Introduction

Fermentation of malt extract by yeast into flavoursome beer, and the use of yeast in baking, are two of the oldest biotechnologies known to humans. Brewing of a typical modern lager consists of a primary fermentation resulting in ethanol, carbon dioxide and much of the aroma, and a secondary fermentation (lagering) to obtain the final taste (Hough et al. 1982; Young and Lewis 1995). Two biochemical and physiological properties of the yeast strains employed in brewing are essential to prevent degradation of the fermentation product, ethanol: their ability to suppress respiratory metabolism in the presence of sugar (the “glucose-repression” regulatory circuit; Gancedo 1998) and the ability of the cells grow anaerobically, i.e., in the absence of oxygen (Viser et al. 1990). The latter capability becomes crucial towards the end of the fermentation, when the sugar pool has been depleted and the level of oxygen is very low but the yeast must remain metabolically active (Hough et al. 1982; Young and Lewis 1995). Similarly, baking yeast is also able to grow in the absence of oxygen. In nature, apparently only two groups of yeasts, those belonging to the Saccharomyces and Brettanomyces/Dekkera clades, have developed such an ability (Subik et al. 1974; Viser et al. 1990). The progenitor of the modern ascomycetous yeast species was most probably completely dependent on the presence of oxygen (Møller et al. 2001). However, upon diversification, a few of the yeast lineages progressively reduced their dependence on oxygen by remodelling several biochemical pathways. One of the crucial requirements for anaerobiosis in yeast is that the fourth step of de novo pyrimidine biosynthesis, the conversion of dihydroorotate to orotate catalysed by dihydroorotate dehydrogenase (DHODase) (Denis-Duphil 1989) is independent of the respiratory chain (Nagy et al. 1992). While aerobic yeasts such as Schizosaccharomyces pombe and Pichia stipitis, posses a respiratory chain dependent mitochondrial (mt) DHODase, the Saccharomyces cerevisiae enzyme is cytosolic (cytDHODase) and transfers the reducing reaction equivalents to fumarate (Nagy et al. 1992). Indeed, transfer of the S. cerevisiae cytDHOD gene into P. stipitis by genetic transformation in the laboratory transformed this yeast species into a facultative anaerobe (Shi and Jeffries 1998). In this report, we argue that the ability of certain yeasts to grow under anaerobic conditions was promoted by the horizontal transfer of a bacterial gene coding for a DHODase which is independent of respiratory activity.

Materials and methods

Southern analysis

Total DNA from each of the yeast species studied here was digested with different restriction enzymes, fractionated on a 1% agarose gel, blotted onto a membrane and hybridised to a probe originating from the S. cerevisiae (plasmid P548) or S. pombe (plasmid P544) DHODase gene. Standard hybridisation and membrane washing procedures (Sambrook et al. 1989) at low temperature (55°C) and high salt concentration were used.

Cloning and analysis of DHOD genes

The S. pombe -like DHOD gene was cloned from an S. kluyveri genomic library (contained in the vector FL44SII; Bonneaud et al. 1991) by complementation of the ura1 mutation carried by the S. cerevisiae Y583 strain (MAT a ura1). The complementing plasmid, P629, was rescued into Escherichia coli, and the insert was shown to contain an ORF coding for a putative DHODase, referred to here as SkmtDHODase (GenBank Accession No. AF452108). The 5′ end of the ORF encodes a putative mitochondrial import signal. A partial sequence of a S. cerevisiae DHOD -like gene was obtained in the course of the S. kluyveri genome sequencing project (Cliften et al. 2001). The full-length ORF encoding SkcytDHODase (Accession No. AF452109) was rescued into the S. cerevisiae strain Y583, by PCR amplification and subsequent transformation of the sub-cloned gene containing the native upstream and downstream sequences (plasmid P652). The transformants were shown to have become prototrophic for pyrimidines. Both S. kluyveri sequences were aligned with other known DHODase sequences using the program Clustal X 1.81 (Jeanmougin et al. 1998), and a phylogenetic analysis was performed by the method of (Van de Peer and De Wachter 1994).

Gene disruption

The dominant drug resistance cassettes based on KanMX3, which confer resistance to geneticin, were constructed by a three-step PCR. In this way longer overhangs, originating from the 5′ and 3′ portions of the DHOD gene(s) were added to the KanMX3 gene. PCR fragments, used for disruption of the two S. kluyveri DHODase genes, were introduced into different S. kluyveri strains, Y090 (MATα thr) and Y091 (MAT a his x) (Møller et al. 2001) by electroporation as described previously (Gojkovic et al. 2000). The cells were then plated, and transformants were selected, on YPD medium containing geneticin (75 mg/l). Putative transformants were checked for the presence of gene disruptions by PCR.

Generation of double mutants

The S. kluyveri strains Y785 (MATα thr cytdhod::KanMX3) and Y981 (MAT a his x mtdhod::KanMX3) were crossed. The diploid prototrophs obtained were then sporulated on plates containing 1% potassium acetate, 0.1% yeast extract and 0.05% glucose. A mixture of sporulated and non-sporulated cells was then resuspended in 5 mM DTT solution containing Zymolyase (0.2 mg/ml), and incubated for 30 min at 25°C. Sterile glass beads were added, and the suspension was shaken for another 30 min. An equal volume of 0.1% Triton X-100 was added, and the suspension was vortexed vigorously for 2 min. The mixture so obtained, consisting of diploid cells, separated and un-separated spores, was then plated on YPD and the resulting colonies were screened for pyrimidine auxotrophs. Several of these were then analyzed by PCR for the presence of deletions within their DHOD genes and for uracil auxotrophy.

DHODase activity assays

The DHOD genes of S. kluyveri, S. cerevisiae and S. pombe were subcloned into the E. coli expression vector P343 (Gojkovic et al. 2001), resulting in the plasmids P628 (coding for SkcytDHODase), P611 (SkmtDHODase), P609 (SccytDHODase) and P610 (SpmtDHODase). The C-terminally His-tagged recombinant enzymes expressed from these plasmids were purified as described by Gojkovic et al. (2001) and their specific activities were determined using 1 mM DHO and electron acceptors at the indicated concentrations. The assays were performed as described previously (Ullrich et al. 2001; Knecht et al. 1996) at 30°C in TRIS buffer [50 mM TRIS-HCl pH 8, 150 mM KCl] containing 0.1% (v/v) Triton X-100. Reduction of the electron acceptors was measured at the wavelengths indicated and using the following extinction coefficients: DCIP, ε600=18800 M−1 cm−1; ferricyanide, ε420=1020 M−1 cm−1; NAD+, ε340=6200 M−1 cm−1; or alternatively tested for the formation of orotate (ε280=7500 M−1 cm−1) using fumarate or O2 as electron acceptor. When quinones were used as electron acceptors, orotate formation was measured at the isobestic wavelength of the quinone: QD, ε300=2950 M−1 cm−1; Q10, ε300=2950 M−1 cm−1; Q6, ε293=4700 M−1 cm−1); Q0, ε287=5680 M−1 cm−1; and plastoquinone, ε280=7500 M−1 cm−1. For each enzyme the specific activity with ferricyanide was set to 100% and the data shown are the means (±SD) of three measurements.

Cell fractionation

S. kluyveri Y090 was first grown under aerobic and anaerobic conditions. Thereafter the yeast cells were resuspended in ice-cold isolation medium (50 mM TRIS-HCl pH 7.5, 0.15 M NaCl, 5 mM EDTA, 0.3 M sucrose) and disrupted by agitation with glass beads using a FastPrep machine FP 120 (Bio101 ThermoSavant). The suspension was centrifuged at 500× g for 10 min to remove unbroken cells and debris. The supernatant was centrifuged at 20,000× g for 10 min. This step was repeated once and the pellets from both centrifugations were dissolved in isolation medium and combined to give the pellet fraction. Hexokinase (Bergmeyer 1974), alcohol dehydrogenase (Tong et al. 2001) and succinate dehydrogenase (Knecht et al. 1996)activities were assayed as described previously. DHODase activity, using DHO as substrate and ferricyanide as electron acceptor, was measured as described above.

Oxygen consumption

Oxygen consumption by the pellet fraction was determined (Knecht et al. 1996)using a Clark-Type electrode (Oxylab 1.81 b System, Biolytik) in the presence of 2 mM DHO or 0.3 mM NADH. This fraction contained predominantly mitochondrial enzymes (as demonstrated by the presence of NADH oxidation activity) and was isolated from fresh aerobically grown S. kluyveri Y090 cells. Following the addition of 1 mM KCN, an inhibitor of the respiratory chain, the residual oxygen consumption activities was measured as a control.

Yeast growth in bioreactors

The S. cerevisiae host strains Y827 and Y829 (ura1) transformed with either the S. kluyveri cytDHOD or S. kluyveri mtDHOD gene, respectively, were cultivated in well controlled bioreactors at 28°C and pH 5.0 under aerobic and anaerobic conditions in defined minimal medium containing 20 g/l glucose (Møller et al. 2001, 2002). For anaerobic growth, the bioreactors were continuously flushed with nitrogen (containing less than 3 ppm O2) at a flow rate of 0.5 l per l of medium per min, beginning 24 h prior to inoculation. The dissolved oxygen concentration was monitored with an autoclavable O2-sensor (Mettler-Toledo) and was below the detection level during anaerobic cultivation and above 30% of saturation during aerobic cultivation. Growth was monitored by measuring optical density at 600 nm.

Results

S. kluyveri has two DHODases

The phylogenetic relationship among Saccharomyces/Kluyveromyces yeasts is now well established (Kurtzman and Robnett 2003). S. bayanus is one of the closest relatives of S. cerevisiae, while S. castellii diverged before the S. cerevisiae and S. bayanus lineages separated. All three species can grow anaerobically, and in general they exhibit very similar physiological properties (Vaughan-Martini and Martini 1998). S. kluyveri is the least related and can also grow anaerobically, but it exhibits only weak glucose repression of the respiratory pathway (Vaughan-Martini and Martini 1998; Møller et al. 2001, 2002). The type strains of these species were analysed by DNA hybridisation for the presence of different types of DHOD genes. The S. cerevisiae URA1 gene, coding for cytDHODase, and the S. pombe gene coding for mtDHODase were used as probes in our hybridisation experiments (Fig. 1a). The genome sequences of Candida albicans and S. pombe, two aerobic yeast species, were also analysed for the presence of DHODases using the BLAST network services at the National Center for Biotechnology Information. Surprisingly, while S. bayanus and S. castellii gave a signal with the cytDHOD gene in the hybridisation experiments, S. kluyveri gave a signal with the mtDHOD gene as well as with the cytDHOD gene probe (Fig. 1a). C. albicans and S. pombe contained only the mtDHOD gene. The two S. kluyveri genes, coding for SkcytDHODase and SkmtDHODase, respectively, were subsequently cloned and the predicted sequences of their products were compared with other DHODase sequences. While SkmtDHODase was grouped with the other eukaryotic enzymes, SkcytDHODase was classified within the sub-family which includes the S. cerevisiae enzyme and DHODase A from the prokaryote Lactococcus lactis (Fig. 1b).

Fig. 1a, b
figure 1

Distribution of different types of DHODases in the type strains of different yeast species ( a) and the phylogenetic relationship between bacterial and eukaryotic DHODases ( b). a There are two types of DHODase, an S. cerevisiae- and an S. pombe-like enzyme, in different yeast species, as deduced from DNA hybridisation patterns. b Phylogenetic tree generated by TreeCon 1.3b (Van de Peer and De Wachter 1994) showing the relationship among different DHODases (based on an alignment of DHODase protein sequences). Note that the two DHODases in S. kluyveri belong to two different enzyme subfamilies. The Accession Nos. of the amino acid sequences and the bootstrap values for nodes (100 bootstrap replications) are shown. The numbers in parentheses (asterisks) indicate the length of the putative mitochondrial leader sequence, which was not used in the alignment

Full size image

The double DHOD mutant of S. kluyveri is a pyrimidine auxotroph

Each of the DHOD genes from S. kluyveri was able to complement the ura1 defect in S. cerevisiae, when placed under the control of its own promoter, and therefore granted uracil prototrophy on the strain Y583 (MAT a ura1). The two Sk DHOD genes and their function were further studied in S. kluyveri by constructing the corresponding deletion mutants. The DNA fragments used for transformation carried at each end ~500 bp of the S. kluyveri DHOD gene sequences. Two S. kluyveri strains, Y785 (MATα thr cytdhod::KanMX3) and Y981 (MAT a his x mtdhod::KanMX3), characterized by deletion in the cytDHOD and mtDHOD genes, respectively, were generated. Both of these strains were uracil prototrophs in the presence of oxygen. Upon mating of Y785 and Y981 the resulting diploids were sporulated and double dhod mutants were isolated. These strains, Y989, Y990, Y991 and Y992, were all uracil auxotrophs. This finding confirms that S. kluyveri has two functional DHODase enzymes. However, as shown below, these two enzymes differ in their biochemical properties.

DHODases and their electron acceptors

The DHOD genes were cloned from S. kluyveri, S. cerevisiae and S. pombe, expressed as recombinant enzymes (SkcytDHODase, SkmtDHODase, SccytDHODase and SpmtDHODase) and their specific activities were determined (Table 1). All four enzymes could use several native and two artificial electron acceptors, ferrycyanide (FeCy) and 2,6-dichlorophenolindophenol (DCIP). FeCy was the best electron acceptor. In the presence of this artificial electron acceptor (at 1 mM) and varying concentrations of DHO, the kinetic parameters measured for SkmtDHODase were as follows: Km=0.204 (±0.006) mM, Vmax=27.78 (±2.72) U/mg, and kcat=23.3/s. The corresponding values for SkcytDHODase were 0.235 (±0.026) mM, 142.46 (±3.85) and 88.1, respectively. A high turnover for all enzymes was observed also in the presence of the soluble quinone Q0. Quinone Q6, a likely physiological acceptor whose availability in vivo is dependent on the integrity of the respiratory chain, was a good electron acceptor only for SkmtDHODase and SpmtDHODase (Table 1). Fumarate, also a likely physiological acceptor, was a good acceptor for SkcytDHODase and SccytDHODase, but not for the other two enzymes, when the high background activity is taken into account (Table 1). NAD+ was not a good acceptor for any of the enzymes (Table 1). However, the presence of atmospheric oxygen also promoted a very low DHODase activity (Table 1), suggesting that molecular oxygen can act as a poor acceptor of electrons. Nevertheless, it is apparent that, while SkmtDHODase is likely to be dependent on electron acceptors which require the integrity of the respiratory chain, SkcytDHODase can catalyse the transfer of electrons to a larger variety of acceptors, including fumarate. The presence of fumarate in the yeast cell is independent of a functional respiratory chain, since it can be generated in the cytoplasm via the glyoxylate cycle (Duntze et al. 1969; Gancedo and Serrano 1989).

Table 1 Relative specific activities of recombinant yeast DHODases with different electron acceptors
Full size table

The regulatory role of oxygen

In the following we tried to elucidate the dependence of the two S. kluyveri DHODase activities on aerobic versus anaerobic growth conditions. A pyrimidine prototrophic strain, Y090 (MATα thr), was cultivated in minimal medium (plus threonine) under aerobic and anaerobic conditions (Møller et al. 2001). After fractionation of the cell lysate, the DHODase activities were determined using DHO as substrate and FeCy as electron acceptor. These activities were compared with the activities of marker enzymes; hexokinase (Bergmeyer 1974) and alcohol dehydrogenase (Tong et al. 2001) as cytosolic markers, and succinate dehydrogenase (Knecht et al. 1996) as a mitochondrial marker. The pellet fraction, containing mitochondrial markers, exhibited the same low DHODase activity regardless of whether it was prepared from aerobically or anaerobically grown cells. However, the supernatant fraction containing the cytoplasmic markers exhibited a much higher DHODase activity when it had been obtained from cells grown under anaerobic conditions (Fig. 2). Lack of oxygen enhanced the activity of SkcytDHODase. Following the addition of succinate (a substrate for complex II of the respiratory chain), the consumption of oxygen and DHO was measured in the two Y090 subfractions isolated from aerobically grown cells. It was detected only in the pellet fraction. DHODase activity co-localised with NADH oxidation by succinate and like NADH oxidation, was inhibited by the addition of KCN, an inhibitor of complex IV of the respiratory chain (Fig. 3). Thus, the SkmtDHOHase activity appears to be functionally linked to the respiratory chain and the presence of oxygen.

Fig. 2
figure 2

Cytoplasmic S. kluyveri DHODase activity is up-regulated under anaerobic growth conditions. The activities of four different enzymes in the supernatant (Sup.; containing predominantly the cytoplasmic enzymes) and pellet (Pel.; obtained by centifugation at 20,000×g, contains predominantly the mitochondrial enzymes) fractions of S. kluyveri Y090 grown under aerobic (+oxygen) and anaerobic (−oxygen) conditions were determined. Two cytoplasmic enzymes, hexokinase and alcohol dehydrogenase, are up-regulated under anaerobic conditions, while succinate dehydrogenase, a mitochondrial enzyme, is up-regulated only in the presence of oxygen

Full size image
Fig. 3
figure 3

The mitochondrial DHODase from S. kluyveri is coupled to the respiratory chain. Oxygen consumption by the pellet fraction (containing predominantly mitochondrial enzymes) in the presence of two substrates (2 mM DHO and 0.3 mM NADH) is shown. The residual activities observed upon addition of 1 mM KCN, an inhibitor of the respiratory chain, are indicated as a percentage of the consumption in the absence of inhibitor. One unit (U) is equivalent to the consumption of one μmol of O2 per min

Full size image

Growth in the absence of oxygen

Both S. kluyveri genes could rescue pyrimidine prototrophy on the S. cerevisiae ura1 (Y583) strain, but their ability to contribute to the anaerobic phenotype differed. Each of the genes, under the control of its native promoter, was introduced into the uracil auxotroph Y583, and the resulting prototrophic strains, Y827 (expressing SkcytDHODase) and Y829 (expressing SkmtDHODase), were cultivated in well controlled bioreactors under aerobic and anaerobic conditions in defined minimal medium (Møller et al. 2001, 2002). Under aerobic conditions both strains exhibited typical diauxic growth with similar specific growth rates during exponential growth on glucose (Fig. 4). In anaerobic conditions Y827 grew at an exponential rate although with a slightly lower specific growth rate. However, Y829 did not show significant growth (Fig. 4). Therefore, only SkcytDHODase confers on the transformed Y583 yeast cells the ability to grow in minimal medium under anaerobic conditions.

Fig. 4
figure 4

The ability of yeast to grow under anaerobic conditions in minimal medium is dependent on the presence of the cytoplasmic DHODase enzyme. S. cerevisiae ura1 strains, carrying either the SkcytDHOD (Y827) or the SkmtDHOD (Y829) gene, were cultivated under aerobic and anaerobic conditions in defined minimal medium and growth was monitored by measuting absorbance at 600 nm

Full size image

Discussion

While aerobic yeasts and other eukaryotes contain a mitochondrial DHODase (Bjornberg et al. 1997) (Fig. 1b), the facultative anaerobe S. cerevisiae contains a cytoplasmic enzyme (Nagy et al. 1992). On the other hand, S. kluyveri , which separated from the S. cerevisiae lineage ~100 million years ago, represents an intermediary evolutionary state, as it has both enzymes (Fig. 1a), which presumably utilize different kinds of electron acceptors (Table 1). In the S. pombe cell the mitochondrial DHODase with its quinone-type electron acceptor requires the action of the respiratory chain for its activity, while in S. cerevisiae the cytoplasmic DHODase and one of its electron acceptors, fumarate, do not (Nagy et al. 1992). Similarly, in S. kluyveri SkmtDHODase is likely to be linked to quinone and the respiratory chain (Fig. 3), while SkcytDHODase uses cytoplasmic fumarate (Table 1). In addition, the SkcytDHODase activity is elevated under non-respiratory conditions (Fig. 2). While both SkcytDHODase and SkmtDHODase can provide pyrimidine prototrophy, only SkcytDHODase permits anaerobic growth in minimal medium (Fig. 4). Therefore, the acquisition of the cytoplasmic-type enzyme must have represented an important step in the evolution of decreased dependence on oxygen among Saccharomyces yeasts. Presumably, the progenitor of the S. cerevisiae and S. kluyveri lineages initially had only the mitochondrial-type enzyme. A bacterial gene coding for a DHODase enzyme similar to the modern Lactococcus lactis A enzyme (Fig. 1b) is likely to have been adopted 100–150 million years ago, just before speciation of Saccharomyces yeasts took place (Fig. 5). Such horizontal transfers from bacteria to eukaryotes have previously been postulated also for other organisms (Andersson et al. 2001). Upon separation of the S. cerevisiae and S. kluyveri lineages the original eukaryote enzyme has apparently been lost in the S. cerevisiae lineage (Fig. 5). Another, although less likely explanation, is that a bacterial DHODase was already present in the eukaryote progenitor, but was afterwards lost in the majority of lineages. It is worth noting that the three DHODase subfamilies (Fig. 1b) originate from a progenitor enzyme, the gene for which was duplicated before the eukaryote and bacterial lineages had separated. However, while all three subfamilies can be found in bacteria, in general eukaryotes have only one, mitochondrial-type, DHODase. The least probable explanation, which is contradicted by our phylogenetic analysis (Fig. 1b), is parallel evolution of fumarate dependent enzymes in some bacteria and yeast. The appearance of oxygen independent DHODase in yeast could coincide with, or precede, the gross duplication of the yeast genome (Wolfe and Shields 1997; Langkjaer et al. 2003). In addition, during evolution further gradual “improvements”, such as the development of effective glucose repression, have followed. Millions of years later, these novel properties were successfully exploited by man for use in baking and for the fermentation of sugars into tasty beverages.

Fig. 5
figure 5

A model depicting the origin of the yeast DHODase enzymes. The rough phylogenetic relationship among the four different yeast species is adopted from Kurtzman and Robnett (2003). The progenitor of the Saccharomyces lineages had only a mitochondrial-like DHODase. This situation can still be found in the majority of yeasts, including S. pombe and C. albicans. However, ~100–200 million years ago, a cytoplasmic-like DHODase was recruited, and this “intermediate” stage with two DHODases is still conserved in S. kluyveri. Upon separation of the S. kluyveri and S. cerevisiae lineages, the latter presumably lost the gene for the mitochondrial-like DHODase enzyme

Full size image