Maltose accumulation-induced cell death in Saccharomyces cerevisiae

Abstract

Pretreatment of lignocellulose yields a complex sugar mixture that potentially can be converted into bioethanol and other chemicals by engineered yeast. One approach to overcome competition between sugars for uptake and metabolism is the use of a consortium of specialist strains capable of efficient conversion of single sugars. Here, we show that maltose inhibits cell growth of a xylose-fermenting specialist strain IMX730.1 that is unable to utilize glucose because of the deletion of all hexokinase genes. The growth inhibition cannot be attributed to a competition between maltose and xylose for uptake. The inhibition is enhanced in a strain lacking maltase enzymes (dMalX2) and completely eliminated when all maltose transporters are deleted. High-level accumulation of maltose in the dMalX2 strain is accompanied by a hypotonic-like transcriptional response, while cells are rescued from maltose-induced cell death by the inclusion of an extracellular osmolyte such as sorbitol. These data suggest that maltose-induced cell death is due to high levels of maltose uptake causing hypotonic-like stress conditions and can be prevented through engineering of the maltose transporters. Transporter engineering should be included in the development of stable microbial consortia for the efficient conversion of lignocellulosic feedstocks.

A hypotonic-like response induced by intracellular maltose accumulation results in cell death in a xylose-fermenting specialist yeast strain.

Introduction

The generation of renewable and sustainable energy is regarded as a viable approach and a potential remedy to circumvent dependence on fossil fuel sources. This strategy holds promise for addressing both climate change and the energy crisis. Lignocellulosic biomass as a major feedstock exists widely on the Earth, including energy crops grown on low-quality soil, and organic waste from industrial, agriculture, and forestry, as well as waste from municipal and green areas (parks and gardens) (Robak and Balcerek 2018). Recently also the development of bioethanol production from starch feedstocks or raw materials (Busic et al. 2018) have gained attention (Bai et al. 2008). Next to fermentable sugars, some other oligosaccharides can be obtained to varying levels due to differences in sources of biomass and pretreatment processes, including the production of maltose (Coulier et al. 2013). In yeast cells, intracellular maltose is degraded into two glucoses molecules by maltase where excess intracellular glucose, when not further metabolized, can be exported via Hxt transporters (Van Leeuwen et al. 1992, Jansen et al. 2002). Maltose metabolism requires the MAL (MAL1-4 and MAL6) loci with MALx1 encoding a maltose permease, MALx2 encoding the maltase enzyme, and MALx3 encoding a maltose transcriptional activator. Other genes encode amylomaltase enzymes that are not important for growth on maltose per se. The five highly homologous MAL loci map to telomere-linked sites are located on five different chromosomes and most yeast strains have more than one MAL locus (Naumov et al. 1994). For example, CEN.PK113-7D has MAL1, MAL2, and MAL3 presented on chromosomes VII, III, and II, respectively (Nijkamp et al. 2012). Additionally, the α-glucoside permeases Mph2 and Mph3 also transport maltose and are subjected to regulation by maltose transcriptional activator Malx3 (Day et al. 2002).

After pretreatment of lignocellulose containing wastes, the availability of fermentable sugar mixtures such as pentoses (i.e. xylose) and hexoses (i.e. glucose), has prompted the development of microbial pathways for efficient xylose or glucose–xylose (co)-consumption. A main host in such applications is the yeast Saccharomyces cerevisiae that is an efficient hexose converter but naturally unable to utilize pentose sugars. This has been solved through the introduction of XYL1 and XYL2 from Pichia stipites encoding xylose reductase (XR) and xylitol dehydrogenase, respectively, which allows S. cerevisiae to successfully utilize xylose as sole carbon source (Kotter and Ciriacy 1993). However, xylose fermentation is limited using this pathway due to the cofactor imbalance and an insufficient capacity of nonoxidative pentose phosphate pathway (Kotter and Ciriacy 1993, Karhumaa et al. 2005, 2007, Lee et al. 2012). Alternatively, xylose isomerase (XI) has been introduced in S. cerevisiae. This enzyme mediates the isomerization of xylose to xylulose, which avoids excessive accumulation of xylitol and prevents the cofactor imbalance associated with the XR pathway (Kuyper et al. 2003, Madhavan et al. 2009). A further engineering target is transport of xylose into S. cerevisiae cells. Saccharomyces cerevisiae lacks a specific xylose transporter, and xylose is typically transported with poor affinity and efficiency by members of the Hxt glucose transporter family. Hence, cells first utilize glucose, and only when the glucose is depleted, xylose utilization commences. Since xylose metabolism is highly susceptible to inhibitors that accumulate in the medium during fermentation, a more robust fermentation requires the cometabolism of glucose and xylose. Thus, heterologous xylose transporters have been expressed in S. cerevisiae, but only low xylose consumption rates were accomplished due to low activity and protein instability (Young et al. 2011, Nijland and Driessen 2019). Hence, direct engineering of endogenous hexose transporters was considered as an alternative option as it ensures proper expression and integration in the cellular regulatory networks. Through directed evolution and saturation mutagenesis, Hxt transporters have been engineered with high specificity for xylose, and this also has led to yeast strains that support efficient glucose and xylose coconsumption (Hou et al. 2017, Nijland and Driessen 2019). In particular, mutation of a conserved asparagine [in the chimeric Hxt36, Hxt7, and Gal2 transporters (N367) and in the cryptic Hxt11 (N366) transporter] resulted in specific xylose transporters unable to transport glucose and in which xylose transport is not inhibited by glucose (Farwick et al. 2014, Nijland et al. 2014, Shin et al. 2015).

Instead of achieving cometabolism of various sugars by a single yeast strain, another strategy is to utilize specialist strains to allow cometabolism in a microbial consortium. Recently, a xylose-fermenting specialist strain IMX730 was constructed from S. cerevisiae CEN.PK113-7D (Verhoeven et al. 2018a). In this strain, the four hexose kinase genes were deleted to completely abolish glucose phosphorylation whereas high xylose assimilation rates could be achieved by introduction of the XI XylA and xylulokinase Xks1 and overexpression of the genes involved in nonoxidative pentose phosphate pathway. Hence, this strain is able to grow on xylose rather than glucose (Fig. 1) (Verhoeven et al. 2018b), although the presence of glucose delays growth likely because of competition between xylose and glucose for transport. An important aspect in using specialist strains is that they should sustain sugar fermentation, even in the presence of high concentrations of alternative sugars. The glucose inhibition of growth likely can be circumvented by the expression glucose-insensitive transporters. Nevertheless, specialist strains should also be insensitive to other sugars present in complex feedstocks.

Figure 1.

Schematic representation of maltose, glucose, and xylose transport and metabolism in IMX730. Arrows in bold represent the overexpression of genes involved in the corresponding reaction. means genes involved in the pathway are deleted. Different color blocks indicate different pathways (glycolysis, the oxidative PPP, and the nonoxidative PPP).

Here, we have examined the effect of maltose on xylose specialist strain S. cerevisiae IMX730.1 growing on xylose and demonstrate a remarkable phenomenon of growth inhibition that can be attributed to the high level accumulation of maltose that triggers a hypotonic-like response and cell death. The implications for the development of specialist strains for sugar utilization in consortia are discussed including a strategy to render such fermentations more robust.

Materials and methods

Strains and cultivations

Saccharomyces cerevisiae IMX730 derived from S. cerevisiae CEN.PK113-7D, was obtained from Professor Dr J.T. Pronk of the Delft University of Technology (Verhoeven et al. 2018a) in which previously (Nijland et al. 2023) the His3 was deleted to facilitate histidine selection. All strains and plasmids used in this work are listed in Table 1 and Table S1 (Supporting Information), respectively. The CRISPR/Cas9 technique was used to perform gene editing (Mans et al. 2015). All oligonucleotides used in this work are listed in Table S2 (Supporting Information). All mutants were made using the His3-based pMEL16 plasmid carrying the gRNA target, which was transformed together with a double stranded repair fragment to delete targeted genes. Similarly, the IMX730.1 control strain was prepared using a modified pMEL16 plasmid without a yeast specific gRNA target. For multiple genes deletion, pMEL16 was removed followed by a next round of gene deletion yielding dMalX2, in which all maltase genes were deleted, and dMalX1, in which all maltose transporters were deleted. The GLK1 overexpression, using plasmid pRS313-P7T7-GLK1 was performed in IMX730.1 and dMalX2 yielding 730OEglk1 and dMalOEglk1, respectively. All yeast transformations were performed following the Elble’s protocol (Elble 1992). Shake flask cultures were grown in the mineral medium supplemented with 25 mM d-xylose (Luttik et al. 2000) at 30°C, 200 rpm. Precultures of cells in the exponential phase were used as inoculum into mineral medium supplemented with different sugars. The initial OD₆₀₀ was set at 0.1. Cell cultures were collected and measured by UV–visible spectrophotometer (Novaspec Plus) at 600 nm at indicated time points. All growth experiments were carried out in at least two biological replicates.

Table 1.

Saccharomyces cerevisiae strains.

Strain	Genotype	Source or reference
CEN.PK113-7D	MATa MAL2-8c SUC2	Nijkamp et al. (2012)
IMX730	MATα ura3-52 his3-1 leu2-3112 MAL2-8c SUC2 glk1::Sphis5, hxk1::KlLEU2 gal1::cas9- amdS gre3::pTDH3_RPE1 pPGK1_TKL1, pTEF1_TAL1 pPGI1_NQM1 pTPI1_RKI1 pPYK1_TKL2 can1::(pTPI_xylA_tCYC)∗9 pTEF1_XKS1 hxk2:: PcaraT)	Verhoeven et al. (2018a)
IMX730-his	IMX730; Sphis5∆	Nijland et al. (2023)
IMX730.1	IMX730-his; modified pMEL16	This study
dMalX2	IMX730-his; MAL12∆; MAL22∆; MAL32∆::his3	This study
dMalX1	IMX730-his; MAL11∆; MAL21∆; MAL31∆; MPH2∆; MPH3∆::his3	This study
CENdMalX2	CEN.PK113-7D; MAL12∆; MAL22∆; MAL32∆::his3	This study
730OEglk1	IMX730.1; pRS313-P7T7-GLK1	This study
dMalOEglk1	dMalX2; pRS313-P7T7-GLK1	This study

d-[¹⁴C] xylose and [¹⁴C] maltose uptake assays

The uptake of d-[¹⁴C] xylose or [¹⁴C] maltose (CAMPRO Scientific GmbH, Veenendaal, the Netherlands) was performed as described earlier with minor modifications (Shin et al. 2017). Cells were collected by centrifugation (3000 rpm, 3 min, 20°C), washed and resuspended into mineral medium. [¹⁴C] xylose stocks were added to the cells supplemented with different concentrations of glucose or maltose. For [¹⁴C] maltose uptake, 1.1 mM [¹⁴C] maltose was used. The uptake reactions were stopped at various time intervals by addition of 5 ml of 0.1 M lithium chloride, and the suspension was filtered (0.45 mm HV membrane filter, Millipore, France). Filters were washed with another 5 ml of lithium chloride and counted with the emulsifier scintillator plus (Perkin-Elmer).

Transcriptome analysis

The strain in which all maltase genes were deleted (dMalX2) (Table 1) was pregrown in mineral medium with 25 mM xylose up to the exponential phase (OD ∼5). The preculture was diluted to OD of 1 in the mineral medium with 25 mM xylose for 4 h growing to OD of ∼2.5. The culture was split into two, and one of the cultures was supplemented with 1.1 mM maltose whereas the other received water as control. After 1.5 h, cells were collected and total RNA was isolated as described (Zhang et al. 2022). The experiment was performed in two biological replicates. The total RNA samples were sent to BGI (Copenhagen, Denmark) for RNA-seq analysis. The raw data with 20 M and paired-end 100 reads obtained from BGI was filtered including adaptor sequences removing contamination and low-quality reads removing. After the quality check with FastQC, transcriptome data was analyzed through the STAR-SAMtools-FeatureCounts pipeline. RNA counts of the various samples were compared by ‘DESeq2’ package in R programing (version 4.1.3). Before further analysis, genes with count ≤ 10 are removed. CEN.PK113-7D genome was used as reference genome and the statistical significance cutoff was set as 0.05 (P < .05).

Cell viability

Cell viability was determined by propidium iodide (PI) staining. Cells were collected and washed twice with phosphate-buffered saline (PBS) buffer. Next, cells were stained with 1 µg PI/ml for 5 min, at 4°C, in dark. Subsequently, the cells were washed and resuspended in PBS, and either analyzed by flow cytometry (BD Accuri^TM C6, Becton Dickinson) or dropped on the glass slide overlaid with an 1% agarose pad and then covered with coverslips in order to capture images using a Nikon Ti-E microscope (Nikon Instruments, Tokyo, Japan) equipped with Hamamatsu Orca Flash 4.0 camera.

Cell dry weight and volume

Cells were grown in mineral medium with 25 mM xylose for 16 h and collected by centrifugation. Cell resuspension was dropped on the glass slide overlaid with an 1% agarose pad and then covered with coverslips. Images were captured by Nikon Ti-E microscope (Nikon Instruments) with 100x objective. Cell volume calculation (Nijland et al. 2019) were performed by using ImageJ 1.48v with a plug-in BudJ v4.3, which draws ellipse fitting cell and subsequently returns the values of major axis (R) and minor axis (r). Thereby, equation (1) was used to calculate the cell volume. All cells from three slides growing under the same condition were chosen to do the calculation.

(1)

In order to determine the dry cell weight (DCW), predried syringe filters (Whatman Puradisc 30 mm/0.2 µm) were employed which were weighed before filtering 50 ml of cells with a measured OD₆₀₀ of ∼5. Subsequently filters, together with cells, were dried overnight at 60°C and weighed again.

Results

Inhibition of growth on xylose by maltose

Saccharomyces cerevisiae IMX730.1 is as a xylose-fermenting specialist that is unable to phosphorylate glucose due to the deletion of the genes encoding glucose kinase activity (Verhoeven et al. 2018b). Further, it utilizes xylose efficiently due to the introduction of a XI and the overexpression of the nonoxidative phosphate pathway. Maltose is a disaccharide that consists of two units of glucose linked with 1,4-α bond and is transported into the cell via maltose/proton symporters and subsequently hydrolyzed into two glucose molecules by α-d-glucosidases, specifically the enzyme maltase (Van Leeuwen et al. 1992). Remarkably, growth of strain IMX370.1 on xylose (25 mM) was inhibited by the presence of maltose (Fig. 2A). Inhibition already occurred at a very low concentration of maltose (∼1 mM) and increases significantly at higher concentrations. Importantly, IMX730.1 is unable to grow on maltose due to the lack of the glucose kinases (Fig. 2A). Although growth was also inhibited by glucose, mostly causing a delay in growth, much higher concentrations of glucose were required as compared to maltose (Fig. S1, Supporting Information). The latter was noticed previously in other glucose/xylose transport studies as discussed in the introduction. In contrast, maltose transport into the cells occurs via specific maltose/proton symporters. These systems transport maltose against a concentration gradient at the expense of the proton motive force. To examine if the xylose growth inhibition by maltose is due to competition for transport, the uptake of xylose was studied in the presence of increasing concentrations of maltose or glucose, respectively. As expected, glucose efficiently competed with xylose for uptake. However, maltose had little effect on xylose uptake (Fig. 2B), which shows that the growth inhibition is not due to interference with xylose uptake.

Figure 2.

Growth inhibition by maltose. (A) Growth profiles of IMX730.1 on xylose supplemented with different concentrations of maltose. The number before letters indicates the concentration of corresponding sugar in mM unit. 25X: 25 mM xylose (▪); 11M: 11 mM maltose (•); 25X + 1.1M: 25 mM xylose + 1.1 mM maltose (◊); 25X + 2.2M: 25 mM xylose + 2.2 mM maltose (∆); and 25X + 11M: 25 mM xylose + 11 mM maltose (○). (B) Xylose uptake of IMX730.1 in the presence of glucose or maltose. Cells were loaded with 25 mM d-[¹⁴C] xylose supplemented with various concentrations of glucose (▪) or maltose (•). The uptake reaction was stopped after 1 min. Data in (A) and (B) are displayed as the average and standard deviation of two biological replicates.

Intracellular maltose causes the growth inhibition

To examine the mechanism of maltose inhibition on xylose growth, two different strains were constructed, i.e. a maltase deletion strain (dMalX2) and a maltose transporter deletion strain (dMalX1). Accordingly, MALx2 including MAL12, MAL22, and MAL32 were deleted from IMX730.1 to yield dMalX2, and in another strategy, the transporter genes MALx1 (including MAL11, MAL21, and MAL31) and MPH2/3 were deleted from IMX730.1 to yield dMalX1. Growth of IMX730.1 and the deletion mutants on xylose in the presence of maltose was analyzed (Fig. 3A). In the dMalX2 strain, lacking all maltase genes, extracellular glucose is undetectable (Fig. S2f, Supporting Information) whereas growth inhibition by maltose was strongly exacerbated. Now already at a concentration of maltose as low as 55 µM, complete growth inhibition is observed (Fig. S2b, Supporting Information). In contrast, strain dMalX1, which lacks all maltose transporter genes, is completely insensitive to maltose inhibition, and can withstand maltose concentrations of up to 22 mM (Fig. S2a, Supporting Information). This strain is completely defective in maltose uptake (Fig. 6A). These data suggest that the maltose inhibition of growth on xylose is due to the accumulation of intracellular maltose.

Figure 3.

Intracellular maltose induces growth inhibition: (A) growth profiles (OD₆₀₀) of IMX730.1 (▪), the maltase deletion strain (dMalX2; •), and the maltose transporter deletion strain (dMalX1; ▴) in 25 mM xylose (X) and 25 mM xylose with 1.1 mM maltose (XM). (B) Growth profiles of IMX730.1 (squares) and dMalX2 (triangles) after with (open symbols) and without (closed symbols) the addition of 1.1 mM maltose at T0 to exponentially growing cell cultures in 25 mM xylose. Data are displayed as the average and standard deviation of two biological replicates.

Figure 6.

A hypotonic-like response induced by intracellular maltose accumulation. (A) Radioactive uptake assay measuring the intracellular maltose accumulation in IMX730.1 (▪), dMalX2 (), and dMalX1 () using 1.1 mM ([¹⁴C]) maltose and 25 mM xylose. The DCW of IMX730.1, dMalX1, and dMalX2 are 189.97 ± 4.33, 203.09 ± 12.62, and 215.75 ± 4.30 mg/1000 OD unit, respectively. (B) Top 10 biological processes of upregulated differential expression genes (DEGs) analyzed by gene ontology based on the corrected P-value. GO analysis was performed using a ‘GO Term Finder’ tool from SGD database (http://www.yeastgenome.org/). Genes with adjusted P-value ≤ .05 and fold change ≥ 2 or ≤ 0.5 were considered as the DEGs. (C) Fluorescence microscope images of PI-stained cells 5 h after growing on the indicated conditions. X: 25 mM xylose, XM: 25 mM xylose + 1.1 mM maltose, XM0.1S: 25 mM xylose + 1.1 mM maltose + 0.1 M sorbitol, XM0.2S: 25 mM xylose + 1.1 mM maltose + 0.2 M sorbitol, and XM0.3S: 25 mM xylose + 1.1 mM maltose + 0.3 M sorbitol. Scale bar = 10 µm. (D) Subset of DEGs discussed in the results section. Data in (A), (B), and (D) are displayed as the average and standard deviation of two biological replicates.

To further define how fast cells respond to maltose inhibition, growth profiles on 25 mM xylose were compared between IMX730.1 and the dMalX2 deletion strain in medium pulsed with maltose (Fig. 3B). After maltose addition, growth (OD₆₀₀) of the dMalX2 strain continues for about 1.5 h before it is arrested. In contrast, a 1.1-mM maltose pulse only slightly reduced the growth of strain IMX730.1 once growth was initiated. This reduced inhibition as compared to the data in Fig. 3(A) could be attributed to the higher cell numbers before maltose was added and because these cells convert maltose into glucose which is subsequently exported via Hxt transporters. Overall, these data suggest that growth inhibition by maltose is a slow process involving several generations of cells.

To determine if the maltose growth inhibition results in cell lysis and/or loss of viability, flow cytometer and fluorescence microscopy were utilized to perform live–dead assays with PI in which also the (average) cell size was determined. Cells were collected 4 h after a pulse of maltose, stained with PI and analyzed by flow cytometer (Fig. 4B). The fluorescence intensity of the dMalX2 strain notably increases in the presence of maltose, while cell count decreases significantly (Fig. 4A). The cell distribution as obtained from forward scatter area (FSC-A) and side scatter area (SSC-A) is altered by the maltose addition (Fig. 4B and C). In the dMalX2 strain, the scattering increased after maltose addition, suggesting cell aggregates or an increase in cell volume. Cell aggregation was confirmed by fluorescence microscope, concomitantly with cell death after the addition of maltose (Fig. 4D). Although the median of cell volume is not changed by the maltose addition (Fig. 4E), there is a minor, although not significant, increase in cell volume from 33.52 ± 15.03 µm³ to 36.93 ± 16.15 µm³ in dMalX2 strain whereas it remained the same (from 31.10 ± 14.50 to 31.15 ± 12.91 µm³) in the IMX730.1 control strain. These data indicate that maltose addition to the dMalX2 strain results in only a minor increase in cell volume but causes cell death (Fig. 4D).

Figure 4.

Maltose induces cell death and cell aggregation. (A) Fluorescence intensity (bars) and cell count (dots) after a pulsed addition of maltose. Cells were collected and stained with PI dye 4 h after a pulsed addition of maltose. Cells were grown on solely 25 mM xylose (X) and 25 mM xylose with a pulsed addition of 1.1 mM maltose (XM). Data are displayed as average and standard deviation of two biological replicates. (B) Distribution plot of side scatter (SSC-A) and forward scatter (FSC-A) of IMX730.1 in X (red dots) and XM (black dots). Samples were collected and analyzed by flow cytometer 4 h after a pulsed addition of maltose. (C) Distribution plot of SSC-A and FSC-A of dMalX2 in X (red dots) and XM (black dots). (D) Fluorescence microscope images of PI-stained cells 5 h after a pulsed addition of maltose (1.1 mM). Scale bar = 10 µm. (E) Cell volume of IMX730.1 and dMalX2 in X and XM. The cell volume was calculated 5 h after a pulsed addition of maltose in which the red dots represent the average cell volume.

Maltose inhibition is a general phenomenon

To determine if the maltose inhibition is specific for growth on xylose or a more general phenomenon, the effect of maltose on growth the IMX730.1 strain on other carbon sources was assessed. The IMX730.1 strain does not show significant growth on acetic acid or glycerol as a sole carbon source. However, using this strain, ethanol can be used as a sole carbon source. Cells grown on ethanol showed a similar inhibition of growth by maltose, and this phenomenon was exacerbated in the dMalX2 strain (Fig. 5A). To restore the ability of glucose utilization, the GLK1 gene encoding glucokinase was expressed in strain IMX730.1 (730OEglk1) and dMalX2 (dMalOEglk1). GLK1 expression restored growth on glucose, but still in the dMalX2-OEglk1 strain, growth was inhibited by maltose (Fig. 5B). This inhibition was not observed in the IMX730-OEglk1 strain, as this strain can now grow on maltose because of the presence of both GLK1 and MALx2. Note that the maltose inhibition of dMalX2-OEglk1 on glucose is less severe as compared to growth on xylose or ethanol as sole carbon source (Figs 3A and 5A, respectively). The parental S. cerevisiae CEN.PK113-7D strain was insensitive to maltose inhibition, but upon the deletion of all maltase genes (CENdMalX2), a similar maltose inhibition, although at different maltose concentrations due to the plate-assay set-up, was observed when cells were grown on glucose (Fig. 5C). Summarizing, these data suggest that inhibition of growth by maltose is ubiquitous, and not specific for growth on xylose only.

Figure 5.

The ubiquity of maltose inhibition. (A) Growth profiles (OD₆₀₀) of IMX730.1 and dMalX2 on 25 mM ethanol (E) or 25 mM ethanol supplemented with 1.1 mM maltose (EM). (B) Growth profiles of strain harboring GLK1 in IMX730.1 (730OEglk1) and dMalX2 (dMalOEglk1) on 25 mM glucose (G) and 25 mM glucose addition with 1.1 mM maltose (GM). Data are displayed as the average and standard deviation of two biological replicates. (C) Spot assays of CEN.PK and CENdMalX2 on MM plates containing 21 mM glucose supplemented with different concentrations of maltose. Serial dilutions of cell culture ranging from 10⁰ to 10⁻³ were spotted on the plates with the indicated combination of glucose or xylose and maltose.

A hypotonic-like response induced by intracellular maltose accumulation

The dMalX2 strain lacks all maltase enzymes and hence these cells are expected to accumulate high concentrations of maltose without the ability to metabolize this molecule. Therefore, the level of intracellular maltose accumulation was determined employing ¹⁴C-labeled maltose. We observed significant intracellular maltose accumulation (∼ 45 nmol/mg DCW) in the dMalX2 strain whereas no accumulation is observed in the maltose transport deficient dMalX1 strain (Fig. 6A). In the IMX730.1 strain, able to convert maltose into glucose, some intracellular radioactivity accumulated within 10 min whereupon the levels decreased. The latter is most likely due to the intracellular maltose conversion into glucose, and subsequent release of the glucose by the cells via Hxt transporters (Jansen et al. 2002). In order to calculate the intracellular maltose concentration accurately, we used a calculated average of cell volume before and after the addition of maltose. As discussed in the previous section, the cell volume slightly increased after the addition of 1.1 mM maltose (Fig. 4E). Taking the increase into account, the intracellular maltose concentration goes up to ∼25 mM, which could potentially lead to a hypotonic-like stress conditions. Notably, in contrast to a typical hypotonic stress caused by the low extracellular osmolyte concentration, high levels of maltose accumulation will lead to an elevated intracellullar osmolarity.

To determine if the process of maltose accumulation indeed induces a hypotonic-like stress condition, or any other transcriptional response, transcriptome analysis was performed with strain dMalX2 1.5 h after the addition of 1.1 mM maltose. Transcriptome data is displayed in a volcano plot (Fig. S3a, Supporting Information) where in total 1653 genes are of statistical significance (P-value ≤ .05) and 487 genes (fold-change ≥ 2 or ≤ 0.5) are considered as differential expression genes (DEGs) as compared to the growth condition without added maltose. All DEGs are listed in Supplementary file 2. Gene ontology analysis of upregulated DEGs showed that six out of top 10 biological processes are related to the biogenesis and organization, and metabolic process of cell wall (Fig. 6B). A total of 186 downregulated DEGs are enriched into Gene Ontology Term ‘cellular porcess’ (GO:0009987) where genes were further analyzed using a ‘GO Slim Mapper’ tool from SGD database (Fig. S3b and Table S3, Supporting Information). These downregulated genes are involved in mitochondrial translation, mitotic cell cycle, transcription by RNA polymerase Ⅱ, and regulation of cell cycle. Notably, the downstream gene transcription levels (fold-change ± SD) of the protein kinase C1 (PKC1) cascade pathway, which activation is a response to hypotonic stress, are altered by the maltose accumulation. These downstream genes constitute genes involved in spore wall assembly and cell wall biosynthesis FKS1 (2.37 ± 0.26), FKS2 (also known as GSC2, 3.66 ± 0.50), and FKS3 (2.81 ± 0.44), genes involved in cell wall stress CRH1 (2.04 ± 0.16) and CHS3 (2.33 ± 0.23) (Sanz et al. 2017), genes involved in cell cycle progression SMI1 (also known as KNR4, 0.41 ± 0.08) (Martin-Yken et al. 2002), genes involved in cell cycle and polarized growth CLN1 (0.48 ± 0.06) and CLN2 (0.36 ± 0.06) (Heinisch et al. 1999). Besides, transcription level of NHP6B is upregulated to 2.21 ± 0.37, which is involved in morphological and cytoskeletal defects (Costigan et al. 1994). Target KDX1 (also known as MLP1) of downstream transcription factor Rlm1 is also found to be upregulated to 4.86 ± 0.54 (Watanabe et al. 1997). Additionally, gene expression changes related to the cytosolic calcium levels, which is a typical response to hypotonic shock are observed. Herein, RCH1 (2.17 ± 0.28) encoding the negative regulator of cytosolic calcium homeostasis, FLC2 (2.83 ± 0.25) encoding the putative calcium channel, and CCH1 (2.71 ± 0.41) encoding the voltage-gated high-affinity calcium channel are upregulated. Meanwhile, we also found the upregulation of REE1 (7.40 ± 0.89), YJR124C (3.07 ± 0.43), and FAR1 (2.53 ± 0.42), which were reported to be induced by calcium shortage (Lombardia et al. 2002). Surprisingly, the downstream genes of the High Osmolarity Glycerol (HOG) signaling pathway are also noted in the transcriptome data. For instance, transcription levels of transcription factor SMP1 and YAP6 are upregulated to 4.43 ± 0.63 and 2.42 ± 0.50, respectively, and targets of transcription factor Sko1, Hot1, and Msn2/4 are changed. For instance, Sko1 targets FAA1 (2.03 ± 0.20), CWP1 (3.92 ± 0.35), and SED1 (3.14 ± 0.49), which are all upregulated while targets PUT4 (0.49 ± 0.08), SOK2 (0.41 ± 0.14), and YAP1801 (0.44 ± 0.07) referred as the paralog of YAP1802 are all downregulated (Proft et al. 2005). Targets of Hot1 NQM1 is upregulated to 2.53 ± 0.23 but GPD1 is downregulated to 0.45 ± 0.04 (Gomar-Alba et al. 2015), while transcription level of Msn2/4 target HSP12 increases to 2.30 ± 0.17 (Martinez-Pastor et al. 1996). Overall, the transcriptome data suggests that cell wall stress induced by maltose accumulation triggers the activation of PKC pathway. Further, this hypotonic-like response might be stimulated by a high intracellular osmolarity as transcriptional levels of genes involved in the HOG pathway sensing ambient hypertonic stress are also altered in our transcriptome date. We hypothesize that cell death is the consequence of this hypotonic-like response.

Osmotic stabilization rescues cells from maltose induced cell death

If maltose induced cell death is indeed related to a cellular osmotic disbalance due to the high level accumulation of maltose, we hypothesize that osmotic stabilization would rescue cells from this fate. For that purpose, maltose induced cell death of xylose grown cells was followed with the live–dead staining and fluorescence microscopy in the absence and presence of various concentrations of sorbitol (Fig. 6C). Sorbitol is not metabolized by yeast cell, and hence its external addition can be used to bring the osmolarity of the medium in balance with the intracellular osmolarity. In the absence of sorbitol, maltose caused the death of a major fraction in the dMalX2 strain. However, in the presence of sorbitol, cell viability is remarkably improved. When the maltose concentration was increased from 1.1 to 2.75 mM, also a higher sorbitol concentration was, experimentally determined, required to protect the cells against maltose induced cell dead, i.e. 0.1 M versus 0.83 M, respectively (Fig. S4a, Supporting Information). Although the addition of sorbitol decreased the maltose induced cell dead (Fig. S4c, Supporting Information), sorbitol also reduced growth in the IMX730.1 strain, at least within the 5 h timeframe (Fig. S4b, Supporting Information). This phenomenon might be due to altered expression of many genes involved in osmotic stress signaling and osmoadaptation (Hohmann 2002). The required sorbitol/maltose ratio at 2.75 mM maltose was significantly higher as compared to 1.1 mM, most likely due to increased maltose uptake rates at higher maltose concentrations. To ensure that the effect of sorbitol is related to osmolarity and not due to competition for sugar transport, the effect of sorbitol on the uptake of ¹⁴C-xylose and ¹⁴C-maltose was examined (Fig. S4c, Supporting Information). Compared to xylose uptake of IMX730.1 in the presence of 1.1 mM maltose (56.4 ± 1.9 nmol/mg DCW), the addition of 0.3 M sorbitol slightly increased xylose uptake to 59.7 ± 0.2 nmol/mg DCW. Likewise, the presence of sorbitol resulted in a slight elevation of xylose uptake in dMalX2 from 41.7 ± 1.4 to 46.5 ± 0.7 nmol/mg DCW. The minor increase in uptake can be attributed to the improved cell viability when cells are exposed to maltose in the presence of sorbitol. Importantly, 0.3 M sorbitol addition does not prevent maltose uptake by the dMalX2 and IMX730.1. Our data, therefore indicate that sorbitol at high concentration does not compete with xylose or maltose transport. Hence, the data demonstrate that sorbitol protects cells against maltose induced death likely through osmotic balancing.

Discussion

To produce bioethanol in a second-generation fermentation, the yeast S. cerevisiae needs to convert various sugars to yield an economically feasible process. While much research is focused on cometabolism of xylose and glucose, plant waste streams may also contain other sugars at various ratios (Guo et al. 2022). Here, we reported on a remarkable phenomenon where maltose inhibits the growth of xylose-specialist strain of S. cerevisiae. Maltose toxicity, or ‘substrate-accelerated death,’ has been observed before (Postma et al. 1990) and the phenomenon was utilized by Tamayo Rojas et al. (2022) to develop a system to analyze glucose efflux. Postma et al. (1990) refer to an osmotic burst however only show maltose uptake experiments and no additional evidence. Here, we show that the maltose-induced growth inhibition is not specific for growth on xylose alone but can to different degrees also be observed when cells are growth on glucose or ethanol as the sole carbon source. Additionally, we have shown that the maltose growth inhibition is strongly stimulated when all maltase genes are deleted, and is completely prevented when all maltose transporters are deleted. This demonstrates that inhibition is due to intracellular accumulation of maltose. Compared to the IMX730.1 strain, maltose inhibition in the parental strain CEN.PK2-1D is rather limited (Fig. 5C), which may be attributed to the glucose repression. Maltose metabolism is rapidly repressed by glucose in various ways. In the presence of glucose, the transcriptional repressor Mig1 binds to the intergenic region of the MAL loci, thereby downregulating the expression of MALx1 and MALx2 (Yao et al. 1994). Furthermore, the maltose permeases (Malx1) are ubiquitinated after phosphorylation and subsequently delivered to the vacuole and degraded (Lucero and Lagunas 1997, Medintz et al. 2000, Dupre et al. 2004). Interestingly, the activity of maltase enzymes keeps constant during glucose repression (M. 1969). Therefore, maltose permease degradation in CEN.PK2-1D likely renders cells more insensitive to maltose inhibition in an analogous manner as the deletion of all maltose transporter genes. Since maltases are absent in the CEN.PK113-7D derivative strain CENdMalX2, likely the intracellular maltose can accumulate to such an extent that there is inhibition of cell growth. Maltose uptake is accompanied by the uptake of 1 proton per maltose molecule and the accumulation of a proton would acidify the intracellular space if not counteracted by proton export via Pma1 at the expense of 1 ATP (Serrano et al. 1986). The uptake of maltose could, therefore drain the ATP levels causing the observed substrate-accelerated death. However, ATP levels were (not significantly) reduced in dMalX2 in 30 min after the pulsed addition of maltose and resumed back to normal level after 90 min (data not shown) suggesting that ATP depletion is not the main mechanism of cell death.

In the maltose inhibited cells, cell aggregation or flocculation is observed. This likely relates to increased cell–cell interaction between flocculins generated by FLO genes expression and cell wall polysaccharides, primarily mannans and glucans. Yeast flocculation is a crucial feature to respond to environmental stress, induced by many factors, changes in cell wall integrity, expression of FLO family, temperature, and others (Westman et al. 2014, Sariki et al. 2019). It has been reported that maltose can lead to flocculation related to an increased glucan content of the cell wall (Nayyar et al. 2017). Consistent with our current work, genes involved in glucan biosynthesis are upregulated in maltose exposed cells, such as, GSC2, BGL2, EXG1, and CRH1, but genes in FLO family, except YHR213W, show no significant differences in the transcriptional response. Recently, activation of Pkc1-Slt2-Rlm1 MAPK pathway has been reported to induce flocculation via upregulation of FLO genes (Sariki et al. 2019). This Pck1-Slk2-Rlm1 pathway is activated in our transcriptome data set, although no upregulation of the FLO genes was observed. The latter may be due to the collecting time of RNA samples that occurred early after the addition of maltose, i.e. 1.5 h whereas cell aggregate starts to occur after 3 h.

With the maltase deficient dMalX2 strain, maltose is accumulated at high levels (Fig. 6A) potentially causing an osmolarity imbalance between extra- and intracellular environments. Osmoregulation is a ubiquitous mechanism to tackle the dynamic environment. Osmotic upshift (hypertonic shock) or downshift (hypotonic shock) could trigger two independent mitogen-activated protein (MAP) kinase signal transduction cascades with a centralized principle that production and accumulation of the solute glycerol is used to maintain a balance between extra- and intracellular osmolarity. The HOG signaling pathway responding to the hypertonic shock consists of two branches, the Sln1 branch and Sho1 branch. Sln1 and Sho1 sense hypertonic shock and transduce the signal to Hog1 which is subsequently phosphorylated by Pbs2. The translocation of Hog1 to nucleus subsequently activates the transcription factor Hot1, Sko1, Msn2/4, and Smp1 of which Hot1 upregulates NAD-dependent glycerol-3-phosphate dehydrogenase GPD1 expression to promote glycerol production, concomitant with a reduced glycerol export rate (Tamas et al. 1999, Hohmann 2002). Upon hypotonic shock, the PKC1 cascade is activated to regulate the expression of genes involved in cell wall metabolism and actin cytoskeleton reorganization (Davenport et al. 1995, Hohmann 2002, Gualtieri et al. 2004). Wsc1 and Mid2, as sensors for cell wall integrity signaling, activate Pkc1 via Rho1, thereby transducing signal to transcription factor Rlm1 and Swi4 through Bck1-Mkk1/2-Mpk1 MAP cascade (Hohmann 2002, Levin 2005). Simultaneously, autophosphorylated Sln1 sensor upon hypotonic shock could accumulate phospho-Skn7, which stimulates the expression of genes involved in cell wall biogenesis, such as OCH1 encoding a mannosyltransferase (Levin 2005). Furthermore, hypotonic shock can induce a transient rise of cytosolic calcium level via an influx of extracellular calcium through the Cch1 or Mid1 channels or by release of calcium from intracellular compartments (Batiza et al. 1996, Levin 2005, Rigamonti et al. 2015).

In the present work, addition of maltose to the dMalX2 strain caused an altered genes expression related to both hypo- and hypertonic responses. In the data set, the upregulation of FLC2, that encodes a putative calcium channel, was noted, This channel is involved in the hypotonic shock-induced calcium response (Rigamonti et al. 2015) and cell wall damage response (Garcia et al. 2004). Further, upregulation occurred of FAR1, which participating in cell growth and division and of SED1 that participates in cell wall organization and this normally associated with a cytosolic calcium shortage (Lombardia et al. 2002). Interestingly, GPD1, the key targets of HOG pathway activation upon hyperosmolarity, is downregulated. Previously, the GPD1 transcription level was reported to decrease 10-fold with 30 min after a hypotonic shock (Rep et al. 1999). Further, the PKC1 signaling and Sho1-Pbs2-Hog1 pathways evidently activated in our transcriptome data. However, the expression levels of genes involved in the activation of Sln1–Skn7 branch were not altered. Our data suggest that there is a change in osmoregulation in the dMalX2 strain, that is reminiscent to moderate hyperosmolarity caused by extracellular environment that induces the HOG signaling pathway and simultaneously tries to counteract the increase in intracellular osmolarity caused by maltose accumulation by triggering the PKC1 kinase pathway. Since the high cell viability is observed upon hyperosmotic stress (Fig. 6C; Fig. S4a, Supporting Information), a surged intracellular osmolarity is the real culprit of cell death in this situation. Moreover, SED1 and CRH1 expression levels, which are upregulated in the data set to 3.14 ± 0.49 and 2.04 ± 0.16, respectively, have been reported to be altered under both hypotonic and hypertonic shocks (Lombardia et al. 2002, Proft et al. 2005, Sanz et al. 2017, Udom et al. 2019). In addition to a response to hypotonic shock, CLN1/2 are also responsible for the delay at the G1 stage under hyperosmolarity (Belli et al. 2001), suggesting that osmotic up- or downshift can induce a similar phenotype, for example, cell wall damage or growth delay.

Maltose induces cell death as shown by the live–dead fluorescent assays and the decreased cell viability. Strikingly, cell viability and growth (Fig. S4b, Supporting Information) are significantly improved by the addition of sorbitol to the medium. In previous studies, 0.4 ∼ 1 M sorbitol was used to induce hyperosmolarity responses (Thelen et al. 2021, Shimasaki et al. 2022, Zhang et al. 2022). Here, lower levels of 0.1 to 0.3 M sorbitol sufficed to counteract the maltose induced cell death (Fig. 6C). These data suggest that cell death is indeed caused by an imbalance between intra- and extraosmolarity induced by the high level of maltose accumulation. However, sorbitol can be transported inside cells via the Hxt transporters (Hxt13, Hxt15, Hxt16, and Hxt17) (Jordan et al. 2016), which potentially might cause competition for uptake of xylose. However, the IMX730.1 and dMalX2 strains are unable to grow on sorbitol while xylose uptake is not affected by extracellular sorbitol. Theoretically, sorbitol could even be added to a consortia of specific sugar consuming specialist strains, to balance osmolarity. However, sorbitol has a negative effect on growth at higher concentrations.

Taken together, we conclude that an increased intracellular osmolarity induced by maltose accumulation results in growth inhibition.

Our observations have important implications for the design of specialist strains for the fermentation of complex sugar mixtures using microbial consortia. With such strains, it is important that their growth on a given sugar is not affected by high levels of other sugars present in the pretreated plant waste materials. To isolate sugar metabolism in specialist strains, an effective strategy will be to prevent such cells from transporting other sugars, and to prevent cross-specificity of the existing sugar transporters for competing sugars. For instance, in our study, deletion of all maltose transporters renders cells completely insensitive to maltose inhibition. Ultimately, engineering of the transporter landscape of specialist strains will contribute to a more robust fermentation of complex sugar mixtures.

Conflict of interest

The authors declare no competing interests.

Funding

This research was supported by the Chinese Scholarship Council (CSC).