The Role of Plasma Membrane Pleiotropic Drug Resistance Transporters in the Killer Activity of Debaryomyces hansenii and Wickerhamomyces anomalus Toxins

The killer strains of Debaryomyces hansenii and Wickerhamomyces anomalus species secrete antimicrobial proteins called killer toxins which are active against selected fungal phytopathogens. In our research, we attempted to investigate the role of plasma membrane pleiotropic drug resistance (PDR) transporters (Pdr5p and Snq2p) in the mechanism of defense against killer toxins. Saccharomyces cerevisiae mutant strains with strengthened or weakened pleiotropic drug resistance due to increased or reduced number of mentioned PDR efflux pumps were tested for killer toxin susceptibility. The present study demonstrates the influence of the Snq2p efflux pump in immunity to W. anomalus BS91 killer toxin. It was also shown that the activity of killer toxins of D. hansenii AII4b, KI2a, MI1a and CBS767 strains is regulated by other transporters than those influencing W. anomalus killer toxin activity. In turn, this might be related to the functioning of the Pdr5p transporter and a complex cross-talk between several regulatory multidrug resistance networks. To the best of our knowledge, this is the first study that reports the involvement of PDR transporters in the cell membrane of susceptible microorganisms in resistance to killer yeasts’ toxins.


Introduction
Killer strains of the species Debaryomyces hansenii and Wickerhamomyces anomalus exhibit antagonistic activity against several yeast species and fungal phytopathogens. Thus, they are considered promising biocontrol agents in plant protection [1][2][3][4][5][6][7][8][9][10][11][12]. Antagonistic activity of D. hansenii and W. anomalus killer strains can be manifested through various mechanisms, including competition for nutrients and space, mycoparasitism, biofilm formation, induction of plants' immune response to pathogens, and the secretion of antifungal substances, such as volatile organic compounds (VOCs), β-glucanases, and killer toxins [13][14][15][16][17][18]. The production of killer toxins, exhibiting antimicrobial activity, is a relatively common phenomenon in yeast. Strains with killer phenotypes produce killer toxins and are immune to their toxin. However, they may be susceptible to killer toxins produced by other killer yeasts [19,20]. The protein structure, specific receptors, and mode of action of killer toxins may differ significantly among yeast genera, species, and even strains. Most killer toxins So far, it has been established that β-glucans in the cell wall of sensitive microorganisms are the primary receptors for killer toxins of D. hansenii and W. anomalus [6,14,16,[40][41][42][43]. However, little is known about the further effect of the killer toxins on the plasma membrane of sensitive microorganisms and potential defence systems. In this research, we attempt to establish whether plasma membrane components, such as PDR transporters, play a role in resistance to D. hansenii and W. anomalus killer toxins. In comparative studies, we analyzed the sensitivity of S. cerevisiae wild type strains and their isogenic mutants with increased or reduced pleiotropic drug resistance to D. hansenii and W. anomalus killer toxins. To the best of our knowledge, this is the first report considering the correlation between killer toxins activity and the function of plasma membrane transporters which contributes to the overall research on killer toxins and killer yeast antagonistic activity in biocontrol of pathogenic fungi.

Results
PDR pumps are responsible for the translocation of small molecule compounds. However, in the present study, only crude killer toxins preparations exhibited the inhibitory effect on tested S. cerevisiae strains, whereas the small-molecule fractions (below 10 kDa), obtained from killer yeast cultures did not. Also, the post-culture fluids obtained from the culture of non-killer control strain D. hansenii CLIB 545 did not inhibit the growth of S. cerevisiae, thus confirming that the differences in the susceptibility of tested S. cerevisiae strains are toxin-dependent since CLIB 545 strain was grown under the same conditions as killer strains but the post culture medium did not contain killer toxins (Figures 2-4).
The growth of S. cerevisiae wild type strain YPH 500 and its isogenic mutants with deleted transporter genes PDR5, SNQ2, and PDR5 together with SNQ2 was inhibited by crude killer toxins of all tested killer yeast strains (AII4b, KI2a, MI1a, CBS 767, and BS91 at varying levels ( Figure 2)). In general, the double deletant (∆pdr5∆snq2) was more susceptible to killer toxins than the parental YPH 500 strain or the single knock-outs ∆pdr5 and ∆snq2 (Figure 2). Deletion of PDR5 resulted in higher sensitivity to killer toxins of D.

Results
PDR pumps are responsible for the translocation of small molecule compounds. However, in the present study, only crude killer toxins preparations exhibited the inhibitory effect on tested S. cerevisiae strains, whereas the small-molecule fractions (below 10 kDa), obtained from killer yeast cultures did not. Also, the post-culture fluids obtained from the culture of non-killer control strain D. hansenii CLIB 545 did not inhibit the growth of S. cerevisiae, thus confirming that the differences in the susceptibility of tested S. cerevisiae strains are toxin-dependent since CLIB 545 strain was grown under the same conditions as killer strains but the post culture medium did not contain killer toxins (Figures 2-4).  Figure 2). Interestingly, ∆snq2 mutant was less sensitive to KI2a, MI1a, and CBS 767 in relation to isogenic wild type strain YPH 500, whereas in the presence of W. anomalus BS91 killer toxin, this strain was more sensitive than the wild type isogenic YPH 500 strain ( Figure 2). Double deletion of genes encoding two main transcription factors PDR1 and PDR3 in S. cerevisiae FY1679-28 mutant resulted in significantly higher sensitivity to crude killer toxins of all tested killer D. hansenii yeast strains: AII4b, KI2a, MI1a, CBS 767, as compared to the wild type strain ( Figure 3). Interestingly, the growth of wild-type FY 1679-28 strain and its isogenic ∆pdr5 mutant strain was not inhibited in the presence of W. anomalus BS9 killer toxin preparation, whereas double knock-out mutant (∆pdr1∆pdr3) appeared to be sensitive ( Figure 3).   The statistical significance of the differences between the means was tested separately for each killer toxin preparation (BS91, KI2a, MI1a, AII4b, and CBS 767) and marked with different letters. The study was performed with a Kruskal-Wallis one-way analysis of variance by ranks at p = 0.05. Means followed by different letters (a,b) are significantly different, while means marked by ab are not significantly different from groups a and b.

Discussion
PDR5, SNQ2, and YOR1 genes, under prevailing control of PDR1 and PDR3 transcription factors, encode Pdr5p, Snq2p, and Yor1p efflux pumps in the cell membrane ( Figure 1). These pumps are one of the most characterized transporters in the ABC family   The statistical significance of the differences between the means was tested separately for each killer toxin preparation (BS91, KI2a, MI1a, AII4b, and CBS 767) and marked with different letters. The study was performed with a Kruskal-Wallis one-way analysis of variance by ranks at p = 0.05. Means followed by different letters (a,b) are significantly different, while means marked by ab are not significantly different from groups a and b.

Discussion
PDR5, SNQ2, and YOR1 genes, under prevailing control of PDR1 and PDR3 transcription factors, encode Pdr5p, Snq2p, and Yor1p efflux pumps in the cell membrane ( Figure 1). These pumps are one of the most characterized transporters in the ABC family The statistical significance of the differences between the means was tested separately for each killer toxin preparation (BS91, KI2a, MI1a, AII4b, and CBS 767) and marked with different letters. The study was performed with a Kruskal-Wallis one-way analysis of variance by ranks at p = 0.05. Means followed by different letters (a,b) are significantly different, while means marked by ab are not significantly different from groups a and b. The growth of S. cerevisiae wild type strain YPH 500 and its isogenic mutants with deleted transporter genes PDR5, SNQ2, and PDR5 together with SNQ2 was inhibited by crude killer toxins of all tested killer yeast strains (AII4b, KI2a, MI1a, CBS 767, and BS91 at varying levels ( Figure 2)). In general, the double deletant (∆pdr5∆snq2) was more susceptible to killer toxins than the parental YPH 500 strain or the single knock-outs ∆pdr5 and ∆snq2 (Figure 2). Deletion of PDR5 resulted in higher sensitivity to killer toxins of D. hansenii AII4b, KI2a, and MI1a strains, and not to the toxin of D. hansenii CBS 767 and W. anomalus BS91 (Figure 2). Interestingly, ∆snq2 mutant was less sensitive to KI2a, MI1a, and CBS 767 in relation to isogenic wild type strain YPH 500, whereas in the presence of W. anomalus BS91 killer toxin, this strain was more sensitive than the wild type isogenic YPH 500 strain (Figure 2).
Double deletion of genes encoding two main transcription factors PDR1 and PDR3 in S. cerevisiae FY1679-28 mutant resulted in significantly higher sensitivity to crude killer toxins of all tested killer D. hansenii yeast strains: AII4b, KI2a, MI1a, CBS 767, as compared to the wild type strain (Figure 3). Interestingly, the growth of wild-type FY 1679-28 strain and its isogenic ∆pdr5 mutant strain was not inhibited in the presence of W. anomalus BS9 killer toxin preparation, whereas double knock-out mutant (∆pdr1∆pdr3) appeared to be sensitive ( Figure 3).
Unexpectedly, the sensitivity of S. cerevisiae hyper-resistant pdr1-3 mutant to all toxins of D. hansenii, was significantly higher, as compared to that of the wild type strain YALA-B1 ( Figure 4). However, deletion of the PDR5 transporter gene in a mutant strain with retained pdr1-3 gain-of-function mutation resulted in a phenotype that was less sensitive to all killer toxins of D. hansenii, still more sensitive than wild-type strain YALA-B1. On the contrary, the pdr1-3 mutant strain was completely resistant to the killer toxin of W. anomalus BS91. Also, there were no growth inhibition zones observed for pdr1-3∆pdr5 isogenic mutant in the presence of BS91 toxin, whereas the growth of the control wild type YALA-B1 strain was inhibited by BS91 killer toxin (Figure 4).

Discussion
PDR5, SNQ2, and YOR1 genes, under prevailing control of PDR1 and PDR3 transcription factors, encode Pdr5p, Snq2p, and Yor1p efflux pumps in the cell membrane ( Figure 1). These pumps are one of the most characterized transporters in the ABC family and exhibit extremely broad, sometimes overlapping, substrate specificity. The joint and compensatory action of these efflux pumps is responsible for the extrusion of various undesired small molecules and peptides from cells [25,44]. In turn, killer toxins derived from D. hansenii and W. anomalus are proteins of the molecular mass exceeding 10 kDa, which exhibit antagonistic activity towards sensitive microorganisms upon binding to their cell wall [1,12,14,17,41]. So far, there have been no reports that PDR transporters could transport large proteins across the cell membrane. In our experiments, only a large molecule fraction containing killer toxin exhibited an inhibitory effect on tested S. cerevisiae, while the small molecular fraction containing peptides and other small molecules did not. Therefore, it can be inferred that PDR pumps may play a role as plasma membrane components that are involved in resistance to killer toxins. However, these pumps may not necessarily work as a killer toxin extrusion system, and may instead modulate their activity.
In the present study, the role of PDR pumps in the killer toxin effect was more pronounced for W. anomalus BS91 killer toxin, as compared to D. hansenii killer toxins ( Figure 5).
ponents that are involved in resistance to killer toxins. However, these pumps may not necessarily work as a killer toxin extrusion system, and may instead modulate their activity.
In the present study, the role of PDR pumps in the killer toxin effect was more pronounced for W. anomalus BS91 killer toxin, as compared to D. hansenii killer toxins ( Figure  5). The activity of W. anomalus BS91 killer toxin towards S. cerevisiae mutants with a single deletion of SNQ2 gene as well as deletion of both SNQ2 and PDR5 genes was enhanced as compared to the wild type strain YPH 500 and its single PDR5 deletant suggesting the possible role of Snq2p in resistance to BS91 killer toxin. In turn, in S. cerevisiae pdr1-3 hyper-resistant mutant, exhibiting enhanced expression of PDR5, SNQ2 and YOR1, the activity of BS91 killer toxin was extinguished. The deletion of the PDR5 gene in the pdr1-3 genetic background did not result in regaining of susceptibility of cells to BS91 killer toxin suggesting that the Pdr5p transporter may not be involved in the inhibition of BS91 killer toxin activity, since the lack of this protein did not change the resistance of pdr1-3 mutant to BS91 killer toxin. This was also shown in FY 1679-28 wild type strain and its PDR5 single deletant. None of them was sensitive to BS91 toxin. A double deletant in PDR1 and PDR3 was significantly more sensitive to BS91 killer toxin than the isogenic wild type S. cerevisiae strain, again indicating the involvement of Snq2p and possibly Yor1p transporters in resistance to BS91 toxin ( Figure 5). The activity of W. anomalus BS91 killer toxin towards S. cerevisiae mutants with a single deletion of SNQ2 gene as well as deletion of both SNQ2 and PDR5 genes was enhanced as compared to the wild type strain YPH 500 and its single PDR5 deletant suggesting the possible role of Snq2p in resistance to BS91 killer toxin. In turn, in S. cerevisiae pdr1-3 hyper-resistant mutant, exhibiting enhanced expression of PDR5, SNQ2 and YOR1, the activity of BS91 killer toxin was extinguished. The deletion of the PDR5 gene in the pdr1-3 genetic background did not result in regaining of susceptibility of cells to BS91 killer toxin suggesting that the Pdr5p transporter may not be involved in the inhibition of BS91 killer toxin activity, since the lack of this protein did not change the resistance of pdr1-3 mutant to BS91 killer toxin. This was also shown in FY 1679-28 wild type strain and its PDR5 single deletant. None of them was sensitive to BS91 toxin. A double deletant in PDR1 and PDR3 was significantly more sensitive to BS91 killer toxin than the isogenic wild type S. cerevisiae strain, again indicating the involvement of Snq2p and possibly Yor1p transporters in resistance to BS91 toxin ( Figure 5).
In turn, killer toxins of D. hansenii exhibited varying activity towards S. cerevisiae PDR mutants and the tendency in their performance differed from that of W. anomalus BS91 killer toxin. As was expected, double deletant in both major transcription factors PDR1 and PDR3 was significantly more susceptible to all killer toxins of D. hansenii. Also, the single deletion of PDR5 and the double deletion of PDR5 and SNQ2 affected the activity of most D. hansenii killer toxins as compared to the wild type YPH 500 strain. Whereas, single deletion of SNQ2 did not, suggesting the role of Pdr5p in resistance to D. hansenii killer toxins in susceptible cells. However, the gain-of-function pdr1-3 mutant, bearing an armour of PDR efflux pumps including Pdr5p, Snq2p, and Yor1p in the plasma membrane, was unexpectedly more sensitive to all D. hansenii killer toxins than the wild type isogenic strain. Single deletion of the PDR5 gene in the pdr1-3 background resulted in a phenotype that was more sensitive to D. hansenii killer toxins than the wild type strain. However, it was still significantly less sensitive than the hyper-resistant pdr1-3 mutant ( Figure 5). It could be therefore inferred that the D. hansenii killer toxin may be either recognized and affected by Pdr5p transporter in the cell membrane with a regular density of PDR transporters, and not with an excess of efflux pumps as it is in the pdr1-3 mutant, or that killer toxin activity may be influenced by other than tested efflux pumps in the cell membrane (possibly under the control of YAP1 or YRR1 transcription factors). Also, according to Kolaczkowska et al. [44], there exists a compensatory activation of one multidrug transporter upon disruption of genes encoding different multidrug transporters, which is accompanied by increased efflux of substrates specific for the activated ones. According to their studies, upon the disruption of the PDR5 gene, the resistance to Yor1p-and Snq2p-specific substrates increases. Since D. hansenii killer toxin activity did not seem to depend on Snq2p in YPH 500 background, the role of Yor1p should be considered. Yor1p is, among others, the most important pump in the extrusion of Aureobasidin A-a cyclic depsipeptide produced by killer yeast Aurobasidium pullulans, toxic to several yeast species and filamentous fungi, such as Botrytis cinerea, Monilinia sp. and Penicillium sp. [45,46], towards which D. hansenii also exhibits antagonistic activity [16,47,48].

Conclusions
The present research, with the use of S. cerevisiae mutants in terms of PDR transporters responsible for the extrusion of substances of low molecular mass from cells, demonstrated that the killer effect may depend on the presence and density of PDR pumps in the plasma membrane of sensitive to killer toxin microorganisms, including Pdr5p and Snq2p ( Figure 5). It also pointed out a new difference in killer phenomenon between killer strains of D. hansenii and W. anomalus species, where W. anomalus killer toxin was inhibited by the PDR transporters under the control of PDR1 and PDR3 transcription factors, whereas D. hansenii killer toxins activity was dependent on different transporters in plasma membrane than BS91 killer toxin, which could not be unambiguously defined. In previous studies, it was already proven that killer toxins derived from D. hansenii and W. anomalus exhibited varying antagonistic activity against several fungal plant pathogens through different mechanisms of action [15,16,18]. A comparative study of killer yeast strains of D. hansenii and W. anomalus species revealed their common feature, which is the recognition of β-glucans in the cell wall of the attacked microorganisms. It also pointed out a differentiating feature, which is the formation and release of antimicrobial volatile organic compounds, that was observed only in W. anomalus, and not in D. hansenii killer strains in combating pathogenic filamentous fungi [9,16]. This biocontrol mechanism of W. anomalus might be in turn attributed to the production of ethyl acetate [13,49] and 2-phenylethanol [50], from which the latter one may be ejected from S. cerevisiae cells by Pdr12p efflux pump, whose PDR12 gene stays under the control of WAR1 transcription factor [51,52]. In relation to biocontrol of phytopathogenic fungi, that contain MDR pumps in plasma membrane responsible for acquired resistance to chemical treatment of crops with azoles, strobilurins and succinate dehydrogenase inhibitors [53][54][55], this may elucidate the varying susceptibility of fungal pathogenic strains to killer toxins of W. anomalus and D. hansenii. As was shown in this study the inhibition of killer toxin produced by these yeast species relies on different PDR transporters. These findings shed new light on the perception of the mechanisms of the killer effect, understood as the antagonistic effect of killer yeasts against sensitive, often pathogenic yeast and microscopic fungi. Our novel findings constitute the foundations for research on the involvement of pleiotropic drug resistance efflux pumps of different families in the antagonistic effect of killer yeasts on sensitive microorganisms.

Yeast Strains
Killer strains of D. hansenii KI2a, MI1a and AII4b were isolated from blue-veined Rokpol cheese [56] and belong to the Culture Collection of the Department of Biotechnology and Food Microbiology at the Wroclaw University of Environmental and Life Sciences (Wroclaw, Poland). Another killer strain of D. hansenii, CBS 767 was obtained from CBS-KNAW Culture Collection. W. anomalus BS91 killer yeast was isolated from naturally fermented olive brine [57] and belongs to the Di3A Culture Collection (University of Catania, Italy). Killer phenotypes of these strains are presented in Table 1. D. hansenii CLIB 545, obtained from Centre International de Ressources Microbiennes (CIRM-Levures, http://www.inra.fr/cirmlevures, 16 August 2021), was used as a control non-killer, nonresistant to killer toxins strain (Table 1). S. cerevisiae yeast strains were chosen based on the characteristics of drug resistance associated with PDR transporters in their plasma membrane and their genotypes are presented in Table 2. YPH500 is an isogenic wild type strain to YKKB-13 strain with a single PDR5 knock-out, YYM5 strain with a single SNQ2 knock-out, and YYM3-with a double PDR5 and SNQ2 knock-outs. FY-1679-28C is an isogenic wild type strain to hyper-sensitive FY ∆pdr1∆pdr3 double knock-out strain and a single PDR5 knock-out strain FY-WT/∆pdr5-2. YALA-B1 is an isogenic strain to YALA-G4 hyper-resistant mutant (pdr1-3) and YZGA 278 hyper-resistant mutant (pdr1-3) with a single PDR5 knock-out. S. cerevisiae YALA-B1 K − R −D,−W "K"-Killer; "R"-Resistant; "D"-D. hansenii killer toxins; "W"-W. anomalus killer toxin. K + R + -produces a killer toxin and is resistant to killer toxins; K + R − -produces a killer toxin and is sensitive to killer toxins produced by other killer strains; K − R − -does not produce a killer toxin and is sensitive to killer toxins.

Killer Toxin Production
Yeasts were cultured on yeast peptone dextrose agar plates (YPDA; yeast extract, 10 g; peptone, 10 g; dextrose, 20 g; agar, 20 g per litre of distilled water) at 25 • C for 48 h. Then, the single colony biomass of each yeast strain was inoculated to 50 mL of YPD broth of pH adjusted to 4.5 with McIlvaine buffer and incubated either at 14 • C for 48 h (D. hansenii strains) or 25 • C (W. anomalus) for 24 h on a rotary shaker at 160 rpm to provide favourable conditions for killer toxin production. Yeast cells from liquid cultures were pelleted at 8000 g for 10 min. The supernatant was filtrated using a sterile syringe filter with a 0.22 µm pore size (Merck Millipore, Burlington, MA, USA, Milex-GP 33 mm, PES membrane). The obtained post-culture media served as crude toxin preparations. In addition, the smallmolecule fractions (<10 kDa) were separated from the post-culture media using a filter with an appropriate cut-off point (Merck Millipore, Amicon ® Ultra-15, 10 kDa). Both crude killer toxin preparations and small-molecule fractions were used in the diffusion assay.

Killer Toxin Diffusion Assay
The killer activity of D. hansenii and W. anomalus toxins was tested in agar diffusion bioassay according to the method described byŻarowska [12]. Twenty-five mL of YPDA medium, supplemented with 0.03% (w/v) methylene blue and 4% (w/v) of NaCl, was adjusted to pH 4.5 with McIlvaine buffer and used in the assay for the evaluation of killer toxin activity. Fifty µL of each tested sample (i) crude killer toxin preparation, (ii) smallmolecule fraction (<10 kDa), was added into wells, sterilely cut in YPDA plates containing S. cerevisiae cells at a final concentration of 5 × 10 5 per mL of the assay medium ( Table 2). The zones of S. cerevisiae growth inhibition [cm] around wells were measured after 48 h of incubation at 20 • C. The experiment was repeated three times with three replicates for each tested sample.

Statistical Analysis
In all repeated experiments, the arithmetic means were calculated and analyzed using Kruskal-Wallis one-way analysis of variance by ranks at p = 0.05. Post-culture media derived from the cultures of D. hansenii and W. anomalus were treated as individual preparations, the activity of which was tested against S. cerevisiae strains. The significant differences in growth inhibition zones between each S. cerevisiae wild type strain (YPH 500; FY-1679-28C and YALA-B1) and its isogenic PDR-mutants (YKKB-13, YYM 5, YMM3; FY∆pdr1∆pdr3, FY-WT/∆pdr5-2 and YALA-G4, YALA 278) were analyzed separately for each killer toxin preparation to demonstrate the role of PDR pumps on the toxic effect of all tested killer toxin preparations. The strength of the killer effect was not compared among killer strains. The results of mean killer activity of each preparation in relation to the tested S. cerevisiae strains, significantly different from each other, were grouped into homogenous groups a and b, while the results without statistically significant differences were assigned to the ab group, which was not significantly different from groups a or b. Data from experiments were analyzed using Statistica package software (Version 12; Statsoft Inc., Tulsa, OK, USA).
Funding: This work was funded by research grant POIG.01.03.01-02-080/12 "The use of yeast Y. lipolytica and D. hansenii, their enzymes and toxins for application in industry and agriculture" co-financed by the European Union from the European Regional Development Found. The APC/BPC is financed/co-financed by Wroclaw University of Environmental and Life Sciences.
Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: The dataset used and/or analyzed during the current study are publicly available.