Lipid droplets (LDs) accumulate in response to diverse cellular stresses. However, their regulation and physiological roles remain poorly understood in most contexts. Here, we show that in budding yeast, chronic hyperosmotic stress induces sustained LD accumulation. Unlike the transient LD response observed during acute osmotic shock, chronic stress triggers prolonged, Dga1-dependent triacylglycerol synthesis. In the absence of triacylglycerol synthesis, cellular fitness is impaired. Lipidomic profiling reveals extensive membrane remodeling during chronic hyperosmotic stress, most notably an increase in phosphatidylcholine (PC) resulting in changes to membrane properties. In LD-deficient cells, the hyperosmotic stress–induced PC increase is blunted, while enhancing PC synthesis improves the fitness of triacylglycerol-deficient cells. Thus, LD accumulation and phospholipid remodeling underlie an adaptive response to chronic hyperosmotic stress.

Lipid droplets (LDs) are unique lipid storage organelles present in most eukaryotic cells (Mathiowetz and Olzmann, 2024; Farese and Walther, 2025; Klemm and Carvalho, 2024). They are composed of a phospholipid monolayer surrounding a hydrophobic core of neutral lipids, primarily triacylglycerols (TAGs) and sterol esters (STEs). TAGs are synthesized through the acylation of diacylglycerol (DAG) with a fatty acid, either derived from an activated acyl-CoA or transferred from a donor phospholipid (Athenstaedt, 2010). STEs are generated by esterifying sterols with acylated fatty acids (Athenstaedt, 2010). LD biogenesis occurs by an evolutionarily conserved mechanism in which seipin oligomers concentrate neutral lipids within the ER bilayer, thereby facilitating their condensation into lens structures that subsequently bud toward the cytoplasm to originate a mature LD (Klug et al., 2024; Thiam and Ikonen, 2021).

Proper LD assembly and regulation are essential for cellular and organismal health by maintaining membrane homeostasis and energy balance (Zadoorian et al., 2023). Defects in LD formation result in lipodystrophies, best exemplified by the seipinopathies (Cartwright and Goodman, 2012), while excessive LD formation is the defining feature of fatty liver disease (Heeren and Scheja, 2021) and is further associated with poor cancer prognosis (Quan et al., 2025). Because many environmental stresses perturb membrane composition and biophysical properties, LD accumulation is a common adaptive response (Pressly et al., 2022; Jin et al., 2023; Safi et al., 2024). Lipids stored in LDs contribute to ATP synthesis and cell adaptation to nutrient limitation (Seo et al., 2017; Rambold et al., 2015; Hariri et al., 2018; Rogers et al., 2021). When membrane expansion is required, such as upon growth resumption from quiescence, LDs can be rapidly turned over to supply lipids for membrane biogenesis (Markgraf et al., 2014). Conversely, recent evidence shows that LDs can buffer excess lipids and membranes (Phan et al., 2025; Wang et al., 2018) and sequester membrane lipids that are susceptible to oxidation (Lange et al., 2025), or otherwise compromise organelle integrity (Nguyen and Olzmann, 2017). Fluctuations in temperature alter membrane properties (Ernst et al., 2016), and the inability to make LDs renders cells sensitive to changes in temperature (Wang et al., 2018). Together, these findings suggest that LDs enable cells to remodel membrane composition and maintain membrane homeostasis under stress conditions.

Here, building on a high-throughput approach to quantify LD abundance in budding yeast, we identified a previously unrecognized LD response to chronic hyperosmotic stress that is required for cellular fitness. Recent reports have demonstrated rapid and transient LD accumulation upon acute (<30 min) hyperosmotic stress (Phan et al., 2025), but longer term adaptations of membranes and organelles to allow cell proliferation are less well studied. We show that LD accumulation is important for proliferation under sustained, chronic (>3 h) hyperosmotic stress. This response occurs in parallel to, but is not directly controlled by, the canonical hyperosmotic adaptation machinery and is necessary for modulating membrane properties. LD accumulation is accompanied by increased phosphatidylcholine (PC) levels, an increase that is blunted in the absence of LDs. Promoting PC synthesis alleviates the requirement for LD formation during chronic hyperosmotic stress. Together, our findings identify LD accumulation as a key adaptive response to chronic hyperosmotic stress, contributing to membrane remodeling and supporting cellular fitness.

Pln1 localizes to the surface of LDs and is rapidly degraded when not bound to LDs (Gao et al., 2017). Fluorescently tagged Pln1 is therefore commonly used as an LD marker (Fig. S1 A) (Gao et al., 2017; Giménez-Andrés et al., 2021; Wang et al., 2024), and its levels are a good surrogate for LD abundance (Rao et al., 2025, Preprint; Gao et al., 2017). To investigate LD accumulation dynamics in a high-throughput manner, we measure the levels of endogenously tagged Pln1-mCherry using flow cytometry. During exponential growth, Pln1-mCherry levels sharply increase as cells approach the stationary phase, mirroring LD abundance as detected by fluorescence microscopy of BODIPY-stained cells (Fig. S1, A and B).

+ Expand view − Collapse view
Figure S1.
Figure S1. Refer to the image caption for details. Panel A shows fluorescence microscopy images of cells stained with BODIPY and expressing Pln1-mCherry. The images are divided into two rows based on optical density: optical density less than 1 and optical density greater than 2. The images show the co-localization of Pln1-mCherry with lipid droplets. Panel B presents a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells grown to varying densities, grouped by ranges of absorbance at optical density at 600 nanometers. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel C displays a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells treated or not treated with 200 micromolar rapamycin for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel D shows a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to temperature shifts from 30 degrees Celsius for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel E presents a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to nutrient restriction for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel F displays a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to 4 millimolar hydrogen peroxide exposure for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel G shows a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to 1 molar potassium chloride hyperosmotic stress for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel H presents a heatmap of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to varying times and severities of potassium chloride treatment in SCD media. The color scale indicates the mean values from three biological replicates. Panel I displays a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells in either SCD or YPD subjected to 1 molar potassium chloride treatment for various times. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel J shows a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to 1.5 molar sorbitol treatment in SCD media for 3 hours. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles.

Screening of LD responses to stress. (A) Pln1-mCherry colocalization with BODIPY-stained LDs in exponential (OD < 1)- and stationary-phase (OD > 2) cells. Z-projection images are representative of multiple replicates. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells grown to varying densities, measured and grouped by the indicated ranges of absorbance at OD600 nm. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 22 (<0.49), 7 (0.5–0.99), 19 (1–1.99), 12 (>2). (C) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells treated or not with 200 µM rapamycin for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (D) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to the indicated temperature shifts from 30°C for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (E) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to the indicated nutrient restriction for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (F) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to 4 mM H2O2 exposure for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (G) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to 1M KCl hyperosmotic stress for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (H) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to varying times and severities of KCl treatment in SCD media. UT = untreated. The color scale indicates the mean values from three biological replicates. (I) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells in either SCD or YPD subjected to 1M KCl treatment for the indicated times. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (J) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to 1.5M sorbitol treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4.

Figure S1.
A multi-panel image depicts the localization and fluorescence of Pln1-mCherry in cells under various conditions. Panel A shows fluorescence microscopy images of cells stained with BODIPY and expressing Pln1-mCherry. The images are divided into two rows based on optical density: optical density less than 1 and optical density greater than 2. The images show the co-localization of Pln1-mCherry with lipid droplets. Panel B presents a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells grown to varying densities, grouped by ranges of absorbance at optical density at 600 nanometers. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel C displays a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells treated or not treated with 200 micromolar rapamycin for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel D shows a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to temperature shifts from 30 degrees Celsius for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel E presents a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to nutrient restriction for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel F displays a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to 4 millimolar hydrogen peroxide exposure for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel G shows a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to 1 molar potassium chloride hyperosmotic stress for various time points. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel H presents a heatmap of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to varying times and severities of potassium chloride treatment in SCD media. The color scale indicates the mean values from three biological replicates. Panel I displays a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells in either SCD or YPD subjected to 1 molar potassium chloride treatment for various times. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles. Panel J shows a violin plot of the median Pln1-mCherry fluorescence measured by flow cytometry from populations of single wild-type cells subjected to 1.5 molar sorbitol treatment in SCD media for 3 hours. The plot shows the spread of data, with dashed lines indicating the median and dotted lines indicating the upper and lower quartiles.

Screening of LD responses to stress. (A) Pln1-mCherry colocalization with BODIPY-stained LDs in exponential (OD < 1)- and stationary-phase (OD > 2) cells. Z-projection images are representative of multiple replicates. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells grown to varying densities, measured and grouped by the indicated ranges of absorbance at OD600 nm. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 22 (<0.49), 7 (0.5–0.99), 19 (1–1.99), 12 (>2). (C) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells treated or not with 200 µM rapamycin for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (D) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to the indicated temperature shifts from 30°C for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (E) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to the indicated nutrient restriction for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (F) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to 4 mM H2O2 exposure for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (G) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to 1M KCl hyperosmotic stress for the indicated time points. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (H) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to varying times and severities of KCl treatment in SCD media. UT = untreated. The color scale indicates the mean values from three biological replicates. (I) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells in either SCD or YPD subjected to 1M KCl treatment for the indicated times. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (J) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells subjected to 1.5M sorbitol treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4.

Close Figure S1.

To further validate the flow cytometry–based results, we used pharmacological inhibition of TORC1 with rapamycin, a treatment known to induce LD accumulation (Madeira et al., 2015; Teixeira et al., 2021; Liu et al., 2019; Hosios et al., 2022; Dubots et al., 2014). As expected, rapamycin treatment of wild-type (WT) cells resulted in a rapid and sustained increase of Pln1-mCherry levels (Fig. S1 C). Therefore, flow cytometry analysis of Pln1-mCherry levels provides a sensitive and accurate readout of LD accumulation.

Chronic hyperosmotic stress induces LD accumulation

We next monitored the levels of Pln1-mCherry over time in exponentially growing cells subjected to a variety of stresses: oxidative, temperature, nutrient, and hyperosmotic. Upon moderate temperature shifts (from 30 to 25°C or 37°C) or acute restriction of glucose or amino acids (2–0.1% glucose or a 10-fold reduction in amino acid levels), cells showed minimal changes in Pln1-mCherry levels over a 6-h period (Fig. S1, D and E). Oxidative stress induced by H2O2 resulted in a mild and gradual Pln1-mCherry accumulation (Fig. S1 F). In contrast, hyperosmotic stress induced by KCl resulted in a rapid increase in Pln1-mCherry accumulation that peaked at 3 h after stress induction and remained high for at least 24 h (Fig. 1, A and B; and Fig. S1 G). Moreover, the increase in Pln1-mCherry levels triggered by KCl-induced hyperosmotic stress was dose- and time-dependent (Fig. S1 H), and occurred in both rich and defined media (Fig. S1 I). Moreover, a similar increase in Pln1-mCherry levels was observed when cells were grown in sorbitol, a nonionic osmolyte (Fig. S1 J). Importantly, a reduction of Pln1-mCherry to baseline levels was observed within 3 h upon removal of the hyperosmotic stress showing this response was reversible (Fig. 1 C). Finally, we confirmed that the increase in Pln1-mCherry levels upon hyperosmotic stress was due to LD accumulation by imaging of BODIPY-labeled cells (Fig. 1 D). This analysis showed that KCl treatment increased both the number and size of LDs in cells (Fig. 1, E and F). Together, these results demonstrate that chronic hyperosmotic stress triggers long-lasting but reversible LD accumulation.

Figure 1.
A multi-panel image depicts lipid droplet accumulation under hyperosmotic stress. Panel A shows histograms of mCherry fluorescence in Pln1-mCherry tagged cells treated with 1 molar potassium chloride for different time points. Panel B presents violin plots of median Pln1-mCherry fluorescence measured by flow cytometry in single cells subjected to 1 molar potassium chloride treatment for various times. Panel C displays violin plots of median Pln1-mCherry fluorescence in cells subjected to 16 hours of 1 molar potassium chloride treatment followed by a washout. Panel D includes microscopy images of lipid droplets stained with BODIPY and imaged by live cell microscopy after varying lengths of 1 molar potassium chloride treatment. Panel E shows violin plots analyzing lipid droplets per cell from samples used in Panel C. Panel F presents violin plots analyzing lipid droplet diameters from samples used in Panel C. Panel G is a schematic of lipid droplet formation. Panel H shows a growth assay of wild-type and mutant cells in SCD and SCD plus 1 molar potassium chloride, with the main line indicating the mean and error bars showing standard error of the mean from four biological replicates.

LD accumulation is a chronic response to hyperosmotic stress. (A) Representative histograms obtained from 10,000 Pln1-mCherry–tagged cells grown in SCD + 1M KCl for the indicated time points. The untagged Pln1 shows the background fluorescence, while UT is untreated. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single cells subjected to 1M KCl treatment in SCD media for the indicated time. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7. (C) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single cells subjected to 16 h of 1M KCl treatment in SCD media which was then washed out and replaced with SCD + 1M KCl (mock washout) or SCD without 1M KCl for 3 h before measurement. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (D) LDs stained with BODIPY and imaged by live-cell microscopy after varying lengths of 1M KCl treatment in SCD. Z-projection images are representative of multiple replicates. UT = untreated. (E) Analysis of LDs per cell from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 85 (UT), 89 (3 h KCl), 64 (6 h KCl) cells from three biological replicates. (F) Analysis of LD diameters from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 165 (UT), 626 (3 h KCl), 427 (6 h KCl) LDs from three biological replicates. (G) Schematic of LD formation. Dga1 or Lro1 converts DAG into TAG, while Are1 or Are2 converts sterols into STEs. TAGs and STEs are concentrated and packaged into LDs by the seipin complex. (H) Growth assay of WT and are1Δare2Δdga1Δlro1Δ (LDΔ) cells in SCD and SCD + 1M KCl. The main line shows the mean, with the error bars showing SEM from four biological replicates.

Figure 1.
A multi-panel image depicts lipid droplet accumulation under hyperosmotic stress. Panel A shows histograms of mCherry fluorescence in Pln1-mCherry tagged cells treated with 1 molar potassium chloride for different time points. Panel B presents violin plots of median Pln1-mCherry fluorescence measured by flow cytometry in single cells subjected to 1 molar potassium chloride treatment for various times. Panel C displays violin plots of median Pln1-mCherry fluorescence in cells subjected to 16 hours of 1 molar potassium chloride treatment followed by a washout. Panel D includes microscopy images of lipid droplets stained with BODIPY and imaged by live cell microscopy after varying lengths of 1 molar potassium chloride treatment. Panel E shows violin plots analyzing lipid droplets per cell from samples used in Panel C. Panel F presents violin plots analyzing lipid droplet diameters from samples used in Panel C. Panel G is a schematic of lipid droplet formation. Panel H shows a growth assay of wild-type and mutant cells in SCD and SCD plus 1 molar potassium chloride, with the main line indicating the mean and error bars showing standard error of the mean from four biological replicates.

LD accumulation is a chronic response to hyperosmotic stress. (A) Representative histograms obtained from 10,000 Pln1-mCherry–tagged cells grown in SCD + 1M KCl for the indicated time points. The untagged Pln1 shows the background fluorescence, while UT is untreated. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single cells subjected to 1M KCl treatment in SCD media for the indicated time. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7. (C) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single cells subjected to 16 h of 1M KCl treatment in SCD media which was then washed out and replaced with SCD + 1M KCl (mock washout) or SCD without 1M KCl for 3 h before measurement. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (D) LDs stained with BODIPY and imaged by live-cell microscopy after varying lengths of 1M KCl treatment in SCD. Z-projection images are representative of multiple replicates. UT = untreated. (E) Analysis of LDs per cell from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 85 (UT), 89 (3 h KCl), 64 (6 h KCl) cells from three biological replicates. (F) Analysis of LD diameters from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 165 (UT), 626 (3 h KCl), 427 (6 h KCl) LDs from three biological replicates. (G) Schematic of LD formation. Dga1 or Lro1 converts DAG into TAG, while Are1 or Are2 converts sterols into STEs. TAGs and STEs are concentrated and packaged into LDs by the seipin complex. (H) Growth assay of WT and are1Δare2Δdga1Δlro1Δ (LDΔ) cells in SCD and SCD + 1M KCl. The main line shows the mean, with the error bars showing SEM from four biological replicates.

Close Figure 1.

LD accumulation is important for cell fitness under hyperosmotic stress

Hyperosmotic stress causes rapid water egress and a drop in plasma membrane tension (Roffay et al., 2021; Phan et al., 2025). The drop in plasma membrane tension leads to acute responses to hyperosmotic stress. These include temporary LD accumulation, dependent upon efficient endocytosis and the lipid transfer proteins Osh2 and Osh3 (Phan et al., 2025). Plasma membrane tension is, however, typically restored within 30–60 min of stress onset (Klipp et al., 2005; Shen et al., 2023; Petelenz-Kurdziel et al., 2013), long before the phenotypes we observe. We therefore hypothesized that LD accumulation may be important for long-term fitness under persistent hyperosmotic stress, beyond its role in restoring plasma membrane tension.

To test this hypothesis, we examined mutants defective in LD biogenesis. LD biogenesis depends on neutral lipid synthesis and is spatially controlled by the seipin complex (Fig. 1 G). Seipin mutants produce aberrant LDs that are defective in recruiting surface proteins including Pln1-mCherry (Grippa et al., 2015). Consistent with this, seipin mutants showed reduced Pln1-mCherry accumulation under hyperosmotic stress (Fig. S2 A), and most cells had only one or two large LDs (Fig. S2, B and C). Since the existing LDs expanded under stress, neutral lipids were still produced (Fig. S2 D). Of note, the seipin-dependent changes in LD size did not compromise cell fitness under hyperosmotic conditions (Fig. S2 E). We next tested the importance of stress-induced neutral lipid synthesis using a strain that lacks all four relevant enzymes (Dga1 and Lro1, which produce TAGs, and Are1 and Are2, which synthesize STEs, referred to here as LDΔ) (Oelkers et al., 2002; Sandager et al., 2002) and consequently is devoid of LDs (Fig. S2, F and G). Unlike seipin mutants, LDΔ cells exhibited a pronounced growth defect under hyperosmotic stress (Fig. 1 H). Together, these findings indicate that a process related to LD accumulation, rather than normal LD morphology per se, is required for adaptation to persistent hyperosmotic stress.

+ Expand view − Collapse view
Figure S2.
Figure S2. Refer to the image caption for details. Panel A shows a violin plot comparing median Pln1-mCherry fluorescence in wild-type (WT) and sei1 cells under untreated and 1 molar potassium chloride conditions. The y-axis represents fluorescence in arbitrary units. Panel B displays microscopy images of lipid droplets stained with BODIPY in WT and sei1 cells after 3 hours of 1 molar potassium chloride treatment. Panel C presents a violin plot of lipid droplets per cell, with the y-axis indicating the number of droplets. Panel D shows a violin plot of lipid droplet diameters, with the y-axis in micrometers. Panel E is a line graph of cell count over time for WT and sei1 cells in SCD and SCD plus 1 molar potassium chloride, with the y-axis representing optical density at 600 nanometers. Panel F shows microscopy images of lipid droplets in WT and LD cells after 3 hours of 1 molar potassium chloride treatment. Panel G presents a violin plot of lipid droplets per cell in WT and LD cells. Panel H shows a violin plot of median Pln1-mCherry fluorescence in WT, dga1lro1, and are1are2 cells. Panel I displays microscopy images of lipid droplets in WT, dga1lro1, and are1are2 cells after 3 hours of 1 molar potassium chloride treatment. Panel J presents a violin plot of lipid droplets per cell in these cell types. Panel K shows a violin plot of lipid droplet diameters in WT, dga1lro1, and are1are2 cells. Panel L presents a violin plot of median Pln1-mCherry fluorescence in WT, are1are2, dga1lro1, dga1, and lro1 cells under 1.5 molar sorbitol treatment. Panel M shows a violin plot of median Pln1-mCherry fluorescence in WT and osh2osh3 cells under 1 molar potassium chloride treatment. Panel N presents a violin plot of median Pln1-mCherry fluorescence in WT and ede1 cells under 1 molar potassium chloride treatment.

LD responses to hyperosmotic stress. (A) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and sei1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (B) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT and sei1Δ cells. Z-projection images are representative of multiple replicates. (C) Analysis of LDs per cell from samples used in B. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 95 (WT UT), 107 (sei1Δ UT), 104 (WT KCl), 92 (sei1Δ KCl) cells from three biological replicates. (D) Analysis of LD diameters from samples used in B. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 251 (WT UT), 276 (sei1Δ UT), 253 (WT KCl), 230 (sei1Δ KCl) LDs from three biological replicates. (E) Growth assay of WT and sei1Δ cells in SCD and SCD + 1M KCl. The graph shows the mean and SEM from four biological replicates. (F) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for WT and are1Δare2Δdga1Δlro1Δ (LDΔ) cells. Z-projection images are representative of multiple replicates. (G) Analysis of LDs per cell from samples in F. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 103 (WT UT), 96 (LDΔ UT), 97 (WT KCl), 108 (LDΔ KCl) cells from three biological replicates. (H) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, dga1Δlro1Δ, and are1Δare2Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (I) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT, dga1Δlro1Δ, and are1Δare2Δ cells. Z-projection images are representative of multiple replicates. (J) Analysis of LDs per cell from samples in I. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 124 (WT UT), 121 (dga1Δlro1Δ UT), 125 (are1Δare2Δ UT), 120 (WT KCl), 113 (dga1Δlro1Δ KCl), 78 (are1Δare2Δ KCl) cells from 4 biological replicates. (K) Analysis of LD diameters from samples in I. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 414 (WT UT), 173 (dga1Δlro1Δ UT), 250 (are1Δare2Δ UT), 504 (WT KCl), 218 (dga1Δlro1Δ KCl), 262 (are1Δare2Δ KCl) LDs from four biological replicates. (L) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, are1Δare2Δ, dga1Δlro1Δ, dga1Δ, and lro1Δ cells subjected to 1.5M sorbitol treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (M) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and osh2Δosh3Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (N) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and ede1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5.

Figure S2.
A multi-panel image depicts lipid droplet responses to hyperosmotic stress. Panel A shows a violin plot comparing median Pln1-mCherry fluorescence in wild-type (WT) and sei1 cells under untreated and 1 molar potassium chloride conditions. The y-axis represents fluorescence in arbitrary units. Panel B displays microscopy images of lipid droplets stained with BODIPY in WT and sei1 cells after 3 hours of 1 molar potassium chloride treatment. Panel C presents a violin plot of lipid droplets per cell, with the y-axis indicating the number of droplets. Panel D shows a violin plot of lipid droplet diameters, with the y-axis in micrometers. Panel E is a line graph of cell count over time for WT and sei1 cells in SCD and SCD plus 1 molar potassium chloride, with the y-axis representing optical density at 600 nanometers. Panel F shows microscopy images of lipid droplets in WT and LD cells after 3 hours of 1 molar potassium chloride treatment. Panel G presents a violin plot of lipid droplets per cell in WT and LD cells. Panel H shows a violin plot of median Pln1-mCherry fluorescence in WT, dga1lro1, and are1are2 cells. Panel I displays microscopy images of lipid droplets in WT, dga1lro1, and are1are2 cells after 3 hours of 1 molar potassium chloride treatment. Panel J presents a violin plot of lipid droplets per cell in these cell types. Panel K shows a violin plot of lipid droplet diameters in WT, dga1lro1, and are1are2 cells. Panel L presents a violin plot of median Pln1-mCherry fluorescence in WT, are1are2, dga1lro1, dga1, and lro1 cells under 1.5 molar sorbitol treatment. Panel M shows a violin plot of median Pln1-mCherry fluorescence in WT and osh2osh3 cells under 1 molar potassium chloride treatment. Panel N presents a violin plot of median Pln1-mCherry fluorescence in WT and ede1 cells under 1 molar potassium chloride treatment.

LD responses to hyperosmotic stress. (A) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and sei1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (B) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT and sei1Δ cells. Z-projection images are representative of multiple replicates. (C) Analysis of LDs per cell from samples used in B. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 95 (WT UT), 107 (sei1Δ UT), 104 (WT KCl), 92 (sei1Δ KCl) cells from three biological replicates. (D) Analysis of LD diameters from samples used in B. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 251 (WT UT), 276 (sei1Δ UT), 253 (WT KCl), 230 (sei1Δ KCl) LDs from three biological replicates. (E) Growth assay of WT and sei1Δ cells in SCD and SCD + 1M KCl. The graph shows the mean and SEM from four biological replicates. (F) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for WT and are1Δare2Δdga1Δlro1Δ (LDΔ) cells. Z-projection images are representative of multiple replicates. (G) Analysis of LDs per cell from samples in F. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 103 (WT UT), 96 (LDΔ UT), 97 (WT KCl), 108 (LDΔ KCl) cells from three biological replicates. (H) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, dga1Δlro1Δ, and are1Δare2Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (I) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT, dga1Δlro1Δ, and are1Δare2Δ cells. Z-projection images are representative of multiple replicates. (J) Analysis of LDs per cell from samples in I. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 124 (WT UT), 121 (dga1Δlro1Δ UT), 125 (are1Δare2Δ UT), 120 (WT KCl), 113 (dga1Δlro1Δ KCl), 78 (are1Δare2Δ KCl) cells from 4 biological replicates. (K) Analysis of LD diameters from samples in I. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 414 (WT UT), 173 (dga1Δlro1Δ UT), 250 (are1Δare2Δ UT), 504 (WT KCl), 218 (dga1Δlro1Δ KCl), 262 (are1Δare2Δ KCl) LDs from four biological replicates. (L) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, are1Δare2Δ, dga1Δlro1Δ, dga1Δ, and lro1Δ cells subjected to 1.5M sorbitol treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (M) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and osh2Δosh3Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (N) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and ede1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5.

Close Figure S2.

Hyperosmotic LD accumulation is driven by Dga1-mediated TAG synthesis

We next assessed the individual contribution of TAG and STE synthesis to cell fitness under hyperosmotic stress. Simultaneous deletion of ARE1 and ARE2, which prevent STE synthesis, did not impact cell growth either in the absence or in the presence of hyperosmotic stress (Fig. 2 A). In contrast, preventing TAG synthesis with dga1Δlro1Δ cells resulted in a growth defect specifically under hyperosmotic stress (Fig. 2 A). Deletion of DGA1, but not LRO1, alone also resulted in a growth defect relative to the WT cells (Fig. 2 A). Therefore, TAG synthesis is required for adaptation to chronic hyperosmotic stress with Dga1 playing a major role.

Figure 2.
A multi-panel image depicts the impact of different genetic deletions on cell growth, lipid droplet formation, and lipid droplet size under hyperosmotic stress. Panel A shows two line graphs depicting cell count over time in hours for different cell types, including WT, are1are2, dga1lro1, dga1, and lro1, under 0 molar potassium chloride and 1 molar potassium chloride conditions. The x-axis represents time in hours, and the y-axis represents cell count as optical density at 600 nanometers. The graphs show that the presence of 1 molar potassium chloride affects cell growth differently across the cell types. Panel B presents a violin plot showing the median Pln1-mCherry fluorescence measured by flow cytometry for different cell types subjected to 1 molar potassium chloride treatment for 3 hours. The x-axis lists the cell types, and the y-axis shows the fluorescence signal in arbitrary units. The plot indicates variations in fluorescence levels among the cell types. Panel C contains microscopy images of lipid droplets stained with BODIPY and imaged by live cell microscopy after 3 hours of 1 molar potassium chloride treatment for different cell types. The images show the distribution and intensity of lipid droplets. Panel D features a violin plot analyzing the number of lipid droplets per cell from the samples used in Panel C. The x-axis lists the cell types and conditions, untreated and treated with potassium chloride, and the y-axis shows the number of lipid droplets per cell. The plot highlights differences in lipid droplet counts among the cell types and conditions. Panel E displays a violin plot analyzing the diameters of lipid droplets from the samples used in Panel C. The x-axis lists the cell types and conditions, and the y-axis shows the lipid droplet diameter in micrometers. The plot reveals variations in lipid droplet sizes among the cell types and conditions. Panel F shows microscopy images visualizing Dga1-mNeonGreen (Dga1-mNG) and lipid droplets stained with MDH in live cells with and without 3 hours of treatment with 1 molar potassium chloride. The images are representative of multiple replicates and show the localization and intensity of Dga1-mNG and lipid droplets.

Dga1-produced TAGs drive chronic hyperosmotic LD accumulation. (A) Growth assay of WT, are1Δare2Δ (a1Δa2Δ), dga1Δlro1Δ (d1Δl1Δ), dga1Δ, and lro1Δ cells in SCD and SCD + 1M KCl. Error bars show SEM from four biological replicates. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, dga1Δlro1Δ, dga1Δ, and lro1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7. (C) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT, dga1Δlro1Δ, dga1Δ, and lro1Δ cells. Z-projection images are representative of multiple replicates. (D) Analysis of LDs per cell from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 124 (WT UT), 125 (dga1Δlro1Δ UT), 130 (dga1Δ UT), 108 (lro1Δ UT), 123 (WT KCl), 116 (dga1Δlro1Δ KCl), 135 (dga1Δ KCl), 115 (lro1Δ KCl) cells from four biological replicates. (E) Analysis of LD diameters from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 326 (WT UT), 141 (dga1Δlro1Δ UT), 379 (dga1Δ UT), 187 (lro1Δ UT), 514 (WT KCl), 275 (dga1Δlro1Δ KCl), 388 (dga1Δ KCl), 373 (lro1Δ KCl) LDs from four biological replicates. (F) Visualization of Dga1-mNG and LDs (stained using MDH) in live cells with and without 3 h of treatment with 1M KCl in SCD. UT = untreated. Images are representative of multiple replicates.

Figure 2.
A multi-panel image depicts the impact of different genetic deletions on cell growth, lipid droplet formation, and lipid droplet size under hyperosmotic stress. Panel A shows two line graphs depicting cell count over time in hours for different cell types, including WT, are1are2, dga1lro1, dga1, and lro1, under 0 molar potassium chloride and 1 molar potassium chloride conditions. The x-axis represents time in hours, and the y-axis represents cell count as optical density at 600 nanometers. The graphs show that the presence of 1 molar potassium chloride affects cell growth differently across the cell types. Panel B presents a violin plot showing the median Pln1-mCherry fluorescence measured by flow cytometry for different cell types subjected to 1 molar potassium chloride treatment for 3 hours. The x-axis lists the cell types, and the y-axis shows the fluorescence signal in arbitrary units. The plot indicates variations in fluorescence levels among the cell types. Panel C contains microscopy images of lipid droplets stained with BODIPY and imaged by live cell microscopy after 3 hours of 1 molar potassium chloride treatment for different cell types. The images show the distribution and intensity of lipid droplets. Panel D features a violin plot analyzing the number of lipid droplets per cell from the samples used in Panel C. The x-axis lists the cell types and conditions, untreated and treated with potassium chloride, and the y-axis shows the number of lipid droplets per cell. The plot highlights differences in lipid droplet counts among the cell types and conditions. Panel E displays a violin plot analyzing the diameters of lipid droplets from the samples used in Panel C. The x-axis lists the cell types and conditions, and the y-axis shows the lipid droplet diameter in micrometers. The plot reveals variations in lipid droplet sizes among the cell types and conditions. Panel F shows microscopy images visualizing Dga1-mNeonGreen (Dga1-mNG) and lipid droplets stained with MDH in live cells with and without 3 hours of treatment with 1 molar potassium chloride. The images are representative of multiple replicates and show the localization and intensity of Dga1-mNG and lipid droplets.

Dga1-produced TAGs drive chronic hyperosmotic LD accumulation. (A) Growth assay of WT, are1Δare2Δ (a1Δa2Δ), dga1Δlro1Δ (d1Δl1Δ), dga1Δ, and lro1Δ cells in SCD and SCD + 1M KCl. Error bars show SEM from four biological replicates. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, dga1Δlro1Δ, dga1Δ, and lro1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7. (C) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT, dga1Δlro1Δ, dga1Δ, and lro1Δ cells. Z-projection images are representative of multiple replicates. (D) Analysis of LDs per cell from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 124 (WT UT), 125 (dga1Δlro1Δ UT), 130 (dga1Δ UT), 108 (lro1Δ UT), 123 (WT KCl), 116 (dga1Δlro1Δ KCl), 135 (dga1Δ KCl), 115 (lro1Δ KCl) cells from four biological replicates. (E) Analysis of LD diameters from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 326 (WT UT), 141 (dga1Δlro1Δ UT), 379 (dga1Δ UT), 187 (lro1Δ UT), 514 (WT KCl), 275 (dga1Δlro1Δ KCl), 388 (dga1Δ KCl), 373 (lro1Δ KCl) LDs from four biological replicates. (F) Visualization of Dga1-mNG and LDs (stained using MDH) in live cells with and without 3 h of treatment with 1M KCl in SCD. UT = untreated. Images are representative of multiple replicates.

Close Figure 2.

Since the growth assays reveal the importance of TAG synthesis for long-term adaptation, we monitored Pln1-mCherry to investigate LD accumulation upon hyperosmotic stress. Consistent with the growth assays, are1Δare2Δ cells displayed no changes in LD levels compared with the WT cells using Pln1-mCherry or BODIPY as markers (Fig. S2, H–K). However, the LD response was strongly blunted in dga1Δlro1Δ cells (Fig. 2, B–E and Fig. S2, H−K). In contrast to acute hyperosmotic stress (Phan et al., 2025), Lro1 was dispensable for LD accumulation induced by chronic hyperosmotic stress (Fig. 2, B–E). On the other hand, dga1Δ cells showed much less Pln1-mCherry, comparable to that observed in dga1Δlro1Δ cells (Fig. 2 B). The predominant role of Dga1 was also seen when stress was induced with the nonionic osmolyte sorbitol (Fig. S2 L). The defect in LD accumulation in dga1Δ cells was further confirmed by imaging LDs after BODIPY labeling (Fig. 2, C–E). Therefore, Dga1 plays a major role in LD accumulation during chronic hyperosmotic stress, suggesting that LDs formed during chronic stress are regulated differently from those produced during the Lro1-dominated acute phase (Phan et al., 2025).

Supporting this being a distinct response, deletion of the endocytic protein Ede1 or the lipid transfer proteins Osh2 and Osh3, required for the acute phase LD formation (Phan et al., 2025), had no impact on Pln1-mCherry accumulation following 3 h of hyperosmotic stress (Fig. S2, M and N). This difference likely reflects distinct upstream cues and physiological functions during acute and chronic responses to hyperosmotic stress.

Dga1 uses DAG and CoA-activated fatty acids to produce TAG (Fig. S3 A) (Oelkers et al., 2002). As an independent test on the important role of Dga1, we restricted the availability of its substrates. Deletion of Nem1 or Faa1 and Faa4, which reduce the levels of DAG (Su et al., 2014) and activated fatty acids (Johnson et al., 1994), respectively, impaired LD accumulation under hyperosmotic stress (Fig. S3, B–J), consistent with a key role for Dga1 in this adaptive response.

+ Expand view− Collapse view
Figure S3.
Figure S3. Refer to the image caption for details.Panel A shows a schematic of lipid metabolism, with Nem1, phosphatidic acid, free fatty acids, Faa1 and Faa4, diacylglycerol, acyl fatty acid, Dga1, and triacylglycerol labeled. Panel B shows violin plots of Pln1-mCherry fluorescence signal in wild-type and nem1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with fluorescence measured in arbitrary units and a p-value of 0.0158. Panel C shows BODIPY stained microscopy images of wild-type and nem1 deletion cells after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel D shows violin plots of lipid droplets per cell for wild-type and nem1 deletion cells under untreated and 1 molar potassium chloride treatment, with p-values less than 0.0001. Panel E shows violin plots of lipid droplet diameter in micrometers for wild-type and nem1 deletion cells under untreated and 1 molar potassium chloride treatment, with p-values of 0.0191 and 0.0003. Panel F shows violin plots of Pln1-mCherry fluorescence signal in wild-type, faa1 deletion, and faa4 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment. Panel G shows violin plots of Pln1-mCherry fluorescence signal in wild-type and faa1/faa4 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with a p-value of 0.0229. Panel H shows BODIPY stained microscopy images of wild-type and faa1/faa4 deletion cells after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel I shows violin plots of lipid droplets per cell in wild-type and faa1/faa4 deletion cells under untreated and 1 molar potassium chloride treatment, with a p-value less than 0.0001. Panel J shows violin plots of lipid droplet diameter in micrometers in wild-type and faa1/faa4 deletion cells under untreated and 1 molar potassium chloride treatment. Panel K shows immunoblot images comparing samples without and with 1 molar potassium chloride treatment for 3 hours, with bands labeled Dga1, Lro1-Flag, Are1-HA, Are2-Myc, and Tubulin, and molecular mass markers shown at 37, 50, 75, and 100 kilodaltons. Panel L shows violin plots of fold protein induction for Dga1, Lro1-Flag, Are1-HA, and Are2-Myc after 3 hours of potassium chloride treatment compared with untreated conditions, with p-values of 0.0276 and 0.0963. Panel M shows a violin plot of Dga1-mNG per cell in arbitrary units at 0 and 3 hours of potassium chloride treatment, with a p-value less than 0.0001. Panel N shows a violin plot of Lro1-mCherry per cell in arbitrary units at 0 and 3 hours of potassium chloride treatment, with a p-value of 0.0010. Panel O shows a violin plot of Are1-GFP per cell in arbitrary units at 0 and 3 hours of potassium chloride treatment. Panel P shows a violin plot of Are2-mCherry per cell in arbitrary units at 0 and 3 hours of potassium chloride treatment. Panel Q shows a plot of messenger RNA expression change as log base 2 of negative delta C T over time in hours for dga1, lro1, are1, and are2, with data shown at 0, 0.5, 1, 2, and 3 hours and p-values of 0.0782 and 0.0245.

DAG and acylated fatty acid availability limits LD accumulation. (A) Schematic of TAG production from PA via DAG and FFAs. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and nem1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated; PA, phosphatidic acid; FFAs, free fatty acids. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 10. (C) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT and nem1Δ cells. Z-projection images are representative of multiple replicates. (D) Analysis of LDs per cell from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 104 (WT UT), 88 (nem1Δ UT), 109 (WT KCl), 95 (nem1Δ KCl) cells from three biological replicates. (E) Analysis of LD diameters from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 267 (WT UT), 333 (nem1Δ UT), 199 (WT KCl), 233 (nem1Δ KCl) LDs from 3 biological replicates. (F) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of WT, faa1Δ, and faa4Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (G) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of WT and faa1Δfaa4Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (H) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT and faa1Δfaa4Δ cells. Z-projection images are representative of multiple replicates. (I) Analysis of LDs per cell from samples used in H. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 89 (WT UT), 82 (faa1Δfaa4Δ UT), 79 (WT KCl), 96 (faa1Δfaa4Δ KCl) cells from three biological replicates. (J) Analysis of LD diameters from samples used in H. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 245 (WT UT), 267 (faa1Δfaa4Δ UT), 459 (WT KCl), 199 (faa1Δfaa4Δ KCl) LDs from three biological replicates. (K) Western blot analysis of Dga1, Lro1-Flag, Are1-HA, and Are2-Myc levels before and after 3 h of 1M KCl treatment in SCD. Images are representative of multiple replicates. (L) Analysis of normalized protein fold change from F. Statistics are compared with 1 (no fold change). UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7. (M) Median Dga1-mNG fluorescence measured by flow cytometry from populations of cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 8. (N) Median Lro1-mCherry fluorescence measured by flow cytometry from cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (O) Median Are1-GFP fluorescence measured by flow cytometry from cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (P) Median Are2-mCherry fluorescence measured by flow cytometry from cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (Q) mRNA levels of DGA1, LRO1, ARE1, and ARE2 compared with the housekeeping gene TUB2 at the indicated times after 1M KCl treatment. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. Source data are available for this figure:SourceData FS3.

Figure S3.
A multi-panel image depicts the effects of different genetic modifications and treatments on lipid droplet accumulation and related protein expression in cells.Panel A shows a schematic of lipid metabolism, with Nem1, phosphatidic acid, free fatty acids, Faa1 and Faa4, diacylglycerol, acyl fatty acid, Dga1, and triacylglycerol labeled. Panel B shows violin plots of Pln1-mCherry fluorescence signal in wild-type and nem1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with fluorescence measured in arbitrary units and a p-value of 0.0158. Panel C shows BODIPY stained microscopy images of wild-type and nem1 deletion cells after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel D shows violin plots of lipid droplets per cell for wild-type and nem1 deletion cells under untreated and 1 molar potassium chloride treatment, with p-values less than 0.0001. Panel E shows violin plots of lipid droplet diameter in micrometers for wild-type and nem1 deletion cells under untreated and 1 molar potassium chloride treatment, with p-values of 0.0191 and 0.0003. Panel F shows violin plots of Pln1-mCherry fluorescence signal in wild-type, faa1 deletion, and faa4 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment. Panel G shows violin plots of Pln1-mCherry fluorescence signal in wild-type and faa1/faa4 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with a p-value of 0.0229. Panel H shows BODIPY stained microscopy images of wild-type and faa1/faa4 deletion cells after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel I shows violin plots of lipid droplets per cell in wild-type and faa1/faa4 deletion cells under untreated and 1 molar potassium chloride treatment, with a p-value less than 0.0001. Panel J shows violin plots of lipid droplet diameter in micrometers in wild-type and faa1/faa4 deletion cells under untreated and 1 molar potassium chloride treatment. Panel K shows immunoblot images comparing samples without and with 1 molar potassium chloride treatment for 3 hours, with bands labeled Dga1, Lro1-Flag, Are1-HA, Are2-Myc, and Tubulin, and molecular mass markers shown at 37, 50, 75, and 100 kilodaltons. Panel L shows violin plots of fold protein induction for Dga1, Lro1-Flag, Are1-HA, and Are2-Myc after 3 hours of potassium chloride treatment compared with untreated conditions, with p-values of 0.0276 and 0.0963. Panel M shows a violin plot of Dga1-mNG per cell in arbitrary units at 0 and 3 hours of potassium chloride treatment, with a p-value less than 0.0001. Panel N shows a violin plot of Lro1-mCherry per cell in arbitrary units at 0 and 3 hours of potassium chloride treatment, with a p-value of 0.0010. Panel O shows a violin plot of Are1-GFP per cell in arbitrary units at 0 and 3 hours of potassium chloride treatment. Panel P shows a violin plot of Are2-mCherry per cell in arbitrary units at 0 and 3 hours of potassium chloride treatment. Panel Q shows a plot of messenger RNA expression change as log base 2 of negative delta C T over time in hours for dga1, lro1, are1, and are2, with data shown at 0, 0.5, 1, 2, and 3 hours and p-values of 0.0782 and 0.0245.

DAG and acylated fatty acid availability limits LD accumulation. (A) Schematic of TAG production from PA via DAG and FFAs. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and nem1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated; PA, phosphatidic acid; FFAs, free fatty acids. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 10. (C) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT and nem1Δ cells. Z-projection images are representative of multiple replicates. (D) Analysis of LDs per cell from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 104 (WT UT), 88 (nem1Δ UT), 109 (WT KCl), 95 (nem1Δ KCl) cells from three biological replicates. (E) Analysis of LD diameters from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 267 (WT UT), 333 (nem1Δ UT), 199 (WT KCl), 233 (nem1Δ KCl) LDs from 3 biological replicates. (F) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of WT, faa1Δ, and faa4Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (G) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of WT and faa1Δfaa4Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (H) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT and faa1Δfaa4Δ cells. Z-projection images are representative of multiple replicates. (I) Analysis of LDs per cell from samples used in H. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 89 (WT UT), 82 (faa1Δfaa4Δ UT), 79 (WT KCl), 96 (faa1Δfaa4Δ KCl) cells from three biological replicates. (J) Analysis of LD diameters from samples used in H. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 245 (WT UT), 267 (faa1Δfaa4Δ UT), 459 (WT KCl), 199 (faa1Δfaa4Δ KCl) LDs from three biological replicates. (K) Western blot analysis of Dga1, Lro1-Flag, Are1-HA, and Are2-Myc levels before and after 3 h of 1M KCl treatment in SCD. Images are representative of multiple replicates. (L) Analysis of normalized protein fold change from F. Statistics are compared with 1 (no fold change). UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7. (M) Median Dga1-mNG fluorescence measured by flow cytometry from populations of cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 8. (N) Median Lro1-mCherry fluorescence measured by flow cytometry from cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 6. (O) Median Are1-GFP fluorescence measured by flow cytometry from cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (P) Median Are2-mCherry fluorescence measured by flow cytometry from cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (Q) mRNA levels of DGA1, LRO1, ARE1, and ARE2 compared with the housekeeping gene TUB2 at the indicated times after 1M KCl treatment. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. Source data are available for this figure:SourceData FS3.

Close Figure S3.

Dga1 protein levels increase upon hyperosmotic stress

Having established the importance of Dga1 for stress-induced LD accumulation, we next examined how Dga1 is regulated during hyperosmotic stress. Hyperosmotic stress resulted in a significant increase in endogenous Dga1 protein abundance, as detected by western blot and by flow cytometry of endogenously tagged Dga1 (Fig. S3, K–M). Importantly, this effect appeared specific as levels of all other neutral lipid–synthesizing enzymes remained unchanged aside from a modest increase in the levels of endogenous Lro1 expressed as a fusion to mCherry (Fig. S3, K–P). Consistent with the increase in protein abundance, DGA1 mRNA levels were also specifically increased (Fig. S3 Q).

We next used fluorescence microscopy of Dga1-mNeonGreen (Dga1-mNG) to determine whether the increase in Dga1 abundance was accompanied by changes in subcellular localization. In unstressed cells, Dga1-mNG was distributed throughout the ER, whereas under hyperosmotic stress, it concentrated at LDs (Fig. 2 F). Given that Dga1 is more long-lived on LDs than in the ER (Ruggiano et al., 2016), this relocalization likely further stabilizes Dga1. Together, these data indicate that Dga1 upregulation and relocalization drive LD accumulation during adaptation to hyperosmotic stress.

The canonical hyperosmotic stress response indirectly limits LD accumulation

Hyperosmotic stress adaptation is predominantly mediated through the high osmolarity glycerol (HOG) pathway (de Nadal and Posas, 2022). This well-characterized signaling cascade involves multiple sensors that converge on activation of the Hog1 MAP kinase (Fig. S4 A), which triggers an adaptive transcriptional program. Among other changes, this program rapidly enhances synthesis of glycerol, which acts as an intracellular osmolyte and facilitates restoration of cell volume (Fig. 3 A). Hog1 activation typically lasts for around 30 min, while downstream protein and glycerol synthesis peak within 60–90 min of stress onset (Petelenz-Kurdziel et al., 2013; Shen et al., 2023; Klipp et al., 2005). To test whether the LD accumulation upon hyperosmotic stress is under HOG pathway control, we analyzed hog1Δ cells. In the absence of stress, hog1Δ cells showed Pln1-mCherry levels comparable to WT controls (Fig. 3 B). Remarkably, under hyperosmotic stress, hog1Δ cells displayed a dramatic increase in Pln1-mCherry levels, far exceeding that observed in WT cells (Fig. 3 B). This pronounced Pln1-mCherry increase in hog1Δ cells reflected the accumulation of more and larger LDs, as confirmed by fluorescence microscopy of BODIPY-stained cells (Fig. 3, C–E). Moreover, manipulation of Hog1 activity by mutations that promote or inhibit its function resulted in a reduction or increase in Pln1-mCherry accumulation, respectively (Fig. S4, B and C). Together, these experiments indicate that the HOG pathway limits LD accumulation during hyperosmotic stress.

+ Expand view − Collapse view
Figure S4.
Figure S4. Refer to the image caption for details. Panel A shows a schematic of the hyperosmotic stress signaling pathway involving Sln1, Ypd1, Ssk1, Pbs2, Hog1, Ptc1, Hkr1, Msb2, and Sho1, leading to osmo adaptation. Panel B shows violin plots of Pln1-mCherry signal in wild-type, sln1-DAmP, and ptc1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with fluorescence measured in arbitrary units and p-values of 0.0318 and 0.0073. Panel C shows violin plots of Pln1-mCherry signal in wild-type, hog1 deletion, pbs2 deletion, and sho1 deletion ssk1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0343, 0.0292, and 0.0045. Panel D shows violin plots of Pln1-mCherry signal in wild-type, hkr1 deletion msb2 deletion, hkr1 deletion ypd1-DAmP, and msb2 deletion ypd1-DAmP cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with a p-value less than 0.0001. Panel E shows violin plots of Pln1-mCherry signal in wild-type, wild-type plus sch9-wild-type, and wild-type plus sch9-ON cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with a p-value of 0.0448. Panel F shows violin plots of Pln1-mCherry signal in wild-type and sit4 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0523 and 0.0006. Panel G shows violin plots of Pln1-mCherry signal in wild-type and mpk1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment. Panel H shows violin plots of Pln1-mCherry signal in wild-type and hac1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment.

Hog1 regulator effect on LD accumulation upon hyperosmotic stress. (A) Schematic of the Hog1 regulatory pathway. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, sln1-DAmP, and ptc1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Ptc1 is the phosphatase which turns off Hog1 activity, while Sln1 constrains the activity of the signaling cascade upstream of Hog1. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (C) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hog1Δ, pbs2Δ, and sho1Δssk1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Pbs2 is the MAP kinase kinase which activates Hog1, while Sho1 and Ssk1 are the MAP kinase kinase kinases which activated Pbs2. These mutations restrict Hog1 activity. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (D) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hkr1Δmsb2Δ, hkr1Δypd1-DAmP, and msb2Δypd1-DAmP cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. These strains each have mutations of the most upstream sensor proteins, restricting activation of the HOG pathway. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 3. (E) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells with no vector, Sch9-3xHA vector, or Sch9-2D3E-3xHA vector (Sch9 ON) subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Expressing the Sch9-ON vector prevents the inhibition of Sch9 signaling following TORC1 inhibition. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 3. (F) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and sit4Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Sit4 is activated upon TORC1 inhibition, and its deletion prevents this activation. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (G) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and mpk1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Mpk1 is the terminally activated kinase in the cell wall integrity pathway, and its deletion prevents this activation. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 3. (H) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and hac1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. HAC1 mRNA undergoes splicing upon UPR activation, with the resulting protein activating transcription of downstream targets. Gene deletion prevents this occurring. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 3. UPR, unfolded protein response.

Figure S4.
A multi-panel image depicts the effect of Hog1 regulators on lipid droplet accumulation under hyperosmotic stress. Panel A shows a schematic of the hyperosmotic stress signaling pathway involving Sln1, Ypd1, Ssk1, Pbs2, Hog1, Ptc1, Hkr1, Msb2, and Sho1, leading to osmo adaptation. Panel B shows violin plots of Pln1-mCherry signal in wild-type, sln1-DAmP, and ptc1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with fluorescence measured in arbitrary units and p-values of 0.0318 and 0.0073. Panel C shows violin plots of Pln1-mCherry signal in wild-type, hog1 deletion, pbs2 deletion, and sho1 deletion ssk1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0343, 0.0292, and 0.0045. Panel D shows violin plots of Pln1-mCherry signal in wild-type, hkr1 deletion msb2 deletion, hkr1 deletion ypd1-DAmP, and msb2 deletion ypd1-DAmP cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with a p-value less than 0.0001. Panel E shows violin plots of Pln1-mCherry signal in wild-type, wild-type plus sch9-wild-type, and wild-type plus sch9-ON cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with a p-value of 0.0448. Panel F shows violin plots of Pln1-mCherry signal in wild-type and sit4 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0523 and 0.0006. Panel G shows violin plots of Pln1-mCherry signal in wild-type and mpk1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment. Panel H shows violin plots of Pln1-mCherry signal in wild-type and hac1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment.

Hog1 regulator effect on LD accumulation upon hyperosmotic stress. (A) Schematic of the Hog1 regulatory pathway. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, sln1-DAmP, and ptc1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Ptc1 is the phosphatase which turns off Hog1 activity, while Sln1 constrains the activity of the signaling cascade upstream of Hog1. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (C) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hog1Δ, pbs2Δ, and sho1Δssk1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Pbs2 is the MAP kinase kinase which activates Hog1, while Sho1 and Ssk1 are the MAP kinase kinase kinases which activated Pbs2. These mutations restrict Hog1 activity. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5. (D) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hkr1Δmsb2Δ, hkr1Δypd1-DAmP, and msb2Δypd1-DAmP cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. These strains each have mutations of the most upstream sensor proteins, restricting activation of the HOG pathway. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 3. (E) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT cells with no vector, Sch9-3xHA vector, or Sch9-2D3E-3xHA vector (Sch9 ON) subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Expressing the Sch9-ON vector prevents the inhibition of Sch9 signaling following TORC1 inhibition. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 3. (F) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and sit4Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Sit4 is activated upon TORC1 inhibition, and its deletion prevents this activation. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (G) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and mpk1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Mpk1 is the terminally activated kinase in the cell wall integrity pathway, and its deletion prevents this activation. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 3. (H) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and hac1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. HAC1 mRNA undergoes splicing upon UPR activation, with the resulting protein activating transcription of downstream targets. Gene deletion prevents this occurring. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 3. UPR, unfolded protein response.

Close Figure S4.
Figure 3.
A multi-panel image depicts hyperosmotic stress signaling and lipid droplet production. Panel A shows a schematic of the hyperosmotic stress response, in which hyperosmotic stress activates Hog1, which acts through Gpd1 to promote glycerol production. Panel B shows violin plots of Pln1-mCherry signal in wild-type and hog1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with fluorescence measured in arbitrary units and a p-value less than 0.0001. Panel C shows BODIPY stained microscopy images of wild-type and hog1 deletion cells after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel D shows violin plots of lipid droplets per cell in wild-type and hog1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with a p-value less than 0.0001. Panel E shows violin plots of lipid droplet diameter in micrometers in wild-type and hog1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0663 and less than 0.0001. Panel F shows violin plots of Pln1-mCherry signal in wild-type, hog1 deletion, and gpd1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values less than 0.0001. Panel G shows violin plots of Pln1-mCherry signal in wild-type and hog1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment in media containing 0 percent glycerol or 5 percent glycerol, with a p-value of 0.0252 for the 0 percent glycerol condition and a non-significant result for the 5 percent glycerol condition. Panel H shows violin plots of Pln1-mCherry signal in wild-type, hog1 deletion, and hog1 deletion dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0232, 0.0333, and 0.0110. Panel I shows violin plots of Pln1-mCherry signal in wild-type, hog1 deletion, hog1 deletion dga1 deletion, and hog1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0069, 0.0011, and 0.0002.

Canonical hyperosmotic stress signaling indirectly limits LD production. (A) Simplified schematic of hyperosmotic stress adaptation signaling. Hyperosmotic stress leads to activation of the Hog1 MAP kinase, which enhances Gpd1 production to create glycerol as a balancing intracellular osmolyte. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and hog1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7. (C) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT and hog1Δ cells. Z-projection images are representative of multiple replicates. (D) Analysis of LDs per cell from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 83 (WT UT), 83 (hog1Δ UT), 49 (WT KCl), 55 (hog1Δ KCl) cells from three biological replicates. (E) Analysis of LD diameters from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 293 (WT UT), 238 (hog1Δ UT), 188 (WT KCl), 396 (hog1Δ KCl) LDs from three biological replicates. (F) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hog1Δ, and gpd1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (G) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and hog1Δ cells subjected to 1M KCl treatment with or without 5% glycerol in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (H) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hog1Δ, and hog1Δdga1Δlro1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (I) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hog1Δ, hog1Δdga1Δ, and hog1Δlro1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5.

Figure 3.
A multi-panel image depicts hyperosmotic stress signaling and lipid droplet production. Panel A shows a schematic of the hyperosmotic stress response, in which hyperosmotic stress activates Hog1, which acts through Gpd1 to promote glycerol production. Panel B shows violin plots of Pln1-mCherry signal in wild-type and hog1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with fluorescence measured in arbitrary units and a p-value less than 0.0001. Panel C shows BODIPY stained microscopy images of wild-type and hog1 deletion cells after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel D shows violin plots of lipid droplets per cell in wild-type and hog1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with a p-value less than 0.0001. Panel E shows violin plots of lipid droplet diameter in micrometers in wild-type and hog1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0663 and less than 0.0001. Panel F shows violin plots of Pln1-mCherry signal in wild-type, hog1 deletion, and gpd1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values less than 0.0001. Panel G shows violin plots of Pln1-mCherry signal in wild-type and hog1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment in media containing 0 percent glycerol or 5 percent glycerol, with a p-value of 0.0252 for the 0 percent glycerol condition and a non-significant result for the 5 percent glycerol condition. Panel H shows violin plots of Pln1-mCherry signal in wild-type, hog1 deletion, and hog1 deletion dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0232, 0.0333, and 0.0110. Panel I shows violin plots of Pln1-mCherry signal in wild-type, hog1 deletion, hog1 deletion dga1 deletion, and hog1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with p-values of 0.0069, 0.0011, and 0.0002.

Canonical hyperosmotic stress signaling indirectly limits LD production. (A) Simplified schematic of hyperosmotic stress adaptation signaling. Hyperosmotic stress leads to activation of the Hog1 MAP kinase, which enhances Gpd1 production to create glycerol as a balancing intracellular osmolyte. (B) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and hog1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7. (C) LDs stained with BODIPY and imaged by live-cell microscopy after 3 h of 1M KCl treatment in SCD for Pln1-mCherry WT and hog1Δ cells. Z-projection images are representative of multiple replicates. (D) Analysis of LDs per cell from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 83 (WT UT), 83 (hog1Δ UT), 49 (WT KCl), 55 (hog1Δ KCl) cells from three biological replicates. (E) Analysis of LD diameters from samples used in C. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 293 (WT UT), 238 (hog1Δ UT), 188 (WT KCl), 396 (hog1Δ KCl) LDs from three biological replicates. (F) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hog1Δ, and gpd1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (G) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT and hog1Δ cells subjected to 1M KCl treatment with or without 5% glycerol in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (H) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hog1Δ, and hog1Δdga1Δlro1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 4. (I) Median Pln1-mCherry fluorescence measured by flow cytometry from populations of single WT, hog1Δ, hog1Δdga1Δ, and hog1Δlro1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 5.

Close Figure 3.

The HOG pathway could limit LD accumulation because when it is active, cells ultimately experience less of the hyperosmotic stress. In this scenario, one would expect that LD accumulation should be similarly modulated by altering glycerol levels. Indeed, deletion of GPD1, a glycerol phosphate dehydrogenase essential for glycerol production, enhanced Pln1-mCherry accumulation to the same extent as HOG1 deletion (Fig. 3 F). Conversely, supplementing hog1Δ cells with glycerol reduced Pln1-mCherry accumulation to WT levels (Fig. 3 G). Blocking activation of the upstream sensor proteins Hkr1, Msb2, and Ypd1 did not prevent LD accumulation (Fig. S4 D), further indicating that LD accumulation is not directly regulated by the HOG pathway. As cells defective in the HOG pathway cannot effectively produce the intracellular osmolytes necessary to resolve the decreased membrane tension upon acute hyperosmotic stress (Petelenz-Kurdziel et al., 2013), LDs may be continually produced to try and increase membrane tension in these cells, as described for the acute phase of hyperosmotic stress (Phan et al., 2025). In agreement with this, although hog1Δ cells relied on Dga1 and Lro1 for full LD accumulation (Fig. 3 H), this was no longer dependent upon Dga1 as in the WT (Fig. 3 I and Fig. 2 B). This suggests a larger role for Lro1, as seen in the acute phase of adaptation (Phan et al., 2025). Together, these results indicate that the HOG pathway indirectly limits LD accumulation by promoting cell adaptation to hyperosmotic stress.

Besides the HOG pathway, the response to hyperosmotic stress can also be modulated by a variety of other signaling pathways such as the unfolded protein response (Burgos et al., 2019), the cell wall integrity (Yoshikawa et al., 2009), and target of rapamycin (Hughes Hallett et al., 2014). However, genetic manipulation of these pathways did not affect Pln1-mCherry upon hyperosmotic stress levels, suggesting they play no role in the LD response (Fig. S4, E–H).

Chronic hyperosmotic stress drives membrane remodeling

While our data identify TAG-containing LDs as important for adaptation to hyperosmotic stress, we next asked whether this response is accompanied by broader organelle remodeling. To assess how widespread organelle remodeling is during hyperosmotic stress, we analyzed the morphology of the ER and mitochondria, organelles that maintain close physical and metabolic connections with LDs (Wu et al., 2018; Henne and Cohen, 2026). Upon hyperosmotic stress, we detected no substantial changes in the morphology or abundance of the ER and mitochondria, monitored, respectively, by endogenously GFP-tagged Sec63 and Ilv6 (Fig. 4, A and B), well-established marker proteins of these organelles (Wang et al., 2018; Huh et al., 2003). Time course analysis revealed that as for other stress conditions (Fedoseeva et al., 2017; Kondo et al., 2022), hyperosmotic stress caused a rapid but transient mitochondrial fragmentation. This was resolved within 1 h in both WT and dga1Δlro1Δ cells (Fig. 4, C–E). Thus, although hyperosmotic stress causes transient mitochondrial fragmentation during the acute phase, chronic hyperosmotic stress does not trigger persistent large-scale organelle remodeling. LD accumulation therefore appears to be a specific long-term lipid response rather than a consequence of gross ER or mitochondrial dysfunction. Importantly, the similar organelle morphology observed in WT and dga1Δlro1Δ cells further indicates that the reduced fitness of TAG-deficient cells under chronic hyperosmotic stress is not caused by gross ER or mitochondrial collapse.

Figure 4.
A multi-panel image showing cellular response to hyperosmotic stress. Panel A shows fluorescence microscopy images of the endoplasmic reticulum, visualized with Sec63-GFP, in wild-type (WT) and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel B shows fluorescence microscopy images of mitochondria, visualized with Ilv6-GFP, in WT and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel C presents fluorescence microscopy images of mitochondria, visualized with Ilv6-GFP, in WT and dga1 deletion lro1 deletion cells under untreated conditions, after 15 minutes of 1 molar potassium chloride treatment, and after 60 minutes of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel D shows a scatter plot of the percentage of cells with fragmented mitochondria for WT and dga1 deletion lro1 deletion cells at 0, 15, 30, and 60 minutes of 1 molar potassium chloride treatment. Panel E shows a scatter plot of the percentage of cells with elongated mitochondria for WT and dga1 deletion lro1 deletion cells at 0, 15, 30, and 60 minutes of 1 molar potassium chloride treatment.

Hyperosmotic stress does not cause lasting changes to ER or mitochondrial morphology. (A) Cellular ER visualized with Sec63-GFP and imaged by live-cell microscopy after 0 (UT) or 3 h of 1M KCl treatment in SCD for WT and dga1Δlro1Δ cells. Images show cell midsections and are representative of multiple replicates. (B) Mitochondria visualized by endogenous Ilv6-GFP expression and imaged by live-cell microscopy after 0 (UT) or 3 h of 1M KCl treatment in SCD for WT and dga1Δlro1Δ cells. Z-projection images are representative of multiple replicates. (C) Mitochondria visualized by endogenous Ilv6-GFP expression and imaged by live-cell microscopy after 0 (UT), 15, or 60 min of 1M KCl treatment in SCD for WT and dga1Δlro1Δ cells. Z-projection images are representative of multiple replicates. (D) Quantification of cells from C with fragmented mitochondria. Each point shows the percentage of cells from a replicate (N = 4) exhibiting solely fragmented mitochondria. Error bars show the mean and SEM. (E) Quantification of cells from C with elongated mitochondria. Each point shows the percentage of cells from a replicate (N = 4) exhibiting solely elongated mitochondria. Error bars show the mean and SEM.

Figure 4.
A multi-panel image showing cellular response to hyperosmotic stress. Panel A shows fluorescence microscopy images of the endoplasmic reticulum, visualized with Sec63-GFP, in wild-type (WT) and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel B shows fluorescence microscopy images of mitochondria, visualized with Ilv6-GFP, in WT and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel C presents fluorescence microscopy images of mitochondria, visualized with Ilv6-GFP, in WT and dga1 deletion lro1 deletion cells under untreated conditions, after 15 minutes of 1 molar potassium chloride treatment, and after 60 minutes of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel D shows a scatter plot of the percentage of cells with fragmented mitochondria for WT and dga1 deletion lro1 deletion cells at 0, 15, 30, and 60 minutes of 1 molar potassium chloride treatment. Panel E shows a scatter plot of the percentage of cells with elongated mitochondria for WT and dga1 deletion lro1 deletion cells at 0, 15, 30, and 60 minutes of 1 molar potassium chloride treatment.

Hyperosmotic stress does not cause lasting changes to ER or mitochondrial morphology. (A) Cellular ER visualized with Sec63-GFP and imaged by live-cell microscopy after 0 (UT) or 3 h of 1M KCl treatment in SCD for WT and dga1Δlro1Δ cells. Images show cell midsections and are representative of multiple replicates. (B) Mitochondria visualized by endogenous Ilv6-GFP expression and imaged by live-cell microscopy after 0 (UT) or 3 h of 1M KCl treatment in SCD for WT and dga1Δlro1Δ cells. Z-projection images are representative of multiple replicates. (C) Mitochondria visualized by endogenous Ilv6-GFP expression and imaged by live-cell microscopy after 0 (UT), 15, or 60 min of 1M KCl treatment in SCD for WT and dga1Δlro1Δ cells. Z-projection images are representative of multiple replicates. (D) Quantification of cells from C with fragmented mitochondria. Each point shows the percentage of cells from a replicate (N = 4) exhibiting solely fragmented mitochondria. Error bars show the mean and SEM. (E) Quantification of cells from C with elongated mitochondria. Each point shows the percentage of cells from a replicate (N = 4) exhibiting solely elongated mitochondria. Error bars show the mean and SEM.

Close Figure 4.

Given the absence of widespread organelle remodeling, we next asked whether hyperosmotic stress instead remodels the cellular lipidome. To address this, we performed whole-cell lipidomic analysis to define global lipid changes. We observed that hyperosmotic stress resulted in a large increase of neutral lipids in both WT and hog1Δ cells but not in LDΔ, as expected (Fig. S5 A). Interestingly, in WT cells this increase was dominated by TAGs, while in hog1Δ cells, TAGs and STEs increased to a similar extent suggesting that the LDs produced in these different genetic backgrounds are distinct (Fig. 5, A and B; and Fig. S5 B). Further analysis of the TAGs accumulated during hyperosmotic stress revealed changes in the acyl-chain composition specific to WT cells, with the constituent fatty acids being, on average, longer and more unsaturated (Fig. S5 C). These changes were primarily due to higher proportions of 16:1 and 18:1 chains (Fig. S5 D). Thus, chronic hyperosmotic stress remodels neutral lipid content and composition.

+ Expand view− Collapse view
Figure S5.
Figure S5. Refer to the image caption for details.Panel A shows a stacked bar graph of lipid amounts, expressed as normalized picomoles, for LD4, WT, and hog1 deletion cells under untreated and 1 molar potassium chloride conditions, with categories including neutral lipids, lysophospholipids, sphingolipids, diacylglycerols, ergosterol, and phospholipids. Panel B shows bar graphs of neutral lipid types, expressed as a percentage of neutral lipids, comparing triacylglycerol and sterol ester levels in WT and hog1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel C shows TAG acyl chain length and unsaturation, with the x-axis representing average fatty acid carbons and the y-axis representing average fatty acid double bonds. Panel D shows lipid amounts as a percentage of TAG acyl chains for fatty acid chain compositions of 16 carbons with 0, 1, or 2 double bonds and 18 carbons with 0 or 1 double bond. Panel E shows phosphatidylethanolamine levels as a percentage of phospholipids in LD4, WT, and hog1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel F shows phosphatidylcholine levels as a percentage of phospholipids in LD4, WT, and hog1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel G shows PE acyl chain length and unsaturation, with the x-axis representing average fatty acid carbons and the y-axis representing average fatty acid double bonds. Panel H shows phosphatidylethanolamine amounts expressed as normalized picomoles for different fatty acid chain compositions, including 16 carbons with 0 or 1 double bond, 16 carbons with 0 or 1 double bond combined with 18 carbons with 1 double bond, 16 carbons with 1 double bond combined with 16 or 18 carbons with 1 double bond, 18 carbons with 0 or 1 double bond, and 18 carbons with 1 double bond. Panel I shows PC acyl chain length and unsaturation, with the x-axis representing average fatty acid carbons and the y-axis representing average fatty acid double bonds. Panel J shows immunoblot images comparing WT and dga1 deletion lro1 deletion cells without and with 3 hours of 1 molar potassium chloride treatment, with Opi3-GFP, free GFP, and Tubulin indicated. Panel K shows messenger RNA expression change as log base 2 of negative delta delta C T for cho2 and opi3 in WT and dga1 deletion lro1 deletion cells under untreated and 1 molar potassium chloride conditions.

Limiting PC catabolism does not impact LDs under hyperosmotic stress. (A) Normalized levels of the major lipid classes for each strain detected by lipidomics with and without 3 h of treatment with 1M KCl. Bars show the mean with SEM from three biological replicates. (B) Proportion of neutral lipids from A that were TAGs and STEs for WT and hog1Δ cells with and without 3 h of treatment with 1M KCl. Bars show the mean with SEM from three biological replicates. (C) Averaged properties of all acyl chains found in TAGs for WT and hog1Δ cells both with and without 3 h of 1M KCl treatment. The x axis shows the average length in carbons, while the y axis shows the average number of double bonds. Each point shows an individual replicate (three for each condition). (D) Proportion of the major acyl-chain classes found in TAGs for WT and hog1Δ cells with and without 3 h of treatment with 1M KCl. Bars show the mean with SEM from three biological replicates. (E) PE species as a percentage of total PLs detected with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (F) PC species as a percentage of total PLs detected with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (G) Averaged properties of all acyl chains found in PE for each strain both with and without 3 h of 1M KCl treatment. The x axis shows the average length in carbons, while the y axis shows the average number of double bonds. Each point shows an individual replicate (three for each condition). (H) Normalized levels of the major classes of PE species for each strain with and without 3 h of treatment with 1M KCl. Bars show the mean with SEM from three biological replicates. (I) Averaged properties of all acyl chains found in PC for each strain both with and without 3 h of 1M KCl treatment. The x axis shows the average length in carbons, while the y axis shows the average number of double bonds. Each point shows an individual replicate (three for each condition). (J) Western blot of Opi3-GFP from populations of single WT and dga1Δlro1Δ cells either untreated or subjected to 1M KCl treatment in SCD media for 3 h. Images are representative of multiple replicates. (K) mRNA levels of CHO2 and OPI3 in WT and dga1Δlro1Δ cells compared with the housekeeping gene TUB2 at the indicated times after 1M KCl treatment. Error bars show the mean and SEM, with four replicates for CHO2 and 3 for OPI3. PLs, phospholipids. Source data are available for this figure: SourceData FS5.

Figure S5.
A multi-panel image depicts lipidomic analysis under hyperosmotic stress.Panel A shows a stacked bar graph of lipid amounts, expressed as normalized picomoles, for LD4, WT, and hog1 deletion cells under untreated and 1 molar potassium chloride conditions, with categories including neutral lipids, lysophospholipids, sphingolipids, diacylglycerols, ergosterol, and phospholipids. Panel B shows bar graphs of neutral lipid types, expressed as a percentage of neutral lipids, comparing triacylglycerol and sterol ester levels in WT and hog1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel C shows TAG acyl chain length and unsaturation, with the x-axis representing average fatty acid carbons and the y-axis representing average fatty acid double bonds. Panel D shows lipid amounts as a percentage of TAG acyl chains for fatty acid chain compositions of 16 carbons with 0, 1, or 2 double bonds and 18 carbons with 0 or 1 double bond. Panel E shows phosphatidylethanolamine levels as a percentage of phospholipids in LD4, WT, and hog1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel F shows phosphatidylcholine levels as a percentage of phospholipids in LD4, WT, and hog1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel G shows PE acyl chain length and unsaturation, with the x-axis representing average fatty acid carbons and the y-axis representing average fatty acid double bonds. Panel H shows phosphatidylethanolamine amounts expressed as normalized picomoles for different fatty acid chain compositions, including 16 carbons with 0 or 1 double bond, 16 carbons with 0 or 1 double bond combined with 18 carbons with 1 double bond, 16 carbons with 1 double bond combined with 16 or 18 carbons with 1 double bond, 18 carbons with 0 or 1 double bond, and 18 carbons with 1 double bond. Panel I shows PC acyl chain length and unsaturation, with the x-axis representing average fatty acid carbons and the y-axis representing average fatty acid double bonds. Panel J shows immunoblot images comparing WT and dga1 deletion lro1 deletion cells without and with 3 hours of 1 molar potassium chloride treatment, with Opi3-GFP, free GFP, and Tubulin indicated. Panel K shows messenger RNA expression change as log base 2 of negative delta delta C T for cho2 and opi3 in WT and dga1 deletion lro1 deletion cells under untreated and 1 molar potassium chloride conditions.

Limiting PC catabolism does not impact LDs under hyperosmotic stress. (A) Normalized levels of the major lipid classes for each strain detected by lipidomics with and without 3 h of treatment with 1M KCl. Bars show the mean with SEM from three biological replicates. (B) Proportion of neutral lipids from A that were TAGs and STEs for WT and hog1Δ cells with and without 3 h of treatment with 1M KCl. Bars show the mean with SEM from three biological replicates. (C) Averaged properties of all acyl chains found in TAGs for WT and hog1Δ cells both with and without 3 h of 1M KCl treatment. The x axis shows the average length in carbons, while the y axis shows the average number of double bonds. Each point shows an individual replicate (three for each condition). (D) Proportion of the major acyl-chain classes found in TAGs for WT and hog1Δ cells with and without 3 h of treatment with 1M KCl. Bars show the mean with SEM from three biological replicates. (E) PE species as a percentage of total PLs detected with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (F) PC species as a percentage of total PLs detected with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (G) Averaged properties of all acyl chains found in PE for each strain both with and without 3 h of 1M KCl treatment. The x axis shows the average length in carbons, while the y axis shows the average number of double bonds. Each point shows an individual replicate (three for each condition). (H) Normalized levels of the major classes of PE species for each strain with and without 3 h of treatment with 1M KCl. Bars show the mean with SEM from three biological replicates. (I) Averaged properties of all acyl chains found in PC for each strain both with and without 3 h of 1M KCl treatment. The x axis shows the average length in carbons, while the y axis shows the average number of double bonds. Each point shows an individual replicate (three for each condition). (J) Western blot of Opi3-GFP from populations of single WT and dga1Δlro1Δ cells either untreated or subjected to 1M KCl treatment in SCD media for 3 h. Images are representative of multiple replicates. (K) mRNA levels of CHO2 and OPI3 in WT and dga1Δlro1Δ cells compared with the housekeeping gene TUB2 at the indicated times after 1M KCl treatment. Error bars show the mean and SEM, with four replicates for CHO2 and 3 for OPI3. PLs, phospholipids. Source data are available for this figure: SourceData FS5.

Close Figure S5.

We next examined changes to phospholipid composition. Most phospholipids remained largely unchanged as a proportion of the phospholipid pool (Fig. 5 C). Strikingly, however, in all three strains hyperosmotic stress increased the proportion of PC with a concomitant drop in phosphatidylethanolamine (PE) (Fig. 5 D; and Fig. S5, E and F). The decrease in PE was comparable between WT and LDΔ cells but more pronounced in hog1Δ cells, whereas the increase in PC was attenuated in LDΔ cells relative to WT and trended slightly lower in hog1Δ (Fig. 5 D). Hyperosmotic stress therefore drives a switch in the major phospholipid classes, a change that appears more robust in WT cells. We then asked whether this switch was accompanied by remodeling of the acyl chains, both within each class and in the phospholipids as a whole. The PE acyl chains became longer following stress in all strains, owing to a depletion of C16-containing species, while the minor di-C18 species were retained (Fig. S5, G and H). PC chains remained similar in length but became more saturated, driven by the emergence of a new species, 16:0_16:1, absent from untreated cells (Fig. 5 E and Fig. S5 I). The overall acyl-chain profile of stressed WT cells shifted toward longer chains but remained closely similar to that of LDΔ cells (Fig. 5 F); because LDΔ cells share this profile yet are impaired for proliferation under stress, acyl-chain remodeling alone cannot account for the fitness defect. Thus, chronic hyperosmotic stress remodels the most abundant phospholipids, shifting the balance from PE to PC and altering the acyl-chain composition within both classes.

Figure 5.
A multi-panel image depicts lipidomic changes under hyperosmotic stress. Panel A shows a scatter plot. The x-axis represents the treatment conditions, untreated and 3 hours of potassium chloride treatment, and the y-axis represents the triacylglycerol level as a percentage. Panel B shows a scatter plot of sterol esters as a percentage of total cellular lipids with and without 3 hours of treatment with 1 molar potassium chloride. The x-axis represents the treatment conditions, untreated and 3 hours of potassium chloride treatment, and the y-axis represents the sterol ester level as a percentage. Panel C shows a bar graph of the proportion of phospholipid species detected by lipidomics as a percentage of total phospholipids. The x-axis represents different strains and conditions, including LD deletion untreated, LD deletion potassium chloride, wild-type untreated, wild-type potassium chloride, hog1 deletion untreated, and hog1 deletion potassium chloride, and the y-axis represents the proportion of phospholipids as a percentage. Panel D shows a scatter plot of log base 2 fold changes in phosphatidylethanolamine and phosphatidylcholine detected by lipidomics following 3 hours of treatment with 1 molar potassium chloride compared with untreated conditions. The x-axis represents different phospholipids, phosphatidylethanolamine and phosphatidylcholine, and the y-axis represents the log base 2 fold change. Panel E shows a bar graph of normalized levels of the major classes of phosphatidylcholine species detected by lipidomics for each strain with and without 3 hours of treatment with 1 molar potassium chloride. The x-axis represents different strains and conditions, and the y-axis represents the phosphatidylcholine amount in picomoles. Panel F shows a scatter plot of averaged properties of all acyl chains found in phospholipids for each strain with and without 3 hours of treatment with 1 molar potassium chloride. The x-axis shows the average length in carbon atoms, while the y-axis shows the average number of double bonds.

Chronic hyperosmotic stress drives LD-dependent lipidome changes. (A) TAGs as a percentage of total cellular lipids detected by lipidomics with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (B) STEs as a percentage of total cellular lipids detected by lipidomics with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (C) Proportion of phospholipid species detected by lipidomics as a percentage of total PLs. PA = phosphatidic acid; PC = phosphatidylcholine; PE = phosphatidylethanolamine; PI = phosphatidylinositol; PS = phosphatidylserine; PLs, phospholipids. N = 3; error bars show the mean and SEM. (D) Log2 fold changes in PE and PC detected by lipidomics following 3 h of treatment with 1M KCl compared with untreated. N = 3; error bars show the mean and SEM. (E) Normalized levels of the major classes of PC species detected by lipidomics for each strain with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (F) Averaged properties of all acyl chains found in phospholipids for each strain both with and without 3 h of 1M KCl treatment. The x axis shows the average length in carbons, while the y axis shows the average number of double bonds. Each point shows an individual replicate (three for each condition).

Figure 5.
A multi-panel image depicts lipidomic changes under hyperosmotic stress. Panel A shows a scatter plot. The x-axis represents the treatment conditions, untreated and 3 hours of potassium chloride treatment, and the y-axis represents the triacylglycerol level as a percentage. Panel B shows a scatter plot of sterol esters as a percentage of total cellular lipids with and without 3 hours of treatment with 1 molar potassium chloride. The x-axis represents the treatment conditions, untreated and 3 hours of potassium chloride treatment, and the y-axis represents the sterol ester level as a percentage. Panel C shows a bar graph of the proportion of phospholipid species detected by lipidomics as a percentage of total phospholipids. The x-axis represents different strains and conditions, including LD deletion untreated, LD deletion potassium chloride, wild-type untreated, wild-type potassium chloride, hog1 deletion untreated, and hog1 deletion potassium chloride, and the y-axis represents the proportion of phospholipids as a percentage. Panel D shows a scatter plot of log base 2 fold changes in phosphatidylethanolamine and phosphatidylcholine detected by lipidomics following 3 hours of treatment with 1 molar potassium chloride compared with untreated conditions. The x-axis represents different phospholipids, phosphatidylethanolamine and phosphatidylcholine, and the y-axis represents the log base 2 fold change. Panel E shows a bar graph of normalized levels of the major classes of phosphatidylcholine species detected by lipidomics for each strain with and without 3 hours of treatment with 1 molar potassium chloride. The x-axis represents different strains and conditions, and the y-axis represents the phosphatidylcholine amount in picomoles. Panel F shows a scatter plot of averaged properties of all acyl chains found in phospholipids for each strain with and without 3 hours of treatment with 1 molar potassium chloride. The x-axis shows the average length in carbon atoms, while the y-axis shows the average number of double bonds.

Chronic hyperosmotic stress drives LD-dependent lipidome changes. (A) TAGs as a percentage of total cellular lipids detected by lipidomics with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (B) STEs as a percentage of total cellular lipids detected by lipidomics with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (C) Proportion of phospholipid species detected by lipidomics as a percentage of total PLs. PA = phosphatidic acid; PC = phosphatidylcholine; PE = phosphatidylethanolamine; PI = phosphatidylinositol; PS = phosphatidylserine; PLs, phospholipids. N = 3; error bars show the mean and SEM. (D) Log2 fold changes in PE and PC detected by lipidomics following 3 h of treatment with 1M KCl compared with untreated. N = 3; error bars show the mean and SEM. (E) Normalized levels of the major classes of PC species detected by lipidomics for each strain with and without 3 h of treatment with 1M KCl. N = 3; error bars show the mean and SEM. (F) Averaged properties of all acyl chains found in phospholipids for each strain both with and without 3 h of 1M KCl treatment. The x axis shows the average length in carbons, while the y axis shows the average number of double bonds. Each point shows an individual replicate (three for each condition).

Close Figure 5.

Interplay between LDs and phospholipid remodeling during chronic hyperosmotic stress

To test whether the increased PC is functionally important, we manipulated PC synthesis. PC is produced by sequential trimethylation of PE in the ER, with the initial step carried out by Cho2 and subsequent steps by Opi3 (McMaster and Bell, 1994) (Fig. 6 A). Alternatively, PC can be synthesized via the Kennedy pathway from DAG and choline, with Pct1 catalyzing the rate-limiting step (Fig. 6 A). The Kennedy pathway can also produce PE by providing ethanolamine instead of choline. We therefore asked whether stimulating PC synthesis could compensate for the loss of TAG in dga1Δlro1Δ cells. Choline supplementation improved the growth of dga1Δlro1Δ cells under hyperosmotic stress (Fig. 6, B and C). Importantly, this effect was dependent on Pct1, indicating that the rescue requires PC synthesis from exogenous choline (Fig. 6 D). In contrast, ethanolamine supplementation did not improve the growth of dga1Δlro1Δ cells under hyperosmotic stress. Thus, elevated PC and TAG levels are both important for adaptation to chronic hyperosmotic stress, and increasing PC levels can compensate for the absence of TAG.

Figure 6.
A multi-panel image depicts cell growth and protein localization under different conditions. Panel A shows a schematic of phospholipid metabolism involving ethanolamine (Eth), diacylglycerol (DAG), choline (Cho), phosphatidylethanolamine (PE), phosphatidylcholine (PC), Cho2, Opi3, and Ptc1. Panel B shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for WT cells under untreated conditions, 1 molar potassium chloride, 1 molar potassium chloride plus choline, and 1 molar potassium chloride plus ethanolamine conditions. Panel C shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for dga1 deletion lro1 deletion cells under the same four conditions. Panel D shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for ptc1 deletion dga1 deletion lro1 deletion cells under untreated conditions, 1 molar potassium chloride, 1 molar potassium chloride plus choline, and 1 molar potassium chloride plus ethanolamine conditions. Panel E shows violin plots of Opi3-GFP fluorescence in arbitrary units for WT and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with Cho2 and Opi3 shown on the x-axis and p-values of 0.0182, 0.0177, 0.0023, and 0.0012. Panel F shows fluorescence microscopy images of Cho2-GFP in WT and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel G shows fluorescence microscopy images of Opi3-GFP in WT and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with 5-micrometer scale bars.

PE and PC must be balanced for hyperosmotic stress adaptation. (A) Schematic of PC biosynthesis from PE conversion. PE and PC can also be made by combining DAG with, respectively, ethanolamine (Eth) or choline (Cho). (B) Growth assay of Pln1-mCherry WT cells in SCD and SCD + 1M KCl with and without supplementation of 2 mM ethanolamine (Eth) or choline (Cho). The main line shows the mean, with the error bars showing SEM from six biological replicates. (C) Growth assay of Pln1-mCherry dga1Δlro1Δ cells in SCD and SCD + 1M KCl with and without supplementation of 2 mM ethanolamine (Eth) or choline (Cho). The main line shows the mean, with the error bars showing SEM from six biological replicates. (D) Growth assay of Pln1-mCherry dga1Δlro1Δpct1Δ cells in SCD and SCD + 1M KCl with and without supplementation of 2 mM ethanolamine (Eth) or choline (Cho). The main line shows the mean, with the error bars showing SEM from six biological replicates. (E) Median fluorescence of endogenously expressed Cho2-GFP and Opi3-GFP measured by flow cytometry from populations of WT and dga1Δlro1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7 for Cho2-GFP, 8 for Opi3-GFP. (F) Localization of Cho2 tagged with GFP in WT and dga1Δlro1Δ cells grown either in SCD media or in SCD supplemented with 1M KCl for 3 h prior to imaging. Images use the same color scale, show a single slice, and are representative of multiple experiments. (G) Localization of Opi3 tagged with GFP in WT and dga1Δlro1Δ cells grown either in SCD media or in SCD supplemented with 1M KCl for 3 h prior to imaging. Images use the same color scale, show a single slice, and are representative of multiple experiments.

Figure 6.
A multi-panel image depicts cell growth and protein localization under different conditions. Panel A shows a schematic of phospholipid metabolism involving ethanolamine (Eth), diacylglycerol (DAG), choline (Cho), phosphatidylethanolamine (PE), phosphatidylcholine (PC), Cho2, Opi3, and Ptc1. Panel B shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for WT cells under untreated conditions, 1 molar potassium chloride, 1 molar potassium chloride plus choline, and 1 molar potassium chloride plus ethanolamine conditions. Panel C shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for dga1 deletion lro1 deletion cells under the same four conditions. Panel D shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for ptc1 deletion dga1 deletion lro1 deletion cells under untreated conditions, 1 molar potassium chloride, 1 molar potassium chloride plus choline, and 1 molar potassium chloride plus ethanolamine conditions. Panel E shows violin plots of Opi3-GFP fluorescence in arbitrary units for WT and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with Cho2 and Opi3 shown on the x-axis and p-values of 0.0182, 0.0177, 0.0023, and 0.0012. Panel F shows fluorescence microscopy images of Cho2-GFP in WT and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with 5 micrometer scale bars. Panel G shows fluorescence microscopy images of Opi3-GFP in WT and dga1 deletion lro1 deletion cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, with 5-micrometer scale bars.

PE and PC must be balanced for hyperosmotic stress adaptation. (A) Schematic of PC biosynthesis from PE conversion. PE and PC can also be made by combining DAG with, respectively, ethanolamine (Eth) or choline (Cho). (B) Growth assay of Pln1-mCherry WT cells in SCD and SCD + 1M KCl with and without supplementation of 2 mM ethanolamine (Eth) or choline (Cho). The main line shows the mean, with the error bars showing SEM from six biological replicates. (C) Growth assay of Pln1-mCherry dga1Δlro1Δ cells in SCD and SCD + 1M KCl with and without supplementation of 2 mM ethanolamine (Eth) or choline (Cho). The main line shows the mean, with the error bars showing SEM from six biological replicates. (D) Growth assay of Pln1-mCherry dga1Δlro1Δpct1Δ cells in SCD and SCD + 1M KCl with and without supplementation of 2 mM ethanolamine (Eth) or choline (Cho). The main line shows the mean, with the error bars showing SEM from six biological replicates. (E) Median fluorescence of endogenously expressed Cho2-GFP and Opi3-GFP measured by flow cytometry from populations of WT and dga1Δlro1Δ cells subjected to 1M KCl treatment in SCD media for 3 h. UT = untreated. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 7 for Cho2-GFP, 8 for Opi3-GFP. (F) Localization of Cho2 tagged with GFP in WT and dga1Δlro1Δ cells grown either in SCD media or in SCD supplemented with 1M KCl for 3 h prior to imaging. Images use the same color scale, show a single slice, and are representative of multiple experiments. (G) Localization of Opi3 tagged with GFP in WT and dga1Δlro1Δ cells grown either in SCD media or in SCD supplemented with 1M KCl for 3 h prior to imaging. Images use the same color scale, show a single slice, and are representative of multiple experiments.

Close Figure 6.

Having established that elevated PC contributes to cell fitness under hyperosmotic stress, we next investigated how the PC increase is triggered. We hypothesized that the key PC biosynthetic enzymes Cho2 and Opi3 might be upregulated upon hyperosmotic stress. Quantification of endogenously GFP-tagged Cho2 (Cho2-GFP) and Opi3 (Opi3-GFP) by flow cytometry showed that chronic hyperosmotic stress increases the levels of both enzymes (Fig. 6 E). This increase did not result in altered localization of either Cho2-GFP or Opi3-GFP (Fig. 6, F and G) although in Opi3-GFP–expressing cells some of the GFP signal could be found within the vacuole. Using western blot, we confirmed both that this vacuolar species was likely free GFP released by autophagy and that the full-length Opi3-GFP was increased following hyperosmotic stress (Fig. S5 J). Surprisingly, in both cases, the increase in protein occurred without a corresponding increase in mRNA (Fig. S5 K), indicating a posttranscriptional mechanism. Increased Cho2 and Opi3 levels following hyperosmotic stress likely enable the phospholipid shift from PE to PC.

To further explore the links between TAG and PC synthesis in response to hyperosmotic stress, we tested how blocking TAG production impacts PC synthesis. Surprisingly, dga1Δlro1Δ cells displayed higher steady-state levels of both Cho2-GFP and Opi3-GFP, which did not increase further following hyperosmotic stress, and were not mislocalized (Fig. 6, E–G and Fig. S5 J). This increase also occurred without a corresponding increase in mRNA (Fig. S5 K). Thus, in the absence of TAG synthesis, PC biosynthetic enzymes are constitutively elevated and already at, or close to, the levels reached by stressed WT cells. This may explain why choline supplementation rescues growth in dga1Δlro1Δ cells: with the enzymes already maximally induced, these cells cannot further increase PC synthesis in response to stress, a limitation that is bypassed by supplying exogenous choline (Fig. 6 C).

Membrane adaptation to hyperosmotic stress requires TAG synthesis

Next, we tested whether the lipid remodeling induced by chronic hyperosmotic stress alters membrane properties. In particular, we focused on membrane fluidity, which has been reported to decrease under hyperosmotic conditions (Laroche et al., 2001). To investigate this, we used fluorescence recovery after photobleaching (FRAP) of the ER-resident membrane protein Sec63-GFP as a readout of ER membrane protein mobility. Consistent with the previous study, Sec63-GFP diffusion in WT cells decreased following hyperosmotic stress (Fig. 7, A–D). Importantly, protein mobility within the ER lumen, detected by the HDEL-GFP reporter, was unaffected showing that the change was specific to the ER membrane (Fig. 7, E and F). Strikingly, this decrease in Sec63-GFP mobility did not occur in dga1Δlro1Δ cells (Fig. 7, G–J). Thus, TAG synthesis is required for the reduction in membrane protein mobility, and by inference membrane fluidity, that accompanies hyperosmotic stress.

Figure 7.
A multi-panel image depicts the effects of hyperosmotic stress on membrane properties and cell growth. Panel A shows fluorescence microscopy images of WT Sec63-GFP cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, at 10 seconds before, 0 seconds, and 10 seconds after the indicated time point, with 3 micrometer scale bars. Panel B shows a line graph of normalized Sec63-GFP intensity over time in seconds for WT cells under untreated and 1 molar potassium chloride conditions. Panel C shows a violin plot of Sec63-GFP intensity recovery rate, expressed as percent per second, for WT cells under untreated and 1 molar potassium chloride conditions, with a p-value of 0.0037. Panel D shows a violin plot of half-time recovery in seconds for WT cells under untreated and 1 molar potassium chloride conditions, with a p-value of 0.0114. Panel E shows a violin plot of HDEL-GFP intensity recovery rate, expressed as percent per second, for WT cells under untreated and 1 molar potassium chloride conditions, with a non-significant result. Panel F shows a violin plot of half-time recovery in seconds for WT cells under untreated and 1 molar potassium chloride conditions. Panel G shows fluorescence microscopy images of dga1 deletion lro1 deletion Sec63-GFP cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, at 10 seconds before, 0 seconds, and 10 seconds after the indicated time point, with 3 micrometer scale bars. Panel H shows a line graph of normalized Sec63-GFP intensity over time in seconds for dga1 deletion lro1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel I shows a violin plot of Sec63-GFP intensity recovery rate, expressed as percent per second, for dga1 deletion lro1 deletion cells under untreated and 1 molar potassium chloride conditions, with a non-significant result. Panel J shows a violin plot of half-time recovery in seconds for dga1 deletion lro1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel K shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for WT cells at 30 degrees Celsius under untreated conditions, 1 molar potassium chloride, 0.1 percent BZA, and 1 molar potassium chloride plus 0.1 percent BZA conditions. Panel L shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for dga1 deletion lro1 deletion cells at 30 degrees Celsius under untreated conditions, 1 molar potassium chloride, 0.1 percent BZA, and 1 molar potassium chloride plus 0.1 percent BZA conditions. Panel M shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for dga1 deletion lro1 deletion cells at 25 degrees Celsius under untreated conditions, 1 molar potassium chloride, 0.1 percent BZA, and 1 molar potassium chloride plus 0.1 percent BZA conditions.

Membrane adaptation to hyperosmotic stress requires TAG synthesis. (A) Representative image of endogenously expressing Sec63-GFP WT cell midsections subjected to FRAP. Cells were subject to 0 or 3 h of treatment with 1M KCl in SCD medium. (B) Normalized fluorescence recovery curves of regions of interest from A. The main line shows the mean, with the error bars showing SEM from 48 (UT) and 54 (KCl) regions subject to FRAP from three biological replicates. (C) IRR calculated from B. This parameter is independent of the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 48 (UT) and 54 (KCl) regions of interest from three biological replicates. (D) Half-time recovery (T1/2) calculated from B. This parameter takes into account the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 45 (UT) and 52 (KCl) regions of interest from three biological replicates. (E) IRR for the luminal ER substrate HDEL-GFP. This parameter is independent of the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 43 (UT) and 40 (KCl) regions of interest from three biological replicates. (F) Half-time recovery (T1/2) for the luminal ER substrate HDEL-GFP. This parameter takes into account the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 39 (UT) and 39 (KCl) regions of interest from three biological replicates. (G) Representative image of endogenously expressing Sec63-GFP dga1Δlro1Δ cell midsections subjected to FRAP. Cells were subject to 0 or 3 h of treatment with 1M KCl in SCD medium. (H) Normalized fluorescence recovery curves of regions of interest from G. The main line shows the mean, with the error bars showing SEM from 59 (UT) and 55 (KCl) regions subject to FRAP from three biological replicates. (I) IRR calculated from H. This parameter is independent of the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 59 (UT) and 55 (KCl) regions of interest from three biological replicates. (J) Half-time recovery (T1/2) calculated from H. This parameter takes into account the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 59 (UT) and 55 (KCl) regions of interest from three biological replicates. (K) Growth assay of Pln1-mCherry WT cells at 30°C in SCD and SCD + 1M KCl with and without supplementation of 0.1% BZA. The main line shows the mean, with the error bars showing SEM from 4 biological replicates. (L) Growth assay of Pln1-mCherry dga1Δlro1Δ cells at 30οC in SCD and SCD + 1M KCl with and without supplementation of 0.1% BZA. The main line shows the mean, with the error bars showing SEM from four biological replicates. (M) Growth assay of Pln1-mCherry dga1Δlro1Δ cells at 25°C in SCD and SCD + 1M KCl with and without supplementation of 0.1% BZA. The main line shows the mean, with the error bars showing SEM from four biological replicates. IRR, initial recovery rates.

Figure 7.
A multi-panel image depicts the effects of hyperosmotic stress on membrane properties and cell growth. Panel A shows fluorescence microscopy images of WT Sec63-GFP cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, at 10 seconds before, 0 seconds, and 10 seconds after the indicated time point, with 3 micrometer scale bars. Panel B shows a line graph of normalized Sec63-GFP intensity over time in seconds for WT cells under untreated and 1 molar potassium chloride conditions. Panel C shows a violin plot of Sec63-GFP intensity recovery rate, expressed as percent per second, for WT cells under untreated and 1 molar potassium chloride conditions, with a p-value of 0.0037. Panel D shows a violin plot of half-time recovery in seconds for WT cells under untreated and 1 molar potassium chloride conditions, with a p-value of 0.0114. Panel E shows a violin plot of HDEL-GFP intensity recovery rate, expressed as percent per second, for WT cells under untreated and 1 molar potassium chloride conditions, with a non-significant result. Panel F shows a violin plot of half-time recovery in seconds for WT cells under untreated and 1 molar potassium chloride conditions. Panel G shows fluorescence microscopy images of dga1 deletion lro1 deletion Sec63-GFP cells under untreated conditions and after 3 hours of 1 molar potassium chloride treatment, at 10 seconds before, 0 seconds, and 10 seconds after the indicated time point, with 3 micrometer scale bars. Panel H shows a line graph of normalized Sec63-GFP intensity over time in seconds for dga1 deletion lro1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel I shows a violin plot of Sec63-GFP intensity recovery rate, expressed as percent per second, for dga1 deletion lro1 deletion cells under untreated and 1 molar potassium chloride conditions, with a non-significant result. Panel J shows a violin plot of half-time recovery in seconds for dga1 deletion lro1 deletion cells under untreated and 1 molar potassium chloride conditions. Panel K shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for WT cells at 30 degrees Celsius under untreated conditions, 1 molar potassium chloride, 0.1 percent BZA, and 1 molar potassium chloride plus 0.1 percent BZA conditions. Panel L shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for dga1 deletion lro1 deletion cells at 30 degrees Celsius under untreated conditions, 1 molar potassium chloride, 0.1 percent BZA, and 1 molar potassium chloride plus 0.1 percent BZA conditions. Panel M shows a line graph of cell count, measured as optical density at 600 nanometers, over time in hours for dga1 deletion lro1 deletion cells at 25 degrees Celsius under untreated conditions, 1 molar potassium chloride, 0.1 percent BZA, and 1 molar potassium chloride plus 0.1 percent BZA conditions.

Membrane adaptation to hyperosmotic stress requires TAG synthesis. (A) Representative image of endogenously expressing Sec63-GFP WT cell midsections subjected to FRAP. Cells were subject to 0 or 3 h of treatment with 1M KCl in SCD medium. (B) Normalized fluorescence recovery curves of regions of interest from A. The main line shows the mean, with the error bars showing SEM from 48 (UT) and 54 (KCl) regions subject to FRAP from three biological replicates. (C) IRR calculated from B. This parameter is independent of the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 48 (UT) and 54 (KCl) regions of interest from three biological replicates. (D) Half-time recovery (T1/2) calculated from B. This parameter takes into account the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 45 (UT) and 52 (KCl) regions of interest from three biological replicates. (E) IRR for the luminal ER substrate HDEL-GFP. This parameter is independent of the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 43 (UT) and 40 (KCl) regions of interest from three biological replicates. (F) Half-time recovery (T1/2) for the luminal ER substrate HDEL-GFP. This parameter takes into account the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 39 (UT) and 39 (KCl) regions of interest from three biological replicates. (G) Representative image of endogenously expressing Sec63-GFP dga1Δlro1Δ cell midsections subjected to FRAP. Cells were subject to 0 or 3 h of treatment with 1M KCl in SCD medium. (H) Normalized fluorescence recovery curves of regions of interest from G. The main line shows the mean, with the error bars showing SEM from 59 (UT) and 55 (KCl) regions subject to FRAP from three biological replicates. (I) IRR calculated from H. This parameter is independent of the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 59 (UT) and 55 (KCl) regions of interest from three biological replicates. (J) Half-time recovery (T1/2) calculated from H. This parameter takes into account the mobile fraction. Violin plots show the spread of data, with the dashed lines indicating the median and the dotted lines the upper and lower quartiles. N = 59 (UT) and 55 (KCl) regions of interest from three biological replicates. (K) Growth assay of Pln1-mCherry WT cells at 30°C in SCD and SCD + 1M KCl with and without supplementation of 0.1% BZA. The main line shows the mean, with the error bars showing SEM from 4 biological replicates. (L) Growth assay of Pln1-mCherry dga1Δlro1Δ cells at 30οC in SCD and SCD + 1M KCl with and without supplementation of 0.1% BZA. The main line shows the mean, with the error bars showing SEM from four biological replicates. (M) Growth assay of Pln1-mCherry dga1Δlro1Δ cells at 25°C in SCD and SCD + 1M KCl with and without supplementation of 0.1% BZA. The main line shows the mean, with the error bars showing SEM from four biological replicates. IRR, initial recovery rates.

Close Figure 7.

Finally, we assessed the physiological significance of these changes in membrane properties. If reduced membrane fluidity contributes to adaptation, then artificially increasing membrane fluidity should impair growth. We therefore treated cells with benzyl alcohol (BZA), which increases membrane fluidity by intercalating into lipid bilayers and disrupting lipid packing (Mukhopadhyay et al., 2002). BZA sensitized both WT and dga1Δlro1Δ cells to hyperosmotic stress but had little effect in its absence (Fig. 7, K and L), demonstrating the importance of membrane homeostasis for growth under these conditions. Conversely, lowering the temperature reduces membrane fluidity (Renne and Ernst, 2023) and should alleviate the growth defect of dga1Δlro1Δ cells. Consistent with this prediction, reducing the growth temperature largely abolished the growth defect of dga1Δlro1Δ cells under hyperosmotic stress (Fig. 7 M). These results are consistent with decreased membrane fluidity being an important adaptive feature toward hyperosmotic stress, which is curtailed in the absence of TAG synthesis, reducing cell fitness.

We showed that chronic hyperosmotic stress triggers sustained and adaptive LD accumulation. Yeast cells frequently encounter sustained hyperosmotic environments in nature, for example, on desiccating fruit substrates and in industrial fermentations such as brewing, where high sugar concentrations impose prolonged osmotic stress. Unlike the rapid and transient LD response observed during acute osmotic shock (Phan et al., 2025), chronic stress induces persistent LD accumulation that depends predominantly on Dga1-mediated TAG synthesis and is required for optimal growth. LD accumulation is not directly driven by HOG signaling, but instead becomes more prominent when osmotic balance is not fully restored. Thus, LD accumulation represents a parallel adaptive response to HOG signaling to support cell growth under hyperosmotic conditions.

Mechanistically, LD accumulation during chronic stress is distinct from the acute phase. Acute osmotic shock primarily engages Lro1-dependent TAG production (Phan et al., 2025), whereas chronic stress relies on Dga1. This shift in enzymatic dependence suggests that chronic hyperosmotic stress activates a distinct lipid storage program, likely reflecting a transition from acute membrane slack buffering to long-term membrane adaptation.

Beyond the increase in LDs, persistent hyperosmotic stress did not cause widespread organelle remodeling, with the morphology of the ER and mitochondria being largely preserved, even in cells unable to form LDs. This specificity argues that LD accumulation does not occur as a secondary consequence of generalized stress. Instead, LD accumulation appears to support the full extent of hyperosmotic stress-triggered lipidome remodeling. Beyond the expansion of neutral lipids, we observed a shift in phospholipid composition, characterized by an elevation in PC and a concomitant reduction in PE. While the PE decrease occurs independently of LD formation, the magnitude of PC elevation is attenuated in cells lacking LDs. Our genetic and metabolic perturbations further demonstrate that increased PC contributes to fitness during chronic hyperosmotic stress, particularly when TAG synthesis is compromised. Together, these data support a model in which TAG synthesis and phospholipid remodeling both contribute to hyperosmotic stress adaptation, with LDs amplifying the increase in PC rather than driving the phospholipid changes as a whole. In the future, it will be interesting to investigate how lipid flux is rerouted during this response, including the source of the acyl chains channeled into TAG and the turnover of individual phospholipids, questions that will require metabolic labeling approaches. Nonetheless, our findings expand the role of LDs beyond neutral lipid storage, implicating them in membrane lipid remodeling during chronic environmental stress.

The selective requirement for Dga1 may be particularly important in this context. Unlike Lro1, which can generate TAG by transferring acyl chains from membrane phospholipids, including PC, Dga1 utilizes DAG and activated fatty acids (Oelkers et al., 2000). Preferential involvement of Dga1 during chronic stress may therefore promote TAG synthesis while preserving membrane phospholipid pools, facilitating lipid remodeling without depleting structural lipids. Consistent with this interpretation, TAG-deficient cells constitutively upregulate PC biosynthetic enzymes, yet remain compromised for growth under stress, suggesting that LD formation supports functional phospholipid adaptation beyond simple enzyme induction.

Interestingly, the enzymes driving these two responses appear to be controlled differently: Dga1 induction under stress involves a transcriptional component, whereas the increase in the PC biosynthetic enzymes Cho2 and Opi3, both during stress and upon TAG depletion, occurs without a corresponding change in mRNA and is therefore posttranscriptional. The molecular basis of this posttranscriptional regulation, and how it is engaged by both hyperosmotic stress and the loss of TAG synthesis, remains to be determined.

Consistent with previous observations that persistent hyperosmotic stress alters membrane physical properties such as fluidity, we observed reduced ER membrane protein mobility under these conditions (Laroche et al., 2001). Membrane fluidity changes are most commonly associated with enhanced lipid saturation and/or length, as well as ergosterol concentration (Renne and Ernst, 2023). Although we observed acyl-chain remodeling within individual phospholipid classes, the bulk saturation and length of membrane phospholipids were not substantially altered, pointing to a different underlying mechanism. The major physiological change upon hyperosmotic stress is elevated intracellular glycerol. Glycerol, and other sugars, can interact directly with phospholipid headgroups to alter membrane properties, decreasing lipid diffusion and increasing stiffness (van den Bogaart et al., 2007; Pocivavsek et al., 2011; Abou-Saleh et al., 2019; Schaich et al., 2020). This change could explain the observed decrease in membrane protein diffusion in the WT cells. We therefore speculate that glycerol acts on the membrane through two opposing effects: interactions at the phospholipid headgroups that reduce diffusion, and, when glycerol penetrates more deeply, intercalation into the bilayer that disrupts lipid packing and would instead tend to increase diffusion (Pereira and Hünenberger, 2008). PC, a fluidity-enhancing lipid with a bulky, weakly interactive headgroup (Seu et al., 2006; Dawaliby et al., 2011; Fajardo et al., 2011), engages glycerol less than PE and may buffer these interactions. In WT cells, the stress-induced rise in PC would favor controlled, headgroup-level interactions, yielding the net decrease in diffusion we observe. In TAG-deficient cells, where the PC increase is blunted, greater glycerol intercalation could counteract this stiffening, consistent with the failure of these cells to reduce membrane protein diffusion under stress. In this way, increased PC may help maintain membrane homeostasis during sustained hyperosmotic challenge.

More broadly, accumulation of osmolytes such as glycerol is a conserved feature of hyperosmotic adaptation across eukaryotes (Burg and Ferraris, 2008). Salt tolerance in plants is enhanced by elevated PC levels and is associated with TAG accumulation (Guo et al., 2019; Mueller et al., 2015; Corti et al., 2023). Renal epithelial cells, which experience chronic hyperosmotic stress, are enriched in both PC and TAG (Toback, 1984; Surma et al., 2021), and LD and PC accumulation are hallmarks of cancer cells subjected to mechanical and osmotic compression (Linke et al., 2024; Quan et al., 2025; Li et al., 2025). Our findings suggest that coordinated LD expansion and phospholipid remodeling may represent a conserved strategy for preserving membrane function during sustained osmotic imbalance.

Yeast growth conditions

All strains used were derived from the BY4741 collection, with vectors sourced from Williams et al. (2024), and are listed in Table S1. Mutants were made either using homologous recombination as described in detail elsewhere, or by crossing and dissection using a Singer SporePlay + microscope, followed by genotype determination. Strains that had not previously been verified were verified by performing PCRs within the deleted genes.

Cells were grown at 30°C on YPD agar plates (10 g/l yeast extract, 20 g/l peptone, 2% glucose, 20 g/l agar) before being used for the described experiments. For all growth assays, cells were grown in SCD (6.7 g/l YNB, 0.6 g/l CSM) in 16-mm-diameter glass tubes at 30°C, 180–200 rpm, for a minimum of 14 h to allow them to enter the logarithmic phase. Cells were diluted to 0.05 OD600 nm in 200 μl of the indicated media in a well of a flat-bottomed 96-well plate. The plate was incubated in a VANTAstar (BMG Labtech) at the indicated temperatures for the indicated times, with 30 s of vigorous shaking followed by OD600 nm measurement every 15 min.

Flow cytometry and analysis

Cells were grown to the logarithmic phase the day before they were due to be analyzed in SCD (16-mm-diameter glass tubes at 30°C, 180–200 rpm, for a minimum of 4 h), unless otherwise indicated. Overnight they were diluted into 25-mm-diameter glass tubes or conical flasks, such that they would grow to ∼0.05–0.5 OD600 nm after 16 h at 30°C, 180–200 rpm. Samples were taken before treatment and stored on ice, then subjected to the indicated treatment with samples taken at the indicated times. These samples were then subject to flow cytometry analysis using a Fortessa (BD Biosciences) controlled by FACSDiva software (BD Biosciences) with a High-Throughput Sampler attachment set to record 10,000 events. Subsequent analysis was performed on FCS3.0 files using FlowJo (v10.10.0) to isolate bona fide single cells through gating, and the median relevant fluorescence value (either GFP or mCherry) for this population was recorded. From this median value, the median value of an unlabeled sample run alongside it was subtracted to obtain the final value.

Microscopy and analysis

Cells were grown as for the flow cytometry experiments, with the exception that the overnight setup aimed for the cells to grow to between 0.2 and 0.5 OD600 nm. Samples were taken at the indicated treatment times and labeled with MDH (0.1 mM AutoDot, SM1000b; Abgent) or BODIPY 493/503 (1 µg/ml, Invitrogen) where indicated and taken straight for imaging without washing. Imaging was performed at room temperature on either a Zeiss Axio Observer with a 63× objective (Plan-APOCHROMAT 1.4) using a digital CMOS camera (ORCA Flash 4.0; Hamamatsu) and SlideBook 6.0 (3i) software, or Olympus FV3000 microscope with a 60× objective (Plan-APO N 1.4) using internal detectors and FV (Olympus) software.

Analysis and image preparation were performed using FIJI. For analysis, images of Ilv6-GFP, as well as MDH- and BODIPY-stained cells, underwent Z-projection using the standard deviation setting. Cho2-GFP, Opi3-GFP, and Sec63-GFP were imaged at the midsection. BODIPY-stained LDs were detected using ComDet (v.0.5.6) to isolate spherical particles and define their cross-sectional area in pixels, which was mathematically converted to diameter in Microsoft Excel. Detected LDs per cell were counted manually following ComDet particle identification, with budding cells counted as one cell if the bud was <50% of the size of the mother, and two separate cells if it was >50% the size of the mother. Any detected particle outside a cell was excluded from further analysis. Mitochondrial fragmentation was scored by eye, with cells having either completely elongated, tubular mitochondria; completely fragmented, punctate mitochondria; or a mixture of both.

FRAP microscopy and analysis

Cells were prepared as for the above microscopy analysis, with the FRAP assays performed at 30°C on a SoRa spinning disk (Olympus) microscope with 60× objective (UPlanApo 1.5) and a Photometrics 95B camera using cellSens software. Cells were imaged for five frames before FRAP was performed in a stripe across the cell, and then, imaging proceeded for another 60 frames, each at 300-ms exposure. Sec63-GFP cells were imaged with a frame every 2 s, while HDEL-GFP had a frame every 0.5 s.

Analysis was performed as previously described (Shyu et al., 2019). Briefly, the intensities over time were taken from membranes subjected to FRAP, and not subjected to FRAP on the same cell using FIJI. The following equations were then used to define the values for each time point, normalizing for photobleaching, the initial fluorescence of the cell, and the lowest fluorescence of the bleached area.
From these data, the initial recovery rates (which show recovery independently of the mobile fraction) and the half-time recovery (which takes into account the mobile fraction) were calculated for each region subjected to FRAP using GraphPad Prism. The initial recovery rate was calculated by performing linear regression analysis to obtain the gradient during the initial stages of recovery where the average was linear (0–10 s for Sec63-GFP, 0–1.5 s for HDEL-GFP). The gradient was then converted into % recovery per second. The half-time recovery (T1/2) was calculated by fitting the entire recovery curve to an exponential plateau curve, yielding values for the Y intersect (Y0, set as 0), the maximum Y value (Ym), and a rate constant, K. The following calculation was then performed to obtain the X value in seconds for a Y value of 0.5 (i.e., halfway recovered). Values where the curve was a poor fit were excluded.

Protein extraction and western blotting

Cells were grown as for the flow cytometry experiments, with the overnight growth occurring in 5–10 ml SCD in a conical flask. Cells were then diluted down to 0.2 OD600 nm in SCD + 1M KCl, and samples of 2 OD600 nm were harvested at the indicated times. UT samples were taken from the overnight culture prior to dilution. Samples were spun down in a benchtop centrifuge (8,000 rpm, 30 sec) and resuspended in 300 μl ice-cold 150 mM NaOH. After incubation for 10 min on ice, the cells were spun down again and resuspended in 100 μl lysis buffer (0.1M Tris-HCl, pH 6.8, 2% SDS). Samples were incubated at 65°C with shaking in an Eppendorf ThermoMixer F1.5 for 15 min, then brief spun to remove cell debris. The protein concentration of the supernatant was determined on a NanoDrop, and the samples were adjusted to the lowest value. 80 μl of sample was then mixed with 20 μl loading dye (50% glycerol, 0.05% bromophenol blue) and stored at −20°C prior to use.

For western blot analysis, ∼100 µg of protein samples was run on 4–20% Criterion TGX Precast SDS-PAGE gels (Bio-Rad) at 0.06 Amps per gel for 45 min. Blotting was performed using the Trans-Blot Turbo semidry system (Bio-Rad) onto PVDF membranes. Gel running and blotting were performed according to the manufacturer’s instructions. After blotting, membranes were blocked in 5% milk for 1 h, then incubated overnight with primary antibodies. Following primary antibody incubation, secondaries were applied for at least 1 h before imaging was performed using Western Lightning ECL Pro Enhanced Luminol Reagent (Revvity) on an Amersham Imager 600 (General Electric). Between each of these steps, membranes were twice rinsed with PBS-T (PBS with 0.1% Tween-20), followed by three 15-min washes. Primary antibodies were all diluted in PBS-T/1% milk, 0.0002% sodium azide and stored at 4°C between uses. Secondary antibodies were diluted 1:10,000 in PBS-T and used once. Primary antibodies used in this study were as follows: α-Dga1 (1:3,000, rabbit, homemade), α-Flag (1:2,000, mouse, A8592-2MG; Sigma-Aldrich), α-HA (1:2,000, rat, 3F10; Roche), α-Myc (1:1,000, mouse, 9E10; Thermo Fisher Scientific), α-tubulin (1:1,000, rat, sc-53030; Santa Cruz), α-PGK1 (1:5,000, mouse, 22C5D8; Thermo Fisher Scientific). Secondary antibodies used in this study were as follows: α-rabbit IgG (mouse, 211-032-171; Jackson ImmunoResearch), α-rat IgG (goat, 112-035-175; Jackson ImmunoResearch), α-mouse IgG (goat, 115-035-174; Jackson ImmunoResearch).

Western blot analysis was performed using FIJI, with the average band intensity determined by drawing a box around it and using the measure feature. Intensity was normalized to that of the loading control (tubulin or PGK1) to compensate for any errors in loading, then to the level of the untreated sample to account for differences between different membranes and imaging times.

Lipidomic sample preparation

Cells were crashed out of frozen stocks onto an YPD agar plate and grown overnight at 30°C. Cells were then resuspended to ∼0.2 OD600 nm in 2 ml SCD in a 16-mm-diameter tube and grown for 4 h at 30°C, 180 rpm. They were then diluted to 0.07 OD600 nm in 5 ml SCD in a 25-mm-diameter tube. After a further 4-h growth at 30°C, 180 rpm, they were diluted into 250 ml SCD in a 1-liter conical flask so they would reach ∼0.4–0.8 OD600 nm after 16 h (WT and hog1Δ: to 0.0008 OD600 nm, are1Δare2Δdga1Δlro1Δ: to 0.001 OD600 nm). In the morning, cells were inoculated in 250 ml SCD in 1-liter conical flasks at a concentration of 0.1 or 0.25 OD600 nm for UT and KCl-treated samples, respectively, and returned to the incubator. After 3 h, equal OD600 nm of all samples was taken and spun down using a JLA10.500 rotor (12 min, 4,000 rpm, 4°C). Pellets were resuspended in 10 ml ice-cold water, transferred to a 50-ml tube, and spun down again using a benchtop centrifuge (5 min, 3,900 rpm, 4°C). The pellet was resuspended in 1 ml of ice-cold water, and the OD600 nm was determined. 20 OD of cells was transferred to individual precooled 1.5-ml tubes. For all except 1 tube, these were spun down again and the pellets flash-frozen in liquid nitrogen before long-term storage at −70°C. For the other sample, the volume in the tube was made up to 500 μl by the addition of ice-cold water and transferred to a 2-ml tube containing 0.5-ml 0.5-mm-diameter glass beads (Thistle Scientific). Tubes were sealed with parafilm and subjected to lysis by bead beating at 4°C (OMNI International Bead Ruptor Elite, 10 × 2.9m/s 1 min on, 3 min off). After lysis, 0.5 ml ice-cold water was added to the samples and mixed by pipetting, and 0.5 ml was transferred without beads to a new 2-ml tube, which was flash-frozen in liquid nitrogen. Samples were stored at −70°C or on dry ice until they were ready to undergo analysis.

Lipidomic analysis

Lipidomic mass spectrometry analysis was performed by Lipotype GmbH. Lipids were extracted using chloroform and methanol (Klose et al., 2012). Samples were spiked with lipid class-specific internal standards prior to extraction. After drying and resuspending in MS acquisition mixture, lipid extracts were subjected to mass spectrometric analysis. Mass spectra were acquired on a hybrid quadrupole/Orbitrap mass spectrometer equipped with an automated nanoflow electrospray ion source in both positive and negative ion modes. Lipid identification was performed using LipotypeXplorer (Herzog et al., 2012) on unprocessed (*.raw format) mass spectra. For MS-only mode, lipid identification was based on the molecular masses of the intact molecules. The MSMS mode included the collision-induced fragmentation of lipid molecules, and lipid identification was based on both the intact masses and the masses of the fragments. Prior to normalization and further statistical analysis, lipid identifications were filtered according to mass accuracy, occupation threshold, noise, and background. Lists of identified lipids and their intensities were stored in a database optimized for the particular structure inherent to lipidomic datasets. Intensity of lipid class-specific internal standards was used for lipid quantification. Data are presented in either mol%, where the total molarity of the defined class adds up to 100%, or normalized pmol, where the picomoles of each lipid are normalized to account for the differing volumes of sample analyzed.

Determination of mRNA levels by RT-qPCR

Cells were grown and treated as described above. 1 OD600nm of cells was lysed per condition by bead beating using the settings described in the lipidomic analysis, and RNA was isolated using the Monarch Spin RNA isolation kit (NEB) according to the manufacturer’s instructions. Triplicate RT-qPCRs were performed using Luna Universal One-Step RT-qPCR Kit (NEB) according to the manufacturer’s instructions using QuantStudio 3 qPCR System (Applied Biosystems) using the Comparative Ct with Melt setting for primers targeted against the indicated mRNAs, and the housekeeping mRNA TUB2 (Table S2). For each mRNA and condition, the difference in Ct was calculated between the mRNA of interest and TUB2 (ΔCt). The difference between those values and the equivalent WT UT was calculated (ΔΔCt) and converted to a fold change in mRNA expression.

Statistical analysis

All statistical analysis and graph generation were performed using GraphPad Prism, with error bars representing the standard error. For dual-variable grouped graphs (Fig. 2 B; Fig. 3, B, F, and G; Fig. 4, A, B, D, and E; Fig. S1, C–G, and I; Fig. S2, A, D, H, and L–N; Fig. S3, B, F–I, and Q; Fig. S4, B–H; Fig. S5, B, D–F, and H), two-way ANOVA between matched replicates was performed, followed by multiple comparisons analysis between samples of the same genotype and samples of the same treatment condition. For single-variable column graphs (Fig. 1, B, C, E, and F; Fig. 2, D and E; Fig. 3, D and E; Fig. S1 B; Fig. S2, C, G, J, and K; and Fig. S3, D, E, I, and J), one-way ANOVA with matched replicates was performed, followed by multiple comparisons between all samples. Paired t tests were performed for Fig. S1 J, Fig. S3, M–P; and Fig. 7, C–F, I, and J. For Fig. S3 L, a one-sample t test against an expected mean of 1 was performed. For Fig. 6 E and Fig. S5 K, Tukey’s multiple comparisons tests were performed.

Online supplemental material

Fig. S1 shows the utility of using Pln1-mCherry as a proxy for LD formation, and how levels of this protein respond to a variety of cellular stresses. Fig. S2 explores the role of the seipin complex, as well as TAG and STE synthesis in regulating LD accumulation upon chronic hyperosmotic stress. It further looks at the roles of endocytic and lipid trafficking proteins reported to be involved in a LD response to acute hyperosmotic stress. Fig. S3 examines the roles of the DAG and acyl-FA machinery in LD accumulation and the regulation at the protein and mRNA level of the proteins which synthesize TAGs and STEs. Fig. S4 shows the effect on Pln1-mCherry levels following hyperosmotic stress of manipulating the HOG and other stress-responsive signaling pathways. Fig. S5 contains more details from the lipidomic experiment which are presented in Fig. 5, alongside an examination of Cho2 and Opi3 mRNA regulation. Tables S1 and S2 show the strains and primers used in this study, respectively.

All data used in this study are available upon request.

We thank R. Hedley and V. Tsioligka for technical assistance in the analysis of flow cytometry experiments at The Don Mason Facility of Flow Cytometry, and A. Wainman for help with imaging at the Light Microscopy Facility. We thank all members of the Carvalho lab for discussions and Z. Ji, L. Lemus, and R. Klemm for critical reading of the manuscript.

T.D. Williams was supported by the Biotechnology and Biological Sciences Research Council (BBSRC) grant (BB/W001519/1) and by an Eric Reid Fund for Methodology grant awarded by the Biochemical Society. P. Carvalho was supported by a BBSRC grant (BB/W001519/1) and an investigator award from The Wellcome Trust (223153/Z/21/Z). Open Access funding provided by University of Oxford.

Author contributions: Thomas D. Williams: conceptualization, data curation, formal analysis, investigation, methodology, project administration, validation, visualization, and writing—original draft, review, and editing. Christian Klose: methodology. Robert Ernst: funding acquisition, project administration, supervision, visualization, and writing—review and editing. Pedro Carvalho: conceptualization, data curation, funding acquisition, project administration, resources, supervision, validation, visualization, and writing—original draft.

Abou-Saleh
,
R.H.
,
J.R.
McLaughlan
,
R.J.
Bushby
,
B.R.
Johnson
,
S.
Freear
,
S.D.
Evans
, and
N.H.
Thomson
.
2019
.
Effects of glycerol on lipid monolayers at the gas-liquid interface: Impact on microbubble physical and mechanical properties
.
Langmuir
.
35
:
10097
–
10105
.
Athenstaedt
,
K.
2010
.
Neutral lipids in yeast: Synthesis, storage and degradation
. In
Handbook of Hydrocarbon and Lipid Microbiology
.
K.N.
Timmis
, editor.
Springer Berlin Heidelberg
,
Berlin, Heidelberg
.
471
–
480
.
van den Bogaart
,
G.
,
N.
Hermans
,
V.
Krasnikov
,
A.H.
de Vries
, and
B.
Poolman
.
2007
.
On the decrease in lateral mobility of phospholipids by sugars
.
Biophys. J.
92
:
1598
–
1605
.
Burg
,
M.B.
, and
J.D.
Ferraris
.
2008
.
Intracellular organic osmolytes: Function and regulation
.
J. Biol. Chem.
283
:
7309
–
7313
.
Burgos
,
J.I.
,
M.
Morell
,
J.I.E.
Mariángelo
, and
M.
Vila Petroff
.
2019
.
Hyperosmotic stress promotes endoplasmic reticulum stress-dependent apoptosis in adult rat cardiac myocytes
.
Apoptosis
.
24
:
785
–
797
.
Cartwright
,
B.R.
, and
J.M.
Goodman
.
2012
.
Seipin: From human disease to molecular mechanism
.
J. Lipid Res.
53
:
1042
–
1055
.
Corti
,
E.
,
S.
Falsini
,
S.
Schiff
,
C.
Tani
,
C.
Gonnelli
, and
A.
Papini
.
2023
.
Saline stress impairs lipid storage mobilization during germination in Eruca sativa
.
Plants
.
12
:
366
.
Dawaliby
,
R.
,
C.
Trubbia
,
C.
Delporte
,
J.M.
Ruysschaert
,
P.
Van Antwerpen
, and
C.
Govaerts
.
2011
.
Phosphatidylethanolamine is a key regulator of membrane fluidity in eukaryotic cells
.
J. Biol. Chem.
291
:
3658
–
3667
.
Dubots
,
E.
,
S.
Cottier
,
M.-P.
Péli-Gulli
,
M.
Jaquenoud
,
S.
Bontron
,
R.
Schneiter
, and
C.
De Virgilio
.
2014
.
TORC1 regulates Pah1 phosphatidate phosphatase activity via the Nem1/Spo7 protein phosphatase complex
.
PLoS One
.
9
:e104194.
Ernst
,
R.
,
C.S.
Ejsing
, and
B.
Antonny
.
2016
.
Homeoviscous adaptation and the regulation of membrane lipids
.
J. Mol. Biol.
428
:
4776
–
4791
.
Fajardo
,
V.A.
,
L.
McMeekin
, and
P.J.
Gervais
.
2011
.
Influence of phospholipid species on membrane fluidity: A meta-analysis for a novel phospholipid fluidity index
.
J. Membr. Biol.
244
:
97
–
103
.
Farese
,
R.V.
, and
T.C.
Walther
.
2025
.
Essential biology of lipid droplets
.
Annu. Rev. Biochem.
94
:
447
–
477
.
Fedoseeva
,
I.V.
,
D.V.
Pyatrikas
,
A.V.
Stepanov
,
A.V.
Fedyaeva
,
N.N.
Varakina
,
T.M.
Rusaleva
,
G.B.
Borovskii
, and
E.G.
Rikhvanov
.
2017
.
The role of flavin-containing enzymes in mitochondrial membrane hyperpolarization and ROS production in respiring Saccharomyces cerevisiae cells under heat-shock conditions
.
Sci. Rep.
7
:
2586
.
Gao
,
Q.
,
D.D.
Binns
,
L.N.
Kinch
,
N.V.
Grishin
,
N.
Ortiz
,
X.
Chen
, and
J.M.
Goodman
.
2017
.
Pet10p is a yeast perilipin that stabilizes lipid droplets and promotes their assembly
.
J. Cell Biol.
216
:
3199
–
3217
.
Giménez-Andrés
,
M.
,
T.
Emeršič
,
S.
Antoine-Bally
,
J.M.
D’Ambrosio
,
B.
Antonny
,
J.
Derganc
, and
A.
Čopič
.
2021
.
Exceptional stability of a perilipin on lipid droplets depends on its polar residues, suggesting multimeric assembly
.
Elife
.
10
:e61401.
Grippa
,
A.
,
L.
Buxó
,
G.
Mora
,
C.
Funaya
,
F.-Z.
Idrissi
,
F.
Mancuso
,
R.
Gomez
,
J.
Muntanyà
,
E.
Sabidó
, and
P.
Carvalho
.
2015
.
The seipin complex Fld1/Ldb16 stabilizes ER-lipid droplet contact sites
.
J. Cell Biol.
211
:
829
–
844
.
Guo
,
Q.
,
L.
Liu
, and
B.J.
Barkla
.
2019
.
Membrane lipid remodeling in response to salinity
.
Int. J. Mol. Sci.
20
:
4264
.
Hariri
,
H.
,
S.
Rogers
,
R.
Ugrankar
,
Y.L.
Liu
,
J.R.
Feathers
, and
W.M.
Henne
.
2018
.
Lipid droplet biogenesis is spatially coordinated at ER-vacuole contacts under nutritional stress
.
EMBO Rep.
19
:
57
–
72
.
Heeren
,
J.
, and
L.
Scheja
.
2021
.
Metabolic-associated fatty liver disease and lipoprotein metabolism
.
Mol. Metab.
50
:
101238
.
Henne
,
W.M.
, and
S.
Cohen
.
2026
.
Heterogeneity, dynamics and organelle interactions of lipid droplets
.
Nat. Rev. Mol. Cell Biol.
27
:
433
–
449
.
Herzog
,
R.
,
K.
Schuhmann
,
D.
Schwudke
,
J.L.
Sampaio
,
S.R.
Bornstein
,
M.
Schroeder
, and
A.
Shevchenko
.
2012
.
LipidXplorer: A software for consensual cross-platform lipidomics
.
PLoS ONE
.
7
:e29851.
Hosios
,
A.M.
,
M.E.
Wilkinson
,
M.C.
McNamara
,
K.C.
Kalafut
,
M.E.
Torrence
,
J.M.
Asara
, and
B.D.
Manning
.
2022
.
mTORC1 regulates a lysosome-dependent adaptive shift in intracellular lipid species
.
Nat. Metab.
4
:
1792
–
1811
.
Hughes Hallett
,
J.E.
,
X.
Luo
, and
A.P.
Capaldi
.
2014
.
State transitions in the TORC1 signaling pathway and information processing in Saccharomyces cerevisiae
.
Genetics
.
198
:
773
–
786
.
Huh
,
W.-K.
,
J.V.
Falvo
,
L.C.
Gerke
,
A.S.
Carroll
,
R.W.
Howson
,
J.S.
Weissman
, and
E.K.
O’Shea
.
2003
.
Global analysis of protein localization in budding yeast
.
Nature
.
425
:
686
–
691
.
Jin
,
Y.
,
Y.
Tan
,
J.
Wu
, and
Z.
Ren
.
2023
.
Lipid droplets: A cellular organelle vital in cancer cells
.
Cell Death Discov.
9
:
254
.
Johnson
,
D.R.
,
L.J.
Knoll
,
D.E.
Levin
, and
J.I.
Gordon
.
1994
.
Saccharomyces cerevisiae contains four fatty acid activation (FAA) genes: An assessment of their role in regulating protein N-myristoylation and cellular lipid metabolism
.
J. Cell Biol.
127
:
751
–
762
.
Klemm
,
R.W.
, and
P.
Carvalho
.
2024
.
Lipid droplets big and small: Basic mechanisms that make them all
.
Annu. Rev. Cell Dev. Biol.
40
:
143
–
168
.
Klipp
,
E.
,
B.
Nordlander
,
R.
Krüger
,
P.
Gennemark
, and
S.
Hohmann
.
2005
.
Integrative model of the response of yeast to osmotic shock
.
Nat. Biotechnol.
23
:
975
–
982
.
Klose
,
C.
,
M.A.
Surma
,
M.J.
Gerl
,
F.
Meyenhofer
,
A.
Shevchenko
, and
K.
Simons
.
2012
.
Flexibility of a eukaryotic lipidome--insights from yeast lipidomics
.
PLoS ONE
.
7
:e35063.
Klug
,
Y.A.
,
J.V.
Ferreira
, and
P.
Carvalho
.
2024
.
A unifying mechanism for seipin-mediated lipid droplet formation
.
FEBS Lett.
598
:
1116
–
1126
.
Kondo
,
W.
,
T.
Kitagawa
,
H.
Hoshida
,
R.
Akada
, and
I.
Miyakawa
.
2022
.
Morphological changes of mitochondria and actin cytoskeleton in the yeast Saccharomyces cerevisiae during diauxic growth and glucose depletion culture
.
Cytologia (Tokyo)
.
87
:
157
–
162
.
Lange
,
M.
,
M.
Wölk
,
V.W.
Li
,
C.E.
Doubravsky
,
J.M.
Hendricks
,
S.
Kato
,
Y.
Otoki
,
B.
Styler
,
S.L.
Johnson
,
C.A.
Harris
, et al
.
2025
.
FSP1-mediated lipid droplet quality control prevents neutral lipid peroxidation and ferroptosis
.
Nat. Cell Biol.
27
:
1902
–
1913
.
Linke
,
J.A.
,
L.L.
Munn
, and
R.K.
Jain
.
2024
.
Compressive stresses in cancer: Characterization and implications for tumour progression and treatment
.
Nat. Rev. Cancer
24
:
768
–
791
.
Liu
,
N.
,
Y.
Yun
,
Y.
Yin
,
M.
Hahn
,
Z.
Ma
, and
Y.
Chen
.
2019
.
Lipid droplet biogenesis regulated by the FgNem1/Spo7-FgPah1 phosphatase cascade plays critical roles in fungal development and virulence in Fusarium graminearum
.
New Phytol.
223
:
412
–
429
.
Laroche
,
C.
,
L.
Beney
,
P.A.
Marechal
, and
P.
Gervais
.
2001
.
The effect of osmotic pressure on the membrane fluidity of Saccharomyces cerevisiae at different physiological temperatures
.
Appl. Microbiol. Biotechnol.
56
:
249
–
254
.
Li
,
L.
,
Y.
Huang
,
Y.
Gui
,
W.
Xiang
,
M.
Yang
,
Y.
Hou
, and
M.
Peng
.
2025
.
The role of phosphatidylcholine metabolism in tumors
.
Med. Oncol.
42
:
450
.
Madeira
,
J.B.
,
C.A.
Masuda
,
C.M.
Maya-Monteiro
,
G.S.
Matos
,
M.
Montero-Lomelí
, and
B.L.
Bozaquel-Morais
.
2015
.
TORC1 inhibition induces lipid droplet replenishment in yeast
.
Mol. Cell. Biol.
35
:
737
–
746
.
Markgraf
,
D.F.
,
R.W.
Klemm
,
M.
Junker
,
H.K.
Hannibal-Bach
,
C.S.
Ejsing
, and
T.A.
Rapoport
.
2014
.
An ER protein functionally couples neutral lipid metabolism on lipid droplets to membrane lipid synthesis in the ER
.
Cell Rep
.
6
:
44
–
55
.
Mathiowetz
,
A.J.
, and
J.A.
Olzmann
.
2024
.
Lipid droplets and cellular lipid flux
.
Nat. Cell Biol.
26
:
331
–
345
.
McMaster
,
C.R.
, and
R.M.
Bell
.
1994
.
Phosphatidylcholine biosynthesis via the CDP-choline pathway in Saccharomyces cerevisiae. Multiple mechanisms of regulation
.
J. Biol. Chem.
269
:
14776
–
14783
.
Mueller
,
S.P.
,
D.M.
Krause
,
M.J.
Mueller
, and
A.
Fekete
.
2015
.
Accumulation of extra-chloroplastic triacylglycerols in arabidopsis seedlings during heat acclimation
.
J. Exp. Bot.
66
:
4517
–
4526
.
Mukhopadhyay
,
K.
,
A.
Kohli
, and
R.
Prasad
.
2002
.
Drug susceptibilities of yeast cells are affected by membrane lipid composition
.
Antimicrob. Agents Chemother.
46
:
3695
–
3705
.
de Nadal
,
E.
, and
F.
Posas
.
2022
.
The HOG pathway and the regulation of osmoadaptive responses in yeast
.
FEMS Yeast Res.
22
:
foac013
.
Nguyen
,
T.B.
, and
J.A.
Olzmann
.
2017
.
Lipid droplets and lipotoxicity during autophagy
.
Autophagy
.
13
:
2002
–
2003
.
Oelkers
,
P.
,
D.
Cromley
,
M.
Padamsee
,
J.T.
Billheimer
, and
S.L.
Sturley
.
2002
.
The DGA1 gene determines a second triglyceride synthetic pathway in yeast
.
J. Biol. Chem.
277
:
8877
–
8881
.
Oelkers
,
P.
,
A.
Tinkelenberg
,
N.
Erdeniz
,
D.
Cromley
,
J.T.
Billheimer
, and
S.L.
Sturley
.
2000
.
A lecithin cholesterol acyltransferase-like gene mediates diacylglycerol esterification in yeast
.
J. Biol. Chem.
275
:
15609
–
15612
.
Pereira
,
C.S.
, and
P.H.
Hünenberger
.
2008
.
Effect of trehalose on a phospholipid membrane under mechanical stress
.
Biophys. J.
95
:
3525
–
3534
.
Petelenz-Kurdziel
,
E.
,
C.
Kuehn
,
B.
Nordlander
,
D.
Klein
,
K.-K.
Hong
,
T.
Jacobson
,
P.
Dahl
,
J.
Schaber
,
J.
Nielsen
,
S.
Hohmann
, and
E.
Klipp
.
2013
.
Quantitative analysis of glycerol accumulation, glycolysis and growth under hyper osmotic stress
.
PLoS Comput. Biol.
9
:e1003084.
Phan
,
J.
,
M.
Silva
,
R.
Kohlmeyer
,
R.
Ruethemann
,
L.
Gay
,
E.
Jorgensen
, and
M.
Babst
.
2025
.
Recovery of plasma membrane tension after a hyperosmotic shock
.
Mol. Biol. Cell
.
36
:
ar45
.
Pocivavsek
,
L.
,
K.
Gavrilov
,
K.D.
Cao
,
E.Y.
Chi
,
D.
Li
,
B.
Lin
,
M.
Meron
,
J.
Majewski
, and
K.Y.C
Lee
.
2011
.
Glycerol-induced membrane stiffening: The role of viscous fluid adlayers
.
Biophys. J.
101
:
118
–
127
.
Pressly
,
J.D.
,
M.Z.
Gurumani
,
J.T.
Varona Santos
,
A.
Fornoni
,
S.
Merscher
, and
H.
Al-Ali
.
2022
.
Adaptive and maladaptive roles of lipid droplets in health and disease
.
Am. J. Physiol. Cell Physiol.
322
:
C468
–
C481
.
Quan
,
J.
,
C.
Zhang
,
X.
Chen
,
X.
Cai
, and
X.
Luo
.
2025
.
Lipid droplet - Organelle crosstalk and its implication in cancer
.
Prog. Biophys. Mol. Biol.
197
:
11
–
20
.
Rambold
,
A.S.
,
S.
Cohen
, and
J.
Lippincott-Schwartz
.
2015
.
Fatty acid trafficking in starved cells: Regulation by lipid droplet lipolysis, autophagy, and mitochondrial fusion dynamics
.
Dev. Cell.
32
:
678
–
692
.
Rao
,
M.J.
,
B.
Folger
,
J.
Reus
,
A.
Toulmay
,
J.
Li
,
R.
Zhang
,
W.
Prinz
,
J.
Goodman
, and
F.
Wang
.
2025
.
Pln1 mediates lipid droplet-vacuole tethering during microlipophagy in Saccharomyces cerevisiae
.
bioRxiv
. .
(Preprint posted May 12, 2025)
.
Renne
,
M.F.
, and
R.
Ernst
.
2023
.
Membrane homeostasis beyond fluidity: Control of membrane compressibility
.
Trends Biochem. Sci.
48
:
963
–
977
.
Roffay
,
C.
,
G.
Molinard
,
K.
Kim
,
M.
Urbanska
,
V.
Andrade
,
V.
Barbarasa
,
P.
Nowak
,
V.
Mercier
,
J.
García-Calvo
,
S.
Matile
, et al
.
2021
.
Passive coupling of membrane tension and cell volume during active response of cells to osmosis
.
Proc. Natl. Acad. Sci. USA
.
118
:e2103228118.
Rogers
,
S.
,
H.
Hariri
,
N.E.
Wood
,
N.O.
Speer
, and
W.M.
Henne
.
2021
.
Glucose restriction drives spatial reorganization of mevalonate metabolism
.
Elife
.
10
:e62591.
Ruggiano
,
A.
,
G.
Mora
,
L.
Buxó
, and
P.
Carvalho
.
2016
.
Spatial control of lipid droplet proteins by the ERAD ubiquitin ligase Doa10
.
EMBO J.
35
:
1644
–
1655
.
Safi
,
R.
,
P.
Menéndez
, and
A.
Pol
.
2024
.
Lipid droplets provide metabolic flexibility for cancer progression
.
FEBS Lett.
598
:
1301
–
1327
.
Sandager
,
L.
,
M.H.
Gustavsson
,
U.
Ståhl
,
A.
Dahlqvist
,
E.
Wiberg
,
A.
Banas
,
M.
Lenman
,
H.
Ronne
, and
S.
Stymne
.
2002
.
Storage lipid synthesis is non-essential in yeast
.
J. Biol. Chem.
277
:
6478
–
6482
.
Schaich
,
M.
,
D.
Sobota
,
H.
Sleath
,
J.
Cama
, and
U.F.
Keyser
.
2020
.
Characterization of lipid composition and diffusivity in OLA generated vesicles
.
Biochim. Biophys. Acta Biomembr.
1862
:
183359
.
Seo
,
A.Y.
,
P.-W.
Lau
,
D.
Feliciano
,
P.
Sengupta
,
M.A.L.
Gros
,
B.
Cinquin
,
C.A.
Larabell
, and
J.
Lippincott-Schwartz
.
2017
.
AMPK and vacuole-associated Atg14p orchestrate μ-lipophagy for energy production and long-term survival under glucose starvation
.
Elife
.
6
:e21690.
Seu
,
K.J.
,
L.R.
Cambrea
,
R.M.
Everly
, and
J.S.
Hovis
.
2006
.
Influence of lipid chemistry on membrane fluidity: Tail and headgroup interactions
.
Biophys. J.
91
:
3727
–
3735
.
Shen
,
W.
,
Z.
Gao
,
K.
Chen
,
A.
Zhao
,
Q.
Ouyang
, and
C.
Luo
.
2023
.
The regulatory mechanism of the yeast osmoresponse under different glucose concentrations
.
iScience
.
26
:
105809
.
Shyu
,
P.
,
B.S.H.
Ng
,
N.
Ho
,
R.
Chaw
,
Y.L.
Seah
,
C.
Marvalim
, and
G.
Thibault
.
2019
.
Membrane phospholipid alteration causes chronic ER stress through early degradation of homeostatic ER-resident proteins
.
Sci. Rep.
9
:
8637
.
Surma
,
M.A.
,
M.J.
Gerl
,
R.
Herzog
,
J.
Helppi
,
K.
Simons
, and
C.
Klose
.
2021
.
Mouse lipidomics reveals inherent flexibility of a Mammalian lipidome
.
Sci. Rep.
11
:
19364
.
Su
,
W.-M.
,
G.-S.
Han
, and
G.M.
Carman
.
2014
.
Yeast Nem1-Spo7 protein phosphatase activity on Pah1 phosphatidate phosphatase is specific for the Pho85-Pho80 protein kinase phosphorylation sites
.
J. Biol. Chem.
289
:
34699
–
34708
.
Teixeira
,
V.
,
T.S.
Martins
,
W.A.
Prinz
, and
V.
Costa
.
2021
.
Target of rapamycin complex 1 (TORC1), protein kinase A (PKA) and cytosolic ph regulate a transcriptional circuit for lipid droplet formation
.
Int. J. Mol. Sci.
22
:
9017
.
Thiam
,
A.R.
, and
E.
Ikonen
.
2021
.
Lipid droplet nucleation
.
Trends Cell Biol.
31
:
108
–
118
.
Toback
,
F.G.
1984
.
Phosphatidylcholine metabolism during renal growth and regeneration
.
Am. J. Physiol.
246
:
F249
–
F259
.
Wang
,
C.-W.
,
R.-H.
Chen
, and
Y.-K.
Chen
.
2024
.
The lipid droplet assembly complex consists of seipin and four accessory factors in budding yeast
.
J. Biol. Chem.
300
:
107534
.
Wang
,
S.
,
F.-Z.
Idrissi
,
M.
Hermansson
,
A.
Grippa
,
C.S.
Ejsing
, and
P.
Carvalho
.
2018
.
Seipin and the membrane-shaping protein Pex30 cooperate in organelle budding from the endoplasmic reticulum
.
Nat. Commun.
9
:
2939
.
Williams
,
T.D.
,
I.
Joshua
,
F.
Soubigou
,
S.M.
Dublanska
,
R.
Bergquist
, and
A.
Rousseau
.
2024
.
Distinct TORC1 signalling branches regulate Adc17 proteasome assembly chaperone expression
.
J. Cell Sci.
137
:
jcs261892
.
Wu
,
H.
,
P.
Carvalho
, and
G.K.
Voeltz
.
2018
.
Here, there, and everywhere: The importance of ER membrane contact sites
.
Science
.
361
:eaan5835.
Yoshikawa
,
K.
,
T.
Tanaka
,
C.
Furusawa
,
K.
Nagahisa
,
T.
Hirasawa
, and
H.
Shimizu
.
2009
.
Comprehensive phenotypic analysis for identification of genes affecting growth under ethanol stress in Saccharomyces cerevisiae
.
FEMS Yeast Res.
9
:
32
–
44
.
Zadoorian
,
A.
,
X.
Du
, and
H.
Yang
.
2023
.
Lipid droplet biogenesis and functions in health and disease
.
Nat. Rev. Endocrinol.
19
:
443
–
459
.

Author notes

Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. C. Klose reported other from Lipotype GmbH during the conduct of the study; and other from Lipotype outside the submitted work. No other disclosures were reported.

This article is available under a Creative Commons License (Attribution 4.0 International, as described at https://creativecommons.org/licenses/by/4.0/).

or Create an Account

Close subscription notice
Close access options