Methylglyoxal synthetase, enol-pyruvaldehyde, glutathione and the glyoxalase system

Article abstract of DOI:10.1021/ja027065h

enol-Pyruvaldehyde (ePY or 2-hydroxypropenal, O=C(H)-C(OH)=CH₂) a transient intermediate in the alkaline decomposition of the triosephosphates to methylglyoxal is now observed by UV and ¹H NMR spectroscopy as the immediate product of the methylglyoxal synthetase (MGS) reaction: dihydroxyacetone-P → P_i + ePY → methylglyoxal (MG). Analysis of ePY formed from 1-¹³C- and (1R, 3S)-[1,3-²H]-DHAP establishes the stereochemical course of its formation by MGS. Its rate of ketonization is much too slow to be in the sequence required for the assay of MGS by coupling of the MG produced to glyoxalase I (Glx I): MG + glutathione (GSH) → (S)-lactylglutathione (D-LG). Instead, ketonization occurs by way of the hemithioacetal (HTA) formed between ePY and GSH, and could be either an enzymatic function of Glx I or occur nonenzymatically at an activated rate. Enzymatic ketonization was ruled out because the methyl group of D-LG formed from specifically labeled ePY is achiral. Chemical ketonization of ePY is activated by general bases, such as acetate, and by thiols such as GSH and 2-mercaptoethanol, which disrupt its stabilizing double bond conjugation as hemithioacetal (HTA) adducts. 2-Mercaptoacetate combines both functions, acting as the HTA adduct of ePY with the appended carboxylate group presumably positioned to promote abstraction of the enol proton and protonation of the enolate carbon at an accelerated rate. In the MGS-Glx I system (dihydroxyacetone-P → ePY, ePY + GSH → GS-ePY, GS-ePY → GS-MG, GS-MG → D-LG), the nonenzymatic 2nd and 3rd steps describe the catalytic role of GSH in the critical ketonization process and set the stage for its participation in the glyoxalase system.

Full text of DOI:10.1021/ja027065h

Published on Web 10/12/2002
Methylglyoxal Synthetase, Enol-Pyruvaldehyde, Glutathione
and the Glyoxalase System
Irwin A. Rose*,† and James S. Nowick‡
Contribution from the Department of Physiology and Biophysics and Department of Chemistry,
UniVersity of California, IrVine, IrVine, California 92717-1700
Received May 27, 2002. Revised Manuscript Received August 30, 2002
Abstract: enol-Pyruvaldehyde (ePY or 2-hydroxypropenal, OdC(H)-C(OH)dCH2) a transient intermediate
in the alkaline decomposition of the triosephosphates to methylglyoxal is now observed by UV and 1H
NMR spectroscopy as the immediate product of the methylglyoxal synthetase (MGS) reaction: dihydroxy-
acetone-P f Pi + ePY f methylglyoxal (MG). Analysis of ePY formed from 1-13C- and (1R, 3S) -[1,3-
2H]-DHAP establishes the stereochemical course of its formation by MGS. Its rate of ketonization is much
too slow to be in the sequence required for the assay of MGS by coupling of the MG produced to glyoxalase
I (Glx I): MG + glutathione (GSH) f (S)-lactylglutathione (D-LG). Instead, ketonization occurs by way of
the hemithioacetal (HTA) formed between ePY and GSH, and could be either an enzymatic function of Glx
I or occur nonenzymatically at an activated rate. Enzymatic ketonization was ruled out because the methyl
group of D-LG formed from specifically labeled ePY is achiral. Chemical ketonization of ePY is activated
by general bases, such as acetate, and by thiols such as GSH and 2-mercaptoethanol, which disrupt its
stabilizing double bond conjugation as hemithioacetal (HTA) adducts. 2-Mercaptoacetate combines both
functions, acting as the HTA adduct of ePY with the appended carboxylate group presumably positioned
to promote abstraction of the enol proton and protonation of the enolate carbon at an accelerated rate. In
the MGS-Glx I system (dihydroxyacetone-P f ePY, ePY + GSH f GS-ePY, GS-ePY f GS-MG, GS-MG
f D-LG), the nonenzymatic 2nd and 3rd steps describe the catalytic role of GSH in the critical ketonization
process and set the stage for its participation in the glyoxalase system.
Introduction
cells not known to contain MGS. The toxicity of MG from
carbohydrate sources in E. coli depends on the content of
The “glyoxalase system” refers to the two enzymes, glyox-
alase I and II, that together with glutathione remove the cytotoxic
keto aldehyde methylglyoxal from most cells: MG + GSH f
D-LG f D-lactate + GSH. MG is a readily permeable product
of unicellular organisms that contain MGS, but is formed in all
cells under physiological conditions due to the inherent instabil-
ity of their triose-P metabolites,1 imperfections in enzymes that
modify them,2,3 and from sources yet to be identified.4 Most of
the MG of cells resides in long-lived amine complexes of
proteins4 and possibly nucleic acids,5 which may explain its
toxicity,6 which in E. coli depends on the growth conditions
even in high MGS expressing cells.7 That inhibitors of glyox-
alase I inhibit tumor growth8 is evidence for MG toxicity in
glyoxalase I, notwithstanding the presence of a glutathione-
independent glyoxalase III and NADH-dependent aldehyde-
reducing enzymes.9
The nonenzymatic and the MGS dependent decomposition
of dihydroxyacetone-P (DHAP) have a common chemistry
(Scheme 1), with the exception that the enzyme is stereospecific
for the pro-S proton of C-1 of DHAP, whereas both are
nonstereospecific in the formation of the methyl group of MG.10
This latter observation suggested that the immediate product
of the MGS reaction was the enol form of MG, which then
ketonized in solution. Harrison and co-workers11-13 have pro-
vided extensive kinetic and structural studies of E. coli MGS,
complexed with analogues of the proposed enediol-P intermedi-
ate, showing Asp-71, the proton-abstracting base, at the 1si face
of the intermediate and the bridge oxygen to PO3 to be out of
the plane, a geometry required for the â-elimination of -OPO3
in the MGS reaction sequence (Scheme 1). On the basis of the
geometry of the bound inhibitors and the positions of neighbor-
ing protein residues the enediol-P intermediate is best interpreted
* To whom inquiries may be addressed. Tel: 949-824-5857. E-mail:
zerose@net-star.net.
† Department of Physiology and Biophysics.
‡ Department of Chemistry.
(1) Richard, J. J. Am. Chem. Soc., 1984, 106, 4926-4936.
(2) Webb, M. R.; Sandring, D. N.; Knowles, J. R. Biochem. 1977, 16, 2738-
2741.
(3) Richard, J. R. Biochem. 1991, 30, 4581-4585.
(4) Chaplen, F. W. R.; Fahl, W. E.; Cameron, D. C. Proc. Nat’l Acad. Sci.
(USA) 1998, 95, 5533-5538.
(9) MacLean, M. J.; Ness, L. S.; Ferguson, G. P.; Booth, I. R. Mol.
Microbiology 1998, 27, 563-571.
(10) Summers, M. C.; Rose, I. A. J. Am. Chem. Soc. 1977, 99, 4475-4478.
(11) Saadat, D.; Harrison, D. H. T. Biochem. 1998, 37, 10 074-10 086.
(12) Saadat, D.; Harrison, D. H. T. Biochem. 2000, 39, 2950-2960.
(13) Marks, G. T.; Harris, T. K.; Massiah, M. A.; Mildvan, A. S.; Harrison, D.
H. T. Biochem. 2001, 40, 6805-6818.
(5) Krymkiewicz, N. FEBS Lett. 1973, 29, 51-54.
(6) Papoulis, A.; al-Abed, Y.; Bucala, R. Biochem. 1995, 34, 648-655.
(7) Totemeyer, S.; Booth, N. A.; Nichols, W. W.; Dunbar, B.; Booth, I. R.
Mol. Microbiol. 1998, 27, 553-562.
(8) Kavarana, M. J.; Kovaleva, E. G.; Creighton, D. J.; Wollman, M. B.;
Eiseman, J. L. J. Med. Chem. 1999, 42, 221-228.
9
10.1021/ja027065h CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 13047-13052
13047

A R T I C L E S
Rose and Nowick
Scheme 1. Methylglyoxal synthetase (MGS) reaction
as a cis-enediol.13 Structural comparisons with trioseP isomerase,
which has a minor MG synthetase side reaction2,3 are instructive.
Surprisingly, the properties of 2-hydroxypropenal seem not
to have been described.14 Because many enols are known to
ketonize at measurable rates,15,16 the availability of recombinant
MGS suggested a simple method for its possible preparation.
Of interest would be the following: (1) a direct demonstration
of ePY as the product of MGS, as would be shown by its
accumulation in a solution of the enzyme and DHAP; (2) a study
of the stereochemistry of its formation by MGS; (3) a study of
the mechanisms of its ketonization; and (4) the possibility that
the introduction of GSH into the glyoxalase system evolved
from a role in the ketonization step.
Figure 1. Absorbance changes upon mixing DHAP and MGS. MGS (36
U) was mixed into 1 mM DHAP, pH 7.5 in H2O (open circles), D2O (filled
circles), or D2O with 0.6 mM Pi (filled squares). The 245 nm absorbance
was recorded at 10 s. intervals.
DHAP with MGS in D2O, or of labeled FDP with aldolase and MGS
in D2O and stored in dry ice or liquid nitrogen, if necessary, prior to
use. 3-13C FDP was used to generate 1-13C ePY. (1S)-[1,3-2H] FDP,
used to generate [1,3-2H]-ePY, was prepared from glucose-d7. NMR
spectra were recorded at 500 MHz on a Bruker instrument at 5°. DSS
(0.03 mM) was included as an internal standard.
Materials and Methods
Materials. The enzymes, MGS from E. coli, the wild-type recom-
binant form described by Saadat and Harrison11 was generously
provided by Dr. David Harrison. The enzyme is a homohexamer, with
kcat ) 220 s-1, Km of DHAP ) 0.2 mM, when assayed at pH 7 with 15
mM GSH and excess Glx I by following the increase in absorbance
due to the final product, D-LG. Methylmalonyl-CoA:pyruvate trans-
carboxylase, was a generous gift of Prof. Harland Wood. The fumarase
of E. coli, prepared in this lab, was made available as a plasmid by Dr.
Mark Donnelly, Argonne National Laboratory. Other enzymes and
chemicals were purchased from Sigma.
To determine the stereochemical course of the formation of lactoyl
glutathione, the D-lactate, derived by treatment with Glx II, was
converted to pyruvate in good yield at pH 7.5 by incubation with NADH
(5 mM), D-lactate dehydrogenase (1 mg), phenazine methosulfate (0.2
mM) and catalase for 30 min at 37° with shaking in the dark.19 Catalase
was included to minimize the accumulation of H2O2 with loss of
pyruvate. The isolated pyruvate was converted to malate using
transcarboxylase, methylmalonyl-CoA, NADH, and malate dehydro-
genase as described previously.20
All enzymatic studies were done at 25 °C.
Methods. (1S)-2H,3H-Fructose-1,6-P2 (FDP) was prepared from
1-3H-glucose (Sigma), hexokinase, phosphoglucose isomerase, phos-
phofructokinase, and pyruvate kinase in D2O with slightly less than 2
equiv. of P-enolpyruvate. The formation of pyruvate was followed in
periodic samples using lactate dehydrogenase and NADH.17 (1S)-H-
FDP-d6 from glucose-d7 (Merck of Canada) and 3-13C-FDP from
3-13C-glucose (Cambridge Isotopes) were prepared in the same way
in H2O. In the former case, addition of phosphofructokinase was delayed
to allow the deuterium at C-2 of the glucose-6-P to fully exchange
with solvent.18 FDP, pyruvate, and malate were purified on Dowex-
1-Cl- columns using 0.2, 0.02, and 0.01 N HCl, respectively, for elution
in about 3, 5, and 4 column volumes, respectively. DHAP was assayed
with glycerol-P dehydrogenase and NADH. FDP was assayed by the
further addition of FDP aldolase. When the isolated FDP was used to
generate the ePY for NMR studies in D2O it was treated with activated
charcoal to remove contaminating ATP, reabsorbed on a short Dowex
1-Cl- column which was next exchanged with D2O before the FDP
was eluted with 0.2 N DCl. The pH was carefully adjusted to ∼6 with
dilute NaOH in D2O for storage at -10 °C until conversion to ePY.
Abbreviations. (Bis-Tris) 2-bis {hydroxyethyl}amino-2-{hydroxy-
methyl}1,3-propanediol, (DHAP) dihydroxyacetone-P, (D-LG) D-
lactoyl glutathione, (DSS) sodium 2,2,-dimethyl-1-silapentane-5-sul-
fonate, (ePY) enol-pyruvaldehyde, (eDP) cis-enediol-3-P, (FDP) fructose-
1,6 bisphosphate, (Glx) glyoxalase, (GSH) glutathione, (GS-ePY) HTA-
adduct of GSH and ePY, (GS-MG) HTA-adduct of GSH and MG,
(HTA-) hemithioacetal, (MG) methylglyoxal, (MGH) methylglyoxal
hydrate, (MGS) MG synthetase, (PEP) phosphoenolpyruvate, (2PG)
2-phosphoglycolate.
Results
Demonstration of an Unstable Product of the MGS
Reaction. To search for an enol product, the MGS reaction was
monitored in the UV, at 245 nm.21 A large amount of enzyme
was used so that a transient product with a signal intensity
expected of an enol could be detected if its decay rate were not
much greater than 20% per sec. Indeed, a rapid increase in
absorbance was observed followed by its decay with t1/2 ) 10
s in H2O and 60 s in D2O, Figure 1. In the presence of Pi, a
strong inhibitor of MGS,22 the synthetic phase was greatly
slowed. A prolonged period during which synthesis and decay
of the intermediate are equal in rate was followed by net decay
1
Samples of ePY for H NMR studies were generated by reaction of
(14) Hart, H. Chem. ReV. 1979, 79, 515-529 reports, footnote 36, a private
communication from J. C. Speck, Jr. of the appearance of a 250 nm
absorbing material from the distillate of an acid solution of glyceraldehyde
which produced pyruvaldehyde (MG) under unspecified conditions.
(15) Capon, B.; Guo, B.-Z.; Kwok, F. C.; Siddhanta, A. K.; Zucco, C. Acc.
Chem. Res. 1988, 21, 135-140.
(19) Willard, J. M.; Rose, I. A. Biochem. 1973, 12, 5241-5246.
(20) Kuo, D. J.; Rose, I. A. J. Am. Chem. Soc. 1982, 104, 3235-3236.
(21) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. In Spectrometric
Identification of Organic Compounds, 5th ed.; J. Wiley & Sons: New York,
1991, 302.
(16) Rapporport, Z. The Chemistry of Enols; J. Wiley & Sons: New York, 1990.
(17) Cohn, M.; Pearson, J. E.; O’Connell, E. L.; Rose, I. A. J. Am. Chem. Soc.
1970, 92, 4095-4098.
(18) Rose, I. A.; O’Connell, E. L. J. Biol. Chem. 1961, 236, 3086-3092.
(22) Hooper, D. J.; Cooper, R. A. Biochem J. 1972, 128, 321-329.
9
13048 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

enol-Pyruvaldehyde: Chemistry and Enzymology
A R T I C L E S
Scheme 2. Stereochemical course of the MGS reaction
synthesis starting with perdeutero-glucose-6-P in H2O).18,27 If
the specifically labeled 3-1H enediol-P intermediate (Scheme
1) were the immediate product of MGS and rapidly decomposed
to the enolaldehyde (k ) 106 s -1 in solution)1, one would expect
1
the C3 H of ePY to be randomly located and the pairs of
doublets (Figure 2A) would appear as two singlets. If the
â-elimination step occurs on the enzyme, only one of the singlets
1
should be H labeled. As observed (Figure 2D) the downfield
Figure 2. 1H NMR spectra of enol-pyruvaldehyde (ePY), methylglyoxal
hydrate (MGH) and labeled ePY. (A) ePY generated from the reaction of
DHAP and MGS; (B) MGH and ePY, formed after incubation of the (above)
sample of ePY at ca. 40 °C for several minutes; (C) 1-13C ePY generated
from the reaction of 1-13C DHAP and MGS; (D) [1,3-2H]-ePY generated
from (1R),(3S)-[1,3-2H]-DHAP and MGS.
doublet has disappeared, is 2H, and the upfield doublet is seen
as a singlet. With further incubation at 25° in D2O, this peak
decreased in intensity. This decrease establishes that this peak
does not derive from the C1-hydrogen of the MGH that would
form; this hydrogen would not give a signal, having been derived
from the C3-deuterium of glucose-d7. By using glucose-d7 for
the synthesis of DHAP the formation of protio-glyceraldehyde-
3-P in the aldolase reaction is avoided. The chemical decom-
position of glyceraldehyde-3-P to MG in D2O leads to perdeutero-
MGH. Therefore, the shoulder at the upfield doublet position,
shown in Figure 2B, would not be expected to increase with
the formation of MGH.
The combination of experiments C and D establishes that
the hydrogen that is cis to the aldehyde of ePY is derived from
the pro-S hydrogen of DHAP, as shown in Scheme 2. This
requires that in the conversion of the cis-enediol-P to ePY,
expulsion of monophosphate must occur from the (1si,2re)-face
of the molecule (Scheme 2). This is the same face from which
Asp 71 must abstract the C1-proton of DHAP.
as the substrate is largely consumed,23 presumably by its
ketonization to MG, which does not absorb significantly at this
wavelength.
Spectral Characterization of the Transient Product. To
determine its spectrum in the ultraviolet, the transient product
of MGS was generated under conditions of slow ketonization
(t1/2 ≈ 27 min in 50 mM Bis Tris buffer, pH 5.8, in D2O) with
a known amount of DHAP. 2-Phosphoglycolate (2PG), 0.3 mM,
a strong inhibitor of MGS,11 was added to terminate synthesis
within 15 s. The absorbance from 225 to 275 nm was recorded,
a sample taken for DHAP determination, and the absorbance
was allowed to decay completely. From the changes in absor-
bance and the measured decrease in DHAP, a peak molar
absorbance of 5000-5400 was determined at pH 5.8 and ∼6400
at pH 7.4. The λmax found, 245 nm, agrees with prediction based
on empirical rules for related R,â-unsaturated carbonyl com-
pounds.21
pH and Buffer Effects on the Ketonization Rate of ePY.
To examine the influence of pH and buffer on the ketonization
rate, a small volume of 2PG was added to quench the enzyme
after a brief synthetic period. The reaction conditions were then
adjusted as desired, and the subsequent decline in absorbance
was monitored in the absence of enzymatic synthesis. Rates of
absorbance change, 0.69/ t1/2, were linear for at least 3 half-
The transient intermediate formed from DHAP was character-
ized by proton NMR under conditions of relative stability (pH
5.8-6.0 and 5° C, t1/2 ) 90 min in D2O), Figure 2A. A peak
for the aldehyde proton at 9.22 ppm and two doublets (2JHH
)
lives. A low buffer-independent rate, ∼3 × 10-4 -1
s , is observed
2.3 Hz) for the geminal hydrogens were observed centered at
5.63 and 5.27 ppm. Coupling of the aldehyde proton to the vinyl
protons is too weak to be seen. A shoulder at the upfield doublet
that grows with time, Figure 2B, is derived from the C1 proton
of MG aldehyde hydrate.24 The transient proton spectrum formed
from 1-13C DHAP generated from 3-13C FDP (Figure 2C) is
also consistent with ePy. The aldehyde proton is split by the
aldehyde 13C nucleus (1JCH ) 183.9 Hz). The 2-fold greater
coupling constant (3JCH ) 12.5 Hz vs 6.1 Hz) establishes the
downfield proton as trans to the aldehyde group.25,26
from pH 2-7. Higher rates are seen at high pH indicating
specific OH- catalysis: kOH- ) 0.05 s-1 at pH 10. Carboxylates,
piperizine alkylsufonates, and Pi provide a linear Brønsted plot
(Figure 3) consistent with general base catalysis with â ≈ 0.8,
similar to the value for ketonization of vinyl alcohol, 0.77.29
However, ketonization of ePY is about 75-fold slower than that
of vinyl alcohol under the same conditions, probably due to
the double bond conjugation of the ground state of the former.
-
HCO3 and Bis-Tris are significantly less effective catalysts
than predicted from their basicity and are therefore preferred
for use as buffers.
Coupling of MGS and Glyoxalase I. MGS is usually
assayed by coupling with Glx I and GSH by following the
increase at 240 nm due to D-LG. The D2O effect of the coupled
To examine the stereochemistry of the MGS reaction, the
DHAP was labeled with 1H at the C3 pro-R position (by virtue
of the pro-R specificity of phosphoglucose isomerase in the
(23) The MGS rate becomes more strongly inhibited as Pi increases and DHAP
decreases as a consequence of sigmoidal kinetics,10,22 to such an extent
that it becomes very difficult to run the reaction to completion.
(24) Creighton, D. J.; Migliorini, M.; Pourmotabbed, T.; Guha, M. K. Biochem.
1988, 27, 7376-7384.
(25) Vogeli, U.; von Philipsborn, W. Org. Magn. Reson. 1975, 7, 617-627.
(26) Brugel, W.; Ankel, Th.; Kruckenberg, F. Z. Electrochem. 1960, 64, 1121-
1155.
(27) Johnson, C. K.; Gabe, E. J.; Taylor, M. R.; Rose, I. A. J. Am. Chem. Soc.
1965, 87, 1802-1804.
(28) Designation of the faces at a trigonal carbon is according to the system of
Hanson, K. R. J. Am. Chem. Soc. 1966, 88, 2731-2742.
(29) Capon, B.; Zucco, C. J. Am. Chem. Soc. l982, 109, 7567-7572.
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13049

A R T I C L E S
Rose and Nowick
Scheme 3. Thiol-catalyzed tautomerization of ePy
the loss of the 245 nm absorbance following the addition of
GSH to newly generated ePY the Glx I reaction was detected
without a lag and to the same extent found if time were allowed
for MG to be formed spontaneously, before addition of the GSH
and Glx I. Therefore, the loss of absorbance found upon addition
of GSH to ePY cannot be the result of a Michael addition. For
the formation of a hemithioacetal adduct of ePY, per se, to be
responsible for the loss of the ultraviolet absorbance, the assay
without a lag suggests that either Glx I catalyzes the ketonization
of the adduct as well as its well-known isomerization reaction,
i.e., GS-ePY f GS-MG f D-LG, or else the ketonization of
the adduct is nonenzymatic and sufficiently fast for coupling
to Glx I.
Figure 3. Brønsted plot. 1. lactate pKa 3.86, 2. acetate pKa 4.76, 3. succinate
pKa 5.59, 4. HCO3- pKa 6.35, 5. Bis-Tris pKa 6.5, 6. PIPES pKa 6.8, 7. Pi
pKa 7.2, and 8. HEPES pK 7.5, where kB- ) 0.69/t1/2‚x(buffer concentra-
tion) x‚fB- and fB- is the fraction of buffer in the basic form.
If the HTA adduct of GSH with stereospecifically formed
ePY were ketonized by action of Glx I, then one might expect
to be able to generate D-LG with an isotopically chiral CH3
group. To examine this possibility, specifically labeled 3-2H,3H-
DHAP was generated from labeled FDP10 and aldolase, and
converted to D-LG by coupling MGS and excess Glx I in H2O.
3
Chirality in the Glx I reaction should result in unequal H
labeling of the geminal protons of malate, a consequence of
isotope discrimination of the transcarboxylase reaction.20,31 The
3
results, based on labilization of H by reaction with fumarase,
were identical to those obtained with the noncoupled system
where chirality is lost in the formation of MG.10 Thus, 49 and
50 ( 1% of the 3H of isolated malate was labilized by fumarase
compared with 37 or 64% found previously20 with pyruvate
formed from Z and E 3H-PEP, respectively. Therefore, keton-
ization would seem to occur prior to the Glx I reaction, and be
nonenzymatic:32 GS-ePY f GS-MG.
Ketonization of Hemithioacetals of ePY. Formation of HTA
adducts of simple aldehydes is known to be very rapid,33 much
faster than the rates seen in Figure 4. Thus, if an HTA-ePY
adduct is an intermediate in ketonization it would be present in
small amount in a preequilibrium leading to the loss of the
slower change of the absorbance noted, Scheme 3.
According to this model, the rate of apparent ketonization
should become nonlinear with respect to thiol concentration
when the concentration of thiol is comparable to or greater than
the dissociation constant of the complex. No departure from
linear dependence on mercaptoethanol concentration was ob-
served up to 40 mM, the limit for reliable measurements at 5°
C. Thus, Kdiss of the mercaptoethanol-ePY adduct will have to
be considerably greater than 40 mM. This should not be
unexpected given that the dissociation constant of adducts of
acetaldehyde with simple thiols is about 40 mM,33 and that the
value for the conjugation-stabilized enolaldehyde should be
much greater. This implies that only a small fraction of ePY
will be present in the HTA-activated forms to obtain the effects
seen in Figure 4. This small fraction will ketonize more rapidly
than the same amount of ePY, if for no other reason than its
loss of conjugate-stabilization.
Figure 4. Activation of ketonization by thiols. ePY was generated in Bis-
Tris (50 mM, pH 5.9) and quenched. Additions were 1. none, 2. oxidized
GSH (2 mM), 3. 2-mercaptoacetate (2 mM), 4. GSH (2 mM).
system with limiting MGS was only 1.5-2, compared with ∼6
for the spontaneous loss of the absorbance due to ePY, Figure
1. No D2O effect was observed in the rate of production of
ePY by MGS assayed by the initial increase in absorbance at
245 nm, nor was a primary effect seen with 12H-DHAP in the
Glx I-coupled reaction.10 Therefore, the rate-limiting step of the
MGS reaction does not involve proton transfer. No lag was seen
in the Glx-coupled system in low buffer at pH 6 with sufficient
Glx I and GSH to complete the assay within 60 s although t1/2
of ePY by the spectral decay assay was ∼300 s. Therefore,
activation of the ketonization of ePY by GSH, Glx I or the
combination of GSH and Glx I is required.
It was noted that GSH (pKa 3.59, 8.75, and 9.65),30 leads to
an increased rate of fall in absorbance of ePY. Using 2 mM
GSH at pH 5.9, t1/2 was 36 s (Figure 4). A comparable amount
of oxidized GSH showed only a small acceleration relative to
the background rate, t1/2 ≈ 380 s. 2-Mercaptoethanol increased
the rate of loss of absorbance ∼5-fold, t1/2 ≈ 69 s at 2 mM.
Is the loss of absorbance at 245 nm in Figure 4 a measure of
the ketonization of ePY? Because thiols are not reported to be
catalysts of ketonization, two alternative explanations were
considered: a Michael (vinyl) addition of the thiol to the ePY
or the formation of a hemithioacetal adduct. Immediately after
(31) Cheung, Y. F.; Fung, C. H.; Walsh, C. Biochem. 1975, 14, 2981-2986.
(32) Rozzell, J. D.; Benner, S. A. J. Am. Chem. Soc. 1984, 106, 4937-4941
report an exceptional if not unique example of an achiral enzyme product:
acetone-CH3 in the acetoacetate decarboxylase reaction.
(30) Jencks, W. P.; Regenstein, J. Handbook of Biochemistry; 1968, J150-J188,
Chemical Rubber Company Press: Cleveland, Ohio, source of all pKs
referred to in the text.
(33) Lienhard, G. E.; Jencks, W. P. J. Am. Chem. Soc. l966, 88, 3982-3998.
9
13050 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

enol-Pyruvaldehyde: Chemistry and Enzymology
A R T I C L E S
Table 1. Functional and Bifunctional Activatorsa
By lowering the temperature to 5°, after its synthesis from
DHAP in D2O at 25°, ePy was sufficiently stable for its NMR
spectrum to be recorded.
concentration
(M × 10-3
kactivator
succinate
effectd
D O
2
activator
)
(M-1s-1 b
)
effectf
The original proposal that enolpyuvaldehyde and not meth-
ylglyoxal is the product of methylglyoxal synthetase was based
on the generalization that enzymatic products should not be
stereospecifically random as was found to be the case in forming
the methyl group of MG.10 An exception to this rule, subse-
quently found with acetoacetate decarboxylase,32 required a
reexamination of the assumption that MG was not the immediate
product of the MGS reaction. This conclusion has now been
verified by identification of a transient intermediate in great
excess of enzyme equivalence (Figure 1), with spectral proper-
ties of ePY (Figure 2). The observation that ePY is stereospe-
cifically formed in the MGS reaction (Figure 2D) made it
possible to evaluate whether its ketonization might be enzyme
catalyzed in the formation of the methyl group of the MGS/
GSH/Glx I system. Because stereospecificity was not observed
in the CH3 group of the D-LG product, it was concluded that
ketonization of ePY is not enzyme catalyzed, again with the
application of the enzyme stereochemical rule. That free ePY
is formed stereospecifically establishes ePy as the primary, and
with Pi, the only product of methylglyoxal synthetase, or more
accurately DHAPase.
Our conclusion that both the general base, Asp 71, and the
eliminated OPO3 group occur from the same face of the enediol-
3-P intermediate (Scheme 2) depends on the stereochemistry
of the proton abstraction of the first step (ref 10) and the
structure of the intermediate. On the basis of X-ray diffraction
studies of the Harrison group, the -NOH and COH groups of
phosphoglycolohydroxamic acid, a very high affinity inhibitor,
are hydrogen bonded to the active site in a cis-geometry.13 The
contribution of the present work, identifying the methylene
protons in relation to the aldehyde group of the enolpyruval-
dehyde product, Figure 2C and D, establishes the stereochemical
course of the enzymatic reaction, shown in Scheme 2.
The uncatalyzed ketonization of ePY, ∼0.25% s-1 at pH 7,
is much too slow to be a step in the MGS:Glx I couple. The
enhanced stability of ePY compared with other enols, and its
very unfavorable association with thiols and H2O compared to
other aldehydes are undoubtedly due to conjugation of the ene-
and aldehyde groups. The absence of evidence for hydration of
the carbonyl of ePY in the proton NMR spectra is consistent
with 0.8% hydration of benzaldehyde.40
GSH
succinate
2
100
2
0.20
0.08
8.4
0.14
4.3
2.4
2.04
6.4
HSCH2CH2-OH
9.7
1.3e
3.0
1.9
HSCH2CO2
37.0
1.24g
3.7
+
HSCH2CH2NH3
97.0c
a enolPyruvaldehyde was generated in 50 mM Bis-Tris, pH 5.8 in H2O,
D2O, or 100 mM succinate in H2O at 25 °C. MGS activity was quenched
with 2PG. Activators were then added to the noted final concentration and
the absorbance change recorded. b The ketonization rate in H2O, corrected
for the rate with BisTris buffer alone (2.2 × 10-3 s-1), was measured with
the concentration of activator noted. “kactivator” is the corrected rate divided
by the activator concentration. c Compared with HOCH2CH2NH3+, 0.006
M-1s-1 d Calculated with correction for the rates with and without 100
.
mM succinate (0.014 and 0.0022 s-1). e The succinate effect with 3-mer-
captopropionate was 1.2-fold. f Corrected for D2O effect in the absence of
activators. g 1.0, determined for 3-mercaptopropionate
Bifunctional Catalysis. If thiols, such as mercaptoethanol,
activate ketonization as the HTA adduct of ePY and buffers,
such as acetate, do so by promoting the enol to keto conversion
one might expect an even greater activation by 2-mercaptoace-
tate, which has the possibility of intramolecular buffer catalysis.
Indeed, mercaptoacetate and 2-mercaptoethylamine are much
better catalysts than mercaptoethanol (Table 1). Using 2 mM
mercaptoethanol, it would require ∼300 mM lactate, a base with
the same pKa as mercaptoacetate, to obtain the rate found with
2 mM mercaptoacetate. It would require 650 mM ethanolamine
with 2 mM mercaptoethanol to achieve the rate found with 2
mM mercaptoethylamine. The smaller enhancements by 100
mM succinate seen with these activators compared with
mercaptoethanol, Table 1, may indicate shielding of the enol
site by the carboxylate and amino groups of these HTA adducts.
The greater rate with GSH and lower succinate effect seen with
GSH than with mercaptoethanol suggest that the C-terminal
glycine carboxyl of GSH may also play a catalytic role.
The much lower D2O effects seen in Table 1 for the
ketonization rate with all HTA-adducts compared with succinate
may not be the result of a decreased kinetic role of proton
transfer. HTA adducts are known to reach ∼2.9-fold greater
concentration at equilibrium in D2O due to inverse fractionation
factor for the HTA adduct (Ø ≈ 1.25) and the unfavorable value
for thiols (Ø ≈ 0.43).34 Therefore, a solvent kinetic isotope effect
in the ketonization of ePY of ∼5.9 should be decreased to ∼2.0,
as found with GSH and mercaptoethanol (Table 1). The larger
effects seen with the mercaptoethylamine adduct, 3.9, may
indicate a decrease in the fractionation factor contribution as
the ketonization rate increases relative to the decomposition of
the adduct.
In support of the HTA adduct as the form in which GSH
activates ketonization of ePY are the following factors: (1) a
similar effect with mercaptoethanol, (2) oxidized glutathione
is inactive, (3) 2-mercaptoacetate and mercaptoethylamine are
manyfold more active than equivalent amounts of mercapto-
ethanol and the corresponding base, appearing to act as bi-
functional activators. An eight-membered ring structure would
provide a linear alignment of donor-proton-acceptor that should
be favorable for proton transfer.38,39 (4) The low solvent isotope
effect found with GSH and mercaptoethanol, ∼2 compared with
that of the uncatalyzed or general base-catalyzed reactions, 5-6,
is readily explained by the inverse solvent effect expected for
the HTA-substrate equilibrium and is not inconsistent with slow
proton transfer.
Discussion
The use of enzymes to synthesize enols is not new. Both as
an enzymatic end product,35,36 and as a bound intermediate
liberated by acid denaturation,37 these enols have been useful
for enzymatic and chemical mechanism studies. The source
enzyme can be inactivated, as in the present case, and conditions
altered for further study within the lifetime of the enol product.
(34) Schowen, R. L. Prog. Phys. Org. Chem. 1972, 9, 275-332.
(35) Kuo, D. J.; O’Connell, E. L.; Rose, I. A. J. Am. Chem. Soc. 1979, 101,
5025-5030.
(36) Miller, B. A.; Leussing, D. L. J. Am. Chem. Soc. l985, 107, 7146-7153.
(37) Jaworowski, A.; Rose, I. A. J. Biol. Chem. 1985, 260, 945-948.
(38) Gandour, R. D. Tetrahedron Lett. 1974, 295-298.
(39) Kirby, A. J.; Fersht, A. R. Prog. Bioorg. Chem. 1971, 1, 1-8.
(40) Guthrie, J. P. J. Am. Chem. Soc. 2000, 122, 5529-5538.
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13051

A R T I C L E S
Rose and Nowick
The following published evidence may be cited to support
the kinetic model in which the rate-limiting step for HTA
catalyzed ketonization is the ketonization step (Scheme 3). Rate-
limiting formation of HTA adducts would require the observa-
tion of an inverse solvent isotope effect, a reflection of a 2-fold
higher concentration of RS- at equilibrium in D2O.33 However,
inverse effects have not been found either in the Glx I coupled
assay or in the direct assays, Table 1. With respect to the last
step of Scheme 3, it has been shown that the decomposition of
HTA adducts of keto aldehydes24 and of acetaldehyde with
weakly acidic thiols,33,41 are specific base, not general base,
catalyzed. The effects of succinate shown in Table 1, as well
as those of the bifunctional activators, are therefore not
consistent with rate-limiting dissociation of the HTA-MG
adduct. It will be useful to determine the relative equilibria for
formation of the several bifunctional HTA adducts of ePY before
their kinetic differences can be attributed to mechanistic factors.
Direct measurement of adduct formation at higher concentrations
of the several mercaptans, by following the loss of 245 nm
absorbance by rapid flow kinetic methods, will be necessary to
establish the rates and equilibria for formation of the adducts.
It should then be possible to compare the intrinsic rates of
ketonization of simple HTA-ePY adducts with that of other enols
and determine if the HTA function of simple thiols plays a role
other than the loss of double bond conjugation. One such role
(as suggested by a reviewer of this paper), might be for the
HTA-hydroxyl group to activate a bridging water molecule for
proton transfer to the enolate-carbon.
capacity of MGS, the ketonization of ePY and the glyoxalase
system.42 With 1-5 mM GSH,43,44 a 5-25% per sec. rate of
ketonization ensures that GS-MG, not MG will be available for
the Glx I reaction. In bypassing free MG, primarily as MGH,
the slow dehydration step, ∼1% per sec,45 is avoided as a rate-
limiting step in reacting with GSH. When MG is observed in
cell growth media9 it is probably due to insufficient Glx I activity
with which to trap the generated GS-MG.
One is left to wonder whether the use of GSH in activating
the otherwise slow ketonization of ePY and in bypassing the
slow dehydration of MGH may have played an evolutionary
role in the selection of GS-MG to be the substrate form of the
glyoxalase system. If so, other glyoxalase systems may exist in
nature. For example, most actinomycetes lack GSH, but contain
mycothiol (made up of linked N-acetylcysteine, glucosamine,
and myoinositol) as their major low molecular weight reducing
agent44 . It will be of interest if these organisms evolved a
glyoxalase system that uses mycothiol, which should be about
as active as glutathione in catalyzing the ketonization of
enolpyruvaldehyde.
Acknowledgment. Grant support was provided by NIH Grant
GM20940 to I.A.R. who acknowledges Dr. Ralph A. Bradshaw
for generously sharing space and facilities. J.S.N. thanks the
Camille and Henry Dreyfus Foundation and the American
Chemical Society for support in the form of awards.
JA027065H
enolPyruvaldehyde will not be a major metabolite in modern
nucleated cells, given their lack of MGS. However, in some
microorganisms, as much as 40% of the glycolytic flux may
utilize the MGS/glyoxalase pathway without exceeding the
(42) Fareleira, P.; Legall, J.; Xavier, A. V.; Santos, H. J. Bacteriol. 1997, 179,
3972-3980.
(43) Miranda-Vizuete, A.; Rodriguez-Ariza, A.; Toribio, F.; Holmgren, A.;
Lopez-Barea, J.; Pueyo, C. J. Biol. Chem. l996, 271, 19 099-19 103.
(44) Newton, G. L.; Arnold, K.; Price, M. S.; Sherrill, C.; Delcardayre, S. B.;
Aharonowitz, Y.; Cohen, G.; Davies, J.; Fahey, R. C.; Davis, C. J. Bacteriol,
1996, 178, 1990-1995.
(41) Barnett, R. E.; Jencks, W. P. J. Am. Chem. Soc. 1969, 91, 6758-6765.
(45) Cliffe, E. E.; Waley, S. G. Biochem. J. 1961, 79, 475-482.
9
13052 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002