Introduction

Debaryomyces hansenii (anamorph: Candida famata) is a halotolerant yeast able to grow in the presence of high NaCl concentrations. This yeast can tolerate as much as 4 Μ of salt, while the growth of the budding yeast Saccharomyces cerevisiae is completely inhibited with 1.7 Μ NaCl (Onishi 1963; Prista et al. 2016). Originally isolated from seawater (Norkrans 1966), D. hansenii can also be found in salty and fermented foods with high osmolarity (Gori et al. 2012; Ramos et al. 2017). Importantly and due to its ability to inhibit fungal growth, this yeast species can be employed in the prevention of food contamination (Medina-Córdova et al. 2018).

The halotolerant behaviour of D. hansenii is the result of several properties acting in concert. These include its particular potassium and sodium homeostasis (Norkrans 1966, 1968), its membrane composition (sterol-to-phospholipid ratio) (Turk et al. 2007), the presence of glycerol and alkali-metal-cation transporters (Prista et al. 1997, 2016; Lages et al. 1999; Ramos et al. 2011; Martínez et al. 2011; Michán et al. 2013), the resistance of some enzymes to salt stress or expression induced by high salt concentrations (Alba-Lois et al. 2004; Sharma et al. 2005; Guerrero et al. 2005; Minhas et al. 2012; Chawla et al. 2017), and energy obtaining pathways that are optimised in the presence of salts (Sánchez et al. 2006, 2008; Calahorra et al. 2009; Cabrera-Orefice et al. 2014a, b) among others.

Highly-active glycerol uptake for the maintenance of an adequate intracellular concentration is one of the main processes that characterises halotolerant species (Lages et al. 1999). Indeed, in D. hansenii, as in many yeast species, glycerol has been found to be a preferential compatible solute (Gustafsson and Norkrans 1976; André et al. 1988; Larsson et al. 1990); however, D. hansenii is also able to efficiently produce trehalose or arabinitol as compatible solutes under high salt concentration (Adler and Gustafsson 1980; Blomberg and Adler 1992; González-Hernández et al. 2005). The intracellular level of glycerol depends on its synthesis, utilisation and transport, responding to various stress solutes besides NaCl, i.e. KCl, sucrose or Na2SO4 (André et al. 1988; Neves et al. 1997). The main metabolic pathways (Adler et al. 1985; André et al. 1988) and the transport systems (Lages et al. 1999; Neves et al. 2004; Klein et al. 2017) involved in glycerol metabolism in D. hansenii are similar to those present in S. cerevisiae, with the exception of the occurrence of a D. hansenii Na+/glycerol symporter that can also use potassium ions to internalise glycerol (Lucas et al. 1990).

Upon hyperosmotic stress, glycerol synthesis in S. cerevisiae as well as in other eukaryotic cells is mainly driven by the HOG1 (High Osmolarity Glycerol) pathway. The exposure to hyperosmotic stress triggers this MAPK pathway, which comprises two sub-pathways: the SHO1 and the SLN1 branches. Each one, by means of a cascade of progressive phosphorylations, activates the Hog1 protein, which internalises into the nucleus and triggers the transcription of genes needed to cope with the osmotic stress, especially those encoding enzymes for the synthesis and accumulation of glycerol (de Nadal et al. 2002; Saito and Posas 2012; Hohmann 2015).

DhGPD1, the gene encoding glycerol-3-phosphate dehydrogenase, which catalyses the first step in glycerol synthesis, was identified, isolated and heterologously expressed in a gpd1Δ strain of S. cerevisiae. It was found that hyperosmotic stress activated the transcription of DhGPD1 gene, restoring glycerol production in the Scgpd1Δ mutant, responding to the induction exerted by the native ScHog1 (Thomé and Trench 1999; Thomé 2004, 2005). In addition, the regulation of DhGPD1 and DhGPP2 genes was analysed in comparison to that of their orthologous in S. cerevisiae grown under different NaCl concentrations, and it was confirmed that these genes are upregulated upon hyperosmotic stress and that they play a significant role in the NaCl tolerance of D. hansenii (Gori et al. 2005).

Saline tolerance is also achieved by the activity of several ion transporters located at the plasma membrane of yeast cells, including uniporters such as Trk1, antiporters such as Nha1 and the Na+ -ATPase Ena1 (Ariño et al. 2019). It has been determined that osmotic stress induced by high sodium concentration, increases the transcriptional activity of the ENA1 gene in S. cerevisiae in a Hog1 dependent manner (Marquez and Serrano 1996). Transcription of the orthologues ENA1 and ENA2 genes of D. hansenii is also increased in high salt concentrations and at least for ENA2 with high pH (Almagro et al. 2001).

At the present time, the HOG1 pathway of D. hansenii has yet to be described. The only reports published include the cloning and sequencing of the DhHOG1 gene and its heterologous expression in a S. cerevisiae hog1Δ mutant (Bansal and Mondal 2000) and a study of the cellular localisation of DhHog1 after hyperosmotic treatment (Sharma et al. 2005). In the former report, the expressed D. hansenii gene conferred to the S. cerevisiae hog1Δ mutant the ability to grow under hyperosmotic stress and to induce glycerol overproduction. In the latter report, it was observed that phosphorylated DhHog1 had a delayed entry into the D. hansenii nucleus and was exported back in the cytoplasm after exposure to severe saline stress in a phosphorylated state. In this study, it was also shown that UV light and oxidative stress additionally induce DhHog1 phosphorylation.

D. hansenii belongs to the monophyletic clade that ambiguously translates CTG into Ser and Leu (Tekaia et al. 2000; Fitzpatrick et al. 2006; Papon et al. 2014). Besides the existence of some reports describing the DNA transformation of D. hansenii (Ricaurte and Govind 1999; Voronovsky et al. 2002; Minhas et al. 2009), gene modification using the conventional tools of molecular biology has been challenging in this yeast (Gerami-Nejad et al. 2009). It has been argued that this may be due to its robust cell wall (Gezelius and Norkrans 1970), or because it has a highly-efficient Non-Homologous End Joining (NHEJ) repair system, which does not allow deletion or modification of a gene of interest.

Given the important role that the HOG1 pathway plays in the halotolerance behavior, in the present study we report for the first time its inactivation by deletion of the DhHOG1 gene in D. hansenii. This was achieved using a suitable plasmid, specially constructed for genetic manipulation of the CTG clade yeast (Defosse et al. 2018). The phenotypical and physiological characteristics of this mutant (Dhhog1Δ) were investigated so that we might be able to understand the role of Hog1 in D. hansenii’s halotolerance behaviour as well as to other stresses. We found that DhHog1 was required for survival to high external osmolarity by regulating the expression of genes required for glycerol accumulation and it also has a role in the response to alkali, oxidative and endoplasmic reticulum stresses.

Materials and methods

Strains, plasmids and culture media

D. hansenii Y7426 was kindly donated by the US Department of Agriculture, Peoria, Illinois, USA. Dhhog1Δ (hog1Δ::SAT1-yeYFP1) is isogenic to strain Y7426 and its construction is depicted in Fig. 3a. S. cerevisiae BY4742 (MATα, his3-Δ1, leu2-Δ0, lys2-Δ0, ura3-Δ0) and its isogenic hog1Δ strain (MATα, his3-Δ1, leu2-Δ0, lys2-Δ0, ura3-Δ0, hog1Δ::kanMX) were obtained from the Yeast Knockout Collection.

Yeast cells were routinely grown at 28 °C in YPD (1% yeast extract, 2% peptone, 2% glucose). For mutant selection and maintenance, YPD was supplemented with 150 µg/ml nourseothricin (clonNAT) (Werner BioAgents GmbH, Germany). Modifications to the media and temperatures are indicated in figure legends of each experiment.

Escherichia coli DH5α strain was used to propagate plasmids. Bacteria were grown at 37 °C in LB (1% tryptone, 0.5% yeast extract, 0.5% NaCl) containing 100 µg/ml ampicillin for plasmid selection.

The pAYCU244 vector (PDhTEF1-SAT1-TMgPGK1-PMgACT1-yeYFP-TMgTRP1), especially adapted to D. hansenii (Defosse et al. 2018) was used for yeast transformation. The pGEM-T Easy vector (Promega ™) was used for cloning and sequencing of PCR products.

Growth assays

Growth curves were determined in YPD medium supplemented with the indicated NaCl concentrations using 250 ml nephelometric flasks. Cultures were incubated at 28 °C with shaking at 250 rpm. OD readings were obtained every hour using a Klett-Summerson colorimeter. Klett units were converted to OD540 by multiplying by 0.002. Equations for linear regressions were calculated at the exponential growth interval in each condition. The relative growth was then calculated as the slope value in each condition in proportion to the slope value of the zero-NaCl condition in each species.

Growth on YPD plates supplemented by the indicated NaCl concentrations (or the indicated oxidative and ER stress agents) was performed by spotting tenfold serial dilutions. Cells were grown overnight in YPD liquid medium and the OD600 was adjusted to 1.0. Dilution series were prepared in YPD and aliquots were spotted onto agar plates. Plates were incubated at 28 °C for 48 h and photographed. When required, incubation temperature and time were modified as indicated.

Gene cloning and disruption

The DhHOG1 gene was amplified by PCR using primers DhHog1F and DhHog1R (Table S1) and genomic D. hansenii DNA as a template. The PCR product was cloned into the pGEM-T Easy Vector.

The D. hansenii hog1Δ mutant was constructed by homologous recombination using a PCR amplicon containing the SAT1-yeYFP cassette and HOG1 recombinant ends (Fig. 3a). The PCR amplicon was synthesised using the hybrid oligonucleotides Dhhog1∆p244hyb_F and R (Table S1) and the pAYCU244 plasmid as a template.

DhHog1-yeYFP fusion

DhHOG1 tagging with the yeast enhanced Yellow Fluorescent Protein (yeYFP) sequence was achieved by successive PCR reactions using plasmid pAYCU244 as starting template (Fig. 3a). A PCR reaction was designed to fuse the yeYFP ORF in frame with the 3′ end of DhHOG1, and to eliminate the HOG1 stop codon at the same time. A separate PCR reaction was carried out to fuse the SAT1 terminator region with the DhHOG1 3′ UTR. These two PCR products were fused in a third PCR reaction that yielded an amplicon of 2899 bp containing 120 bp of the DhHOG1 3′ ORF, followed by the yeYFP gene (with the TRP1 terminator), the SAT1 gene (flanked by the TEF1 promoter and the PGK terminator) and 120 bp of the DhHOG1 3′ UTR. The primers used for the PCR reactions are listed in Table S1. The PCR product was sequenced and used for yeast transformation.

Yeast transformation

D. hansenii cells were grown in 10 ml YPD for 24 h. 1 ml of this culture was used to inoculate 50 ml of YPD and allowed to grow for 24 h. Cells were harvested by centrifugation (4500g, 10 min, RT), suspended in 6 ml of 50 mM phosphate buffer pH 7.5 to which fresh 25 mM DTT was added, and incubated at 30 °C for 15 min. After centrifugation (4000g, 10 min, 4 °C), cells were washed twice with 20 ml ice-cold water, once with 20 ml ice-cold 1 M sorbitol, and finally harvested by centrifugation (4000g, 10 min, 4 °C). Cells were resuspended in 3 ml ice-cold 1 M sorbitol. 1 µg DNA (suspended in 10 µl water) was added to a 200 µl cell suspension and transferred to a 0.2 cm gap width electroporation cuvette. A BioRad gene pulser was used for electroporation with an electric pulse of 2.3 kV, 25 µF and 200 ohms. The cells were immediately washed out from the cuvette with 1 ml YPD. This cell suspension was used to inoculate 10 ml of YPD and incubated for 24 h (28 °C, 150 rpm) to allow cell regeneration. 100 µl were plated onto YPD supplemented with 150 µg/ml clonNAT and incubated at 28 °C until colonies appeared (2–3 days). Transformants were transferred to fresh YPD plus 150 µg/ml of clonNAT and analysed for DNA integration.

Phosphorylation assays and immunoblotting

Western blots were performed as reported by Vázquez-Ibarra et al. (2018). Cells grown overnight in YPD were collected, washed, resuspended in fresh medium and treated with different concentrations of NaCl for different periods of time (as indicated in each figure). Cells were fixed with 85% trichloroacetic acid for 10 min at room temperature (RT) and washed twice with 1 ml water. Cells were lysed with SB-DTT buffer and glass beads as previously described (Velázquez-Zavala et al. 2015). Protein samples (20 µl of cell-free extracts for D. hansenii and 3 to 5 µl for S. cerevisiae) were resolved by SDS/PAGE and transferred to an Immobilon polyvinylidene difluoride membrane (Millipore, Edo. de México, México). Immunoblotting was performed using anti-Hog1 (1:1000 y-215; Santa Cruz Biotechnology, Dallas, TX, USA) for total Hog1 detection, and anti-phospho-p38 (1:1000; Cell Signaling Technology, Danvers, MA, USA) for phospho-Hog1 detection. Reconstituted goat anti-rabbit HRP antibody from Invitrogen, G21234 (0.1 mg/ml in 50% glycerol-PBS, 1:10,000) was used as a secondary antibody. Immobilon™ Western Chemiluminescent HRP substrate (Millipore, MA., USA) was used in detect the chemiluminescence. All western blots were repeated three times and representative blots are shown in figures.

Fluorescence microscopy

Cells containing the integrative SAT1-yeYPF cassette into the DhHOG1 locus were grown overnight in YPD plus 150 µg/ml clonNAT. Cells were harvested and washed with water and observed with an LSM 710-Zeiss microscope with the Zen black 2.3 software. Images were obtained by confocal microscopy in the double photon mode. yeYFP was detected at 514–527 nm, excitation and emission wavelengths respectively.

The DhHog1-yeYFP fusion was detected in cells grown overnight in YPD, collected and suspended at OD600 = 1.0. Cells were then centrifuged and resuspended in fresh YPD plus 2 M NaCl and incubated at 28 °C. 1 ml samples were taken at different times and fixed with 150 ul 37% formaldehyde for 15 min. Fixed cells were pelleted and washed three times with 1 ml PBS. For nucleus staining, DAPI (4′,6-diamidino-2-pheylindole) was added at a final concentration of 90 nM, incubated for 30 min in the dark at RT, washed three times with PBS and suspended in 100 ul PBS. yeYFP was visualised using a NIKON epifluorescence microscope with a green filter. Images were taken and processed using QCapture Pro (version 6.0) software.

Measurements of glycerol production

Overnight cultures were used to prepare 2.0 OD600 suspensions in YPD. For each experimental sample, 10 ml of the suspension were centrifuged and re-suspended in 5 ml of fresh YPD supplemented or not with the indicated NaCl concentrations and incubated for 4 h in a rotator at 28 °C. For determination of external glycerol, 1 ml of each suspension was centrifuged at 1625g and the supernatant kept on ice. For total glycerol, 1 ml of each suspension was boiled for 15 min in capped tubes, placed on ice for 10 min, vortexed for 2 min and centrifuged; glycerol content was determined in 10 µl of each supernatant with the enzymatic Glycerol kit (Boehringer Mannheim/R-Biopharm) and adapted to be performed in 96 well plates. A POLARstar Omega microplate reader was used for measurement of absorbance at 340 nm. Glycerol content was normalised to the yeast dry weight of each sample. The internal glycerol was calculated by means of the subtraction of the external glycerol value from the total glycerol value in each condition. Data shown are average values ± SD. Every sample was measured in duplicate from 3 biological replicates.

RNA extraction

Overnight cultures of wild type D. hansenii and Dhhog1Δ mutant were adjusted to 0.5 OD600 in 100 ml YPD media and incubated for 3 h at 28 °C with shaking. 50 ml of each culture was centrifuged (5 min, 1750g) and resuspended in 10 ml of YPD supplemented or not with 2 M NaCl and incubated for 60 min under the same conditions. After incubation, cells were centrifuged and resuspended in 1 ml of AE buffer (50 mM sodium acetate and 10 mM EDTA) to perform the RNA extraction protocol according to Schmitt et al. (1990). RNA integrity was verified by electrophoresis in a 1% denaturing agarose gel.

Analysis of gene expression

Total RNA was digested with DNaseI (RQ1 RNase-Free DNase, Promega) to remove any contaminating genomic DNA. cDNA synthesis reactions were performed using the ImProm-II™ Reverse Transcription System (Promega kit). Quantitative real-time PCR (qRT-PCR) was performed using the standard curve method with specific primers for the genes DhGPD1 (ID: 2903610), DhSTL1 (ID: 2902951) and DhACT1 (ID: 2901278), encoding putative NAD+-dependent glycerol-3-phosphate dehydrogenase, the glycerol/H+ symporter of the plasma membrane and actin, respectively. Oligonucleotides were initially screened for the absence of dimers formation and cross-hybridisation. Oligonucleotide primer pairs with 100% of amplification efficiencies were used (Table S1). Control reactions without reverse transcriptase were performed. qRT-PCR analysis was performed using a Rotor-Gene Q (Qiagen) machine. The detection dye used was SYBR Green (2 × SYBR Select Mastermix from Applied Biosystems). qRT-PCR was carried out as follows: 95 °C for 5 min (1 cycle), 95 °C for 15 s, 58 °C for 20 s, and 72 °C for 20 s (35 cycles). Transcripts were normalised relative to the DhACT1 transcript quantities. Relative expression levels (fold induction) were evaluated with respect to wild type cells (control) by the standard curve method. Data shown are the mean values ± SD of four biological replicates.

Results

Determination of the halotolerance level of D. hansenii

To determine the maximum NaCl tolerance of D. hansenii, we established growth curves in YPD medium supplemented with different salt concentrations. To get a clear idea of the salt resistance of D. hansenii, this and the following experiments were conducted in comparison to the halosensitive S. cerevisiae strain. Due to the preferred respiratory metabolism of D. hansenii (Sánchez et al. 2008), the growing parameters were obtained in well-aerated flasks, to ensure optimal O2 availability. We detected that as the NaCl concentration increased, the growth velocity of both strains decreased (Fig. 1); however, 1 M NaCl only showed a moderate effect on the D. hansenii growth (Fig. 1a) in contrast to S. cerevisiae, which at such concentration had a severe growth defect (Fig. 1b). This is more clearly observed when comparing the relative growth values (obtained with the slope of the linear regression) (Fig. 1, inserts). The S. cerevisiae growth was reduced by 70% at 1 M NaCl, while the same NaCl concentration decreased growth by only 20% in D. hansenii. At higher NaCl concentrations, S. cerevisiae did not grow, while D. hansenii was still able to divide. 50% reduction in D. hansenii growth velocity was detected with a NaCl concentration as high as 1.5 M. Although 2 M NaCl displayed a significant effect on the growth properties of D. hansenii, it accumulated biomass and reached the stationary phase in twice the time it took the culture with no salt (not shown). These observations indicate that D. hansenii can grow adequately in NaCl concentrations that are harmful to S. cerevisiae and that under these growth conditions it can tolerate as much as 2 M NaCl.

Fig. 1
figure 1

Growth curves of Debaryomyces hansenii (a) and Saccharomyces cerevisiae (b) under different NaCl concentrations. Cultures were grown in YPD with the indicated NaCl concentrations, at 28 °C, in a rotary shaker at 250 rpm. The initial inoculum with freshly harvested overnight YPD cells, was adjusted to 0.1 OD600. OD readings were taken every hour in a Klett- Summerson colorimeter. Linear regressions were adjusted, and the relative growth was calculated with the slope values normalised to the condition with zero NaCl. A representative graph of 3 different biological replicas is shown

Full size image

Kinetics of DhHog1 phosphorylation upon NaCl treatment

In S. cerevisiae, hyperosmotic stress leads to the activation of the HOG pathway through transient phosphorylation of the Hog1 MAPK. The time-lapse of phosphorylated Hog1 depends on the severity of the hyperosmotic stimulus, for instance ranging from 40 to 60 min under moderate stress conditions (0.4 to 0.6 M NaCl) (Van Wuytswinkel et al. 2000; Vázquez-Ibarra et al. 2018). To analyse the contribution of Hog1 to the salt tolerance of D. hansenii, we determined the phosphorylation kinetics of DhHog1 in NaCl concentrations where the cell growth was either unaffected or compromised. DhHog1 phosphorylation was monitored with an anti-phospho-p38 antibody that detects the dually phosphorylated form. When D. hansenii cells were exposed to increasing NaCl concentrations for 10 min, DhHog1 was barely phosphorylated at 0.3 M, but this gradually increased until a maximum plateau was reached at 1–2 M (Fig. 2a). Although DhHog1 phosphorylation dropped significantly at concentrations higher than 2 M it was still detected at 3 M NaCl (Fig. S1). In contrast, ScHog1 reached its maximum at 0.5 M NaCl and then abruptly declined until being undetected at 2 M NaCl. A time-course assay was subsequently performed using the NaCl concentration at which the peak of phosphorylation was detected in the previous experiment. In this assay, the level of phosphorylated DhHog1 increased with time reaching its maximum at 60 min and then decreased gradually (Fig. 2b), but it was still detected after 180 min exposure to NaCl. In S. cerevisiae, the time-lapse of phospho-Hog1 under these conditions was much more transient than that observed in D. hansenii. Compared to ScHog1, the right shift of phosphorylated DhHog1 abundance in both the dose-dependent response and the time course indicates that DhHog1 is phosphorylated at higher NaCl concentrations than ScHog1 and for longer periods. This observation is in agreement with the salt tolerance level that D. hansenii showed in the cell growth assay and suggests that Hog1 can be required to some extent to induce resistance to high salt concentration.

Fig. 2
figure 2

Dose-dependent and time-course Hog1 phosphorylation in D. hansenii and S. cerevisiae. Cells were grown in YPD to mid-log phase and treated with NaCl. a Dose–response. Cells were treated with the indicated NaCl concentrations for 10 min (D. hansenii) or 5 min (S. cerevisiae). Protein extracts from the opposite strain treated with the indicated NaCl concentration (10 min for Sc and 5 min for Dh) were included in each blot. Extended NaCl dosage to 3 M is shown in Fig S1. b Time course. Cells were treated with 2 M (D.hansenii) or 0.5 M (S. cerevisiae) NaCl for the indicated times. Relative phospho-Hog1 abundance was determined by densitometry of three independent blots. Values represent the quotient of phospho-Hog1 divided by total Hog1 and multiplied by 10. Bars represent the standard error. Total Hog1 was detected with anti-Hog1 antibody and phosphorylated Hog1 was detected in the same stripped membrane with an anti-phospho p38 antibody. Representative blot images are presented

Full size image

Construction of the DhHOG1 null mutant

These previous observations suggest that the Hog1 MAPK may play an important role in the halotolerance mechanism in D. hansenii. The DhHOG1 gene was cloned and sequenced (Bansal and Mondal 2000) and shown to encode a protein highly similar to Hog1 from a variety of species, including non-conventional yeast species (Fig. S2). The deduced DhHog1 protein is predicted to be 387 amino acid residues in length and shares 86% identity with ScHog1. It displays a conserved phosphorylatable TGY motif within the activation loop, and the two Pbs2 binding domains: the common docking (CD) domain and the PBD-2 domain (Murakami et al. 2008). The C-terminus of DhHog1 shows extensive differences in length and sequence compared to that of ScHog1; however, DhHog1 conserves the region that is involved in the regulation of ScHog1 autophosphorylation (Maayan et al. 2012). Interestingly, alignment shown in Fig. S2 indicates that species related to the CTG clade display globally a reduced Hog1 sequence in length concerning their C-terminus. This could reflect that, as previously observed in Saccharomycotina two-component systems (Hérivaux et al. 2018), some clade-specific evolution paths in stress response cell circuitries have likely intervened in budding yeasts.

To assess the contribution of DhHog1 to the halotolerance of D. hansenii, we designed a strategy for the construction of a HOG1-deleted strain. Conventional tools for gene replacement and inactivation are not useful in D. hansenii, since this yeast belongs to the CTG clade which ambiguously translates the CUG codon into Ser or Leu (Tekaia et al. 2000). In this work the pAYCU244 plasmid (Fig. 3a) was used, which was previously optimised for genetic engineering the D. hansenii codon usage (Defosse et al. 2018). A cassette containing the SAT1 gene (which confers nourseothricin resistance) and the yeYFP reporter gene (yeast-enhanced Yellow Fluorescent Protein) was used to replace the DhHOG1 locus by homologous recombination (Fig. 3a). Only 9 clones (out of 624 nourseothricin-resistant clones) that were sensitive to 2 M NaCl were isolated. These 9 clones were subjected to a series of PCR reactions to explore the structure of their HOG1 loci (Fig. 3b). Three clones appeared to have integrated the SAT1-yeYFP cassette in the right locus (Clones 2, 5 and 8), however, clones 5 and 8 presented also the wild type HOG1 amplicon. Only in clone 2 was the SAT1-yeYFP cassette correctly integrated by a double-crossing over in the HOG1 locus, suggesting that this was indeed a bona fide HOG1 null mutant. This was further confirmed with two more observations: clone 2 showed a high expression of the yeYFP reporter (Fig. 3c) and it did not show cross reaction with the anti-Hog1 antibody nor with the anti-p38 antibody after treatment with 1 M NaCl (Fig. 3d). Taken together, the previous observations suggest that double-crossing over is not a frequent event in D. hansenii and that most integrations occur ectopically as previously observed in other unconventional yeast species (Schorsch et al. 2009; Kretzschmar et al. 2013; Oguro et al. 2017).

Fig. 3
figure 3figure 3

a pAYCU244 plasmid map and DhHOG1 deletion strategy. Only relevant genes and regulatory sequences of pAYCU244 (Defosse et al. 2018) are depicted. The SAT1 gene confers resistance to nourseothricin and the yeYFP gene encodes the yeast-enhanced yellow fluorescent protein. Primers to generate the SAT1-yeYFP cassette flanked by 40 bp recombinant tails (purple) are indicated with small black arrows. Cassette integration and DhHOG1 gene replacement are indicated with tiny dotted grey lines. Small colored arrows indicate primers used to verify gene replacement. b PCR analysis of putative recombinant clones. Genomic DNA was used as a template. Coloured arrows (same code as a) on top of gels indicate the primers used in each reaction. Dh denotes PCR products from the WT strain. Expected PCR products size (bp) is indicated. Asterisk indicates clone with the right integration. c Microscopic images of D. hansenii WT strain (control) and clone No. 2. Fresh cells grown in YPD or YPD + nourseothricin (clone No. 2) were observed by confocal microscopy in the double photon mode with an LSM 710-Zeiss microscope. yeYFP was observed at 514–527 nm, excitation and emission wavelengths respectively. Control images were visualised in brightfield mode. d Hog1 and phospho-Hog1 detection in putative recombinant clones by western-blotting. Cells grown to mid-log phase were incubated for 10 min in 1 M NaCl and protein extracts were prepared. Total Hog1 was detected with an anti-Hog1 antibody and phospho-Hog1 was detected in the same stripped membrane with an anti-p38 antibody. D. hansenii (Dh) and S. cerevisiae (Sc) protein extracts from cells treated in the same conditions were loaded in the same gel

Full size image

DhHog1 is required for halotolerance in D. hansenii

The participation and contribution of Hog1 in D. hansenii halotolerance were determined by spotting tenfold serial dilutions onto YPD plates containing increasing NaCl concentrations. In this assay, the wild-type (WT) D. hansenii strain grew adequately under all NaCl concentrations tested (Fig. 4) which is in agreement with the growth kinetics depicted in Fig. 1. Compared to the Schog1Δ mutant, which is highly sensitive to mid-NaCl concentrations, the growth of the Dhhog1 mutant on 0.5 M NaCl was indistinguishable from that of the WT strain. However, inactivation of DhHog1 caused a strong reduction of growth at 1 M and higher NaCl concentrations. The effect of sorbitol, which imposes a different sort of osmotic stress than NaCl, was also tested. Growth of WT D. hansenii cells on sorbitol was similar to that on NaCl, however, the Dhhog1Δ mutant was very sensitive to high sorbitol concentrations. Close examination of plates with osmo-equivalent concentrations of NaCl and sorbitol (e.g. 1.0 M NaCl vs 1.5 M sorbitol) (Almagro et al. 2000), show a more pronounced requirement of DhHog1 to deal with high sorbitol concentrations. These observations indicate that DhHog1 is needed for optimal growth at high osmotic conditions, which is consistent with the observed phosphorylation peak which occurred at 2 M NaCl, and that it is required for halotolerance in D. hansenii.

Fig. 4
figure 4

Effect of NaCl and Sorbitol dosage in cell growth. The indicated strains were grown to mid-log phase and adjusted to OD600 = 1.0. Aliquots of tenfold serial dilution were spotted on YPD or YPD containing the indicated NaCl or Sorbitol concentrations. Plates were incubated for 72 h at 28 °C and photographed. DhHog1Δ corresponds to clone 2 of Fig. 3

Full size image

DhHog1 is partially translocated to the nucleus after hyperosmotic treatment

In S. cerevisiae, phosphorylation induces translocation of Hog1 from the cytoplasm to the nucleus, and this is achieved in WT cells after 5 min of exposure to hyperosmotic stress (Ferrigno et al. 1998). To determine whether DhHog1 would translocate into the nucleus after 2 M NaCl exposure, DhHog1 was tagged with the yeYFP protein. A similar strategy as for the construction of the HOG1 null mutant was designed except that the yeYFP gene was cloned in frame into the 3′ end of DhHOG1. Integration of the HOG1-yeYFP-SAT cassette into the DhHOG1 locus was also done by homologous recombination. With this strategy, it was detected that DhHog1 is in the cytoplasm under isosmotic conditions, but after different incubation times under hyperosmotic shock (2 M NaCl), a fraction of the yeYFP signal was detected also into the nucleus (Fig. 5). In our experiments we observed that DhHog1 was evenly distributed throughout the cytoplasm and nucleus from 10 to 180 min (Fig. S3) after the hyperosmotic treatment, although by 180 min the amount of phospho-DhHog1 had decreased significantly (Fig. 2b).

Fig. 5
figure 5

DhHog1-yeYFP localisation after hyperosmotic treatment. Cells carrying the integrated DhHOG1-yeYFP hybrid gene were grown to mid-log phase and treated with or without (T0) 2 M NaCl for 10 min (T10) or 60 min (T60). Representative images of cells showing DhHog1-yeYFP and DAPI (nucleus) are shown. Images were acquired by epifluorescence microscopy. An extended time course (from 5 to 180 min) is shown in Fig S3

Full size image

DhHog1 regulates transcription of DhGPD1 and DhSTL1 under hyperosmotic stress

Hyperosmotic stress induces changes in the transcription of several genes required for short- and long-term responses. To determine whether D. hansenii Hog1 has transcriptional activity on canonical osmo-regulated genes, expression of DhGPD1 (encoding the NAD+-dependent glycerol-3-phosphate dehydrogenase) and DhSTL1 (encoding the glycerol-proton symporter of the plasma membrane) under hyperosmotic stress in both WT and Dhhog1Δ strains was measured. High osmotic shock led to a strong induction of transcription in the WT strain, in which both genes were upregulated about tenfold under 2 M NaCl (Fig. 6). In the Dhhog1Δ mutant, the expression of DhGPD1 and DhSTL1 dropped under osmotic stress twofold and fourfold respectively. Interestingly, while DhSTL1 expression was fully dependent on DhHog1 under stress, around 50% of the DhGPD1 expression did not depend on DhHog1 in high NaCl concentration.

Fig. 6
figure 6

Effect of hyperosmotic stress and DhHog1 inactivation on gene expression. WT and Dhhog1Δ cells were grown in YPD to OD = 0.5 and treated or not treated with 2 M NaCl for 1 h. Total RNA was obtained and subjected to qRT-PCR analysis. Relative transcript levels of DhGPD1 or DhSTL1 were obtained from the mean value of three independent experiments (± SD) normalised to the transcript levels of the DhACT1 gene

Full size image

DhHog1 partially contributes to glycerol production upon hyperosmotic stress

In S. cerevisiae, Hog1 regulates glycerol concentration through two mechanisms: rapid accumulation due to direct modulation of metabolic enzymatic activity, and a long-term transcriptional induction of genes required for glycerol production and accumulation (Schaber et al. 2012). We studied the contribution of DhHog1 to long-term glycerol production and extracellular accumulation, which could explain the cellular adaptation to high salt environments. Cells were incubated with increasing NaCl concentrations and total and external glycerol were measured. The hyperosmotic treatment induced a gradual increase in total glycerol concentration, which reached its maximum at 1.5–2 M NaCl (Fig. 7). The amount of glycerol at these NaCl concentrations was about threefold higher compared to the untreated cells. At 2.5 M NaCl the amount of glycerol decreased, most probably due to pleiotropic defects in cell physiology and viability. The total glycerol accumulation in the Dhhog1Δ mutant followed similar kinetics compared to the WT strain, except that the maximum accumulation in 2 M NaCl was twofold higher than the condition with no stress and this value represents 60% of the maximum amount of glycerol accumulated in the WT strain (Fig. 7). The internal glycerol followed the same kinetics as the total glycerol in both, the WT strain and the Dhhog1Δ strain (Fig. 7, insert). These observations indicate that the contribution of DhHog1 to total glycerol accumulation is about 40% under hyperosmotic conditions. This result is in line with the previous observation through which it was detected that the expression of DhGPD1 has a component that is independent of DhHog1 under hyperosmotic conditions. In contrast to total glycerol, the amount of external glycerol in the WT strain remained constant until 1.5 M NaCl, and then dropped moderately at 2.0 and at 2.5 M (Fig. 7). A similar kinetics for external glycerol under the NaCl dose curve was detected in the Dhhog1Δ mutant, indicating that DhHog1 does not contribute to the regulation of glycerol flux through the membrane under these conditions.

Fig. 7
figure 7

Effect of hyperosmotic stress and DhHog1 inactivation on glycerol accumulation. WT and Dhhog1Δ cells were grown in YPD to OD = 2.0 and treated or not treated with the indicated NaCl concentrations for 4 h. Cell suspension was divided into two samples; one was boiled for 15 min and used for total glycerol determination; the second was centrifuged and the supernatant was used for external glycerol determination. Internal glycerol (insert) was calculated by means of the subtraction of external glycerol values from total glycerol values. Glycerol content is the mean value (± SD) of three independent experiments and referred to yeast dry weight

Full size image

Involvement of DhHog1 in the response to a variety of stressors

To date, it is clear that HOG pathways have pleiotropic roles in a variety of stressful conditions (Hernández-Elvira et al. 2019). Such is the case of the Schizosaccharomyces pombe Sty1 pathway, which is analogous to the S. cerevisiae HOG pathway. In S. pombe, Sty1 participates in a variety of stresses such as osmotic stress, heat shock, oxidative stress, UV light, nitrogen starvation and some other stress conditions (Hohmann 2002). To assess whether DhHog1 plays a role in stress conditions other than hyperosmotic stress, we determined cell growth by spotting serial dilutions of cell suspension onto plates containing agents that induce oxidative stress, endoplasmic reticulum (ER) stress, and alkaline stress, and also under cold and heat stress. Under conditions of oxidative stress induced with either hydrogen peroxide (H2O2) or t-butyl-hydroperoxide (t-BHP) we were able to make two general observations: First, compared to S. cerevisiae, D. hansenii was highly sensitive to H2O2, but surprisingly resistant to stress induced with t-BHP (Fig. 8a). Secondly, the Dhhog1Δ mutant was moderately sensitive to oxidative stress induced with H2O2 but not with t-BHP, while the Schog1Δ mutant was resistant to both agents (Fig. 8a). These observations suggest that DhHog1 may participate in the response to oxidative stress which is in agreement with the DhHog1 phosphorylation induced with H2O2 (Sharma et al. 2005). Under ER stress conditions induced with the antibiotic tunicamycin (Tn) (Torres-Quiroz et al. 2010), it was found that the Dhhog1Δ mutant was moderately sensitive to Tn, at a concentration where the Schog1Δ mutant shows severe growth impairment (Fig. 8a).

Fig. 8
figure 8

a Effect of oxidative and ER stress on cell growth. The indicated strains were grown to mid-log phase and adjusted to OD600 = 1.0. Aliquots of tenfold serial dilution were spotted on YPD or YPD containing the indicated concentrations of hydrogen peroxide (H2O2), t-butyl-hydroperoxide (t-BHP) or tunicamycin (Tn). Plates were incubated for 72 h at 28 °C and photographed. Oxidative and ER stress assays were performed with the same cultures at the same time, so the control plate has been duplicated for better comparison. b Effect of alkali and hyperosmotic stress in cell growth. The indicated strains were grown and platted as indicated in a. The pH of YPD medium was adjusted to the indicated values prior to sterilisation. 50 mM MES-TEA, 50 mM HEPES-TEA or 50 mM TAPS-TEA were used to buffer at pH 6.0, 7.0 and 8.0, respectively. Plates were incubated for 72 h at 28 °C and photographed

Full size image

Regarding the stress imposed by pH, the alkali-tolerant nature of D. hansenii is known. It can grow reasonably well at a pH of 8.0, a condition where S. cerevisiae is highly sensitive (Fig. 8b; Sánchez et al. 2018). Accordingly, the ability of the Dhhog1Δ mutant to grow on plates buffered at pH 8.0 compared with neutral and acidic pH (7.0 and 6.0 respectively) was tested. First, as expected, it was found that the WT strain of D. hansenii grew adequately in the three conditions, while the growth of S. cerevisiae at pH 8.0 was impaired (Fig. 8b). Interestingly, the Dhhog1Δ mutant showed an unexpected high sensitivity to pH 8.0 when combined with 1 M NaCl, which is a mid-stress condition for D. hansenii (Fig. 8b). The effect of hyperosmotic stress on the Dhhog1Δ mutant was not observed at pH 6.0 and 7.0. This indicates that under high salt concentrations, Hog1 is required to cope with alkali stress in D. hansenii.

Finally, it was explored whether DhHog1 is involved in response to cold stress and heat stress. D. hansenii Hog1 has no apparent role in this sort of stress (Fig. S4). However, it was detected that D. hansenii is cold-resistant, based on its ability to grow at 16 °C in short periods and at 4 °C in long periods (which is not observed in S. cerevisiae) and that it is sensitive to heat stress, due to its inability to grow at 33 °C or higher temperatures, at which S. cerevisiae still grows adequately.

Discussion

D. hansenii is one of the most halotolerant and osmotolerant yeasts, which under some conditions, it can grow even in concentrations as high as 4.0 M NaCl (Onishi 1963; Norkrans 1966). Given this characteristic, D. hansenii has high biotechnological potential in the food industry, which makes it of high interest to understand the molecular mechanisms that contribute to its halotolerant behaviour. Although some studies have addressed the participation of the HOG pathway in the response to hyperosmotic stress (Sharma and Mondal 2005; Sharma et al. 2005), some conclusions have been obtained by inter-species gene expression. Here we report for the first time the phenotypic characterization of a null HOG1 mutant in D. hansenii that allowed us to unequivocally determine the contribution of this MAPK to the halotolerance of this yeast. We have also given an accurate description of the activation kinetics of DhHog1, its transcriptional activity in canonical osmo-responsive genes, its participation in long-term glycerol production, and its involvement under different stress conditions.

Our observations indicate that WT D. hansenii can grow in NaCl concentrations ≥ 2 M, which is consistent with the previous classification of this species as a yeast belonging to the 3 M salt-stress class (Lages et al. 1999), and that deletion of DhHOG1 induces sensitivity to that range of concentrations. Although at a slower rate than the WT strain, a mutant devoid of DhHog1 was still able to grow in 1 M NaCl, a concentration that imposes a strong negative effect on S. cerevisiae as detected in this and other works (Blomberg 1997; Rodríguez-González et al. 2017).

Our data indicate that strong Hog1 phosphorylation depends on high external osmolarity and that in low osmolarity, where ScHog1 is phosphorylated at a high level, DhHog1 is not activated. Accordingly, high expression of DhSTL1 and DhGPD1 was detected at 2 M NaCl and the long-term glycerol production peaked at this concentration. These observations indicate that the D. hansenii HOG pathway is set to be activated with high external osmolarity and that in low and mid-osmotic conditions, other protection systems should be active. A similar situation is present in the extreme halotolerant yeast Hortaea werneckii, where its two redundant paralogous Hog1A and Hog1B are fully active at salinities ≥ 3 M, but not at moderate NaCl concentrations (Kejžar et al. 2015). The fact that the HOG pathway can be activated at high NaCl concentrations is supported by the ability of D. hansenii to accumulate a large amount of Na+ inside the cell, which is a trait that prompted some authors to describe it as a ‘sodium includer’ (Norkrans and Kylin 1969; Prista et al. 1997; Prista and Loureiro-Dias 2007).

In S. cerevisiae, phosphorylation is necessary to allow translocation of Hog1 from the cytoplasm to the nucleus, and this is achieved in wild-type cells after 5 min of exposure to hyperosmotic stress (Ferrigno et al. 1998). Nuclear localisation under this condition is transient and retro-translocation to cytoplasm occurs through the action of protein phosphatases, including nuclear Tyr-phosphatase Ptp2 (Mattison and Ota 2000). Thus, translocation to the nucleus of DhHog1 was detected within 10 min after hyperosmotic stress was imposed with 2 M NaCl and it was still detected inside the nucleus after 180 min of treatment. This behaviour is consistent with the phosphorylation time course obtained with 2 M NaCl. Interestingly, DhHog1 was also detected in the cytoplasm throughout the entire time course. These observations show some differences with previous studies in which DhHog1 was visualised by indirect immunostaining (Sharma et al. 2005). In those experiments, the nuclear accumulation of DhHog1 was delayed 90 min after hyperosmotic shock, and then re-entry into the cytoplasm occurred 150 min later. The discrepancy with our observations could be explained by differences in the detection level between indirect immunofluorescence and detection of the hybrid DhHog1-yeYFP, and/or differences in strain’s genetic backgrounds. Although we do not know whether cytoplasmic DhHog1 is active in our conditions, its long nuclear retention not only matches with its phosphorylation status but also its role under extreme hyperosmotic conditions in which sustained long activity is expected. Although a decline in phosphorylation occurred in long periods, DhHog1 was still present in the nucleus. Disregarding an effect of increased protein turnover, this observation suggests that a mechanism of downregulation, mediated perhaps by nuclear phosphatases (Mattison and Ota 2000) may be present in D. hansenii. Interestingly a study of comparative genomics of the HOG signalling system predicts lack of Ptp2 phosphatase in D. hansenii (Krantz et al. 2006), suggesting that Ptp3 (which is present) may take both roles.

The peak in glycerol accumulation at 2 M NaCl correlates very well with the high expression of DhGPD1 and DhSTL1 detected at this concentration. While expression of DhSTL1 was fully dependent on DhHog1, the DhHog1-dependent expression of DhGPD1 was 60%. The high expression of DhSTL1 is strongly correlated with the high glycerol/H+ symporter activity that has been determined in D. hansenii (Pereira et al. 2014). The expression of both DhGPD1 and DhSTL1 in D. hansenii, followed a pattern of behaviour similar to that found in S. cerevisiae. STL1 is an osmo-responsive gene whose expression completely depends on Hog1 (O’Rourke and Herskowitz 2004), while GPD1 not only depends on Hog1 under hyperosmotic stress but also on other signalling systems (Rep et al. 1999). It is worth considering that under hyperosmotic stress, GPD1 can also be transcriptionally regulated by a Cyc8/Tup1-SUMOylation-dependent des-repression mechanism (Nadel et al. 2019), thus several mechanisms can be acting in concert to achieve a tight control of DhGPD1 expression. Our observations indicate that a significant expression of DhGPD1 was induced by osmostress in the Dhhog1Δ mutant, suggesting that other pathways may be participating or that a fraction of DhGPD1 expression is controlled by a general stress mechanism in D. hansenii. It would be interesting to determine whether the stress caused specifically by Na+ ions in D. hansenii can activate the Ca++-Calcineurin pathway and to what extent this may regulate the response to hyperosmotic stress.

In S. cerevisiae, Hog1 regulates glycerol concentration through two distinct mechanisms. One is a rapid accumulation of glycerol which includes the closure of the Fps1 glycerol channel to diminish glycerol efflux (Lee et al. 2013) and the increase of glycerol synthesis by upregulation of both GPD1 (Westfall et al. 2008), and PFK2 (which encodes 6-phosphofructo-2-kinase) (Dihazi et al. 2004). The second is the transcriptional regulation of a large set of genes required for glycerol synthesis and accumulation (O’Rourke and Herskowitz 2004). Here we found that DhHog1 contributes to long term responses. Indeed, the amount of accumulated glycerol gradually increased with the increase in osmolarity until reaching a plateau at 1.5–2 M NaCl, in both, total and internal glycerol. This long-term glycerol production was just partially dependent on Hog1 since the maximum accumulation dropped just 1.5 fold in the Dhhog1Δ. The DhHog1-independent glycerol accumulation has not only the transcriptional component as discussed above, but also the increase in enzymatic activity directly exerted by NaCl (Adler et al. 1985; André et al. 1991). In fact, glycerol production can be enhanced by increased activity of Gpd1 and dihydroxyacetone kinase (Dak1,2) promoted by NaCl, which appears to modulate the specific activity of these enzymes. Interestingly, D. hansenii Gpd1 displays 10-folds more specific activity in the presence of NaCl than that of S. cerevisiae (Adler et al. 1985; André et al. 1991; Thomé 2005). Although the role of glycerol as a suitable compatible osmolyte in D. hansenii has been determined (Adler and Gustafsson 1980; Lages et al. 1999; Gori et al. 2005), it is important to consider that other polyols like arabinitol and trehalose may also contribute to the osmo-protection through a combined interplay (Adler and Gustafsson 1980; González-Hernández et al. 2005). The increase in total and internal glycerol accumulation and the contribution of Hog1 in D. hansenii coincide with those of Kluyveromyces lactis, where glycerol increases fivefold under hyperosmotic stress and KlHog1 contributes 50% to this increase (Rodríguez-González et al. 2017). Given the halo-sensitive nature of K. lactis, the common feature between both species is that they share highly respiratory metabolisms.

Regarding the amount of external glycerol under increasing NaCl concentrations, we found that this did not follow the same kinetics as total and internal glycerol. In fact, although no significant differences were detected, external glycerol showed a tendency to diminish as the concentration of NaCl rose. One likely explanation was that this was due to increased glycerol uptake mediated by high Stl1 symporter activity. We have disregarded this possibility since the profile of the mutant strain, where expression of DhSTL1 is negligible, mirrored that of the wild type strain. However, it is important to mention that D. hansenii can accumulate glycerol against a strong concentration gradient (Alder et al. 1985). Whether this is due only to the DhStl1 activity and/or to other transport systems remains to be determined. Another possibility could be that glycerol efflux is a low capacity process in D. hansenii, due to the lack of a canonical aquaglyceroporin (Fps1 in S. cerevisiae) capable of releasing glycerol under hyperosmotic stress (Petterson et al. 2005; Sabir et al. 2016). It is worth noting, however, that D. hansenii contains a gene that is orthologous to the S. cerevisiae aquaporins Aqy1 and Aqy2, which are orthodox water channels that apparently may transport not only water but also solutes (Petterson et al. 2005). Orthodox aquaporins have been poorly studied in yeast species. In S. cerevisiae, functional expression of aquaporins is strongly dependent on the strain. When expressed, aquaporins appear to play a role in freeze tolerance, but Aqy1 seems to be specific for sporulation, while Aqy2 is expressed in vegetative cells but is downregulated by hyperosmotic stress in a Hog1-dependent manner (Ahmadpour et al. 2014). The genome of D. hansenii appears to encode only one aquaporin (which is slightly more related to ScAqy2) (Sabir et al. 2016). Whether this putative DhAqy2 participates in the regulation of glycerol efflux under hyperosmotic stress in D. hansenii, it is necessary to speculate that its regulation should largely differ from that of S. cerevisiae. It would be interesting to determine whether DhAqy2 has a role in the cold tolerance of D. hansenii detected in this work. Finally, the total and external glycerol reduction observed in concentrations over 2 M NaCl, which coincide with the drop in DhHog1 phosphorylation, cannot be explained as a regulatory mechanism exerted by DhHog1, but rather by an altered metabolic condition promoted by the hyperosmotic stress.

In addition to its participation in the hyperosmotic stress response, ScHog1 has been shown to have pleiotropic activities in a variety of conditions; for example, it has a role in the endoplasmic reticulum (ER) stress response (Torres-Quiroz et al. 2010; Hernández-Elvira et al. 2019), in the general stress response (de Nadal and Posas 2010) and even in mitophagy (Mao and Klionsky 2011). According to the results reported in this work, DhHog1 also appears to play a role in response to oxidative stress, which is in agreement with the transcriptional induction, triggered by NaCl, of genes for oxidative protection (Segal-Kischinevzky et al. 2011; Calderón-Torres et al. 2011; Michán et al. 2013; Ramos-Moreno et al. 2019); DhHog1 may also play a role in the response to ER stress induced by tunicamycin, which inhibits protein N-glycosilation. Since this is extranuclear Hog1 activity (Torres-Quiroz et al. 2010; Hernández-Elvira et al. 2019), it was not surprising to detect DhHog1 in the cytoplasm during hyperosmotic stress. Our results also suggest that DhHog1 does not participate in cold stress nor in heat stress. Although we performed a wide screening at a variety of temperatures, we do not discard DhHog1 cross-activation by simultaneous induction of temperature stress and hyperosmotic or oxidative stress.

An interesting finding in this work was the strong sensitivity displayed by the Dhhog1Δ mutant to combined hyperosmotic stress and alkali stress. The resistance of D. hansenii to alkaline conditions is a property well known since the first isolated strains came from the ocean (pH around 8.0) (Norkrans 1966, 1968), and another strain has been isolated from soy sauce with a pH around 10 (Kurita and Yamazaki 2002). In natural environments, where a myriad of stressful conditions are simultaneously present, the finding that Hog1 is essential to cope with alkali stress in a mid-osmotic condition is quite relevant to the understanding of the physiological characteristics of an alkali-halotolerant yeast. In fact, it has been demonstrated that D. hansenii has a high capacity to acidify a medium buffered at pH 8.0 in the presence of high salt concentration (Sánchez et al. 2018). Additionally, D. hansenii expresses an alkali-metal-cation/H+ antiporter of broad spectrum (Velkova and Sychrova 2006; Ramos et al. 2011) and two Na+-ATPases which are overexpressed in high NaCl concentrations and one of them also in high pH (Almagro et al. 2001). It now, appears that this alkali-resistance is at least in part regulated by the Hog1 MAPK in D. hansenii.

Our study points to the scenario where the HOG pathway of D. hansenii is activated with high external osmolarity, and that under low and mid-osmotic conditions other protection systems are active. Several studies have identified factors that may play a role in the halotolerance of this yeast, including the increase of ergosterol in the plasma membrane in high salt concentrations (Turk et al. 2007) and the induction of expression and activity of several transporters such as the K+–Na+ symporter (Martínez et al. 2011). Thus, several osmoregulatory mechanisms must act in a coordinated fashion to permit an adaptive cellular response to salty environments.

Here we described for the first time the construction of a HOG1 null mutant by homologous recombination in D. hansenii, a technical approach that has been elusive for some time, and most importantly we have determined the threshold of salt concentration that D. hansenii may tolerate without Hog1 and the contribution that this MAPK plays in the halotolerance of this species.