The Dysferlin Domain-Only Protein, Spo73, Is Required for Prospore Membrane Extension in Saccharomyces cerevisiae

Prospore membrane formation consists of de novo double-membrane formation, which occurs during the developmental process of sporulation in Saccharomyces cerevisiae. Membranes are formed into their proper size and shape, and thus, prospore membrane formation has been studied as a general model of membrane formation. We identified SPO73, previously shown to be required for spore wall formation, as an additional gene involved in prospore membrane extension. Genetic and cell biological analyses suggested that Spo73 functions on the prospore membrane with other factors in prospore membrane extension, counteracting the bending force of the prospore membrane. Spo73 is the first dysferlin domain-only protein ever analyzed. The dysferlin domain is conserved from yeast to mammals and is found in dysferlin proteins, which are involved in dysferlinopathy, although the precise function of the domain is unknown. Continued analysis of Spo73 will contribute to our understanding of the function of dysferlin domains and dysferlinopathy.

T he yeast Saccharomyces cerevisiae initiates the developmental process of sporulation under conditions of nitrogen deprivation and the presence of a nonfermentable carbon source. Four spores are formed inside a diploid cell, which turns into an ascus. In this process, four new membranes called prospore membranes are formed in the cytosol of the diploid cell during meiosis II, encapsulate the four nuclei, respectively, and become plasma membranes of each spore (1)(2)(3).
A prospore membrane grows through a series of discrete morphological stages (Fig. 1A). (i) It appears as a small horseshoe-shaped structure at the cytoplasmic surface of the spindle pole body by fusion of post-Golgi vesicles; (ii) it expands into a small round shape; (iii) as meiosis II progresses, it extends along the nuclear envelope and is shaped into a tube-like or cashew nut-like structure engulfing the nucleus, cytoplasm, and organelles; and (iv) at the end of the meiotic process, it quickly turns into a round shape of mature size, just following or at the same time as the closure of its leading edge (4)(5)(6). Prospore membrane formation is a process of de novo membrane formation into a proper size and shape; thus, it is considered to be a model of biological membrane formation.
Previous studies have shown that VPS13 and SPO71 function in prospore membrane extension. VPS13 was originally identified as a gene involved in trans-Golgi networkendosomal cycling of a subset of membrane proteins and was shown to be required for vacuolar protein sorting (7). VPS13 also has a function in sporulation; vps13⌬ mutant cells can only form small prospore membranes and are sporulation defective (8,9). SPO71, which encodes a dual pleckstrin homology (PH) domain protein, was identified in screening for genes upregulated during and required for sporulation (10). spo71⌬ mutant cells also form small prospore membranes, although the defect looks less severe than that in vps13⌬ mutant cells (11). An analysis of the relationship between these proteins revealed that Spo71 interacts with Vps13 and recruits it to the prospore membrane, where both proteins colocalize (12). Further, Vps13 and Spo71 are suggested to have a function in regulating the levels of phosphatidylinositol 4-phosphate (PI4P) in prospore membranes; PI4P pools of prospore membranes are reduced in vps13⌬ and spo71⌬ mutants (12). PI4P has a variety of cellular roles, including the regulation of membrane traffic through the Golgi apparatus (13). Thus, decreased levels of PI4P in the prospore membrane of these mutant cells could cause some defect in prospore membrane extension. However, the overall mechanism of prospore membrane extension is still unknown.
The spore membrane-bending pathway and the leading-edge complex (LEC), both of which are important for the shape of the prospore membrane, could also contribute to proper prospore membrane extension. The spore membrane-bending pathway consists of a phospholipase B, Spo1; a membrane protein, Sma2; and glycophosphatidylinositol-anchored proteins, including Spo19 (14). These proteins are suggested to exert a bending force toward the nucleus during prospore membrane growth to facilitate its proper shaping; in spo1⌬ and sma2⌬ mutants, prospore membranes become wide and open. The LEC is a protein complex that localizes to the lip of each prospore membrane and consists of at least four proteins, including Don1 and an essential component, Ssp1 (15)(16)(17). Deletion of the SSP1 gene causes the formation of unnaturally elongated prospore membranes that stick tightly to the nuclear envelope (15), suggesting that the LEC counteracts the bending force of the prospore membrane until it is removed at the onset of closure of the prospore membrane (4,18).
In this study, we performed a screening for additional genes involved in prospore membrane extension and identified SPO73, which was previously suggested to be involved in spore wall formation (19). spo73⌬ mutant cells show phenotypes similar to those of spo71⌬ mutant cells. We also show that a Spo73 protein carrying mutations in a surface basic patch mislocalizes to the cytoplasm and overexpression of Spo71 can partially rescue localization to the prospore membrane. Additionally, we find that, as is the case with SPO71 (11), SPO73 genetically interacts with SPO1 and SMA2; double mutants partially recover sporulation, suggesting that they have opposing functions in prospore membrane extension. Further, our bioinformatic analysis revealed that Spo73 is a dysferlin domain-only protein. Dysferlin domains are found in a subset of ferlin family proteins, including dysferlin, which is associated with the plasma membrane, t-tubule network, and intracellular vesicles in skeletal muscle and is involved in Ca 2ϩdependent plasma membrane repair (20). Mutations in this domain of the dysferlin protein cause dysferlinopathies, including limb girdle muscular dystrophy type 2B and Miyoshi myopathy (20)(21)(22)(23), although the precise function of the domain is not known. Thus, analysis of Spo73 will help our understanding of the function of dysferlin domains and may contribute to understanding of the molecular basis of dysferlinopathies.

RESULTS
Screening for mutants defective in prospore membrane extension identified the spo73⌬ mutant. Previous screening for mutants defective in prospore membrane formation identified eight mutants (8). One of them, the vps13⌬ mutant, formed small round prospore membranes, indicating that VPS13 is required for proper prospore membrane extension. However, the overall mechanism of prospore membrane extension was almost unknown. To obtain additional mutants defective in prospore membrane extension, we expanded the screening to 293 mutants (for a list of corresponding genes, see Table S1 in the supplemental material) that were selected from the deletion collection based on previous reports of genome-wide analyses of sporulation defects (24)(25)(26). Mutants analyzed in previous screenings, mutants with changes in functionally characterized genes, and strains that did not grow well were excluded from this study, and the vps13⌬ mutant was used as a control. A green fluorescent protein (GFP)-tagged Spo20 lipid-binding domain (27) was expressed as a prospore membrane marker in those mutants and observed under a microscope during sporulation, one by one. We identified two mutants, the spo71⌬ and spo73⌬ mutants, that showed only a small prospore membrane pattern suggestive of a prospore membrane extension defect (Fig. 1B). During the course of our study, Spo71 was reported to be involved in prospore membrane extension (11). Therefore, we focused on Spo73. SPO73 encodes a 143amino-acid protein with a dysferlin C-terminal domain in its C-terminal region (19). Analysis with HHpred revealed that the Spo73 protein also contains a dysferlin N-terminal domain in its N-terminal region; thus, Spo73 is a dysferlin domain-only protein. As all of the dysferlin domain-containing proteins characterized so far contain other domains (20,28), such as transmembrane and/or c2 domains, Spo73 is the first example of a dysferlin domain-only protein. Spo73 was first identified in a genomewide screening for genes upregulated during and involved in sporulation and has been shown to be required for spore wall formation (10,19). No relatedness of Spo73 to prospore membrane formation has previously been reported.
The small prospore membrane phenotype of the spo73⌬ mutant was confirmed by time-lapse video microscopy. To observe sporulation more precisely, fast-sporulating SK1 background strains were used. In wild-type cells, prospore membranes appeared as four horseshoe-shaped membrane structures, enlarged to form four small round shapes, extended to form tube-like shapes, and finally closed and quickly attained their spherical shapes ( Fig. 1C; see Movie S1 in the supplemental material). In spo73⌬ mutant cells, prospore membranes appeared and enlarged to form four small round membranes, as observed in wild-type cells ( Fig. 1C; see Movie S2 in the supplemental material). However, prospore membrane extension stopped at this stage and extension to tube-like structures was never observed. This phenotype was similar to that of spo71⌬ mutant cells ( Fig. 1C; see Movie S3 in the supplemental material). These results indicated that a dysferlin domain-only protein, Spo73, is involved in prospore membrane extension.
Capture of the nuclei by prospore membranes is partially defective in spo73⌬ mutant cells. Small prospore membranes in spo73⌬ mutant cells sometimes failed to capture each nucleus, although it has been reported that small prospore membranes in spo71⌬ mutant cells capture each nucleus efficiently (11). Capture of nuclei by prospore membranes was examined in wild-type and spo73⌬ mutant cells. Histone 2B tagged with mCherry was used as a nuclear marker and observed with a prospore membrane marker during sporulation. In contrast to wild-type cells, in which the vast majority of the nuclei were captured by prospore membranes, 42% of the prospore membranes failed to capture nuclei in spo73⌬ mutant cells ( Fig. 2A). In addition, we found that, in contrast to a previous report, 40% of the prospore membranes failed to capture nuclei in spo71⌬ mutant cells ( Fig. 2A). To confirm that the spo73⌬ mutant has a defect in nuclear capture, time-lapse analysis was performed. While nuclear capture occurred efficiently in wild-type cells ( Fig. 2B; see Movie S4 in the supplemental material), in spo73⌬ mutant cells, we easily found cells in which prospore membranes failed to capture nuclei ( Fig. 2B; see Movie S5). In these cells, the nuclei, marked by Htb2-mCherry, stuck at the leading edge of the prospore membrane and could not be engulfed. Thus, spo73⌬ mutant cells have a partial defect in nuclear capture by prospore membranes, and a similar phenotype is also seen in spo71⌬ mutant cells.
Phenotype of the spo73⌬ mutant is similar to that of spo71⌬ and vps13⌬ mutants. We saw similar defects in prospore membrane extension and nuclear capture in spo73⌬ and spo71⌬ mutant cells. spo71⌬ mutant cells have been shown to form prospore membranes with bubbles or intraluminal vesicles between double membranes, have a partial defect in removal of the LEC, and form prospore membranes with decreased levels of PI4P (12). Thus, we examined spo73⌬ mutant cells for these phenotypes. First, we performed electron microscopy (EM) analysis. Wild-type and spo73⌬ mutant cells were sporulated and analyzed by EM. In wild-type cells, all of the prospore membranes looked to be closely apposed double membranes except for those in cells with apparent spore walls (Fig. 3A). In contrast, about 40% (68 in 167) of spo73⌬ mutant prospore membranes showed bubbles or intraluminal vesicles ( Fig. 3B to D), which have previously been reported in spo71⌬ and vps13⌬ mutant cells (12). Next, we assessed the removal of the LEC, which is required for prospore membrane closure. Postmeiotic cells selected by 4=,6-diamidino-2-phenylindole (DAPI) staining were observed to detect LECs by using Don1-GFP as a marker. While most LECs were removed in wild-type postmeiotic cells, the LECs of 30% of the prospore membranes were still present in postmeiotic spo73⌬ mutant cells, suggesting a partial defect in prospore membrane closure (Fig. 3E). Finally, PI4P levels at the prospore membrane were examined by using GFP-PH OSH2 (29) as a PI4P biosensor and Dtr1 fused to red fluorescent protein (RFP) (9) as a prospore membrane marker. Although the original procedure was mostly reproducible, we revised the experimental procedure to observe sporulating cells at 7 to 9 h at 30°C to focus on cells around the time of prospore membrane formation. In all three strains, about 80% of the RFP-positive prospore membranes were GFP positive, indicating that PI4P levels were not changed in wildtype and spo73⌬ and spo71⌬ mutant cells (Fig. 3F). This is in contrast to an earlier report indicating that PI4P levels are reduced in spo71⌬ mutant cells (12). The reason for this discrepancy is unclear; nonetheless, these results indicate that the spo73⌬ mutant has prospore membrane extension defects similar to those of the spo71⌬ and vps13⌬ mutants.
Spo73 localizes to the prospore membrane independently of Spo71 and Vps13. Coluccio et al. have shown that hemagglutinin-tagged Spo73 localizes to dots on prospore membranes in fixed cells (19). Subcellular localization of Spo73 protein was observed in live cells by using N-terminally GFP-tagged Spo73 expressed from a low-copy-number vector. Although the plasmid rescued the sporulation defect of the spo73⌬ mutant, no apparent localization pattern was observed (data not shown). Therefore, GFP-Spo73 was overexpressed during sporulation by using a multicopy vector. Upon overexpression, a clear prospore membrane pattern was observed (Fig. 4A). Colocalization with a prospore membrane marker confirmed that Spo73 localizes to prospore membranes (Fig. 4B).
Vps13 localization to prospore membranes is dependent on Spo71 (12). Thus, interdependency of localization among Spo73, Spo71, and Vps13 was analyzed. A  GFP-tagged version of each gene was expressed in mutants with the other two genes deleted and examined during sporulation. Spo73-GFP localized to prospore membranes in spo71⌬ and vps13⌬ mutant cells (Fig. 4C), indicating that Spo73 localization to the prospore membrane is not dependent on Spo71 or Vps13. Similarly, prospore membrane localization of Spo71-GFP was not disturbed in spo73⌬ and vps13⌬ mutant cells (Fig. 4C), indicating that Spo71 localization to the prospore membrane is not dependent on Spo73 and Vps13. As to Vps13-GFP, although some cytosolic background fluorescence was observed, a prospore membrane pattern was observed in spo73⌬ mutant cells (Fig. 4D). This indicates that localization of Vps13 to the prospore membrane is not dependent on Spo73, which is distinct from the dependence of Vps13 localization on Spo71.
Mutations in a surface patch of basic amino acid residues of Spo73 disrupt its localization to the prospore membrane. Our analysis revealed that Spo73 consists mostly of the dysferlin domain. The three-dimensional (3D) structure of Spo73 was modeled on the structure of the dysferlin domain of myoferlin (Protein Data Bank [PDB] code 2K2O A chain) (30) (Fig. 5A). A similar 3D model was obtained on the basis of the 3D structure of the dysferlin domain of dysferlin (PDB code 4CAH B chain; data not shown) (23). Besides the antiparallel ␤-sheet structure formed by the N-and C-terminal dysferlin domains, a surface basic loop was found in both dysferlin domains, which may be important for electrostatic interactions with acidic materials, such as protein surfaces or phospholipids. Thus, we mutated those amino acids in GFP-Spo73 to alanines (R119A, K121A, K123A) (Fig. 5B). The mutant version of GFP-Spo73 (GFP-Spo73AAA) was overexpressed in spo73⌬ mutant cells and observed during sporulation. GFP-Spo73AAA-expressing cells did not show the prospore membrane pattern; instead, they showed a uniform cytosolic pattern (Fig. 5C). However, this strain sporu- lated (61%) at a level only slightly lower than that of the spo73⌬ mutant strain overexpressing GFP-Spo73 (78%), suggesting that the overall structure of Spo73 is not disrupted by the mutations and that overexpression may be masking a more severe sporulation defect. Therefore, we expressed GFP-Spo73AAA from a centromere (CEN)based low-copy-number vector by using its own promoter. This strain sporulated to a significantly lesser extent (7%) than the strain expressing wild-type GFP-Spo73 from a low-copy-number vector (58%) (Fig. 5D). These results indicate that the basic loop of Spo73 is required for its localization to the prospore membrane and that spo73AAA is partially functional.
Overexpression of Spo71 can localize Spo73AAA to the prospore membrane. Taking advantage of spo73AAA, which is partially functional, we examined the relationship between Spo73 and Spo71 or Vps13. Suppression of spo73AAA was tested by using cells expressing spo73AAA from a CEN-based low-copy-number vector. The SPO71 and VPS13 genes were each overexpressed by using a 2m-based vector, and percent sporulation was examined. Also, the effect on the localization of Spo73AAA was tested by using cells overexpressing GFP-Spo73AAA and Spo71 or Vps13 from 2mbased vectors. We could not detect apparent suppression of spo73AAA by overexpression of SPO71 or VPS13 (Fig. 5E). However, in contrast to the cytosolic pattern observed in cells overexpressing only GFP-Spo73AAA or both GFP-Spo73AAA and Vps13, an apparent prospore membrane pattern was observed in cells overexpressing both GFP-Spo73AAA and Spo71 (Fig. 5F). Considering a previous report showing that Spo71-GFP localizes to the prospore membrane, which was detectable only when overexpressed (12), the present observation suggests that overexpression of Spo71 can recruit detectable levels of mutant GFP-Spo73 to the prospore membrane. The results also imply that the function of Spo73 in prospore membrane extension is not necessarily dependent on its localization to the prospore membrane.
SPO73 has a genetic interaction with SPO1 and SMA2. Parodi et al. showed that SPO71 has a weak genetic interaction with SPO1 (11). Also, although no data were shown, a genetic interaction between SPO73 and SPO1 has been described (31). Therefore, we looked for a genetic interaction of SPO73 with SPO1 and SMA2, which code for proteins involved in the spore membrane-bending pathway, together with SSP1, which codes for the essential component of the LEC. While all of the single mutants, i.e., the spo1⌬, sma2⌬, and spo73⌬ mutants, were completely sporulation defective, the spo1⌬ spo73⌬ and sma2⌬ spo73⌬ double mutants sporulated reproducibly (spo1⌬ spo73⌬ mutant, 9%; sma2⌬ spo73⌬ mutant, 5%) (Fig. 6A). No sporulation of the ssp1⌬ spo73⌬ double mutant was observed.
To assess the mechanism of suppression, the phenotype of the spo1⌬ spo73⌬ double mutant was analyzed in comparison with that of single mutants. As reported by Parodi et al. (11), Htb2-mCherry and GFP-Spo20 51-91 were expressed in the mutant cells during sporulation, and postmeiotic cells were classified into two classes. Cells without apparent prospore membranes were designated class I. Cells with at least one prospore membrane were designated class II. In addition, the capture of the nuclei by prospore membranes and prospore membrane perimeters was assessed. As is the case with the spo71⌬ mutant, the spo73⌬ mutant partially suppressed the defect of the spo1⌬ mutant. The spo1⌬ spo73⌬ double mutant exhibited more cells in class II, indicating a defect less severe than that of the spo1⌬ mutant (Fig. 6B). No apparent suppression of the nuclear capture rate was observed (Fig. 6C). As to prospore membrane size, the spo73⌬ mutant cells showed a small prospore membrane phenotype; however, apparent suppression was not observed in the spo1⌬ spo73⌬ double mutant (Fig. 6D). But we found some prospore membranes in the double mutant larger than those in the spo73⌬ mutant, consistent with the 9% sporulation of this mutant. Our results indicate that SPO73 genetically interacts with the genes involved in spore membrane bending, similar to SPO71.

DISCUSSION
Previous studies have shown the involvement of Vps13 and Spo71 in prospore membrane extension (9,11,12); however, the overall mechanism of the process has remained to be elucidated. Here, we identified Spo73 as an additional factor involved in this process.
The spo73⌬ mutant showed phenotypes similar to those of the spo71⌬ and vps13⌬ mutants; it formed small prospore membranes with intraluminal vesicles, was partially defective in nuclear capture by prospore membranes, and was also partially defective in the removal of the LEC. All three proteins localized to the prospore membrane. Analyses of interaction and localization dependency in a previous study showed that Spo71 interacts with and recruits Vps13 to the prospore membrane (12). Our analysis showed that localization of Spo73 to the prospore membrane is not dependent on Spo71 and Vps13 and vice versa. However, a mutant Spo73 protein that showed a cytosolic localization pattern was targeted to the prospore membrane when Spo71 was overexpressed. Although no genetic suppression was detected, this rescue of localization could be explained by interaction between Spo73 and Spo71. Further analysis is required to show a physical interaction between these proteins.
Membrane-bending pathway mutants form wide and aberrant prospore membranes (14), which are quite the opposite of the small prospore membranes formed by the spo73⌬ mutant. A genetic interaction was observed between SPO73 and the genes involved in spore membrane bending, SPO1 and SMA2. Our analysis of prospore membranes showed that the spo1⌬ mutant phenotype was partially suppressed by spo73⌬, although apparent suppression of the spo73⌬ mutant phenotype by spo1⌬ was not observed. SPO71 has a genetic interaction with SPO1 as well (11), although it is weaker than the interaction between SPO73 and SPO1 or SMA2. Thus, our results support the idea that Spo73 functions with Spo71 in prospore membrane extension and that this function somehow works to counteract the bending force exerted by the spore membrane-bending pathway.
In our results, the rate of prospore membrane capture in the mutants was somewhat lower than reported (11). This could be caused by a difference in the strain background or some experimental conditions. It has been reported that PI4P levels on prospore membranes of vps13⌬ and spo71⌬ mutant cells are lower than on those of wild-type cells (9,12). Thus, Vps13 and Spo71 are considered to have a role in regulating the PI4P levels on prospore membranes. This was based on the observation of a PI4P biosensor in cells that sporulated overnight at room temperature, which could contain cells at various stages during sporulation. To focus on cells just around meiosis, we changed the method to sporulating the cells at 30°C for 7 to 9 h. Using this new method, we could not detect any apparent difference in PI4P levels in spo71⌬ and spo73⌬ mutant cells from those in wild-type cells, suggesting that PI4P levels are not affected or are affected at undetectable levels in those mutants around meiosis II. The reasons for the discrepancy between the observations in this and previous studies are unclear, but one of the reasons could be that observing cells at later stages amplified the difference to detectable levels in the original procedure.
Bioinformatic analysis revealed that Spo73 is a dysferlin domain-only protein. Dysferlin domains are found in a subgroup of the ferlin family proteins (20). In humans, three out of the six ferlin family proteins contain dysferlin domains, including dysferlin and myoferlin. Dysferlin is a protein considered to function in Ca 2ϩ -dependent membrane repair in muscles, and mutations in its gene cause dysferlinopathies, including limb girdle muscular dystrophy type 2B and Miyoshi myopathy (21)(22)(23). Myoferlin is involved in myoblast fusion (32). In S. cerevisiae, dysferlin domains are found in a subset of peroxins, Pex30, Pex31, and Pex32, that function in the control of peroxisome size and number (33). All of the dysferlin domain-containing proteins analyzed so far are involved in membrane-related processes, although the precise role of this domain is still unknown. In this study, we revealed that dysferlin domain-only protein Spo73 is required for proper prospore membrane formation. Prospore membrane formation of S. cerevisiae is a process of dynamic reorganization of intracellular membrane structure and has been studied as a model of de novo membrane formation. Given that other dysferlin domain proteins contain specific domains in addition to this domain (28), dysferlin domain-only protein Spo73 is the most suitable protein to elucidate the conserved function of dysferlin domains.
Our analysis revealed that a mutant Spo73 protein that has mutations in a conserved basic loop cannot localize to the prospore membrane. Thus, Spo73 may interact with acidic materials such as phospholipids on the prospore membrane through its conserved basic loop. Although more analysis is required, interaction with membranes could be an important feature of the dysferlin domain. Interestingly, one of the mutations causing muscular dystrophy resides in the basic loop of the inner dysferlin domain of the dysferlin protein (R1022Q) (23), suggesting that this mutation could affect the interaction of the domain with the membrane, which could cause the disease phenotype. Further analysis of Spo73 could provide insights into the conserved function of dysferlin domains and the pathogenesis of dysferlinopathy.

MATERIALS AND METHODS
Yeast strains and media. Standard media and genetic techniques were used unless otherwise noted (34). The genes corresponding to the mutants selected for screening from the deletion collection in the BY4743 background (Invitrogen, San Diego, CA) are listed in Table S1 in the supplemental material. Other yeast strains used in this study are listed in Table 1. All of the strains in Table 1 were derived from the SK1 background. PCR-based gene alterations were performed as previously described (35). Strains were constructed with the primers and plasmids in Table S2. The primers used in this study are listed in Table S3. All PCR-based integrations and disruptions were confirmed by genomic PCR. All of the deletions constructed in this study were confirmed to be rescuable by expression of the genes deleted.
Plasmids. The plasmids used in this study are listed in Table S4 in the supplemental material. To generate 306-DON1-GFP, DON1-GFP was cut out of pSB8 (35) and cloned into pRS306. To construct pSB1 (pFA6a-His3MX6-P SPO20 -GFP), the GAL1 promoter of pFA6a-His3MX6-PGAL1-GFP (36) was replaced with the SPO20 promoter (35). To generate 424-GFP-SPO73, the promoter region of a chromosomal copy of SPO73 was replaced with a P SPO20 -GFP fusion by a PCR-based method with primers OK13 and OK14 by using pSB1 as a template. P SPO20 -GFP-SPO73 was then amplified from genomic DNA by using HT369 and OK15 and cloned into pRS424. To construct 424-P SPO20 -GFP-SPO73AAA, site-directed mutagenesis was performed by the QuikChange method with primers OK29 and OK30. To construct 314-P SPO73 -GFP-SPO73 and 314-P SPO73 -GFP-SPO73AAA, GFP-SPO73 and GFP-SPO73AAA without a promoter were cloned into pRS314 together with a SPO73 promoter fragment amplified with HT421 and HT422. To construct 426-SPO71 and 426-VPS13, SPO71 and VPS13 were amplified from SK1 genomic DNA by PCR with primers OK11 and HT424 and primers TN194 and TN197, respectively, and cloned into pRS426. To generate 316-HTB2-mCherry, a chromosomal copy of HTB2 was fused to mCherry by PCR-mediated transformation. HTB2-mCherry was then amplified from genomic DNA with primers TN62 and HT66E and then cloned into pRS316. All amplified regions were sequenced.
Sporulation. Sporulation was performed as previously described (37). For sporulation on plates, cells grown on yeast extract-peptone-dextrose or synthetic dextrose medium plates were shifted to sporulation plates (1% potassium acetate) and incubated at 30°C.
Nuclear capture counting and prospore membrane perimeter measurement were performed as previously described (11). Cells expressing both HTB2-mCherry and GFP-SPO20 51-91 were sporulated for 9 h, and postmeiotic cells were identified by the pattern of Htb2-mCherry localization. The percentage of prospore membranes capturing nuclei is shown for each strain. Prospore membrane perimeters were measured by ImageJ (http://rsb.info.nih.gov/ij/). PI4P levels at the prospore membrane were tested as previously described (12), except that sporulation was induced for 7 to 9 h at 30°C. Cells carrying both pRS426-P PRC1 -GFP-PH OSH2 and pRS424-DTR1-RFP were observed under the microscope, and GFP-PH OSH2 /Dtr1-mRFP colocalization was quantified.
Time-lapse imaging was performed as previously described (38). Images were captured on a Zeiss Axiovert 100 microscope equipped with a CoolSNAP HQ camera (Photometrics) at 2-min intervals with IPLab 3.6.5a software (Scanalytics). The temperature was held at 28°C during image collection. 3D stacks were performed with IPLab 3.6.5a. This study EM. Sporulating cells were collected, fixed with 2% paraformaldehyde-2% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) at 4°C overnight, and washed three times with 0.1 M phosphate buffer. Cells were then treated with 4% KMnO 4 at room temperature for an hour, washed with distilled water, and embedded in 1.5% low-melting-point agarose. Samples were dehydrated with acetone and infiltrated with increasing concentrations of Spurr's resin in propylene oxide and finally with 100% Spurr's resin. After polymerization with dimethylaminoethanol, ultrathin sections were cut on a Leica Ultracut UCT microtome and stained with uranyl acetate and lead citrate. The sections were then viewed with a JEOL 2000EX electron microscope.