Genetic and Physiological Characterization of Fructose-1,6-Bisphosphate Aldolase and Glyceraldehyde-3-Phosphate Dehydrogenase in the Crabtree-Negative Yeast Kluyveromyces lactis

The milk yeast Kluyveromyces lactis degrades glucose through glycolysis and the pentose phosphate pathway and follows a mainly respiratory metabolism. Here, we investigated the role of two reactions which are required for the final steps of glucose degradation from both pathways, as well as for gluconeogenesis, namely fructose-1,6-bisphosphate aldolase (FBA) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). In silico analyses identified one gene encoding the former (KlFBA1), and three genes encoding isoforms of the latter (KlTDH1, KlTDH2, KlGDP1). Phenotypic analyses were performed by deleting the genes from the haploid K. lactis genome. While Klfba1 deletions lacked detectable FBA activity, they still grew poorly on glucose. To investigate the in vivo importance of the GAPDH isoforms, different mutant combinations were analyzed for their growth behavior and enzymatic activity. KlTdh2 represented the major glycolytic GAPDH isoform, as its lack caused a slower growth on glucose. Cells lacking both KlTdh1 and KlTdh2 failed to grow on glucose but were still able to use ethanol as sole carbon sources, indicating that KlGdp1 is sufficient to promote gluconeogenesis. Life-cell fluorescence microscopy revealed that KlTdh2 accumulated in the nucleus upon exposure to oxidative stress, suggesting a moonlighting function of this isoform in the regulation of gene expression. Heterologous complementation of the Klfba1 deletion by the human ALDOA gene renders K. lactis a promising host for heterologous expression of human disease alleles and/or a screening system for specific drugs.


Introduction
Glycolysis has been studied for more than a century, with a crucial impact on biochemistry, ever since the observation of cell-free fermentation in extracts from the wine, beer, and baker's yeast Saccharomyces cerevisiae [1]. For this yeast, glycolysis is essential, as a lack of enzyme activity in most steps leads to growth inhibition on glucose as a sole carbon source [2][3][4]; reviewed in [5]. S. cerevisiae is highly specialized in alcoholic fermentation, to which it owes its biotechnological importance. To this end, the reactions of pyruvate decarboxylase and alcohol dehydrogenase convert the glycolytic end-product pyruvate into carbon dioxide and ethanol [6,7].
In humans, glycolysis is also of utmost importance in central carbohydrate metabolism, as exemplified by so-called glycogen-storage diseases associated with impaired enzyme activities, e.g., phosphofructokinase (glycogen-storage disease VII; Tarui's disease [8]) or fructose-1,6-bisphosphate aldolase (glycogen-storage disease XII [9]). Moreover, cancer cells generally increase the flux through glycolysis, designated as the so-called Warburg which may also contain GAPDH [36]. Moreover, a physical interaction of the aldolase with RNA polymerase II was suggested to modulate the transcriptional activity towards tRNA genes [37]. In K. lactis, FBA has been suggested to participate in the regulation of hypoxic metabolism [38]. A plethora of moonlighting functions have been assigned to FBA in mammalian cells, many related to its interaction with cytoskeleton components (reviewed in [39]). A broad overview of FBA functions apart from its role in central carbohydrate metabolism throughout all biological kingdoms has recently been published [40].
Regarding GAPDH, a number of non-glycolytic functions were discovered for extracellular enzymes in different yeast species. In S. cerevisiae, all three isoforms were detected in the cell wall, suggesting an extracellular function [41]. Moreover, the three isoforms were identified as the source of antifungal peptides claimed to ensure the predominance of S. cerevisiae in the later stages of wine fermentations [42]. GAPDH from the opportunistic pathogen Candida albicans was found in the cell wall as an immunoactive protein, antibodies which can prevent systemic infections caused by this yeast [43]. A homolog of the enzyme was also found in the cell wall of the industrial yeast Kluyveromyces marxianus and to mediate flocculation [44]. Other moonlighting functions are suggested by the association of both FBA and GAPDH with mitochondria in high-throughput studies of S. cerevisiae [45].
Here, we investigated deletion mutants in KlFBA1 and the three genes-encoding GAPDH isoforms in K. lactis. We showed that KlFba1 is not essential for growth on glucose, but GAPDH activity, mediated by the combined action of KlTdh1 and KlTdh2, is. Promoter fusions with a lacZ reporter gene demonstrated that the gene encoding the third GAPDH isoform is repressed by glucose, but expressed on medium containing ethanol as a sole carbon source. The mutants obtained herein provide the basis for heterologous expression of human genes encoding isoforms and disease-related alleles, as exemplified by the human ALDOA gene.

Deletions of KlFBA1 or KlTDH2 Affect Growth on Glucose
A comparison in the yeast gene order browser (http://ygob.ucd.ie/ accessed on 10 December 2021) revealed syntenic homologs of both FBA1 (encoding fructose-1,6-bisphosphate aldolase) and TDH2 (encoding a major isoform of glyceraldehyde-3-phosphate dehydrogenase) in the genomes of Saccharomyces cerevisiae and Kluyveromyces lactis. Consequently, these genes were designated as KlFBA1 (systematic name KLLA0E07569g) and KlTDH2 (systematic name KLLA0A11858g). Due to the whole-genome duplication in S. cerevisiae, the latter is also syntenic to ScTDH3. A further homolog of the dehydrogenase gene annotated in the K. lactis genome does not appear to be syntenic with S. cerevisiae or the common ancestor, and was designated as KlTDH1 (systematic name KLLA0F20988g). The deduced protein sequences of ScFba1 and KlFba1 showed 75% identical amino acid residues. KlTdh1 and KlTdh2 were only 80% identical to ScTdh1, but yielded higher values amongst each other and compared to the other two S. cerevisiae isoforms (KlTdh1/KlTdh2 = 83%, KlTdh1/ScTdh2 = 82% identity, KlTdh1/ScTdh3 = 81%; KlTdh2/ScTdh2 = 85%, KlTdh2/ScTdh3 = 86%). On the other hand, a gene encoding the previously described NADP + -dependent isoform of GAPDH found exclusively in K. lactis, KlGDP1 (systematic name KLLA0F09141g [34]), displayed only approximately 56% identity with the NAD +dependent isoforms of both yeasts.
To investigate the in vivo function of the encoded enzymes, we proceeded by deleting the four genes from the K. lactis genome, as described in Materials and Methods and schematically represented in Figure 1a. Tetrad analyses performed with heterozygous diploids carrying a wild type and a deleted allele of each gene showed that deletions of KlFBA1 and KlTDH2 still grow on glucose, albeit yielding smaller colonies than wild-type segregants (Figure 1a). The growth defect was much more pronounced in Klfba1 deletions than in segregants lacking KlTDH2. In contrast, neither deletion of KlTDH1 nor that of KlGDP1 affected growth of the respective segregants on glucose (Figure 1a), indicating that the encoded isoforms are dispensable for glucose utilization under standard growth conditions. Figure 1. Tetrad analyses of deletion mutants. (a) Strains with heterozygous deletions of the genes indicated in the schematic drawings were sporulated and subjected to tetrad analyses on rich medium (YEPD). After micromanipulation, plates were incubated for 2 days (images with smaller colonies) or 3 days (Kltdh2/KlTDH2) at 30 • C. Growth was documented by scanning the plates and adjustment of brightness using Corel Photo-Paint. Nine representative tetrads from a total of at least eighteen separated from each diploid are shown. Light-blue circles mark the segregants carrying the respective deletions, as judged from replica plating onto selective marker plates. Note that in the case of the heterozygous diploid for the Klgdp1 deletion, another deletion, Kltrp1::kanMX, carried the same marker cassette. Therefore, only segregants unequivocally carrying the deletion, i.e., those that were both prototrophic for tryptophan and resistant to G418, are highlighted. Strains employed were KHO483 and KHO505 (top row), and KHO455 and KHO461 (bottom row). (b) Tetrad analysis of a heterozygous diploid (KHO476) obtained from crossing a Kltdh2::kanMX deletion with a Klgdp1::SpHIS5 deletion. Segregants carrying a single Kltdh2 deletion are highlighted with a light-blue circle. Double deletions are indicated by yellow circles. The plate was incubated for 3 days at 30 • C after micromanipulation and all 18 tetrads separated are shown. (c) Tetrad analysis of a diploid strain heterozygous for both the Kltdh1 and the Klgdp1 deletion, but homozygous for the Kltdh2 deletion (KHO506). After micromanipulation, tetrads were allowed to grow on rich medium with 2% glycerol and 2% ethanol as carbon sources (YEPGE) for 3 days at 30 • C. Segregants were picked onto a master plate of the same medium and allowed to grow overnight. Numbers refer to the tetrad analyzed and signs from top to bottom designate segregants A to D as indicated. Only 6 representative tetrads from a total of 14 tetrads with 4 viable segregants are shown. Growth was documented after replica plating onto the indicated media and incubation over night at 30 • C. The relevant genotype for each segregant can be deduced from the table given below. Segregants carrying Kltdh1 deletions in combination with Kltdh2 and/or Klgdp1 were assigned from their ability to grow on synthetic medium lacking histidine and with 2% glycerol (w/v) as a sole carbon source. Klgdp1 deletions were assigned by PCR using flanking oligonucleotides as primers. SC = synthetic complete medium with either 2% glucose (w/v) or 2% ethanol (v/v).
In order to determine the importance of the different GAPDH isoforms in vivo, deletions were combined by crossing and the resulting diploids were sporulated and subjected to tetrad analyses. While deletions of KlGDP1 combined with Kltdh2 did not produce any additive phenotype with respect to growth on glucose (Figure 1b,c), no viable segregants of the type Kltdh1 Kltdh2 could be obtained on standard rich medium with 2% glucose (Figure 1c). This suggested that GAPDH activity is required for growth on glucose and that KlGdp1 does not contribute significantly to glucose consumption in vivo. However, Kltdh1 Kltdh2 double deletions yielded viable segregants on media with non-fermentable carbon sources (2% ethanol plus 2% glycerol; note that glycerol alone suffices as a sole carbon source to promote growth of the double deletions), with similar colony sizes to those of wild-type and single-deletion segregants. Upon replica plating, segregants with the double deletion failed to grow on synthetic media containing 2% glucose but retained the ability to grow on plates with 2% ethanol as a sole carbon source (Figure 1c). In contrast, triple Kltdh1 Kltdh2 Klgdp1 deletion mutants lacked growth on either glucose or ethanol. Thus, KlTdh1 and KlTdh2 catalyze the reactions from and towards glyceraldehyde-3-phosphate depending on the provided carbon source, while in their absence KlGdp1 activity is sufficient for gluconeogenesis from ethanol.

KlFBA1
Encodes the Sole Fructose-1,6-Bisphosphate Aldolase in K. lactis and KlTdh2 Is the Major Isoform of Glyceraldehyde-3-Phosphate Dehydrogenase To further substantiate the conclusions drawn from the genetic analyses, specific enzyme activities were determined in crude extracts from different strains obtained in the tetrad analyses. As shown in Table 1, Klfba1 deletions did not retain detectable FBA activity, while wild-type segregants yielded specific activity of approximately 300 mU/mg protein. This confirms that KlFBA1 is the only encoding gene and that the enzyme is important, but not essential, for glucose utilization. Kltdh2∆ 1 Specific activities were routinely measured in crude extracts from strains grown in rich medium with 2% glucose (w/v) as a carbon source (YEPD). In the case of Kltdh1 Kltdh2 double deletions, which cannot grow on glucose, the hexose in the growth medium was substituted by 2% ethanol (v/v) as a carbon source. GAPDH activities were determined in the reverse direction to glycolysis as described in materials and methods. Activities given represent the mean values of at least three independent determinations from each crude extract with the standard deviations (±).
On the other hand, Kltdh2 deletions produced specific GAPDH activities of approximately 25% compared to those of the wild-type strains after growth on rich medium with glucose (Table 1). Kltdh1 deletions retained more than 50% of wild-type GAPDH activity, confirming that KlTdh2 is the major isoform under standard growth conditions, but KlTdh1 contributes significantly. No GAPDH activity could be detected in Kltdh1 Kltdh2 double deletions when grown on glycerol or transferred to media with either glucose or ethanol as a sole carbon source (Table 1, and data not shown). Besides lacking any detectable NAD + -dependent GAPDH activities, substitution of NADH for NADPH in the reverse enzyme assays did not yield detectable activities in the Kltdh1 Kltdh2 double deletions or the wild type, even when cells were transferred to ethanol medium 6 h prior to the preparation of crude extracts. This indicates that KlGdp1 activity, although it must be present to allow for growth of the double mutants on ethanol, is extremely low compared to those conferred by KlTdh1 and KlTdh2.
In order to determine if the specific enzyme activities correlate with the expression of the encoding genes, and if the latter depends on the carbon source, we proceeded by fusing the respective gene promoters to the bacterial lacZ reporter gene and determined specific ß-galactosidase activities. To this end, the plasmid-encoded reporter constructs were introduced into a recipient K. lactis strain lacking its endogenous Kllac4 gene (Table 2). KlTDH1p KlTDH2p 1 Strains were grown after inoculation from a preculture for 6 h in logarithmic phase on selective synthetic medium lacking uracil with the indicated carbon sources. Specific ß-galactosidase activities were measured twice in crude extracts from two independent transformants, each. Standard deviations are given (±). KHO209-8A was used as a recipient strain for all constructs. It lacks intrinsic ß-galactosidase activity due to a deletion of the KlLAC4 gene.
Expression from the KlFBA1 promoter was highest when cells were grown on selective media with glucose as a sole carbon source, with approximately twice the activities driven by the KlTDH1 and KlTDH2 promoters, and a very low expression from the promoter of KlGDP1. Transfer for six hours to medium with ethanol as a sole carbon source led to an almost eight-fold increase in reporter activity for cells harboring the KlGDP1p-lacZ fusion (Table 2). Interestingly, both KlTDH1-lacZ and KlTDH2p-lacZ constructs also displayed a drastic increase in ß-galactosidase activities on ethanol compared to those on glucose (approximately three-and two-fold, respectively), while those of the KlFBA1-lacZ construct slightly decreased (Table 2). This indicates that expression of all genes encoding GAPDH isoforms is subject to different degrees of glucose repression, and that KlTdh1 may play a more important role under gluconeogenic than fermentative growth conditions.

Heterologous Genes Complement the Growth Defects of a Klfba1 Deficiency
In order to establish yeast as a host for heterologous expression of genes encoding FBA enzymes, we introduced the homolog from S. cerevisiae under the control of its native promoter in the Klfba1 deletion and vice versa KlFBA1 in the Scfba1 deletion. Both yeast mutants were also employed to introduce the gene encoding a human FBA isoform (AL-DOA) under the control of the ScPFK2 promoter. In tetrad analyses on rich media with glucose, glycerol or ethanol, the Scfba1 deletion was lethal in our hands. Both the native ScFBA1 gene and the K. lactis homolog restored full growth to the respective segregants in Scfba1, when introduced on a CEN/ARS plasmid, presumed to reside in 1-2 copies per cell, as judged by colony sizes (Figure 2a). However, S. cerevisiae deletions carrying the human ALDOA gene did not produce viable progeny when introduced on the low-copy vector but yielded smaller colonies than the wild type when expressed from the ScPFK2 promoter from the high-copy number plasmid (Figure 2a; lower panel).
Since a K. lactis strain lacking its endogenous KlFBA1 gene can still grow on rich medium with glucose as a carbon source, the two yeast genes under the control of their native promoters were introduced directly into the haploid recipient strain. Transformants were able to complement the growth retardation of the deletion when streaked out for single colonies on synthetic medium with glucose as a carbon source (Figure 2b; note that the same shuttle vectors were employed as for S. cerevisiae, but that they are present in higher copy numbers in K. lactis due to their pKD1 origin of replication). Accordingly, the genes from both yeast species conferred increased FBA activity to the recipient strain as compared to a wild-type control (Figure 2c). Importantly, the human ALDOA gene introduced on the same vector and expressed from the ScPFK2 promoter also conferred FBA activity, albeit reaching only approximately 18% of that of the wild-type control. This serves as a proof of principle that K. lactis is a suitable host for the expression of aldolase genes from various organisms and different modes of catalysis (see Section 3). , or used as a recipient for the plasmids with the genes indicated below each plate (all allowed to germinate on YEPD). Plates were incubated for 2 days at 30 • C, with the exception of the one carrying the ALDOA gene on a multicopy vector (pJJH3069), which was incubated for 3 days. Note that growth of more than two segregants per tetrad indicates complementation of the lethal Scfba1∆ phenotype. Replica plating onto synthetic media lacking the respective auxotrophic requirements confirmed that segregants carrying the deletion were only able to grow if they carried a plasmid with a complementing gene. (b) The haploid K. lactis strain KHO477-1D was used to introduce the plasmids indicated and transformants were selected on synthetic medium lacking uracil, with 2% glucose (w/v) as a carbon source. After preparation of a master plate, cells were streaked out for single colonies on the same medium and incubated at 30 • C for 4 days. Note that the pronounced growth retardation of the recipient strain was complemented by both the native FBA1 gene and its homolog from S. cerevisiae. (c) Specific FBA activities were measured in crude extracts from two independent strains, each carrying the plasmids indicated. Cells were grown on synthetic medium with glucose lacking uracil for selection of plasmid maintenance. Values given are the mean of two biological replicas, obtained from at least three technical replicas for each crude extract, with the respective standard deviations (±), which is also represented by the error bars. Strains with a Klfba1 deletion fail to grow in liquid synthetic medium with glucose and were therefore grown in YEPD, yielding specific activities below detectable levels, as listed in Table 2.

KlTdh2, but Not KlFba1, Accumulates in the Nucleus upon Oxidative Stress
Both FBA and GAPDH enzymes were reported to partially associate with mitochondria under oxidative stress in the Saccharomyces Genome Database (SGD) database (https:// www.yeastgenome.org/; accessed on 21 December 2021). Therefore, functional C-terminal GFP fusions were constructed for KlFba1 and KlTdh2 encoded at their native loci, and employed in colocalization studies with a mitochondrial marker, KlIdp1-mCherry, in lifecell fluorescence microscopy. As is evident from Figure 3, both enzymes showed a cytosolic distribution under standard growth conditions. (a) Distribution of KlFba1-GFP in a strain carrying a genomic fusion construct (KlIDP1-mCherry) producing a mitochondrial marker (mit-mCherry; upper two panels) or a marker for the nuclear pore complex (NUP60-mCherry, nuc-mCherry; lower two panels). Hydrogen peroxide was added at a concentration of 4.4 mM where indicated and images were taken after 20 min of incubation. Distribution of all three marker proteins remained the same after a prolonged incubation for 90 min with hydrogen peroxide. (b) Distribution of KlTdh2-GFP in strain backgrounds as in (a), i.e., mitochondrial KlIdp1-mCherry (upper-two panels) and nuclear Nup60-mCherry; lower-two panels), without and with 4.4 mM hydrogen peroxide. The latter images were taken after 15 min of incubation with the stressor, in which case the nuclear accumulation of KlTdh2-GFP was still observed in samples taken after 1 h of incubation. Visual inspection of more than 100 cells without and with hydrogen peroxide revealed that nuclear accumulation occurred in more than 75% of the cells exposed to oxidative stress. Strains employed were: KHO474-1A, KHO465-6C, KHO472-1D, KHO466-1A. Size bars correspond to 1 µm for all images in the same panel.
Application of hydrogen peroxide did not provoke a notable accumulation of either enzyme at mitochondria. However, for KlTdh2-GFP a distinct accumulation in a round structure was observed under oxidative stress. Therefore, we employed strains also carrying a genomic NUP60-mCherry fusion labeled in a pore protein that resides in the nuclear membrane. The corresponding images confirmed that KlTdh2-GFP accumulated in the nucleus upon oxidative stress (Figure 3b). In contrast, KlFba1-GFP did not display a distinct accumulation in any subcellular compartment.
In silico analyses suggested that only one gene encoding an FBA was present in the K. lactis genome (KlFBA1), whereas two homologs encoding putative NAD + -dependent GAPDH isozymes were detected (KlTDH1 and KlTDH2), in addition to a gene encoding a previously reported isoform capable of using either NAD + or NADP + (KlGDP1; [34]). Deletion of the entire open reading frames of all four genes combined with determinations of specific enzyme activities in crude extracts confirmed that indeed, all isoforms present in K. lactis were covered by this study. FBA activity in Klfba1 deletions was below detectable levels, paralleled by a severe growth defect of such mutants on media with glucose as a sole carbon source. We attribute the slow growth to the fact that the PPP feeds into glycolysis at the levels of fructose-6-phosphate (i.e., prior to the FBA reaction) and glyceraldehyde-3phosphate (i.e., prior to GAPDH; Figure 4). Some of the carbon from glucose may thus still be directly channeled into the lower part of glycolysis by the PPP in Klfba1 mutants, while a major part needs to be recycled through the phosphoglucose isomerase reaction and the oxidative part of the PPP.
On the other hand, GAPDH activity is required for sugar degradation by both major pathways, and strains lacking this activity consequently did not grow on glucose. This phenotype also demonstrates that the NAD + /NADP + -dependent isoform encoded by KlGDP1 does not contribute under these conditions, most likely because the gene is not expressed, as observed in the reporter fusions. Consistent with previous reports obtained by Northern blots [34], KlGDP1 transcription is subject to glucose repression. Interestingly, our expression studies revealed that KlGDP1 is expressed on ethanol medium, albeit only at approximately 10% of the levels observed for the two KlTDH gene promoters. This explains why Kltdh1 Kltdh2 double mutants still grow on ethanol as a sole carbon source, whereas triple Kltdh1 Kltdh2 Klgdp1 deletions do not. Clearly, KlGdp1 activity alone allows for sufficient gluconeogenesis to support growth in the absence of the two major GAPDH isoforms. It should be noted that we were unable to measure any enzyme activity for KlGdp1 in crude extracts, even when substituting NADH in the reverse enzyme assay for NADPH. We conclude that either the in vitro test conditions were not adequate for the detection of this GAPDH activity, which must be present in vivo, or the enzyme is not sufficiently stable in crude extracts.
Unexpectedly, we observed that the two KlTDH genes also appear to be subject to some glucose repression, although to a lower extent than KlGDP1. Thus, reporter ß-galactosidase activites from promoter fusions were increased by a factor of 3 (KlTDH1p) or two-fold (KlTDH2p) when cells were grown on ethanol instead of glucose as a sole carbon source. This indicates an increased requirement for the encoded isoforms under gluconeogenic growth conditions, possibly to allow for sufficient input of metabolites into the PPP, which runs at a higher capacity in K. lactis than in S. cerevisiae (reviewed in [18]). Replenishment from glyceraldehyde-3-phosphate and fructose-6-phosphate would also avoid excessive production of NADPH by the oxidative part of the PPP, if it was fed only from glucose-6phosphate generated as the final product of gluconeogenesis while growing on ethanol (Figure 4).
How do Kltdh1 Kltdh2 Klgdp1 triple mutants generate sufficient energy to grow on glycerol as a sole carbon source? The mitochondrial glycerol-3-phosphate dehydrogenase, KlGut2, which converts glycerol-3-phosphate to dihydroxyacetone phosphate, employs FAD instead of NAD + as a cofactor, generating reducing power directly fed into the respiratory chain (Figure 4). Glucose-6-phosphate and ATP can therefore be generated from glycerol, which represses the glyoxylate cycle and the consumption of ethanol in K. lactis [47,48]. The GAPDH isoforms in S. cerevisiae were shown to form heteromeric complexes, which may well occur with the K. lactis enzymes, too [42]. Concerning the importance of the different isoforms in vivo, KlTdh2 seems to be more reminiscent of ScTdh3, as a lack of the latter also displayed a slight growth defect on glucose media [42]. The authors also proposed another role for GAPDH in yeast sphingolipid metabolism, whose conservation in K. lactis remains to be determined.
With regard to the link of carbohydrate metabolism with regulation of gene expression it is rather intriguing that KlTdh2 can accumulate in the nucleus under conditions of oxidative stress. This indicates a possible moonlighting function, which could play a role in transcriptional regulation on top of its function in glycolysis and gluconeogenesis. It will thus be interesting to pursue this hypothesis in future transcriptome and proteome analyses, taking advantage of the deletion mutants generated herein. Another role of GAPDH was observed in the biotechnological application of S. cerevisiae for the production of alcoholic beverages [49]. All three GAPDH isoforms were shown to serve as a source for antimicrobial peptides, generated by the metacaspase Mca1 (= Yca1), and were proposed to enhance competition with other fungi and bacteria in wine fermentations. Interestingly, K. lactis disposes of a syntenic gene to ScMCA1 and the two core sequences of the antimicrobial peptides identified in S. cerevisiae, SWYDNEYGYS and FRVPTVDVSVVD [49], are conserved in both KlTdh1 and KlTdh2. Whether K. lactis shares this antimicrobial capacity and if it could be exploited in its biotechnological applications would be another interesting line of future investigation.
With regard to FBA, the severe growth defect and the lack of detectable enzymatic activity in the Klfba1 deletion predestine K. lactis as a heterologous expression host. The haploid deletion strain, although growing poorly on glucose, can be directly propagated on standard medium and used for the introduction of the homologs from other organisms. Restoration of fast growth can then be directly related to a high complementation capacity, whereas FBA assays in crude extracts can still detect less potent isoforms. This was demonstrated by the expression of ScFBA1 and the human ALDOA gene-which encodes the major isoform in muscle tissue and erythrocytes [50]-in the current study. In this context it is important to note that aldolases come in two fundamentally different varieties: whereas bacteria and fungi use a class II type aldolase which requires a divalent metal (Zn 2+ in S. cerevisiae and presumably also in KlFBA) as a cofactor, insects and vertebrates harbor class I-type enzymes that employ a Schiff-base intermediate in their catalysis [51]. It is thus not surprising that the activity produced by the human clone in the K. lactis expression system accounts for less than 20% of the native enzyme with a different mode of catalysis. Nevertheless, it demonstrates that yeasts are a suitable host for expression of genes encoding such fundamentally different enzymes. In Scfba1 deletions, strong overexpression of the clone was required for complementation. This is consistent with previous reports on the heterologous expression of an FBA gene from flies (Drosophila melanogaster), where the strong glycolytic gene promoter PGI1p and a multicopy plasmid were employed [52]. In the latter work the authors worked with a Scfba1 deletion able to grow on rich medium with acetate as a carbon source. In the genetic background employed herein, the heterozygous deletion did not produce any viable progeny in tetrad analyses on this or the various other tested media with non-fermentable carbon sources. Although the growth deficiency in S. cerevisiae may be overcome by buffering a rich medium with ethanol as a carbon source to pH 5.5 [36], such adaptations are not required for the Klfba1 strains. We conclude that with the ease of handling and the respiratory metabolism being more similar to that of human cells, K. lactis could serve as a better host to study the function of FBA isoforms and alleles associated with human diseases. Since aldolase has been shown to reside in the cell walls of different fungal pathogens, such as Candida albicans and related species [53][54][55] and Cryptococcus neoformans [56], which contribute to virulence by binding and activating human plasminogen, the enzyme can be a target for drugs and vaccines (reviewed in [40]). With the results presented herein, K. lactis could thus be a valuable host with a GRAS (generally regarded as safe) status for high-level production of the respective proteins.
We conclude that both FBA and the different isoforms of GAPDH serve essential functions in the degradation of glucose by the respiratory yeast K. lactis. GAPDH activity is further required for gluconeogenesis, for which the isoform encoded by the glucoserepressed KlGDP1 gene suffices. The nuclear localization of KlTdh2 under oxidative stress provides the first hint to an important moonlighting function of this isoform, which is 1 Strains KHO46-2A, KHO69-8C, and KHO70 have been described previously [57]. All other strains were constructed in this work.
Yeast cell culture and genetic techniques followed standard procedures described for S. cerevisiae [60]. Rich medium (YEPD) contained 1% yeast extract, 2% Bacto peptone (Difco Laboratories Inc., Detroit, MI, USA), and 2% glucose (all w/v). Synthetic media were prepared with Difco yeast nitrogen base with ammonium sulfate as described in [60], with the addition of amino acids and bases using a mixture provided by MP Biomedicals (Eschwege, Germany; CSM-His-Leu-Trp-Ura) supplemented as required for selection of plasmids or deletion markers, and 2% glucose (w/v) as carbon source (SCD). Two percent glycerol (w/v) and/or 2% ethanol (v/v) were used as alternative carbon sources in rich or synthetic media as indicated, and 100 mg/L of G418 was added for selection of the kanMX marker. Hydrogen peroxide was added at the concentrations indicated to induce oxidative stress.
Crossing of K. lactis strains was performed on malt agar plates (5% malt extract, 3% agar, w/v), incubated for 1-2 days at 30 • C. Diploids were selected by complementation of auxotrophic markers after replica plating onto appropriate synthetic media. They were grown for 8-10 h at 30 • C in 3 mL of YEPD, collected by centrifugation and dropped onto potassium acetate plates (1% potassium acetate, 3% agar, w/v) for sporulation. After 1-2 days of incubation at 30 • C, tetrad analyses were performed on YEPD plates using zymolyase 20T for digestion of the ascus walls and a Singer Instruments micromanipulator as described in [61]. Strains heterozygous for the Kltdh1 Kltdh2 double and the Kltdh1 Kltdh2 Klgdp1 deletions were separated and allowed to germinate on rich medium containing glycerol and ethanol (YEPGE). Plates were incubated for 2-3 days at 30 • C and colony formation was documented by scanning. Brightness and contrast were adjusted for entire plates using Corel Photo Paint.
For manipulations in E. coli, strain DH5α was employed with media as described previously [62].

Construction of Plasmids, Gene Fusions and Deletions
For cloning and expression of glycolytic genes in yeast, the triple K. lactis/ S. cerevisiae/E. coli shuttle vector pCXs22 [57] was generally employed. It confers ampicillin resistance to E. coli and carries ScURA3 as a selection marker for both yeasts. A CEN/ARS sequence ensures its low-copy number in S. cerevisiae, and high-copy numbers in K. lactis are conferred by the pKD1 origin of replication. For heterologous gene expression, pJJH2075 was constructed by inserting an optimized ScPFK2 promoter, described previously [63] as an EcoRI/BamHI fragment, into pCXs22. YEplac181 [64] with a 2 µm origin of replication and an ScLEU2 marker was used for high-copy number expression of the human ALDOA gene in S. cerevisiae. For promoter fusions with the bacterial lacZ reporter gene, plasmid pJJH2031 was constructed. It carries the entire lacZ open reading frame cloned as a PCR-generated BamHI/EcoRI fragment into pCXs18, a low-copy number vector for K. lactis from the same series as pCXs22 described above [57,65]. Promoters of genes to be tested were generated by PCR using oligonucleotides with appropriate cloning sites and ligated into the respective sites of pJJH2031 preceding the translational start codon of lacZ. All oligonucleotides were custom-made and purchased from Biolegio (Nijmegen, The Netherlands). A concise list of the plasmids derived from these vectors and constructed in this work is given in Table 4.
For assembly and alignments of DNA and protein sequences, the Clone Manager 9 program (Scientific and Educational Software, Denver, USA) was employed. The integrated ClustalW function was used to determine the percentage of sequence identities.
For gene deletions, the entire open reading frames were substituted by selective marker cassettes obtained from pUG6 (kanMX) or pUG27 (SpHIS5) described in [59]. For this purpose, cassettes were amplified with appropriate oligonucleotide pairs, containing at least 40 nucleotides corresponding to the target sequence at their 5 -end and 20-22 nucleotides of homology to the marker cassettes. PCR products were introduced for homologous recombination into either the diploid recipient strain KHO70 or the haploid strain KHO69-8C, both carrying deletions in the KU80 gene to reduce the frequency of non-homologous recombination. If this direct strategy of one-step gene deletion failed, the PCR products were employed for in vivo recombination with plasmids carrying the respective genes and longer flanking sequences in S. cerevisiae, plasmids were isolated from transformants grown on selective media, amplified in E. coli, and fragments with longer homologous regions were introduced into the K. lactis strains selecting for the deletion marker. Deletions were verified by PCR with flanking oligonucleotides and obtained in the haploid state for KHO70 by sporulation and tetrad analysis. All strains were routinely back-crossed to KU80 wild-type strains, sporulated and again subjected to tetrad analyses to avoid genomic instability. All deletions were finally verified by three independent PCRs to ensure their correct integration. Complete sequences of the respective loci are available upon request. KlTDH1 promoter cloned as a HindIII/BamHI fragment Genomic fusions to genes encoding fluorescence tags followed the same basic strategy as applied for gene deletions, i.e., one-step replacements using PCR products obtained with appropriate oligonucleotide pairs providing homology to the target sequences. Fluorescent marker genes were fused in frame to the 3 -end of the respective coding sequences by eliminating the translational stop codon of the target gene. The fusion cassettes were equipped with a selectable marker, i.e., kanMX for GFP (pJJH1619) and SkHIS3 for mCherry (pJJH1525), with the template vectors described in [66]. Strains with different fluorophores were combined as required for colocalizations by crossing and tetrad analyses.

Fluorescence Microscopy
The setup used for fluorescence microscopy consisted of a Zeiss Axioplan 2 (Carl Zeiss, Jena, Germany) equipped with a 100x alpha-Plan Fluor objective (NA 1.45) and differentialinterference contrast. Sample handling and image processing were described in detail in [67]. The setup was controlled by the Metamorph v6.2 program (Universal Imaging Corporation, Downingtown, PA, USA). Brightfield images were acquired as single planes using DIC. Images were scaled using Metamorphs scale image command (Metamorph Offline Version 7.8.0.0, 64-bit). Deconvolution of the fluorescence images was performed by Huygens Remote Manager v3.5. The processed images were overlaid using Metamorphs overlay images command.

Preparation of Crude Extracts and Enzymatic Determination
For preparation of crude extracts, cells were incubated in liquid media overnight with shaking at 30 • C in 5-10 mL of precultures. They were inoculated in 10-20 mL of the media indicated at an OD600 of approximately 1, and incubated for another 6 h. Cells were harvested by centrifugation and washed twice with 2.5 mL ice-cold potassium phosphate buffer (50 mM, pH7.0). Next, 0.5 g of glass beads with a diameter of 0.25-0.5 mm (Carl Roth GmbH, Karlsruhe, Germany) were added together with 0.5 mL of the phosphate buffer and vigorously shaken for 10 min at 4 • C. After addition of another 0.5 mL of ice-cold phosphate buffer, the supernatant was transferred to 1.5 mL Eppendorf tubes kept on ice, and cell debris was pelleted by centrifugation at 13,000× g and 4 • C for 10 min. The supernatant was removed into fresh Eppendorf tubes and kept on ice as a crude extract for determination of enzymatic activities. Total protein content was determined by the Micro-Biuret method [68] using bovine serum albumin as a standard. Specific ß-galactosidase activities expressed in nmoles ortho-nitrophenol-ß-D-glucopyranoside (ONPG) hydrolyzed per min and mg protein at 30 • C (mU) were determined, as described previously [69]. FBA and GAPDH activities were determined by coupling the conversion of substrates to the oxidation of NADH by ancillary reactions, with the addition of salts and cofactors as listed in [70]. Briefly, fructose-6-phosphate was used at 2 mM as a substrate for FBA, and the fructose-1,6bisphosphate produced was converted by the ancillary enzymes triosephosphate isomerase and glycerol-3-phosphate dehydrogenase, consuming NADH. For GAPDH assays, the reverse reaction to glycolysis was employed. Thus, 3-phosphoglycerate was used at 10 mM concentration in conjunction with 1 mM ATP to be converted by the ancillary enzyme phosphoglycerate kinase to 1,3-bisphosphoglycerate, which served as a substrate for the GAPDH reaction. Note that in the current assays, 5 mM cysteine was replaced by 20 mM dithiothreitol as a reducing agent. All specific activities are expressed as nmoles of substrate converted per min and mg of protein at 30 • C (mU/mg). To test for KlGdp1 activity, NADH was replaced by NADPH. All enzyme assays were carried out in a Beckman DU800 photometer (Beckman-Coulter, Krefeld, Germany) with an automated cell changer kept at 30 • C, recording the decrease in adsorption at 340 nm for 10-15 min.