Transcriptome profiling of Issatchenkia orientalis under ethanol stress

Yeast strains, media, and growth conditions

Issatchenkia orientalis strain CBS 12547 was originally isolated from tropical fruit and food sources, and is involved in the fermentation of some traditional African foods (Greppi et al. 2013; Pedersen et al. 2012). The strain was maintained in the Food Biotechnology Laboratory at Ningbo University. Yeast was initially cultured for 24 h in YPD medium (1% yeast extract, 2% peptone, and 2% glucose) at 30 °C with agitation at 150 rpm. 1 mL was withdrawn, added to 100 mL fresh YPD medium, and incubated as before until the culture reached exponential phase (8 h). For the RNA-Seq experiment, ethanol was added to a final concentration of 10% (v/v), and incubation continued for another 4 h. Three cultures were treated in parallel with ethanol (TE1/TE2/TE3) and three untreated cultures were used as negative controls (T1/T2/T3). Yeast cells were harvested by centrifugation at 4 °C, 2000×g and stored at − 80 °C.

Scanning electron microscopy (SEM)

Issatchenkia orientalis was cultured in medium with 10% ethanol for 24 h. Cells were collected by centrifugation at 1000×g, 4 °C for 10 min and washed three times with physiological saline. Cells were then resuspended in 2.5% glutaraldehyde for 4 h at 4 °C and washed three times with 0.1 M PBS (pH = 7.4) for 15 min per wash. The cells were transferred through a series of ethanol solutions (30, 50, 70, 80, 90, 95 and 100%; 10 min each), and then through a series of tert-butanol-anhydrous ethanol mixtures (ratio 1:3, 1:1, 3:1, 3:0; 10 min each). Finally, the cells were dried and coated with a gold/palladium alloy (40:60) to a thickness of 10–20 nm and observed with a Hitachi S3400N scanning electron microscopy system.

Determination of trehalose concentration

Issatchenkia orientalis was cultured in medium with 10% ethanol for 0, 4, 12, 24, and 48 h. Cells were collected by centrifugation and washed with ultrapure water. Collected cells were frozen in liquid nitrogen and freeze-dried at − 20 °C. 100 mg of dried cells were resuspended in 1 mL ice-cold 0.5 mol/L trichloroacetic acid solution by brief treatment with ultrasound, and then maintained in the same solution at room temperature for 45 min in order to extract the trehalose from the cells. 250 µL extract was incubated with 1 mL 80% sulfuric acid solution containing 0.2% anthrone in a boiling water bath for 5 min. Absorbance at 620 nm was measured and compared with samples containing known concentrations of trehalose (Sigma-Aldrich) (Kitichantaropas et al. 2016; Mahmud et al. 2009).

Determination of ergosterol concentration

Issatchenkia orientalis was cultured in medium with 10% ethanol for 0, 4, 12, 24, and 48 h. Cells were collected by centrifugation, washed with ultrapure water, then frozen in liquid nitrogen and freeze-dried at − 20 °C. 100 mg of dried cells were resuspended in 3 mL ethanol containing 25% potassium hydroxide (m/v; 25 g KOH dissolved in 35 mL pure water, add ethanol to 100 mL) and incubated at 85 °C for 1 h. The entire sample was mixed with 3 mL n-heptane and extracted by vortexing for 3 min. Finally, absorbance of the supernatant at 282 nm was measured and compared with samples containing known concentrations of ergosterol (Sigma-Aldrich) (Arthington-Skaggs et al. 1999).

RNA extraction, library construction and sequencing

As noted earlier, I. orientalis was cultured in medium with 10% ethanol for 4 h before cells were harvested for RNA extraction. Total RNA from each sample was isolated using TRIZOL (Aidlab Biotech, Beijing, China). RNA concentration was quantified using a Qubit^® RNA Assay Kit and a Qubit^® 2.0 Fluorometer (Life Technologies, CA, USA). RNA integrity and purity were evaluated using the RNA Nano 6000 Assay Kit and the NanoPhotometer^® spectrophotometer (IMPLEN, CA, USA).

Construction of cDNA libraries and RNA sequencing were performed by Beijing BioMarker Technologies (Beijing, China). In brief, poly-A mRNA was isolated using poly-T oligomer bound to magnetic beads. mRNA was fragmented using divalent cations at elevated temperature. RNA fragments were copied as cDNA using random primers, and second strand cDNA synthesis was then performed. Double-stranded cDNAs were ligated to a single ‘A’ base and the sequencing adapters. Fragments (200 ± 25 bp) were then separated by agarose gel electrophoresis and selected for PCR amplification as sequencing templates. Finally, the library was constructed for sequencing on the Illumina HiSeq™ 4000 sequencing platform.

Quality control and read mapping

To obtain high-quality data, raw mRNA-Seq reads were processed using in-house Perl scripts. Reads were discarded if they were spoiled by adaptor contamination, contained ambiguous (N) base calls, or if more than 10% of bases had quality values < 30. The minimum acceptable length was 60 bp to avoid sequencing artifacts. All subsequent analyses were based on the filtered data set. Reads were mapped to the I. orientalis reference genome (NCBI Accession Number: GCA_000764455.1) using TopHat2 (http://ccb.jhu.edu/software/tophat/index.shtml). Gene names were assigned to sequences based on matches with the highest score.

Functional annotation

Issatchenkia orientalis genes were aligned to annotated sequences using BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi) and the following databases: Nr (NCBI non-redundant protein sequences, Nt (NCBI non-redundant nucleotide sequences, Pfam (Protein family), KOG/COG (Clusters of Orthologous Groups of proteins), Swiss-Prot (manually annotated and reviewed protein sequence database), protein data bank (PDB), KO (KEGG Ortholog database), and GO (Gene Ontology). The Blast2GO suite (Götz et al. 2008) was used to assign GO terms for molecular function, biological process, and cellular component.

Analysis of differential expression

Differential expression analysis of data from the two experimental conditions (ethanol stress vs. control) was performed using the DESeq R package (1.10.1). P-values were adjusted using Benjamini and Hochberg’s approach for controlling the false discovery rate (FDR). Differentially expressed genes (DEGs) were defined as those with fold change > 3 (P < 0.05) and FDR < 0.01.

KEGG and GO enrichment analyses for DEGs

KEGG (http://www.genome.jp/kegg/) is a database resource for understanding functions and utilities of the biological system from molecular-level information, especially large-scale molecular datasets generated by genome sequencing. KEGG is often used to tentatively assign functions and other properties to genes. We used KOBAS (Mao et al. 2005) to determine if any differentially expressed genes were significantly enriched in KEGG pathways. To determine which Gene Ontology (GO) categories were statistically overrepresented among the DEGs, topGO and Cytoscape version 3.4.0 with BiNGO plugin version 3.0.3 (Maere et al. 2005) were used to identify significantly enriched biological networks and to output the results as graphs.

Quantitative PCR for selected DEGs

Real time quantitative PCR (qPCR) primers for selected DEGs were designed using Primer 5.0 (Additional file 1: Table S1). A Tiangen FastQuant RT Kit (with gDNase) and a KAPA SYBR FAST Universal qPCR Kit were used for reverse transcription and qPCR, respectively. All qPCR reactions were performed using a QuantStudio™ 7 Flex Real-Time PCR system (Applied Biosystems, Thermo Fisher Scientific). The PCR reaction was conducted following the manufacturer’s instructions, and three biological replicates were used in all experiments. Negative controls (template consisting of ultrapure water) were run for each gene. Run-time control of the PCR instrument, baseline correction, and determination of Cq values were performed using QuantStudio™ 7 Flex Real-Time PCR Software v1.2 (Applied Biosystems, Thermo Fisher Scientific).

Intracellular trehalose and ergosterol concentration and SEM imaging

Intracellular concentrations of trehalose and ergosterol, measured after 4, 12, 24 and 48 h of ethanol stress, are shown in Fig. 1. Compared with unstressed controls, ethanol-stressed yeast cells contained higher levels of both compounds. Carbohydrates such as trehalose and glycogen are compatible solutes that resist osmotic pressure across the cytoplasmic membrane and prevent yeast cells from dehydration. Ergosterol is an important component of the yeast cytoplasmic membrane and is also thought to be involved in stress response.

SEM images were captured after ethanol stress for 24 h (Fig. 2). Stressed cells formed large flocs containing hundreds of connected cells.

Library construction and RNA-Sequencing

Libraries (NCBI Accession: PRJNA413795) were constructed for RNA-Seq from three control samples (T1-T3; NCBI Accessions: SRX3277329, SRX3277330 and SRX3277331), and three 10% ethanol-stressed samples (TE1-TE3; NCBI Accessions: SRX3277326, SRX3277327 and SRX3277328). After setting aside reads with adaptor contamination, ambiguous base calls, insufficient length, or unacceptable numbers of low quality base scores, 27.15 Gb of high-quality data were obtained with average quality values ≥ 30 for more than 85% of the reads. 76.01% of reads from the control libraries, and 77.17% of reads from the ethanol-stressed libraries, mapped to the I. orientalis genome, indicating successful library construction.

Identification of differentially expressed genes

Gene expression levels were calculated using FPKM values. As shown in Fig. 3, DEGs detected between control and ethanol stressed transcriptomes were required to meet criteria for fold change > 3 (P < 0.05) as well as FDR < 0.01. Of 502 transcripts with threefold or greater change, 451 were more abundant and 51 less abundant in ethanol-stressed cells. With four exceptions, all successfully matched with entries in the nr (498), Swiss-Prot (379), KEGG (205), or GO (224) databases, yielding a total of 498 unique and annotated DEGs.

Transcript levels for a subset of DEGs in several functional groups were determined by real time quantitative PCR. The results are shown in Fig. 4.

KEGG pathway analysis

DEGs were annotated using KEGG to identify orthologous genes, and KOBAS was used to test for statistically significant enrichment of DEGs in KEGG pathways. Q-values (Storey 2003) were generated by KOBAS, and are analogous to P-values in the context of our analysis. As shown in Fig. 5, DEGs were significantly enriched in pathways used for protein processing in endoplasmic reticulum (ko04141, q < 0.01) and meiosis (ko04113 q < 0.05). Pathways involving lipoic acid metabolism (ko00785) and steroid biosynthesis (ko00100) also had high enrichment scores, but with q values > 0.05.

GO annotation and analyses

Figure 6 shows molecular interaction graphs for the three GO ontology classifiers, generated using Cytoscape-BiNGO. In the biological process ontology (Fig. 6a), statistically overrepresented GO categories can be divided into five groups (process in response to stimuli, protein folding and refolding, sugar transport, DNA repair and flocculation). In the molecular function ontology (Fig. 6b), overrepresented GO categories can be divided into four clusters. The largest group consists of binding functions, specifically nucleotide binding, protein binding, and sugar binding. Other groups involved ATP hydrolase activities, ubiquitin protein ligase activities, and sugar transmembrane transport activities. In the cellular component ontology (Fig. 6c), the overrepresented GO categories included cell wall, ER, and plasmid membrane.

A protein–protein interaction (PPI) network was generated to identify key proteins involved in the response made by I. orientalis to ethanol stress (Additional file 2: Figure S1). Five proteins in this network (DSK2, HSP82, HSA1, BiP, and SMK1) are significantly and differentially expressed. These may play important roles in the stress response.

Ergosterol biosynthesis

KEGG enrichment analysis identified the steroid biosynthesis pathway (ko00100) as highly enriched (Fig. 5) including many DEGs associated with steroid biosynthesis (especially ergosterol biosynthesis). In S. cerevisiae, ergosterol protects cell membrane integrity and enhances membrane fluidity in response to stress (Chi and Arneborg 2000; Ren et al. 2014), but genes associated with ergosterol biosynthesis are transcriptionally down-regulated (Alexandre et al. 2001).

We found that ergosterol accumulates after ethanol stress (Fig. 1). Transcripts for the ergosterol biosynthesis genes ERG2, ERG3, and ERG27 are significantly more abundant in ethanol-stressed cells, in contrast to results reported for these genes in S. cerevisiae. ECM22, which encodes a sterol element-binding transcription factor that regulates sterol uptake and sterol biosynthesis (Woods and Höfken 2016), is also more abundant. ERG25 is an exception, and is less abundant under ethanol stress. The results confirm the role of ergosterol in I. orientalis as an important cytoplasmic membrane protectant in response to ethanol stress.

Trehalose metabolism

Analyses (Table 1, Fig. 4) show that genes involved in trehalose and glycogen metabolism are up-regulated during ethanol stress. The intracellular carbohydrates trehalose and glycogen are compatible solutes that resist osmotic pressure across the cytoplasmic membrane. Trehalose is involved in ethanol tolerance in S. cerevisiae (Mahmud et al. 2009; Wang et al. 2013; Yi et al. 2016). The up-regulation of trehalose and glycogen synthesis genes, and the accumulation of trehalose (Fig. 1), are consistent with this role. Stress tolerance in yeast may rely on trehalose-6p synthase (TPS1), the first enzyme in trehalose biosynthetic pathway, rather than on trehalose itself (Petitjean et al. 2015). In fact, we found that several genes in trehalose biosynthetic pathway, including TPS1, are up-regulated during ethanol stress. We conclude that the regulation of the trehalose pathway plays an important role in protecting cells against ethanol stress in I. orientalis.

Response to stress and stimulus

Genes involved in the response to biotic and abiotic stimulus, including heat and pH, were also enriched (Table 1, Fig. 6). Up-regulation of heat stress response genes, such as LRE1, WSC1, SGT2, and a variety of heat shock proteins, was observed in all samples in response to ethanol. In stress-tolerant S. cerevisiae strains, intracellular trehalose accumulates and heat shock protein genes are continuously induced in response to stresses that damage proteins, including heat, ethanol, osmotic, and oxidative stress (Kitichantaropas et al. 2016).

Expression of RIM101, a pH-response transcription factor, was up-regulated in response to ethanol. The homologous gene in S. cerevisiae regulates response and resistance to low pH and acidic conditions (Mira et al. 2009). In S. cerevisiae, high concentrations of ethanol affect the integrity of the cell membrane, changing proton permeability and causing intracellular acidification (Rosa and Sá-Correia 1996; Teixeira et al. 2009). Vacuolar acidification is a potential mechanism to recover cytosolic homeostasis after ethanol-induced intracellular acidification in S. cerevisiae (Martínez-Muñoz and Kane 2008). Similar mechanisms in I. orientalis may help I. orientalis maintain pH stability in the presence of ethanol.

HSP90, HSP70, and ubiquitin

Genes associated with protein folding and refolding (Fig. 6) are up-regulated under ethanol stress, such as HSP42, HSP78, and HSP104 (Fig. 4). PPI analysis suggests an important role for HSP82 (homolog of yeast HSP90) and HSA1 (HSP70 1) in protein folding and refolding (Additional file 2: Figure S1). Based on our RNA-Seq results, other genes encoding HSP binding proteins and co-chaperones such as STI1, AHA1, SSE1, MAS5, FES1, and SIS1 are also up-regulated.

In eukaryotes, HSP90 proteins are conserved, abundant molecular chaperones involved in many essential cellular processes (Li et al. 2012). Two cytosolic HSP90 isoforms exist in yeast: an inducible form HSP82, and a constitutive form HSC82. The association of HSP90 with HSP70 and a variety of co-chaperones generates large dynamic multi-chaperone complexes known as HSP90/HSP70 machinery. These play critical roles in the recruitment and assembly of client proteins, and also work in concert with the ubiquitin–proteasome system (UPS), directing misfolded proteins for degradation (Li et al. 2012). HSP42, HSP78, and HSP104, which were mentioned above, also help process aggregations of unfolded or misfolded proteins (Glover and Lindquist 1998).

Cytoscape-BiNGO analysis suggests that proteins with ubiquitin-protein ligase activity are up-regulated, including genes encoding ubiquitin-associated proteins (UBP16, BUL2, TOM1, HUL4, BRE1, and CUE2). The UPS degrades proteins that have exceeded their functional lifetime and destroys most unfolded and misfolded proteins (Amm et al. 2014). Proteins with ubiquitin-protein ligase activity, mainly E3 ligases, often work with HSP90/HSP70 chaperone systems and recognize misfolded proteins (Berndsen and Wolberger 2014; Petrucelli et al. 2004). The gene encoding ubiquitin domain-containing protein DSK2, which involved in the ubiquitin–proteasome proteolytic pathway and in spindle pole body duplication, was identified by PPI analysis as a key factor in the response to ethanol stress (Additional file 2: Figure S1).

The up-regulation of genes encoding HSP proteins and E3 ubiquitin ligases suggests that protein misfolding occurs under ethanol stress, possibly affecting proteins that help maintain plasma membrane integrity and function. Since the accumulation of improperly folded proteins is toxic, the HSP90/HSP70 based chaperone machinery and the ubiquitin–proteasome proteolytic pathway may be essential in the response to ethanol stress.

Starvation effect and transport

Genes associated with meiosis, reproduction, sporulation, ascospore cell wall assembly, and membrane biogenesis were up-regulated (Fig. 6, Table 1). For example, RRT12 encodes a spore wall-localized subtilisin-family protease required for spore wall assembly (Suda et al. 2009). GAS4 encodes a 1,3-beta-glucanosyltransferase that elongates 1,3-beta-glucan chains during spore wall assembly (Ragni et al. 2007). FLO1 encodes a cell wall protein that participates directly in adhesive cell–cell interactions during yeast flocculation (Fichtner et al. 2007). IFF6 encodes a GPI-anchored cell wall protein involved in cell wall organization and hyphal growth. Finally, CZF1 is a transcription factor involved in the regulation of filamentous growth in yeasts that responds to temperature and carbon source (Brown et al. 1999; Vinces et al. 2006). It is possible that CZF1 is involved in the flocculation of I. orientalis cells that we observed under ethanol stress (Fig. 2).

Genes with transporter activities were also up-regulated. These include genes involved in amino acid and peptide transport (transporter specific for methionine, cysteine and oligopeptide), carbohydrate transport (transporter specific for hexose such as mannose, fructose and glucose) and transmembrane transport. In addition, genes involved in protein transport, coenzyme transport, lipid transport, a-factor pheromone transport, and genes in the major transporter facilitator superfamily (MFS) were up-regulated.

Nitrogen starvation in S. cerevisiae induces meiosis, pseudohyphal growth, and sporulation. The presence of ethanol may affect the transmembrane transport of nutrients, leading to a pseudo-starvation state that elicits a nitrogen starvation response by the cell (Chandler et al. 2004; Kasavi et al. 2016; Stanley et al. 2010). Consistent with this hypothesis, up-regulation of meiosis, sporulation, and transportation-associated genes suggests that I. orientalis responds to ethanol stress as if it were experiencing nitrogen starvation. In effect, I. orientalis cells mistakenly perceive that they are growing in a nutrient-deficient environment, rather than in a nutrient-complete culture medium. The up-regulation of transmembrane transport genes is thus an attempt by the cell to cope with the pseudo-starvation state caused by ethanol stress.

The pseudo-starvation state may be due to the lack of coenzymes such as NAD + and coenzyme A (CoA). NAD + is an important cofactor for the glycolysis enzyme glyceraldehyde 3-phosphate dehydrogenase (GAPDH), while CoA is required for fatty acid metabolism and the oxidation of pyruvate in the citric acid cycle. We found that several genes encoding NAD(P) + -dependent enzymes were up-regulated, which implies that demand for NAD(P) + had increased. This is consistent with the transcriptional activation of Liz1 (Stolz et al. 2004), which encodes a plasma membrane-localized transport protein for the uptake of pantothenate, the precursor of coenzyme A (CoA). A lack of pantothenate would result in slow growth, delayed septation, and mitotic defects.

In conclusion, our data provide a global view of transcriptional changes in I. orientalis under ethanol stress. The changes are likely to reflect adaptation to stressful conditions at multiple levels. We observed modifications in the trehalose and ergosterol biosynthetic pathways, and also activation of various genes related to stress. Examples include heat shock proteins and their co-chaperones, which refold aggregated and misfolded proteins, and the ubiquitin–proteasome system, which targets misfolded proteins for degradation. Finally, ethanol stress appears to induce a nutrition starvation effect, which is associated with changes in cellular uptake, pseudohyphal growth, and sporulation. These results provide a basis for future investigations of the mechanisms that regulate ethanol stress in I. orientalis.