Red1p, a MEK1-dependent Phosphoprotein That ... - Semantic Scholar
THE JOURNAL OF BIOLOGICAL CHEMISTRY © 1999 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 274, No. 3, Issue of January 15, pp. 1783–1790, 1999 Printed in U.S.A.
Red1p, a MEK1-dependent Phosphoprotein That Physically Interacts with Hop1p during Meiosis in Yeast* (Received for publication, July 9, 1998, and in revised form, October 20, 1998)
Teresa de los Santos and Nancy M. Hollingsworth‡ From the Department of Biochemistry and Cell Biology, Institute for Cell and Developmental Biology, SUNY Stony Brook, Stony Brook, New York 11794-5215
The synaptonemal complex (SC) is a proteinaceous structure formed between pairs of homologous chromosomes during prophase I of meiosis. The proper assembly of axial elements (AEs), lateral components of the SC, during meiosis in the yeast, Saccharomyces cerevisiae, is essential for wild-type levels of recombination and for the accurate segregation of chromosomes at the first meiotic division. Genetic experiments have indicated that the stoichiometry between two meiosis-specific components of AEs in S. cerevisiae, HOP1 and RED1, is critical for proper assembly and function of the SC. A third meiosis-specific gene, MEK1, which encodes a putative serine/threonine protein kinase, is also important for proper AE function, suggesting that AE formation is regulated by phosphorylation. In this paper, we demonstrate that Mek1p is a functional kinase in vitro and that catalytic activity is an essential part of the meiotic function of Mek1 in vivo. Immunoblot analysis revealed that Red1p is a MEK1-dependent phosphoprotein. Co-immunoprecipitation experiments demonstrated that the interaction between Hop1p and Red1p is enhanced by the presence of MEK1. Thus, MEK1-dependent phosphorylation of Red1p facilitates the formation of Hop1p/Red1p hetero-oligomers, thereby enabling the formation of functional AEs.
Sexual reproduction requires the formation of haploid gametes so that fertilization can reconstitute the diploid chromosome number of the organism. The chromosome number of the cell is divided precisely in half by meiosis, a highly conserved and specialized type of cell division. Specifically, the reduction in chromosome number is accomplished by the segregation of homologous chromosomes to opposite poles at the first meiotic division. Proper disjunction requires crossovers (typically manifested cytologically as chiasmata) that provide the physical connections between homologs necessary for proper orientation on the Meiosis I spindle (1). Mutants that interfere with the formation of crossovers produce chromosomally imbalanced gametes that are either inviable themselves or fuse to create inviable zygotes (2). In most organisms, two requirements must be met for wildtype levels of crossovers to occur during meiosis. First, there must be recombination machinery to physically exchange the DNA from non-sister chromatids. Second, a tripartite protein-
* This work was supported by a grant from the Pew Charitable Trusts, as well as National Institutes of Health Grant GM50717. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ To whom correspondence should be addressed. Tel.: 516-632-8581; Fax: 516-632-8575; E-mail: [email protected]. This paper is available on line at http://www.jbc.org
aceous structure called the synaptonemal complex (SC)1 must form between homologous chromosomes (3, 4). SC formation is initiated by the condensation of sister chromatids along a protein core called an axial element (AE). Synapsis occurs when AEs are held together by the introduction of central region components to make mature SCs. In Saccharomyces cerevisiae, deletion of RED1, a gene encoding a component of AEs, eliminates AE formation and therefore red1 mutant phenotypes may indicate functions for AEs (5). In red1 diploids, spore viability is drastically reduced, despite only a 4-fold reduction in crossing over, suggesting a role for AEs in packaging crossovers in such a way that the chiasmata can be used for directing Meiosis I chromosome segregation (5, 6). RED1 also plays a role in a checkpoint that monitors whether recombination is progressing correctly or not. For example, mutations in the meiosis-specific recA homolog DMC1 cause arrest prior to the first meiotic division, presumably as a result of the formation of aberrant recombination intermediates (7). In the absence of RED1, this arrest is abolished despite the fact that the aberrant recombination intermediates are still being generated. It appears, therefore, that the packaging of recombination intermediates into some sort of RED1-dependent chromosomal structure is necessary for the recombinationmonitoring checkpoint to work (8). Understanding AE assembly is clearly a key to understanding axial element function. Genetic and cytological experiments have identified two other meiosis-specific genes, HOP1 and MEK1 (also known as MRE4), that are important for formation of functional AEs in yeast. Evidence that the stoichiometry of Red1p to Hop1p is important in axial element formation comes from genetic experiments in which overexpression of RED1 in the presence of limiting amounts of Hop1p created a dominant negative phenotype (9). This dominant negative phenotype can also be produced by overexpressing RED1 in the presence of wild-type levels of HOP1 but with limiting amounts of functional Mek1p (10). Mek1p therefore appears to influence the Hop1p:Red1p stoichiometry. This paper presents biochemical data supporting the model that the interaction of Hop1p and Red1p with each other is promoted by phosphorylation of Red1p by Mek1p (10). EXPERIMENTAL PROCEDURES
Yeast Strains and Media—The genotypes of the strains used in this study are shown in Table I. All the strains are isogenic derivatives of the SK1-related diploid NH144 (11). Diploid strains mutated for RED1 or MEK1 were created by first transforming S2683 and RKY1145, the two haploid parents of NH144, with the appropriate DNA fragments (12) followed by mating. For the red1::LEU2 mutation, a 4.9-kb BglII/ PstI fragment was purified from pNH119 (13) and used for transforma-
1 The abbreviations used are: SC, synaptonemal complex; AE, axial element; HA, hemagglutinin; GST, glutathione S-transferase; PCR, polymerase chain reaction; kb, kilobase(s); PMSF, phenylmethylsulfonyl fluoride.
1783
1784
Red1p, a MEK1-dependent Phosphoprotein TABLE II Plasmid list
TABLE I Saccharomyces cerevisiae SK1 strains Name
RKY1145 S2683 NH144 YTS1 YTS1::pLP36 YTS1::pLP37 YTS3 YTS3::pTS23 YTS4 NH177 NH211 DW10
Genotype
Name
MATa leu2D::hisG his4-x lys2 ura3 hoD::LYS2 MATa leu2-K lys2 ura3 hoD::LYS2 arg4-Nsp MATa leu2D::hisG his4-x lys2 ura3 hoD::LYS2 ARG4 MATa leu2-K HIS4 lys2 ura3 hoD::LYS2 arg4-Nsp same as NH144 except mek1D::LEU2 mek1D::LEU2 same as YTS1 except ura3::mek1-K199R::URA3 same as YTS1 except ura3::MEK1::URA3 same as NH144 except red1::LEU2 red1::LEU2 same as YTS3 except ura3::RED1–3HA::URA3 ura3::RED1–3HA::URA3 same as NH144 except mek1D::URA3 mek1D::URA3 same as NH144 except red1::LEU2 mek1D::LEU2 red1::LEU2 mek1D::LEU2 same as NH144 except mek1–974 ade2-Bgl mek1–974 ade2-Bgl same as NH144 except hop1::LEU2 RED1–3HA hop1::LEU2 RED1–3HA
pNH83 p1b-1 pNH218 pNH219 pNH124 pB131 pTS21 pTS1 pRD56 pNH168 pVZ1 pNH205 pNH206 pDW14 pDW15 pDW16 pR4C4 pLP31 pLP34 pLP35 pRS306 pLP36 pLP37 pLP41 pNH235 pNH208 SK-HA3 pNH209 pNH212 pTS23
tion. This insertion allele exhibits a null phenotype with respect to spore viability (other phenotypes have not been tested). A 3.6-kb XbaI/ BamHI fragment from pTS1 carries the mek1D::LEU2 allele, and a 2.9-kb BamHI fragment from pTS21 was used to introduce mek1D::URA3. These deletions remove 75 base pairs upstream of the ATG as well as approximately 75% of the MEK1 coding sequence. The presence of the red1::LEU2, mek1D::LEU2, and mek1D::URA3 alleles was confirmed both genetically and by Southern blot analysis. The ade2-Bgl mutation was introduced by two-step gene replacement (12) using pR943 (generously supplied by B. Rockmill, Yale University) cut with SpeI. Haploids containing the ade2-Bgl mutation and appropriate other markers were obtained by crosses. These haploids were then mated to give homozygous diploids. The mek1–974 mutation was introduced into two different mek1D::LEU2 haploids by two-step gene replacement using pNH235 digested with BssHI to target integration to the mek1D::LEU2 locus. Leu2 5FoaR colonies were then mated to create the mek1–974 diploid NH211. The red1::LEU2 mek1D::LEU2 homozygous diploid NH177 was constructed by crossing a MATa mek1D::LEU2 haploid with MATa red1::LEU2 haploid and dissecting tetrads from the resulting heterozygous diploid. Leu1 spore colonies that exhibited a non-parental di-type pattern for LEU2 were then mated to produce the double homozygote. To make the RED1–3HA homozygous diploid YTS3::pTS23, S2683red1 and RKY1145red1 were transformed with pTS23 digested with StuI to target integration of the plasmid to URA3. The resulting transformants were mated to create the diploid. For YTS1:pLP36 and YTS1::pLP37, the plasmids pLP36 and pLP37 were digested with StuI to target integration to URA3 and transformed into YTS1. Standard yeast genetic methods were used (14). Solid media have been described (15). YPA contains 1% yeast extract, 2% bactopeptone, and 2% potassium acetate. Sporulation was performed using 2% potassium acetate. Plasmids—Plasmids for this study were made by standard procedures (16) using the Escherichia coli strain BSJ72 and are listed in Table II. To construct the mek1D::URA3 and mek1D::LEU2 deletion alleles, the polymerase chain reaction (PCR) was used to generate fragments of URA3 or LEU2 with SacI and PvuII ends. The PCR fragments were digested with SacI and PvuII and ligated to pB131 (17) cut with SacI and HpaI to generate pTS21 (mek1D::URA3) and pTS1 (mek1D::LEU2), respectively. To fuse the MEK1 gene to GST (glutathione S-transferase) (18), PCR was employed to generate a fragment containing the entire MEK1 gene. The resulting fragment was digested with BglII and ligated into the BamHI site of pRD56 to create pNH168. The mek1-K199R mutation was introduced into the GST-MEK1 fusion gene by oligonucleotide site-directed mutagenesis (19). The 2.9-kb NotI/SalI fragment from pNH168 containing GAL1p-GST-MEK1 was cloned into NotI/SalI-digested pVZ1 to generate pNH205. The presence of the mutation creates an FspI restriction site that was used for screening plasmids with the desired mutation. The plasmid containing the mek1-K199R allele was designated pNH206. To place the GST-MEK1 and GST-mek1-K199R alleles under control of the MEK1 promoter, the MEK1 promoter was first cloned into the 2m vector, YEplac181. A 924-bp fragment starting at position 2896 rela-
a b
Relevant yeast genotype
HOP1 URA3 2m RED1 URA3 2m RED1 URA3 2m RED1 HOP1 URA3 2m RED1 MEK1 URA3 CEN4 ARS1 mek1D::URA3 URA3 CEN4 ARS1 mek1D::LEU2 URA3 CEN4 ARS1 GAL1p-GST URA3 CEN6 ARSH4 GAL1p-GST-MEK1 URA3 CEN6 ARSH4 pBS1 with extended polylinker GAL1p-GST-MEK1 GAL1p-GST-mek1-K199R MEK1p LEU2 2m MEKp1-GST-MEK1 LEU2 2m MEK1p-GST-mek1-K199R LEU2 2m MEK1 URA3 ARS1 CEN4 MEK1 mek1-K199R MEK1 URA3 mek1-K199R URA3 MEK1 URA3 mek1–974 URA3 CEN4 ARS1 mek1–974 URA3 RED1 (EcoRI site in place of stop codon) 3 HA epitopes in SK2 RED1–3HA RED1–3HA URA3 2m RED1–3HA URA3
Source
(13) (13) (10) (10) (10) (17) This work This work A. Neimana This work S. Henikoffb This work This work This work This work This work (17) This work This work This work (38) This work This work This work This work This work (21) This work This work This work
SUNY Stony Brook. Fred Hutchinson Cancer Research Center.
tive to the MEK1 ATG and ending at 128 was amplified by PCR. The oligonucleotide was engineered to put an NdeI site at the ATG of MEK1 with a SalI site downstream. The fragment was digested with BamHI and SalI and ligated into BamHI/SalI-digested YEplac181 to make pDW14. The 2.9-kb NdeI/SalI fragments containing either GST-MEK1 (pNH168) or GST-mek1-K199R (pNH206) were then cloned into pDW14 cut with NdeI and SalI to make pDW15 and pDW16, respectively. The MEK1 and mek1-K199R alleles were completely sequenced to confirm that no unknown mutations were introduced by either the PCR or site-directed mutagenesis. Sequencing was performed using the ABI Amplitaq sequencing kit (Perkin-Elmer) followed by electrophoresis on an ABI model 373A automated sequencer. To assay for complementation of the mek1D by mek1-K199R, both the MEK1 and mek1-K199R alleles were put under control of the MEK1 promoter without the GST fusion. To accomplish this, a 5.5-kb EcoRI fragment from pR4C4 (10) was first cloned into EcoRI-digested pVZ1 to generate pLP31. After digestion of pLP31 with SpeI and SalI, the vector backbone was ligated with the 1.2-kb SpeI/SalI fragment from either pNH206 or pNH168, resulting in the substitution of approximately 80% of the MEK1 coding sequence. The resulting plasmids, pLP34 (mek1K199R) and pLP35 (MEK1) were then digested with NotI and SalI, and the 2.6-kb fragments were ligated into pRS306 to make pLP36 and pLP37, respectively. The mek1–974 mutation was rescued onto a plasmid by gap repair (20). The MEK1 plasmid, pB131, was cut with SacI and HpaI and transformed into the mek1–974 diploid NH104 (10). The resulting plasmid, pLP41, was used in sequencing reactions with primers located at staggered intervals along the MEK1 gene. To make the mek1–974 integrating plasmid, pNH235, a 3.2-kb EcoRI/SalI fragment from pLP41 was ligated into the URA3 vector, pRS306, digested with EcoRI and SalI. The first step in the construction of the RED1–3HA allele was the introduction of an EcoRI site in place of the RED1 stop codon by PCR. A 1.3-kb fragment containing the 39-half of RED1 was amplified, digested with BglII and EcoRI, and ligated into BglII/EcoRI-digested pNH124 (10) creating pNH208. A 3.4-kb SalI/EcoRI fragment from pNH208 carrying the entire RED1 gene with ;1.1 kb of upstream sequences was cloned into SalI/EcoRI SK-HA3 (21). This ligation fuses the full-length RED1 gene in-frame at the 39-end with three copies of the hemagglutinin (HA) epitope in pNH209. The tagged allele was moved either to a high copy 2m plasmid (pNH212) or to a URA3integrating plasmid (pTS23) by ligating the 3.5-kb SalI/SacI fragment
Red1p, a MEK1-dependent Phosphoprotein from pNH209 into YEp352 and pRS306, respectively. Antibodies—The mouse monoclonal a-HA antibody is contained in the 12CA5 ascites fluid that was purchased from Babco. Affinity-purified a-Gstp antibodies were generously provided by D. Kellogg (University of California Santa Cruz). Antibodies directed against the fulllength native Hop1 protein were generated as follows. The Hop1 protein was overexpressed in vegetative yeast cells and purified to .95% homogeneity (22). 0.5 mg of purified Hop1 protein was injected into two rabbits followed by a boost of 0.25 mg three weeks later (performed by Babco). The resulting sera were found to cross-react with a band of ;67 kDa, the predicted molecular mass of Hop1p, when cells carrying the galactose-inducible GAL10p-HOP1 gene (22) were grown in galactose. This band was absent when the cells were grown on glucose. The 67-kDa band was also not observed using the preimmune sera from these rabbits. Further evidence that these antibodies are specific for Hop1p is that the 67-kDa protein is only observed in sporulating cultures when HOP1 is present.2 Sporulation—2-ml Yeast extract-peptone-dextrose cultures were inoculated with single colonies of the appropriate diploid and grown to saturation at 30 oC with aeration. Strains containing plasmids were inoculated with colonies grown on selective medium. Plasmid stability was monitored immediately prior to transfer to sporulation medium and .80% of the cells were typically found to contain plasmid. The yeast extract-peptone-dextrose cultures were diluted 1:1000 –1:3000 into yeast peptone acetate medium and grown for approximately 16 h with aeration at 30 oC. When cultures reached an OD660 5 1–2, the cells were pelleted, washed with water, and resuspended in 2% potassium acetate to a final concentration of 3 3 107 cells/ml. The cells were shaken at 30 oC for 3 h and then harvested by centrifugation for 10 min at 5000 rpm in a GSA rotor. Prior to centrifugation, a 1-ml aliquot was removed to the 30 oC roller, and the percent sporulation was assessed by light microscopy the following day. Sporulation typically reached levels of .90%. Kinase Assays—After 3 h in sporulation medium, 20-ml aliquots of cells were washed with water, and the cell pellets were quick frozen at 270 oC. Kinase assays were performed as described (23). Immunoprecipitation of Hop1p and Red1–3HAp from Meiotic Extracts—Sporulating cells were harvested, washed one time with water, and resuspended in 0.3 ml of lysis buffer (25 mM Tris-HCl, pH 7.5, 200 mM NaCl, 5 mM EDTA) per 20 ml of cells. To reproducibly detect Red1–3HAp, it was necessary to prepare extracts the same day as the cells were sporulated, i.e. the cells could not be frozen first. The reason for this phenomenon is not understood but is probably not because of degradation of the Red1–3HA protein. After extracts were prepared, the proteins could be frozen at 270 oC and thawed with no loss in the ability to detect Red1–3HAp by immunoblot analysis. Protease inhibitors (1 mM PMSF, 1 mg/ml aprotinin, 1 mg/ml leupeptin, 1 mg/ml pepstatin) and phosphatase inhibitors (10 mM NaF, 10 mM Na4P2O7, 1 mM Na3VO4) were present throughout the lysis and wash steps. One gram of glass beads was added, and the tubes were vortexed three times for 30 s with 2-min intervals on ice. The lysate was removed to a clean tube, and detergents were added to final concentrations of 1% Triton X-100, 0.5% sodium deoxycholate, and 0.1% SDS. The lysate was rocked for 15 min at 4 oC and cleared by centrifugation at 14,000 3 g for 10 min. The supernatant was divided into two tubes to which either 1–2 ml of a-Hop1p or a-HA antibodies was added, followed by rocking for 2 h at 4 oC. To precipitate the immune complexes, 40 ml of protein A-Sepharose (Amersham Pharmacia Biotech) slurry (equilibrated 1:1 with lysis buffer) were added. The tubes were then rocked again for an additional hour at 4 oC. The Sepharose beads were washed four times with 0.5 ml of lysis buffer and resuspended in 30 ml of 23 SDS protein sample buffer. The beads were heated at 95 oC for 5 min and loaded in duplicate onto a 6% SDS-polyacrylamide gel. After fractionation, the proteins were blotted to nitrocellulose and probed for Hop1p and Red1–3HAp using a-Hop1p and a-HA antibodies, respectively. The presence of the antibodies was determined using the ECL kit from Amersham Pharmacia Biotech. The amount of Hop1p and Red1–3HAp present in the immunoprecipitates was quantitated as described in the instructions for the ECL kit (Amersham Pharmacia Biotech). A series of dilutions from the soluble extracts used in the experiment was fractionated on the same gel and probed simultaneously with the appropriate antibodies. After the ECL reaction, the films were scanned using a Bio-Rad imaging densitometer. A standard curve of protein concentration was generated for Hop1p and Red1–3HAp using the Molecular Analyst, Version 1.1
2
N. Hollingsworth, unpublished observations.
1785
software (Bio-Rad). The amount of Hop1p and Red1–3HAp present in the immunocomplexes was then determined using the standard curves. To compare the efficiency of the co-immunoprecipitation in MEK1 and mek1D extracts, the following formula was used. ~Co-IP proteinMEK1/IP proteinMEK1)/ (Co-IP proteinmek1D/IP proteinmek1D)
(Eq. 1)
Phosphatase Experiments—For each sample, 20 ml of sporulating YTS3/pNH212 cells were used for immunoprecipitation of Red1–3HAp. The immunoprecipitates were treated with l protein phosphatase (New England Biolabs) as described (24). RESULTS
Overexpression of HOP1 Suppresses mek1–974 in the SK1 Background whereas Overexpression of RED1 Exacerbates the mek1–974 Mutant Phenotype—Genetic interactions between MEK1, HOP1, and RED1 in the A364A strain background led to the proposal that phosphorylation of either Hop1p or Red1p by Mek1p facilitates the formation of Hop1p/Red1p heterooligomers, thereby ensuring the proper balance between the proteins in the assembly of axial elements during meiosis (10). To test this model biochemically, it was desirable to use the SK1 strain background in which sporulation proceeds rapidly, relatively synchronously, and with high efficiency (25). To ensure that the genetic dosage effects of HOP1 and RED1 overexpression in the presence of the leaky mek1–974 mutant occur in SK1 as well as A364A, it was necessary to repeat the dosage experiments in SK1. The mek1–974 mutation was cloned by gap repair using the diploid NH104 (10), and the entire gene was sequenced. A single transition mutation of Gly to Ala at nucleotide 502 was found. This mutation creates a valine to methionine substitution at amino acid 168, which is located immediately before the GXG motif in the first conserved kinase domain (26). The mek1–974 allele was introduced into the SK1 background by two-step gene replacement, thereby creating the homozygous diploid, NH211. This diploid was transformed with the same plasmids used in the previous analysis (10), the transformants were sporulated, and tetrads were dissected to assess spore viability. As in the A364A background, the mek1–974 mutant is partially functional in SK1, producing 33.2% viable spores compared with ;1% for the deletion (Ref. 27 and Table III). The spore viability defect is complemented by MEK1 on a CEN plasmid (Table III, NH211). The presence of excess HOP1 greatly improved the mutant phenotype, resulting in 74% viable spores. In contrast, overexpression of RED1 decreased the number of viable spores from 33.2 to 16.8% (Table III, NH211). Both the HOP1 suppression and the RED1 exacerbation of the mek1–974 spore inviability phenotype are statistically significant by x2 analysis (p , 0.001). The one difference observed between SK1 and A364A was the result obtained when both HOP1 and RED1 were co-overexpressed on the same plasmid. In A364A, HOP1/RED1 co-overexpression restored the spore viability observed for vector alone. In SK1, a suppression phenotype similar to that for HOP1 alone was found (Table III, NH211). Detection of Red1p in Meiotic Cells Using the Hemagglutinin Epitope Tag—Because attempts to use polyclonal antibodies directed against Red1p (28) to detect the Red1p protein by immunoblot analysis were unsuccessful,3 an epitope-tagged version of the protein was created. The RED1 gene was fused in frame at the 39-end to three copies of the hemagglutinin epitope. The fusion gene was then integrated at URA3 in two different isogenic SK1 red1::LEU2 haploids. This insertion allele of RED1 exhibits a null phenotype with regard to spore 3
T. de los Santos, unpublished observations.
1786
Red1p, a MEK1-dependent Phosphoprotein TABLE III Spore viability assayed by tetrad dissection
Strain/plasmid
NH211 NH211/pR4C4 NH211/pNH83 NH211/pNH218 NH211/pNH219 YTS3/YEp352 YTS3::pTS23 YTS3/pNH212 YTS3/p1b-1 YTS4 YTS4/pDW15 YTS4/pDW14 YTS1 YTS1::pLP37 YTS1::pLP36
Relevant yeast genotype
% Spore viability
No. asci
mek-974 mek-974 mek1–974/MEK1 CEN mek1–974 mek1–974/HOP1 2m mek1–974 mek1–974/RED1 2m mek1–974 mek1–974/RED1 HOP1 2m mek1–974 red1::LEU2 red1::LEU2 red1::LEU2 RED1–3HA red1::LEU2 RED1–3HA red1::LEU2/RED1–3HA 2m red1::LEU2 red1::LEU2/RED1 2m red1::LEU2 mek1D::URA3 mek1D::URA3 mek1D::URA3/GST-MEK1 2m mek1D::URA3 mek1D::URA3/GST-mek1-K199R mek1D::URA3 mek1D::LEU2 mek1D::LEU2 mek1D::LEU2 MEK1 mek1D::LEU2 mek1D::LEU2 mek1-K199R mek1D::LEU2
33.2
104
92.3
52
74.0
104
16.8
104
69.0
105
3.8
53
,0.5
57
50.0
72
83.5
50
0.3
71
76.8
42
2.1
71
5.0
55
94.9
74
1.8
56
viability (13). The haploids were mated to create a red1::LEU2 diploid that is homozygous for the RED1–3HA allele (YTS3::pTS23). Alternatively, the RED1–3HA gene on a high copy number 2m vector was transformed into YTS3 (YTS3/ pNH212). After 3 h in sporulation medium, a time point when Red1–3HAp and Hop1p concentrations are maximal,3 protein extracts were made from YTS3/YEp352, YTS3::pTS23, and YTS3/pNH212. No protein of the appropriate molecular weight for Red1–3HAp was detected with the a-HA antibodies in YTS3 containing the vector alone (Fig. 1A). These cells were proceeding through meiosis, as evidenced by there being a wild-type amount of Hop1p in the YTS3/YEp352 extract (Fig. 1B). A RED1–3HA-dependent a-HA cross-reacting band was observed in YTS3::pTS23. This protein represents Red1–3HAp as this band is absent from extracts of NH144, an isogenic diploid homozygous for the untagged RED1 gene, as well as from vegetative cultures of YTS3::pTS23.3 Interestingly, in YTS3/ pNH212, where the RED1–3HA allele is present in high copy number, several Red1–3HAp bands are apparent (Fig. 1A). Tetrad dissection was performed to determine the ability of the RED1–3HA allele to complement the spore viability defect of red1::LEU2 in YTS3. YTS3 carrying vector alone produced 3.8% viable spores. When RED1–3HA is homozygous in YTS3::pTS23, no complementation was observed (Table III, YTS3/YEp352). Over-expression of the RED1–3HA allele restored a significant amount of RED1 although the complementation was not as good as overexpression of the untagged allele of RED1 (Table III, YTS3/YEp352). Because the overexpressed RED1–3HA allele was detectable and provided a substantial level of RED1 function (50% spore viability versus 3.8%), this construct was used in all subsequent experiments. Red1–3HAp Is a Phosphoprotein—To determine whether the slower migrating species of Red1–3HAp are because of phosphorylation, the Red1–3HA protein was immunoprecipitated and tested for sensitivity to l protein phosphatase (New England Biolabs). After washing, the immunoprecipitates were resuspended in phosphatase buffer, treated with or without
FIG. 1. Detection of Red1–3HAp and Hop1p in meiotic extracts. The red1::LEU2 diploid, YTS3, was transformed with vector alone (YEp352), pTS23 (two integrated copies of RED1–3HA), or pNH212 (2m RED1–3HA). Cells were sporulated for 3 h, and total protein extracts were made. 20 mg of total extract were fractionated by SDS-PAGE, blotted to nitrocellulose, and probed with either a-HA antibodies (panel A) or a-Hop1p antibodies (panel B).
phosphatase, fractionated by SDS-PAGE, and blotted, and the Red1–3HAp was detected using the a-HA antibody. In the sample where no phosphatase was added, three discrete bands were observed (Fig. 2). When l protein phosphatase was added, the slower mobility bands were no longer visible. The addition of phosphatase inhibitors prior to the phosphatase prevents the loss of the shifted bands (Fig. 2), indicating that it is the phosphatase activity which accounts for the disappearance of the modified bands and not a contaminating protease. One concern was that the phosphorylation is occurring on the HA epitope rather than Red1p. That this is not the case is demonstrated by the findings of Bailis and Roeder (29) who have recently shown that Red1p is a phosphoprotein using antibodies against the endogenous Red1p. Mek1p Has Kinase Activity in Vitro and the Kinase Activity Is Essential for the Meiotic Function of MEK1—A good candidate for the kinase which phosphorylates Red1p is Mek1p. To determine whether the Mek1p kinase homology is meaningful, Mek1p kinase activity was tested in vitro. The MEK1 gene was fused to the 39-end of the affinity tag GST and the GST-MEK1 fusion allele was placed under the control of the MEK1 promoter. As a negative control, an allele of MEK1 (mek1-K199R) was created in which the codon for an invariant lysine residue present in domain II of the kinase (amino acid 199) was mutated to arginine. This lysine to arginine mutation has been shown to reduce or abolish kinase activity in a number of other kinases (30). The mek1D::URA3 diploid YTS4 was transformed with YEplac181 (2m), pDW15 (2m GST-MEK1), or pDW16 (2m GST-mek1-K199R). Cells were sporulated for 3 h, protein ex-
Red1p, a MEK1-dependent Phosphoprotein
1787
FIG. 2. Detection of a MEK1-dependent phosphorylated form of Red1–3HAp in meiotic extracts. Red1–3HAp was immunoprecipitated from 3 mg of soluble protein extracted from YTS3/pNH212 after 3 h in sporulation medium. Immunocomplexes were treated with and without l protein phosphatase and phosphatase inhibitors. Red1–3HAp was detected by immunoblot analysis using a-HA antibodies.
tracts were made, and the Gst-Mek1p and Gst-mek1-K199R fusion proteins were purified using glutathione-Sepharose beads. Kinase activity was assayed by autophosphorylation. The relative amount of Gst-Mek1p and GST-mek1-K199R present during the assay was determined by probing the same blot used for autoradiography with antibodies directed against Gstp. A radioactive band migrating at the predicted molecular mass for Gst-Mek1p was observed; this band is absent in YTS4/ YEplac181 (Fig. 3A). Very little autophosphorylation activity was detectable for the Gst-mek1-K199R protein despite the fact that more protein was present than for Gst-Mek1 (Fig. 3B). The reduction in phosphorylation of the mek1-K199R protein demonstrates that most of the observed kinase activity is because of Mek1p and not to a co-precipitating kinase or to the Gst moiety. Complementation tests were performed to assay the function of the two MEK1 alleles. The GST-MEK1 gene in YTS4 complemented the mek1D spore viability defect while the GSTmek1-K199R allele did not (Table III, YTS4). GST-MEK1 also complemented when integrated into the chromosome, showing that overexpression is not a requirement for activity.2 In addition, MEK1 and mek1-K199R (without GST) were integrated at URA3 in the mek1D::LEU2 diploid YTS1. Full complementation of the spore viability defect was seen in the YTS1::pLP37 diploid carrying MEK1 compared with YTS1 alone (94.5% versus 5.0% viable spores; Table III, YTS4). In contrast, the mek1K199R allele fails to complement (Table III, YTS4). Therefore, mutation of the invariant lysine to arginine at position 199 creates an allele of MEK1 that exhibits a null phenotype with regard to spore viability. Because this mutation also greatly reduces Mek1p kinase activity in vitro, we conclude that the kinase activity of Mek1p is essential for its meiotic function. The Bulk of Red1–3HAp Phosphorylation Is Dependent on MEK1—To test whether Red1–3HAp is a potential substrate for Mek1p, the protein was examined in MEK1 (YTS3/ pNH212)and mek1D (NH177/pNH212) diploids. Extracts were prepared from cells after 3 h in sporulation medium. Because the total amount of Red1–3HAp present in mek1D diploids is reduced,3 nearly three times as much total protein from the mek1D was loaded onto the gel compared with MEK1. Although there is a similar amount of the unphosphorylated form of Red1–3HAp in both strains, the vast majority of phosphorylated Red1–3HAp is absent in the mek1D diploid (Fig. 4). Red1– 3HAp phosphorylation is therefore dependent on MEK1. This experiment does not, however, address whether this dependence is because of direct phosphorylation of Red1–3HAp by Mek1p or whether an intermediate kinase is involved. When the soluble fraction is examined, a low level of phosphorylated Red1–3HAp is sometimes observed. Because the spore viability of the mek1D diploid overexpressing RED-3HA is very low (1.3%, 77 asci), any phosphorylated Red1–3HAp that is present is apparently non-functional. Whether this “promiscuous” phosphorylation of Red1–3HAp is because of the overexpres-
FIG. 3. Autophosphorylation of Gst-Mek1p from meiotic extracts. Cells were sporulated for 3 h and extracts from YTS4 (mek1D/ mek1D) containing vector alone (YEplac181), GST-MEK1 (pDW15), or GST-mek1-K199R (pDW16) were treated with glutathione-Sepharose to affinity purify the GST fusion proteins. The purified proteins were treated with [g-32P]ATP:ATP to assay for autophosphorylation. The proteins were examined by phosphoimager analysis to detect incorporated radioactivity (panel A). The same blot was subsequently probed with a-Gstp antibodies (panel B).
FIG. 4. Detection of Red1–3HAp in MEK1 and mek1D diploids. Total cell extracts from YTS3/pNH212 (MEK1/MEK1) and NH177/ pNH212 (mek1D/mek1D) were examined for Red1–3HAp using immunoblot analysis and a-HA antibodies. For YTS3/pNH212, 35 mg were loaded on the gel; for NH177/pNH212, 90 mg were used.
sion of the protein or whether it normally occurs is not known. Hop1 and Red1–3HAp Co-immunoprecipitate from Meiotic Cell Extracts—Co-immunoprecipitation experiments were performed to determine whether Hop1p and Red1–3HAp are physically associated in meiotic cells. YTS3/pNH212 was sporulated for 3 h, protein extracts were made, and equal amounts of protein were incubated with either a-HA antibodies or a-Hop1p antibodies to immunoprecipitate Red1–3HAp and Hop1p, respectively. The specificity for Hop1p immunoprecipitation was shown by the fact that no Hop1 is detected in immunoprecipitates from sporulating cells deleted for HOP1 (Fig. 5, compare lanes 5 and 6) nor is Hop1 precipitated if the a-Hop1p antibody is omitted (Fig. 5, lane 4). Similarly, no Red1–3HAp is observed when the a-HA antibody is used for immunoprecipitation in diploids carrying the untagged RED1 gene (Fig. 5, compare lanes 2 and 3) or if the a-HA antibody is left out when precipitating from diploids containing the RED1–3HA gene (Fig. 5, lane 4). The Hop1p and Red1p immunoprecipitated samples were
1788
Red1p, a MEK1-dependent Phosphoprotein
FIG. 5. Specificity of Hop1 and Red1–3HAp immunoprecipitation. Cells from various diploids were sporulated for 3 h and soluble extracts prepared. Hop1p and Red1–3HAp were immunoprecipitated from these extracts using either a-Hop1p or a-HA antibodies. Lanes 1, 2, 4, and 5, YTS3/pNH212 (HOP1/HOP1 red1::LEU2/red1::LEU2/2 m RED1–3HA); lane 3, YTS3/p1b-1 (HOP1/HOP1 red1::LEU2/ red1::LEU2/2 m RED1); lane 6, DW10::pTS23 (hop1::LEU2/hop1 ::LEU2 RED1–3HA/RED1–3HA; and lane 7, YTS3::pTS23 (HOP1/ HOP1 red1::LEU2/red1::LEU2 RED1–3HA/RED1–3HA. Lanes 1–3 were probed with a-HA antibodies to detect Red1–3HAp; and lanes 4 –7 were probed with a-Hop1p antibodies.
divided in half, fractionated by SDS-PAGE, transferred to nitrocellulose and probed with either a-HA antibodies or a-Hop1p antibodies to look for co-immunoprecipitation of Red1–3HAp with Hop1p and vice versa. In YTS3/pNH212, Hop1p co-immunoprecipitated with Red1–3HAp (Fig. 6A, lane 3). The converse is also true; when Hop1p was first precipitated with a-Hop1p antibodies, Red1–3HAp was detected in the complex (Fig. 6B, lane 1). The Absence of MEK1 Reduces the Efficiency of the Hop1p/ Red1–3HAp Co-immunoprecipitation—To test whether the MEK1-dependent phosphorylation of Red1–3HAp has any effect on the ability of Red1–3HAp to interact with Hop1p, the co-immunoprecipitation experiment was performed using extracts from the mek1D strain, NH177/pNH212. Because there is less soluble Hop1p and Red1–3HAp in the absence of MEK1, twice as much protein was used for the immunoprecipitation compared with the MEK1 diploid. When the a-Hop1p antibodies were incubated with the protein extracts, nearly equivalent amounts of Hop1p were precipitated from both strains (Fig. 6B, lanes 3 and 4). However, when these precipitates were probed with the a-HA antibodies, a 4.3-fold reduction in the amount of Red1–3HAp co-precipitating was observed with the mek1D extract (Fig. 6B, lane 2). In three experiments where the amount of Red1–3HAp co-immunoprecipitating with Hop1 was quantitated, a 4.3-, 4.1-, and 3.6-fold reduction was observed in the mek1D diploid compared with MEK1. The reduction in the Red1–3HAp co-IP was not because of a lack of Red1–3HAp in the mek1D diploid as evidenced by the amount of Red1–3HAp present in the a-HA immunocomplexes from the mek1D diploid compared with MEK1 (Fig. 6A, lanes 1 and 2). The amount of Hop1p co-immunoprecipitating with Red1–3HAp varied from a reduction of 1.8- (the value for the experiment shown in Fig. 6, lanes 3 and 4), 1.5-, to 0-fold change in the mek1D diploid compared with MEK1. To rule out a loading artifact, blots probed with a-Hop1p antibodies were stripped and probed with a-HA antibodies and vice versa, and the same results were observed.3 The presence of MEK1 therefore facilitates the Hop1p/Red1–3HAp interaction in vivo. DISCUSSION
Genetic studies in the A364A background led to a model which says that Mek1p phosphorylation facilitates the assem-
FIG. 6. Hop1p and Red1–3HAp co-immunoprecipitation from meiotic extracts. Cells from YTS3/pNH212 (MEK1/MEK1) and NH177/pNH212 (mek1D/mek1D) were sporulated for 3 h. Hop1p and Red1–3HAp were immunoprecipitated from soluble extracts prepared from these cells (2.8 mg for YTS3/pNH212 and 6.6 mg for NH177/ pNH212) using a-HA antibodies (panel A) or a-Hop1p antibodies (panel B). Immunoprecipitates were then probed with either a-HA (to detect Red1–3HA) or a-Hop1p antibodies as noted beneath the pictures. The cross-reacting band at the bottom of panel B, lanes 3 and 4 is because of detection of the antibody used in the immunoprecipitation.
bly of Red1p and Hop1 in a particular stoichiometry necessary to generate functional AEs during SC formation (10). To test this model biochemically, it was desirable to use the SK1 strain background, because the cells sporulate efficiently and synchronously. However because mutants in MEK1 and HOP1 exhibit more severe phenotypes in the SK1 background compared with A364A, it was possible that the model based on the findings from the A364A strain background would not apply to the SK1 background. For mek1D diploids, spore viability is reduced to # 1% in SK1 (27) compared with 15% in A364A (10). For hop1 mutants, spore viability is ,1% in both strain backgrounds, but no residual recombination is observed in SK1 (6, 31, 32). To interpret the biochemical results, it was therefore important to determine whether the same dosage relationships between mek1–974 and overexpression of HOP1 and RED1 observed in A364A exist in SK1. The mek1–974 allele used for the dosage experiments in A364A was cloned by gap repair and introduced into SK1. The mek1–974 allele is partially functional in SK1, the same as in A364A. Overexpression of HOP1 in the SK1 mek1–974 diploid resulted in partial suppression of the spore viability defect. In contrast, overexpression of RED1 decreased spore viability. Co-overexpression of HOP1 and RED1 was previously shown to result in a phenotype similar to neither gene being present, consistent with the idea that the two proteins can titrate out each other. In SK1, the HOP1/RED1 co-overexpression phenotype more resembled that of HOP1 alone, suggesting that this strain background may be more sensitive to the levels of HOP1, a result consistent with the more severe recombination phenotype exhibited by hop1 mutants in this strain. The reproducibility of the dosage results in SK1 indicated that it would be an
Red1p, a MEK1-dependent Phosphoprotein appropriate strain in which to test the stoichiometry model biochemically. A basic assumption of the stoichiometry model is that Mek1p is a serine/threonine protein kinase. This assumption was proven to be correct by demonstrating that Mek1p has autophosphorylation activity in vitro. This activity is greatly reduced by substitution of a conserved lysine in the catalytic domain for arginine—a mutation that abolishes function in vivo. Therefore, the kinase activity of Mek1p is an essential component of its meiotic function. To determine whether either Hop1 or Red1 is a MEK1-dependent phosphoprotein, the two proteins were analyzed by immunoblot analysis for mobility shifts that may be indicative of phosphorylation. Extensive attempts to detect a phosphorylated form of Hop1p were unsuccessful. In contrast, a modified form of Red1–3HAp that is sensitive to phosphatase treatment was observed. The reduced amounts of this modified form in mek1D diploids suggest that Red1–3HAp may be the substrate for Mek1p although the presence of an intermediate kinase cannot be ruled out. In MEK1 strains, the Red1–3HA phosphoprotein was observed only when the RED1–3HA allele was overexpressed, suggesting that the presence of the HA tag may partially inhibit the action of the kinase. The Red1–3HA phosphoprotein is likely to represent the functional form of the protein, as it is only when this species is present that RED1 activity is restored in vivo. Genetic evidence exists for Hop1p and Red1p homo-oligomers as well as for Hop1p/Red1p hetero-oligomers (10, 13, 31). The Hop1p homo-oligomer has recently been demonstrated biochemically using purified protein (22). To test whether Hop1p and Red1p interact in meiotic cells, co-immunoprecipitation experiments were performed. The Hop1p antibodies coprecipitated both the phosphorylated and unphosphorylated forms of Red1–3HAp. The presence of unphosphorylated Red1– 3HAp in the Hop1p immune complex is not surprising given that Mek1p phosphorylation is not essential for the Hop1p/ Red1p interaction because stretches of SC are observed even in mek1D diploids (17). In addition, although the stoichiometry between Hop1p and Red1p is as yet undetermined, there is usually more Red1–3HAp co-immunoprecipitating with Hop1p than Hop1p that co-precipitates with Red1–3HA (e.g. there is an 8-fold difference in Fig. 6). This is the result expected if the ratio of Red1p:Hop1p is greater than 1:1. If this is the case then when a molecule of Hop1p is precipitated, it may bring down not only the phosphorylated Red1–3HAp with which it was associated but also unphosphorylated Red1–3HAp which was interacting with the Red1–3HA phosphoprotein. It is important to note that these experiments were all performed with an allele of RED1 that is being overexpressed. Because Smith and Roeder (28) have shown that overexpression of RED1 can result in a more continuous staining pattern of Red1p along the chromosome, it may be that the stoichiometries observed here do not reflect the exact stoichiometries present in strains containing endogenous levels of Red1p. Although Mek1p phosphorylation of Red1–3HAp is not essential for the Hop1p/Red1p interaction, we have proposed that it facilitates the formation of the hetero-oligomer in the proper stoichiometry (10). This idea was tested by performing the Hop1p/Red1–3HAp co-immunoprecipitation experiments in a mek1D diploid. While it is possible to immunoprecipitate equivalent amounts of Hop1p from meiotic extracts made from MEK1 and mek1D diploids, the level of Red1–3HAp co-precipitating was reduced approximately 4-fold in the absence of MEK1. When Red1–3HAp was immunoprecipitated from mek1D and MEK1 extracts, approximately 2-fold less Hop1p was observed in the complex. The presence of MEK1 clearly
1789
TABLE IV Comparison of S. cerevisiae Hop1p and Mek1p with different potential homologs S. c., Saccharomyces cerevisiae; D. m., Drosophila melanogaster; C. a., Candida albicans; A. t., Arabidopsis thaliana; aa, amino acids; I, identity; S, similarity.
S. c. Mek1 3 D. m. loki S. c. Mek1 3 S. p. Mek1 S. c. Mek1 3 C. a. Mek1a
Kinase domain (302 aa)
FHA domain (85 aa)
31% I; 50% S 40% I; 58% S 53% I; 67% S
27% I; 50% S 31% I; 56% S no data
250 aa Hop1 N-terminal domain
S. c. Hop1 3 A. t. Hop1 no. 1 S. c. Hop1 3 A. t. Hop1 no. 2 S. c. Hop1 3 C. a. Hop1b a b
25% I; 50% S 24% I; 46% S 29% I; 57% S
For C. albicans Mek1p, only a 165-amino acid fragment is available. For C. albicans Hop1p, the fragment is 177 amino acids long.
promotes the Hop1p/Red1p interaction. Whether hetero-oligomer formation is enhanced because the Red1p-3HA phosphoprotein has decreased affinity for Red1–3HAp or an increased affinity for Hop1p remains to be determined. Co-immunoprecipitation of Hop1p and Red1–3HAp demonstrates that the two proteins physically interact in meiotic cells, but these experiments do not address whether the interactions are occurring on chromosomes or not. Given that Hop1p and Red1p co-localize on AEs (28), this idea seems likely. The genetic data presented in (9) are consistent with the interpretation of the co-IP experiment that MEK1 facilitates Hop1p/Red1p hetero-oligomer formation. However, the possibility cannot be ruled out that the diminished co-immunoprecipitation of Red1–3HA with Hop1p and vice versa in the mek1D is an artifact resulting from the decreased levels of soluble Red1–3HAp observed when MEK1 is absent. In the latter case, it may be that the MEK1-dependent phosphorylation of Red1p serves some other, as yet undefined, function during meiosis. One way to distinguish between these two possibilities is to identify and mutate the phosphorylation sites in Red1p and test to see if such mutants have decreased affinity for Hop1p. Such studies are currently underway. The mechanism for assembling AEs by phosphorylation-facilitated protein-protein interactions may be conserved in evolution. Homologs for Mek1p and Hop1 exist in S. pombe and C. albicans (Table IV). Within the amino terminus of Mek1p, but outside of the kinase homology, is a conserved sequence of 56 amino acids called the forkhead-associated domain (FHA) (33). This domain is found almost exclusively in nuclear proteins, including Dun1p and Spk1p, two protein kinases of S. cerevisiae involved in the cellular response to DNA damage (33). Interestingly, in Drosophila there is an ovarian-specific protein kinase called loki that also contains an FHA domain.4 It is intriguing to speculate that loki is the functional homolog of Mek1p in flies, especially given that SC formation only occurs in Drosophila females. The Asy1 gene of Arabidopsis thaliana encodes a protein with significant homology to Hop1p (Table IV), and asy1 mutant plants exhibit phenotypes (greatly reduced synapsis and chiasma formation)5 similar to hop1 mutants in yeast (32). Cor1/SPC3 may be the structural analog to Red1p in mammalian cells. Both proteins are meiosis-specific and are predicted to contain extensive coiled coil domains in their C termini (34, 35). Furthermore, the C termini of both proteins mediate homodimerization in the two-hybrid system (10, 36). SPC3 has been shown to be phosphorylated and, 4
S. Larochelle and B. Suter, personal communication. A. P. Caryl, F. C. H. Franklin, and G. H. Jones, personal communication. 5
1790
Red1p, a MEK1-dependent Phosphoprotein
furthermore, that the pattern of phosphorylation is dynamic throughout meiotic prophase (37). It is possible that the early phosphorylation events observed for SPC3 fulfill the same role in establishing the correct protein stoichiometry in mammalian AEs that it plays in the Hop1p/Red1p interaction. Acknowledgments—We thank Breck Byers, Neta Dean, JoAnne Engebrecht, and Aaron Neiman for helpful discussions and useful comments on the manuscript. We are also grateful to Ann Sutton and Simon Rudge for advice. Aaron Neiman provided the protein alignments given in Table IV. Neta Dean, Doug Kellogg, and Beth Rockmill provided plasmids and/or antibodies. We thank Lisa Ponte and Dana Woltering for excellent technical support. REFERENCES 1. Bascom-Slack, C. A., Ross, L. O., and Dawson, D. S. (1997) Adv. Genet. 35, 253–284 2. Roeder, G. S. (1997) Genes Dev. 11, 2600 –2621 3. Von Wettstein, D., Rasmussen, S. W., and Holm, P. B. (1984) Annu. Rev. Genet. 18, 331– 413 4. Heyting, C. (1996) Curr. Opin. Cell Biol. 8, 389 –396 5. Rockmill, B., and Roeder, G. S. (1990) Genetics 126, 563–574 6. Mao-Draayer, Y., Galbraith, A. M., Pittman, D. L., Cool, M., and Malone, R. E. (1996) Genetics 144, 71– 86 7. Bishop, D. K., Park, D., Xu, L., and Kleckner, N. (1992) Cell 69, 439 – 456 8. Xu, L., Weiner, B. M., and Kleckner, N. (1997) Genes Dev. 11, 106 –118 9. Friedman, D. B., Hollingsworth, N. M., and Byers, B. (1994) Genetics 136, 449 – 464 10. Hollingsworth, N. M., and Ponte, L. (1997) Genetics 147, 33– 42 11. Hollingsworth, N. M., Ponte, L., and Halsey, C. (1995) Genes Dev. 9, 1728 –1739 12. Rothstein, R. (1991) Methods Enzymol. 194, 281–301 13. Hollingsworth, N. M., and Johnson, A. D. (1993) Genetics 133, 785–797 14. Rose, M. D., Winston, F., and Heiter, P. (1990) Methods in Yeast Genetics: A laboratory course manual, Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, NY 15. Vershon, A. K., Hollingsworth, N. M., and Johnson, A. D. (1992) Mol. Cell. Biology 12, 3706 –3714 16. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y 17. Rockmill, B., and Roeder, G. S. (1991) Genes Dev. 5, 2392–2404 18. Smith, D. B., and Johnson, K. S. (1988) Gene 67, 31– 40 19. Smith, M. (1985) Annu. Rev. Genet. 19, 423– 462 20. Orr-Weaver, T. L., and Szostak, J. W. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 4417– 4421 21. Neiman, A. M., Mhaiskar, V., Manus, V., Galibert, F., and Dean, N. (1997) Genetics 145, 637– 645 22. Kironmai, K. M., Muniyappa, K., Friedman, D. B., Hollingsworth, N. M., and Byers, B. (1998) Mol. Cell. Biol. 18, 1424 –1435 23. Neiman, A. M., and Herskowitz, I. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 3398 –3402 24. Altman, R., and Kellogg, D. (1997) J. Cell Biol. 138, 119 –130 25. Padmore, R., Cao, L., and Kleckner, N. R. (1991) Cell 66, 1239 –1256 26. Hunter, T. (1991) in Methods in Enzymology (Hunter, T., and Sefton, B. M., eds) Vol. 200, pp. 3– 81, Academic Press, Inc., San Diego, CA 27. Leem, S.-H., and Ogawa, H. (1992) Nucleic Acids Res. 20, 449 – 457 28. Smith, A. V., and Roeder, G. S. (1997) J. Cell Biol. 136, 957–967 29. Bailis, J. M., and Roeder, G. S. (1998) Genes Dev., 12, 3551–3563 30. Hanks, S. K., Quinn, A. M., and Hunter, T. (1988) Science 241, 42–52 31. Ajimura, M., Leem, S., and Ogawa, H. (1993) Genetics 133, 51– 66 32. Hollingsworth, N. M., and Byers, B. (1989) Genetics 121, 445– 462 33. Hofmann, K., and Bucher, P. (1995) Trends Biochem. Sci. 20, 347–349 34. Dobson, J. J., Pearlman, R. E., Karaiskakis, A., Spryopoulos, B., and Moens, P. B. (1994) J. Cell Sci. 107, 2749 –2760 35. Lammers, J. H. M., Offenberg, H. H., van Aalderen, M., Vink, A. C. G., Deitrich, A. J. J., and Heyting, C. (1994) Mol. Cell. Biol. 14, 1137–1146 36. Tarsounas, M., Pearlman, R. E., Gasser, P. J., Park, M. S., and Moens, P. B. (1997) Mol. Biol. Cell 8, 1405–1414 37. Lammers, J. H. M., van Aalderen, M., Peters, A. H. F. M., van Pelt, A. A. M., Gaemers, I. C., de Rooij, D. G., de Boer, P., Offenberg, H. H., Dietrich, A. J. J., and Heyting, C. (1995) Chromosoma (Berl) 104, 154 –163 38. Sikorski, R. S., and Hieter, P. (1989) Genetics 122, 19 –27
Vol. 274, No. 3, Issue of January 15, pp. 1783–1790, 1999 Printed in U.S.A.
Red1p, a MEK1-dependent Phosphoprotein That Physically Interacts with Hop1p during Meiosis in Yeast* (Received for publication, July 9, 1998, and in revised form, October 20, 1998)
Teresa de los Santos and Nancy M. Hollingsworth‡ From the Department of Biochemistry and Cell Biology, Institute for Cell and Developmental Biology, SUNY Stony Brook, Stony Brook, New York 11794-5215
The synaptonemal complex (SC) is a proteinaceous structure formed between pairs of homologous chromosomes during prophase I of meiosis. The proper assembly of axial elements (AEs), lateral components of the SC, during meiosis in the yeast, Saccharomyces cerevisiae, is essential for wild-type levels of recombination and for the accurate segregation of chromosomes at the first meiotic division. Genetic experiments have indicated that the stoichiometry between two meiosis-specific components of AEs in S. cerevisiae, HOP1 and RED1, is critical for proper assembly and function of the SC. A third meiosis-specific gene, MEK1, which encodes a putative serine/threonine protein kinase, is also important for proper AE function, suggesting that AE formation is regulated by phosphorylation. In this paper, we demonstrate that Mek1p is a functional kinase in vitro and that catalytic activity is an essential part of the meiotic function of Mek1 in vivo. Immunoblot analysis revealed that Red1p is a MEK1-dependent phosphoprotein. Co-immunoprecipitation experiments demonstrated that the interaction between Hop1p and Red1p is enhanced by the presence of MEK1. Thus, MEK1-dependent phosphorylation of Red1p facilitates the formation of Hop1p/Red1p hetero-oligomers, thereby enabling the formation of functional AEs.
Sexual reproduction requires the formation of haploid gametes so that fertilization can reconstitute the diploid chromosome number of the organism. The chromosome number of the cell is divided precisely in half by meiosis, a highly conserved and specialized type of cell division. Specifically, the reduction in chromosome number is accomplished by the segregation of homologous chromosomes to opposite poles at the first meiotic division. Proper disjunction requires crossovers (typically manifested cytologically as chiasmata) that provide the physical connections between homologs necessary for proper orientation on the Meiosis I spindle (1). Mutants that interfere with the formation of crossovers produce chromosomally imbalanced gametes that are either inviable themselves or fuse to create inviable zygotes (2). In most organisms, two requirements must be met for wildtype levels of crossovers to occur during meiosis. First, there must be recombination machinery to physically exchange the DNA from non-sister chromatids. Second, a tripartite protein-
* This work was supported by a grant from the Pew Charitable Trusts, as well as National Institutes of Health Grant GM50717. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ To whom correspondence should be addressed. Tel.: 516-632-8581; Fax: 516-632-8575; E-mail: [email protected]. This paper is available on line at http://www.jbc.org
aceous structure called the synaptonemal complex (SC)1 must form between homologous chromosomes (3, 4). SC formation is initiated by the condensation of sister chromatids along a protein core called an axial element (AE). Synapsis occurs when AEs are held together by the introduction of central region components to make mature SCs. In Saccharomyces cerevisiae, deletion of RED1, a gene encoding a component of AEs, eliminates AE formation and therefore red1 mutant phenotypes may indicate functions for AEs (5). In red1 diploids, spore viability is drastically reduced, despite only a 4-fold reduction in crossing over, suggesting a role for AEs in packaging crossovers in such a way that the chiasmata can be used for directing Meiosis I chromosome segregation (5, 6). RED1 also plays a role in a checkpoint that monitors whether recombination is progressing correctly or not. For example, mutations in the meiosis-specific recA homolog DMC1 cause arrest prior to the first meiotic division, presumably as a result of the formation of aberrant recombination intermediates (7). In the absence of RED1, this arrest is abolished despite the fact that the aberrant recombination intermediates are still being generated. It appears, therefore, that the packaging of recombination intermediates into some sort of RED1-dependent chromosomal structure is necessary for the recombinationmonitoring checkpoint to work (8). Understanding AE assembly is clearly a key to understanding axial element function. Genetic and cytological experiments have identified two other meiosis-specific genes, HOP1 and MEK1 (also known as MRE4), that are important for formation of functional AEs in yeast. Evidence that the stoichiometry of Red1p to Hop1p is important in axial element formation comes from genetic experiments in which overexpression of RED1 in the presence of limiting amounts of Hop1p created a dominant negative phenotype (9). This dominant negative phenotype can also be produced by overexpressing RED1 in the presence of wild-type levels of HOP1 but with limiting amounts of functional Mek1p (10). Mek1p therefore appears to influence the Hop1p:Red1p stoichiometry. This paper presents biochemical data supporting the model that the interaction of Hop1p and Red1p with each other is promoted by phosphorylation of Red1p by Mek1p (10). EXPERIMENTAL PROCEDURES
Yeast Strains and Media—The genotypes of the strains used in this study are shown in Table I. All the strains are isogenic derivatives of the SK1-related diploid NH144 (11). Diploid strains mutated for RED1 or MEK1 were created by first transforming S2683 and RKY1145, the two haploid parents of NH144, with the appropriate DNA fragments (12) followed by mating. For the red1::LEU2 mutation, a 4.9-kb BglII/ PstI fragment was purified from pNH119 (13) and used for transforma-
1 The abbreviations used are: SC, synaptonemal complex; AE, axial element; HA, hemagglutinin; GST, glutathione S-transferase; PCR, polymerase chain reaction; kb, kilobase(s); PMSF, phenylmethylsulfonyl fluoride.
1783
1784
Red1p, a MEK1-dependent Phosphoprotein TABLE II Plasmid list
TABLE I Saccharomyces cerevisiae SK1 strains Name
RKY1145 S2683 NH144 YTS1 YTS1::pLP36 YTS1::pLP37 YTS3 YTS3::pTS23 YTS4 NH177 NH211 DW10
Genotype
Name
MATa leu2D::hisG his4-x lys2 ura3 hoD::LYS2 MATa leu2-K lys2 ura3 hoD::LYS2 arg4-Nsp MATa leu2D::hisG his4-x lys2 ura3 hoD::LYS2 ARG4 MATa leu2-K HIS4 lys2 ura3 hoD::LYS2 arg4-Nsp same as NH144 except mek1D::LEU2 mek1D::LEU2 same as YTS1 except ura3::mek1-K199R::URA3 same as YTS1 except ura3::MEK1::URA3 same as NH144 except red1::LEU2 red1::LEU2 same as YTS3 except ura3::RED1–3HA::URA3 ura3::RED1–3HA::URA3 same as NH144 except mek1D::URA3 mek1D::URA3 same as NH144 except red1::LEU2 mek1D::LEU2 red1::LEU2 mek1D::LEU2 same as NH144 except mek1–974 ade2-Bgl mek1–974 ade2-Bgl same as NH144 except hop1::LEU2 RED1–3HA hop1::LEU2 RED1–3HA
pNH83 p1b-1 pNH218 pNH219 pNH124 pB131 pTS21 pTS1 pRD56 pNH168 pVZ1 pNH205 pNH206 pDW14 pDW15 pDW16 pR4C4 pLP31 pLP34 pLP35 pRS306 pLP36 pLP37 pLP41 pNH235 pNH208 SK-HA3 pNH209 pNH212 pTS23
tion. This insertion allele exhibits a null phenotype with respect to spore viability (other phenotypes have not been tested). A 3.6-kb XbaI/ BamHI fragment from pTS1 carries the mek1D::LEU2 allele, and a 2.9-kb BamHI fragment from pTS21 was used to introduce mek1D::URA3. These deletions remove 75 base pairs upstream of the ATG as well as approximately 75% of the MEK1 coding sequence. The presence of the red1::LEU2, mek1D::LEU2, and mek1D::URA3 alleles was confirmed both genetically and by Southern blot analysis. The ade2-Bgl mutation was introduced by two-step gene replacement (12) using pR943 (generously supplied by B. Rockmill, Yale University) cut with SpeI. Haploids containing the ade2-Bgl mutation and appropriate other markers were obtained by crosses. These haploids were then mated to give homozygous diploids. The mek1–974 mutation was introduced into two different mek1D::LEU2 haploids by two-step gene replacement using pNH235 digested with BssHI to target integration to the mek1D::LEU2 locus. Leu2 5FoaR colonies were then mated to create the mek1–974 diploid NH211. The red1::LEU2 mek1D::LEU2 homozygous diploid NH177 was constructed by crossing a MATa mek1D::LEU2 haploid with MATa red1::LEU2 haploid and dissecting tetrads from the resulting heterozygous diploid. Leu1 spore colonies that exhibited a non-parental di-type pattern for LEU2 were then mated to produce the double homozygote. To make the RED1–3HA homozygous diploid YTS3::pTS23, S2683red1 and RKY1145red1 were transformed with pTS23 digested with StuI to target integration of the plasmid to URA3. The resulting transformants were mated to create the diploid. For YTS1:pLP36 and YTS1::pLP37, the plasmids pLP36 and pLP37 were digested with StuI to target integration to URA3 and transformed into YTS1. Standard yeast genetic methods were used (14). Solid media have been described (15). YPA contains 1% yeast extract, 2% bactopeptone, and 2% potassium acetate. Sporulation was performed using 2% potassium acetate. Plasmids—Plasmids for this study were made by standard procedures (16) using the Escherichia coli strain BSJ72 and are listed in Table II. To construct the mek1D::URA3 and mek1D::LEU2 deletion alleles, the polymerase chain reaction (PCR) was used to generate fragments of URA3 or LEU2 with SacI and PvuII ends. The PCR fragments were digested with SacI and PvuII and ligated to pB131 (17) cut with SacI and HpaI to generate pTS21 (mek1D::URA3) and pTS1 (mek1D::LEU2), respectively. To fuse the MEK1 gene to GST (glutathione S-transferase) (18), PCR was employed to generate a fragment containing the entire MEK1 gene. The resulting fragment was digested with BglII and ligated into the BamHI site of pRD56 to create pNH168. The mek1-K199R mutation was introduced into the GST-MEK1 fusion gene by oligonucleotide site-directed mutagenesis (19). The 2.9-kb NotI/SalI fragment from pNH168 containing GAL1p-GST-MEK1 was cloned into NotI/SalI-digested pVZ1 to generate pNH205. The presence of the mutation creates an FspI restriction site that was used for screening plasmids with the desired mutation. The plasmid containing the mek1-K199R allele was designated pNH206. To place the GST-MEK1 and GST-mek1-K199R alleles under control of the MEK1 promoter, the MEK1 promoter was first cloned into the 2m vector, YEplac181. A 924-bp fragment starting at position 2896 rela-
a b
Relevant yeast genotype
HOP1 URA3 2m RED1 URA3 2m RED1 URA3 2m RED1 HOP1 URA3 2m RED1 MEK1 URA3 CEN4 ARS1 mek1D::URA3 URA3 CEN4 ARS1 mek1D::LEU2 URA3 CEN4 ARS1 GAL1p-GST URA3 CEN6 ARSH4 GAL1p-GST-MEK1 URA3 CEN6 ARSH4 pBS1 with extended polylinker GAL1p-GST-MEK1 GAL1p-GST-mek1-K199R MEK1p LEU2 2m MEKp1-GST-MEK1 LEU2 2m MEK1p-GST-mek1-K199R LEU2 2m MEK1 URA3 ARS1 CEN4 MEK1 mek1-K199R MEK1 URA3 mek1-K199R URA3 MEK1 URA3 mek1–974 URA3 CEN4 ARS1 mek1–974 URA3 RED1 (EcoRI site in place of stop codon) 3 HA epitopes in SK2 RED1–3HA RED1–3HA URA3 2m RED1–3HA URA3
Source
(13) (13) (10) (10) (10) (17) This work This work A. Neimana This work S. Henikoffb This work This work This work This work This work (17) This work This work This work (38) This work This work This work This work This work (21) This work This work This work
SUNY Stony Brook. Fred Hutchinson Cancer Research Center.
tive to the MEK1 ATG and ending at 128 was amplified by PCR. The oligonucleotide was engineered to put an NdeI site at the ATG of MEK1 with a SalI site downstream. The fragment was digested with BamHI and SalI and ligated into BamHI/SalI-digested YEplac181 to make pDW14. The 2.9-kb NdeI/SalI fragments containing either GST-MEK1 (pNH168) or GST-mek1-K199R (pNH206) were then cloned into pDW14 cut with NdeI and SalI to make pDW15 and pDW16, respectively. The MEK1 and mek1-K199R alleles were completely sequenced to confirm that no unknown mutations were introduced by either the PCR or site-directed mutagenesis. Sequencing was performed using the ABI Amplitaq sequencing kit (Perkin-Elmer) followed by electrophoresis on an ABI model 373A automated sequencer. To assay for complementation of the mek1D by mek1-K199R, both the MEK1 and mek1-K199R alleles were put under control of the MEK1 promoter without the GST fusion. To accomplish this, a 5.5-kb EcoRI fragment from pR4C4 (10) was first cloned into EcoRI-digested pVZ1 to generate pLP31. After digestion of pLP31 with SpeI and SalI, the vector backbone was ligated with the 1.2-kb SpeI/SalI fragment from either pNH206 or pNH168, resulting in the substitution of approximately 80% of the MEK1 coding sequence. The resulting plasmids, pLP34 (mek1K199R) and pLP35 (MEK1) were then digested with NotI and SalI, and the 2.6-kb fragments were ligated into pRS306 to make pLP36 and pLP37, respectively. The mek1–974 mutation was rescued onto a plasmid by gap repair (20). The MEK1 plasmid, pB131, was cut with SacI and HpaI and transformed into the mek1–974 diploid NH104 (10). The resulting plasmid, pLP41, was used in sequencing reactions with primers located at staggered intervals along the MEK1 gene. To make the mek1–974 integrating plasmid, pNH235, a 3.2-kb EcoRI/SalI fragment from pLP41 was ligated into the URA3 vector, pRS306, digested with EcoRI and SalI. The first step in the construction of the RED1–3HA allele was the introduction of an EcoRI site in place of the RED1 stop codon by PCR. A 1.3-kb fragment containing the 39-half of RED1 was amplified, digested with BglII and EcoRI, and ligated into BglII/EcoRI-digested pNH124 (10) creating pNH208. A 3.4-kb SalI/EcoRI fragment from pNH208 carrying the entire RED1 gene with ;1.1 kb of upstream sequences was cloned into SalI/EcoRI SK-HA3 (21). This ligation fuses the full-length RED1 gene in-frame at the 39-end with three copies of the hemagglutinin (HA) epitope in pNH209. The tagged allele was moved either to a high copy 2m plasmid (pNH212) or to a URA3integrating plasmid (pTS23) by ligating the 3.5-kb SalI/SacI fragment
Red1p, a MEK1-dependent Phosphoprotein from pNH209 into YEp352 and pRS306, respectively. Antibodies—The mouse monoclonal a-HA antibody is contained in the 12CA5 ascites fluid that was purchased from Babco. Affinity-purified a-Gstp antibodies were generously provided by D. Kellogg (University of California Santa Cruz). Antibodies directed against the fulllength native Hop1 protein were generated as follows. The Hop1 protein was overexpressed in vegetative yeast cells and purified to .95% homogeneity (22). 0.5 mg of purified Hop1 protein was injected into two rabbits followed by a boost of 0.25 mg three weeks later (performed by Babco). The resulting sera were found to cross-react with a band of ;67 kDa, the predicted molecular mass of Hop1p, when cells carrying the galactose-inducible GAL10p-HOP1 gene (22) were grown in galactose. This band was absent when the cells were grown on glucose. The 67-kDa band was also not observed using the preimmune sera from these rabbits. Further evidence that these antibodies are specific for Hop1p is that the 67-kDa protein is only observed in sporulating cultures when HOP1 is present.2 Sporulation—2-ml Yeast extract-peptone-dextrose cultures were inoculated with single colonies of the appropriate diploid and grown to saturation at 30 oC with aeration. Strains containing plasmids were inoculated with colonies grown on selective medium. Plasmid stability was monitored immediately prior to transfer to sporulation medium and .80% of the cells were typically found to contain plasmid. The yeast extract-peptone-dextrose cultures were diluted 1:1000 –1:3000 into yeast peptone acetate medium and grown for approximately 16 h with aeration at 30 oC. When cultures reached an OD660 5 1–2, the cells were pelleted, washed with water, and resuspended in 2% potassium acetate to a final concentration of 3 3 107 cells/ml. The cells were shaken at 30 oC for 3 h and then harvested by centrifugation for 10 min at 5000 rpm in a GSA rotor. Prior to centrifugation, a 1-ml aliquot was removed to the 30 oC roller, and the percent sporulation was assessed by light microscopy the following day. Sporulation typically reached levels of .90%. Kinase Assays—After 3 h in sporulation medium, 20-ml aliquots of cells were washed with water, and the cell pellets were quick frozen at 270 oC. Kinase assays were performed as described (23). Immunoprecipitation of Hop1p and Red1–3HAp from Meiotic Extracts—Sporulating cells were harvested, washed one time with water, and resuspended in 0.3 ml of lysis buffer (25 mM Tris-HCl, pH 7.5, 200 mM NaCl, 5 mM EDTA) per 20 ml of cells. To reproducibly detect Red1–3HAp, it was necessary to prepare extracts the same day as the cells were sporulated, i.e. the cells could not be frozen first. The reason for this phenomenon is not understood but is probably not because of degradation of the Red1–3HA protein. After extracts were prepared, the proteins could be frozen at 270 oC and thawed with no loss in the ability to detect Red1–3HAp by immunoblot analysis. Protease inhibitors (1 mM PMSF, 1 mg/ml aprotinin, 1 mg/ml leupeptin, 1 mg/ml pepstatin) and phosphatase inhibitors (10 mM NaF, 10 mM Na4P2O7, 1 mM Na3VO4) were present throughout the lysis and wash steps. One gram of glass beads was added, and the tubes were vortexed three times for 30 s with 2-min intervals on ice. The lysate was removed to a clean tube, and detergents were added to final concentrations of 1% Triton X-100, 0.5% sodium deoxycholate, and 0.1% SDS. The lysate was rocked for 15 min at 4 oC and cleared by centrifugation at 14,000 3 g for 10 min. The supernatant was divided into two tubes to which either 1–2 ml of a-Hop1p or a-HA antibodies was added, followed by rocking for 2 h at 4 oC. To precipitate the immune complexes, 40 ml of protein A-Sepharose (Amersham Pharmacia Biotech) slurry (equilibrated 1:1 with lysis buffer) were added. The tubes were then rocked again for an additional hour at 4 oC. The Sepharose beads were washed four times with 0.5 ml of lysis buffer and resuspended in 30 ml of 23 SDS protein sample buffer. The beads were heated at 95 oC for 5 min and loaded in duplicate onto a 6% SDS-polyacrylamide gel. After fractionation, the proteins were blotted to nitrocellulose and probed for Hop1p and Red1–3HAp using a-Hop1p and a-HA antibodies, respectively. The presence of the antibodies was determined using the ECL kit from Amersham Pharmacia Biotech. The amount of Hop1p and Red1–3HAp present in the immunoprecipitates was quantitated as described in the instructions for the ECL kit (Amersham Pharmacia Biotech). A series of dilutions from the soluble extracts used in the experiment was fractionated on the same gel and probed simultaneously with the appropriate antibodies. After the ECL reaction, the films were scanned using a Bio-Rad imaging densitometer. A standard curve of protein concentration was generated for Hop1p and Red1–3HAp using the Molecular Analyst, Version 1.1
2
N. Hollingsworth, unpublished observations.
1785
software (Bio-Rad). The amount of Hop1p and Red1–3HAp present in the immunocomplexes was then determined using the standard curves. To compare the efficiency of the co-immunoprecipitation in MEK1 and mek1D extracts, the following formula was used. ~Co-IP proteinMEK1/IP proteinMEK1)/ (Co-IP proteinmek1D/IP proteinmek1D)
(Eq. 1)
Phosphatase Experiments—For each sample, 20 ml of sporulating YTS3/pNH212 cells were used for immunoprecipitation of Red1–3HAp. The immunoprecipitates were treated with l protein phosphatase (New England Biolabs) as described (24). RESULTS
Overexpression of HOP1 Suppresses mek1–974 in the SK1 Background whereas Overexpression of RED1 Exacerbates the mek1–974 Mutant Phenotype—Genetic interactions between MEK1, HOP1, and RED1 in the A364A strain background led to the proposal that phosphorylation of either Hop1p or Red1p by Mek1p facilitates the formation of Hop1p/Red1p heterooligomers, thereby ensuring the proper balance between the proteins in the assembly of axial elements during meiosis (10). To test this model biochemically, it was desirable to use the SK1 strain background in which sporulation proceeds rapidly, relatively synchronously, and with high efficiency (25). To ensure that the genetic dosage effects of HOP1 and RED1 overexpression in the presence of the leaky mek1–974 mutant occur in SK1 as well as A364A, it was necessary to repeat the dosage experiments in SK1. The mek1–974 mutation was cloned by gap repair using the diploid NH104 (10), and the entire gene was sequenced. A single transition mutation of Gly to Ala at nucleotide 502 was found. This mutation creates a valine to methionine substitution at amino acid 168, which is located immediately before the GXG motif in the first conserved kinase domain (26). The mek1–974 allele was introduced into the SK1 background by two-step gene replacement, thereby creating the homozygous diploid, NH211. This diploid was transformed with the same plasmids used in the previous analysis (10), the transformants were sporulated, and tetrads were dissected to assess spore viability. As in the A364A background, the mek1–974 mutant is partially functional in SK1, producing 33.2% viable spores compared with ;1% for the deletion (Ref. 27 and Table III). The spore viability defect is complemented by MEK1 on a CEN plasmid (Table III, NH211). The presence of excess HOP1 greatly improved the mutant phenotype, resulting in 74% viable spores. In contrast, overexpression of RED1 decreased the number of viable spores from 33.2 to 16.8% (Table III, NH211). Both the HOP1 suppression and the RED1 exacerbation of the mek1–974 spore inviability phenotype are statistically significant by x2 analysis (p , 0.001). The one difference observed between SK1 and A364A was the result obtained when both HOP1 and RED1 were co-overexpressed on the same plasmid. In A364A, HOP1/RED1 co-overexpression restored the spore viability observed for vector alone. In SK1, a suppression phenotype similar to that for HOP1 alone was found (Table III, NH211). Detection of Red1p in Meiotic Cells Using the Hemagglutinin Epitope Tag—Because attempts to use polyclonal antibodies directed against Red1p (28) to detect the Red1p protein by immunoblot analysis were unsuccessful,3 an epitope-tagged version of the protein was created. The RED1 gene was fused in frame at the 39-end to three copies of the hemagglutinin epitope. The fusion gene was then integrated at URA3 in two different isogenic SK1 red1::LEU2 haploids. This insertion allele of RED1 exhibits a null phenotype with regard to spore 3
T. de los Santos, unpublished observations.
1786
Red1p, a MEK1-dependent Phosphoprotein TABLE III Spore viability assayed by tetrad dissection
Strain/plasmid
NH211 NH211/pR4C4 NH211/pNH83 NH211/pNH218 NH211/pNH219 YTS3/YEp352 YTS3::pTS23 YTS3/pNH212 YTS3/p1b-1 YTS4 YTS4/pDW15 YTS4/pDW14 YTS1 YTS1::pLP37 YTS1::pLP36
Relevant yeast genotype
% Spore viability
No. asci
mek-974 mek-974 mek1–974/MEK1 CEN mek1–974 mek1–974/HOP1 2m mek1–974 mek1–974/RED1 2m mek1–974 mek1–974/RED1 HOP1 2m mek1–974 red1::LEU2 red1::LEU2 red1::LEU2 RED1–3HA red1::LEU2 RED1–3HA red1::LEU2/RED1–3HA 2m red1::LEU2 red1::LEU2/RED1 2m red1::LEU2 mek1D::URA3 mek1D::URA3 mek1D::URA3/GST-MEK1 2m mek1D::URA3 mek1D::URA3/GST-mek1-K199R mek1D::URA3 mek1D::LEU2 mek1D::LEU2 mek1D::LEU2 MEK1 mek1D::LEU2 mek1D::LEU2 mek1-K199R mek1D::LEU2
33.2
104
92.3
52
74.0
104
16.8
104
69.0
105
3.8
53
,0.5
57
50.0
72
83.5
50
0.3
71
76.8
42
2.1
71
5.0
55
94.9
74
1.8
56
viability (13). The haploids were mated to create a red1::LEU2 diploid that is homozygous for the RED1–3HA allele (YTS3::pTS23). Alternatively, the RED1–3HA gene on a high copy number 2m vector was transformed into YTS3 (YTS3/ pNH212). After 3 h in sporulation medium, a time point when Red1–3HAp and Hop1p concentrations are maximal,3 protein extracts were made from YTS3/YEp352, YTS3::pTS23, and YTS3/pNH212. No protein of the appropriate molecular weight for Red1–3HAp was detected with the a-HA antibodies in YTS3 containing the vector alone (Fig. 1A). These cells were proceeding through meiosis, as evidenced by there being a wild-type amount of Hop1p in the YTS3/YEp352 extract (Fig. 1B). A RED1–3HA-dependent a-HA cross-reacting band was observed in YTS3::pTS23. This protein represents Red1–3HAp as this band is absent from extracts of NH144, an isogenic diploid homozygous for the untagged RED1 gene, as well as from vegetative cultures of YTS3::pTS23.3 Interestingly, in YTS3/ pNH212, where the RED1–3HA allele is present in high copy number, several Red1–3HAp bands are apparent (Fig. 1A). Tetrad dissection was performed to determine the ability of the RED1–3HA allele to complement the spore viability defect of red1::LEU2 in YTS3. YTS3 carrying vector alone produced 3.8% viable spores. When RED1–3HA is homozygous in YTS3::pTS23, no complementation was observed (Table III, YTS3/YEp352). Over-expression of the RED1–3HA allele restored a significant amount of RED1 although the complementation was not as good as overexpression of the untagged allele of RED1 (Table III, YTS3/YEp352). Because the overexpressed RED1–3HA allele was detectable and provided a substantial level of RED1 function (50% spore viability versus 3.8%), this construct was used in all subsequent experiments. Red1–3HAp Is a Phosphoprotein—To determine whether the slower migrating species of Red1–3HAp are because of phosphorylation, the Red1–3HA protein was immunoprecipitated and tested for sensitivity to l protein phosphatase (New England Biolabs). After washing, the immunoprecipitates were resuspended in phosphatase buffer, treated with or without
FIG. 1. Detection of Red1–3HAp and Hop1p in meiotic extracts. The red1::LEU2 diploid, YTS3, was transformed with vector alone (YEp352), pTS23 (two integrated copies of RED1–3HA), or pNH212 (2m RED1–3HA). Cells were sporulated for 3 h, and total protein extracts were made. 20 mg of total extract were fractionated by SDS-PAGE, blotted to nitrocellulose, and probed with either a-HA antibodies (panel A) or a-Hop1p antibodies (panel B).
phosphatase, fractionated by SDS-PAGE, and blotted, and the Red1–3HAp was detected using the a-HA antibody. In the sample where no phosphatase was added, three discrete bands were observed (Fig. 2). When l protein phosphatase was added, the slower mobility bands were no longer visible. The addition of phosphatase inhibitors prior to the phosphatase prevents the loss of the shifted bands (Fig. 2), indicating that it is the phosphatase activity which accounts for the disappearance of the modified bands and not a contaminating protease. One concern was that the phosphorylation is occurring on the HA epitope rather than Red1p. That this is not the case is demonstrated by the findings of Bailis and Roeder (29) who have recently shown that Red1p is a phosphoprotein using antibodies against the endogenous Red1p. Mek1p Has Kinase Activity in Vitro and the Kinase Activity Is Essential for the Meiotic Function of MEK1—A good candidate for the kinase which phosphorylates Red1p is Mek1p. To determine whether the Mek1p kinase homology is meaningful, Mek1p kinase activity was tested in vitro. The MEK1 gene was fused to the 39-end of the affinity tag GST and the GST-MEK1 fusion allele was placed under the control of the MEK1 promoter. As a negative control, an allele of MEK1 (mek1-K199R) was created in which the codon for an invariant lysine residue present in domain II of the kinase (amino acid 199) was mutated to arginine. This lysine to arginine mutation has been shown to reduce or abolish kinase activity in a number of other kinases (30). The mek1D::URA3 diploid YTS4 was transformed with YEplac181 (2m), pDW15 (2m GST-MEK1), or pDW16 (2m GST-mek1-K199R). Cells were sporulated for 3 h, protein ex-
Red1p, a MEK1-dependent Phosphoprotein
1787
FIG. 2. Detection of a MEK1-dependent phosphorylated form of Red1–3HAp in meiotic extracts. Red1–3HAp was immunoprecipitated from 3 mg of soluble protein extracted from YTS3/pNH212 after 3 h in sporulation medium. Immunocomplexes were treated with and without l protein phosphatase and phosphatase inhibitors. Red1–3HAp was detected by immunoblot analysis using a-HA antibodies.
tracts were made, and the Gst-Mek1p and Gst-mek1-K199R fusion proteins were purified using glutathione-Sepharose beads. Kinase activity was assayed by autophosphorylation. The relative amount of Gst-Mek1p and GST-mek1-K199R present during the assay was determined by probing the same blot used for autoradiography with antibodies directed against Gstp. A radioactive band migrating at the predicted molecular mass for Gst-Mek1p was observed; this band is absent in YTS4/ YEplac181 (Fig. 3A). Very little autophosphorylation activity was detectable for the Gst-mek1-K199R protein despite the fact that more protein was present than for Gst-Mek1 (Fig. 3B). The reduction in phosphorylation of the mek1-K199R protein demonstrates that most of the observed kinase activity is because of Mek1p and not to a co-precipitating kinase or to the Gst moiety. Complementation tests were performed to assay the function of the two MEK1 alleles. The GST-MEK1 gene in YTS4 complemented the mek1D spore viability defect while the GSTmek1-K199R allele did not (Table III, YTS4). GST-MEK1 also complemented when integrated into the chromosome, showing that overexpression is not a requirement for activity.2 In addition, MEK1 and mek1-K199R (without GST) were integrated at URA3 in the mek1D::LEU2 diploid YTS1. Full complementation of the spore viability defect was seen in the YTS1::pLP37 diploid carrying MEK1 compared with YTS1 alone (94.5% versus 5.0% viable spores; Table III, YTS4). In contrast, the mek1K199R allele fails to complement (Table III, YTS4). Therefore, mutation of the invariant lysine to arginine at position 199 creates an allele of MEK1 that exhibits a null phenotype with regard to spore viability. Because this mutation also greatly reduces Mek1p kinase activity in vitro, we conclude that the kinase activity of Mek1p is essential for its meiotic function. The Bulk of Red1–3HAp Phosphorylation Is Dependent on MEK1—To test whether Red1–3HAp is a potential substrate for Mek1p, the protein was examined in MEK1 (YTS3/ pNH212)and mek1D (NH177/pNH212) diploids. Extracts were prepared from cells after 3 h in sporulation medium. Because the total amount of Red1–3HAp present in mek1D diploids is reduced,3 nearly three times as much total protein from the mek1D was loaded onto the gel compared with MEK1. Although there is a similar amount of the unphosphorylated form of Red1–3HAp in both strains, the vast majority of phosphorylated Red1–3HAp is absent in the mek1D diploid (Fig. 4). Red1– 3HAp phosphorylation is therefore dependent on MEK1. This experiment does not, however, address whether this dependence is because of direct phosphorylation of Red1–3HAp by Mek1p or whether an intermediate kinase is involved. When the soluble fraction is examined, a low level of phosphorylated Red1–3HAp is sometimes observed. Because the spore viability of the mek1D diploid overexpressing RED-3HA is very low (1.3%, 77 asci), any phosphorylated Red1–3HAp that is present is apparently non-functional. Whether this “promiscuous” phosphorylation of Red1–3HAp is because of the overexpres-
FIG. 3. Autophosphorylation of Gst-Mek1p from meiotic extracts. Cells were sporulated for 3 h and extracts from YTS4 (mek1D/ mek1D) containing vector alone (YEplac181), GST-MEK1 (pDW15), or GST-mek1-K199R (pDW16) were treated with glutathione-Sepharose to affinity purify the GST fusion proteins. The purified proteins were treated with [g-32P]ATP:ATP to assay for autophosphorylation. The proteins were examined by phosphoimager analysis to detect incorporated radioactivity (panel A). The same blot was subsequently probed with a-Gstp antibodies (panel B).
FIG. 4. Detection of Red1–3HAp in MEK1 and mek1D diploids. Total cell extracts from YTS3/pNH212 (MEK1/MEK1) and NH177/ pNH212 (mek1D/mek1D) were examined for Red1–3HAp using immunoblot analysis and a-HA antibodies. For YTS3/pNH212, 35 mg were loaded on the gel; for NH177/pNH212, 90 mg were used.
sion of the protein or whether it normally occurs is not known. Hop1 and Red1–3HAp Co-immunoprecipitate from Meiotic Cell Extracts—Co-immunoprecipitation experiments were performed to determine whether Hop1p and Red1–3HAp are physically associated in meiotic cells. YTS3/pNH212 was sporulated for 3 h, protein extracts were made, and equal amounts of protein were incubated with either a-HA antibodies or a-Hop1p antibodies to immunoprecipitate Red1–3HAp and Hop1p, respectively. The specificity for Hop1p immunoprecipitation was shown by the fact that no Hop1 is detected in immunoprecipitates from sporulating cells deleted for HOP1 (Fig. 5, compare lanes 5 and 6) nor is Hop1 precipitated if the a-Hop1p antibody is omitted (Fig. 5, lane 4). Similarly, no Red1–3HAp is observed when the a-HA antibody is used for immunoprecipitation in diploids carrying the untagged RED1 gene (Fig. 5, compare lanes 2 and 3) or if the a-HA antibody is left out when precipitating from diploids containing the RED1–3HA gene (Fig. 5, lane 4). The Hop1p and Red1p immunoprecipitated samples were
1788
Red1p, a MEK1-dependent Phosphoprotein
FIG. 5. Specificity of Hop1 and Red1–3HAp immunoprecipitation. Cells from various diploids were sporulated for 3 h and soluble extracts prepared. Hop1p and Red1–3HAp were immunoprecipitated from these extracts using either a-Hop1p or a-HA antibodies. Lanes 1, 2, 4, and 5, YTS3/pNH212 (HOP1/HOP1 red1::LEU2/red1::LEU2/2 m RED1–3HA); lane 3, YTS3/p1b-1 (HOP1/HOP1 red1::LEU2/ red1::LEU2/2 m RED1); lane 6, DW10::pTS23 (hop1::LEU2/hop1 ::LEU2 RED1–3HA/RED1–3HA; and lane 7, YTS3::pTS23 (HOP1/ HOP1 red1::LEU2/red1::LEU2 RED1–3HA/RED1–3HA. Lanes 1–3 were probed with a-HA antibodies to detect Red1–3HAp; and lanes 4 –7 were probed with a-Hop1p antibodies.
divided in half, fractionated by SDS-PAGE, transferred to nitrocellulose and probed with either a-HA antibodies or a-Hop1p antibodies to look for co-immunoprecipitation of Red1–3HAp with Hop1p and vice versa. In YTS3/pNH212, Hop1p co-immunoprecipitated with Red1–3HAp (Fig. 6A, lane 3). The converse is also true; when Hop1p was first precipitated with a-Hop1p antibodies, Red1–3HAp was detected in the complex (Fig. 6B, lane 1). The Absence of MEK1 Reduces the Efficiency of the Hop1p/ Red1–3HAp Co-immunoprecipitation—To test whether the MEK1-dependent phosphorylation of Red1–3HAp has any effect on the ability of Red1–3HAp to interact with Hop1p, the co-immunoprecipitation experiment was performed using extracts from the mek1D strain, NH177/pNH212. Because there is less soluble Hop1p and Red1–3HAp in the absence of MEK1, twice as much protein was used for the immunoprecipitation compared with the MEK1 diploid. When the a-Hop1p antibodies were incubated with the protein extracts, nearly equivalent amounts of Hop1p were precipitated from both strains (Fig. 6B, lanes 3 and 4). However, when these precipitates were probed with the a-HA antibodies, a 4.3-fold reduction in the amount of Red1–3HAp co-precipitating was observed with the mek1D extract (Fig. 6B, lane 2). In three experiments where the amount of Red1–3HAp co-immunoprecipitating with Hop1 was quantitated, a 4.3-, 4.1-, and 3.6-fold reduction was observed in the mek1D diploid compared with MEK1. The reduction in the Red1–3HAp co-IP was not because of a lack of Red1–3HAp in the mek1D diploid as evidenced by the amount of Red1–3HAp present in the a-HA immunocomplexes from the mek1D diploid compared with MEK1 (Fig. 6A, lanes 1 and 2). The amount of Hop1p co-immunoprecipitating with Red1–3HAp varied from a reduction of 1.8- (the value for the experiment shown in Fig. 6, lanes 3 and 4), 1.5-, to 0-fold change in the mek1D diploid compared with MEK1. To rule out a loading artifact, blots probed with a-Hop1p antibodies were stripped and probed with a-HA antibodies and vice versa, and the same results were observed.3 The presence of MEK1 therefore facilitates the Hop1p/Red1–3HAp interaction in vivo. DISCUSSION
Genetic studies in the A364A background led to a model which says that Mek1p phosphorylation facilitates the assem-
FIG. 6. Hop1p and Red1–3HAp co-immunoprecipitation from meiotic extracts. Cells from YTS3/pNH212 (MEK1/MEK1) and NH177/pNH212 (mek1D/mek1D) were sporulated for 3 h. Hop1p and Red1–3HAp were immunoprecipitated from soluble extracts prepared from these cells (2.8 mg for YTS3/pNH212 and 6.6 mg for NH177/ pNH212) using a-HA antibodies (panel A) or a-Hop1p antibodies (panel B). Immunoprecipitates were then probed with either a-HA (to detect Red1–3HA) or a-Hop1p antibodies as noted beneath the pictures. The cross-reacting band at the bottom of panel B, lanes 3 and 4 is because of detection of the antibody used in the immunoprecipitation.
bly of Red1p and Hop1 in a particular stoichiometry necessary to generate functional AEs during SC formation (10). To test this model biochemically, it was desirable to use the SK1 strain background, because the cells sporulate efficiently and synchronously. However because mutants in MEK1 and HOP1 exhibit more severe phenotypes in the SK1 background compared with A364A, it was possible that the model based on the findings from the A364A strain background would not apply to the SK1 background. For mek1D diploids, spore viability is reduced to # 1% in SK1 (27) compared with 15% in A364A (10). For hop1 mutants, spore viability is ,1% in both strain backgrounds, but no residual recombination is observed in SK1 (6, 31, 32). To interpret the biochemical results, it was therefore important to determine whether the same dosage relationships between mek1–974 and overexpression of HOP1 and RED1 observed in A364A exist in SK1. The mek1–974 allele used for the dosage experiments in A364A was cloned by gap repair and introduced into SK1. The mek1–974 allele is partially functional in SK1, the same as in A364A. Overexpression of HOP1 in the SK1 mek1–974 diploid resulted in partial suppression of the spore viability defect. In contrast, overexpression of RED1 decreased spore viability. Co-overexpression of HOP1 and RED1 was previously shown to result in a phenotype similar to neither gene being present, consistent with the idea that the two proteins can titrate out each other. In SK1, the HOP1/RED1 co-overexpression phenotype more resembled that of HOP1 alone, suggesting that this strain background may be more sensitive to the levels of HOP1, a result consistent with the more severe recombination phenotype exhibited by hop1 mutants in this strain. The reproducibility of the dosage results in SK1 indicated that it would be an
Red1p, a MEK1-dependent Phosphoprotein appropriate strain in which to test the stoichiometry model biochemically. A basic assumption of the stoichiometry model is that Mek1p is a serine/threonine protein kinase. This assumption was proven to be correct by demonstrating that Mek1p has autophosphorylation activity in vitro. This activity is greatly reduced by substitution of a conserved lysine in the catalytic domain for arginine—a mutation that abolishes function in vivo. Therefore, the kinase activity of Mek1p is an essential component of its meiotic function. To determine whether either Hop1 or Red1 is a MEK1-dependent phosphoprotein, the two proteins were analyzed by immunoblot analysis for mobility shifts that may be indicative of phosphorylation. Extensive attempts to detect a phosphorylated form of Hop1p were unsuccessful. In contrast, a modified form of Red1–3HAp that is sensitive to phosphatase treatment was observed. The reduced amounts of this modified form in mek1D diploids suggest that Red1–3HAp may be the substrate for Mek1p although the presence of an intermediate kinase cannot be ruled out. In MEK1 strains, the Red1–3HA phosphoprotein was observed only when the RED1–3HA allele was overexpressed, suggesting that the presence of the HA tag may partially inhibit the action of the kinase. The Red1–3HA phosphoprotein is likely to represent the functional form of the protein, as it is only when this species is present that RED1 activity is restored in vivo. Genetic evidence exists for Hop1p and Red1p homo-oligomers as well as for Hop1p/Red1p hetero-oligomers (10, 13, 31). The Hop1p homo-oligomer has recently been demonstrated biochemically using purified protein (22). To test whether Hop1p and Red1p interact in meiotic cells, co-immunoprecipitation experiments were performed. The Hop1p antibodies coprecipitated both the phosphorylated and unphosphorylated forms of Red1–3HAp. The presence of unphosphorylated Red1– 3HAp in the Hop1p immune complex is not surprising given that Mek1p phosphorylation is not essential for the Hop1p/ Red1p interaction because stretches of SC are observed even in mek1D diploids (17). In addition, although the stoichiometry between Hop1p and Red1p is as yet undetermined, there is usually more Red1–3HAp co-immunoprecipitating with Hop1p than Hop1p that co-precipitates with Red1–3HA (e.g. there is an 8-fold difference in Fig. 6). This is the result expected if the ratio of Red1p:Hop1p is greater than 1:1. If this is the case then when a molecule of Hop1p is precipitated, it may bring down not only the phosphorylated Red1–3HAp with which it was associated but also unphosphorylated Red1–3HAp which was interacting with the Red1–3HA phosphoprotein. It is important to note that these experiments were all performed with an allele of RED1 that is being overexpressed. Because Smith and Roeder (28) have shown that overexpression of RED1 can result in a more continuous staining pattern of Red1p along the chromosome, it may be that the stoichiometries observed here do not reflect the exact stoichiometries present in strains containing endogenous levels of Red1p. Although Mek1p phosphorylation of Red1–3HAp is not essential for the Hop1p/Red1p interaction, we have proposed that it facilitates the formation of the hetero-oligomer in the proper stoichiometry (10). This idea was tested by performing the Hop1p/Red1–3HAp co-immunoprecipitation experiments in a mek1D diploid. While it is possible to immunoprecipitate equivalent amounts of Hop1p from meiotic extracts made from MEK1 and mek1D diploids, the level of Red1–3HAp co-precipitating was reduced approximately 4-fold in the absence of MEK1. When Red1–3HAp was immunoprecipitated from mek1D and MEK1 extracts, approximately 2-fold less Hop1p was observed in the complex. The presence of MEK1 clearly
1789
TABLE IV Comparison of S. cerevisiae Hop1p and Mek1p with different potential homologs S. c., Saccharomyces cerevisiae; D. m., Drosophila melanogaster; C. a., Candida albicans; A. t., Arabidopsis thaliana; aa, amino acids; I, identity; S, similarity.
S. c. Mek1 3 D. m. loki S. c. Mek1 3 S. p. Mek1 S. c. Mek1 3 C. a. Mek1a
Kinase domain (302 aa)
FHA domain (85 aa)
31% I; 50% S 40% I; 58% S 53% I; 67% S
27% I; 50% S 31% I; 56% S no data
250 aa Hop1 N-terminal domain
S. c. Hop1 3 A. t. Hop1 no. 1 S. c. Hop1 3 A. t. Hop1 no. 2 S. c. Hop1 3 C. a. Hop1b a b
25% I; 50% S 24% I; 46% S 29% I; 57% S
For C. albicans Mek1p, only a 165-amino acid fragment is available. For C. albicans Hop1p, the fragment is 177 amino acids long.
promotes the Hop1p/Red1p interaction. Whether hetero-oligomer formation is enhanced because the Red1p-3HA phosphoprotein has decreased affinity for Red1–3HAp or an increased affinity for Hop1p remains to be determined. Co-immunoprecipitation of Hop1p and Red1–3HAp demonstrates that the two proteins physically interact in meiotic cells, but these experiments do not address whether the interactions are occurring on chromosomes or not. Given that Hop1p and Red1p co-localize on AEs (28), this idea seems likely. The genetic data presented in (9) are consistent with the interpretation of the co-IP experiment that MEK1 facilitates Hop1p/Red1p hetero-oligomer formation. However, the possibility cannot be ruled out that the diminished co-immunoprecipitation of Red1–3HA with Hop1p and vice versa in the mek1D is an artifact resulting from the decreased levels of soluble Red1–3HAp observed when MEK1 is absent. In the latter case, it may be that the MEK1-dependent phosphorylation of Red1p serves some other, as yet undefined, function during meiosis. One way to distinguish between these two possibilities is to identify and mutate the phosphorylation sites in Red1p and test to see if such mutants have decreased affinity for Hop1p. Such studies are currently underway. The mechanism for assembling AEs by phosphorylation-facilitated protein-protein interactions may be conserved in evolution. Homologs for Mek1p and Hop1 exist in S. pombe and C. albicans (Table IV). Within the amino terminus of Mek1p, but outside of the kinase homology, is a conserved sequence of 56 amino acids called the forkhead-associated domain (FHA) (33). This domain is found almost exclusively in nuclear proteins, including Dun1p and Spk1p, two protein kinases of S. cerevisiae involved in the cellular response to DNA damage (33). Interestingly, in Drosophila there is an ovarian-specific protein kinase called loki that also contains an FHA domain.4 It is intriguing to speculate that loki is the functional homolog of Mek1p in flies, especially given that SC formation only occurs in Drosophila females. The Asy1 gene of Arabidopsis thaliana encodes a protein with significant homology to Hop1p (Table IV), and asy1 mutant plants exhibit phenotypes (greatly reduced synapsis and chiasma formation)5 similar to hop1 mutants in yeast (32). Cor1/SPC3 may be the structural analog to Red1p in mammalian cells. Both proteins are meiosis-specific and are predicted to contain extensive coiled coil domains in their C termini (34, 35). Furthermore, the C termini of both proteins mediate homodimerization in the two-hybrid system (10, 36). SPC3 has been shown to be phosphorylated and, 4
S. Larochelle and B. Suter, personal communication. A. P. Caryl, F. C. H. Franklin, and G. H. Jones, personal communication. 5
1790
Red1p, a MEK1-dependent Phosphoprotein
furthermore, that the pattern of phosphorylation is dynamic throughout meiotic prophase (37). It is possible that the early phosphorylation events observed for SPC3 fulfill the same role in establishing the correct protein stoichiometry in mammalian AEs that it plays in the Hop1p/Red1p interaction. Acknowledgments—We thank Breck Byers, Neta Dean, JoAnne Engebrecht, and Aaron Neiman for helpful discussions and useful comments on the manuscript. We are also grateful to Ann Sutton and Simon Rudge for advice. Aaron Neiman provided the protein alignments given in Table IV. Neta Dean, Doug Kellogg, and Beth Rockmill provided plasmids and/or antibodies. We thank Lisa Ponte and Dana Woltering for excellent technical support. REFERENCES 1. Bascom-Slack, C. A., Ross, L. O., and Dawson, D. S. (1997) Adv. Genet. 35, 253–284 2. Roeder, G. S. (1997) Genes Dev. 11, 2600 –2621 3. Von Wettstein, D., Rasmussen, S. W., and Holm, P. B. (1984) Annu. Rev. Genet. 18, 331– 413 4. Heyting, C. (1996) Curr. Opin. Cell Biol. 8, 389 –396 5. Rockmill, B., and Roeder, G. S. (1990) Genetics 126, 563–574 6. Mao-Draayer, Y., Galbraith, A. M., Pittman, D. L., Cool, M., and Malone, R. E. (1996) Genetics 144, 71– 86 7. Bishop, D. K., Park, D., Xu, L., and Kleckner, N. (1992) Cell 69, 439 – 456 8. Xu, L., Weiner, B. M., and Kleckner, N. (1997) Genes Dev. 11, 106 –118 9. Friedman, D. B., Hollingsworth, N. M., and Byers, B. (1994) Genetics 136, 449 – 464 10. Hollingsworth, N. M., and Ponte, L. (1997) Genetics 147, 33– 42 11. Hollingsworth, N. M., Ponte, L., and Halsey, C. (1995) Genes Dev. 9, 1728 –1739 12. Rothstein, R. (1991) Methods Enzymol. 194, 281–301 13. Hollingsworth, N. M., and Johnson, A. D. (1993) Genetics 133, 785–797 14. Rose, M. D., Winston, F., and Heiter, P. (1990) Methods in Yeast Genetics: A laboratory course manual, Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, NY 15. Vershon, A. K., Hollingsworth, N. M., and Johnson, A. D. (1992) Mol. Cell. Biology 12, 3706 –3714 16. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y 17. Rockmill, B., and Roeder, G. S. (1991) Genes Dev. 5, 2392–2404 18. Smith, D. B., and Johnson, K. S. (1988) Gene 67, 31– 40 19. Smith, M. (1985) Annu. Rev. Genet. 19, 423– 462 20. Orr-Weaver, T. L., and Szostak, J. W. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 4417– 4421 21. Neiman, A. M., Mhaiskar, V., Manus, V., Galibert, F., and Dean, N. (1997) Genetics 145, 637– 645 22. Kironmai, K. M., Muniyappa, K., Friedman, D. B., Hollingsworth, N. M., and Byers, B. (1998) Mol. Cell. Biol. 18, 1424 –1435 23. Neiman, A. M., and Herskowitz, I. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 3398 –3402 24. Altman, R., and Kellogg, D. (1997) J. Cell Biol. 138, 119 –130 25. Padmore, R., Cao, L., and Kleckner, N. R. (1991) Cell 66, 1239 –1256 26. Hunter, T. (1991) in Methods in Enzymology (Hunter, T., and Sefton, B. M., eds) Vol. 200, pp. 3– 81, Academic Press, Inc., San Diego, CA 27. Leem, S.-H., and Ogawa, H. (1992) Nucleic Acids Res. 20, 449 – 457 28. Smith, A. V., and Roeder, G. S. (1997) J. Cell Biol. 136, 957–967 29. Bailis, J. M., and Roeder, G. S. (1998) Genes Dev., 12, 3551–3563 30. Hanks, S. K., Quinn, A. M., and Hunter, T. (1988) Science 241, 42–52 31. Ajimura, M., Leem, S., and Ogawa, H. (1993) Genetics 133, 51– 66 32. Hollingsworth, N. M., and Byers, B. (1989) Genetics 121, 445– 462 33. Hofmann, K., and Bucher, P. (1995) Trends Biochem. Sci. 20, 347–349 34. Dobson, J. J., Pearlman, R. E., Karaiskakis, A., Spryopoulos, B., and Moens, P. B. (1994) J. Cell Sci. 107, 2749 –2760 35. Lammers, J. H. M., Offenberg, H. H., van Aalderen, M., Vink, A. C. G., Deitrich, A. J. J., and Heyting, C. (1994) Mol. Cell. Biol. 14, 1137–1146 36. Tarsounas, M., Pearlman, R. E., Gasser, P. J., Park, M. S., and Moens, P. B. (1997) Mol. Biol. Cell 8, 1405–1414 37. Lammers, J. H. M., van Aalderen, M., Peters, A. H. F. M., van Pelt, A. A. M., Gaemers, I. C., de Rooij, D. G., de Boer, P., Offenberg, H. H., Dietrich, A. J. J., and Heyting, C. (1995) Chromosoma (Berl) 104, 154 –163 38. Sikorski, R. S., and Hieter, P. (1989) Genetics 122, 19 –27
