Chemical transformation of the multibudding yeast, Aureobasidium pullulans

Aureobasidium pullulans is a ubiquitous polymorphic black yeast and is one of the most common fungi cultivated in indoor environments (Fig. 1 A) (Adams et al., 2013; Humphries et al., 2017; Nonnenmann et al., 2012). It is also frequently found on plant surfaces (Andrews et al., 2002; Grube et al., 2011; Olstorpe et al., 2010) and in more extreme environments including hypersaline waters (Gunde-Cimerman et al., 2000) and glaciers (Branda et al., 2010). This fungus is well known for its ability to produce pullulan, a polysaccharide with applications in the food, pharmaceutical, and materials industries (Singh et al., 2008). A. pullulans also protects fruit from post-harvest rot caused by molds, extending its applications to the agricultural industry (Di Francesco et al., 2020; Sharma et al., 2009). Now, A. pullulans is gaining attention among cell biologists for its unusual mode of proliferation (Mitchison-Field et al., 2019). A. pullulans yeast cells are often multinucleate and can produce multiple (>10) buds in a single cell cycle (Fig. 1 A). This makes it distinct from both ascomycete and basidiomycete model budding yeasts that produce only a single bud per cell cycle, including Saccharomyces cerevisiae, Candida albicans, Cryptococcus neoformans, and Ustilago maydis. Multibudding growth poses new questions that did not arise in model yeasts. For example, it is unclear how mother cells determine the number of buds that they will make in a given cycle or whether they deliver similar numbers of nuclei and organelles to each bud. In addition, the wide variability in cell size and nuclear number among A. pullulans yeast cells provides a contrast to the care with which other yeasts regulate their volume and nuclear endowment. Thus, A. pullulans is a promising emerging model for expanding our understanding of strategies for successful proliferation.

In a recent work to develop tools for cell biological investigation in this unconventional yeast, Agrobacterium-mediated gene transfer (AMT) was used to transform constructs expressing fluorescent probes for nuclear and cytoskeletal elements into A. pullulans (Petrucco et al., 2024). Interestingly, several commonly used fluorescent proteins were not detectable in A. pullulans, possibly due to the presence of TTA Leucine codons that are very rare in that system (Petrucco et al., 2024). Because AMT frequently leads to non-homologous integration of DNA into the genome, many transformants affect gene expression in unintended ways. In contrast, the most popular yeast model systems, Saccharomyces cerevisiae and Schizosaccharomyces pombe, are typically transformed using lithium acetate (LiAc)/single-stranded DNA (ssDNA)/polyethylene glycol (PEG) to generate chemically competent cells (Gietz and Woods, 2002; Rai et al., 2018). The popularity of this approach is due to its use of inexpensive and readily available reagents, production of transformants in just 3–4 days, and applicability to intact cells requiring no spheroplasting or electroporation. The tractability of the model yeast systems is also greatly enhanced by the efficiency with which these yeasts employ homologous recombination rather than non-homologous end joining, enabling targeted homology-mediated manipulation of the genome. Here, we show that A. pullulans can also be transformed using an adapted LiAc/ssDNA/PEG protocol that includes glucose during incubation with DNA. We also use homologous recombination for effective genome manipulation and introduce codon-optimized GFP and mCherry that can be used as bright probes. Together, the protocols and tools we describe greatly increase the ease and speed with which new genome-manipulated strains can be constructed in this system.

Studies on yeast model systems have established an almost perfect correlation between DNA content (ploidy) and cell size (Galitski et al., 1999). In addition, nuclear volume is tightly coupled to cell volume and DNA content (Cantwell and Nurse, 2019; Jorgensen et al., 2007; Lemière et al., 2022; Neumann and Nurse, 2007). Model yeasts have control networks that couple cell cycle transitions to cell size so as to maintain size homeostasis in cell populations (Rhind, 2021; Schmoller and Skotheim, 2015). In contrast, we found that A. pullulans yeast cells can vary almost 70-fold in cell size. Using our new tools to construct strains with fluorescent markers allowing segmentation of nuclei and cytoplasm, we found that A. pullulans cells varied considerably in the number of nuclei and in the relation between nuclear number (DNA content) and cell size. However, the total nuclear volume was well correlated with cell size. These findings support recent proposals for a universal osmotic mechanism to couple nuclear volume to cytoplasmic volume (Lemière et al., 2022).

Because there are no known episomal plasmids for this fungus, we aimed to integrate heterologous DNA into the A. pullulans genome. A cassette with the hygromycin B phosphotransferase gene flanked by the Chytrid Spizellomyces punctatus histone HTB2 promoter and the S. cerevisiae ADH1 terminator conferred hygromycin resistance following AMT of A. pullulans (Petrucco et al., 2024). We cloned this cassette into a plasmid with ∼1 kb flanking sequences from the genomic A. pullulans URA3 gene (Fig. 1 B). Similar cassettes have been used to target integration into the A. pullulans genome following protoplasting or electroporation (Chen et al., 2023; Chi et al., 2012; Li et al., 2016; Slightom et al., 2009; Thornewell et al., 1995; Uchiyama et al., 2018). In A. pullulans, ura3 mutants are uracil auxotrophs (Rose et al., 2000), allowing rapid screening to determine whether hygromycin-resistant transformants integrated correctly at the targeted URA3 locus (Fig. 1 C). Introduction of the ura3::HYG^R cassette by AMT resulted in hygromycin-resistant transformants, of which 14.5 ± 7.7% (mean ± s.d., n = 8 experiments) were uracil auxotrophs. This demonstrates that homologous recombination can be targeted with AMT, although most transformants integrated elsewhere. We then PCR-amplified the cassette and flanking sequences to generate linear DNA for the transformation of chemically competent cells.

We began with a published LiAc/ssDNA/PEG protocol (see Materials and methods) that efficiently transforms S. cerevisiae (Fig. 1 D) (Gietz and Woods, 2002). We transformed 10⁸ cells with 10 µg linear DNA but obtained no hygromycin-resistant colonies in three independent trials. To determine how well A. pullulans cells survive the protocol, cells were plated on non-selective YPD media before and after transformation to determine the number of viable cells (colony-forming units or CFUs). Viability decreased by >99% for A. pullulans transformed with the S. cerevisiae protocol, and roughly 70% of that drop was due to incubation in 40% PEG and the rest due to the 42°C heat shock (Fig. S1 A). A 37°C heat shock was much better tolerated, with viability reduced to ∼95% (Fig. S1 A), but transformation efficiency remained very low (one hygromycin-resistant transformant from three trials).

For S. cerevisiae, adding dimethyl sulfoxide (DMSO) or ethanol was reported to increase the transformation efficiency of some strains by as much as 25-fold (Hill et al., 1991; Lauermann, 1991). However, the addition of 5% DMSO or 10% ethanol to the transformation buffer just prior to heat shocking at 37°C did not increase the number of transformants (Fig. S1 B). However, both additives reduced cell viability over 100-fold (Fig. S1 C). Thus, loss of viability may mask any improvement in the transformation of A. pullulans.

In S. cerevisiae, nutrient supplementation with amino acids was reported to improve transformation efficiency nearly 10-fold (Yu et al., 2016). For A. pullulans, the addition of amino acids to the Competence and Transformation Buffers (Materials and methods) had no significant effect on the number of transformants or cell viability (Fig. 2, A–C). However, the addition of 2% glucose increased the number of transformants dramatically (Fig. 2 A). Most of these transformants, ∼89%, failed to grow on media without uracil, indicating that the hygromycin cassette primarily integrated into the genome via homologous recombination at the URA3 locus (Fig. 2 B). Glucose addition only slightly increased cell viability, <twofold (Fig. 2 C). Thus, glucose addition appears to improve the transformation of A. pullulans through mechanisms beyond simply increasing cell viability.

To determine the optimal time to add glucose, we tested the effects of adding glucose to the Competence Buffer in step (i), the transformation buffer in step (ii) before the 1-h incubation, or the transformation buffer in step (iii) just before the heat shock (see Fig. 1 D for details on protocol steps). All conditions resulted in >100 transformants with most integrating the hygromycin cassette at the URA3 locus (Fig. 2, D and E). We achieved the most transformants when glucose was added in two steps, i and iii. Adding glucose in these two steps increased the number of transformants and cell viability by ∼1.5-fold compared with adding glucose to step (i) alone (Fig. 2, D and E). Interestingly, while glucose is key to successful transformation in A. pullulans, it is not beneficial in S. cerevisiae (Fig. S2).

With glucose supplementation, we next tested the importance of other protocol parameters. In S. cerevisiae, the addition of single-stranded carrier DNA improves the transformation efficiency ∼1,000-fold (Schiestl and Gietz, 1989). Similarly, we found that single-stranded carrier DNA increased the number of transformants dramatically for A. pullulans (Fig. 3 A) without affecting cell viability (Fig. 3 B).

In S. cerevisiae, heat shock improves transformation ∼twofold (Hayama et al., 2002; Ito et al., 1983). In A. pullulans, the heat shock step was not required but a 15-min heat shock at 37°C did increase the number of transformants ∼twofold (Fig. 3 C). Heat shock had no effect on the percentage of colonies that integrated the cassette at the targeted URA3 locus (Fig. 3 D). However, heat shock did reduce cell viability (Fig. 3 E).

To determine the best time to harvest A. pullulans cells for transformation, we grew cells in YPD to a range of densities and then transformed 10⁸ cells for each condition. We determined how cell density relates to OD₆₀₀ (Fig. S3 A) and monitored the increase in OD₆₀₀ as cells proliferated (Fig. S3 B). Cells harvested at the early log phase (0.5–1 × 10⁷ cells/ml) yielded two to fourfold more transformants than cells harvested at higher densities (Fig. 3 G), with no difference in the efficiency of integration at URA3 or cell viability (Fig. 3, G and H).

In the experiments above, cells were allowed to recover overnight in 50% YPD with 0.5 M sorbitol before hygromycin exposure. Omitting the recovery step resulted in very few transformants, and the number of transformants increased with recovery time (Fig. 4 A). There was no cell proliferation in the initial 4 h of recovery, but cell number increased ∼fivefold following overnight recovery (Fig. 4 B). This suggests that resistance develops during the first 4 h, after which the cells simply proliferate so that the fraction of resistant cells did not increase after 4 h recovery (Fig. 4 C). Recovery times did not affect integration efficiency at URA3 (Fig. 4 D).

In the model yeasts, S. cerevisiae and S. pombe, exogenous DNA can be targeted to specific genomic sites using homologous flanking sequences as short as 35–80 bp (Bähler et al., 1998; Baudin et al., 1993; Lorenz et al., 1995; McElver and Weber, 1992). To ask how the length of the flanking homologous sequences affects integration efficiency in A. pullulans, we varied the length of the homologous sequence from ∼50 bp to 1,000 bp. Transformation efficiency was similar for 800 bp and 1 kb of flanking homology. Homologous flanking sequences shorter than 800 bp resulted in fewer transformants, and those shorter than 200 bp produced very few hygromycin-resistant colonies (Fig. 4 E), most of which were not correctly integrated at the targeted URA3 locus (Fig. 4 F).

Based on these findings, we recommend harvesting cells at the early log phase, adding glucose and single-stranded carrier DNA, exposing cells to a 15-min heat shock at 37°C, allowing 4 h of recovery to develop hygromycin resistance, and using >800 bp flanking homologous sequences in integration cassettes (Fig. 1 D). These optimized conditions resulted in a transformation efficiency of 59 ± 31 transformants per μg of DNA, with 88.5 ± 8.4% of transformants correctly integrated at URA3 (mean ± sd across four experiments).

For many analyses, it is desirable to express fluorescently tagged proteins in cells. This was recently achieved in A. pullulans using AMT, yielding variable colony morphologies presumably due to different integration sites (Petrucco et al., 2024). If the genomic integration site is responsible for the variable outcome, then targeted integration at URA3 should minimize variability between transformants. This strategy has been used in several systems, including human cells, to target integration at “genomic safe harbor” loci (Papapetrou and Schambach, 2016).

The ura3::HYG^R uracil auxotrophs generated above consistently showed a slightly reduced growth rate in YPD compared with wild-type cells. We reasoned that restoring URA3 could allow selection for uracil prototrophy and allow integration of transgenes adjacent to URA3. A construct was designed introducing a neighboring transgene that expresses the A. pullulans homolog of CIT1 (Fehrenbacher et al., 2004; Suissa et al., 1984) from the S. cerevisiae ACT1 promoter with a C-terminal tdTomato tag (Fig. 5 A). We transformed the ura3::HYG^R strain using the same protocol as above except that the recovery step was omitted. Transformation yielded hundreds of uracil prototrophs (Fig. 5 B), ∼90% of which failed to grow on media with hygromycin (Fig. 5 C). Transformants grew comparably to wild-type cells on YPD, suggesting that this site constitutes a “genomic safe harbor” (Fig. 5, D and E) (Papapetrou and Schambach, 2016).

We selected six transformants that grew on media lacking uracil but not on plates with hygromycin, from three independent transformations, and imaged them using confocal microscopy. All six showed similar levels of red fluorescence signal decorating a tubulated network highly reminiscent of mitochondria in S. cerevisiae (Fig. 5, F and G) (Fehrenbacher et al., 2004). This suggests that all transformants integrated only a single copy of the CIT1-tdTomato construct. We confirmed that this signal was mitochondrial by staining the cells with the mitochondrial dye, MitoTracker. MitoTracker signal overlapped with the Cit1-tdTomato signal (Fig. 5 H). Thus, as in S. cerevisiae (Fehrenbacher et al., 2004; Suissa et al., 1984), the A. pullulans CIT1 homolog can be used as a mitochondrial marker.

To test if the uracil auxotrophs remained competent after storage at −80°C, cells were grown to the early log phase, rinsed, resuspended in Competence Buffer, and frozen by directly placing them at −80°C. A week later, the cells were thawed at room temperature and transformed using the same protocol as for freshly grown cells. This resulted in hundreds of transformants indicating A. pullulans competent cells can be prepared ahead of time and stored at −80°C for later use (Fig. 5, I and J).

A. pullulans yeast cells are unusually morphologically plastic and vary in volume (∼70-fold range) and number of nuclei (∼20-fold range). This natural variation presents a unique opportunity to investigate the scaling of nuclear and cell volume with ploidy in unperturbed cells. To accomplish this, we built two vectors designed to integrate at the native URA3 locus. The first drives dual expression of a nuclear GFP probe (NLS-GFP) and a cytoplasmic 3xmCherry probe under the control of the bidirectional Chytrid S. punctatus histone promoter (Fig. 6 A). The second uses the same design to drive dual expression of NLS-3xGFP and S. punctatus histone 2B-mCherry. We introduced both constructs using the optimized LiAc/ssDNA/PEG method and selected two independent transformants with integration at the URA3 locus as described above. All grew comparably with wild-type (Fig. S4 A).

Because previous findings indicated that TTA codons might limit the expression of common fluorescent reporters in A. pullulans (Petrucco et al., 2024), we synthesized codon-optimized versions of GFP and mCherry and readily detected the probes (Fig. 6 B). We used those signals and automated segmentation to measure the nuclear and cellular volumes of 1,002 cells (Fig. 6, B and C).

In agreement with work from other systems, the total nuclear volume (sum of all nuclear volumes) scaled with cell volume, maintaining a relatively constant nuclear/cell volume ratio (nuclear/cell volume 0.093 ± 0.036, mean ± SD, n = 1,002 cells) (Fig. 6 C). This is partially due to the presence of more nuclei in larger cells (Fig. 6 D). However, cells with the same number of nuclei can vary considerably in volume. For example, cells with two nuclei had volumes between 76 and 604 µm³, and cells with eight nuclei had volumes between 681 and 2,035 µm³ (Fig. 6 D). Leveraging this volume variability, we compared the volume of an individual nucleus to the cell volume for cells with the same number of nuclei (Fig. 6 E). This shows that the volume of an individual nucleus depends on the number of other nuclei in the same cell. This is also apparent by comparing cells that are similar in volume but contain different numbers of nuclei; for a cell with more nuclei, each nucleus is smaller (Fig. 6, F and G). Thus, A. pullulans appears to maintain a constant ratio of the total nuclear volume (sum of nuclear volumes of all nuclei) to cellular volume.

The scaling of nuclear volume to cell volume appears to be adjusted on a rapid timescale during mitosis. Whereas nuclei in premitotic mother cells all had similar volumes, postmitotic nuclei delivered to buds did not (Fig. 6, H and I; and Videos 1 and 2). When similar-sized buds received one versus two nuclei, nuclear volume adjusted accordingly as soon as it could be measured; because A. pullulans undergoes semi-open mitosis in which the NLS probe is released to the cytoplasm, we were unable to quantify nuclear volumes during mitosis (Videos 1 and 2) (Petrucco et al., 2024). While nuclei in daughter cells that received only one nucleus were larger than nuclei in daughter cells that received two, the amount of histone (spH2B-mCherry signal) in all nuclei across daughters was the same, suggesting each nucleus had the same amount of DNA (Fig. 6 J). Consistent with that conclusion, staining of DNA with SYTOX Green in fixed cells suggested that most nuclei (but see below for exceptions) had a similar DNA content, approximately double that of S. cerevisiae haploid cells stained in parallel (Fig. S4, E and F). Given that the A. pullulans genome (30 MB [Gostinčar et al., 2014]) is roughly double the size of the S. cerevisiae genome (12 MB), this is consistent with a haploid genome content in A. pullulans nuclei.

Interestingly, the concentration of NLS-3GFP (average GFP signal) in the nuclei of one-nucleus and two-nucleus daughter cells differed, with two-nucleus daughter cells showing ∼40% lower nuclear NLS signal (Fig. 6 K). The implications of this finding are considered in the Discussion.

In most multinucleate cells, all nuclei were similar in volume (CV = 0.16 ± 0.15, mean ± SD, n = 623 cells). However, occasionally (8.6% of 623 cells) cells had “mini-nuclei” (arbitrarily defined as nuclei <40% the volume of the largest nucleus in the same cell) (Fig. S4, B–D). We did not observe “macro-nuclei” much larger than their sister nuclei. The mini-nuclei occurred more frequently in larger multinucleate cells (Fig. S4 D). Nuclear volume variability can result from differential access to the cytoplasmic volume (Neumann and Nurse, 2007; Windner et al., 2019). However, mini-nuclei in A. pullulans were mixed in among the larger neighboring nuclei (Fig. S4 B). This suggests that mini-nuclei would have similar access to cytoplasm as their larger neighbors. Mini-nuclei were also detected with SYTOX staining (Fig. S4 E) and with the histone probe (Fig. S4 G), suggesting that they contain DNA. Micronuclei in animal cells result from DNA damage or chromosomal segregation errors during mitosis (Fenech, 2000; Heddle, 1973; Schmid, 1975). While the origin of mini-nuclei in A. pullulans is unknown, the observation that they are more common in multinucleate cells suggests that if they are the result of mitotic errors, these may occur more frequently in larger cells.

Here, we present a LiAc/ssDNA/PEG protocol using intact A. pullulans cells that yields a transformation efficiency of ∼60 transformants per μg of DNA in just 3 days. Impressively, this efficiency is comparable to or higher than that reported for more laborious spheroplasting (Guo et al., 2017; Zhang et al., 2019) and electroporation (Guo et al., 2017) protocols. Efficient transformation with LiAc/ssDNA/PEG required the addition of 2% glucose during the transformation. In S. cerevisiae, the addition of amino acids increases transformation efficiency, potentially by stimulating substrate-induced endocytosis of amino acid transporters at the plasma membrane (Yu et al., 2016). Consistent with that hypothesis, a subset of genes involved in endocytosis is required for the efficient transformation of S. cerevisiae via LiAc/ssDNA/PEG (Kawai et al., 2004, 2010). In contrast, the addition of amino acids did not improve the transformation of A. pullulans. Conversely, addition of glucose, which is key in A. pullulans, did not improve S. cerevisiae transformation. Thus, while the effect of different nutrients varies between fungi, nutrient supplementation can improve transformation by more than an order of magnitude. Testing the effects of different nutrients may enable PEG-based transformation of a wider range of fungi including those that previously proved recalcitrant to this technique.

Our DNA constructs include flanking sequences with homology to the A. pullulans genome, and with 1 kb homology arms, ∼90% of transformants integrated the constructs at the targeted locus. Using this method, we built integration vectors designed to introduce different fluorescent probes for chromatin, nuclei, cytoplasm, and mitochondria at the URA3 locus. All probes were expressed well and had no detectable effect on growth. Our vectors can be easily modified to express different tagged proteins. Thus, the URA3 locus is a safe harbor for the expression of transgenes in A. pullulans, and our ura3::Hyg^R strain and optimized vectors provide a starter kit to enable such expression. Together, our protocol for efficient LiAc/ssDNA/PEG transformation, homologous recombination to target expression of transgenes at a safe harbor, and fluorophores optimized for A. pullulans provide a large step toward making this unconventional fungus tractable for cell biological investigations.

As proof of principle, we built strains to explore novel cell biology in A. pullulans. Previous studies in yeast and other systems had suggested that cell size scales with DNA content and nuclear volume scales with cell size (Cantwell and Nurse, 2019; Fankhauser, 1939; Galitski et al., 1999; Jorgensen et al., 2007; Kondorosi et al., 2000; Lemière et al., 2022; Neumann and Nurse, 2007; Windner et al., 2019). In yeast cells, powerful size control mechanisms maintain a uniform cell size (Chen and Futcher, 2021), so studies on scaling either have a limited size range or use experimental perturbations to create artificial variability. The natural variability in size of A. pullulans provided an opportunity to ask how DNA content, nuclear volume, and cell volume scale over a wider range without applied perturbations. In particular, we wondered whether nuclei with the same DNA content would all be similar in size or whether nuclear volume would scale with cell volume instead. We found that nuclear volume scales with cell volume over a ∼70-fold range in cell volume, rather than with DNA content. This is consistent with findings in animal cells and mononucleate yeasts, where nuclear volume depends on the available volume of the cytoplasm and on the synthesis and transport of macromolecules between the nucleus and cytoplasm (Ganguly et al., 2016; Kume et al., 2017; Lemière et al., 2022; Neumann and Nurse, 2007). Nuclear import/export was proposed to determine nuclear volume by regulating the macromolecule concentration of the nucleus relative to the cytoplasm thereby determining the osmotic forces on the nucleus that set its volume (Lemière et al., 2022).

The scaling of nuclear volume with cell volume seems to develop on a rapid timescale after mitosis. In cases where similar-sized daughters from the same mother received one or two nuclei after mitosis, each nucleus in the two-nucleus cells was about half the volume of each nucleus in the 1-nucleus cells (Fig. 6, H and I) In contrast, when S. pombe cells are induced to divide asymmetrically, it takes an entire cell cycle to restore the nucleus:cytoplasmic ratio (Cantwell and Nurse, 2019; Lemière et al., 2022). The closed mitosis of S. pombe generates two equal-sized nuclei. However, mitosis in A. pullulans is semi-open, with an NLS-containing probe released into the cytosol during mitosis (Petrucco et al., 2024). This allows post-mitotic nuclear protein import to depend on the cytoplasmic “catchment area” available to each nucleus, leading to rapid re-scaling of nuclear volume to cell volume. Rapid rescaling may be unnecessary for an organism like S. pombe where cell size is uniform, but advantageous for the morphologically plastic A. pullulans.

An intriguing and unexpected finding was that nuclei in comparably-sized two-nucleus daughters were not only smaller than those in one-nucleus daughters but also had lower concentrations of the NLS-3GFP probe (Fig. 6, H and K). As probe accumulation into post-mitotic nuclei occurs prior to cytokinesis (Videos 1 and 2), the nuclei in mother and daughter cells presumably compete for the cytoplasmic NLS probe. To account for the probe concentration differences noted above, we speculate that smaller nuclei import the probe more slowly than larger nuclei. This would give larger nuclei an edge in the competition for the NLS probe, leading to the observed concentration differences that would then get “locked in” during cytokinesis.

Semi-open mitosis implies that many nuclear proteins would become cytoplasmic during mitosis and that post-mitotic nuclei would compete for reimport of such proteins. This could lead to different concentrations of specific proteins in nuclei of different size, with smaller nuclei having lower concentrations, as observed for the NLS probe. Conversely, proteins like histones that stay associated with the DNA during mitosis would be inherited in equal amounts by all nuclei and would therefore be present at higher concentrations in smaller nuclei (Fig. 6 J). The picture that emerges from these findings is that in A. pullulans, the same haploid genomic DNA can find itself in nuclei with significantly different volume and protein composition, raising the interesting question of how the amounts and concentrations of various DNA-interacting proteins like histones, chromatin regulators, transcription factors, helicases, and polymerases vary between nuclei, and what consequences such variation may have for gene expression, cell cycle progression, and other processes.

Strains used in this study are listed in Table S1. All experiments were conducted with A. pullulans strain EXF-150 (Gostinčar et al., 2014). Unless otherwise indicated, A. pullulans was grown at 24°C in standard YPD medium (2% glucose, 2% peptone, 1% yeast extract) with 2% BD Bacto agar (214050; VWR) in plates.

To replace the endogenous A. pullulans URA3 coding region (protein ID 283826) with a hygromycin-resistance cassette, we used DLB4621. This plasmid contains the hygromycin resistance gene, aph(4)-Ia, from plasmid pGI3EM22C (#135488; Addgene) driven by the Chytrid S. punctatus histone H2B promoter and the S. cerevisiae ADH1 terminator, previously shown to confer hygromycin resistance in A. pullulans (Petrucco et al., 2024). This HYG^R cassette is flanked by 1,000 bp sequences upstream and downstream of the URA3 locus from the A. pullulans genome. DNA fragments were obtained by restriction digest or PCR and assembled into the pGI3EM22C plasmid backbone using the NEBuilder HiFi DNA Assembly Mater Mix (E2621L; New England Biolabs).

To replace ura3::HYG^R with URA3:CIT1-tdTomato, we used DLB4702. This plasmid contains the same backbone and URA3 flanking sequences as DLB4621 but inserts the A. pullulans URA3 coding region and S. cerevisiae TEF1 terminator, followed by the S. cerevisiae ACT1 promoter driving expression of the A. pullulans CIT1 coding region (protein ID 294441) fused in frame with tdTomato, followed by the S. cerevisiae ADH1 terminator. DNA fragments were obtained by restriction digest or PCR and assembled into the DLB4621 plasmid backbone using the NEBuilder HiFi DNA Assembly Mater Mix (E2621L; New England Biolabs).

To replace ura3::HYG^R with URA3:3xmCherry:NLS-GFP, we used DLB4781. This plasmid contains the same backbone and URA3 flanking sequences, URA3 coding sequence, and S. cerevisiae TEF1 terminator as DLB4702, but inserts a dual expression cassette with the bidirectional Chytrid S. punctatus histone H2A/B promoter driving expression of three tandem copies of mCherry with the S. cerevisiae CYT1 terminator and NLS-GFP with the S. cerevisiae ADH1 terminator. The NLS sequence is derived from a biosensor with two tandem NLS sequences shown previously to label nuclei in A. pullulans (Petrucco et al., 2024). The mCherry and GFP sequences were codon-optimized to eliminate all TTA codons and reduce rare codons (Petrucco et al., 2024). Optimized mCherry and GFP sequences were synthesized by Twist Bioscience. DNA fragments were obtained by restriction digest or PCR and assembled into the DLB4702 plasmid backbone using the NEBuilder HiFi DNA Assembly Mater Mix (E2621L; New England Biolabs).

To replace ura3::HYG^R with URA3:HTB2-mCherry:NLS-3xGFP, we used DLB4815. This plasmid contains the same backbone and URA3 flanking sequences, URA3 coding sequence, and S. cerevisiae TEF1 terminator as DLB4702, but inserts a dual expression cassette with the bidirectional Chytrid S. punctatus histone H2A/B promoter driving expression of H2B-mCherry with the S. cerevisiae CYT1 terminator and NLS-3xGFP (three tandem copies of GFP) with the S. cerevisiae ADH1 terminator. GFP and mCherry sequences were codon-optimized as described above. DNA fragments were obtained by restriction digest or PCR and assembled into the DLB4702 plasmid backbone using NEBuilder HiFi DNA Assembly Mater Mix (E2621L; New England Biolabs).

All plasmids were confirmed by whole plasmid sequencing (Plasmidsaurus). Plasmids and full sequences are available through Addgene.

Linear transforming DNA was prepared by PCR or restriction digest. In most experiments, to knock out URA3 and replace it with the Hyg^R cassette, the ura3::Hyg^R cassette (with 1,000 bp URA3-homologous flanking sequences) was amplified from DLB4621 using iProof HF Master Mix (#1725310; BioRad) with primers CP001 (5′-GGC​AAA​ACG​AAA​CGG​GCG​C-3′) and CP002 (5′-GGA​GTG​CAG​ACA​TGA​CTA​TG-3′). To test integration efficiency using homologous flanking sequences of various lengths, additional primer pairs were used to amplify the ura3::Hyg^R cassette: 800 bp Ura3HA_800 bp_F (5′-AGA​GTG​GAG​TGT​TCT​CAG​GCA​G-3′) and Ura3HA_800 bp_R (5′-TAG​TTT​GAC​AGA​ATC​GAA​ACC​GAG​C-3′), 400 bp Ura3HA_400 bp_F (5′-AGC​AGC​CTC​GCT​CTT​TGC​TG-3′) and Ura3HA_400 bp_R (5′-CCG​TTT​CTC​AAT​CAT​TGT​GGA​GC-3′), 200 bp Ura3HA_200 bp_F (5′-TGG​TGC​TCC​TCG​GTC​GTC-3′) and Ura3HA_200 bp_R (5′-GAT​GAA​TCC​CAA​GCT​GCT​CTA​CAC-3′), 100 bp Ura3HA_100 bp_F (5′-TTC​GAC​AGC​TAC​CCA​GTC​GC-3′) and Ura3HA_100 bp_R (5′-CGG​AGG​CTT​CGA​TGT​TTC​CTG​TC-3′), and 50 bp Ura3HA_50 bp_F (5′-TTC​GAC​AGC​TAC​CCA​GTC​GC-3′) and Ura3HA_50 bp_R (5′-CGG​AGG​CTT​CGA​TGT​TTC​CTG​TC-3′).

To introduce the mitochondrial marker (URA3:CIT1-tdTomato), the NLS and cytoplasmic markers (URA3:NLS-GFP;3xmCherry), and the NLS and histone markers (URA3:HTB2-mCherry:NLS-3xGFP) DLB4702, DLB478, and DLB4815 (respectively) were linearized with restriction enzymes ApaI and NruI-HF. Linear DNA was purified and concentrated by adding 0.1 volume of 3 M NaOAc and 2.5 volumes of 100% ethanol followed by incubation on ice for 10 min. The DNA–ethanol mixture was then added to a silica DNA binding column (T1017-2; New England Biolabs), rinsed with 400 µl wash buffer (80% EtOH 20 mM NaCl 2 mM Tris pH 8), allowed to dry for 1 min, and then eluted in ∼40–60 µl 10 mM Tris pH 8 at a final concentration of ∼1,500 ng/µl.

To make competent cells, a single A. pullulans colony was inoculated into 5 ml of YPD (4% glucose) and grown overnight (∼18 h) at 24°C with agitation. The following day, the overnight culture was diluted to a final density of ∼5 × 10⁶ cells/ml in 50 ml of YPD (4% glucose) in a 250-ml flask. Cells were grown at 22–24°C with agitation at 200 rpm for ∼2–3 h until reaching a density of ∼10⁷ cells/ml, unless indicated otherwise. Cells were harvested via centrifugation at 2,254 rcf for 8 min in 50 ml conical tubes and then resuspended in 1 ml sterile room-temperature deionized water and transferred to a 1.5 ml Eppendorf tube before pelleting at 9,391 rcf for 10 s. The cells were rinsed once in 500-µl sterile Competence Buffer (10 mM Tris pH 8, 1 mM EDTA, 1 M sorbitol) and pelleted at 9,391 rcf for 10 s. Finally, cells were resuspended at a final density of ∼2 × 10⁹ cells/ml in Competence Buffer, usually ∼250 µl, and divided into aliquots in 1.5 ml Eppendorf tubes with ∼10⁸ cells in each aliquot (∼60–65 µl in each). For conditions with amino acids supplemented into the Competence Buffer, Complete Supplement Mixture minus uracil (1004-100; Sunrise Science Products) and uracil (600-621-9; Sigma-Aldrich) were added to the Competence Buffer at a final concentration of 0.96 and 0.025 g/l, respectively. This corresponds to an amino acid concentration 1.25× the concentration used in a standard complete synthetic medium previously shown to improve the transformation of S. cerevisiae (Yu et al., 2016). Competence Buffer with amino acids was prepared fresh just before use and filter-sterilized. To make frozen competent cells for future use, aliquots of cells in Competence Buffer were transferred to a −80°C freezer.

Prior to transformation, 10 mg/ml single-stranded fish-sperm DNA (11467140001; Roche) in 10 mM Tris-HCl, 10 mM NaCl, and 1 mM EDTA, pH 8.0, was boiled for 5 min, briefly vortexed, and stored on ice for use later that day or at −20°C for long-term storage. After cooling on ice, 10 µl carrier DNA and 10 µg linear DNA (7–10 µl) were added to an aliquot of competent cells. If using previously frozen competent cells, cells were allowed to thaw at room temperature for ∼5 min before use. For conditions with 2% glucose in the Competence Buffer, 2.5 µl 40% glucose was also added to the competent cells. After each addition, cells were gently mixed by flicking the side of the tube. Next, 600 µl Transformation Buffer (10 mM Tris pH 8, 1 mM EDTA, 40% [wt/vol] PEG 3350) was added, and the cells were inverted quickly several times to mix. Transformation Buffer was always made fresh just before use. For conditions with 2% glucose in the Transformation Buffer, 30 μl 40% glucose was added before or after the 1-h incubation at 24°C as indicated. For conditions with amino acids added to the Transformation Buffer, Complete Supplement Mixture minus uracil (1004-100; Sunrise Science Products), and uracil (600-621-9; Sigma-Aldrich) were added at a final concentration of 0.96 and 0.025 g/l, respectively. Cells in the Transformation Buffer were incubated at 24°C for 1 h with mixing on a culture wheel and then heat shocked at 37°C or 42°C in a heat block for 15 or 30 min as indicated. For conditions with DMSO or ethanol, 70 µl anhydrous DMSO (final concentration 10%) or 45 µl 100% ethanol (final concentration 5%) was added just before the heat shock. Following heat shock, cells were pelleted at 9,391 rcf for 10 s and the PEG solution was removed. For selection on Hygromycin B, cells were resuspended in 1 ml 1 M sorbitol and added to 1 ml YPD in a glass culture tube. Cells were then allowed to recover for the indicated time at 24°C with gentle shaking at 100 rpm before plating on YPD plates supplemented with Hygromycin B (400051-1MJ; Millipore) at a final concentration of 174 µl/l (∼70.4 mg/l). For selection on drop-out uracil medium, cells were resuspended in 250 µl 1 M sorbitol and plated on media lacking uracil (6.71 g/l BD Difco Yeast Nitrogen Base without Amino Acids, BD291940; Thermo Fisher Scientific, 0.77 g/l Complete Supplement Mixture minus uracil, 1004-100; Sunrise Science Products, 2% glucose, and 2% BD Bacto agar, 214050; VWR). Plates were incubated at 24°C and colonies typically appeared after 2–3 days.

To screen transformants for integration at the targeted site, 30 individual colonies were struck out onto selection media, allowed to grow for 2 days at 24°C, and replica were plated onto test media (either YPD with Hygromycin B or plates lacking uracil) to assess whether markers were swapped (i.e., whether hygromycin resistance was associated with uracil auxotrophy, or whether uracil prototrophy was associated with hygromycin sensitivity).

To test if glucose addition improves the transformation of S. cerevisiae, haploid S. cerevisiae (W303) was grown overnight in YPD (2% glucose) at 24°C. Overnight cultures were back diluted and allowed to grow to a density of ∼1 × 10⁷ cells/ml in 50 ml YPD at 24°C. Cells were transformed with 5 µl (7,500 ng) XcmI linearized pRS306 integration vector using the same protocol described for A. pullulans above with and without the addition of glucose. Cells were selected on -Ura plates and the number of colonies was counted after 3 days at 24°C.

To quantify cell viability during the transformation protocol, 1% of the volume of the cell suspension was removed at the indicated steps and serially diluted 1,000× (dilution 1) and 500,000× (dilution 2) in sterile water. 50 µl dilution 1 and 150 µl dilution 2 were plated on YPD and incubated at 24°C for 48 h. The number of colonies (CFUs) was counted.

To compare the cell growth of A. pullulans strains, a single colony was inoculated into 5 ml of YPD (2% glucose) and grown at 24°C for 48 h. Cultures were serially diluted five times in sterile deionized water in a 96-well plate and transferred onto solid YPD plates using a pin-frogger. Plates were grown for 48 h at 24°C and imaged on an Amersham Imager 680 (General Electric Company) using the colorimetric Epi-white settings. To quantify cell growth, the width of the spot from the first serial dilution was measured in Fiji (Schindelin et al., 2012).

To grow yeast for imaging experiments, a single colony was used to inoculate 5 ml of YPD (2% glucose). Cultures were grown overnight at 24°C to a density of 1–5 × 10⁶ cells/ml. Cells were pelleted at 9,391 rcf for 10 s and resuspended at a final density of ∼7 × 10⁷ cells/ml in complete synthetic media, CSM (6.71 g/l BD Difco Yeast Nitrogen Base without Amino Acids, BD291940; Thermo Fisher Scientific, 0.77 g/l Complete Supplement Mixture minus uracil, 1004-100; Sunrise Science Products, 20 mg/l uracil, 600-621-9; Sigma-Aldrich, and 2% glucose). Approximately, 2 × 10⁵ cells were mounted on a 500 µl 1.5% agarose (97062-250; VWR) pad made with CSM. All experiments were conducted at room temperature (20–22°C).

To stain mitochondria, 0.5 µl MitoTracker Green FM (M7514; Thermo Fisher Scientific) was added to 100 μl overnight culture in YPD and incubated with agitation at 24°C for 10 min. Before imaging, cells were rinsed twice with CSM and resuspended at a final density of ∼7 × 10⁷ cells/ml.

To measure fluorescence levels, cells were imaged on a Nikon Ti2E inverted microscope with a CSU-W1 spinning-disk head (Yokogawa), CFI60 Plan Apochromat Lambda D 60× Oil Immersion Objective (NA 1.42; Nikon Instruments), and a Hamamatsu ORCA Quest qCMOS camera controlled by NIS-Elements software (Nikon Instruments). The entire cell volume was acquired using 75 Z-slices (at 0.2 μm steps). To image cells expressing Cit1-tdTomato, exposure times of 100 ms at 50% laser power (excitation 561 nm) were used. For cells stained with MitoTracker Green, exposure times of 50 ms at 10% laser power (excitation 488 nm) were used. For cells coexpressing 3xmCherry and NLS-GFP, exposure times of 200 ms at 50% and 70% laser power were used for the 561 and 488 nm lasers, respectively. Time series of strains expressing NLS-3xGFP and H2B-mCherry were acquired using triggered acquisition with 50 ms exposure time and 8% and 5% laser power for the 561 and 488 nm lasers, respectively.

To quantify Cit1-tdTomato expression levels, z-stacks were average intensity projected, the cells were segmented, and the mean fluorescence inside each cell and outside all cells (background) was measured for each image using NIS-Elements General Analysis 3 (GA3, Nikon Instruments) software. The average Cit1-tdTomato signal for each cell was calculated by subtracting the mean background signal from the mean cell signal. To measure the cell and nuclear volume, cells expressing 3xmCherry and NLS-GFP were 3D segmented and the volumes were measured using NIS-Elements General Analysis 3 (GA3; Nikon Instruments) software.

To quantify DNA content using SYTOX Green (S7020; Thermo Fisher Scientific), wild-type A. pullulans and haploid S. cerevisiae (YEF473) cells were grown overnight at 24°C in YPD and then 1 × 10⁷ cells were fixed in 4 ml 70% ethanol for 1 h. Cells were rinsed once with deionized H₂O and incubated with 2 mg/ml RNaseA (EN0531; Thermo Fisher Scientific) for 3 h at 37°C. Cells were then rinsed with 50 mM Tris pH 7.5, and the S. cerevisiae and A. pullulans cells were mixed in a final volume of 500 µl 50 mM Tris pH 7.5, S. cerevisiae and A. pullulans cells were stained together in the same tube with 10 μl SYTOX Green (5 mM stock, S7020; Thermo Fisher Scientific) for 1 h. Following staining, cells were rinsed three times with 50 mM Tris pH 7.5 and directly mounted on glass slides.

Images of SYTOX stained cells were acquired with an Andor Revolution XD spinning-disk confocal microscope (Andor Technology) with a CSU-X1 5,000-rpm confocal scanner unit (Yokogawa) and a UPLSAPO 100 ×/1.4 oil-immersion objective (Olympus) controlled by MetaMorph software (Molecular Devices). Z stacks with 17-z steps of 0.2 μm were captured by an iXon 897 EMCCD camera (Andor Technology). Exposure time was 100 ms with 20% laser power. All imaging was done at room temperature (22–25°C).

Image analysis was done in FIJI (Schindelin et al., 2012). Z-stacks were made into sum intensity projections, and the total SYTOX signal (total DNA content) per nucleus was calculated by multiplying the mean SYTOX signal by the area of each nucleus. Values are reported relative to the average total SYTOX signal across all S. cerevisiae nuclei.

All statistical analysis was done using GraphPad Prism. Unless indicated otherwise, data distributions were assumed to be normal, but this was not formally tested. Statistical comparison between indicated conditions was conducted using the two-sided Student’s t test or one-way ANOVA, as indicated in the figure legends. After running an ANOVA, the Tukey test was used to compare the mean with every other mean. Differences were considered significant if the P value was <0.05.

Fig. S1 shows the effects of different heat shock conditions or additives (ethanol or DMSO) on A. pullulans transformation efficiency and viability. Fig. S2 shows the effect of adding 2% glucose to S. cerevisiae during chemical transformation. Fig. S3 shows a growth curve of A. pullulans and the relationship between the cells/ml and OD₆₀₀. Fig. S4 relates to Fig. 6 in the main text and shows the growth of cells expressing NLS-GFP, 3xmCherry or NLS-3xGFP, H2B-mCherry compared with WT as well as examples of cells with mini-nuclei and associated quantification of mini-nuclei sizes and the % of cells with mini-nuclei. Fig. S4 also shows quantification of DNA content per nucleus in S. cerevisiae and A. pullulans.Table S1 provides a list of yeast strains used in this study. Videos 1 and 2 provide example time series of mitosis in cells expressing NLS-GFP and H2B-mCherry.

The data generated in this study are available from the corresponding author upon reasonable request.

We thank Alex Crocker for help with codon optimization, and Amy Gladfelter, Audrey Williams, Clara Fikry, and Alex Crocker for comments on the manuscript. Thanks to the Lew and Gladfelter labs for stimulating discussions. Some of the imaging experiments were performed with instruments from the Duke Light Microscopy Facility, and we thank Lisa Cameron and Yasheng Gao for their help with microscopy.

This work was funded by National Institutes of Health/National Institute of General Medical Sciences grant R35GM122488 to D.J. Lew.

Author contributions: A.C.E. Wirshing: Conceptualization, Data curation, Formal analysis, Investigation, Visualization, Writing - original draft, Writing - review & editing, C.A. Petrucco: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing - review & editing, D.J. Lew: Conceptualization, Funding acquisition, Project administration, Supervision, Writing - review & editing.