Transition State Analogs for Thiamin Pyrophosphate-dependent Enzymes*

pyruvate dehydrogenase pyrophosphate first order analysis of the decrease in the rates of inactivation caused by thiamin pyrophosphate indicates that thiamin thiazolone pyrophosphate binds at the coenzyme sites. We have also synthesized thiamin thiothiazolone pyrophosphate obtained very similar results with this compound.

For most thiamin-PP'-dependent enzymic reactions, the major covalent changes that occur during catalysis are well established (1). A common feature of these mechanisms is that the high energy intermediates, unlike thiamin-PP itself, have structures in which the sulfur-containing ring is uncharged. For example, the initial steps in the thiamin-PP-dependent decarboxylation of pyruvate involve two metastable intermediates -the ylide, which is formed by abstraction of a proton from carbon 2 of the thiazolium nucleus, and the enamine, which is the immediate product of the decarboxylation of the pyruvate adduct. The transition states leading to the formation of these reactive intermediates should resemble the intermediates in structure and energy (2). These considerations led us to view 'ITPP and 'PPTPP as transition state analogs for thiamin-PP-dependent enzymic reactions.

TTPP, X=O;TTTPP, X=S
In this report we describe the synthesis of these compounds and provide evidence that, in accord with the prediction for transition state analogs (3,4), they bind to the thiamin-PP sites of Escherichiu coli pyruvate dehydrogenase complex much more strongly than does thiamin-PP itself.

MATERIALS AND METHODS
Thiamin disulfide was purchased from ICN Pharmaceuticals.  (8). The thiazolone (1.35 g) was phosphorylated with pyrophosphoryl chloride by the procedure that Hecht and Hawrelak (9) have described for the phosphorylation of 2'-0-methyladenosine. lTP was isolated from the aqueous extract of the reaction mixture at pH 7 by adsorption on Bio-Rad AGI resin in the formate form. The resin was washed with water; and the compound, which bears no net charge at pH 2.5 due to protonation of the pyrimidine ring (pK about 5 (IO)), was then released from the resin by adjusting the pH of the suspension to 2.5 with formic acid. The acidic extract was concentrated by rotary evaporation and then lyophilized. This yielded 1.1 g of TTP that was, on the basis of its ultraviolet spectrum and total phosphorus content, about 70% pure by weight.
This reaction was carried out by the procedure that Kozarich et al.
with the exception that lTP was treated with 2 mol of tri-n-butylamine.
The residue that remained upon evaporation of the dimethyl formamide was dissolved in water; and the aqueous solution, after adjustment to pH 6 with HCl, was applied to a column (1.2 x 9 cm) of Bio-Rad AG2-X8 (acetate) resin. lTPP was eluted between 0. 8 1. Titration of the pyruvate dehydrogenase complex with Tl'PP. Enayme at a concentration of 150 Fg/ml was incubated with the stated concentrations of l"l'PP in 0.5 rn~ MgCl,/lO mm potassium phosphate, pH 6.6, at room temperature.
ARer 20 min (0) and 1 hour (01, 10.~1 aliquots were assayed for activity; AAlmin is the initial rate of change in absorbance at 340 am. The rate of loss of activity in the presence of 'ITPP was followed through the assay of samples from the reaction mixture at time intervals. With an excess of 'M'PP the decrease in activity was first order (Fig. 2, closed circles). The values of the pseudo-first order rate constants for the inactivation of 8 pg/ml of pyruvate dehydrogenase by 1.25,2.5, and 5.0 x 10m7 M TI'PP, at 3" in 0.5 mM MgCl,/lO mM potassium phosphate, pH 6.6, were found to be 0.066, 0.15, and 0.29 min', respectively. Thus, the rate constant is directly proportional to the concentration of TTPP, and the reaction is second order. The average value of the second order rate constant is 5.7 x lo5 M-I min'.
Thiamin-PP decreased the rate of inactivation of pyruvate dehydrogenase complex by TTPP (Fig. 2). The simplest scheme for the interpretation of this effect is one in which the inhibitor (Z) can combine with each pyruvate dehydrogenase subunit (E) (14,15) of the pyruvate dehydrogenase complex only when thiamin-PP (S) is not bound to the subunit, presumably because of competition for the same site: Here K, is the equilibrium dissociation constant for thiamin-PP and ki is the second order rate constant for inactivation.
If it is assumed that the rate of combination of enzyme with thiamin-PP is much larger than that of enzyme with TTPP so that the equilibrium between E and ES is maintained during the inactivation, the observed first order rate constant for formation of EZ is given by The inserted plot in Fig. 1 shows that this equation describes the data. The value for K,, given by the ratio of the intercept to the slope, is 9 x lo-" M.

Reversibility of Reaction of Pyruvate Dehydrogenase
Complex with TTPP-Enzyme at a concentration of 15 pg per ml that had been inactivated by treatment with 1.25 x 10m7 M 'PTPP was dialyzed against 1000 volumes of 0.5 mM MgCl,/lO mM potassium phosphate, pH 6.6, at 3" for 27 hours, and then further dialyzed for 24 hours against a second 1000 volumes of buffer. During this period, the activity of the enzyme rose from less than 0.25 to about 0.75 unit per mg. Identical treatment of enzyme that had not been inactivated resulted in a fall in activity from 15 to 12 units/mg. Thus only about 5% of the initial activity was recovered during dialysis. This minimal reactivation is not unexpected in view of the values of the rate constant for inactivation and the upper limit for the dissociation constant given above. From these, we can calculate an upper limit for the first order rate constant for dissociation of bound TTPP, the value of which corresponds to a half-time of 40 hours. The nature of the interaction between TI'PP and the enzyme appears to be noncovalent, since TI'PP was released under denaturing conditions. Three milligrams of pyruvate dehydrogenase complex (16 nmol of TTPP binding sites) in 1 ml of 0.5 mM MgCl,/lO mM potassium phosphate was treated with 10 nmol of 'ITPP. This solution was dialyzed against water in order to remove the buffer and any unreacted TTPP. The dialysate wasshaken with 0.10 ml of chloroform/ethanol (l/l), and the mixture placed in a boiling water bath for 5 min. After removal of the precipitated protein by centrifugation, the supernatant was concentrated under nitrogen and then subjected to thin layer chromatography in the system described under "Materials and Methods." An ultraviolet-absorbing compound with the mobility of TTPP and the intensity expected for about 10 nmol was detected.2 Enzyme that had not been treated with 'ITPP yielded no compound in this region of the chromatogram. TTTPP -Experiments identical with those described above, with the exception of the titration of the binding sites and the recovery of inhibitor after denaturation, were also performed with TTTPP. The results, including the values of the upper limit for the dissociation constant and of the second order rate constant for inactivation, were substantially the same as those obtained with TI'PP (data not presented, except for that in Table I

Analogs of Thiumin Pyrophosphate Transition States
Hammes (7) have found that the complex contains about 24 independent binding sites for thiamin-PP, with a single intrinsic dissociation constant. The value of this dissociation constant at 3" in 0.5 mM MgCl,/lO mM potassium phosphate, pH 6.2, is 12 x 10m6 M. Since our conditions differ from these only in the DH (6.6 rather than 6.2). it seems likelv that the value of the di&ociation constant for thiamin-PP under our conditions should not differ by more than a factor of 2 or 3; and the value of 9 x 10M6 M that we obtain from the kinetic analysis is thus a reasonable one. A comparison of this value with the upper estimate of the dissociation constants for !M'PP and 'PITPP shows that these inhibitors bind at least 20,000 times more tightly than the coenzyme. This finding is evidence that they are transition state analogs (3,4) and furthermore, that the mechanisms that provided the basis for their design are correct. The remarkable difference between the affinity of the protein for the coenzyme and its affinity for the transition state analogs may be due to the positioning of the sulfurcontaining ring of these compounds either in a nonpolar region of the apoenzyme or in proximity to a positively charged group of the apoenzyme. Transfer of the positively charged thiazolium nucleus from water to either a hydrophobic or a positively charged environment should be energetically much less favorable than transfer of the uncharged thiazolone nucleus. Spectroscopic studies (16, 17) have shown that the thiamin binding site of the related enzyme, pyruvate decarboxylase, is hydrophobic.
On the other hand, the relative affinities for various coenzyme analogs have led to the suggestion that there is a positively charged group at the thiamin site of the enzyme transketolase (18).
Because of the common features of the mechanisms of thiamin-PP-dependent enzymic reactions (11, 'Pl'PP and TPl'PP should prove to be potent inhibitors of other enzymes that function with this cofactor. The compounds should be useful for the titration of thiamin-PP binding sites and for metabolic studies in which the inhibition of a thiamin-PP-dependent enzyme is desired.