High-resolution live cell imaging to define ultrastructural and dynamic features of the halotolerant yeast Debaryomyces hansenii

Debaryomyces hansenii is a halophilic and osmotolerant yeast species that has been of interest for basic and applied research due to its unique ability to withstand high osmotic pressure, high salinity, and low water activity (Breuer and Harms, 2006). This yeast is notably recognised for its involvement in the ripening processes of meats and cheeses, as well as its role in the synthesis of valuable chemicals like xylitol and riboflavin (Prista et al., 2016; Yaguchi et al., 2017). Despite its potential, some structural and physiological peculiarities together with the slow development of genetic tools for D. hansenii have limited its scientific use and biotechnological potential. Early studies considered D. hansenii as a possible emergent pathogen (St-Germain and Laverdière, 1986; Wagner et al., 2005), but this was most probably due to misidentification of the strains, since this yeast is phylogenetically related to some Candida species and distinguishing between species was previously difficult (Desnos-Ollivier et al., 2008; Nishikawa et al., 1996, 1997).

Beyond this, studies in D. hansenii have been further hindered due to its intractability to typical techniques used for other yeast species (e.g. disruption of the cell wall), potentially low fermentive capacity, and its ability to adapt and grow in the presence of different antibiotics (Ruiz–Pérez et al., 2023). Moreover, and importantly, genetic heterogeneity and distinct chromosome polymorphism has been argued to explain contradictory findings (Huang et al., 2021; Petersen and Jespersen, 2004). The slow development of genetic tools for D. hansenii has been affected by several factors. One of the biggest drawbacks has been the absence of efficient transformation protocols until recent years (Minhas and Biswas, 2019; Minhas et al., 2009). Furthermore, the availability of only a few circular plasmids capable of propagation in this yeast posed a major challenge, as these plasmids relied on auxotrophic markers and were applicable only to specific laboratory strains (Maggi and Govind, 2004; Ricaurte and Govind, 1999). To complicate matters, the genetic engineering of D. hansenii has been severely limited due to its membership in the CTG clade, which encodes serine instead of leucine and the preference for non-homologous end joining in integration of DNA in chromosomes, making it extremely difficult to introduce precise alterations to the genome through homology-directed repair (Kawaguchi et al., 1989; Prista et al., 2016). Nevertheless, recent advances have been made in the genetic engineering of D. hansenii, particularly with the development of a CRISPR/Cas9 method that facilitates efficient oligo-mediated gene editing (Spasskaya et al., 2021; Strucko et al., 2021). Although that method is not yet widely used, it is plausible that this progress will be significant as it allows for precise gene editing, which is essential for the development of D. hansenii as a cell factory for various biotechnological applications (Navarrete et al., 2022).

Due to the potential of D. hansenii it is unfortunate that the above-described barriers have limited understanding the ultrastructural organisation and physiological regulation of this yeast species. Beyond some original electron microscopy analyses (Gezelius and Norkrans, 1970; Rij and Veenhuis, 1975), little has been learnt about organelle organisation since; even less is known about dynamic processes in living D. hansenii cells. This is particularly frustrating given that other yeast species, such as the budding yeast Saccharomyces cerevisiae, have an extensive arsenal of tools and reagents that have been used to define eukaryotic cell biology and drive industrial applications (Duina et al., 2014; Mattanovich et al., 2014). To overcome this paucity, we have optimised the use of fluorescent dyes with super-resolution microscopy, in addition to label-free holotomography, to elucidate various cell biological aspects of D. hansenii. Our work reveals many physiological details, including a 3D segmentation-based identification of cell cycle stages, intracellular organisation of different organelles, trafficking through the vacuolar degradation and endosomal recycling pathways, and observations of organelle inheritance.

Initial brightfield imaging experiments using differential interference contrast (DIC) confirmed that Debaryomyces hansenii are markedly smaller in size than Saccharomyces cerevisiae (Fig. 1A). To quantify this difference, measurements of cellular area segmented from brightfield micrographs showed a mean area of 17.2±5.8 µm² for S. cerevisiae compared with 10.2±5.8 µm² for D. hansenii (Fig. 1B). We have previously noted that this type of analysis from dividing cells results in a wide range of areas (Laidlaw et al., 2022a), so we sampled more than 600 cells per yeast species. To underscore this difference, we cultured a strain of S. cerevisiae expressing a GFP tagged flavodoxin-like protein Ycp4 (Grandori and Carey, 1994), which localises to eisosomes to promote quiescence (Megarioti et al., 2023). S. cerevisiae cells expressing GFP-Ycp4 from a constitutive promoter were cultured in parallel with unlabelled D. hansenii then mixed. Confocal imaging of the co-culture, with S. cerevisiae cells identified by expression of GFP-Ycp4, was used to compare to the relatively smaller D. hansenii cells in the same micrographs (Fig. 1C).

In an effort to analyse cellular aspects of yeast without expressing fluorescent proteins we first optimised holotomography for imaging of S. cerevisiae cells based on their relative refractive index (Fig. 2A). The yeast cells could be clearly visualised from this rapid imaging technique, and using the 3D analysis software individual cells could be identified and segmented for downstream processing (Fig. 2B). We imaged large fields of view of cultured S. cerevisiae and D. hansenii cells and segmented to create holotomograms for individual cells. We manually categorised these cells based on cell cycle stage, with cells assigned to groups of single cells with no buds, cells with small daughter cells budding, and cells later in the cell-division process with large buds (Fig. 2C). Encouragingly, the described holotomography analysis could be performed in rich YPD media, which is essentially incompatible with fluorescence microscopy due to its auto-fluorescence (Brouwer et al., 2020). Holotomography allows for a range of physical parameters to be measured effortlessly from these segmented cells. For example, estimates of dry mass, volume and length confirmed S. cerevisiae cells are consistently larger than the D. hansenii counterparts at the same stage of the cell cycle (Fig. 2D). To ascertain stages of the cell cycle in an automated manner, we compared sphericity measurements from the pre-defined categories of S. cerevisiae single cells (e.g. a mother cell in G1 phase or a newly budded daughter cell) or cells in the active process of budding (Fig. 2E). Single cells have sphericity >0.84 and dividing cells have sphericity <0.82. As with other parameters, when the sphericity of non-budding S. cerevisiae cells were compared between YPD and SC media, no significant difference was observed, supporting the notion that holotomography can be performed equally well in rich YPD culture media. We then confirmed cell cycle stage of D. hansenii using sphericity with single cells >0.75 and diving cells <0.72 (Fig. 2F).

Although D. hansenii clearly exhibits different physical and physiological features to S. cerevisiae, we considered analysing D. hansenii proteins using antibodies raised against S. cerevisiae proteins. For this, we used antibodies that had previously been used to recognise: the E3 ubiquitin ligase Rsp5 (Stamenova et al., 2004), the eisosme factor Pil1 (Walther et al., 2007), and the endosomal sorting complex required for transport (ESCRT) associated protein Ist1 (Tan et al., 2015). The respective orthologues were identified using protein-protein BLASTP (Camacho et al., 2009) against Debaromyces hansenii (taxid:4959): Rsp5/DEHA2D03718; Pil1/DEHA2F01694; Ist1/DEHA2C12408 (Fig. 3A). Most D. hansenii counterparts are similar in predicted molecular weight but with a range of sequence similarity. However, even Ist1 with lowest homology shares many regions across diverse yeast species (Fig. 3B-C). Normalising lysates based on optical density did not create equal loading of lysates, presumably due to difference in cellular size, but loading 3-fold more (3×) D. hansenii lysate provided relatively similar proteome profiles, albeit with very distinct protein enrichment profiles (Fig. 3D). Using this harvesting process, we confirmed that orthologues of Rsp5, Pil1 and Ist1 could all be detected at predicted molecular weights, although levels of the D. hansenii Rsp5 orthologue DEHA2D03718 was comparatively low (Fig. 3E).

Having compared D. hansenii to its better studied budding yeast counterpart S. cerevisiae, we next sought to optimise established tools from S. cerevisiae work to reveal details about the cellular architecture and dynamic processes in D. hansenii. The first structure we targeted for localisation was the nucleus. Nuclear stains have been used for many decades and are useful for demarcating individual cells, providing information on sub-nuclear concentrations of DNA, and for simple analysis of nuclear morphology and size. For this, we used Hoechst 33342, a blue fluorescent dye that binds to AT-rich DNA sequences (Latt and Wohlleb, 1975). Labelling of D. hansenii nuclei was very inefficient in SC media, therefore, we briefly incubated the cells in Tris buffer at pH 7.2 to increase nuclei labelling efficiency. Using this optimised protocol, the majority of cellular nuclei were defined (Fig. 4A). Beyond this, focus on individual cells allowed clear indication of nuclear localisation that can be distinguished with vacuoles observable from DIC imaging (Fig. 4B). These micrographs also demonstrate a small level of mitochondrial DNA separate from the nucleus. We next wanted to label mitochondria, which is highly important for metabolism of budding yeasts (Malina et al., 2018). For this, the carbocyanine-based dye, MitoTracker Green, was used to label living D. hansenii cells in SC media (Chazotte, 2011). MitoTracker Green permeates the surface membrane and accumulates in active mitochondria (Fig. 4C). As with S. cerevisiae, many cells exhibited a characteristic ribbon morphology with structures dispersed across mother and daughter cells during cell division (McFaline–Figueroa et al., 2011; Westermann and Neupert, 2000). Additionally, some other cells exhibited a more punctate localisation of mitochondria.

We next set out to define the cell wall of D. hansenii using Calcofluor-white (CFW), a blue fluorescent compound that binds cellulose and chitin of fungal cell walls and is enriched in chitin ‘bud-scar’ regions following separation of the budding daughter cell (Cabib and Bowers, 1971; Hayashibe and Katohda, 1973). These regions are typically observed as ring-shaped structures on the cell wall. As expected, CFW provides a bright signal at the cell wall, and allows bud scars to be observed (Fig. 5A). Older cells exhibit more bud scars (Bartholomew and Mittwer, 1953), which was easier to assess by taking 45×150 nm z-slices, with individual scars only apparent at certain regions (Fig. 5B) or from construction of 3D volumes (Movie 1). We also performed structured illumination microscopy (SIM) to obtain 3D images of CFW stained D. hansenii cells, to better highlight the discrete bud scars (Fig. 5C). Finally, we tested the capacity of the blue dye CMAC (7-amino-4-chloromethylcoumarin) to label the vacuolar lumen of D. hansenii cells. Labelling of vacuoles was very efficient in minimal media, even following a brief incubation with CMAC (Fig. 5D). The size and number of vacuoles per cell was easy to ascertain from CMAC staining, with fission-fusion vacuolar projections indicative of vacuolar dynamics observable (Fig. 5E).

To gain dynamic information about trafficking in the endolysosomal system, we used the styryl dye FM4-64, which is useful for visualising membranes in yeast, as it binds phospholipids of the plasma membrane and robustly increases fluorescence (Vida and Emr, 1995). We have recently employed an FM4-64 pulse chase technique combined with Airyscan2 to survey the endosomal system of S. cerevisiae (Andreev et al., 2023). A similar strategy was used to initially label the plasma membrane of D. hansenii cells, then temporally label endocytic trafficking to the lysosome/vacuole (Fig. 6A). At first, the dye is exclusively localised to the surface membrane. Only 2 min of the label-free chase was sufficient to observe significant endosomal signal in addition to the PM. Further time points show FM4-64 traffics from the PM to exclusively in endosomes at 10-min chase, followed by successful trafficking to the vacuole at 20 min and increasing over the remaining pulse period. This serves as a useful kinetic marker of endocytosis, in addition to spatially labelling the endolysosomal system of D. hansenii cells. We recently identified an endosomal recycling route that bypasses the trans-Golgi network in S. cerevisiae cells (Laidlaw et al., 2022b; MacDonald and Piper, 2017), efficiency of which can be measured by assessing the efflux of endosomal FM4-64 (Wiederkehr et al., 2000). We loaded D. hansenii endosomes with FM4-64 and measured efflux as a decrease in fluorescence by flow cytometry, showing 58.3±3.8% internalised dye is secreted via endosomal recycling during the first 10 min (Fig. 6B). We attribute this to active recycling mechanisms as the pre-treatment of FM4-64 endosome loaded cells with sodium azide, to inhibit active transport processes, entirely inhibits dye efflux.

As holotomography was so efficient at rapid imaging entire cells, we next tested if intracellular refractive index (RI) contrast correlated with the vacuole. We expressed a GFP tagged nuclear marker in S. cerevisiae (GFP-Taf3) and imaged fluorescence simultaneously with holotomography measurements. Highlighting regions of particular RI contrast allowed the vacuole to be clearly distinguished from the GFP tagged nucleus (Fig. 7A). Additionally, membranous elements of the plasma membrane and vacuole could be segmented separately. We then applied similar technique to image D. hansenii cells and this time, without a GFP integrated marker, we labelled the vacuoles with CMAC. Again, based only on refractive indices, the vacuoles of cells could be clearly observed (Fig. 7B). This highlights the power of using holotomography as a label-free technique, that can analyse specific organelles like the vacuole, for rapid imaging in live cells, without any prior genetic manipulations.

Finally, organelles are inherited in budding yeast, with the sequential series of vacuole, mitochondria and nucleus being in part directed from the mother cell to the budding daughter cell in S. cerevisiae (Li et al., 2021). We therefore used the above-described methods to capture these inheritance events in D. hansenii. Using pulse-chased FM4-64, Hoechst, and MitoTracker, these dyes could be used to capture organelle inheritance events of the vacuole, nucleus and mitochondria, respectively (Fig. 8A-C). Indicating that these biological events are also accessible to investigation in D. hansenii.

There is growing interest in harnessing the biotechnological potential of the budding yeast Debaryomyces hansenii. For example, its extreme tolerance to salt and growth in alkaline conditions allows for low-cost sterile production in non-specialised conditions (Prista et al., 1997). D. hansenii metabolism has been shown as a valuable source for an array of products, such as xylitol, trehalose, D-arabitol, and several vitamins (Gírio et al., 2000; Nobre and da Costa, 1985; Voronovsky et al., 2002). Beyond this, as D. hansenii also has a very high capacity to synthesize and accumulate lipids (Merdinger and Devine, 1965), it is an attractive microbial cell factory model system for production of both natural and non-natural chemical products (Breuer and Harms, 2006). Finally, D. hansenii can be used in fermentation processes, due to its osmotolerance and production of desirable flavour additives (Flores et al., 2004). Despite all these applications, little is known about the architecture of D. hansenii cells and study of its cell biology and biochemistry has been limited.

In this study we have made progress in the characterisation of D. hansenii by allowing observation of different biological features: (1) using live cell imaging combined with different fluorescent dyes; (2) employing label-free holotomography; (3) use of antibodies generated against S. cerevisiae proteins for immunodetection of D. hansenii orthologues. Indeed, this study was initiated by comparison of D. hansenii with S. cerevisiae. Although these yeasts are relatively distant in evolutionary terms, and clearly exhibit different sizes and physical features (Figs 1–2), we did identify orthologues of S. cerevisiae proteins that were validated in D. hansenii lysates by immunoblotting (Fig. 3). For example, the ESCRT-like protein required for endosomal recycling Ist1 (Laidlaw et al., 2022b), a pathway we were able to observe in D. hansenii using a lipid dye efflux assay (Fig. 6). Similarly, we were able to immunoblot the E3 ubiquitin ligase Rsp5, a master regulator of endosomal and vacuolar trafficking in S. cerevisiae (MacDonald et al., 2020; Sardana and Emr, 2021), which can be coupled to observations of endosomal trafficking kinetics by pulse chase experiments and vacuolar morphology observations using CMAC, FM4-64 and holotomography (Figs, 4, 6 and 7). Finally, although we were not able to visualise eisosome compartments using fluorescent methods, we were at least able to detect levels of the core eisosome factor Pil1. As Pil1 is post-translationally regulated in response to nutrient stress (Laidlaw et al., 2021; Paine et al., 2023), and eisosomes are generally associated with stress conditions (Athanasopoulos et al., 2019; Babst, 2019), it would be interesting to know whether eisosomes are involved in D. hansenii salt tolerance. These are just some tractable inferences that can be made despite being unable to genetically modify and visualise trafficking pathways and biological processes. Similar approaches would be possible for any cross-species analysis, assuming there was sufficient conservation to monitor protein levels. We note that using antibodies raised against orthologues can lead to misinterpretation if comparing, for example, absolute levels across different species. Indeed, even the proteome profiles from lysates of D. hansenii and S. cerevisiae differ significantly (Fig. 3D). However, these results do give an initial foothold to study such proteins in D. hansenii in different physiological conditions, where comparisons can be interpreted.

We also labelled several intracellular structures clearly with fluorescent dyes. For example, MitoTracker Green efficiently labelled yeast mitochondria in media, allowing for live cell imaging and morphological analyses (Fig. 4). This is important, because modulation of respiratory activity in D. hansenii allows for adaptation to high salt conditions (Garcia-Neto et al., 2017). We also speculate that the high number of cells exhibiting fragmented or punctate mitochondrial morphologies, as opposed to the ribbon structures, might indicated that even our optimised culturing conditions may be inferior and stress inducing. Vacuole labelling was very effective, using both CMAC to label the vacuolar lumen and an FM4-64 pulse-chase protocol to label the limiting membrane of the vacuole (Figs 5,6). In the past we have struggled to effectively label S. cerevisiae nuclei with DAPI staining, resorting to fixation (Paine et al., 2021). Therefore, we tried various optimisation strategies using the DNA stain Hoechst 33342. This was successful with cells at different stages in the cell cycle having clearly identifiable nuclei (Fig. 4). The cell wall staining procedure CFW was highly successful. This suggests the integrity of the D. hansenii cell wall could be ascertained through CFW treatment, as has been done for other fungal species (Hill et al., 2006; Ram et al., 1994). Beyond this, the chitin deposit scars created following asymmetrical divisions of budding daughter cells were easily visualised with different 3D imaging approaches, including Airyscan2 and lattice SIM (Figs 5, Movie 1). Although replicative life span can be assessed through micro-manipulation of dividing cells (Kennedy et al., 1994) simply labelling with CFW and categorising cells based on how many scars are observed is a simple and effective method (Sinclair, 2013) now accessible to D. hansenii experiments. Unlike our other organelle labelling strategies in D. hansenii that allowed for efficient labelling and downstream imaging in media, staining of nucelli was only efficient in Tris buffer at pH 7.2. Although imaging of nuclei at steady state can provide various useful information, it is limited for live cell imaging in media. Although we can observe inheritance of the nuclei, only live cell imaging of FM4-64 labelled vacuoles and MitoTracker Green labelled mitochondria can be tracked in real time. In S. cerevisiae nuclear inheritance has been tracked by live cell imaging following transformation or integration of fluorescently tagged nuclear proteins, such as Nup59 or Nrd1 (Lecinski et al., 2024 preprint; Li et al., 2021). Therefore, it may be that further genetic developments will be required to convincingly track nuclear inheritance in D. hansenii.

In addition to fluorescence microscopy, we show holotomography based refractive index measurement, for example between the cellular contents and the extracellular media, works equally well in rich media compared to minimal media (Fig. 2E). Therefore, holotomography provides a very powerful imaging technique to complement other microscopy approaches, by allowing imaging in commonly used experimental media that is never typically used for imaging. In addition, holotomography allows for assessment of cells based on their stage in the cell cycle. We were able to ascertain whether cells were budding based solely on the sphericity measurements from the analysis software. Holotomography provides an exciting alternative for studying yeast species in conditions that mimic their natural environment, with the additional advantage of imaging cells very rapidly and without the requirement of tagging with fluorescent proteins, which can also potentially perturb the experimental system (Fei et al., 2011; Hoebe et al., 2008).

In summary, we have created a pathway to study many cell biological and biochemical aspects of the yeast species D. hansenii. The approaches we have employed, such as re-purposing antibodies, using fluorescent dyes, and harnessing label free imaging is a very powerful combination. Only a tiny fraction of known bacteria species are amenable to transformation (Johnston et al., 2014), systematic assessment of photosynthetic algae has only recently been achieved using a model species and most non-model species are inaccessible to investigation (Fauser et al., 2022; Mosey et al., 2021), and patient derived primary human cells are inherently limited due to the nature of their source and growth requirements (Richter et al., 2021). Therefore, arguably the most important biological cells for human health and survival remain inaccessible to scientific investigation and methods to overcome these challenges have high potential.

Wild-type strain BY4742 (Brachmann et al., 1998) was used for all Saccharomyces cerevisiae experiments apart from the GFP-Ycp4 strain (Fig. 1C) and GFP-Taf3 (Fig. 7A) tagged strains that are BY4741 background (Weill et al., 2018). The CB1 isolate of Debaryomyces hansenii (Ramos et al., 2017) was used for all other experiments.

Both species of yeast cells were grown in either YPD (1% yeast extract, 2% peptone, 2% dextrose) or synthetic complete (SC) minimal media (2% glucose, 0.675% yeast nitrogen base without amino acids, plus appropriate amino acid dropouts for plasmid selection) (Formedium, Norfolk, UK). S. cerevisiae cells were routinely cultured at 30°C and D. hansenii cells were cultured at 25°C with shaking. Typically, a 2-fold serial dilution of was prepared and grown overnight so that mid-log phase cells could be harvested and used for experimentation the following morning.

For microscopy, S. cerevisiae and D. hansenii cells from log-phase culture were labelled prior to concentration (∼10 fold) in SC media and then visualised at room temperature. For confocal microscopy, a laser scanning Zeiss LSM 980 equipped with an Airyscan 2 detector and a 63× / 1.4 NA objective lens was used. For Structured illumination microscopy (SIM), a Zeiss Elyra 7 system was used with a Plan-Apochromat 40× / 1.4 NA oil objective. Imaging in the blue channel for Hoechst 33342, Calcofluor-white, and CMAC was excited with a 405 nm laser (emission 422-497 nm), imaging in the green channel for GFP and MitoTracker-Green was excited with a 488 nm laser (emission 495-550 nm), and imaging in the red channel for FM4-64 was excited with a 561 nm laser (emission 570-620 nm). Airyscan2 and SIM² processing was performed using Zen Blue (version 3.4.1) and images were modified for brightness and pseudo-colouring using ImageJ software (NIH).

All D. hansenii cells were grown overnight to mid-log phase at 25°C prior to different labelling procedures. Where possible, labelling was performed in SC media, with the exception of Hoechst 33342 that only efficiently labelled yeast nuclei when labelling was performed in Tris buffer at pH 7.2. For nuclear staining, cells were concentrated ∼10 fold and washed once in Tris buffer at pH 7.2. The concentrated yeast sample was resuspended in 100 µl buffer containing 2 µl Hoechst 33342 from a 1 mg/ml frozen aliquot. For mitochondrial staining, 200 µl of culture was incubated with 4 µl MitoTracker Green stock at 100 µM concentration. For CMAC vacuolar staining, 200 µl of culture was incubated with 2 µl CMAC from a 10 mM stock. For cell wall staining, 200 µl of culture was incubated with 2 µl CFW, washed twice then imaged. FM4-64 pulse chase microscopy experiments were pulsed with 8 µM dye for 2 min followed by label-free media chases at indicated time points.

Yeast cells were grown to log phase before imaging by holotomography at room temperature. Most experiments were acquired by 3D quantitative phase imaging (QPI) using an HT-2H instrument (Tomocube Inc), employing Mach–Zehnder interferometry and a digital micromirror device. 2D holographic images were acquired from 48 azimuthally symmetric directions with a polar angle (64.5°) controlled by a digital micromirror device (DMD) and used to construct a 3D tomogram based on refractive index. The diffracted beams from the sample were collected using a high numerical aperture objective lens (NA=1.2, UPLSAP 60XW, Olympus) and recorded using a complementary sCMOS image sensor (Blackfly S BFS-U3-28S5 M, FLIR Systems Inc.). For holotomography coupled to fluorescence imaging, of GFP-Taf3 S. cerevisiae and CMAC-labelled D. hansenii cells, correlative 3D fluorescence images were acquired using the HT-2 T instrument. Blue fluorescence data were acquired with an excitation filter of 392 nm±12 nm (381∼404 nm) and an emission filter of 432 nm±18 nm (414∼450 nm), while the green fluorescence was done with an excitation filter of 474 nm±14 nm (461∼487 nm) and an emission filter of 523 nm±23 nm (500∼546 nm). Semrock filters (IDEX Health & Science LLC) were used. The fluorescence z-stacks were acquired and processed with deconvolution software (AutoQuant by Media Cybernetics). The data was imaged and visualized using a commercial software (TomoStudio version 3.3.9, Tomocube Inc., Korea).

Mid-log phase D. hansenii cells were concentrated 10× in 200 µl YPD containing 40 µM FM4-64 dye {N-(3-Triethylammoniumpropyl)-4-[6-(4-(Diethylamino) Phenyl] (Hexatrienyl) Pyridinium Dibromide} dye (Invitrogen) to label endosomes for 8 min. Cells were then harvested in a pre-chilled centrifuge and supernatant fluid containing FM4-64 discarded. Cells were then washed three times in ice cold SC media. A concentrated sample of cold washed D. hansenii cells (∼20 µl) was then added to a flow cytometer tube and 3 ml of room temperature SC media at room temperature was added. Tube was immediately loaded on to a LSR Fortessa instrument (BD Biosciences) and fluorescence measurements from >1000 events per second were recorded for 10 min. FM4-64 fluorescence during cytometry was excited with a 561 nm laser, and data collected from an emission filter 710-750 nm. Data was visualised by converting fluorescence measurements to a percentage of the average fluorescence recorded in the first 10 s. Flow data were analysed using FCS Express (version 7.06.0015; De Novo Software).

Log phase cultures of D. hansenii and S. cerevisiae were harvested, treated with 0.2 N NaOH for 5 min before resuspension in buffer (8 M urea, 10% glycerol, 50 mM Tris-HCl pH 6.8, 5% SDS, 0.1% bromophenol blue and 10% 2-mercaptoethanol) to generate lysates. Lysates were resolved by SDS-PAGE then transferred to a nitrocellulose membrane using the iBlot2 transfer system (Invitrogen). Ponceau S and Coomassie staining was used to confirm successful transfer and equal loading, respectively. Rabbit polyclonal primary antibodies raised against Rsp5 (Stamenova et al., 2004), Pil1 (Walther et al., 2007), Ist1 (Tan et al., 2015) were individually incubated at 1/1000 concentration in 5% milk PBST (Phosphate-Buffered Saline, 0.1% Tween) overnight with rotation at 4°C. Six 5 min washes in PBST were performed before incubation with 1/5000 IgG anti-rabbit secondary antibodies conjugated to HRP (Thermo Fisher Scientific) for 1 h. Membranes were again washed six times 5 mins prior to visualisation. Membranes were visualised using a ChemiDoc Imager (Bio-Rad) in combination with enhanced chemiluminescence (ECL) Super Signal Pico Plus (Thermo Fisher Scientific).

For cell area measurements, segmentation from brightfield micrographs was performed using the Cell Magic Wand Plugin for ImageJ. For holotomography analyses, cells were manually segmented and parameters isolated from the refractive index measurements using TomoStudio analysis Software version 3.3.9 (Tomocube Inc). Unpaired Student's t-tests were performed using GraphPad Prism version 10.1.0 to compare the statistical significance between experimental conditions, with P-values included. An asterisk (*) is used to denote significance P<0.05.

We would like to thank staff at the York Bioscience Technology Facility for technical assistance. This research was supported by a Sir Henry Dale Research Fellowship from the Wellcome Trust and the Royal Society 204636/Z/16/Z (C.M.).