Strains and approaches for genetic crosses in the oleaginous yeast Lipomyces starkeyi

Abstract The oleaginous yeast Lipomyces starkeyi is a powerful lipid producer with great industrial potential. Recent studies have reported the isolation of mutant L. starkeyi cells with higher lipid producing capacity. Although genetic engineering strategies have been applied to L. starkeyi, classical genetic approaches are lacking. The development of tools that facilitate genetic crosses in L. starkeyi would not only make it possible to build improved lipid‐producing strains but also facilitate molecular biological analysis of this species. In this study, I report a set of strains and approaches useful for performing genetic crosses with L. starkeyi. The homothallic L. starkeyi reportedly forms an ascus containing two to 20 spores. These spores were resistant to glusulase and could be dissected using a micromanipulator, suggesting that random spore and tetrad (spore dissection) analysis can be adapted for L. starkeyi. Additionally, to isolate a pair of heterothallic strains useful for genetic crosses, the homothallic strain was exposed to UV irradiation, and 10 self‐sterile strains were crossed with one another. One of these combinations, Ls75 and Ls100, sporulated stably. Moreover, to detect genetic recombination, I introduced a different drug resistance marker into each strain and crossed them. The resulting progeny exhibited Mendelian segregation of the resistance markers. Altogether, the work reported here provides a powerful resource for genetic analysis in L. starkeyi.

of a single meiotic event, such that genetic analysis of the spores ('tetrad analysis') can reveal linkage relationships among genes. Tetrad analysis is also necessary for constructing strains useful for other genetic and biological studies. One reason similar methods have not been applied in L. starkeyi is that sporulation in this species is peculiar.
Like other yeasts, L. starkeyi form asci; however, each ascus can contain from two to 20 spores (Kurtzman, 2011). Moreover, the frequency of sporulation of L. starkeyi is extremely low. Indeed, plating on commonly used sporulation media (V8, YM, or 1/10 YM) induces sporulation at a frequency of only 2%, and the frequency on L. starkeyi-specific sporulation medium (AF medium) is 14.6% (Watanabe et al., 1997).
Another barrier to genetic analysis in L. starkeyi is that available strains behave as homothallic strains, whereas heterothallic strains are preferable for genetic strategies. In Saccharomycotina species, cell type is determined by the presence of MATa or MATα alleles at the mating type (MAT) locus. The stability of the mating type classifies yeast strains as members of one of two groups, heterothallic or homothallic. Heterothallic strains have a stable mating type and only mate with strains of different mating types, and therefore behave as selfsterile. In contrast, homothallic strains can exist in either of two different forms, a primary and a secondary form (Hanson & Wolfe, 2017;Riley et al., 2016). Secondary homothallism by mating type switching has been well studied in S. cerevisiae and Sz. pombe. Additionally, methylotrophic yeasts, such as Ogataea polymorpha, Ogataea minuta, and Komagataella phaffii (Hanson et al., 2014;Maekawa & Kaneko, 2014;Yoko et al., 2019), contain one copy each of MATa and MATα loci, flanked by a pair of inverted repeat (IR) sequences. One of the MAT loci is proximal to the heterochromatic region of the genome (centromere or telomere), and its transcription is selectively repressed. Mating type is switched by invasion of the chromosomal region between the two MAT loci, swapping the positions of the active and repressed MAT loci. L. starkeyi is classified as a primary homothallic species.
According to previous reports (Hanson & Wolfe, 2017;Krassowski et al., 2019;Riley et al., 2016), the L. starkeyi genome harbors two MATa genes (a1 and a2) and two MATα genes (α1 and α 2) but lacks DNA repeats near these genes, which are located close together on the same chromosome. The functions of the MAT genes and pheromone-receptor interactions of L. starkeyi have not been elucidated.
L. starkeyi has the ability to form spores and its genome sequence suggests that it is homothallic (McNeil & Stuart, 2018;Riley et al., 2016). Heterothallic strains are able to mate with opposite mating types, thus allowing for genetic crossbreeding. However, heterothallic strains of L. starkeyi have not been reported. In this study, I isolated strains of L. starkeyi useful for genetic analysis. The strains Ls75 and Ls100 were isolated from cells that survived exposure to UV irradiation. When Ls75 and Ls100 were crossed, sporulation occurred practically and stably. Moreover, the progeny of a cross between these strains exhibited Mendelian segregation. The crossable strains of L. starkeyi newly generated in this work will be useful for genetic analysis and may provide insights into non-conventional yeasts that form multi-spores.

| Spore dissection
Sporulated cells were scraped and suspended in 20 μl of sterile water.
The suspended cells were streaked on YPD plates. Spores on YPD plates were dissected using a micromanipulator (SporePlay+, Singer).
Enzyme treatment (e.g., zymolyase) was not required, because the spores are released from the mature ascus (Kurtzman, 2011).

| UV irradiation
The cells were spread on YPD plates, and the plates were exposed to a dose of 0 to 100 J/m 2 of UV light (FUNA UV Crosslinker). After UV irradiation, the plates were incubated at 26 C. UV sensitivity was normalized by defining the number of colonies obtained without irradiation as 100%. Since the cell viability decreased when the UV dose was high, the numbers of spread cells were 300 and 3000 for 0 and 25 J/m 2 , and were 3000 and 30,000 cells for 75 and 100 J/m 2 , respectively.

| BODIPY staining and microscopy analysis
S medium culture cells were collected by centrifugation (200Âg, 3 min, 25 C) and then washed with PBS. The cell pellets were suspended in 25 μl of 0.5 μg/ml BODIPY™ 493/503 (Invitrogen, USA) and incubated for 10 min at room temperature. After washing with PBS, the cells suspended in PBS were observed using a TCS SP8 confocal microscope equipped with a 63Â objective lens (N.A. 1.40) (Leica Microsystems GmbH).
The reported sporulation frequency of L. starkeyi CBS1807 is extremely low; YM and 1/10 YM induced sporulation at a rate of 2% and AF induced it at a rate of 14.6% (Watanabe et al., 1997). Disappointingly, there have been few reports on L. starkeyi sporulation. Identification of strains and/or conditions under which spore formation by L. starkeyi is stable would facilitate genetic studies. To accomplish this, I first confirmed whether the homothallic strain L. starkeyi NBRC1289 (also called CBS1807) could be induced to sporulate. L. starkeyi NBRC1289 cells grown on YPD plates were streaked on YM, 1/10 YM, or AF solid media and incubated at 26 C. After 1 month, sporulation was not observed on YM medium; however, I did observe a few sporulation events on 1/10 YM (3.0 ± 2.9%). The overall frequency of sporulation on AF medium was 17.3 ± 2.5% at 2 weeks after plating and did not increase much after further incubation, presumably as most spores had already been released by mature asci (Kurtzman, 2011). Consistent with a previous report (Watanabe et al., 1997), AF medium efficiently induced L. starkeyi sporulation. Thus, I primarily used this medium in subsequent experiments.
Unlike the asci of other fungi, which produce four or eight spores, the L. starkeyi asci contain 2-20 spores ( Figure 1a, upper panel) (Kurtzman, 2011). This multi-spore appearance was the same as that observed for teliospores of Rhodotorula species (Newell & Hunter, 1970).

| Spore analysis of L. starkeyi
In yeast species, spore dissection (tetrad analysis) is useful not only for linkage studies, but also for building strains with multiple mutations. However, whether spores in the multi-spore asci of L. starkeyi could be dissected was not known. To ascertain this, I induced L. starkeyi NBRC1289 cells to form spores on AF solid medium, then dissected ascospores onto YPD plates using a micromanipulator. I chose asci containing three to six spores as it was difficult to completely separate spores in asci with 10 or more spores. Most separated spores were viable (Figure 2a), and these segregants sporulated when cells were incubated on AF solid medium, indicating that the segregants maintained homothallism.
Spore dissection is a classical method for genetic analysis; however, this method requires use of a specialized instrument and is not appropriate when handling an ascus with many spores. By contrast, random spore analysis is a convenient and high-throughput method.
To be appropriate for random spore analysis, spore cells must be resistant to treatment with glusulase, an enzyme prepared from snail intestinal juices that can digest the ascus wall and kill vegetative cells. To test the suitability of this approach, I ascertained whether spores of L. starkeyi were resistant to glusulase. A 5-mm diameter colony grown on AF solid medium was scraped and treated with glusulase. The treated cells were able to grow on YPD plates (Figure 2b, spore). In contrast, for a colony scraped from a YPD plate, glusulase-treated cells did not grow (Figure 2b, vegetative). These results indicate that L. starkeyi spores are resistant to glusulase. Altogether, these results show that both spore dissection and random spore analysis can be performed with L. starkeyi.

| Isolation of crossable strains
Heterothallic strains mate with cells of an opposite mating type. These strains are very useful for genetic studies with different genotypes. To obtain crossable strains, I decided to select self-sterile cells following exposure to UV irradiation. To do this, I first determined the UV sensitivity of the L. starkeyi NBRC1289 strain. The cells were spread on YPD plates and exposed to UV irradiation. The UV sensitivity of L. starkeyi was the same as that for S. cerevisiae OC-2 ( Figure 3a).
Lipid-producing mutants have been isolated following optimal UV irradiation resulting in about 5% cell viability (Tapia et al., 2012). Therefore, I next irradiated cells with 75 or 100 J/m 2 , which resulted in a cell viability of 6.8 ± 1.8% or 1.2 ± 1.2%, respectively. The 111 survivor cells were then streaked on AF medium. The well-sporulated L. starkeyi colonies changed to brown (Figure 3b). S. cerevisiae did not sporulate on AF medium. To select heterothallic cells unable to mate and sporulate with themselves, whitish colonies were picked up, and a lack of spore formation was confirmed by microscopy. Ten strains showed sterility or extremely low sporulation efficiency, and I made all possible crosses among these 10 strains. If heterothallic-like strains were included, then some combinations would form spores. Consistent with this idea, some of the 10 strains were fertile when crossed with others. One of these combinations, Ls75 and Ls100, was able to stably induce sporulation on AF medium but not on water plates. I did not detect Ls75 spores alone by microscopy; however, glusulase-resistant cells grew. Therefore, Ls75 alone did sporulate, albeit at a low frequency. By contrast, for Ls100, I did not detect the formation of glusulase-resistant cells, suggesting that it never sporulates. Notably, however, when Ls75 and Ls100 were mixed, they reproducibly formed spores on AF medium (Figure 3c). These results suggest that Ls75 and Ls100 behave as heterothallic strains.

| L. starkeyi crossable strains maintain lipid production capability
Ls75 and Ls100 strains were isolated after UV irradiation, which is known to induce mutations that may affect growth and lipid production. To define the growth phenotypes of these crossable strains, After treatment with glusulase, spores were grown on YPD plates containing hygromycin or zeocin to select for drug-resistant progeny, which were subsequently streaked on AF medium to test for self-sterility. After crossing Ls100 with the hygromycin-and zeocinresistant strain, I identified a strain 'Ls100H' that was resistant to hygromycin but not to zeocin. This strain was self-sterile but could be crossed with Ls75, resulting in sporulation. Likewise, a self-sterile zeocin-resistant strain, Ls75Z, was obtained after crossing the doublemutant strain with Ls75. Ls75Z could be crossed with Ls100, similar to the parental strain Ls75. Notably, the phenotypes of these heterothallic drug-resistant strains, Ls100H and Ls75Z, were different from either of the parental phenotypes, suggesting that meiotic recombination occurred when the homothallic Δlsku80::hph Δlslig4::Sh ble strain was crossed with the heterothallic Ls100 or Ls75.
Next, spores from a cross between Ls75Z and Ls100H were treated with glusulase, and I evaluated viable colonies for drug resistance phenotypes (Figure 5a). The proportion of cells exhibiting the parental phenotype (i.e., resistant to either zeocin or hygromycin) was 54%, and the proportion exhibiting a recombinant phenotype (i.e., resistant to both drugs or sensitive to both) was 46%. The genome sequence of L. starkeyi has been determined (Riley et al., 2016), but all contigs have not yet been joined. Thus, the physical distance between the markers (LsKU80 and LsLIG4) was not clearly defined.
Lastly, to determine if genetic recombination analysis by spore dissection is possible, ascospores resulting from a cross between Ls75Z and Ls100H were dissected using a micromanipulator Of the 10 viable progeny, six had parental phenotypes and four had recombinant phenotypes. I infer that the inviable spores were of a recombinant type that is sensitive to both drugs. These findings clearly demonstrate that genetic recombination occurred and was accurately detected following a cross of strains Ls75Z and Ls100H.
Taken together, these results show that genetic analysis can be performed by crossing the L. starkeyi derivate strains Ls75 and Ls100.

| DISCUSSION
L. starkeyi is a remarkable organism with the ability to accumulate lipids at a level that comprises up to 70% of its dry weight using various available carbon sources (McNeil & Stuart, 2018;. Despite L. starkeyi having excellent characteristics for industrial use, classical genetic cross strategies have not previously been reported. The identification of genetically crossable strains of L. starkeyi would facilitate the combining of mutations, such as mutations that have a positive impact on lipid production. In this report, I show that L. starkeyi, which forms multi-spores, can be adapted for genetic analysis, and that the strains Ls75 and Ls100 together provide a powerful resource for genetic analysis. L. starkeyi asci contain two to 20 spores, and the number of spores depends on incubation time and nutrient conditions (Figure 1).
This multi-spore appearance is the same as for the teliospores of Rhodotorula species (Newell & Hunter, 1970); however, the process of multi-spore formation has not been elucidated. A previous report showed that when over-replication at premeiotic S-phase was induced by ectopic activation of cyclin B/cyclin-dependent kinase, the asci contained up to 20 spores. Moreover, the formation of an ascus containing more than four nuclei occurred immediately following the execution of meiosis II and reached a maximum shortly thereafter (Strich et al., 2004). L. starkeyi predominantly formed hexads (five to six spores) at 14 days and octads (seven to eight spores) at 28 days ( Figure 1b, left panel). Thus, the number of spores gradually increased with longer incubation times. If the multi-spore asci of L. starkeyi form via over-replication at the premeiotic S-phase, then meiotic division and spore formation would be more rapid. In a study of Sz. octosporus, the number of spores was found to be dependent on nutrient conditions, and eight spores were not always formed (Seike & Niki, 2017).
The authors mention that if eight spores are formed within an ascus in some fission yeast species due to an extra post-meiotic mitosis, the regulation of spore number probably occurs at the level of commitment to post-meiotic mitosis. Consequently, I propose that the multispore asci of L. starkeyi might be the result of repeated rounds of post-meiotic mitosis.
I found that an L. starkeyi homothallic strain can form spores not only on AF medium but also on water solid medium ( Figure 1). Little is known about how water solid medium induces the sporulation of L. starkeyi. It has been reported that sporulation of the sake yeast Kyokai No. 7 is promoted by the presence of trace amounts of divalent cations present in agar powder (Suizu, 1996). If sporulation is promoted by divalent cations, sporulation frequency should be increased by the addition divalent cations. However, I found that adding magnesium or calcium ions or both to water solid medium did not increase the sporulation frequency and that the frequency was not affected when I used agar powder from another supplier (Wako) (data not shown). In addition, on water solid medium, I never observed sporulation after mixing Ls75 and Ls100. The sporulation frequency of the cross between Ls75 and Ls100 was lower than that of the homothallic strain (Figures 2b and 3c). Even homothallic L. starkeyi have a lower sporulation frequency on water solid medium (6.0 ± 2.3%) than on AF solid medium (17.3 ± 2.5%). Thus, it is possible that Ls75 and Ls100 can form spores on water solid medium, but the rate is so exceptionally low that these events were not detected.
This study shows that together, Ls75 and Ls100 provide a powerful resource for genetic analysis of L. starkeyi. Because the sites of mutation in these strains have not been identified, it is unclear if Ls75 and/or Ls100 have mutations at the MAT locus. Some of the progeny of a cross between Ls75 and Ls100 have a homothallic phenotype, suggesting incomplete linkage between these mutation sites. The L. starkeyi genome harbors MATa (a1 and a2 genes) and MATα (α1 and α 2 genes) loci that are close to one another ($20 kb) on the same chromosome (Hanson & Wolfe, 2017;Krassowski et al., 2019;Riley et al., 2016). The MAT loci of methylotrophic yeasts, such as Ogataea species and K. phaffii (Hanson et al., 2014;Maekawa & Kaneko, 2014;Yoko et al., 2019), are similar to that of L. starkeyi, and mating-type switching has been identified. The mating type of these yeast is changed by swapping the positions of the active and repressed MAT loci, a process that requires the IR sequence (Maekawa & Kaneko, 2014).
However, IR sequences have not been found in L. starkeyi. The mechanisms by which the mating type can be switched in the absence of an IR have not yet been revealed (Hanson & Wolfe, 2017). The authors of a previous study pointed out that mating type switching, which requires a mechanism to repress non-expressed copies of the MAT genes, evolved in the Saccharomyces lineage relatively soon after the methylation of heterochromatin at the enhanced histone H3 at lysine 9 (H3K9) was lost (Riley et al., 2016). L. starkeyi is the only Saccharomycotina species that contains orthologs of genes involved in the formation of heterochromatin via H3K9 methylation. In particular, I have reported that the L. starkeyi heterochromatin protein Lsw1 can associate with heterochromatin in Sz. pombe in an H3K9 methylation-dependent manner, suggesting that the expression of MAT locus genes might be repressed by heterochromatin via methylation of H3K9 (Takayama, 2020). Thus, analysis of the mechanisms of mating type switching using the newly isolated crossable strains, Ls75 and Ls100, might reveal a novel MAT determination system that is not dependent upon an IR sequence.