During stress, many organisms accumulate compatible solutes. These solutes must be eliminated upon return to optimal conditions as they inhibit cell metabolism and growth. In contrast, enzyme interactions optimize metabolism through mechanisms such as channeling of substrates. It was decided to test the (compatible solute) trehalose-mediated inhibition of some yeast glycolytic pathway enzymes known to associate and whether inhibition is prevented when enzymes are allowed to associate. Trehalose inhibited the isolated glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and hexokinase (HXK), but not aldolase (ALD) nor phosphoglycerate kinase (PGK). When these enzymes were mixed in pairs, both GAPDH and HXK were protected by either ALD or PGK acquiring the inhibition behavior of the resistant enzyme. GAPDH was not protected by HXK, albumin or lactate dehydrogenase (LDH). Also, ALD did not protect glucose 6-phosphate dehydrogenase (G6PDH), suggesting that protection is specific. In yeast cell extracts, fermentation was resistant to trehalose inhibition, suggesting all enzymes involved in the glucose-dependent production of ethanol were stabilized. It is suggested that during the yeast stress response, enzyme association protects some metabolic pathways against trehalose-mediated inhibition.
Whether enzyme association exists in vivo and what its physiological significance is, is a matter of debate (Vas & Batke, 1981; Ashmarina, 1984; Sukhodolets, 1987; Malhotra, 1995; Ovádi & Srere, 1996). When under stress, yeast accumulates nearly 0.5 M trehalose, increasing cytoplasmic viscosity (Hottiger, 1994; Singer & Lindquist, 1998). Trehalose accumulation has a dual effect, protecting, but also inhibiting, different enzymes (De Virgilio, 1994; Singer & Lindquist, 1998; Sampedro, 2002; Uribe & Sampedro, 2003; Sebollela, 2004; Esmann, 2008). If enzyme association helps enzymes to resist the trehalose-mediated inhibition, this would constitute an advantage during stress. The results would be useful for the understanding of survival in the presence of compatible solutes.
Compatible solutes stabilize the activity and structure of diverse proteins and membranes, probably through mechanisms that include trapping of specific water molecules at the protein surface and/or substitution of hydration-water (Crowe, 1984; Timasheff, 1993). In addition, the high viscosity promoted by crowding inhibits protein vibrations around the native structure (Kaushik & Bhat, 2003; Sampedro & Uribe, 2004). Viscosity may also be inhibitory to enzymes where the catalytic activity depends on large interdomain motions (Hottiger, 1994; Singer & Lindquist, 1998; Sampedro, 2002; Sampedro & Uribe, 2004; Faber-Barata & Sola-Penna, 2005; Esmann, 2008); for example, in the H+-ATPase from Kluyveromyces lactis, a correlation between viscosity and a decrease in Vmax and in catalytic efficiency (Vmax/S0.5) was detected (Sampedro, 2002; Uribe & Sampedro, 2003). The viscosity-dependent inhibitory effect of trehalose and glycerol has been confirmed in glutathione reductase and the Na+/K+-ATPase (Sebollela, 2004; Esmann, 2008). Avoiding high viscosity is perhaps the reason why yeasts rapidly eliminate compatible solutes when stress is over (Attfield, 1987): failure to do so would lead to decreased survival (Wera, 1999).
During stress, cells maintain a high flux rate in the metabolic pathways that provide the energy required for survival (Hounsa, 1998; Hoffmann & Holzhütter, 2009). Thus we thought it would be worth to explore whether compatible solute-mediated inhibition is prevented in mixtures of enzymes known to associate, for example glyceraldehyde-3-phosphate dehydrogenase (GAPDH) with aldolase (ALD) (Ashmarina, 1984; Malhotra, 1995) or with phosphoglycerate kinase (PGK) (Ovadi & Keleti, 1978; Batke & Tompa, 1986). Furthermore, enzyme association has been described throughout the glycolytic pathway (Green, 1965; Weber & Bernhard, 1982; Aman & Wang, 1986; Srere, 1987; Ovádi & Saks, 2004), in the Krebs cycle and in oxidative phosphorylation (Green, 1965; Robinson, 1987). Four isolated yeast glycolysis enzymes known to associate (Ovadi & Keleti, 1978; Ashmarina, 1984; Batke & Tompa, 1986; Malhotra, 1995) were tested for trehalose-mediated inhibition. It was observed that when assayed alone, GAPDH and hexokinase (HXK), but not ALD or PGK, were inhibited by trehalose. Interestingly, when mixed with either ALD or PGK, both GAPDH and HXK were protected against inhibition. In addition, in a yeast cell extract, fermentation was not inhibited by trehalose. From the above, it is suggested that enzyme association confers resistance to compatible solute-mediated inhibition. This resistance would constitute an advantage of enzyme association not identified to date.
Materials and methods
All the reagents were of analytical quality. Glucose, trehalose, HEPES, EDTA, 2-[N-morpholino] ethanesulfonic acid (MES), triethanolamine (TEA), MgCl2, glycerol, 2-mercapto-ethanol, sodium dodecyl sulfate (SDS), hydrazine hemisulfate salt, ammonium molybdate, l-ascorbic acid, dl-dithiothreitol (DTT), ALD, l-lactic dehydrogenase (LDH) from rabbit muscle, HXK, PGK, GAPDH, alcohol dehydrogenase (ADH) from baker's yeast, d-fructose 1,6 bis-phosphate sodium salt, d-glucose-6-phosphate (G6P) disodium salt, d-(−)-3-phosphoglyceric acid disodium salt, dl-glyceraldehyde-3-phosphate (GA3P), β-NADP+, NADH, NAD and ATP were acquired from Sigma Co. (St. Louis, MO). KOH, Mg2+ and K+ acetate, Na+ phosphate salts, H2SO4, H3PO4, trichloroacetic acid (TCA), NaOH and K2HPO4 were from J.T. Baker (Ecatepec, Mexico). Na+ acetate was from Merck (Darmstadt, Germany). Yeast extract was from Difco and peptone from Bioxon. The protease inhibitor cocktail (Complete) and glucose-6-phosphate dehydrogenase (G6PDH) baker's yeast were from Roche (Basel, Switzerland). Proalbumin was from Celliance (San Francisco, CA).
GAPDH activity (Bergmeyer, 1983) was measured in 20 mM HEPES, 1 mM MgCl2, 1 mM DTT and 1 mM EDTA, pH 7.5, in the presence of different trehalose concentrations (0, 0.2, 0.4, 0.6, 0.8 and 1.0 M), 0.81 mM NAD+, 1.33 mM Pi and 1.7 μM GAPDH. After 15 s, 50 μM GA3P was added. A340 nm was measured in an Aminco/Ollis spectrophotometer at 20 °C.
HXK activity (Bergmeyer, 1983) was measured in 40 mM MES-TEA, pH 7.6, plus different concentrations of trehalose (0, 0.2, 0.4, 0.6, 0.8 and 1 M), 50 mM glucose, 8 mM MgCl2, 0.64 mM ATP and 10 μg mL−1 HXK. After 10 min, 0.1 mL of 30% TCA was added and the samples were centrifuged for 20 min. The supernatants were neutralized with 1 mM NaOH as needed. The samples were mixed with 0.9 mM NADP+ and 0.55 U mL−1 G6PDH. The reaction was followed, measuring the A340 nm of NADPH in an Aminco-Olis DW 2,000 spectrophotometer at 20 °C.
ALD activity was measured in 10 mM sodium phosphate, 0.1 mM EDTA, pH 7.6, trehalose (0, 0.2, 0.4, 0.6, 0.8 and 1 M). Then 3.5 mM hydrazine hemisulfate, 0.1 mM fructose 1,6-diphosphate and 9 μg mL−1 ALD were added. Temperature was 20 °C. Absorbance was quantified in an Aminco-Ollis DW 2000 spectrophotometer. The concentration of hydrazone was measured at 240 nm using an extinction coefficient of 2.73 × 103 M−1 cm−1 (Sygusch & Beaudry, 1984).
PGK activity (Katewa & Katyare, 2003) was measured in 20 mM HEPES, 1 mM MgCl2, 1 mM DTT and 1 mM EDTA, pH 7.5, with different concentrations of trehalose (0, 0.2, 0.4, 0.6, 0.8 and 1 M), 0.2 mM ATP, 1 nM PGK and 6 mM d-(−)-3-phosphoglyceric acid. Samples were incubated for 10 min at 20 °C and TCA 30% was added, centrifuged for 20 min and neutralized with NaOH. Then 0.16 mM NADH and 1.7 μM GAPDH were added. Samples were incubated in a shaker bath (New Brunswick) for 1 h at 20 °C. The reaction was stopped with 5% SDS, leaving the samples on ice. To determine the Pi concentration, 0.5 mL 3 N H2SO4 and 0.1 mL 2.5% ammonium molybdate were added. Finally, hydrazine hemisulfate/ascorbic acid reducing reagent was added and incubated for 30 min. A820 nm was read in a 650 DU Beckman spectrophotometer. Results are μmol Pi mg−1 protein min−1.
G6PDH activity (Bergmeyer, 1983) was measured in 2 mM MES-TEA, pH 7.0, plus different concentrations of trehalose (0, 0.2, 0.4, 0.6, 0.8 and 1 M) with 0.9 mM NADP+, 0.55 U mL−1 G6PDH and, after 30 s, 24 mM G6P. A340 nm of NADPH was measured in an Aminco-Olis DW 2000 spectrophotometer, temperature 20 °C.
Yeast cell cytoplasmic extracts
An industrial strain of Saccharomyces cerevisiae (La Azteca, S.A.) was used. Cells were precultured in 50 mL YPD (1% yeast extract, 5% gelatin peptone, 5% dextrose) for 16 h and the preculture then mixed into 1000 mL YPD and further incubated for 10 h. Cells were harvested by centrifugation, resuspended in 1000 mL water and fasted under aeration at 3 L min−1 for 16 h. The cells were washed twice by centrifuging at 1020 g for 5 min in a GSA rotor (Sorval) and resuspending in sterile water. The last pellet was weighed and resuspended to 50% (w/v) in lysis buffer (100 mM HEPES/KOH, pH 7.5, 600 mM K-acetate, 10 mM Mg2+-acetate, 1 mM EDTA, 20% glycerol, 8 mM 2-mercapto-ethanol and a tablet of ‘Complete’). Cells were disrupted in a Bead-Beater (Biospec Products) with 0.5-mm-diameter glass beads for 20 s × 3. Finally the homogenate was centrifuged at 24 500 g in a 60 TI rotor and an Optima XLK100 Beckman-Coulter ultracentrifuge for 1 h at 4 °C and the supernatant was saved and kept at −80 °C (Lebedeva, 2006).
Glucose-supported production of ethanol
Fermentation by cell cytoplasmic extracts was measured in 0.1 M MES-TEA, pH 7.0, 20 mM glucose and 0, 0.2, 0.4, 0.6, 0.8 or 1 M trehalose. Protein was 11.75 mg mL−1. Samples were incubated at 20 °C for different periods (0, 1, 2, 4, 8 and 16 min). The reaction was terminated with 0.1 mL of 30% TCA and was centrifuged at 200 g for 25 min in a clinical centrifuge. The supernatant was neutralized with NaOH as needed. Ethanol was then quantified spectrophotometrically at 340 nm (Aminco/Ollis DW 2000) in a mixture containing a 0.05-mL aliquot of the sample with 2 mL of 114 mM K2HPO4, pH 7.6, 1.8 mM NAD+. A340 nm was measured and, after 10 s, ADH 30 μg mL−1 (9420 U mL−1) was added at 20 °C.
Trehalose inhibited some, but not all, of the isolated glycolysis enzymes tested
During stress, yeasts use large quantities of ATP to maintain their ion gradients (Attfield, 1987; Hounsa, 1998; Sampedro, 1998; Singer & Lindquist, 1998; Hoffmann & Holzhütter, 2009). In yeast, a major source of ATP is glycolysis, which therefore should be very active during stress, even in the presence of the high trehalose concentrations observed (Hottiger, 1994). The effect of different concentrations of trehalose on four different isolated glycolysis enzymes was evaluated. It was observed that two enzymes were inhibited by trehalose, whereas two were resistant to inhibition, as follows (Fig. 1): HXK and GAPDH were inhibited in proportion to the concentration of trehalose. In contrast, ALD was not inhibited, whereas for PGK the effects were dual: PGK was stimulated at 0.2–0.6 M trehalose, whereas it was inhibited at 0.8 and 1.0 M trehalose (Fig. 1).
The pattern of trehalose inhibition changed when tested in enzyme mixtures
The trehalose-mediated inhibition observed for HXK and GAPDH (Fig. 1) indicated that the fermentation pathway should be inhibited by trehalose unless there is a mechanism to resist inhibition. One such mechanism could be the association of glycolysis enzymes, which in some cases has resulted in resistance against denaturing (Raïs, 2000). Thus, it was decided to test whether mixing enzymes in pairs would result in resistance against the trehalose-mediated inhibition. Each of the two trehalose-sensitive enzymes was mixed with each of the resistant enzymes and tested for trehalose-mediated inhibition. For GAPDH, association to either ALD or PGK resulted in decreased trehalose effects (Fig. 2). In the GAPDH/ALD mixture, the first obvious effect was that GAPDH activity increased over 20 times as compared with the isolated enzyme (Fig. 2a). Then, when in the presence of increasing trehalose concentrations, only at 0.4 M trehalose and higher was a slight inhibition of GAPDH observed (Fig. 2a). To facilitate comparison, the isolated ALD activity data are also included in the figure (Fig. 2a) and it was observed that the pattern of GAPDH inactivation changed to resemble the pattern of ALD inactivation. In the GAPDH/PGK mixture (Fig. 2b) GAPDH activity increased only slightly as compared with the isolated enzyme. However, in the presence of trehalose, the GAPDH activity was stabilized, as follows: from 0.2 to 0.6 M trehalose a small increase in activity was observed, whereas at 0.8 and 1.0 M trehalose, inhibition was observed (Fig. 2b). To facilitate comparison, the activity of the isolated PGK was included in the figure and it was observed that the GAPDH inactivation pattern changed to resemble that of PGK.
ALD, GAPDH and PGK are very close in the middle of the glycolytic sequence, thus it was of interest to determine whether ALD and/or PGK conferred any protection to HXK, which catalyzes the first reaction of the glycolytic pathway. In the HXK/ALD mixture, the HXK activity became resistant to inhibition at all the trehalose concentrations tested (Fig. 2c), i.e. its inhibition pattern became similar to isolated ALD (also included) (Fig. 2c). When the HXK/PGK mixture was tested, HXK activity was also resistant to trehalose except at the highest concentration, where a slight inhibition was observed (Fig. 2d), i.e. HXK resistance to inhibition became similar to that of PGK (Fig. 2d). Thus, when the trehalose-sensitive enzymes GAPDH or HXK were mixed with either of the resistant enzymes ALD or PGK, both became resistant to inhibition by trehalose, opening the possibility that a specific association results in activity protection.
The GAPDH sensitivity to trehalose was not modified by proteins other than ALD or PGK
To define whether the PGK- or ALD-promoted protection of GAPDH was specific, we mixed GAPDH with other proteins, namely albumin, HXK and LDH, at concentrations similar to those used for PGK or ALD (1 nM in all cases). None of these proteins protected GAPDH against trehalose (Fig. 3). This suggests that the interaction of GAPDH with either PGK or ALD is specific and it may be speculated that this interaction might occur in vivo and perhaps play a physiological role protecting glycolytic activity during stress.
G6PDH, an enzyme from a metabolic pathway different to glycolysis is inhibited by trehalose and is not protected by ALD
To test whether the ALD-mediated protection against trehalose was specific for glycolytic pathway enzymes, we decided to analyze the trehalose effects on an enzyme from the pentose-phosphate pathway. It was observed that G6PDH was inhibited by all the trehalose concentrations tested (Fig. 4). Then, when G6PDH was mixed with ALD the same pattern of inhibition was observed, i.e. ALD did not protect G6PDH (Fig. 4), suggesting that the ALD-mediated protection does exhibit some degree of specificity.
Inhibition of the complete fermentation pathway occurs only at very high trehalose concentrations
Our data on the trehalose-mediated inhibition of some isolated glycolysis enzymes suggested that some enzymes are sensitive to inhibition, while others are resistant. Then, when mixed with the resistant enzymes, the sensitive enzymes acquire resistance. To define whether this protection does play a physiological role, it was decided to measure the glucose-supported rate of O2 consumption in intact yeast. In yeast cultures subjected to stress, it is known that trehalose concentrations reach 0.5 M (Hottiger, 1994) and thus, the sensitive enzymes should be inhibited unless they are protected. In our hands the rate of oxygen consumption was the same in cells subjected to stress and controls (results not shown). This result suggested that in vivo the accumulation of trehalose resulting from stress did not affect the glycolytic/Krebs/oxidative phosphorylation pathways.
To further explore this lack of effect, we decided to measure the effect of trehalose on fermentation in a cytoplasmic extract from yeast. In this system, we were able to manipulate the concentration of trehalose while we measured the glucose-dependent production of ethanol. It was observed that in the absence of trehalose, ethanol was produced rapidly during the first 4 min and then stabilized, reaching a concentration of 0.025 μmol mg−1 protein at 8–15 min (Fig. 5). When 0.2–0.6 M trehalose was included in the assay mixture, the initial production of ethanol was slightly less than in the control but eventually reached comparable concentrations (Fig. 5). However, in the presence of the highest concentrations of trehalose tested, a strong (0.8 M trehalose) to complete (1 M trehalose) inhibition of fermentation was observed (Fig. 5). In the cytoplasmic mixture, all the glycolysis enzymes were active at up to 0.6 M trehalose, suggesting that any sensitive enzyme was stabilized in the mixture.
Enzyme association within the cytoplasm has long been proposed by the classical studies of Green (1965), Clegg (1964) and Fulton (1982). In the mitochondrial matrix, association of the Krebs cycle enzymes into a metabolon was reported by Paul Srere (1987). In some organisms, notably trypanosomes the glycolytic pathway is enclosed within organelles called glycosomes (Aman & Wang, 1986). A specific enzyme organization would probably minimize side reactions, making metabolism more efficient (Ellis, 2001). Also, association/dissociation may control metabolic activity (Vas & Batke, 1981; Ashmarina, 1984; Sukhodolets, 1987; Malhotra, 1995; Ovádi & Srere, 1996). Here, we explored another possible reason for enzymes from a given specific pathway to associate: the protection that association might provide against compatible solute-mediated inhibition: enzyme association of the whole glycolytic pathway would allow the cell to maintain a high-energy output during stress. This is suggested by the lack of inhibition of oxidative phosphorylation in intact cells (not shown) and of fermentation in cytoplasmic extracts (Fig. 5). It was interesting to see that the sensitive enzyme acquired the same resistance as the enzyme to which it was associated, i.e. both GAPDH and HXK became as resistant as ALD or PGK depending on the mixture.
During stress, yeasts accumulate trehalose and/or glycerol; then as soon as stress is over, the compatible solutes are rapidly eliminated (Singer & Lindquist, 1998). Failure to either produce (during stress) or eliminate (after stress) trehalose, results in decreased survival (Wera, 1999). It has been speculated that during stress, trehalose protects a number of isolated enzymes from damage (Kaushik & Bhat, 2003), whereas it inhibits other enzymes (Hottiger, 1994; Sampedro & Uribe, 2004; Sebollela, 2004; Esmann, 2008). It would be very interesting to analyze whether some metabolic pathways other than glycolysis are sensitive to compatible solutes. One might imagine that for sensitive enzymes that do not associate (anabolic reactions?), compatible solutes would need to be eliminated as soon as the stress situation is over for cells to survive and reproduce (Wera, 1999). Determining whether there are pathways where enzymes fail to associate would provide further proof that enzyme association is important. Acquisition of compatible solute resistance from the isolated to the in situ condition would probably prove to be a useful tool in discriminating which enzymes associate in situ. At any rate, inside the cell, the high sensitivity of some enzymes to compatible solutes would probably explain why cells need to eliminate the solute in order to survive.
Some glycolysis enzymes seem to perform more than one activity in the cell. It has been reasoned that being the most ancient energy production system, glycolysis evolved early and its enzymes both optimized and diversified their functions (Kim & Dang, 2005). Among these, ALD is remarkable in its ability to interact with other proteins both from within the glycolytic cycle and from the cytoskeleton (Keleti, 1989; Walsh, 1989; Vértessy, 1997). Its ability to protect other enzymes was already illustrated by the reversal of PFK inactivation observed when this enzyme is associated to a mixture of ALD with actin microfilaments (Raïs, 2000). Together with GAPDH (Beth, 1986), ALD binds to band 3 in the erythrocyte membrane (Strapazon & Steck, 1976; Uribe, 1990). In the case of ALD, the addition of the substrate, 1,6-diphosphofructose or the competitive inhibitor butanediol-1,4-diphosphate modifies the conformation of the enzyme, releasing it from its binding site in band 3 (Uribe, 1990).
The possibility that both ALD and/or PGK might constitute the glycolytic scaffold, i.e. a structure where the other glycolysis enzymes bind, has been proposed before (Walsh, 1989; Vértessy, 1997; Raïs, 2000; Campanella, 2005). The sensitivity of diverse enzymes to the mild, reversible compatible solute-mediated inhibition reported here might provide a tool to explore this possibility. At least for trehalose, sucrose and glycerol, the mechanism of enzyme activity inhibition has been proposed to be the increase in viscosity that compatible solutes generate (Sampedro, 1998, 2002). If the basal viscosity of the crowded cytoplasm is considered (Ellis, 2001), then after compatible solute accumulation, viscosity is probably very high and it probably inhibits enzyme motility (Sampedro & Uribe, 2004). Thus, it is interesting to speculate whether inhibition-resistant enzymes are rigid, and therefore would not have to suffer large conformational (viscosity-sensitive) changes during their catalytic cycle. In this scenario, very motile enzymes undergoing conformational changes during catalysis would be very sensitive to viscosity (Sampedro & Uribe, 2004). In addition, it might also be speculated that protection would occur by stiffening the highly motile-sensitive enzymes through their association with rigid enzymes, without forcibly inhibiting catalysis (Singer & Lindquist, 1998; Sampedro, 2002; Esmann, 2008). Indeed, in the ALD/GAPDH mixture, the activity of GAPDH increased about 20-fold (Fig. 2a) and became as resistant to inhibition as ALD.
This study was partially funded by CONACYT grant 79989 and DGAPA-UNAM grant IN217109-3. D.A.O. is a CONACYT fellow enrolled in the Biochemical Science PhD Program at UNAM. We thank Martha Calahorra, Ramón Méndez, Rocío Romualdo and Norma S. Sánchez for their expert technical help.