Diversification of the kinetic properties of yeast NADP‐glutamate‐dehydrogenase isozymes proceeds independently of their evolutionary origin

Abstract In the yeast Saccharomyces cerevisiae, the ScGDH1 and ScGDH3 encoded glutamate dehydrogenases (NADP‐GDHs) catalyze the synthesis of glutamate from ammonium and α‐ketoglutarate (α‐KG). Previous kinetic characterization showed that these enzymes displayed different allosteric properties and respectively high or low rate of α‐KG utilization. Accordingly, the coordinated action of ScGdh1 and ScGdh3, regulated balanced α‐KG utilization for glutamate biosynthesis under either fermentative or respiratory conditions, safeguarding energy provision. Here, we have addressed the question of whether there is a correlation between the regulation and kinetic properties of the NADP‐GDH isozymes present in S. cerevisiae (ScGdh1 and ScGdh3), Kluyveromyces lactis (KlGdh1), and Lachancea kluyveri (LkGdh1) and their evolutionary history. Our results show that the kinetic properties of K. lactis and L. kluyveri single NADP‐GDHs are respectively similar to either ScGDH3 or ScGDH1, which arose from the whole genome duplication event of the S. cerevisiae lineage, although, KlGDH1 and LkGDH1 originated from a GDH clade, through an ancient interspecies hybridization event that preceded the divergence between the Saccharomyces clade and the one containing the genera Kluyveromyces, Lachancea, and Eremothecium. Thus, the kinetic properties which determine the NADP‐GDHs capacity to utilize α‐KG and synthesize glutamate do not correlate with their evolutionary origin.

enzyme displaying sigmoidal kinetics for α-KG utilization, this isoform contributes significantly to NADP-GDH activity during growth on ethanol as sole carbon source (Avendano et al., 2005;DeLuna et al., 2001) and becomes the predominant isoform during stationary phase (Lee et al., 2012). Accordingly, transcription of the ScGDH3 gene is strongly induced during growth on ethanol and is nearly absent on glucose. This carbon-mediated regulation is overimposed to the transcriptional activation by low nitrogen availability (Avendano et al., 2005). Although transcription of the ScGDH1 gene is not repressed on ethanol, the relative contribution of the ScGdh1 enzyme to the overall NADP-GDH activity is much lower than that of ScGdh3 under this condition (DeLuna et al., 2001;Riego, Avendano, DeLuna, Rodríguez, & González, 2002). It is worth mentioning that the NADP-GDHs are not involved in glutamate catabolism, instead, the NAD-dependent glutamate dehydrogenase (1.4.1.2) catalyzes the deamination of glutamate to ammonium and α-KG in yeast (Miller & Magasanik, 1990).
It has been proposed that ScGdh1 and ScGdh3 kinetic differences control α-KG utilization for biosynthetic purposes without compromising flux trough the tricarboxylic acid cycle for energy production during growth on ethanol as sole carbon source (DeLuna et al., 2001). The non-redundant roles of ScGdh1 and ScGdh3 may be the result of an evolutionary process in which duplication of an ancestral gene and divergence of the resulting paralogous led to specialization in glutamate production under different conditions associated with the peculiar facultative metabolism of S. cerevisiae (Avendano et al., 2005).
It has been proposed that in the S. cerevisiae lineage, a whole genome duplication (WGD) event took place (Wolfe & Shields, 1997) and that a selected group of the resulting duplicated genes have been retained in two copies among which are the paralogous ScGDH1 and ScGDH3 genes (Seoighe & Wolfe, 1999). However, the evolutionary studies of the fungal NADP-GDHs have not addressed the characteristics of the pre-WGD ancestral-type genes which did not originate through WGD, and those present in the Saccharomycetes, which arose through WGD.
The Saccharomycetales (or Hemyascomycetes) group includes species closely related to S. cerevisiae for which the genome sequence and genetic manipulation resources are available, representing a valuable tool for functional evolutionary studies. The yeasts Kluyveromyces lactis and Lachancea kluyveri descend from the pre-WGD ancestor, and have a single NADP-GDH-encoding gene, suggesting that no sporadic duplications have occurred in this gene. With regard to the carbon metabolism operating in these yeasts, it is evident that each one shows different levels of adaptation to the fermentative lifestyle: K. lactis metabolism is constitutively respiratory, for this reason, it cannot grow anaerobically and does not produce respiratory-deficient mutants (Breunig et al., 2000). L. kluyveri displays an intermediate fermentative capacity between K. lactis and S. cerevisiae, it can grow anaerobically and produce respiratory-deficient mutants on sugar-rich media, but it only ferments in the absence of oxygen (Moller, Olsson, & Piskur, 2001;Moller et al., 2002), whereas in S. cerevisiae fermentative metabolism predominates whenever high sugar concentration is available regardless of oxygen disponibility. It even represses respiratory metabolism in the presence of high glucose or fructose concentration, through carbon catabolite repression (Gancedo, 1998). This yeast can grow anaerobically and produce respiratory-deficient mutants (Gancedo, 1998). One of the most prominent features of baker′s yeast is the rapid conversion of sugars to ethanol and carbon dioxide under both anaerobic an aerobic conditions; this phenomenon is called Crabtree effect (Hagman, Säll, & Piskur, 2014) and is present in yeast species well adapted to the fermentative life style (Pfeiffer & Morley,2014). According to this classification, S. cerevisiae and L. kluyveri are Crabtree positive, whereas K. lactis is Crabtree negative.
This work addresses the question of whether the evolutionary origin of S. cerevisiae ScGdh1 and ScGdh3 NADP-GDH and their corresponding orthologs in K. lactis and L. kluyveri has influenced their kinetic and transcriptional regulation. Our results show that such regulation does not correlate with the evolutionary origin of the corresponding genes, confirming that gene duplication and further functional diversification play a key role in metabolic remodeling and evolution, regardless of the origin of paralogous gene pair. Table 1 describes the characteristics of the strains used in the present work. All strains constructed for this study were derivatives of CLA1 (ura3 leu2),  or KlWM37-1 (his3 ura3) for S. cerevisiae, L. kluyveri and K. lactis, respectively. Mutants in Scgdh1Δ::kanMx4 (CLA2), Scgdh3Δ::LEU2 (CLA3), Scglt1Δ::URA3 (CLA4), Scgdh1Δ::kanMx4 Scgdh3Δ::LEU2 (CLA5), Scgdh1Δ::kanMx4

| Growth conditions
Strains were routinely grown on minimal medium (MM) containing salts, trace elements, and vitamins following the formula of yeast nitrogen base (Difco
Transformants were selected for uracil prototrophy on MM.

| Cloning and expression
The ScGDH1 and ScGDH3 genes were PCR amplified using the deoxyoligonucleotides pairs 125/126 and 127/128, respectively, using genomic DNA of the CLA1 WT strain as a template. PCR products and the pET-28a(+) plasmid were NheI/XhoI digested and after gel purification were ligated. The LkGDH1 gene was amplified with the deoxyoligonucleotides 129 and 130 using genomic DNA of the Lk156-1 WT strain as a template. PCR product and the pET-28a(+) plasmid were NdeI/BamHI digested and after gel purification were ligated. The KlGDH1 gene was amplified with the deoxyoligonucleotides 131 and 132 using genomic DNA of the KlWM37-1 WT strain as a template.
PCR product and the pET-28a(+) plasmid were NheI/BamHI digested and after gel purification ligated.
Ligations were transformed into the DH5α E. coli strain. After plasmid purification, correct cloning was verified by sequencing. For heterologous expression, the BL21 E. coli strain was transformed. Selected clones were grown in LB medium supplemented with 30 μg ml −1 of kanamycin incubated at 37°C with shaking (250 rpm). When the cultures reached an OD of 0.6 at 600 nm, the expression of the proteins was induced with 100 μmol/L of IPTG (Iso-Propil-Tio-Galactoside), incubated 4 hr at 30°C with shaking (250 rpm), harvested by centrifugation at 1100g for 15 min, and the cellular pellet was stored at −70°C until used.

| Whole cell soluble protein extract
Cells were thawed and resuspended in 20 ml of 30 mmol/L imidazol, 1 mmol/L EDTA, 1 mmol/L dithiothreitol, 1 mmol/L phenylmethylsulfonylfluoride (PMSF). Protein extracts were obtained by sonication (Ultrasonic Processor Model: VCX 130) with a tip sonicator keeping the tubes on ice; five cycles (60% amplitude, one second on and one second off for 1 min) with 1 min of incubation on ice between each cycle. After centrifugation at 1100g for 20 min at 4°C, the supernatant was stored at −20°C until used.

| Affinity Chromatography
To purify the NADP-GDH proteins, the supernatant was loaded on an equilibrated nickel column (Ni-NTA Agarose 100, Thermo Fisher Scientific), which was then washed 10 times with 30 mmol/L imidazol. The protein was eluted with 500 mmol/L imidazol and stored at −20°C until used. Homogeneity of proteins was verified with a polyacrylamide gel electrophoresis 12% (SDS-PAGE) stained with Coomassie Blue (Fig. S2).

| Enzyme assay and protein determination
Whole yeast cell soluble protein extracts were prepared by sonication lysis of cell pellets harvested during exponential growth. The NADP-GDH activity was assayed by the method of Doherty (Doherty, 1970).
Protein was measured by the method of Lowry (Lowry, Rosebrough, Farr, & Randall, 1951), using bovine serum albumin as a standard.
At every glutamic concentration, the α-KG, ammonium chloride and NADPH were fixed (8 μmol/L, 100 mmol/L, and 250 μmol/L, respectively). In order to select the inhibition model, the data were fitted to different models, with the program Dynafit. IC 50 results were globally obtained with program GraphPad Prism 7.00 (Software Inc.).

| Northern blot analysis
Northern blot analysis was carried out as previously described Struhl K. and Davis RW (1981

| Nucleosome scanning assay (NuSA)
The nucleosome scanning assay was made to see the chromatin organization ScGDH1, ScGDH3, LkGDH1 y KlGDH1 promoter, and the procedure to the study of the positioning of nucleosomes on promoters was made as described by Infante et al. 2011 (Table S3) of ScGDH1, ScGDH3, LkGDH1 y KlGDH1 locus whose coordinates are given relative to the ATG (+1).

| Metabolite extraction and analysis
Cell extracts were prepared from exponentially growing cultures.
Samples used for intracellular amino acid determination were treated as previously described (Quezada et al., 2008).

| Phylogenetic analysis
A total of 26 taxa were used in the analysis, including two ascomycetes as outgroup (

| NADP-GDH is the main glutamate-producing pathway in S. cerevisiae, L. kluyveri, and K. lactis
To analyze the relative contribution of glutamate dehydrogenases (NADP-GDH) and glutamate synthase (GOGAT) to glutamate biosynthesis, mutant strains were constructed in which the genes encoding for the NADP-GDH (GDH1/GDH3) or GOGAT (GLT1) were inactivated (Table 1). Growth rates of these mutants were determined on minimal media with glucose or ethanol as carbon sources and ammonium as nitrogen source (Table 2). In the three yeast species, inactivation of the NADP-GDH-encoding genes resulted in a strong reduction in growth rate on both carbon sources (from 60% to 80% relative to the corresponding WT strains) indicating that the proteins ScGdh1/ScGdh3 in S. cerevisiae, LkGdh1 in L. kluyveri and KlGdh1 in K. lactis, are the main contributors to glutamate production under the conditions studied. The glutamine synthetase-GOGAT pathway in S. cerevisiae made a marginal contribution to glutamate production under the conditions studied because the glt1Δ mutant strain grew as well as the wild-type strain (Table 2). However, in L. klyveri and K. lactis, the GOGAT pathway made a significant contribution since inactivation of the GLT1 genes, resulted in reduction of growth rates ranging from 25% to 60% (Table 2). As expected, the mutants lacking NADP-GDH and GOGAT-encoding genes were full glutamate auxotrophs (Table 2). In the pre-WGD species, only one gene is responsible for the NADP-GDH activity because inactivation of either LkGDH1 or KlGDH1 resulted in complete lack of this activity (Table 2). In agreement with previous reports (DeLuna et al., 2001), the contribution of ScGdh3 was evident on ethanol but not on glucose.
When glutamate was supplemented to the growth media, gdhΔ mutant strains recovered wild-type growth (Table 2). However, this was not the case for the L. kluyveri and K. lactis glt1Δ mutant strains, which did not recover wild-type growth rate by glutamate addition. As previously reported, in addition to glutamate biosynthesis, GOGAT plays other role, which has been found to be critical for the maintenance of the redox balance and cytosolic NADH homeostasis (Guillamon et al., 2001).

| Glutamate is not a negative regulator of the S. cerevisiae, L. kluyveri and K. lactis, NADP-GDHs
To further analyze the regulation of the NADP-GDH enzymes, specific activities in the presence of glutamate were determined. The use of 5 mmol/L glutamate as nitrogen source did not result on a reduction in NADP-GDH specific activities as compared to those observed when ammonium sulfate was used (Table 2). Furthermore, when clarified extracts from the WT strains were analyzed in the presence of increasing glutamate concentrations, the half inhibitory concentrations were in the range of 376-681 mmol/L (Fig. S1); this range is much higher than the estimated cytosolic glutamate concentration, 9-46 mmol/L (Table 4). These results indicate that glutamate does not trigger strong negative regulatory mechanisms (e.g., repression of transcription or feedback inhibition) of glutamate biosynthesis under the conditions studied.

| S. cerevisiae, L. kluyveri, and K. lactis showed different patterns of carbon source-dependent transcriptional regulation of NADP-GDH encoding genes
In order to deepen the studies with regard to carbon sourcedependent regulation of the NADP-GDH enzymes, specific activities, transcript levels of the corresponding genes and nucleosome positioning on the promoter regions were analyzed. As previously reported, transcription of the ScGDH3 gene was higher on ethanol as carbon source compared to that observed on glucose ( Figure 1a) (Avendano et al., 2005 andRiego et al., 2002). This was accompanied by nucleosome clearance on the −754 bp to −128 bp ScGDH3 promoter region (Figure 1c) (Avendano et al., 2005). Transcript levels of the ScGDH1 gene and nucleosome positioning on the promoter region were similar on both carbon | 7 of 18 sources (Figure 1a and b). Albeit in S. cerevisiae, the overall activity was similar on both carbon sources (Table 2), and the relative contribution of the ScGdh3 isoform was higher on ethanol than on glucose. When the mutant strain Scgdh1Δ was grown on ethanol, the ScGdh3-specific activity was 10-fold increased (Table 2).
This is in accordance with previous results (DeLuna et al., 2001) demonstrating the observed differential contributions of each enzyme to growth rates (Table 2). When L. kluyveri and K. lactis were grown on ethanol as carbon source, activities were increased 80% and 50%, respectively, compared to those observed on glucose (Table 2). In L. kluyveri, transcription of the LkGDH1 gene was slightly increased on ethanol as carbon source concomitantly with nucleosome clearance on the −738 bp to −336 bp promoter region (Figure 2a and b), whereas that transcriptional levels and nucleosome positioning of KlGDH1 gene no change in carbon sources studied (Figure 2c and d).

| ScGDH1/LkGDH1 and ScGDH3/KlGDH1 gene pairs showed distinctive heterologous complementation patterns
To determine to what extent the various NADP-GDHs were specialized to the metabolic peculiarities of the species they belong to, het-  Figure 3d and h). Interestingly, expression of the ScGDH1 and LkGDH1 genes also showed a similar trend when expressed on K. lactis grown on ethanol. In this case, however, the corresponding growth rates were significantly lower than those of the WT strain (third and fifth bars in Figure 3h), which showed the opposite trend to that observed upon expression on S. cerevisiae or L. kluyveri.
Ectopic expression of the ScGDH3 and KlGDH1 genes also showed a similar trend in the complementation experiments. When expressed on L. kluyveri, they did not improve growth of the Lkgdh1Δ mutant strain (fourth and sixth bars in Figure 3c and g). When expressed on S. cerevisiae, they showed significant complementation but the growth rates were still lower than those of the corresponding WT strains (fourth and sixth bars in Figure 3a

F I G U R E 3 Complementation tests.
Growth rates values are shown relative to the WT strains carrying the empty plasmid: for S. cerevisiae 0.23 hr −1 and 0.15 hr −1 ; for L. kluyveri, 0.12 hr −1 and 0.07 hr −1 ; for K. lactis, 0.35 hr −1 and 0.14 hr −1 on glucose and ethanol, respectively. In all cases, standard deviations of at least three independent cultures were less than three percent. For ectopic expression, the plasmid pRS416 was used for S. cerevisiae, the pLk-EE for L. kluyveri and the YEpKD352 for K. lactis as described in Experimental procedures encoded proteins, in a separate group from the ScGDH3 and KlGDH1 genes. To determine to what extent the amount of active enzyme is responsible of this effect, the NADP-GDH specific activities were determined ( Figure 4). As expected, the non-complemented Scgdh1Δ Scgdh3Δ double-mutant strain, as well as the Lkgdh1Δ and Klgdh1Δ single mutants did not show detectable activity (lack of a second bar in Figure 4b, c, d, f, g and h). Growth of these mutant strains, which are devoid of NADP-GDH activity, may have involved the GOGATdependent glutamate-producing pathway (Magasanik, 2003). In the Scgdh1Δ single mutant strain, however, presence of the ScGDH3 gene was responsible of a very low activity on glucose and a significant activity on ethanol (second bars in Figure 4a and  Unexpectedly, NADP-GDH activity was not detected in three complemented strains: the Lkgdh1Δ mutant strain grown on glucose bearing the ScGDH3 and KlGDH1 genes (lack of fourth and sixth bars in Figure 4C), and the Klgdh1Δ mutant strain grown on ethanol bearing the ScGDH1 gene (lack of a third bar in Figure 4h). This suggested that expression of the heterologous genes in these strains was very low and below the detection limit. Interestingly, heterologous complementation of K. lactis with the S. cerevisiae ScGDH1 and ScGDH3 genes resulted in higher activities of ScGdh3 than those observed for ScGdh1 (third and fourth bars in Figure 4d and h) and this effect was similar to the relative contribution of the two isoforms observed in the WT strain grown on ethanol (Table 2).

| K. lactis KlGdh1 and S. cerevisiae ScGdh3 isoforms showed cooperativity for α-KG utilization, whereas L. kluyveri LkGdh1 and S. cerevisiae ScGdh1 isoforms showed hyperbolic kinetics
In order to analyze the NADP-GDHs biochemical characteristics and find additional elements that could contribute to better understand the results obtained in the complementation tests, substrate utilization kinetics was studied. The His-tagged ScGdh1, ScGdh3, LkGdh1, and KlGdh1 enzymes were purified to electrophoretic homogeneity after heterologous expression in E. coli (Fig. S2). Apparent molecular masses of the ScGdh1, ScGdh3, LkGdh1, and KlGdh1 monomers, respectively, were: 53 kD, 55 kD, 54 kD, and 54 kD. In all cases, values are close to those expected. Initial velocity measurements were made at different substrate concentrations (α-KG, NADPH and ammonium chloride); when the amount of one substrate was varied, the other two were kept at saturating concentrations. The reaction was assayed as previously reported (DeLuna et al., 2001).
The responses to increasing substrate concentrations were heterogeneous: they varied from hyperbolic for the three substrates in the S. cerevisiae isoform ScGdh1 and the L. kluyveri enzyme LkGdh1, to sigmoidal for the three substrates in the K. lactis enzyme KlGdh1 ( Figure 5, S3 and S4). The S. cerevisiae isoform ScGdh3, however, showed hyperbolic response for the three substrates at pH 7.5, but sigmoidal for α-KG at pH 5.8, in agreement with a previous study in which the nontagged homologous purified protein was used (DeLuna et al., 2001). This pH-dependent difference in the shape of the α-KG saturation curve was only observed for the ScGdh3 isoform ( Figure 5, S3 and S4). Experimental data from the hyperbolic or sigmoidal responses were fitted to the Michaelis-Menten or the Hill equations, respectively, and the resulting kinetic parameters are shown in Table 3. At pH 7.5, the four enzymes showed similar turnover numbers, which indicate the catalytic events per unit of time (k cat ), and similar affinities for NADPH and ammonium (Table 3). The ScGdh1, ScGdh3, and LkGdh1 isoforms also showed similar affinities for α-KG, showing similar K m-αKG values. However, KlGdh1 showed lower affinity and strong cooperativity for α-KG utilization, its affinity constant (S 0.5 ) for α-KG was about eight times higher than the K m-αKG of the other enzymes (Table 3) and the fitted Hill number (n H-α-KG ) resulted to be 4.4. These KlGdh1 characteristics were also observed at pH 5.8 and, at this pH value, were also shared by the ScGdh3 isoform. At the acidic pH, the k cat value of the K. lactis enzyme resulted to be three times lower than that observed at pH 7.5 (Table 3). Enzymatic activities of the LkGdh1 and KlGdh1 enzymes were assayed at different pH values ranging from 5 to 9. Maximum activities were observed at pH 7.0 (data not shown), which is close to the 6.8 value reported for the S. cerevisiae isoforms (DeLuna et al., 2001).
In order to get some insight into the in vivo kinetic regulation of the NADP-GDHs, the α-KG and glutamate intracellular pools were determined (Table 4). Similar α-KG values were detected in the three yeast species, except for L. kluyveri grown on ethanol which showed much lower α-KG content, which was probably associated with the very low growth rate observed for this yeast ( Table 2). The estimated α-KG cytosolic concentrations in S. cerevisiae and L. kluyveri were in the range of 0.1-1.14 mmol/L which is close to 0.4-1.76 mmol/L, the reported values for S. cerevisiae (Cueto-Rojas, Maleki Seifar, Ten Pierick, Heijnen, & Wahl, 2016;Hans, Heinzle, & Wittmann, 2003). To estimate these concentrations, a cell volume of 29 μm 3 of which 75% of corresponded to cytosol was used (Kitamoto, Yoshizawa, Ohsumi, & Anraku, 1988), and to compare them with the reported values, a cell volume of 1.7 ml/g cell dry weight was used (Zhang et al., 2015).  (Tables 3 and 4). At these concentrations, the catalytic rates of the ScGdh1, ScGdh3, and LkGdh1 enzymes were highly responsive to changes in substrate concentration (Figure 3). In K. lactis, the α-KG estimated concentration was 0.7 mmol/L, which is around one-fourth of the S 0.5-α-KG (Tables 3 and 4). At this concentration, and because of its strong cooperativity, the KlGdh1 enzyme catalytic rate was not highly responsive to changes in substrate concentration (Figure 3).
The intracellular glutamate content observed on ethanol as carbon source was similar to that observed on glucose in S. cerevisiae and K. lactis, although the pools in the former were higher than in the latter (Table 4). In L. kluyveri, however, the intracellular glutamate pool was much higher on ethanol than on glucose. This could be the results of the high-affinity LkGdh1 has for α-KG, driving intermediate flux to glutamate biosynthesis, which could result in decreased growth rate due to a restriction of carbon flow through Krebs Cycle and limited energy production resulting in low α-KG intracellular concentration (Table 4).
The estimated glutamate intracellular concentrations was found in the range of 9-46 mmol/L (Table 4), which are near to the 30-80 mmol/L, reported values for S. cerevisiae (Cueto-Rojas et al., 2016;Hans et al., 2003). At these concentrations, glutamate does not exert a significant inhibitory effect over the NADP-GDH activity as indicated by the high IC 50 values, which were detected (Fig. S1).

| Phylogenetic analysis of NADP-GDH-sequences revealed that the evolutionary origin of these proteins does not correlate with their kinetic properties
To analyze whether the observed ScGdh1, ScGdh3, LkGdh1, and KlGdh1 regulatory and kinetic properties correlated with the evolutionary origin of these proteins, a phylogenetic tree was constructed with NADP-dependent Gdh sequences from fungal representatives of the different taxonomical classes ( Figure 6).
Overall, GDH phylogenies resembled the taxonomical classifications; however, three major aspects should be highlighted. 1) ScGdh1 (Scer Gdh1) and ScGdh3 (Scer Gdh3) were grouped in a separated clade together with Gdh1 and Gdh3 sequences from

| Kinetic properties comparisons allow grouping ScGdh1 and LkGdh1 enzymes in a peculiar group, different from ScGdh3 and KlGdh1
Hyperbolic kinetics and high affinity for α-KG are shared properties of the ScGdh1 and LkGdh1 enzymes, while cooperativity for α-KG utilization is a shared property of the ScGdh3 isoform with the KlGdh1 enzyme (Table 3, Figure 5). However, the physiological significance of ScGdh3 cooperativity is not clear since it has been only observed at pH 5.8 (Table 3) or lower (DeLuna et al., 2001). Intracellular pH during exponential growth has been reported to be close to neutrality: 7.2 on glucose and 6.8 on a mix of 2% ethanol and 2% glycerol (Orij, Postmus, Ter Beek, Brul, & Smits, 2009).
In non-growing glucose-starved cells, however, it drops to 5.5-6.0 (Orij et al., 2009). These last conditions may reflect the stationary T A B L E 3 Kinetic parameters of the studied NADP-GDHs  phase context in which ScGdh3 plays a significant role (Lee et al., 2012). It is possible that ScGdh3 cooperativity is a reminiscent feature of the ancestor protein without a true physiological and metabolic role in vivo; however, it can also be the case that an unknown allosteric effector induces cooperativity during exponential growth on ethanol, whose effect may be mimicked by acidic pH "in vitro".  (5)  18 Concentrations in mmol/L represent the cytosolic pool and were estimated considering a cell volume of 29 μm 3 of which 75% of corresponded to cytosol (Kitamoto et al., 1988). Numbers in parentheses are standard deviations.
F I G U R E 6 Evolutionary relationships of NADP-GDHs from yeasts. The phylogeny was constructed using the neighbor-joining method (Saitou & Nei, 1987). The optimal tree with the sum of branch length = 1043.39892578 is shown. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (500 replicates) are shown next to the branches (Felsenstein, 1985). The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the number of differences method (Nei & Kumar, 2000) and are in the units of the number of amino acid differences per sequence. The analysis involved 31 amino acid sequences. All ambiguous positions were removed for each sequence pair. There were a total of 484 positions in the final dataset. Evolutionary analyses were conducted in MEGA6 (Tamura et al., 2013). Scer Gdh1, Saccharomyces cerevisiae ScGdh1 (red letters and square bracket); Scer Gdh3, Saccharomyces cerevisiae ScGdh3 (dark blue letters and square bracket); Lklu Gdh1, Lachancea kluyveri LkGdh1 (green letters and square bracket); Klac Gdh1, Kluyveromyces lactis KlGdh1 (light blue letters and square bracket). Post-WGD, post-whole genome duplicacion clade (light brown group); ZT, Zygosaccharomyces-Torulaspora clade (light yellow group); KLE, Kluyveromyces-Lachancea-Eremothecium clade (light blue group). Tbla Gdh1, Torulaspora blattae Gdh1 (red square). NADP-GDH sequence accession numbers, taxa used and their corresponding abbreviations are included in   (Table 3). The estimated α-KG physiological concentration in this yeast was 0.7 mmol/L (Table 4) and at this concentration, the KlGdh1 catalytic rate is not highly responsive to changes in α-KG ( Figure 3). This suggests that α-KG availability does not determine glutamate synthesis in K. lactis. However, the low catalytic rate observed for the KlGdh1 enzyme at 0.7 mmol/L (Figure 3), may not be compatible with the fast growth observed for K. lactis (Table 2). It is possible that unknown activators contribute to modulation of the K.
lactis enzyme in vivo. Interestingly, the S. cerevisiae and L. kluyveri enzymes are highly responsive to changes in α-KG at the physiological concentrations (Table 4 and Figure 3) this suggests that the rate of glutamate synthesis is highly influenced by α-KG availability as was previously proposed (Quezada et al., 2013). Worth of mention is the fact that there is growing evidence indicating that α-KG plays a role in metabolic regulation. Thus, modulation of the intracellular α-KG levels could constitute important mechanisms of metabolic control.
In this regard, it has been proposed that in Caenorhabditis elegans, α-KG is a key metabolite mediating longevity by dietary restriction (Chin et al., 2014). Intracellular α-KG/succinate levels can contribute to the maintenance of cellular identity and have a mechanistic role in the transcriptional and epigenetic state of mouse stem cells (Carey, Finley, Cross, Allis, & Thompson, 2015). Most interestingly, recent studies of Gdh1 function has revealed that gdh1 mutants show enhanced N-terminal histone H3 proteolisis, suggesting that α−KG has a key regulatory role in telomere silencing in S. cerevisiae (Su & Pillus, 2016).
Reported NADPH intracellular concentration is around 286 μmol/L (Zhang et al., 2015; authors considered a cellular volume of 1.7 ml/g cell dry weight), which corresponds to 6-7 K m-NADPH or S 0.5-NADPH (Table 3). At this concentration, the activities of the herein studied NADP-GDH enzymes were not responsive to changes on the NADPH concentration (Fig. S4). This indicates that the physiological concentration of this substrate is close to saturation and does not determine the NADP-GDH activity in vivo. By contrast, the reported intracellular ammonium concentration is 2.2 mmol/L (Cueto-Rojas et al., 2016; considering a cellular volume of 1.7 ml/g cell dry weight as in Zhang et al., 2015). This value is well below the K m-NH4+ or S 0.5-NH4+ shown in Table 3, which indicates that the ammonium availability modulates the NADP-GDH activity in vivo. Thus, glutamate synthesis by NADP-GDH seems to be mainly determined by α-KG and ammonium availability and not by product inhibition by glutamate.
Most interesting was the fact that, the kinetic behavior of the enzymes present in the two yeast species which show significant fermentative capacity when grown on high glucose media (ScGdh1 in S. cerevisiae and LkGdh1 in L. kluyveri) was hyperbolic, showing high affinity for α-KG (K m-α-KG ≈ 0.4 mmol/L). The enzyme present in the yeast with a predominantly respiratory metabolism, KlGdh1 from K. lactis, and the ScGdh3 isoform whose contribution to glutamate synthesis increases during respiratory metabolism in S. cerevisiae (Table 2), were cooperative and showed low affinity for α-KG (S 0.5KlGdh1, pH7.5 = 3.61 mmol/L and S 0.5ScGdh3, pH5.8 = 1.95 mmol/L).
Assuming that during S. cerevisiae evolution, L. kluyveri and K. lactis, selective pressures drove changes in the NADP-GDHs, these enzymes could have changed from cooperative to hyperbolic. As cooperativity was observed in the NADP-GDH from K. lactis and in the ScGdh3 isoform, it seems possible that the common ancestor of the three yeast species had a cooperative NADP-GDH and that this property was lost two times: one in the L. kluyveri lineage after divergence of the S. cerevisiae and L. kluyveri branches, and the other after the WGD event which resulted in the conservation of the hyperbolic ScGdh1 and the cooperative ScGdh3. This suggests that NADP-GDH kinetics may be related to adaptation to the fermentative or respiratory lifestyles, and further research in various yeast species is needed to explore this possibility.