Intracellular NADPH Levels Affect the Oligomeric State of the Glucose 6-Phosphate Dehydrogenase

Analysis of G6PDH in respiratory and fermentative mutants.We reported the presence of an additional G6PDH activity band in cellular extracts from the K. lactis MW179-1D strain grown in ethanol (33). To further investigate this finding, we studied the G6PDH activity pattern in K. lactis respiratory mutants partially or totally impaired in ethanol utilization. Since these strains showed no or poor growth in minimal medium supplemented with ethanol, they were cultivated in ethanol-supplemented rich medium to study the effect of this carbon source on G6PDH. These mutants were deleted for KlNDE1, KlNDI1, KlSDH1, or KlCOX14, genes involved in the synthesis of components of the respiratory transport chain (Fig. 1). In particular, KlNde1 and KlNdi1 (35, 37, 38) are two transdehydrogenases located on the inner mitochondrial membrane, with the active site facing the outer membrane and the matrix, respectively. These two activities stem from single rotenone-insensitive proteins that in yeast substitute for the respiratory transport chain complex I of higher eukaryotes. Different from the corresponding enzymes of S. cerevisiae that use NADH as the substrate, KlNde1 but also KlNde2, a second transdehydrogenase similar to KlNde1, can reoxidize both NADH and NADPH (37, 38) (Fig. 1). The Klsdh1Δ and Klcox14Δ mutants are unable to assemble complex II and IV (11, 32), respectively. Both parental and mutant strains showed a single band of G6PDH activity when grown in glucose (Fig. 2A, lanes 1, 3, 5, 7, and 9). In the presence of ethanol, Klnde1Δ and Klsdh1Δ showed two bands, as did the parental strain (Fig. 2A, lanes 4 and 8 versus lane 2), whereas Klndi1Δ and Klcox14Δ only showed the upper band of G6PDH activity (Fig. 2A, lanes 6 and 10). Since Klcox14Δ, differently from Klsdh1Δ (32), is a respiratory-deficient mutant (11) and Klndi1Δ is devoid of the major NADH reoxidation activity (35), one can conclude that the expression of the second G6PDH band in the presence of ethanol requires respiration and efficient coenzyme reoxidation.

• Open in new tab
• Download powerpoint

Fig 1

Glucose catabolic routes in wild-type (A) and adh null mutant (B) strains. The model depicts glycolysis, the PPP, mitochondrial respiratory transport chain, and major cytosolic NAD(P)⁺/NAD(P)H reoxidation activities. GPD1, NAD-dependent glycerol 3-P dehydrogenase; GUT2, FAD-dependent glycerol-3 P dehydrogenase; NDE1 and NDI1, external and internal inner mitochondrial membrane transdehydrogenases; PDC1, pyruvate decarboxylase; PGI1, phosphoglucoisomerase; SDH, succinate dehydrogenase complex; Q, ubiquinone; bc1, complex III; cox, cytochrome oxidase. Pathways/activities involved in the accumulation/reoxidation of NAD(P)H are highlighted in boxes.

• Open in new tab
• Download powerpoint

Fig 2

G6PDH activities separated on native gels for wild-type and mutant strains. (A) G6PDH from the MW179-1D strain (wt) and its isogenic respiratory mutants. (B) G6PDH from glycolytic and respiratory mutants in the CBS2359 context. (C) G6PDH from CBS2359 grown on increasing concentrations of glucose and from the MW179-1D strain transformed with the RAG1 multicopy plasmid. Extracts were prepared from late-exponential-phase cultures grown on YP medium containing glucose (D), ethanol (E), or glycerol (G) at 2% or at the concentration of glucose indicated in the figure. Glucose at 0.5% plus 2% ethanol (DE) was added to cultures of Klsdh1Δ and Klcox14Δ mutants unable to grow on respiratory carbon sources. Cell extracts were fractioned on polyacrylamide gels and stained for G6PDH activity. The values within the boxes in panels B and C indicate the amount (as a percentage) of each G6PDH band in each lane, as determined by densitometric analysis using the Molecular Imaging software (Kodak).

Extending such an analysis to other K. lactis reference strains, we obtained unexpected results. One of these strains, namely, CBS2359, in contrast with MW179-1D, showed both bands of activity in extracts from glucose cultures as well as from ethanol cultures (Fig. 2B, lanes 1 and 2), while showing only the upper band with glycerol (Fig. 2B, lane 3). To test the influence of fermentative and respiratory metabolism on the presence of the two activities, we analyzed three isogenic mutants impaired for those metabolisms. Klpda1Δ, one of these mutants, lacks the E1a subunit of the pyruvate dehydrogenase complex and, therefore, is unable to dissimilate pyruvate through mitochondria (36, 42). This respiratory mutant showed a G6PDH activity pattern identical to that of the wild type (Fig. 2B, lanes 10 to 12), although both the wild type and Klpda1Δ mutant showed evidence of an elongated staining shadow below the faster-migrating G6PDH band (Fig. 2B, lanes 2, 10, and 11), suggesting an instability of this activity when present at higher levels. On the contrary, Klpgi1Δ and Klpdc1Δ, two glycolytic mutants devoid of phosphoglucoisomerase and the pyruvate decarboxylase activities, respectively (Fig. 1) (6, 13), when grown in the presence of glucose, expressed only the upper G6PDH band (Fig. 2B, lanes 4 and 7). Since these mutants showed the same pattern as MW179-1D (Fig. 2A, lane 1), we speculated that the slightly different kinetics of accumulation/oxidation of ethanol by CBS2359 and MW179-1D during fermentation (31, 33) (unpublished results) could be the basis for their different G6PDH patterns (Fig. 1A). To confirm this hypothesis, we analyzed the G6PDH pattern in CBS2359 cultures grown with increasing concentrations of glucose, i.e., progressively higher fermentative capabilities. Indeed, as shown in Fig. 2C, this strain had only one G6PDH band in samples from cultures containing glucose at concentrations below 0.6% (lanes 1 to 3). A faint, faster-migrating band of activity appeared with 0.6% glucose and increased in intensity when the glucose reached 4%, at which point the two bands were expressed at identical levels as determined by densitometric analysis (Fig. 2C, lanes 4 to 9). The link between accumulation/oxidation of ethanol during fermentation (Fig. 1A) and the presence of two G6PDH bands was also confirmed in the MW179-1D strain. In fact, the transformation of this strain with the plasmid containing the low-affinity glucose transporter RAG1 gene, which has been reported to increase glycolytic flux (22, 40), led to the appearance of the lower band of activity (Fig. 2C, lane 10). These results indicate that the presence of the faster-migrating band is related to the presence of ethanol, whether produced/oxidated during fermentation or added to the culture medium (Fig. 1A).

Analysis of G6PDH in adh mutants.Since alcohol dehydrogenases (ADH) are enzymes involved in the production and oxidation of ethanol (Fig. 1), we analyzed the G6PDH activity bands in extracts from cultures of adh mutants. In Fig. 3A is shown the G6PDH activity pattern from strain MW98-8C and its isogenic Kladh1Δ, Kladh2Δ, and Kladh4Δ mutants grown in glucose, ethanol, or glycerol medium. The parental strain has reduced fermentative capability (12) and, in line with previous findings, showed a single G6PDH band of activity in glucose (Fig. 3A, lane 1). The G6PDH patterns of the three adh mutants grown in ethanol or glycerol were identical to that of the parental strain (Fig. 3A, lanes 5 and 6, 8 and 9, and 11 and 12, versus lanes 2 and 3). Unexpectedly, all mutants displayed, though faint, a second band of activity also in glucose extracts (Fig. 3A, lanes 4, 7, and 10).

• Open in new tab
• Download powerpoint

Fig 3

G6PDH activities separated on native gels for wild-type and mutant strains. (A) G6PDH from MW98-8C and its single isogenic adh mutants. (B) G6PDH from the adh° mutant, from the same strain with the KlADH1 or KlADH2 gene reintegrated into its chromosomal locus or from the adh° mutant transformed with a multicopy plasmid containing KlADH1. (C) G6PDH from the CBS2359 strain transformed with multicopy plasmids containing the genes indicated in each lane. Cell growth, cell extracts, gel conditions, and staining for G6PDH, as well as the values in the boxes, are as those described for Fig. 2.

To test whether the number of ADH genes expressed in the cell influenced the presence of the two activity bands, we analyzed the G6PDH pattern in the adh° mutant. Indeed, as shown in Fig. 3B, the null strain, devoid of all ADH activities and unable to ferment (15), displayed the two bands of G6PDH (lane 1) as reported for the highly fermentative strain CBS2359. Moreover, the adh° mutant, unable to grow in minimal medium containing ethanol, still grew in rich ethanol medium (YPE); extracts prepared under this condition only showed the G6PDH upper band (Fig. 3B, lane 2). The important role played by ADH activities on the control of the G6PDH pattern was confirmed by the reintroduction of either KlADH1 or KlADH2 into the respective chromosomal locus of the adh° mutant (15), which allowed the reappearance of the lower G6PDH activity band with ethanol (Fig. 3B, lanes 5 and 8). However, the reintroduction of KlADH1 or KlADH2 did not lead to the disappearance of the faster-migrating G6PDH band in glucose-grown cells (Fig. 3B, lanes 4 and 7). On the contrary, the overexpression of KlADH1 and KlADH2 (data not shown) from a multicopy plasmid resulted in the presence of the G6PDH upper band alone in glucose or glycerol medium (Fig. 3B, lanes 10 and 12) and of two bands in ethanol (Fig. 3B, lane 11). These data confirmed that the intracellular amount of ADH activity is crucial for the appearance of one or both G6PDH bands when growth is carried out in glucose, while in ethanol even a small amount of ADH activity allows the expression of both bands (Fig. 3B, lanes 5, 8, and 11). Finally, the role of ADH in the control of G6PDH activity was confirmed by overexpressing in the CBS2359 strain other genes involved in the production and utilization of ethanol. These genes, KlADH1, KlADH3, KlALD4, and KlALD6, encode cytosolic (KlAdh1 and KlAld6) and mitochondrial (KlAdh3 and KlAld4) ADH and aldehyde dehydrogenase (ALD) activities, respectively (Fig. 1). As can be seen in Fig. 3C, the presence of both cytosolic activities altered the G6PDH glucose pattern of CBS2359, as the appearance of the lower band was inhibited (lanes 3 and 5). On the contrary, the mitochondrial KlAld4 and KlAdh3 were unable to change the G6PDH pattern observed in the wild type (Fig. 3C, lanes 2 and 4 versus lane 1). Moreover, this role was limited to KlAdh1 and KlAld6, in that the overexpression of other cytosolic or mitochondrial dehydrogenases, directly involved in the maintenance of the NAD(P)⁺/NAD(P)H redox balance, were unable to change the G6PDH expression pattern (Fig. 3C, lanes 6 to 10). These results indicate that the balance between cytosolic ADH and ALD activities and/or of the corresponding NAD⁺-NADH/NADP⁺-NADPH redox ratio control the appearance of the G6PDH faster-migrating band (Fig. 1).

NADP⁺ affects the G6PDH pattern in vitro.It has been reported that G6PDH activity is under strong feedback inhibition by NADPH (4), while NADP⁺ plays a major role in the tetramer-dimer interconversion of human G6PDH (41). To test in vitro the putative effect of NADP⁺(H) on G6PDH, we incubated the protein extracts, prepared from MW179-1D and CBS2359 glucose-grown cells, with increasing concentrations of NADP⁺(H). Cellular extracts of MW179-1D, expressing only the upper G6PDH band (Fig. 4A, lane 5), when incubated with NADP⁺ showed the appearance of the faster-migrating band at increasing concentrations of the cofactor (lanes 6 to 9). Moreover, when the same extract was assayed spectrophotometrically in the presence of 0.3 mM NADP⁺, in the absence of substrate, the appearance of a peak at 340 nm, indicative of NADPH formation, was observed (Fig. 4B, left panel). Since the incubation of NADP⁺ with the extract prepared from the MW179-1D/Klzwf1Δ null strain gave no evidence for NADP⁺ reduction to NADPH (data not shown), we concluded that NADPH formation in the extract of MW179-1D was specifically due to G6PDH. Indeed, the absorbance at 340 nm was dependent on the NADP⁺ added to the cellular extract, with the initial velocities (v_o) following the Michaelis-Menten equation (K_m = 0.029 ± 0.005 mM [mean ± standard deviation]) (Fig. 4B right panel). In contrast, incubation of the CBS2359 extracts with NADP⁺ showed no effect on the G6PDH activity pattern in either the native gel or absorption spectra (Fig. 4A, lanes 5 to 9, and C). The specific effect of NADP⁺ on the G6PDH activity pattern was further confirmed in both extracts from glycerol-grown CBS2359 cultures and from ethanol-grown MW179-1D cells, which expressed a single band of activity (Fig. 2B, lane 3) and two bands (Fig. 2A, lane 2), respectively. In fact, the incubation of glycerol-grown CBS2359 cell extract with NADP⁺ gave rise to the appearance of the faster-migrating G6PDH band in native gel, to the 340 nm peak of absorbance of NADPH, and to a variation of v_o at increasing concentrations of NADP⁺ almost identical to those from the glucose-grown MW179-1D extract (data not shown). Conversely NADP⁺ was unable to modify the G6PDH activity pattern in the extract from ethanol-grown MW179-1D cells or the absorption spectra of the extract itself (data not shown). Notably, the effect of NADP⁺ was specific, since incubation of the extracts from either wild-type strain with NADPH showed no alteration of the G6PDH activity pattern (Fig. 4A, lanes 1 to 4). Therefore, the appearance of the faster-migrating band in glucose-grown MW179-1D or glycerol-grown CBS2359 extracts could only be explained by the concomitant G6PDH-dependent reduction of NADP⁺ to NADPH. It follows that the faster-migrating G6PDH band, evident in vivo in the glucose-grown CBS2359 extracts (Fig. 2C, lanes 4 to 9) or in ethanol-grown extracts from other strains, originates from the upper one following cytosolic accumulation of NADPH.

• Open in new tab
• Download powerpoint

Fig 4

(A) G6PDH activities separated on native gels following in vitro incubation with increasing concentration of NADPH and NADP⁺. The extracts (about 10 μg), prepared from MW179-1D and CBS2359 cultures grown on YPD, were incubated overnight on ice with the cofactors, fractioned on polyacrylamide gels, and stained for G6PDH. (B, left panel) Absorbance spectra from MW179-1D extract following the addition of 0.3 mM NADP⁺. (Right panel) Effect of concentration of NADP⁺ on the initial velocity of NADPH reduction. The solid line is the best fit, according to the Michaelis-Menten equation. (C) Absorbance spectra from CBS2359 extract following the addition of 0.3 mM NADP⁺. (D) An experiment as described for panel A but with extracts from an S. cerevisiae culture grown on YPD and human rabdomiosarcoma cell lines. The values in boxes beneath the gels indicate the percentage of each G6PDH band in each lane, as for Fig. 2.

Because the amino acid sequences of G6PDH have been highly conserved during evolution (2, 26), we generated, via the Swiss-Model server, two models of K. lactis G6PDH (data not shown), using as the templates the human enzyme (1, 16, 28). Based on the homology model, the amino acid residues involved in tetramerization, in the binding of glucose 6-phosphate, coenzyme, and structural NADP⁺ (Fig. 5), as well as in the dimerization (data not shown) appear fully conserved in nature and in space not only in K. lactis (Fig. 5) but also in S. cerevisiae. We therefore suggest that a common mechanism is involved in the control of the G6PDH isoforms in all eukaryotes. Thus, we also tested the effect of NADP⁺ on S. cerevisiae and human G6PDH activities in vitro. Cellular extracts prepared from S. cerevisiae and from human rhabdomyosarcoma cells (a cancer made up of cells that normally develop into skeletal muscles) expressed a single band of G6PDH (Fig. 4D, lanes 1 and lanes 9 to 11, respectively). However, when incubated with increasing concentrations of NADP⁺, these extracts showed the presence of two bands (Fig. 4D, lanes 5 to 8 and 14 to 18) that were clearly visible in both extracts, starting from 6 mM NADP⁺ (Fig. 4D, lanes 5 and 15). This concentration of the cofactor was similar to that responsible for the appearance of the faster-migrating G6PDH band in K. lactis (Fig. 4A, lanes 6 and 7). Interestingly, the addition of NADP⁺ to human extracts produced a stabilizing effect on the two G6PDH activities compared to the extract incubated without cofactor or with just small amounts of cofactor (Fig. 4C, lanes 13 to 18 versus lane 12). Moreover, the single band present in human extract (Fig. 4D, lanes 9 to 12) seemed to correspond to the faster-migrating band when compared to K. lactis and S. cerevisiae G6PDH bands.

• Open in new tab
• Download powerpoint

Fig 5

Sequence alignment of G6PDH from K. lactis and H. sapiens, performed using Clustal X. The two proteins share 46% sequence identity. The fully conserved residues are shown by asterisks. The conservation of residues involved in the interaction with the coenzyme NADP⁺ (inverted black triangle), the substrate glucose-6-phosphate (inverted white triangle), the structural NADP⁺ (open circles), and in tetramerization (black circles) is shown. The structural model of K. lactis G6PDH was automatically built using as a template either one of the two crystal structures available for the human enzyme (PDB IDs 2BH9 or 1QKI). Both structures are from proteins missing the first 11 (1QKI) or the first 26 (2BH9) residues in the sequence.

K. lactis G6PDH bands correspond to the tetramer-dimer isoforms of human activity.The role played by NADP⁺ in the appearance of the second G6PDH band in yeasts and humans and the high conservation in amino acid residues responsible for G6PDH activity and structural integrity (2, 26) suggested that the two native bands detected in K. lactis corresponded to the dimer and tetramer G6PDH forms observed in the human enzyme. During expression of the chimeric KlZWF1GFP gene in the Klzwf1Δ mutant, we noticed not only that the chimeric G6PDH/GFP activities restored the growth defects of the mutant (data not shown), but also that cellular extracts prepared from these cultures showed the presence of G6PDH activity bands with reduced migrating properties, compared to the pattern of the null strain transformed with the KlZWF1 gene (Fig. 6, lanes 1 to 4). Therefore, the expression of this chimeric gene in the CBS2359 strain, harboring KlZWF1 on its chromosomal locus, could lead to the assembly of homo- and heterodimers or tetramers of the G6PDH and G6PDH/GFP activities. Indeed, cellular extracts from glycerol-grown cultures of CBS2359 harboring the KlZWF1GFP gene on a multicopy plasmid showed the presence of five different G6PDH bands of activity (Fig. 6, lanes 5 and 6) that, compared to the parental untransformed strain pattern (lane 8), were interpreted as homo- and heterotetramers of the KlZWF1 and KlZWF1GFP gene products (Fig. 6, lane 5, scheme). On the contrary, cellular extracts from ethanol cultures showed eight different G6PDH bands of activities in the transformed strain (Fig. 6, lanes 7, 10, and 11) and the expected two bands in the control strain (lane 9). The presence of three additional bands from ethanol-grown cells, compared to glycerol-grown cell extracts, could be assigned to the homodimers of G6PDH and G6PDH/GFP and to the heterodimer G6PDH-G6PDH/GFP (Fig. 5, lane 10, scheme). The amount of each G6PDH band, determined by densitometric analysis (Fig. 6, lanes 5 and 11), might reflect both the level of expression of the chromosomal and plasmidic gene but also the differential stability of the wild type compared to the chimeric homo- and hetero-oligomeric activities.

• Open in new tab
• Download powerpoint

Fig 6

Chimeric G6PDH/GFP activities separated on native gels from the Klzwf1Δ and CBS2359. The strains were transformed with centromeric (for Klzwf1Δ) and multicopy (CBS2359) plasmids containing the indicated genes. Lanes 5 and 10 depict the assembly of oligomers; lanes 5 and 7, which are identical to lanes 6 and 10, respectively, contain a 2× amount of cell extract. Cultures were grown on YP containing glycerol (G) or ethanol (E). Staining for G6PDH and numbering to the right of lanes 5 and 11 were as described for Fig. 2.

The faster-migrating G6PDH band acts as a marker of cytosolic accumulation of NADPH.In S. cerevisiae, it has been reported that pgi1 mutants are unable to grow on glucose because diversion of the glycolytic flux toward the PPP leads to a toxic accumulation of NADPH that, by inhibiting G6PDH, may explain the observed cell growth arrest (7, 23). Moreover, in Dicentrarchus labrax liver, it has also been reported that accumulation of NADPH in the cytosol may lead to inhibition of G6PDH activity (4).

Since the main function of G6PDH is to guarantee readily available NADPH in the cell, the G6PDH patterns were analyzed under stress conditions to test whether the faster-migrating band of K. lactis could represent a sign of cytosolic accumulation of NADPH and/or of G6PDH inhibition. Therefore, CBS2359 and MW179-1D cultures were grown in the presence of hydrogen peroxide and acetaldehyde to induce oxidative stress that would require the NADPH produced in the PPP for the glutathione S-transferase (GSH)/glutaredoxin- and thioredoxin-dependent neutralization systems (9, 14, 24, 38). As shown in Fig. 7, the faster-migrating G6PDH activity band of CBS2359 decreased in intensity in the presence of H₂O₂ (lane 1) or in the presence of increasing concentrations of acetaldehyde (lanes 3 to 5), compared to the unstressed control cells (lane 2). On the contrary, the same compounds increased the amount of the single lower-migrating G6PDH band in strain MW179-1D (Fig. 7, lane 6 and 8 to 10). The latter finding suggests that G6PDH expression increases to better respond to the increased NADPH demands, compared to the unstressed control cells (Fig. 7, lane 7). Together, these results provide evidence that the appearance of the faster-migrating G6PDH is a direct consequence of NADPH accumulation in the cell, and it disappears when NADPH is rapidly consumed during oxidative stress.

• Open in new tab
• Download powerpoint

Fig 7

G6PDH patterns from CBS2359 and MW179-1D extracts obtained from YPD cultures with and without acetaldehyde and H₂O₂ at the indicated concentrations. Cells growth, cells extracts, gel conditions, staining for G6PDH, and the meaning of the values in the box under the gel are as described for Fig. 2.

In order to assess if there is a direct link between NAD(P)H accumulation and the presence of two activity bands of G6PDH, we assayed the NAD(P)H content in cellular extracts from ethanol-grown cultures (two bands on the native gel) and from glycerol-grown cultures (a single G6PDH activity band on the native gel) in both the CBS2359 and MW179-1D strains. The amounts of the reduced cofactors in the extracts, determined for three independent samples under each growth condition, were performed by fluorescence measurements and showed in both strains an NAD(P)H content 1.60 to 1.76 times higher in the ethanol-grown compared to glycerol-grown cell extracts.