SPO73 and SPO71 Function Cooperatively in Prospore Membrane Elongation During Sporulation in Saccharomyces cerevisiae

In the yeast Saccharomyces cerevisiae, cells undergoing sporulation form prospore membranes to surround their meiotic nuclei. The prospore membranes ultimately become the plasma membranes of the new cells. The putative phospholipase Spo1 and the tandem Pleckstrin Homology domain protein Spo71 have previously been shown to be required for prospore membrane development, along with the constitutively expressed Vps13 involved in vacuolar sorting. Here, we utilize genetic analysis, and find that SPO73 is required for proper prospore membrane shape and, like SPO71, is necessary for prospore membrane elongation. Additionally, similar to SPO71, loss of SPO73 partially suppresses spo1Δ. Spo73 localizes to prospore membranes and complexes with Spo71. We also find that phosphatidylserine localizes to the prospore membrane. Our results suggest a model where SPO71 and SPO73 act in opposition to SPO1 to form and elongate prospore membranes, while VPS13 plays a distinct role in prospore membrane development.


Introduction
In the budding yeast Saccharomyces cerevisiae, diploid cells can respond to starvation by triggering the developmental process of sporulation. During this process, the cell undergoes significant changes to both its ploidy and its cellular architecture [1][2][3]. The ability of the cell to properly complete sporulation is coupled to formation of prospore membranes, which become the plasma membranes for the future spores [4,5]. Prospore membrane growth is coordinated by multiple mechanisms to specify precise dynamic shapes that ensure encapsulation of genomic and cytosolic content, and also forms the template for spore wall morphogenesis [6][7][8]. Prospore membrane morphogenesis occurs in four stages: (i) initiation, (ii) elongation, (iii) rounding, and (iv) closure. While many genes have been linked to prospore membrane development, how these genes act together to regulate membrane morphogenesis is not fully understood.
Similar to VPS13 and SPO71, the putative phospholipase-A2/B encoded by SPO1 is upregulated during sporulation and required for its completion [11,12,17]. However, spo1Δ alleles exhibit a prospore membrane defect distinct from either vps13Δ or spo71Δ, as the loss of SPO1 reduces the overall formation of prospore membranes, as assayed by a prospore membrane reporter, Spo20 51-91 [16]. Many spo1Δ cells fail to form prospore membranes, and instead exhibit clustering of phosphatidic acid throughout the cytoplasm. Those spo1Δ cells that do form prospore membranes have prospore membranes that grow inappropriately straight, producing a grossly elongated morphology [13,17]. SPO1 has been proposed to be part of the sporulation membrane bending pathway, which acts to provide an inward bending force upon the prospore membrane [17]. Another gene required for sporulation, SPO73, encodes a Dysferlin domain protein [18,19]. SPO73 had previously been investigated for its role in prospore membrane development but was reported to be dispensable for proper prospore membrane shape [19].
We sought to determine how SPO73 contributed to sporulation and found that, contrary to earlier reports, it is required for the proper formation of prospore membranes. We find that the requirement for SPO73 in prospore membrane development is similar to that of SPO71, with both genes being required for prospore membrane elongation. Furthermore, we find that SPO71 and SPO73 act downstream of the early-acting prospore membrane gene VPS13, and that SPO73, SPO71, and SPO1 have a complex genetic relationship necessary for sporulation.

Strains used in this study
All strains used in this study are derivatives of the highly efficient sporulating SK1 strain [20], and are listed in S1 Table. Gene knockouts were created using standard yeast genetic techniques [21]. spo73Δ isolates were constructed by replacing the SPO73 open reading frame in a wild type MATa strain, LH175 with either the HIS3 C.g. gene amplified from pCgHIS (which contains the Candida glabrata HIS3 gene) [16] to produce spo73::HIS3 C.g. , or the LEU2 gene amplified from pCgLEU2 (which contains the Candida glabrata LEU2 gene) [22] to produce spo73::LEU2 C.g. . SPO73-ENVY was constructed by inserting the GFP variant, Envy, immediately before the stop codon of SPO73 using PCR mediated recombination from PCR products amplified from pFA6a-link-Envy-SpHis5 [23]. Transformants were confirmed for proper tagging/gene replacement using PCR, and subsequently backcrossed to a MATα strain. MATa and MATα segregants were verified using both auxotrophic marker identification and PCR genotyping and then mated to produce homozygous diploid strains, The vps13Δ homozygous diploid was generated by crossing HI28 [14] with the MATα strains LH899 and LH176, sporulating the heterozygotes, and dissecting haploids to generate haploids containing the vps13:: his5+ allele with and without the HTB2-mCherry alleles. As with all strains, segregants were verified using auxtrophic marker identification and subsequent PCR confirmation. from Addgene) [25] was used to detect phosphatidylserine localization. GFP-Spo14 [26] was used for Spo14 localization.

Sporulation
Sporulation was performed as described previously [16]. For all assays involving sporulation efficiency, meiotic kinetics were monitored, and counts were only included for cultures that were undergoing sporulation efficiently, as assayed by having at least 50% of the cells entering meiosis by 8 hours post-sporulation induction. Meiosis was monitored by counting cells that had 1 nucleus, 2 nuclei, or > 2 nuclei, using either the fluorescently tagged Htb2 protein or DAPI staining.

Fluorescence microscopy
All strains were imaged at 100x magnification through a 1.45 N.A. with the Axioskop Mot2 widefield microscope (Zeiss). Images were collected using an Orca-ER CCD camera (Hamamatsu) and Openlab 4.04 (Perkin Elmer) software. Image processing was performed using Ima-geJ1.46r (NIH), [27]. For prospore membrane analysis, multiple z-slices were summed to visualize all prospore membranes in each cell. Fluorescent images with were de-convolved using the Iterative Deconvolve plugin for ImageJ [28]. Prospore membrane measurements and statistical comparisons were performed as previously described [16].

Phenotypic Assignment and Statistical Analysis
A minimum of three independent sporulations from each isolate was performed for all quantified phenotypes described. For comparison between genotypes, statistical comparisons were assessed using one-way analysis of variance and subsequent Tukey multiple comparison tests (GraphPad).

Immunoprecipitation
Samples for immunoprecipitation were prepared from 60 OD 600 of cells. Cell pellets were lysed in a MiniBeadBeater8 (Biospec) at 4°C with glass beads in IP buffer (150 mM KCl, 1 mM EDTA, 50 mM HEPES pH 7.6, 1 mM DTT and 0.5% Nonidet P-40) containing a mixture of protease and phosphatase inhibitors as previously described [30]. Lysates were clarified three times at 17,000 rcf at 4°C. Clarified lysates were mixed with 40 μl blocked agarose beads (Bulldog Bio), incubated at 4°C for 30 minutes. Beads were removed from the lysates by centrifugation at 500 rcf, and lysates were mixed with 20 μl of GFP-Trap A beads (Bulldog Bio) and incubated 2 hours at 4°C. GFP-Trap complexes were washed three times in IP buffer and re-suspended in 2x SDS-PAGE sample buffer, boiled for 5 minutes, separated by SDS-PAGE, and detected for proteins as described above.

Results
SPO73 is required for prospore membrane elongation As SPO73 resembled a regulator of the prospore membrane morphogenesis SPO71 both in its expression profile and its dispensability for meiosis [11,12,18,19], we examined prospore membranes in spo73Δ cells. Prospore membrane analysis using the prospore membrane marker G20 (GFP-Spo20 51-91 ) [13], shows that SPO73 is required for the formation of properly sized prospore membranes. spo73Δ cells make prospore membranes that are terminally smaller than those made in wild-type cells (Fig 1A). We see this similar phenotype using two independently constructed alleles of spo73 (spo73::HIS3 C.g. and spo73::LEU2 C.g. ). The mean (+/-SEM) prospore membrane perimeter sizes for the spo73::HIS3 C.g. allele (5.11+/-0.08 μm) was not statistically distinct from that of the spo73::LEU2 C.g. allele, (5.46+/-0.15 μm; p-value~0.2). Intriguingly, the spo73Δ prospore membrane phenotype is similar to that seen in cells lacking SPO71 (Fig 1B). Like spo71Δ cells, spo73Δ cells fail to exhibit the characteristic elongated tube phase of prospore membrane development. The elongated prospore membranes seen in wildtype cells were not seen in either single mutant or the double mutant.
To test whether SPO73 and SPO71 may act in the same or parallel pathways, we assessed the prospore membrane shape and size in a spo71Δ spo73Δ double mutant. We found that loss of both genes resulted in prospore membranes that are equivalent in size and shape as compared to those formed in either single mutant (Fig 1). The prospore membrane perimeters among mutants are not statistically different from each other, while the spo71Δ and spo73Δ single mutants and the spo71Δ spo73Δ double mutants are all statistically distinct from wild type (Tukey HSD, alpha = 0.01). This similar and not additive phenotype is consistent with SPO73 and SPO71 functioning within the same pathway for prospore membrane elongation.

Spo73 localizes to prospore membranes and complexes with Spo71
To assess SPO73 localization in living cells, we tagged the endogenous SPO73 at its C-terminus using the optimized GFP, Envy [23]. Strains homozygous for the SPO73-ENVY allele sporulate at wild-type levels, confirming that the tagged allele is functional. Using this tagged allele, we see Spo73 colocalizing with the prospore membrane localized Dtr1 [32] (Fig 2A). Thus, Spo73 localizes to the prospore membrane, consistent with a role in its development.
Western blotting indicated that the Spo73-Envy protein was produced during sporulation, with levels peaking when the majority of cells have completed meiosis (Fig 2B). Because our genetic analysis suggests that SPO71 and SPO73 function together, we tested whether the two proteins encoded by these genes physically interact. Immunoprecipitation of Spo73-Envy coimmunoprecipitates Spo71-myc ( Fig 2C). As Spo71 was reported as localizing to the prospore membrane [15], this coimmunoprecipitation result is consistent with both proteins localizing to prospore membranes.

SPO71, SPO73 and SPO1 have a complex genetic relationship
Prior work has found that SPO71 acts antagonistically to the putative phospholipase SPO1 during prospore membrane development [16]. As SPO71 and SPO73 have similar effects on prospore membrane size, we examined whether SPO73 might also antagonize SPO1. Previously, we have seen that loss of SPO1 results in a marked decrease in the formation of prospore membranes, with most cells exhibiting a clustering of prospore membrane materials throughout their cytoplasm [16]. We now find that, that similar to SPO71, the loss of SPO73 also improves prospore membrane development in cells lacking SPO1 (Fig 3A). Thus, SPO73 and SPO71 have a similar antagonistic relationship to SPO1.
In the simplest scenario, one might anticipate that the simultaneous removal of both SPO73 and SPO71 activity would cause a similar or an even greater suppression of the spo1Δ defect. Surprisingly, the spo1Δspo71Δspo73Δ triple mutant exhibits a prospore membrane formation defect most similar to the spo1Δ single mutant (Fig 3A). Thus SPO71 and SPO73 do not appear to form a simple linear pathway.
To further confirm these findings, we sought a separate quantitative measure of spore formation. While most spo1Δ cells are unable to form spores [33,34], we do see that spo1Δ cells produce a low-level (1.67%) of refractile spore-like structures twenty-four to thirty-six hours following induction of sporulation, although they do not typically produce four spores within the ascus (Fig 3B). Refractile structures appear during spore development as the outer prospore membrane bilayer is removed and the outer chitosan and dityrosine layers are being deposited [19]. While it is important to note that the appearance of refractile structures does not necessarily mean the completion of spore development, it provided a tool for further confirming the genetic relationships among SPO1, SPO71, and SPO73. Visual analysis of single mutants revealed that like spo1Δ, loss of SPO73 also corresponded to a low-level (0.54%) appearance of refractile spore-like structures, while loss of SPO71 never resulted in the appearance of such structures (Fig 3B). All refractile spore-like structures we counted were observed to surround a nucleus, as would be expected for a maturing spore. When we examine spo71Δspo73Δ double mutant, no refractile spore-like structures were observed, similar to the spo71Δ mutant.
Consistent with the antagonistic relationship between SPO1 and either SPO71 or SPO73 that we observed by analyzing prospore membrane development, loss of SPO73 or SPO71 also suppresses spo1Δ for the refractile spore phenotype. spo1Δspo71Δ (13.82%) and spo1Δspo73Δ (30.86%) mutants exhibited a significant increase in the appearance of refractile spore-like structures compared to any single mutant. Furthermore, the triple mutant, spo1Δ spo71Δ spo73Δ formed few refractile structures (1.24%), similar to the spo1Δ single mutant. This triple mutant phenotype is consistent with what we see when we examine the prospore membrane, . Pgk1 protein abundance serves as a loading control. Ndt80 expression coincides with expression of mid-sporulation genes, and protein abundance serves as a monitor for progression through sporulation. Meiotic counts were conducted from the same cultures used for protein isolation; 200 cells from each time point were counted. Nuclei were visualized using HTB2-mCherry. % 2 nuclei represents cells entering meiosis and may include cells that will sporulate as dyads, %>2 nuclei represents cells who have progressed into MII. (D) Spo71-myc coimmunoprecipitates with Spo73-Envy. Sporulated cells from wild-type (LH177), Spo71-myc (LH901), Spo73-Envy (LH938), and Spo71-myc+Spo73-Envy (LH1039) strains were immunoprecipitated with GFP and probed for Spo713 using and anti-GFP antibody and for Spo71 using an anti-myc antibody. where we see that the spo1Δspo71Δspo73Δ mutant produces a prospore membrane phenotype most similar to spo1Δ mutant (Fig 3A). These data confirm that while the loss of either SPO71 or SPO73 can compensate to some extent for the loss of SPO1, loss of both SPO71 and SPO73 eliminates this effect.
Phosphatidylserine and Spo14 localization is not altered in spo73, spo71, or spo1 mutants Previous studies suggested that the prospore membrane marker GFP-Spo20 51-91 may bind to phosphatidic acid [35]. However, recent work has demonstrated that this construct can interact with multiple anionic lipids [36]. Because we had assayed prospore membranes using GFP-Spo20 51-91 , it is possible that the prospore membrane phenotypes we see may be due to mislocalization of one or multiple GFP-Spo20 51-91 interacting anionic lipids, rather than a complete loss in formation of prospore membranes. Thus, we assayed the localization of phosphatidylserine, an anionic phospholipid found in eukaryotic plasma membranes. While phosphatidylserine is present in a polarized fashion at the bud neck and bud cortex of vegetatively growing yeast cells [37,38], its localization during sporulation has not been previously reported. Using the phosphatidylserine reporter Lact-C2-GFP [25], we found that phosphatidylserine localizes to the developing prospore membrane (Fig 4A). We also see that phosphatidylserine localizes to the aberrant prospore membranes in the spo1Δ, spo71Δ, and spo73Δ mutants. These results suggest that the prospore membrane defect we see using the phosphatidic acid sensor does reflect a prospore membrane defect and not just a problem with the localization of certain anionic lipids.
To further assess whether the observed defects reflect a membrane morphogenesis rather than lipid biogenesis defect, we examine the localization of Spo14, a phospholipase-D that catalyzes the production of phosphatidic acid from phosphatidylcholine, and localizes to the prospore membrane [26,39]. We see Spo14 localization as expected, localizing in developmentally appropriate patterns along the prospore membrane (Fig 4B). Taken together, the prospore membrane marker GFP-Spo20 51-91 , the phosphatidylserine marker Lact-C2-GFP, and GFP-Spo14 all show that spo73 mutants have a prospore membrane defect.
vps13Δ mutants have a more severe prospore membrane defect compared to spo71Δ and spo73Δ mutants The phenotypes of spo71Δ and spo73Δ mutants suggest that SPO71 and SPO73 act during the elongation phase of prospore membrane development. VPS13 has previously been shown to be required for the formation of prospore membranes of appropriate size [13], and has been proposed to act as a partner for SPO71 [15]. We find that the prospore membrane defect of vps13Δ cells is more severe than that observed in spo71Δ or spo73Δ cells (Fig 5), with a comparison of perimeter sizes among mutants revealing that vps13Δ mutant perimeter sizes are statistically distinct (Tukey HSD, alpha = 0.01). vps13Δ mutants produce tiny prospore membranes that are smaller than those seen in spo73 and spo71. Prospore membranes in vps13Δ mutants lacking either or both SPO73 and SPO71 exhibit tiny prospore membranes, like those observed in vps13Δ single mutants (Fig 5). Although Vps13 has been suggested to perform at least one of its functions with Spo71 [15], these epistasis results suggest that Vps13 has a role in prospore membrane development that is distinct from Spo71 and Spo73.

Discussion
Taken together, our data show that SPO73, an essential regulator of sporulation whose role in the process was previously unclear, acts to promote prospore membrane elongation. The effect of SPO73 on prospore membrane elongation resembles those previously reported for the Localization of phosphatidylserine (PS) in wild-type, spo71Δ, spo73Δ and spo1Δ cells. PS was detected using Lact-C2-GFP-p416 [3]. Both elongated and rounded prospore membrane stages are shown for wildtype cells; elongated prospore membranes are not typically seen in spo71Δ and spo73Δ cells (B) Localization of Spo14 in wild-type, spo71Δ, spo73Δ and spo1Δ cells. Spo14 was detected using GFP-Spo14 [26]. For spo1Δ, Class I represents cells without prospore membranes; Class II cells represents cells that make at least one prospore membrane, like in Pleckstrin Homology domain protein SPO71 [16]. We further find that Spo73 localizes to prospore membranes, similar to the reported localization of Spo71 at elongated and rounded prospore membranes [14]. Our localization of Spo73 differs from the previously reported cytoplasmic and punctate localization [19]; however, those previous studies determined Spo73 localization in fixed cells using immunofluorescence techniques which are more prone to artifacts and complicate stage-specific localization. The interaction of SPO71 and SPO73 is also seen biochemically, as Spo71 and Spo73 can be coimmunoprecipitated from sporulating cells. Our data also show that while SPO73, like SPO71, acts antagonistically to the putative phospholipase SPO1 during spore morphogenesis, removing both SPO73 and SPO71 eliminates the effect, suggesting that perhaps the balance of SPO73 and SPO71 activity is important. Finally, we our results suggest that the vacuolar protein VPS13 exerts a broader impact on the process of sporulation than either SPO73 or SPO71, suggesting an additional, different role for VPS13 from SPO73 and SPO71.
SPO73 and SPO71 function during prospore membrane elongation, distinct from VPS13 The dynamic shapes of the developing prospore membrane provide important clues to determine the stage(s) in which regulators of prospore membrane morphogenesis function. This work suggests that SPO73 and SPO71 act together to promote prospore membrane elongation, as the spo71Δ and spo73Δ mutants show defects in prospore membrane elongation, as the spo71Δspo71Δ double mutant phenotype is no worse than either single mutant, and as Spo73 and Spo71 physically interact, as assayed by coimmunoprecipitation. Interestingly, the spo71Δ phenotype is more severe than that seen for spo73Δ, as assayed by the appearance of refractile spore-like structures, suggesting that SPO71 plays a greater role than SPO73.
Our results suggest that SPO71 and SPO73 play a distinct role in prospore membrane development from VPS13. The extreme reduction in prospore membrane size exhibited by vps13Δ cells seen by both others [14] and ourselves suggests that loss of VPS13 causes prospore membranes to prematurely arrest during development. These results are consistent with VPS13 acts during an initial growth phase of the prospore membrane, when the membrane starts to grow from the meiotic outer plaque and when it is crucial for the cell to determine the direction of membrane growth relative to the nucleus. From our results, it is possible that SPO71 and SPO73 act as negative regulators of VPS13, that SPO71 and SPO73 act downstream of VPS13, or, merely that SPO71 and SPO73 work independently of VPS13. Further experiments are needed to clarify the relationship between these genes. Opposing pathways promote sporulation Our study demonstrates that SPO73 and SPO71 exhibit similar genetic relationships with SPO1 during sporulation. We see this relationship using two assays: by the examination of prospore membrane sensors during prospore membrane morphogenesis, and by examining the formation of refractile structures following sporulation. We had previously demonstrated that clustering of prospore membrane materials that occurs in the majority of spo1Δ cells is suppressed by loss of SPO71. Here, we observe a similar phenomenon, where spo73Δ acts as a suppressor of spo1Δ, promoting the formation of prospore membranes and reducing inappropriate phosphatidic acid clustering. When we assay the formation of refractile spore-like structures, we find that both spo1Δspo71Δ and spo1Δspo73Δ mutants exhibited a significant increase in the appearance of refractile structures compared to all of the single mutants. Furthermore, the refractile structures seen in the double mutants appear much closer to wild-type spores, as compared to the structures seen in either the spo1Δ or spo73Δ mutant, further consistent with their antagonistic effects on SPO1.
Our analysis of the spo1Δspo71Δspo73Δ triple mutant yielded a surprising result, where the triple mutant was more defective than the spo1Δspo71Δ and spo1Δspo73Δ double mutants. The triple mutant exhibited a prospore membrane formation defect most similar to the spo1Δ single mutant. One possible explanation for this surprising result is that spo71Δspo73Δ is more defective than either single mutant in some fashion. As SPO71 activity is present in the spo1Δspo73Δ double mutant and SPO73 activity is present in the spo1Δspo71Δ double mutant, these activities work to produce prospore membranes. However, in the spo1Δspo71Δspo73Δ triple mutant, this balance of defects is destroyed, creating a more severe defect.
Supporting Information S1 Table. S. cerevisiae strains used in this study. (PDF)