Acid Base Catalytic Mechanism of the Dihydropyrimidine ...

Vol. 268, No. 5, Issue of February 15, pp. 3407-3413, 1993 Printed in 11.S.A.

THEJ O U R N A L OF BIOLOGICAL CHEMISTRY

0 1993 by The American Society for Biochemistry and Molecular Biology, Inc

Acid Base Catalytic Mechanismof the Dihydropyrimidine Dehydrogenase frompH Studies* (Received for publication, July 16, 1992)

Beate PodschunS, Karin JahnkeS, KlausD. SchnackerzS, and Paul F. CookBY From the $Theodor-Boueri-Institutfur Biowissenschaften (Biozentrum) der Uniuersitat Wiirzburg, Physiologische ChemieI, a m Hubland, 0-8700 Wiirzburg, Germany and the §Department of Microbiology and Immunology, Texas College of Osteopathic Medicine, Fort Worth, Texas 76107-2699

efficiently degraded according to this catabolic pathway, nePrimary deuterium (NADPH(D)), solvent deuterium, and multiple isotope effects and thepH dependence of cessitating the application of extremely high drug doses. For kinetic parameters havebeen used to probe the mech- example, drug doses in colon cancer therapy may be as high anism of the dihydropyrimidine dehydrogenase from as 100 mg/kg 5-fluorouracil separately or in combinationwith pig liver. Isotope effect and pH-rate data suggest a folinicacid (Peters et al., 1989). The main effect of these rate-determining reductive half-reaction in whichre- chemotherapeutic drugs is the inhibition of t,hymidylate synduction of the flavinby NADPH has only a minor rate thase (Rustum, 1989). About 70-80% of the applied 5-fluolimitation ("V- "(VIKNADpH) l.l),while protonation rouracil is degraded in vivo to fluorinate 0-alanine(Woodcock of the flavin at N-1 occurring in a step following re- et al., 1980). Any inhibition of the catabolic pyrimidine path= 3,while Dzo( V/KN,"pH) = 2). An way could potentiate the therapeutic effectiveness of 5-fluoduction is slow (DzOV enzymatic general acid witha pK of 8.2 is required to rouracil, probably leading to the application of lower drug protonate N-1of the flavin. In the second half-reaction, doses and diminishing drugside effects. As a precursor to the uracil is reduced at C-6 by flavin and protonated on of the opposite face at C-5 by an enzymatic general acid design of this kind of inhibitor, the mechanistic properties dihydropyrimidine dehydrogenase are being studied. The enwith a pK of 9. The hydride transfer from N-5of the Weber,1981), pig flavin to C-5 of uracil is facilitated by an enzymatic zymes purifiedfrom rat(Shiotaniand (Podschun et al., 1989), and bovine (Porter et al., 1991) livers general base with a pK of 5.6 that accepts a proton from N- 1of the flavin. There is also evidence from the are composed of two similar subunits of M , 107,000. The pig pH dependence of V and the V/K for reduced dinucle- liver enzyme contains multiple flavins and iron-sulfur prosotide substrates that a second enzyme residue with a thetic groups (Podschun et al., 1989). The kineticmechanism pK of 6.4 must be unprotonated for optimum activity, of the pig liver enzyme has been shown to be a nonclassical but is not essential for activity. Noneof the functional two-site ping-pong mechanism using initial velocity studies groups reflected in the V/KNaDpH pH-rate profile have in the absence and presence of product and dead-end inhibia role in binding, while both of those observed in the tors (Podschun et al., 1990). Thus, two separate binding sites V/Kuracil profile have a role in bindingas shown by the are available for each substrate/product pair, namely pH dependence of the dissociation constants for the NADPH/NADP' and uracil/5,6-dihydrouracil.The reduced competitive inhibitors ATP-ribose and 2,6-dihydroxdinucleotide reduces the enzyme a t site 1, and electrons are ypyridine. transferred to site2 to reduce uracil to 5,6-dihydrouracil. The predicted mechanism is corroborated by isotopicexchange experiments. (Thevery similar dihydroorotate dehydrogenase from bovine liver mitochondria (Hines and Johnston, 1989), Dihydropyrimidinedehydrogenase is thefirstandratean enzyme that catalyzes the two-electron oxidation of 5,6limiting enzymein the three-step sequenceforuracil and dihydroorotatetoorotate,has also been shown to have a thymine degradation. In a strictly NADPH-dependent step, nonclassical two-siteping-pong mechanism.) By using 'H these pyrimidines are reduced to 5,6-dihydropyrimidines. NMR spectroscopy and stereospecifically deuterated coenAfter ring cleavage by the second enzyme, ammonium ions zymes, it has been determined that dihydropyrimidine dehyand CO, are formed in a third enzymatic reaction,leaving as drogenasefrom pig liver specifically abstractsthepro-Sproducts /3-alanine and 0-aminoisobutyrate, respectively hydrogen of NADPH,makingit amember of theB-side (Wasternack, 1980).Cytotoxic compoundssuchas 5-fluo- stereospecific class of dehydrogenases (Podschun, 1992). The rouracil,which are clinically usedincancertherapy,are stereospecificity of the second half-reaction was studied by Gani and Young (1985) using a partially purified preparation * This work was supported in part by Grant GM 36799 from the of bovine liver dihydropyrimidine dehydrogenase. Using 'H National Institutes of Health and Grant B-1031 from the Robert A. were stereospeWelch Foundation(to P. F. C.),GrantSchn 139/11-2from the NMR techniques and 0-alanine samples that cifically labeled in each of the four C-H bonds in an unamDeutscheForschungsgemeinschaft(to K. D. S.), andGrantCRG 900519 from the NATO Scientific Affairs Division (to P. F. C. and biguous manner, it was shown that uracil is reduced with K. D. S.).The costs of publication of this article were defrayed in overall anti-addition of hydrogen at the si face of C-6 and at part by the payment of page charges. This article must therefore be the siface of C-5. hereby marked "aduertisement" in accordance with18 U.S.C. Section Little is presently known of the location of rate-determining 1734 solely to indicate this fact. steps in the dehydrogenase reaction or of the acid base chem7l To whom correspondence should be addressed Dept. of Microistry catalyzed by the enzyme. In this study,we report the pH biology and Immunology, Texas College of Osteopathic Medicine, 3500 Camp Bowie Blvd., Fort Worth, T X 76107-2699. dependence of kinet.ic parameters in an effort to establish a

-

3407

3408

Acid Base Catalytic Mechanism

working catalytic mechanism. In addition, isotope effects are to cNADP+).The reaction mixture contained the following in a final used to identify the slowest of the two half-reactions and a volume of 1 ml:80 mM buffer, 0.8 mM dithioerythritol, varying concentrations of NADPH and uracil, and enzyme. Typical concennumber of the rate-determining steps along the reaction pathtrations of the fixed substrate were20 and 60 p~ for uracil and way. NADPH, respectively. Assays were carried out by pre-equilibrating NADPH at 30 "C with enzyme for 2 min and initiating the reaction with uracil. The experimental reaction mixture containing uracil was measured against a blank containing the identical reaction mixture Materials-NADP+, NADPH, glucose 6-phosphate, and glucose-6- without uracil. This procedure minimized inhibition by breakdown phosphate dehydrogenase were obtained from Boehringer Mannheim. products of NADPH and NADP+ at acidic and basic pH values, Uracil, ATP-ribose, cNADP+,' glucose dehydrogenase, fl-D-glucose- respectively. The enzyme activity was then calculated as the differ1-d, Mes, Mops, Ches, triethanolamine, Tricine, DZO, and dithio- ence between the rate observed for the full reaction mixture minus erythritol were purchased from Sigma, while 2,6-dihydroxypyridine that of the blank. For pH studies, the following buffers were used was from Aldrich. All other chemicals and reagents were obtained over the pH ranges indicated Mes, 5-6; Mops, 6-7.5; triethanolamine or Tricine, 7.5-8.5; and Ches, 8.5-10. When buffers were changed, from commercial sources and were of the highest purity available. Synthesis of cNADPH-cNADP+ was enzymatically reduced using kinetic parameters were redetermined at a pH value that overlapped bovine liver glucose-6-phosphate dehydrogenase. The reaction mix- with the previous buffer to correct for potential buffer inhibition. To ture contained the following in a final volume of 10 ml: 1mM glucose normalize data collected on different days, enzyme stock solution 6-phosphate, 0.25 mM cNADP+, 87.5 units of glucose-6-phosphate activities were standardized by comparing rates with 28 mM potasdehydrogenase,and 70 mM potassium phosphate, pH 7.4. The mixture sium phosphate, pH 7.3,0.8mM dithioerythritol, 60 p~ NADPH, and was incubated at 37 "C for 1 h. The enzymes were then removed by 50 p M uracil. Determination of Kinetic IsotopeEffects-Primary kinetic subultrafiltration through aPM-10 membrane, andthesupernatant strate deuterium isotope effects were measured by direct comparison contained 0.168 mM cNADPH. The reaction product was purified by of the initial velocities obtained with 20 mM uracil at varied concenchromatography on Fractogel EMD DEAE (see below). trations of either NADPH or B-side NADPD. For solvent isotope Preparation of Stereospecifically Deuterated NADPD-B-side effects, all reactant and enzyme solutions were prepared in D,O, and NADPD ([(4S)-D,(4R)-H]NADPD)was prepared by enzymatic oxi- pD-rate profiles were obtained. The ratio of the pH(D)-independent dation of P-D-glucose-1-din the presence of NADP+ using the glucose values of the parameters of interest gives the solvent deuterium dehydrogenasereaction. The reaction mixture contained the following isotope effect. The dihydropyrimidine dehydrogenase was concenin a final volume of20ml:6.25 mM fl-D-glucose-1-d,0.225 mM trated -10-fold at 4 "C using a Centricon PM-10 microconcentrator NADP+, 4.2 units of Bacillus megaterium glucose dehydrogenase, 0.8 (Amicon Corp.) and then brought to the original volume with DZO. mM dithioerythritol, and 80 mM potassium phosphate, pH 7.4. The After repeating this procedure five times, the enzyme solution was used for kinetic determinations. All buffers and reactant solutions mixture was incubated at 37 "C, and reaction progress was followed by monitoring the absorbance change at 340 nm. After completion of prepared in DzO were adjusted to the appropriate pD with KOD or the reaction (1 h) and removing the enzyme by filtration through a DC1. The nomenclature is that of Northrop (1977) as modified by PM-10 membrane, thesupernatant contained 0.21 mM B-side Cook and Cleland (1981), where a leading superscript depicts the NADPD. The stereospecifically deuterated NADPD was purified as isotope on the parameter of interest, and a following subscript indidescribed for cNADPH by chromatography on Fractogel EMD DEAE cates the constant presence of a second label in the experiment. Thus, reflect substrate deuterium isotope effects measured (see below), and fractions havingAZ60/A340 ratios of -2.5 were pooled. "VHz0and DVDzo The NADPD was precipitated with a 10-fold excess of cold acetone in HzO and D20, respectively. Deuterium Washin-The rate of deuterium incorporation into Cand lyophilized. The extent of deuterium incorporation was deter4 of the nicotinamide ring of NADPH was measured under conditions mined by 'H NMR spectroscopy and found to be >95%. of turnover by obtaining 'H NMR spectra as a function of time. Purification of Reduced Dinucleotides on Fractogel EMD DEAESpectra were obtained using 1-ml samples in 30 mM potassium The cNADPH and B-side NADPD prepared as described above were phosphate, pD 7.3, at 25 "C with 5 mM NADPH, 1.5 mM uracil, and purified via an HPLC separationprocedure on Fractogel EMD DEAE -10 pg of dihydropyrimidine dehydrogenase (specific activity of 20.6 (25-40 pm; E. Merck AG, Darmstadt, Federal Republic of Germany). pmol/h/mg). Control samples were identical, but excluded dihydroThe HPLC equipment consisted of an L-6210 intelligent pump pyrimidine dehydrogenase. Spectra were collected on a Bruker spec(Merck/Hitachi), an LCD-500 detector (GAT-7; Analysentechnik trometer at 300 MHz for 'H collecting 8000 data points with a sweep GmbH), and aD-2500 chromatointegrator (Merck).The column was width of 5000 Hz and a pulse angle of 45". A total of 36 scans were equilibrated at pH 7.0 using 1 M potassium phosphate and then accumulated per time point, and all spectra were run at 20 "C. The washed with water. After equilibration in water, samples were injected protons at C-4 of the nicotinamide ring of NADPH resonate as a at a flow rate of 1 ml/min. Buffer salts andnon-nucleotide substrates doublet of doublets with the major resonances centered at 2.9 ppm remaining after completion of the above enzymatic preparations were (pro-R) and 2.86ppm (pro-S) (Fig. 1). Exchange of the pro-Seluted with water. Separation ofNADP' and B-side NADPD (or hydrogen in the remaining NADPH is observed as a decrease in the relative intensities of the pro-S doublet with respect to that reprecNADP+ and cNADPH) was performed using a linear gradient of 00.5 M ammonium hydrogen carbonate, pH 7.0. Under the experimen- senting the pro-R doublet (Fig. 1). Theamount of exchange increases tal conditions used, NADP+ elutes at 15.7 min and NADPH at 20.9 with time as evidenced by 16, 23, and 48% decreases at 20, 186, and 1280 min, respectively. Thesedata, when plotted as In (percent min. Absorbance at 280 nm was monitored, and fractions of 500 pl remaining) versus time, give a first-order rate constant of 6 X lo-' were collected. Enzyme Purification-Dihydropyrimidine dehydrogenase was pu- s-l. No change in the relative intensities of the two doublets was rified from pig liver cytosol by fractionation with ammonium sulfate, observed over the same time period in the minus-enzyme control. Data Processing-Reciprocal initial velocities were plotted versu treatment with acetic acid, and chromatography on DEAE-cellulose and 2',5'-ADP-Sepharose as described (Podschun et al., 1989). Usu- reciprocal substrate concentrations, and the experimental data were fitted using Equations 1-8 by a nonlinear least-squares method and ally the purified enzyme has specific activities ranging from 25 to 30 the Fortranprograms of Cleland (1979). units/mg of protein. One unit is defined as the amount of enzyme that causes the disappearance of 1 pmol of reduced coenzyme/h. Assay Conditions-Enzyme activity was determined at 30 "C using a Varian/Cary 3 spectrophotometer linked to an Epson personal computer. The decrease in absorbance at 340 nmwas monitored accompanying the conversion of NADPH to NADP+ (or cNADPH EXPERIMENTALPROCEDURES

The abbreviations used are: cNADP+, P-nicotinamide-adenine dinucleotide 2':3'-monophosphate; cNADPH, reduced form of cNADP', Ches, 2-(cyclohexylamino)ethanesulfonic acid; 2,6-DHP, 2,6-dihydroxypyridine; Mes, 2-(N-morpholino)ethanesulfonic acid; Mops, 3-(N-morpholino)propanesulfonic acid; Tricine, N tris(hydroxymethy1)methylglycine; HPLC, high pressure liquid chromatography; 4, proton fractionation factor.


3409

A

B

I

I

1

1

3.2

3.0

2.8

PPM

I

2.6

FIG. 1. 'H NMR spectra of dihydropyrimidine dehydrogenase reaction mixture with time after addition of enzyme. Conditions are described under "Experimental Procedures." A , minus dihydropyrimidine dehydrogenase control; B , reaction mixture after 1280 min. The doublet of doublets centered at 2.88 ppm represents the C-4 protons of NADPH. Substrate saturation curves were fitted using Equation 1. Data for substrate deuterium isotope effects obtained by direct comparison of initial velocities were fitted using Equation 2, while data conforming to linear competitive and noncompetitive inhibition were fitted using Equations 3 and 4, respectively. pH profiles decreasing at low or high pH with limiting slopes of tl or -1 were fitted using Equations 5 and 6, while profiles decreasing at both low and high pH with linear slopes of +1 and -1 were fitted using Equation 7, respectively. Data exhibiting a change from a constant value at high pH to a lower constant value at lower pH were fitted using Equation 8. In Equations 1-4, u is the initial velocity; V is the maximum velocity; K, and Kb are K,,, values for reactants A and B, respectively; K,, K,, and K;;are inhibition constants for reactant A and inhibitor I; F, is the fraction of deuterium in the labeled substrate; and Ev,K and E v are the isotope effects minus 1 on V / K and V , respectively. In Equations 5-8, Y is the value of the parameter of interest measured at any pH, C is the pH-independent value of Y , H is the concentration ofH', Kl and K2 are acid dissociation constants for enzyme groups, and Y L and YH are the pH-independent values of Y at low and high pH, respectively, in the case where a partial change is observed.

FIG. 2. pH dependence of kinetic parameters for pig liver dihydropyrimidine dehydrogenase obtained at 30 "C. A , pH dependence of V obtained in H,O (0,O) varying NADPH a t 20 PM uracil (0)and uracil at 60 PM NADPH (0)and in D20 (A,A) varying NADPH at 20 FM uracil (A) and uracil at 60 p M NADPH (A).Data were fitted using Equations 7 and 9. B , pH dependence of V/KNADpH obtained in H 2 0 (0)and in D20 (A) varying NADPH at 20 PM uracil. Data were fitted using Equation 8. C, pH dependence of V/Kurai, obtained in H 2 0 (0)and in D,O (A)varying uracil at 60 PM NADPH. Data were fitted using Equation 8. In all cases,points areexperimental values, while curues are theoretical based on the fitted parameters.

constant with apK of 5.6. The pKcalculatedfrom the decrease at high pH is 8.8. The pH-independentvalues of VIE, at high and low pH are 0.5 and 0.06 s-I, respectively. A repeat of the entire p H profile in D 2 0 gives increases in the pK values by -0.4-0.6 pD units compared to those obtained inwater (Fig. RESULTS 2 A ) , as expected due to the equilibrium solvent deuterium effect on the pK (Quinn and Sutton,1991). The ratio of the pH Dependence of Kinetic Parameters and Solvent DeutepH(D)-independent values gives a DzoVof 3.3 k 0.2. rium Isotope Effects-Kinetic parameters for the dihydropyrThe pH dependence of V/KNADPH obtained by varying the imidine dehydrogenase reaction were measured over the pH range 5-9.8. Whether NADPH concentrations were varied concentration of NADPH at 20 PM uracil is shown in Fig. 2B. with uracil levels fixed at 20 pM or uracil concentrations were The V/K decreases at bothlow and high pH, giving pK values varied with NADPH concentrations fixed a t 60 p M , the value of 8.2 and 5.8, respectively. The pH-independentvalue of V/ for V was observed to decrease at both low and high pH (Fig. KNADpHEt is 6.4 X lo5 M" s-'. A repeat of the V/K profile in D20 again gives the expected increase in the pKvalues and a 2 A ) . The decrease at low pH, however, is partial, indicating that although a group is required unprotonated for optimum Dzo V/KNADpH of 2.2 0.4, slightly smaller than that observed activity, it is not essential for activity. The decrease on the on V. The pH dependence of V/Kuracilobtained by varying the basic side appears to be all or none, suggesting that a second essential group is required protonated for activity. The pKfor concentration of uracil at 60 pM NADPH is shown in Fig. 2C. the partial change at low pH is 6.5, and V goes to a lower The V/K decreases at both low and high pH, giving pK values

Mechanism Catalytic

3410

Base

Acid

of 9.1 and 5.6, respectively. The pH-independentvalue of V/ K,,,,,,E, is 8.5 X lo5M-' s-'. A repeat of the V/K profile in D 2 0 again gives the expected increase in thepK values and a L)~vV/K,,,,ilof 0.67 f 0.13, opposite the normal values observed on V and v/KNAopH. A summary of the pK values is given in Table I. pH Dependence of V and V/K with cNADPH-The pK reflected in the v/KNADpH profile could result from the ionization of the 2"phosphate of NADPH. Using cNADPH, a substrate in which the 2"phosphate is cyclicized, data very similar to those observed with NADPH are obtained. The V decreases at low pH with a pK of 6.2 and levels off at a lower constant value with a pK of5.7 (Fig. 3A). The V profile is also beginning to decrease at high pH, asalso seen for V with NADPH. Datahave been collected forcNADPH to just below the pKin the VNADpH profile, so the decrease is not as evident as itis in Fig. 1A. However, the pKvalues obtained at low pH are identical for VNADpH and VcNADpH, providingconfidence that the high pK is also present. The V/K profile decreases at low and high pH (Fig. 3B), but the change at low pH is partial as in the V profiles with NADPH and cNADPH. The low pK is 6.4, and V/K levels off at a lower constant value with apK of 5.7, while the high pK is 8.4. The pH-independent values of VIE, at high and low pH are 0.4 and 0.07 s-', TABLEI Summary of pK values fromthe pH dependence of the kinetic Darameters Parameter

PK.

V

PKb

6.5 f 0.04 5.6 f 0.05" 5.8 0.3 5.6 f 0.2 6.4 f 0.2 5.7 f 0.2" 5.7 0.2 6.2 f 0.2" constant 6.0 f 0.3

*

8.8

* 0.1

8.2 f 0.3 9.1 f 0.2 8.4 f 0.1

*

9.1

* 0.2

The decrease at low pH is partial, and thereported pK is the one for the function leveling off at a lower constant value.

m

0 -

5

6

7

8

9

8

9

pH

m

0 -

5

6

7 PH

FIG. 3. pH dependence of kinetic parameters for pig liver dihydropyrimidine dehydrogenase obtained by varying cNADPH at 20 PM uracil at 30 O C . A , pH dependence of V. Data were fitted using Equation 9. B, pH dependence of V/KcNADpH. Data were fitted using Equations 7 and 9.

"i $ 5

Y

a

I

A

A *

* r.

P

-

A

A

A

4

4

,

.

,

5

.

m

.

,

7

6

,

t

.

8

9

m

*

10

pH

1

9

I 5

I

6

I

7

I

I

I

0

9

10

PH

FIG. 4. pH dependence of dissociation constants for competitive inhibitors obtained at 30 "C. A , pH dependence of l/Ki for ATP-ribose, competitive uersw NADPH; B, pH dependence of 1/Ki for 2,6-DHP, competitive versus uracil.

respectively, and those of V/KcNADpHEt are 4 X io5and 1.8 X lo4 M-' s-', respectively. These findingsclearly show that ionization of the 2"phosphate of NADPH does not affect V/ K. The pKvalues are included in Table I. Inhibition Profiles-The dissociation constant for ATPribose, an inhibitor competitive versus NADPH, is pH-independent from 5.5 to 9.5 (Fig. 4A). The dissociation constant for2,6-dihydroxypyridine, aninhibitor competitive versus uracil, is pH-dependent, giving pK values of -6 and 9 (Fig. 4B). The pH-independent values for K:(ATp.ribose)and Ki(DHp) are 30 PM and 40 nM, respectively. Kinetic Isotope Effects withB-side NADPD in H 2 0 and D,O-Primary substratekineticdeuterium isotopeeffects were measured at three pHvalues varying the concentration of NADPH(D) at 20 K M uracil. At pH 7.2, isotope effects are small and within experimental error; equal values of 1.09 k 0.03 are measured on V and v/KNADpH. Since V and V/K were found to decrease at both low and high pH, the isotope effects were also measured at pH 5.6 and 8.6 near the pK values observed in the v/KNADpH profile. At both pH values, the isotope effects on the two kinetic parameters are within experimental error, equal to one another. At pH 5.6, DV and DV/K are 1.08 f 0.03, while at pH 8.6, a value of 1.19 f 0.08 is estimated. The theory andmethodology of multiple isotope effects on enzyme-catalyzed reactions developed by Hermes etal. (1982) allow one to determine whether the solvent deuterium and substrate deuterium isotopeeffectsreflectasingle step or separate steps. The substrate deuterium isotope effect was thus repeated in D 2 0 a tpD 7.3, giving a value of 1.03 f 0.03. The decrease in the values ofDVD,O and "( V/KNADPH)DpO z ~"(~/KNADPH)H,O suggests that the compared to D V ~and substrate deuterium and solvent deuterium isotopeeffects reflect different steps. Although there is a slow incorporation of deuterium into the pro-S position at C-4 of the nicotinamide ring of NADPH, the rateof this incorporationis SO slow as to be insignificant under initial velocity conditions. Thus, the small substrate deuterium isotope effects do not reflect the washoutof deuterium from C-4,and thesolvent deuterium isotope effects do not reflect the washin. DISCUSSION

Location of Rate-limiting Steps-The reduction of the flavin by NADPH is the slowest of the two half-reactions based on

Acid Base Catalytic Mechanism the following information. The solventdeuterium isotope effectson V and v/KNADpH are large,while the substrate deuterium isotope effects on V and V/KNADWare small and identical, suggesting common rate limitationfor both of these rate constants. In supportof this suggestion, the pH dependencies andpK valuesobservedforVwith NADPH (or cNADPH)areidenticaltothose observedfor V/K,NmIw. These data further suggest that the pH dependence ofV/ KNAD~Hwill also likely reflect a partial change atlow pH. The reason that the partial change is not observed in the latter is probably a result of some stickiness‘ of NADPH. Consistent with this suggestion, the value of ”(V/KNADpH) increases from -1.1 at pH 7.2 to -1.2 at pH 8.6, a value near the pK(Cook, 1991). Some information on the identity of the slow step(s) within the two half-reactions isalso provided from the isotope effect data.The isotope effect on reduction of the flavin by NADPH(D) is small,with a maximum observed value of 1.2. (A slightly larger value should be observed at pH values >9, where the chemical pathway will completely limit the halfreaction (Cook, 1991).) The small primary deuterium isotope effect indicates thathydride transfer from C-4 of the nicotinamide ring to N-5 of the flavin is either 1) not a major ratedetermining step or 2) has an early (or late) transition state. Thesolventdeuterium isotopeeffect of -3 on V/KNADpH argues for one or more protons transferred in the rate-determining transition state. The substrate deuterium isotope effect (NADPH(D)) decreases to a value of 1 when it is measured in D,O; and thus, the solvent deuterium and substrate deuterium isotopeeffectsreflect separatesteps (Cleland, 1991).A possible explanation for separate hydride and proton transfer steps will be considered below. An inversesolventdeuterium isotopeeffect of -0.7 is observed in the second half-reaction. Possible explanations for a n inverse isotope effect of this magnitude include ionization of a thiol (4 = 0.4-0.6), hydrolysis of metal-water (4 = 0.7-0.9) and metal-bound hydroxide (4 0.7), and the existence of a low barrier hydrogen bond (4 = 0.3-0.6) (Quinn and Sutton, 1991). Allof the above are possible in the case of dihydropyrimidine dehydrogenase. A thiol of bovine dihydropyrimidine dehydrogenase is trapped by the suicide inactivator 5-iodouracil, which may implicate this enzyme residue in the reduction of uracil (Porter et al., 1991). There isalso the potential for the involvement of metal-water (or metal-hydroxide) as a result of the abundanceof non-heme iron present in dihydropyrimidine dehydrogenase (Podschun et al., 1989). Finally, a tight hydrogen bond could exist between an enzyme residue and flavin. An explanation of theinverse solvent deuterium isotope effect will have to await further study. As statedunder“Results,”there is a smallamount of incorporation of deuterium from solventinto C-4 of the reduced nicotinamide ring. Therate of washin (6 X s-’) is very low compared to the rate of the overall reaction (0.4 s”), suggesting that washoutof deuterium from C-4 does not contribute to the smallobserved isotope effect. Since washin is observed, however, the release of NADPmustbe slow enough totrapthedeuterium exchanged intoN-5 of the reduced flavin. In addition, these data suggest that electron transfer to site 2probablydoes not occur until NADP is released. Interpretation of p H Profiles-The V/K for NADPH (or cNADPH) reflects the first half-reactionof the two-site ping-

-

A substrate is termed “sticky”when partitioning of the ES complex toward product is favored compared to dissociation of ES to give free reactant and enzyme.

3411

pong kinetic mechanism. As a result, the pH dependence of the V/K for the reduced dinucleotide substrate provides information on the acid base chemistry and optimum protonation state for reactant binding in the reduction of flavin at site 1 (Podschun et al., 1990). The VIKcNAUPH decreases from a constant value at pH 7.5 to a lower constant value at low pH, and thisbehavior is also observed in the pHdependence of V with either NADPH or cNADPH as the dinucleotide substrate. As stated above, the samebehavior is also expected in the V/KNADpH pH profile and is likely not observed as a result of the stickiness of NADPH. The only titrable group on thedinucleotide is the 2”phosphate.However, even though this group is in esterlinkage withthe 3”hydroxyl in cNADPH and the second phosphate pK is not observed, the same pH behavior is observed for v and V/KcNADpH. Thus, the functional group titrated in these pH profiles must be an enzyme residue. This residue is not essentialfor catalysis as evidenced by the partial change in V and V/K. The actual role of this group is presently notknown, but. it is not involved in binding the dinucleotide substrate since thedissociation constant for the competitive inhibitor ATP-ribose is pH-independent. The group couldbe involved in maintaining the structural integrity a t site 1,but this aspectwill have to await further study. The V/K values for reduced dinucleotides (and V) also decrease at high pH, giving a pK of8.2-8.8. Since reactant has no titrable groupsin this pH range, thepKmust reflect an enzyme group that must be protonated for activity. The V/K for uracil decreases at high and low pH, giving pK values of -5.5 and 9. The pHdependence of the dissociationconstant for 2,6-DHP,aninhibitor competitivewith uracil, indicates that both of the above groups are required correctly protonated for uracil binding. Uracil has no titrable groups in the pH range studied, indicating that the groups with pK values of 5.5-6 and 9 must be enzymeresidues involved in binding andcatalysis. Chemical Mechanism-A reasonable chemical mechanism can be postulated taking into account the above discussion. The binding of NADPH to site 1 is such that the pro-Sside (Podschun, 1992) of the nicotinamide is directed toward N-5 of the flavin isoalloxazine ring (Scheme 1, I). Binding of NADPH apparently requires only weak interactions with the 2’-phosphate since the V/K values forcNADPH and NADPH are nearlyidentical. Thetransfer of the hydride toN-5 generates reduced flavin (11),which is then protonated at N1 by an enzymatic general acid with a pK of8.2-8.4 (III).:$ That hydride transfer and protonation of the reduced flavin occur in separat,e steps is indicated by the multiple isotope effect data exhibiting a decrease in D( V/&A”pH) from a value of 1.1 in H 2 0 t oa value of 1.0 in DnO.For the solvent isotope effect to reflect a step after hydride transfer and be observed on the V/K for NADPH, it must. occur either concomitant with or prior to the release of the product, NADP+, which constitutes an irreversible step under initial velocity conditions. Electron transfer to site 2 then occurs presumably via FeS clusters. Use of 5-iodouracil as a suicide inactivator of the dihydropyrimidinedehydrogenasefrom bovine liver results in the capture of an active-site thiol at C-5 (Porter al., et 1991). The thiol could function as a general acid required to protonate uracil upon reduction by flavin. Using bovine liver dihydropyrimidine dehydrogenase, Gani and Young (1985) showed 3Alternatively, electron transfer to a flavin at site 2 could occur immediately upon reduction of the flavin a t site 1 by NADPH followed by protonation of N-1 of the reduced flavin at site 2, with both occurring prior torelease of NADP+.

3412


R

-1

II

H H

O

I

I

I

111 R

R

-1

H

-0,

-%

SCHEME1. Acid base chemical mechanism of dihydropyrimidine dehydrogenase based on pH and isotope effect studies. anti-addition of hydride and proton to uracil. Finally, work by Hines and Johnston (1989) on the similar dihydroorotate dehydrogenase using multiple isotope effects suggests a concerted addition of hydride and proton. Taking into account the above information, the binding of uracil at site 2 likely occurs such that the si face a t C-6 is directed to N-5 of the

reduced flavin and the si face at C-5 is directed toward an enzymatic generalacid (IV). Reduction of uracil isthen accompanied by deprotonation of the flavin at N-1 and protonation of uracil at C-5 to generate 5,6-dihydrouracil (V). Alternatively, the group with apK of 5.5-6 may serve to bind uracil by hydrogen bonding toN-1.


3413

REFERENCES Peters G. J. Wilt, van der C. L., van Groeningen, C. J., Nadal J. C., Laurensse, E., i n d Pinedo, H. M. (1989)Int. Con r Symp. Ser. 158,il-35 Cleland, W. W. (1979)Methods Enzymol. 63, 103-138 Podschun, R. (1992)Bcochem. Ewphys. 5 e s . Commun. 182,609-616. Podschun, R., Wahler, G., and Schnackerz, K. D. (1989)Eur. J. Biochem. 185, Cleland, W. W. (1991)in Enzyme Mechanism from Isotope Effects (Cook, P. F., pp. ed) 247-265, Press, CRC Inc., Boca Raton, FL 219-224 Podschun, B., Cook, P. F., and Schnackerz, K. D. (1990)J. Biol. Chem. 265, Cook, P. F. (1991)in Enzyme Mechanism from Isotope Effects (Cook, P. F., ed) pp. Raton, 231-245, Boca Inc., Press, CRC FL 12966-12972 Porter, D. J. T., Chestnut, W. G., Taylor L. C. E., Merrill, B. M., and Spector, Cook, P. F., and Cleland, W. W. (1981)Biochemistry 20,1790-1796 Gani, D., and Young, D. W. (1985)J. Chem. SOC. PerkinTrans. I1.1355-1362 Mechanism from Isotope Hermes, J. D., Roeske, c. A., O’Leaw, M. H., andCleland, w. w. (1982) (Co& p, F,, ed) 73 123, CRC press, I ~ ~B~~~ , , R ~FL ~ ~ Biochemistry 21,5106-5114 Rustum, Y. M. (1989)Int. &ngcSym . Ser. 158, 11-19 Hines, V., and Johnston, M. (1989)Biochemistry 28,1227-1234 Shiotani T., and Weber, G. (1981)J. h o l . Chem. 256, 219-224 Northrop, D. B. (1977)in Isotope Effects on Enzyme-catalyzed Reactions (Cle- Wastern& c. (1980)Pharmacal. & ~ h8 , 629-651 ~ ~ . land, W. W., O’Leary, M. H., andNorthrop, D. B., eds)pp. 122-154, Woodcock, T.M., Martin, D. S., Damin, L. E. M., Kemeny,N.E.,and Young, Baltimore Press, University Park C. W. (1980)Cancer (Phcla.) 45,1135-1143

;,(,‘:!g ~ ~ ;~ ~ n 2 , 6 . 6 ~ ~ g ~ ~ ~ ~ ~ ~ n ~ ‘:iffe& ~

,