Mdy2, a ubiquitin-like (UBL)-domain protein, is required for efficient mating in Saccharomyces cerevisiae

Haploid S. cerevisiae cells occur in two mating types, a and α, and mating begins when cells of opposite mating type recognize each other by means of pheromones. MATα cells secrete α-factor, whereas MATa cells secrete a-factor. The receptors for a-factor and α-factor are encoded by STE3 in MATα cells and STE2 in MATa cells respectively, and belong to the large family of G-protein-coupled receptors that have seven transmembrane domains (Burkholder and Hartwell, 1985; Sprague et al., 1983). The binding of pheromone to receptor stimulates three major responses: (1) transcriptional induction of genes involved in mating, (2) arrest of the cells in G₁ phase, and (3) changes in cell morphology (for a review, see Herskowitz, 1995). The targets affected include genes whose functions are required for pheromone production and the pheromone response, and genes for proteins that facilitate or actually participate in cell pairing, cell cycle arrest and subsequent recovery from the arrested state, and mediate the morphological changes required for mating (Sprague and Thorner, 1992). One of the earliest effects of pheromone binding is cell cycle arrest late in the G₁ phase, prior to bud emergence, spindle pole body (SPB) duplication and DNA synthesis. Morphological changes ensue because mating cells sense the direction of the source of pheromone emanating from a mating partner (Segall, 1993). The receiving cell then undergoes polarized growth, forming a projection that is directed towards the signal source. This chemotropic response is thought to involve the generation of an internal landmark that reflects the direction of incidence of the external pheromone signal and overrides the spatial cues that normally control bud formation (Arkowitz, 1999; Chant, 1999).

Formation of the mating projection is mediated by the actin cytoskeleton and requires many other proteins (e.g. Spa2, Pea2, Bem1 and Cdc42) that are otherwise involved in bud emergence (Chenevert et al., 1994; Johnson and Pringle, 1990; Leeuw et al., 1995; Sheu et al., 1998). Components of the cytoskeleton, secretory system, plasma membrane and cell wall become asymmetrically reorganized along the axis defined by the pheromone source (Herskowitz et al., 1995; Matheos et al., 2004; Nern and Arkowitz, 2000; Read et al., 1992; Roemer et al., 1996), and the mating projection serves to concentrate the proteins involved in signalling (pheromones and pheromone receptors), cell adhesion (agglutinins) and fusion (Fus2) in the area of future cell contact and fusion. Thus, the proteins (α-factor, a-factor, Ste2, Ste3) that function specifically in these processes are all highly localized to the mating projections or their tips (Elion, 1995; Sprague and Thorner, 1992; Trueheart and Fink, 1989). Polarized mating cells, which are referred to as `shmoos', signal to one another through their projections, and thereby mutually restrict growth to the sites of cell contact and fusion. Many polarization-related genes also function in the cell fusion pathway, as indicated by the cell fusion defects observed in strains that are mutant for such genes (e.g. SPA2, PEA2, BNI1, RVS161, PRM1) (Dorer et al., 1997; Heiman and Walter, 2000).

Cell fusion usually occurs at the tips of the projections by formation of a conjugation tube or bridge. Nuclear fusion (karyogamy) is the last stage in the mating process and involves at least two steps. First, cytoplasmic microtubules emanating from the SPB bring the nuclei into close proximity, in a process called congression. The second step, karyogamy, entails the fusion of the nuclear membranes (for review, see Rose, 1996). The zygote then undergoes meiosis, and subsequently re-enters the vegetative cell cycle.

Although the cytological events involved in yeast mating have been well described, the molecular components and mechanisms important for mating cell morphogenesis, cell fusion and nuclear fusion are not well understood.

In the context of the EUROFAN project, deletion of YOL111c was shown to cause defects in mating (Iwanejko et al., 1999) (our unpublished data). In a subsequent study, Saeki et al. (Saeki et al., 2002) investigated whether Yol111c could bind to 20S proteasomes or ubiquitin chains, but could not find evidence for either. Here we present the results of an analysis of YOL111c, which we have renamed MDY2 (for mating-deficient yeast). MDY2 encodes a 212-residue protein, which contains an ubiquitin-like (UBL) domain. Its closest homologue is the 157-amino acid protein GdX found in humans. Here we present a phenotypic analysis of mdy2 mutants and conclude that Mdy2 directly or indirectly plays a role in shmoo formation and nuclear migration in the pre-zygote.

The MDY2 gene was identified in the course of the EUROFAN project on the basis of the mating defect observed in a mutant in which the ORF yol111c was disrupted in strain S288c. The first step in the characterization of MDY2 was the complete deletion of the gene using the LoxP-KanMX-LoxP gene disruption cassette (Gueldener et al., 1996) in the W303 strain background, and the subsequent analysis of mating efficiency using a quantitative mating assay (see Materials and Methods). As shown in Fig. 1A, the efficiency of diploid formation in the cross MATamdy2 × MATα MDY2 was 4±0.4% and in the reciprocal cross 6±1.3%; the corresponding value for the bilateral mutant cross was 3±0.9%. The wild-type strain shows a mating efficiency of 19±2.9%. Thus deletion of the MDY2 gene results in a five to sevenfold reduction in mating efficiency. Taken together, mating against a wild-type strain gave a higher mating efficiency (four- to fivefold) than mating involving two mdy2 mutants (five- to sevenfold). The effect of MDY2 deletion on the mating efficiency of MATa cells was slightly stronger than that on MATα cells.

Mating pheromones block the S. cerevisiae cell cycle in G₁ at START, and thus allow the cells to enter the mating pathway. G₁-arrested mating cells become large and pear-shaped (forming shmoos) as a result of mating projection formation, which allows them to aggregate, fuse and undergo karyogamy (Andrews and Herskowitz, 1990; Cross et al., 1988). To identify the step affected by deletion of MDY2, we first examined G₁ arrest and shmoo formation in the mdy2 mutant. Fig. 1B shows that the rate of α-factor-induced shmoo formation in MATamdy2 cells was reduced compared to wild-type cells. The incidence of G₁ arrest was decreased only slightly (data not shown). In the wild-type strain a maximum of 73% of the cells have been transformed into shmoos after 2 hours of exposure to α-factor, whereas only about 47% of mdy2 mutant cells were converted into shmoos by such treatment. The rest of the cells showed no projections. The mutant thus shows an approximate 30% reduction in shmoo formation in comparison to the wild-type strain. The ability to recover from pheromone exposure is also impaired in the mutant. The fact that G1 arrest is not significantly altered indicates that the signal transduction pathway activated by pheromone reception is functional in mdy2 cells.

Next, production of a-factor and α-factor was assessed by assaying for growth inhibition of pheromone-supersensitive strains in plate halo assays. Fig. 1C shows that the halo formed by the MATα mdy2 mutant was smaller than that induced by the wild-type strain after 48 hours of incubation at 30°C (wild-type MATa was used as control). This observation is consistent with the results obtained with an mdy2 deletion in the S288C background (results not shown) and suggests that the MATα mdy2 mutant is partially defective in the production or secretion of α-factor. The difference in halo size induced by MATα cells was not a consequence of reduced growth rate since the mutants did not show a significant growth defect with respect to the wild-type isogenic strains (data not shown). We then investigated whether the defect in α-factor production can be compensated for by mating to a supersensitive strain (sst2 or bar1). mdy2 crosses with MATα sst2 (HGX133) or MATabar1 (HZH350) showed a three- to fourfold reduction in mating efficiency relative to crosses of wild-type strains with MATα sst2 or MATa bar1, compared to the four- to fivefold reduction seen with non-supersensitive strains. Production of a-factor appeared to be normal when tested on a lawn of sensitized MATα sst2 cells (supplementary material Fig. S1). Taken together, these findings demonstrate that changes in pheromone levels cannot account for the mdy2 mating phenotype.

In the quantitative mating assay mdy2 mutants showed a five to sevenfold reduction in mating efficiency relative to wild type. Since the effects of the mdy2 mutation on the pheromone response pathway were relatively weak, the subsequent steps in the mating process were studied.

We first determined the number of mating pairs (pre-zygotes and zygotes) formed in mating mixtures of mdy2 and wild-type cells. After 2-4 hours of mating, mdy2 cells had formed 30-40% fewer mating pairs than the wild-type control (Fig. 2A). The difference in the numbers of mating pairs was even more marked (60%) after 5 hours. This could be largely explained as a direct consequence of a failure to form shmoos.

In wild-type mating mixtures, zygotes accounted for 10, 18, 16 and 10% of total cells after 2, 3, 4 and 5 hours of incubation, respectively, whereas only 3, 6, 5 and 3% of mdy2 mutants were zygotes at those time points (Fig. 2B).

To look specifically at the cell fusion phenotype in cells that formed shmoos, we performed a microscopic analysis of zygotes. Cell fusion was normal in mdy2 zygotes, as judged by the quantitative analysis of cell fusion (Gammie and Rose, 2002) based on monitoring the redistribution of cytosolic GFP present in the MATa or MATα partner following the formation of mating pairs, and the appearance of a septum in unbudded zygotes (supplementary material Figs S4 and S5). The first mitotic division of the zygotic nucleus is not impaired in mdy2 mutants.

To address the possibility that the deletion of MDY2 results in an accumulation of pre-zygotic stages (zygotes that have not undergone karyogamy), we classified mating pairs into three groups: pre-nuclear fusion zygotes, normal zygotes and abnormal zygotes. More than 200 mating pairs formed by mixtures of mdy2 and wild-type cells after a 3.5-hour incubation were analysed. In wild-type mating mixtures, pre-nuclear fusion zygotes accounted for nearly 17% of mating pairs, whereas 48% of mdy2 mutant mating pairs were found to be of this type (Fig. 2C). In wild-type mating mixtures 74% of mating pairs were zygotes at this point, in contrast to 43% for mdy2. Moreover, the positions of the nuclei in mdy2 pre-zygotes differ from those in the wild-type strain. In wild-type pairs the nuclei were closely juxtaposed, but in the mdy2 zygotes they remained further apart (Fig. 3A). The difference was quantified by measuring the distance between the two haploid nuclei: in wild-type zygotes the approximate mean value was 3.6 μm and in mdy2 zygotes 5.2 μm. This difference indicates that mdy2 cells that have fused successfully are impaired in zygote formation (Fig. 3B).

To determine the subsequent fate of the nuclei in mdy2 zygotes, the mating mixtures were analysed after 5.5 hours on YPD plates (Fig. 4). The zygotes were grouped into one of three classes according to the disposition of their nuclei: pre-nuclear fusion zygotes (class I, top row Fig. 4), post-nuclear fusion zygotes (class II, centre row), or zygotes in which mitotic division of haploid nuclei had occurred, indicating complete failure of karyogamy (class III, bottom row). As expected, after 5.5 hours most wild-type zygotes had completed nuclear fusion successfully and were in class II. By contrast, mdy2 exhibited a higher percentage of pre-nuclear fusion zygotes (class I). Moreover, many more mdy2 zygotes were in class III; i.e. karyogamy does not take place at all and the haploid nuclei undergo mitotic division (Fig. 4). Clearly, the mdy2 defect leads to a marked delay in karyogamy, as demonstrated by the fact that 44% of the mutant zygotes belong to Classes I and III compared to only 12% of wild-type zygotes.

Cytoductants are haploid cells containing one parental nucleus in a mixed cytoplasm contributed by both parents. In wild-type matings the number of cytoductants produced is very low. However, in Kar^– zygotes, the presence of two unfused nuclei greatly increases the frequency of cytoductant buds. In order to test whether this occurs in matings involving mdy2 cells, cytoductant analysis (Gammie and Rose, 2002) was performed (Table 1). As expected, mdy2 × wild-type crosses showed a significantly higher cytoductant:diploid ratio (0.13 for MATa and 0.11 for MATα) than wild-type × wild-type crosses (0.0053 for MATa and 0.0065 for MATα).

We next examined whether nuclear migration into pheromone-induced shmoos is defective in the mdy2 mutant. mdy2 and wild-type a cells were arrested by treatment with α-factor and scored for nuclear position using DAPI staining. In wild-type shmoos, the nucleus normally moves to the neck of the pear-shaped shmoo in preparation for mating (Miller and Rose, 1998; Read et al., 1992; Rose, 1991). In agreement with this observation, we found that in 40% of wild-type shmoos with projections (n=409) the nuclei were at the neck of, or in, the shmoo and in 4% of cells the nuclei were still on the opposite side of the cell from the shmoo (Fig. 5A). By contrast, in only 20% of mdy2 shmoos were the nuclei found in or at the shmoo neck, while 14% were on the side opposite the shmoo. Similar results were observed in strain S288c (data not shown). Since the nuclear envelope may be at some distance from the chromatin, we also determined the position of the nuclear envelope using a GFP-Nup116 fusion (Fig. 5B): Nup116 localizes to the nuclear pores (Fabre and Hurt, 1997). In 75% of wild-type shmoos (n>300) GFP-Nup116 was at the neck of, or in, the projection; in 3% of cells the GFP-Nup116 signal was on the opposite side of the cell from the shmoo. By contrast, only in 40% of mdy2 shmoos was the GFP-Nup116 found at the neck or in the shmoo, whereas 9% were on the side opposite the shmoo (Fig. 5B). These results are consistent with those obtained using DAPI staining, and confirm the presence of a defect in nuclear migration.

Thus the lack of Mdy2 seems to have two consequences. First, α-factor-induced shmoo formation is impaired, and second, migration of the nucleus is delayed or fails to occur. Since shmoo formation itself does not seem to be dependent on nuclear migration or attachment (Maddox et al., 2003), this might indicate the involvement of Mdy2 in an earlier function that affects both processes.

During mating, cytoplasmic microtubules are seen to enter the mating projection and apparently make contact with cortical sites (Carminati and Stearns, 1997; Miller and Rose, 1998). Thus, it seems likely that interactions between cortical proteins (like Kar9) and the cytoplasmic microtubules are a key element in nuclear migration and orientation. The cortical localization of Kar9 is particularly notable because it frequently intersects the ends of astral microtubules (Matheos et al., 2004; Miller and Rose, 1998). Since mdy2 mutants exhibit defects in nuclear migration, we wanted to examine microtubule patterns (using GFP-Tub1) and Kar9 localization (using GFP-Kar9) (Fig. 6). In the wild-type shmoos, 77% of cells had a single bundle of cytoplasmic microtubules extending into the shmoo tip; 9% contained a single bundle of cytoplasmic microtubules directed toward the shmoo tip and a second bundle pointing away from the shmoo tip; 14% exhibited misoriented microtubules that were not near the shmoo tip. By contrast, only 35% of the mdy2 cells had a single bundle of cytoplasmic microtubules going to the shmoo tip and 29% contained a single bundle of cytoplasmic microtubules going to the shmoo tip plus an additional bundle directed away from the shmoo tip. There was an increase (36%) in the number of cells with misoriented microtubules (Fig. 6A).

In cells expressing a GFP-Kar9 fusion, the fluorescent signal localizes most frequently to a single spot at the tip of the growing bud and the mating projection (Miller et al., 1999). As expected, in the wild-type shmoos, GFP-Kar9 was localized most frequently (67%) as a dot at the shmoo tip; in a fraction of the cells (22%) the signal formed a line emanating from the shmoo tip (Lee et al., 2000; Matheos et al., 2004; Miller et al., 1999). By contrast, in the mdy2 mutant 70% of shmoos exhibiting mislocalized GFP-Kar9 (Fig. 6B). From these results, we concluded that the defect in nuclear migration of mdy2 mutants could be due to a lesion in the process leading to localization of Kar9.

In order to confirm that lack of the MDY2 function is responsible for the mutant phenotypes observed, mdy2 mutants of both mating types were transformed with 2 μ MDY2 plasmids. Wild-type cells transformed with the empty 2 μ vector served as controls. The MATamdy2 mutant bearing 2 μ MDY2 plasmids was crossed to the MATα mdy2 mutant bearing 2 μ MDY2 plasmids; this cross resulted in diploid formation with an efficiency of 23±3.7%. The MATaMDY2 × MATα MDY2 cross had an efficiency of 16±2.7%. Thus the overexpression of MDY2 in the mdy2 deletants suppressed the mutant phenotype and indeed slightly enhanced mating efficiency relative to the wild-type strain (Fig. 7A). mdy2 mutants harbouring 2 μ MDY2 formed haloes of nearly the same size as the wild-type strain after 48 hours of incubation at 30°C (Fig. 7B). Hence the introduction of 2 μ MDY2 plasmids into MATα mdy2 cells also complements the defect in α-factor production observed in mdy2 deletants.

In order to investigate how overexpression of MDY2 affects the shmoo index, MATamdy2 cells were transformed with 2 μ MDY2 plasmids or the 2 μ vector. The cells were then treated with 5 μM α-factor, samples were taken at different times, and the numbers of shmoos were counted. The results showed that overexpression of MDY2 leads to a slight increase in the rate of formation of shmoos as compared with the wild-type strain (Fig. 7C).

To study the subcellular localization of Mdy2, we constructed a plasmid carrying MDY2 fused to the gene for GFP. The GFP-Mdy2 fusion protein was functional in mating assays and complemented the mating defect of the mdy2 mutant. In vegetatively growing cells, GFP-Mdy2 is localized in the nucleus (Fig. 8A). In cells treated with α-factor, GFP-Mdy2 was also found to be predominantly nuclear (Fig. 8B). Since overexpression may influence the cellular localization of a protein of interest, GFP-MDY2 was also expressed under the control of the MDY2 promoter. No change was noted in the localization of the protein (Fig. 8C). Furthermore, after treatment with α-factor, GFP-Mdy2 was still found to be nuclear (Fig. 8D).

To study the possible influence of pheromone on the level of Mdy2, a plasmid encoding a GST-tagged version of Mdy2 under the control of the GAL1 promoter was constructed. A yeast strain bearing this allele was grown to early log phase in SRG medium (3% raffinose and 1% galactose), which permits transcription of GST-MDY2. The cells were then transferred to a medium containing glucose (which blocks further transcription of GST-MDY2) and 5 μM α-factor (final concentration). Whole cell extracts were prepared from aliquots of the culture at the indicated time points after the shift, and GST-MDY2 was detected by immunoblot analysis with anti-GST antibodies. After galactose depletion and pheromone induction for 90 minutes there is no significant difference in the level of GST-Mdy2, but after 150 minutes a decrease is observed (supplementary material Fig. S2). These data indicate that Mdy2 is relatively stable in the presence of pheromone.

Ubiquitin and UBL proteins, such as NEDD8 or SUMO, are proteolytically processed such that an internal conserved Gly is exposed at the C terminus; this residue is used for their covalent attachment to other proteins. The MDY2 open reading frame (residues 74-149) exhibits some similarity to ubiquitin and 33% similarity to GdX (Fig. 9A), but lacks glycines proximal to its predicted C terminus. It is, however, conceivable that one of the other conserved residues located near the C terminus is used for processing and conjugation. If Mdy2 acts as a modifier protein, it should be processed to remove non-conserved residue(s) at the C-terminal end. To examine whether processing of Mdy2 occurs, plasmids encoding Myc-Mdy2-H6, GST-Mdy2-H6 and GFP-Mdy2-H6 were constructed. The His-tagged fusion proteins were functional in mating assays and complemented the mating defect of the mdy2 mutant. Western analysis revealed that extracts of mdy2Δ cells expressing Mdy2-His6 from a plasmid retain the C-terminal 6xHis tag, which is detectable with an anti-His antibody in both unstimulated cells and cells treated with α factor (Fig. 9B, left panel). In strains carrying vectors encoding Myc, GST and GFP alone, no band was detected at this level (Fig. 9B and data not shown). It is noteworthy that all three antibodies detect two forms of each fusion protein (Fig. 9B and data not shown). Mdy2H6 has a molecular mass of 24.5 kDa, but the anti-His antibody labels bands of 24-26 kDa, which may result from non specific partial processing of protease cleavage sites in the spacer region between the N-terminal tags and the Mdy2H6 protein. The intracellular localisation of the Mdy2-His6 variant was determined by immunofluorescence experiments using anti-His antibody (Fig. 9C). Like GFP-Mdy2, the C-terminally His-tagged protein is localized in the nucleus. This result indicates that the tag is not removed by C-terminal processing.

The function of the MDY2 gene was characterized in detail because the corresponding loss-of-function mutant shows a strong impairment in mating and the gene product contains an ubiquitin-like domain (the UBL domain). In quantitative mating assays mdy2 mutants showed an approximately five to sevenfold reduction in the efficiency of diploid formation. While mdy2 cells display slight defects in the pheromone signal transduction pathway (data not shown), they have much more severe defects in the formation of shmoos, mating pairs and zygotes. Deletion of MDY2 causes a 30% reduction in both shmoo and mating pair formation. The strongest effects are seen at the level of zygote formation and mating frequency. The deletion phenotype first becomes manifest during shmoo formation. We find that in mdy2 cells migration of the nucleus toward the shmoo is delayed or fails to occur. Since shmoo formation itself appears not to be dependent on nuclear migration or attachment (Maddox et al., 2003), Mdy2 might participate in both processes.

Mdy2 is not inducible by pheromone, and is therefore likely to be involved in other processes besides mating. Exposure of cells to pheromone for 90 minutes did not detectably alter levels of GFP-Mdy2, and only a slight decrease was noted after 150 minutes. We tested several proteins from the pheromone response pathway (Ste50, Ste11, Ste5, Ste7, Fus3, and Kss1) for interaction with Mdy2 but none were detected.

Overexpression of MDY2 complemented the defects of the mdy2 mutant such as reduced shmoo formation and α-factor production. The mdy2 deletion mutant harbouring MDY2 on a high-copy-number plasmid also showed a slightly higher efficiency of diploid formation than the wild-type strain bearing the empty vector.

In addition, overexpression of MDY2 under the control of the GAL1 promoter in MATa cells led to an increase in the level of induction of FUS1 in the presence of pheromone (supplementary material Fig. S2), suggesting that Mdy2 can act upon the transcriptional response to mating pheromone.

The S. cerevisiae shmoo tip is a model system for analysing the mechanism that couples force production to microtubule plus ends. The SPB is a trilaminar structure containing a central plaque which is physically embedded in the nuclear envelope, and inner and outer plaques that nucleate the nuclear and cytoplasmic microtubules, respectively (Byers and Goetsch, 1974; Byers and Goetsch, 1975). Dynamic plus ends of astral microtubules, emanating from the SPB, interact with the shmoo tip in mating yeast cells, positioning the nuclei for karyogamy (Maddox et al., 1999). The movement of the nucleus toward the preshmoo and into the shmoo tip occurs via microtubule polymerisation/depolymerisation at the plus end. Individual microtubules grow and shorten in a random search-and-capture process like that described for vegetative G1 cells (Shaw et al., 1997). Once astral microtubules encounter the shmoo tip, subsequent microtubule dynamics results in pushing and/or pulling forces that tether the SPB and nucleus to the shmoo tip (Maddox et al., 1999). Several genes are required for correct positioning of the nuclei. Thus mutants deficient for KAR3 and KAR9 are defective in nuclear migration during pheromone induction. Kar9, a putative cortical protein, is required for cytoplasmic microtubule orientation and is thought to anchor the protein Bim1, which attaches to growing microtubule plus ends at the shmoo tip (Miller et al., 2000; Miller and Rose, 1998). Deletion of KAR3 prevents the attachment of astral microtubules to the shmoo tip, and Kar3 has therefore been proposed to maintain the attachment of depolarising microtubules (Maddox et al., 2003). The fact that nuclear migration toward the developing shmoo in mating cells, and nuclear congression and karyogamy in zygotes, are impaired in mdy2 cells strongly implies that Mdy2 is involved in the process of nuclear migration. Since mitosis itself is not affected in mdy2 cells, Mdy2 function seems to be restricted to mating cells, a feature it shares with Kar3. However, unlike Kar3, Mdy2 is mainly localized in the nucleus. Only upon overexpression is a small fraction of the GFP-Mdy2 fusion protein seen in the mating projections. During the pheromone response, mdy2 mutants were defective in Kar9 localization, and in microtubule alignment. The defect in Kar9 localization and microtubule orientation may indicate that Mdy2 interacts with other proteins that affect the cortical capture of microtubules and spindle orientation. Therefore, Mdy2 may directly or indirectly participates in events at the SPB or other nuclear sites during the migration process; however, a function at the shmoo tip cannot be excluded.

The functional analysis of Mdy2 in this study clearly shows that it is required for efficient mating in S. cerevisiae and, unlike most UBL proteins, it is apparently not subject to C-terminal processing. We could not detect any conjugation product of Mdy2 and therefore speculate that it may function by interacting with the UBA domain of a putative target protein. The closest homologue of Mdy2 is the human ubiquitin-like protein GdX, which consists of 157 amino acids.

The shared regions of homology show 34% identity and encompass residues 74-212 of Mdy2 and the N-terminal 123 residues of GdX. Thus the homology extends outside the UBL region (76 amino acids), which shares approx. 20% identity and 20% similarity with ubiquitin (Fig. 9A). GdX is a single-copy gene that is conserved in evolution and is regarded as a `housekeeping' gene that encodes a protein similar to ubiquitin (Toniolo et al., 1988). A growing family of ubiquitin-like proteins has been uncovered in yeast, including Rub1, Dsk2, Rad23, Apg12, Smt3 and Hub1 (Dittmar et al., 2002; Hochstrasser, 1996; Hochstrasser, 2000). The functional consequences of modification by ubiquitin-like proteins appear to be distinct from those associated with ubiquitinylation, in that the ubiquitin-like species do not typically induce the degradation of their protein targets.

Currently, ubiquitin-like proteins are divided into two subclasses (Jentsch and Pyrowolakis, 2000). Type-1 ubiquitin-like polypeptides (UBLs) consist essentially of only the UBL domain and function as modifiers like ubiquitin, being ligated to target proteins in a process similar, but not identical, to the ubiquitylation pathway; they include SUMO, NEDD8 and UCRP/ISG15. In type-2 UBLs (also referred to as UDPs, ubiquitin-domain proteins) the UBL domain is found as part of a larger protein, which obviously has other functions. Among type-2 UBLs are Rad23, Dsk2, Elongin B and GdX, which do not form conjugates with other proteins (Jentsch and Pyrowolakis, 2000). One subclass of UDPs, including Rad23 and Dsk2, contains a UBL domain at the N-terminal end and an ubiquitin association (UBA) domain at the C terminus (Hofmann and Bucher, 1996). Such proteins appear to function as adapter proteins, interacting with both polyubiquitinylated proteins and the 26S proteasome (Chen and Madura, 2002; Elsasser et al., 2004; Funakoshi et al., 2002; Verma et al., 2004). However, the UBA domain may stabilize the protein by interacting with the internal UBL domain (Ortolan et al., 2004). DSK2 was isolated as a suppressor of kar1, which is defective in spindle pole duplication (Biggins et al., 1996). Dsk2 can interact with the proteasome, and an adaptor function in the ubiquitin-proteasome pathway has been discussed (Funakoshi et al., 2002). Our results indicate that Mdy2 is not subject to C-terminal processing or conjugation to other proteins. Mdy2 does not contain a UBA domain, but could interact with UBA domains of other UDPs such as Dsk2. However, Mdy2 was not detected in a two-hybrid screen using an N-terminally truncated Dsk2 (Funakoshi et al., 2002).

In summary, we have found that Mdy2 plays a role in shmoo formation and nuclear congression during the mating process in S. cerevisiae. Additional work will be required to unravel its function and the significance of its ubiquitin-like domain.

Recombinant DNA techniques were performed according to standard protocols (Sambrook and Russell, 2001). Yeast cells were grown either in YPD medium (2% glucose, 2% peptone, 1% yeast extract) or in synthetic complete (SC) medium (0.67% yeast nitrogen base and 2% glucose), each containing the required nutrient supplements. The strains and plasmids used are listed in Table 2. The mdy2 mutants were constructed by replacing the entire MDY2 open reading frame with the KanMX4 marker. To construct pZH149 (pRS426-MDY2), the MDY2 gene (990 bp) was inserted into the SpeI and KpnI sites of the vector pRS426. Plasmid pZH80 (pGREG576GFP-MDY2) was constructed by in vivo homologous recombination between the MDY2 ORF (obtained by PCR) and the vector pGREG576 (Jansen et al., 2005). Plasmid pZH81 (pGREG546GST-MDY2) was constructed by in vivo homologous recombination between the MDY2 ORF and pGREG546. To construct pZH152, the MDY2 promoter in pGREG576GFP-MDY2 was replaced by the GAL1 promoter. The plasmids p526MDY2H6 (Myc-Mdy2-H6), p546MDY2-H6 (GST-Mdy2-H6) and p576MDY2H6 (GFP-Mdy2-H6) were constructed by in vivo recombination between the 0.7-kb MDY2-6xHis fragment and the pGREG vectors (Jansen et al., 2005). PCR fragments were obtained using pYCG-YOL111c (Cognate Gene, EUROSCARF) as template together with the primers 5′-GAATTCGATATCAAGCTTATCGATACCGTCGACAATGAGCACATCCGCCAGCGG-3′ and 5′-CATGACTCGAGGTCGACTTAGTGATGGTGATGGTGATGTTTGGCCAGAGACCAGCC-3′.

Quantitative mating assays were performed as described previously (Elion et al., 1990; Rad et al., 1992). Cultures of opposite mating type (MATa and MATα) were grown to log phase in SD medium. Equal numbers (3×10⁶) of cells of each mating type were mixed and pelleted by centrifugation at 20°C and 10,200 g for 4 minutes. The cell pellets were then transferred onto nitrocellulose membranes (NC) placed on YPD plates. The plates were incubated at 30°C for 4.5 hours, after which the cells were transferred to Eppendorf tubes for serial 10-fold dilution. Aliquots (100 μl) of each dilution were spread on SD plates and the plates were incubated at 30°C for 2 days. The total numbers of cells (N_t) and the numbers of diploid cells (N_d) were then counted. The mating efficiency is calculated as the number of colonies of a/α diploids (N_d) divided by the sum of a/α diploids plus haploid colonies (N_t). Crosses between two wild-type strains were used as controls. All tests were carried out in duplicate.

Projection formation and budding were assayed essentially as described previously (Xu et al., 1996). Cultures of mutant (MATamut) and wild-type (MATa) strains were grown to log phase (OD₆₀₀ 0.2-0.4). α-Factor was added to a final concentration of 5 μM. Samples were taken after incubation with α-factor for 0, 30, 60, 90, 120, 180, 240 and 300 minutes and fixed immediately with 3.7% formaldehyde. Cells were washed twice with 1× PBS buffer prior to microscopy. At least 300 cells per sample were counted and the numbers of shmoos and unbudded cells were recorded. The shmoo index is defined as the percentage of shmoos relative to the total number of cells.

The assay for zygote formation was performed essentially as described previously (Elion et al., 1990; Rad et al., 1992). Cultures of opposite mating type (MATa and MATα) were grown to log phase in YPD medium. Equal numbers (3×10⁶) of cells of each mating type were mixed and pelleted by centrifugation at 20°C and 10,200 g for 4 minutes. The cell pellets were transferred onto NC filters on YPD plates, and incubated at 30°C for various times (1, 2, 3, 4 and 5 hours). The cells on each membrane were then washed off with 1 ml of sterile water, and pelleted by centrifugation at 20°C and 10,200 g for 4 minutes. The pellets were resuspended in 1× PBS buffer and fixed with 70% ethanol. Cells were washed twice with 1× PBS buffer prior to observation under the microscope. The proportions of zygotes were counted in samples containing at least 800 cells, and the efficiency of zygote formation was calculated. Nuclei were stained with DAPI (4′,6-diamidino-Z-phenylindole; 1 μg/ml), and the different stages of zygote development were analyzed under the microscope. The numbers of pre-zygotes (with unfused nuclei), zygotes and abnormal zygotes in samples containing at least 200 zygotes were determined. Cytoductant analysis and microscopic analysis of the zygotes (Berlin et al., 1991; Gammie and Rose, 2002) were performed to detect defects in cell and nuclear fusion. ρ⁰ strains were obtained by growing cultures in aluminium foil-wrapped flasks containing YPD and 10 μg/ml ethidium bromide at 30°C overnight. The cultures were spread on YPD plates and incubated at 30°C. After a 2-day incubation the colonies were replicated onto YEP glycerol plates and incubated at 30°C. ρ⁰ strains (completely lacking mitochondrial DNA) were identified by DAPI staining. ρ⁰ (can^R) and ρ⁺ (CAN^S) strains were grown to log phase and approximately equal numbers of cells (3×10⁶) were mixed and mated for 4 hours on an NC filter (0.45 μm pore size). Samples were collected and 10-fold serial dilutions (10^–1, 10^–2, 10^–3 and 10^–4) were prepared using sterile water. Samples (100 μl) were plated onto synthetic complete medium minus arginine containing 3% glycerol and canavanine (60 μg/ml). The total numbers of cells and diploids were determined by plating dilutions of the mating mixture onto minimal media. The efficiency of nuclear fusion is expressed as the ratio of cytoductants to diploids (C:D).

Cultures of strains harbouring a plasmid carrying the GFP-tagged genes under the control of the GAL1 promoter were grown to log phase in SRG medium. Cells were harvested and washed with 1× PBS buffer (pH 7.4) in Eppendorf tubes. The cells were sonicated briefly to break up aggregates, and fixed with 70% ethanol for 12 minutes. Then the cells were washed again with 1× PBS buffer, stained with DAPI at room temperature for 12 minutes, and washed again with 1× PBS buffer. The distribution of tagged fusion protein and nuclei was investigated by fluorescence microscopy. Cells were processed for immunofluorescence essentially as described previously (Pringle et al., 1989). Cells were grown to exponential phase (OD₆₀₀=0.5-0.8, 3-4×10⁷/ml) and immediately fixed with formaldehyde (final concentration 5% for 4 hours). The fixed cells were spheroplasted and extracted with 0.1% Triton X-100 for 5 minutes and then attached to a multiwell slide treated with 0.1% poly-lysine (Sigma, St Louis, MO, USA). The cells were first incubated with an anti-His mouse monoclonal antibody (Qiagen; 1:200 dilution in PBS plus 1 mg/ml bovine serum albumin) for 90 minutes and then for 90 minutes with fluorescein isothiocyanate (FITC)-conjugated anti-mouse antibodies (Dianova, Hamburg, Germany; 1:300 dilution in PBS/BSA). Finally, the cells were incubated for 5 minutes with DAPI (1 μg/ml). Fluorescence was visualized with an Axioskop fluorescence microscope (Zeiss, Oberkochen, Germany) using a FITC filter set, and images were obtained with a Zeiss AxioCam digital camera.

Gene disruption in S. cerevisiae was performed according to the method described previously (Gueldener et al., 2002). The fragment carrying the LoxP-kanMX-LoxP disruption cassette was amplified by PCR. PCR products were precipitated with ethanol, recovered by centrifugation and dissolved in 10 μl of water. Yeast cells were transformed with the LoxP-kanMX-LoxP disruption cassette by the high-efficiency LiAc transformation method. Single colonies were isolated from YPD plates containing G418 and prepared for further analysis. A well separated colony growing on an YPD plate containing G418 was picked with a pipette tip and resuspended in 40 μl of 0.02 M NaOH solution in an Eppendorf tube. The tube was heated in a microwave for 90 seconds, and then incubated at room temperature for 30 minutes. The sample was then stored at –20°C for 1 hour. PCR was subsequently performed to check the length of the fragments.

The galactose depletion assay was performed according to a previously described protocol (Esch and Errede, 2002). Cultures were first grown to early log phase in the appropriate selective medium containing 3% raffinose and 1% galactose to induce expression of pGAL-GST-MDY2. Cells were then harvested, washed, and transferred to medium containing 2% glucose to inhibit further transcription of the fusion gene. After the removal of a control sample, incubation of the cultures was continued with or without (5 μM) α-factor to activate the mating pheromone response pathway. Whole cell extracts were then prepared for immunoblot analysis.

We wish to thank J. H. Hegemann, G. Jansen, M. D. Rose and K. Bloom for plasmids and strains. We also thank P. Hardy for reading the manuscript and for helpful discussions. Z. Hu was the recipient of a DAAD doctoral fellowship.