
The zip1-M1 mutation has a deletion in the N-terminal half of the coiled-coil domain (aa 244–511). Previous studies indicated that the zip1-M1 mutant (formerly referred to as zip1-H2) makes apparently normal SC (SYM and ROEDER 1995 ). However, in this study, we found that homologous chromosomes are paired but not synapsed in the zip1-M1 mutant, based on electron microscopic observations (Figure 5E). Furthermore, analysis of Zip1-M1p localization revealed Zip1p antibody staining in regions of parallel axial elements that are unsynapsed (Figure 4K). Similar results were obtained in an SK1 strain background and in cells carrying the zip1-M1 allele on a multicopy vector (Figure 4L and Figure 5F). The simplest explanation for the difference between our results and those published previously is that the SC formed in the zip1-M1 mutant is highly unstable and therefore very sensitive to subtle (and unknown) variations in the spreading procedure.

Sporulation is independent of chromosome synapsis:
Sporulation was assessed in the same strains used for cytological analyses. The zip1-MC2 and zip1-N1 strains sporulate to the same extent as wild type, although sporulation is slightly delayed (Figure 3 and Figure 7A; Table 3). Sporulation in the zip1-NM1 mutant is temperature sensitive. The sporulation efficiency is 55% at 27° (similar to wild type), but it decreases to 11% at 30° (20% of the wild-type level; Figure 3 and Figure 7B; Table 3). zip1-NM2 and zip1-C1 fail to sporulate, just like the null mutant (Figure 3; Table 3). In the mutants just described above, there is a rough correlation between the ability to make SC and the ability to sporulate. However, other mutants provide exceptions to this rule. The zip1-M2 and zip1-MC1 mutants do not sporulate, even though they make fully synapsed chromosomes (Figure 3; Table 3). The fully synapsed chromosomes observed in these mutants persist even after 42 hr in sporulation medium (data not shown). Sporulation in the zip1-M1 mutant is substantially delayed but eventually reaches nearly wild-type levels, even though no stable SC is assembled (Figure 3 and Figure 7A; Table 3). These results demonstrate that the zip1 defects in chromosome synapsis and sporulation can be uncoupled.

View larger version (19K): In this window In a new window Download PPT slide   Figure 7. —Kinetics of spore formation in zip1 mutants. Spore formation was monitored in wild type, zip1-MC2, zip1-N1, and zip1-M1 at 30° (A) and in wild type and zip1-NM1 at 27° (B). Cells were examined in the light microscope to assess spore formation at 2-hr intervals from 16–42 hr after transfer to sporulation medium.

Spore viability and chromosome segregation in zip1 mutants:
To investigate the effects of zip1 deletion mutations on spore viability, chromosome segregation, and meiotic recombination, tetrad analysis was carried out in an SK1 strain in which the centromere of one chromosome III is marked with TRP1 and the centromere of the homolog is marked with URA3. Meiosis in wild type produces four viable spores, of which two are Trp⁺ and the other two are Ura⁺. If a pair of homologs segregates to the same pole during meiosis I (homolog nondisjunction), two of the resulting spores will be disomic for that chromosome (Trp⁺ Ura⁺ in the case of chromosome III), and the other two will be inviable. If one chromosome and one chromatid from the homologous chromosome segregate to the same pole at meiosis I [precocious separation of sister chromatids (PSSC)], the result will be a three-spore-viable tetrad in which one spore is disomic (Ura⁺ and Trp⁺ in the case of chromosome III).

In the zip1-MC2, zip1-N1, and zip1-NM1 strains, which make apparently normal or nearly normal SC, spore viability is similar to that of wild type (Figure 3; Table 3). In the zip1-M2 and zip1-MC1 mutants, which make SCs that are significantly narrower than wild-type complexes, spore viability is decreased to ~90% of wild type (Figure 3; Table 3). In the zip1-NM2 mutant, which undergoes incomplete synapsis, spore viability is decreased even more, but it is still higher than that of the null mutant (Figure 3; Table 3). In the zip1-M1 mutant, in which Zip1p localizes to unsynapsed chromosomes, spore viability is slightly better than that of the zip1 null mutant (Table 3). Spore viability of zip1-C1, which fails to make any SC, is close to that of the null mutant (Table 3). These data indicate a correlation between SC formation and spore viability (Figure 3).

The distribution of four-, three-, two-, one-, and zero-spore-viable tetrads and the occurrence of spores disomic for chromosome III were also analyzed. In all mutants, the proportion of two- and zero-spore-viable tetrads increases significantly as spore viability decreases (Table 3). Similarly, the frequency of tetrads with a pair of spores disomic for chromosome III increases as spore viability decreases (Table 3). These results indicate that the decrease in spore viability in the deletion mutations is caused largely by homolog nondisjunction at meiosis I, as shown previously for the zip1 null mutant (SYM and ROEDER 1994 ). The exceptions to this rule are the zip1-M2 and zip1-MC1 mutants, in which the frequency of three-spore-viable tetrads is much higher than in wild type or other mutants (Table 3), suggesting an additional defect in chromosome segregation. Unexpectedly, the increase in three-spore-viable tetrads is not associated with an increase in the frequency of PSSC for chromosome III (Table 3; see DISCUSSION).

Crossover frequency and distribution:
Crossing over was measured in the MAT-CEN3 and CEN3-HIS4 intervals on chromosome III by tetrad analysis (Table 4). The map distances determined for the two intervals were added together to obtain an overall measure of crossover frequency in each strain (Figure 3). The zip1-MC2 strain displays the highest crossover frequency (61.7 cM), which is close to the map distance observed in wild type (58.1 cM). In contrast, the amount of crossing over is reduced more than threefold in the zip1 null (12.7 cM), zip1-M1 (17.8 cM), and zip1-C1 (18.0 cM) mutants. Crossing over in the other mutants is at intermediate levels (Table 4; Figure 3). Overall, there is a rough correlation between the frequency of crossing over and the level of spore viability (Figure 3).

In the zip1-MC2, zip1-N1, zip1-NM1, zip1-M2, and zip1-MC1 strains, crossing over is increased or unchanged in the MAT-CEN3 interval, but it is decreased in the CEN3-HIS4 interval (Table 4). For all other mutants, including the zip1 null mutant, the decrease in crossing over in the CEN3-HIS4 interval is much greater than the decrease in the MAT-CEN3 interval. These data demonstrate that zip1 affects crossing over to different extents in different regions of the genome. The modest increase in crossing over in the MAT-CEN3 interval in the zip1-MC2, zip1-N1, and zip1-M2 mutants might be caused by reduced interference from crossovers in the adjacent CEN3-HIS4 interval, where crossing over is significantly decreased.

Crossover interference:
Crossover interference can be measured in terms of the frequency of double crossovers in a given interval. A double crossover involving all four chromatids results in a nonparental ditype (NPD) tetrad. Therefore, crossover interference can be quantitated in terms of the NPD ratio (SNOW 1979 ), which is the proportion of NPDs observed relative to the proportion of NPDs expected in the absence of interference. No interference results in an NPD ratio of 1.0, and absolute interference results in an NPD ratio of zero. NPD ratios in the zip1 null mutant are ~1.0 (SYM and ROEDER 1994 ).

NPD ratios were calculated for the MAT-CEN3 and CEN3-HIS4 intervals (Table 4). In the zip1-MC2, zip1-N1, zip1-M2, and zip1-MC1 strains, both intervals show statistically significant levels of positive interference. The NPD ratios in these mutants range from 0.278 to 0.370 for MAT-CEN3 and from 0.130 to 0.230 for CEN3-HIS4. Interference is slightly decreased in the zip1-NM1 mutant and even more in zip1-NM2. The apparent negative interference observed in the CEN3-HIS4 interval in zip1-NM2 (NPD ratio of 3.1) is probably caused by random variation in recombination frequency resulting from limited sample size (SALL and BENGTSSON 1989 ) rather than by a bona fide effect of the zip1-NM2 mutation. In the zip1-M1 mutant, NPD ratios for the MAT-CEN3 and CEN3-HIS4 intervals are 1.000 and 1.176, respectively, suggesting no interference. In summary, all the mutants that make stable, full-length SC display approximate wild-type levels of interference, while mutants that make little or no stable SC exhibit reduced interference (Figure 3).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The C terminus, but not the N terminus, of Zip1p is essential for localization to chromosomes:
Our results demonstrate that the C terminus of Zip1p is important for its function. The zip1-C1 mutation, which removes only 34 aa, completely eliminates the ability of Zip1p to localize to chromosomes and thus abolishes SC formation. Consistent with the cytological data, the defects in sporulation, spore viability, and crossing over caused by zip1-C1 are similar to those caused by the null mutation. Of the 34 aa deleted in zip1-C1, the first nine overlap with the deletions in the zip1-MC1 and zip1-MC2 mutants, which do not affect Zip1p localization to chromosomes. Thus, the 25 aa from 800 to 824 are necessary for Zip1p assembly onto chromosomes.

In contrast, the N terminus of Zip1p is not critical for function. The zip1-N1 and zip1-NM1 mutations have only minor effects on chromosome synapsis. Furthermore, a glutathione S-transferase-ZIP1 fusion gene that lacks the first 20 codons of the ZIP1 open reading frame complements the zip1 null mutant for SC formation and sporulation (data not shown). Together, these mutants cover the entire N-terminal globular domain of Zip1p, yet none has a significant effect on SC formation or Zip1p localization to chromosomes. Even the zip1-NM2 mutant, which is the sum of zip1-N1 and zip1-NM1, makes a significant amount of SC. Consistent with the cytological results, spore viability and crossing over in the zip1-N1 and zip1-NM1 mutants are close to the wild-type levels, and the zip1-NM2 mutant displays much higher spore viability and crossing over than the null mutant. Based on these data, the function of the N-terminal globular domain of Zip1p remains obscure. The temperature-sensitive sporulation defect of the zip1-NM1 strain suggests that the mutant protein and/or the SC formed in the mutant are unstable. The discontinuity of Zip1p staining in the zip1-N1 and zip1-NM2 mutants might reflect a defect in maintaining Zip1p on chromosomes.

Model for the organization of Zip1p in the SC:
Our results indicate that the N and C termini of Zip1p play different roles in SC formation. The C terminus of Zip1p is essential for chromosomal localization, suggesting that this domain interacts with the lateral elements of the SC. The N terminus is neither necessary nor sufficient for the localization of Zip1p to chromosomes. The differential importance of the two Zip1p globular domains, as well as data on SC width discussed below, lead us to propose a model for the organization of Zip1p within the SC. We assume that Zip1p, like other characterized coiled-coil proteins, forms dimers in which the two monomers are in parallel orientation and exact register (reviewed by STEINERT and ROOP 1988 ). Then, two dimers form a tetramer in which the two dimers are in antiparallel orientation and partially overlapping (Figure 8A). It is this tetramer that is the building block of the SC and spans the width of the SC from one lateral element to the other. The C termini are attached to the lateral elements, and Zip1p dimers from opposite lateral elements overlap for their N termini and part of the coiled-coil region (Figure 8A). This organization for Zip1p within the SC is similar to that proposed for the Scp1 and Syn1 proteins, based on high-resolution epitope mapping (DOBSON et al. 1994 ; SCHMEKEL et al. 1996 ). Consistent with our model, epitope mapping of Zip1p shows that the C terminus of each molecule lies in the lateral element region, and the N terminus localizes to the central region (H. DONG and G. S. ROEDER, unpublished results).

View larger version (19K): In this window In a new window Download PPT slide   Figure 8. —Model for the organization of Zip1p within the SC. Each Zip1p dimer is represented by a pair of parallel horizontal lines (for the coiled-coil region) flanked by a circle (for the N-terminal globular domain) and a square (for the C-terminal globular domain). The vertical gray lines indicate the lateral elements of the SC. The relative positions of different regions in the coiled-coil domain of Zip1p are indicated by letters a–f. Figure is not drawn to scale.

Zip1p tetramers could be formed before they localize to chromosomes, or they could be formed in situ with Zip1p dimers that have already localized to chromosomes. The N terminus of Zip1p may contribute to the stability of the SC by stabilizing the interaction between Zip1p dimers attached to opposing lateral elements. It is also possible that there is no direct interaction between Zip1p dimers, and that the dimers are connected by other components of the central region of the SC.

A previous study suggested that a single Zip1p dimer spans the width of the SC from one lateral element to the other (SYM and ROEDER 1995 ). This proposal was based on the observation that duplication or triplication of a restriction fragment encoding part of the Zip1p coiled coil increased the width of the SC by amounts corresponding to the predicted increases in the length of the Zip1p dimer (SYM and ROEDER 1995 ). In the previous study, it was assumed that the entire region that was duplicated (or triplicated) would adopt a coiled-coil conformation (SYM and ROEDER 1995 ). However, an improved program for coiled-coil analysis (see MATERIALS AND METHODS) suggests that nearly 40% of the affected amino acids are within linker regions and, thus, should not make a significant contribution to the length of the coiled-coil domain. The newly predicted increases in the length of Zip1p dimers are much less than the observed increases in the width of the SC. Thus, the behavior of the duplication and triplication mutants is not consistent with the model that a single Zip1p dimer spans the width of the SC.

The width of the SC can be decreased by deletions in the Zip1p coiled coil:
A previous study showed that the width of the SC can be increased by insertions in the coiled coil of Zip1p (SYM and ROEDER 1995 ). Our results demonstrate that the SC width can also be decreased by deletions in the coiled-coil region of Zip1p. According to our model, a short deletion at the N-terminal end of the Zip1p coiled-coil domain should not cause a decrease in the width of the SC, though it should shorten the region of overlap between dimers (Figure 8B). Consistent with this prediction, the width of the SC is not decreased in the zip1-NM1 mutant, although the mutant Zip1p dimer is predicted to be shorter than the wild-type dimer by 7.4 nm. In contrast, a deletion at the C-terminal end of the coiled coil should decrease the width of the SC by twice the change in length of the Zip1p dimer (Figure 8C). In agreement with this prediction, the decrease in the width of the SC (14 nm) in the zip1-MC2 mutant is twice the decrease in the predicted length of the dimer (7.1 nm). Also, the difference in the length of the dimer between the zip1-MC1 and zip1-M2 mutants is predicted to be 7.1 nm, and the difference in the width of the SC is 14 nm. Thus, these results are consistent with the predictions made by our model.

For mutants that have large deletions in the coiled-coil region of Zip1p, such as zip1-MC1 and zip1-M2, it is difficult to predict their effects on the width of the SC because of uncertainty regarding the extent of overlap between Zip1p dimers. Furthermore, studies of the structure and assembly of keratin, which shares structural homology to Zip1p, have shown that different forms of tetramers can be formed depending on the alignment of dimers (AEBI et al. 1988 ; STEINERT 1991 ; STEINERT et al. 1993 ; reviewed by STEWART 1993 ). By analogy, it is possible that tetramerization of Zip1p dimers in the zip1-MC1 and zip1-M2 mutants uses an alternative region of overlap.

Chromosome synapsis and meiotic cell cycle control:
Temporal studies have demonstrated that mature crossover products arise near the end of pachytene, as the SC disassembles (PADMORE et al. 1991 ). In zip1-null strains, chromosomes fail to synapse, and there is a defect (or a delay) in the production of crossover products (SYM et al. 1993 ; SYM and ROEDER 1994 ; STORLAZZI et al. 1996 ). These observations raise the possibility that meiotic cells monitor the status of SC formation, preventing the resolution of recombination intermediates until chromosomes are synapsed. Such coordination would be important if the SC plays a role in regulating the distribution of crossovers along and among chromosomes (reviewed by EGEL 1995 ; ROEDER 1997 ). In this case, crossing over in the absence of SC would lead to a deregulation of crossover distribution. As a result, some chromosome pairs would fail to recombine and therefore nondisjoin at meiosis I, generating inviable meiotic products.

In zip1-MC1 and zip1-M2 strains, which make complete SC that is narrower than wild-type SC, meiotic arrest might be caused by the irregular morphology of the complex. This hypothesis is consistent with a previous study suggesting that a meiosis-specific surveillance system monitors the status of recombination complexes in a specific chromosomal context that includes the SC-related proteins Red1p and Mek1p (XU et al. 1997 ). Red1p is a component of lateral elements and is essential for SC formation (SMITH and ROEDER 1997 ), while Mek1p is a meiosis-specific protein kinase required for normal SC morphogenesis (ROCKMILL and ROEDER 1991 ). The defect (or delay) in sporulation caused by zip1 can be alleviated by a red1 or mek1 mutation (XU et al. 1997 ; J. M. BAILIS and G. S. ROEDER, unpublished results), suggesting that Red1p and Mek1p are essential for the function of the proposed surveillance system. Our data suggest that the checkpoint machinery responds not only to the absence of SC, but also to aberrations in SC morphology.

The observation that zip1-M1 cells sporulate indicates that the irregular structure of chromosomes in this mutant is not recognized as abnormal by the surveillance system. One explanation is that an interaction between Zip1p and some other component(s) of meiotic chromosomes is required to produce the signal that activates the recombination machinery (or blocks the inhibition of this machinery). According to this model, the relevant protein-protein interaction occurs in the zip1-M1 mutant, but not in the zip1-MC1 and zip1-M2 mutants.

Chromosome synapsis and crossing over:
The decrease in crossing over in the zip1-null mutant could be caused by the absence of SC, which in turn is caused by the absence of Zip1p. Alternatively, Zip1p may play a role in meiotic recombination that is independent of its role in SC formation, as proposed by STORLAZZI et al. 1996 . According to this model, it might be possible to isolate zip1 mutants that are proficient in crossing over but defective in synapsis.

The results of this study demonstrate a correlation between the extent of synapsis and the level of crossing over. If zip1 deletion mutations are ranked with respect to the extent of chromosome synapsis and with respect to the amount of crossing over, there is an excellent correspondence between the two rank orders (Figure 3). The simplest interpretation of these results is that the SC promotes crossing over, perhaps by providing a context that favors the resolution of recombination intermediates in the direction of crossing over. We have been unsuccessful in identifying any mutations that affect synapsis but not crossing over. Nevertheless, our results do not exclude the possibility that Zip1p affects recombination independent of its function in SC polymerization.

Interval-dependent effects on crossing over:
In all of the zip1 mutants examined, crossing over in the CEN3-HIS4 interval is decreased more than in the MAT-CEN3 interval (Table 4). These results suggest that crossovers in the CEN3-HIS4 interval are more sensitive to variations in SC morphology or abundance.

In yeast, most or all meiotic recombination is initiated by DSBs, and meiotic recombination hotspots are preferential sites for DSB formation (WU and LICHTEN 1994 ; LICHTEN and GOLDMAN 1995 ; ROEDER 1995 ). It has been shown that the distribution of DSBs is similar to the distribution of meiotic crossovers in the CEN3-LEU2 region contained within the CEN3-HIS4 interval (WU and LICHTEN 1994 ), suggesting that the distribution of crossovers is determined at the initiation stage. However, it was recently reported that DSBs are rare in the CRY1-PGK1 interval (most of the MAT-CEN3 interval), even though there is a significant level of crossing over (BAUDAT and NICOLAS 1997 ). Several possibilities have been suggested to explain this unexpected result (BAUDAT and NICOLAS 1997 ): (1) the genetic distance may be overestimated; (2) the frequency of DSBs may be underestimated, perhaps because the breaks do not occur at discrete sites (hotspots); and (3) some crossovers might be initiated by lesions other than DSBs. If possibility (2) or (3) is the case, the interval-dependent effects on crossing over observed in the zip1 mutants might be DSB related. Perhaps Zip1p (or the SC) affects only recombination events that are initiated by DSBs at recombination hotspots.

zip1-M2 and zip1-MC1 mutants increase the frequency of three-spore-viable tetrads:
The proportion of three-spore-viable tetrads in the zip1-M2 and zip1-MC1 mutants is two- to fivefold higher than in wild-type and other zip1 mutants (Table 3), suggesting that the frequency of PSSC is increased in these mutants. However, the frequency of PSSC for chromosome III in zip1-M2 and zip1-MC1 is not correspondingly higher (Table 3). It is possible that the increase in three-spore-viable tetrads in these mutants is caused by chromosome missegregation at meiosis II or by chromosome loss, but not by PSSC at meiosis I. Alternatively, the increase in three-spore-viable tetrads may be caused by an increase in PSSC that is not detectable in our assay. Chromosome III is one of the smallest chromosomes in yeast, raising the possibility that the defect that causes PSSC in these mutants is more pronounced on large chromosomes, whose missegregation was not scored. The zip1-M2 and zip1-MC1 mutants both make SC that is significantly narrower than wild-type SC (Figure 6). It is possible that the aberrant SC presents a difficulty for SC disassembly, which perturbs subsequent chromosome segregation. Bigger chromosomes might be affected more because of the greater length of SC that must be disassembled.

Chromosome synapsis and crossover interference:
Most models for crossover interference suggest that the SC is involved in the transmission of an inhibitory signal along the chromosome from one crossover site to nearby potential sites of crossing over (EGEL 1978 ; MAGUIRE 1988 ; KING and MORTIMER 1990 ; SYM and ROEDER 1994 ; for review, see JONES 1984 and ROEDER 1995 , ROEDER 1997 ). The observation that chromosomes fail to synapse and crossover interference is completely eliminated in zip1 null mutants has provided molecular evidence for a functional relationship between the SC and interference (SYM and ROEDER 1994 ).

In this study of zip1 deletion mutations, we found that all of the mutants that make stable, full-length SC display crossover interference. Even in zip1-M2 and zip1-MC1 strains, which decrease the width of the SC and reduce crossing over, interference is unaffected. In the zip1-NM1 mutant, interference appears to be decreased (although the difference between wild type and the mutant is not statistically significant). However, short stretches of unsynapsed chromosomes have been observed in spread chromosomes of zip1-NM1, raising the possibility that any decrease in interference is caused by the occasional failure in synapsis or in maintaining SC structure. The zip1-NM2 mutant makes considerably less SC than wild type, and the zip1-C1 mutant makes no SC; interference is reduced or absent in these mutants. The zip1-M1 mutant does not exhibit interference, suggesting that the aberrant structure and decreased stability of the SC assembled in this mutant (if it does make SC) renders the complex nonfunctional for transmitting the inhibitory signal responsible for interference. Overall, these data establish a correlation between crossover interference and the formation of stable SC and are thus consistent with the hypothesis that the SC is required for interference.

	ACKNOWLEDGMENTS

We thank PENELOPE CHUA, ANITA HOPPER, JANET NOVAK, BETH ROCKMILL, and ALBERT SMITH for comments on the manuscript. This work was supported by National Institutes of Health grant GM28904 to G.S.R., the Howard Hughes Medical Institute, and by a postdoctoral fellowship (DRG-1318) from the Cancer Research Fund of the Damon Runyon-Walter Winchell Foundation to K.-S.T.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AEBI, U., M. HÄNER, J. TRONCOSO, R. EICHNER, and A. ENGEL, 1988  Unifying principles in intermediate filament (IF) structure and assembly. Protoplasma 145:73-81.

ALANI, E., R. PADMORE, and N. KLECKNER, 1990  Analysis of wild-type and rad50 mutants of yeast suggests an intimate relationship between meiotic chromosome synapsis and recombination. Cell 61:419-436[Medline].

BAUDAT, F. and A. NICOLAS, 1997  Clustering of meiotic double-strand breaks on yeast chromosome III. Proc. Natl. Acad. Sci. USA 94:5213-5218[Abstract/Free Full Text].

BISHOP, D., D. PARK, L. XU, and N. KLECKNER, 1992  DMC1: a meiosis-specific yeast homolog of E. coli recA required for recombination, synaptonemal complex formation, and cell cycle progression. Cell 69:439-456[Medline].

BURNS, N., B. GRIMWADE, P. B. ROSS-MACDONALD, E.-Y. CHOI, and K. FINBERG et al., 1994  Large-scale analysis of gene expression, protein localization and gene disruption in Saccharomyces cerevisiae.. Genes Dev. 8:1087-1105[Abstract/Free Full Text].

CHUA, P. R. and G. S. ROEDER, 1997  Tam1, a telomere-associated meiotic protein, functions in chromosome synapsis and crossover interference. Genes Dev. 11:1786-1800[Abstract/Free Full Text].

DOBSON, M. J., R. E. PEARLMAN, A. KARAISKAKIS, B. SPYROPOULOS, and P. B. MOENS, 1994  Synaptonemal complex proteins: occurrence, epitope mapping and chromosome disjunction. J. Cell Sci. 107:2749-2760[Abstract].

DRESSER, M. E. and C. N. GIROUX, 1988  Meiotic chromosome behavior in spread preparations of yeast. J. Cell Biol. 106:567-578[Abstract/Free Full Text].

EGEL, R., 1978  Synaptonemal complexes and crossing-over: structural support or interference? Heredity 41:233-237[Medline].

EGEL, R., 1995  The synaptonemal complex and the distribution of meiotic recombination events. Trends Genet. 11:206-208[Medline].

GOYON, C. and M. LICHTEN, 1993  Timing of molecular events in meiosis in Saccharomyces cerevisiae: stable heteroduplex DNA is formed late in meiotic prophase. Mol. Cell. Biol. 13:373-382[Abstract/Free Full Text].

ITO, H., Y. FUKADA, K. MURATA, and A. KIMURA, 1983  Transformation of intact yeast cells treated with alkali cations. J. Bacteriol. 153:163-168[Abstract/Free Full Text].

JONES, G. H., 1984 The control of chiasma distribution, pp. 293–320 in Controlling Events in Meiosis, edited by C. W. EVANS and H. G. DICKINSON. The Company of Biologists Ltd., Cambridge.

KING, J. S. and R. K. MORTIMER, 1990  A polymerization model of chiasma interference and corresponding computer simulation. Genetics 126:1127-1138[Abstract].

KLECKNER, N., 1996  Meiosis: how could it work? Proc. Natl. Acad. Sci. USA 93:8167-8174[Abstract/Free Full Text].

LICHTEN, M. and A. S. H. GOLDMAN, 1995  Meiotic recombination hotspots. Annu. Rev. Genet. 29:423-444[Medline].

LIU, J.-G., L. YUAN, E. BRUNDELL, B. BJÖRKROTH, and B. DANEHOLT et al., 1996  Localization of the N-terminus of SCP1 to the central element of the synaptonemal complex and evidence for direct interactions between the N-termini of SCP1 molecules organized head-to-head. Exp. Cell Res. 226:11-19[Medline].

LOIDL, J., F. KLEIN, and H. SCHERTHAN, 1994  Homologous pairing is reduced but not abolished in asynaptic mutants of yeast. J. Cell Biol. 125:1191-1200[Abstract/Free Full Text].

LUPAS, A., M. VAN DYKE, and J. STOCK, 1991  Predicting coiled coils from protein sequences. Science 252:1162-1164[Medline].

MAGUIRE, M. P., 1988  Crossover site determination and interference. J. Theor. Biol. 134:565-570[Medline].

MEUWISSEN, R. L. J., H. H. OFFENBERG, A. J. J. DIETRICH, A. RIESEWIJK, and M. VAN IERSEL et al., 1992  A coiled-coil related protein specific for synapsed regions of meiotic prophase chromosomes. EMBO J. 11:5091-5100[Medline].

MEUWISSEN, R. L. J., I. MEERTS, J. M. N. HOOVERS, N. J. LESCHOT, and C. HEYTING, 1997  Human synaptonemal complex protein 1 (SCP1): isolation and characterization of the cDNA and chromosomal localization of the gene. Genomics 39:377-384[Medline].

MOSES, M. J., 1968  Synaptinemal complex. Annu. Rev. Genet. 2:363-412.

NAG, D. K. and T. D. PETES, 1993  Physical detection of heteroduplexes during meiotic recombination in the yeast Saccharomyces cerevisiae.. Mol. Cell. Biol. 13:2324-2331[Abstract/Free Full Text].

NAG, D. K., H. SCHERTHAN, B. ROCKMILL, J. BHARGAVA, and G. S. ROEDER, 1995  Heteroduplex DNA formation and homolog pairing in yeast meiotic mutants. Genetics 141:75-86[Abstract].

PADMORE, R., L. CAO, and N. KLECKNER, 1991  Temporal comparison of recombination and synaptonemal complex formation during meiosis in S. cerevisiae.. Cell 66:1239-1256[Medline].

PETES, T. D., R. E. MALONE and L. S. SYMINGTON, 1991 Recombination in yeast, pp. 407–521 in The Molecular and Cellular Biology of the Yeast Saccharomyces: Genome Dynamics, Protein Synthesis, and Energetics, edited by J. R. BROACH, J. R. PRINGLE and E. W. JONES. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

ROCKMILL, B. and G. S. ROEDER, 1990  Meiosis in asynaptic yeast. Genetics 126:563-574[Abstract].

ROCKMILL, B. and G. S. ROEDER, 1991  A meiosis-specific protein kinase homolog required for chromosome synapsis and recombination. Genes Dev. 5:2392-2404[Abstract/Free Full Text].

ROCKMILL, B., M. SYM, H. SCHERTHAN, and G. S. ROEDER, 1995  Roles for two RecA homologs in promoting meiotic chromosome synapsis. Genes Dev. 9:2684-2695[Abstract/Free Full Text].

ROEDER, G. S., 1995  Sex and the single cell: meiosis in yeast. Proc. Natl. Acad. Sci. USA 92:10450-10456[Abstract/Free Full Text].

ROEDER, G. S., 1997  Meiotic chromosomes: it takes two to tango. Genes Dev. 11:2600-2621[Free Full Text].

ROTHSTEIN, R., 1991  Targeting, disruption, replacement and allele rescue: integrative DNA transformation in yeast. Methods Enzymol. 194:281-301[Medline].

SAGE, J., L. MARTIN, F. CUZIN, and M. RASSOULZADEGAN, 1995  DNA sequence of the murine synaptonemal complex protein 1 (SCP1). Biochim. Biophys. Acta 1263:258-260[Medline].

SÄLL, T. and B. O. BENGTSSON, 1989  Apparent negative interference due to variation in recombination frequencies. Genetics 122:935-942[Abstract/Free Full Text].

SCHMEKEL, K. and B. DANEHOLT, 1995  The central region of the synaptonemal complex revealed in three dimensions. Trends Cell Biol.