Indiscriminate Binding by OMP Decarboxylase of UMP Derivatives with Bulky Anionic C₆-Substituents

Abstract

OMP decarboxylase appears to act upon its substrate without the intervention of metals or other cofactors and without the formation of covalent bonds between the enzyme and the substrate. Crystallographic information indicates that substrate binding forces the substrate’s scissile carboxylate group into the neighborhood of several charged groups at the active site. It has been proposed that binding might result in electrostatic stress at the substrate’s C₆ carboxylate group in such a way as to promote decarboxylation by destabilizing the enzyme-substrate complex in its ground state. If that were the case, one would expect UMP derivatives with bulky anionic substituents at C₆ to be bound weakly compared with UMP, which is unsubstituted at C₆. Here, we describe the formation of anionic 5,6-dihydro-6-sulfonyl derivatives by spontaneous addition of sulfite to UMP and to OMP. These sulfite addition reactions, which are slowly reversible and are not catalyzed by the enzyme, result in the appearance of one (or in the case of OMP, two) bulky anionic substituents at the 6-carbon atom of UMP. These inhibitors are bound with affinities that surpass the binding affinity of UMP. We are led to infer that the active site of OMP decarboxylase is remarkably accommodating in the neighborhood of C₆. These are not the properties that one would expect of an active site with a rigid structure that imposes sufficient electrostatic stress on the substrate to produce a major advancement along the reaction coordinate.

In the final step of pyrimidine nucleotide biosynthesis (Scheme 1), the dimeric enzyme orotidine 5′-phosphate decarboxylase (ODCase; E. C. 4.1.1.23) converts orotidine 5′-phosphate (OMP) to uridine 5′-phosphate (UMP) with the release of CO₂. Unlike most decarboxylases, ODCase contains no metal ions or other cofactors that might contribute to catalysis, and there is no evidence that this reaction proceeds through a covalent intermediate (1). One mechanism that has been proposed (2) for this reaction involves electrostatic stress between the scissile 6-carboxylate group of OMP and the carboxylate groups of two active site residues (Asp-91 and Asp-96 of the yeast sequence). These residues recur, with Lys-59 and Lys-95, as members of a conserved quartet of charged residues that are present in ODCase in organisms from every kingdom (3). In the crystal structures of nucleotide complexes with ODCase, the 5′-phosphoryl group appears to brace the substrate within the active site in such a way as to force the scissile carboxylate group into the neighborhood of the charged quartet. Removal of the 5′-phosphoryl group from OMP reduces its susceptibility to enzymatic decarboxylation by a startling factor of 10¹¹, indicating a contribution to the rate enhancement whose magnitude exceeds the effects of a single substituent that appear to have been recorded for any other enzyme reaction (4,5).

Scheme 1.

Decarboxylation of OMP to UMP catalyzed by ODCase.

It is not a simple matter to test the electrostatic stress hypothesis directly. Enzyme complexes with the product UMP have been characterized extensively. But enzyme-bound OMP is kinetically and thermodynamically unstable, and has been observed only after mutation of active site residues that are essential for catalysis (6, 7). An alternative approach, adopted in the present work, is to examine the enzyme’s affinity for substrate analogues whose structures would be expected to accentuate or alleviate stress. Earlier, 6-methylaminoUMP, with a potential positive charge at the position normally occupied by the scissile carboxylate group of OMP, was shown to be bound in uncharged form with an affinity that approaches the affinity of OMP itself (8).

If electrostatic stress were to destabilize the enzyme-OMP complex in such a way as to reduce the thermodynamic barrier that must be surmounted to reach the transition state, then non-reacting analogues bearing negatively charged substituents at C₆ would be expected to be less tightly bound by the enzyme than analogues lacking those substituents. Pyrimidine derivatives have been shown to undergo 6-addition of the sulfite anion, an exceptionally reactive nucleophile, with trans addition of a solvent proton at C₅ (9 – 16) (Scheme 2A). Sulfite addition to UMP seemed to offer an opportunity to prepare an analogue of OMP equipped with a bulkier anionic substituent than the 6-carboxylate group of OMP. Examining the scope of this reaction, we were surprised to find that OMP also undergoes sulfite addition at C₆, albeit less readily than UMP, yielding a molecule with two anionic substituents at the 6-position (Scheme 2B). Here, we describe the preparation of these analogues and their behavior as inhibitors of ODCase.

Scheme 2.

Sulfite addition to (A) UMP and (B) OMP

Materials and Methods

Materials

Nucleotides and other reagents were purchased from Sigma-Aldrich Chemical Corp. Yeast ODCase C155S, a mutant enzyme that is more stable than the wild-type enzyme but retains its catalytic properties, was kindly provided by Dr. Steven Short of GlaxoSmithKline, Research Triangle Park, North Carolina 27709.

Preparation of Adducts

Sulfite mixtures containing equimolar NaHSO₃ and Na₂SO₃ (pH 7.0) were used in addition reactions. Addition of sulfite to UMP was conducted in H₂O or D₂O to produce adducts with two protons at C₅, or one proton and one deuteron at C₅, respectively. Adducts of UMP were prepared by exposing UMP (0.025 M) to sulfite (0.6 M) for 24 hours at room temperature, allowing addition to proceed to completion as described for the corresponding uracil derivatives (9,13). Sulfite addition to OMP required higher sulfite concentrations and higher temperatures than sulfite addition to UMP. OMP adducts were prepared by exposing OMP (~0.010 M) to sulfite (2.0 M) at 40^oC for three days. Unexpectedly, this reaction was found to proceed to completion after the resulting solution had been stored frozen for several days (see Results).

¹H NMR Spectroscopy

¹H NMR spectra were obtained using Varian Unity Inova 500 and 600 MHz spectrometers (Palo Alto, CA) controlled by Solaris 9 software. Samples were generally analyzed at a concentration of ~3 x 10⁻³ M in D₂O at 25 ^oC, acquiring 16 transients, and spectra were referenced to the chemical shift of the residual water proton resonance at 4.77 ppm.

The time course of formation of the two diastereomeric adducts of UMP and OMP was followed by ¹H NMR. Integrated intensities of the resonances arising from the C₆, C₅, and C_1′ protons of UMP, and the C₅, and C_1′ protons of OMP, for each diastereomer and unreacted nucleotide, were substituted into Equation 1(17) to calculate the concentrations of each diastereomer and unreacted nucleotide at each time point. For example, to calculate the concentration of diastereomer I,

$$[\text{I}]=([{\text{Int}}_{\text{I}}]/[{\text{Int}}_{\text{I}}+{\text{Int}}_{\text{II}}+{\text{Int}}_{\text{nuc}}])[{\text{nucleotide}}_{\text{t}}]$$ (Equation 1)

where Int is the integrated intensity of the signal arising from diastereomer I, Int_II + Int_nuc are those of diasteromer II and unreacted nucleotide, and nucleotide_t is the concentration of nucleotide at the start of the experiment.

At each time point, the concentration of sulfite remaining was calculated by subtracting the sum of the concentrations of the two diastereomers of the adduct from the initial sulfite concentration. Changes in the concentration of sulfite, which was present in large excess, were negligible. At this stage, no correction was made for the state of ionization of sulfite at pH 7. From the above concentrations, apparent equilibrium constants for the formation of each adduct, at each time point, were calculated using Equation 2(17).

$${\text{K}}_{\text{app}}={[\text{X}]}_{\text{t}}/{[\text{nucleotide}]}_{\text{t}}{[\text{sulfide}]}_{\text{t}}$$ (Equation 2)

where K_app is the apparent equilibrium constant for formation of the adduct at time t. [X]_t, [nucleotide]_t, and [sulfite]_t are the concentrations of adduct, unreacted nucleotide and sulfite at that time point.

Kinetics of Sulfite Addition

To reduce interference between the UV absorbance of sulfite and the UV absorbance of the nucleotides in experiments to determine the rate of sulfite addition to UMP and OMP spectrophotometrically, it was necessary to reduce the sulfite concentration to values lower than those used in the preparative work,. Reaction mixtures containing UMP or OMP (10⁻⁴ M) and sulfite (pH 7.0, 0.06–0.50 M) were monitored at 276 nm, using a Hewlett-Packard 8452A diode array spectrophotometer.

ODCase Inhibition

The action of ODCase was monitored spectrophotometrically at 279 nm by published procedures (5, 18) with minor modification. The reaction was initiated by adding ODCase (10 μL, ~3 x 10⁻⁹ M in monomer, in 0.1 M NaCl) to a mixture (1 ml) containing OMP (1 – 8 x 10⁻⁵ M) and MOPS buffer (0.01 M, pH 7.2). Hewlett-Packard UV-Visible ChemStation software was used to obtain a zero order fit to absorbance data. Plots of [substrate]/velocity vs. [substrate] were constructed using Excel. Solving the equations for y = 0 yielded the x-intercept, representing –K_m (or – K_p when the inhibitor was present) and the slope yielded 1/V. In each case, inhibition was found to be competitive (see Results), and inhibition constants were calculated using equation 3:

$${\text{K}}_{\text{i}}=\text{i}/[({\text{K}}_{\text{p}}/{\text{K}}_{\text{m}})-1]$$ (Equation 3)

where: i is the concentration of the adduct being tested, K_p is the apparent K_m value of OMP in the presence of inhibitor, and K_m is the K_m value of OMP in the absence of inhibitor.

Results

¹H NMR of UMP and OMP Adducts

The proton resonances observed for the sulfite addition products of UMP and OMP are listed in Tables 1 and 2. In the case of UMP-SO₃, the signals arising from the vinyl protons at C₅ and C₆ of the pyrimidine ring are replaced by those of the aliphatic protons of the 5,6-dihydro species. Because chiral centers are present in both substituent ribose and C₆ of the saturated pyrimidine ring, diastereomeric products would be expected to result from sulfite addition (9, 17, 19) and were observed in the present experiments. Here, we refer to these diastereomers as I and II, the resonances of I appearing downfield from those of II. The non-equivalent protons at C₅ (A and B) and the single proton at C₆ (X) constitute an ABX spin system.

Table 1.

¹H NMR resonances for sulfite addition to UMP ^a,^b

	5(H), 5(H), 5,6-dihydro-UMP-6-SO3 (H-UMP-SO3)
	Diastereomer I (49.4%)				Diastereomer II (50.6%)
	C1′H	C6HX	C5HA	C5HB	C1′H	C6HX	C5HA	C5HB
Chemical Shift Coupling, Hz	5.81(d) 5.5	5.04(d) 7.1	3.30(d d) 7.1, 17.7	3.07(d) 17.7	5.25(d) 1.8	4.91(d) 6.7	3.18(d d) 6.8, 17.7	3.03(d) 17.7
Separation of C5Hs, Hz			115				75
	5(D), 5(H), 5,6-dihydro-UMP-6-SO3 (D-UMP-SO3)
	Diastereomer I (52.2%)				Diastereomer II (47.8%)
	C1′H	C6HX	C5HA	C5HB	C1′H	C6HX	C5HA	C5HB
Chemical Shift	5.81(s)	5.04(s)		3.07(s)	5.26(s)	4.91(s)		3.05(s)
Coupling, Hz	5.0	N/A		N/A	N/A	N/A		N/A

^aChemical shifts, multiplicity, and coupling constants for comparable UMP protons are: C_1′H, 5.96(d) 6.05 Hz; C₅H, 5.95(d) 8.5 Hz; C₆H, 8.1(d) 8.5 Hz. s = singlet, d = doublet, dd = doublet of doublets.

^bComposition of adducts mixture is given in the percentages for each diastereomer.

Table 2.

¹H NMR resonances for sulfite addition to OMP ^a,^b

	5(H), 5(H), 5,6-dihydro-OMP-6-SO3 (H-OMP-SO3)
	Diastereomer I (35.6%)			Diastereomer II (64.4%)
	C1′H	C5HA	C5HB	C1′H	C5HA	C5HB
Chemical Shift Coupling, Hz	5.18(d) 2.3	3.32(d) 17.4	3.27(d) 17.4	5.08(s) N/A	3.40(d) 17.6	3.26(d) 17.6
Separation of C5Hs, Hz		24.8			66.2
	5(D), 5(H), 5,6-dihydro-OMP-6-SO3 (D-OMP-SO3)
	Diastereomer I (29.4%)			Diastereomer II (70.6%)
	C1′H	C5HA	C5HB	C1′H	C5HA	C5HB
Chemical Shift	5.14(d)		3.24(s)	5.02(s)		3.2(s)
Coupling, Hz	2.3		N/A	N/A		N/A

^aChemical shifts, multiplicity, and coupling constants for comparable OMP protons are: C_1′H, 5.45(d) 2.4 Hz; C₅H, 5.5.71(s). s = singlet, d = doublet.

^bComposition of adducts mixture is given in the percentages for each diastereomer.

^cThe deuterated OMP-SO₃ contained a small amount of protio OMP-SO₃, rendering it difficult to obtain good integrated areas for C₅ protons. However, the relative integrated intensities of the C_1′and C₅ protons were comparable in magnitude

Sulfite addition to UMP was accompanied by a change in the chemical shift of the C₆ proton, and two new resonances were observed. The signals attributed to C₅ (A and B) each exhibited a doublet structure, with a characteristically large geminal coupling constant (~17 Hz) like those of the C₅ protons of 5,6-dihydroorotic acid and its methyl ester (20). One of the C₅ protons (C_5A) was also coupled to the C₆ proton, generating a doublet of doublets. Figures 1A and 1B show the 600 MHz spectrum of the UMP adducts.

Figure 1.

600 MHz ¹H NMR spectrum of sulfonate adducts of UMP. A: Protio-UMP-SO₃ prepared in H₂O, spectrum recorded in D₂O. B: Deutero-UMP-SO₃ prepared in D₂O, spectrum recorded in D₂O. The two diastereomers of each adduct are indicated by I and II. Resonances of the C₁′, C₅, and C₆ protons are identified for each diastereomer. Note the doublet of doublets in the protio adduct for the C_5A proton which was added concurrently at C5 with the addition of SO₃ at C₆. The C_5A proton is coupled to the C_5B (large splitting) and to C₆ proton (small splitting). In the deutero adduct, the C_5A proton was replaced by deuterium in the course of the addition reaction.

Table 1 shows the ¹H NMR data for the protons at C₅, C₆ and C_1′ of the ribose of the UMP adducts along with the relative concentrations of the two diastereomers of the adduct, based on comparison of the integrated resonances corresponding to the protons at C₆, C₅, and C_1′. For simplicity, the 2′, 3′, 4′ and 5′ ribose protons are not included in Table 1.

When the sulfonate adduct of UMP was prepared in D₂O, the proton NMR spectrum (Figure 1B) showed that deuteration had occurred at C₅. The complex doublet of doublets, corresponding to the C_5A proton of each diastereomer was lost and two singlets remained at 3.07 and 3.05 ppm, corresponding to the C_5B protons. In addition, the resonances of the C_6X protons collapsed to singlets, because the coupling to the deuteron at C₅ is very small. That observation confirmed the assignment of the complex doublet of doublets at 3.35 and 3.25 ppm to the proton added during sulfite addition. The anomeric C_1′ proton was unaffected by deuteration at C₅. The change in the relative populations of the two diastereomers, between the proteo and deutero adducts, was very slight, amounting to ~ 3%.

The spectra of the sulfonate adducts of OMP show the absence of a resonance for a C₆ proton, or coupling to a C₅ proton, confirming that sulfite addition occurred at C₆ (as in the case of UMP) rather than at C₅. Proton resonances were observed that correspond to two diastereomers, for the anomeric ribose C_1′ proton and the C₅ protons (Figure 2A). The C₅ protons of the adducts also showed the ~17 Hz geminal coupling constant observed for the UMP-SO₃ adducts in the present work, and as reported earlier for 5,6 dihydroorotic acid (20). In the OMP adducts, the signals arising from the C₅ protons (A and B) of the two diastereomers overlapped extensively, with less separation between the doublets of the two C₅ protons than in the case of the UMP adducts, so that the diastereomer I doublets lie within the larger doublets of diastereomer II. The preparation of OMP-SO₃ in D₂O resulted in the expected loss of the C_5A protons of the two diastereomers; and the C_5B proton doublets each collapsed to a single resonance, shifted from their original positions in the protio OMP-SO₃ adducts as shown in Figure 2A and B and Table 2. Unlike UMP-SO₃ in which the two diastereomers were present in essentially equal amounts, the OMP-SO₃ adducts exhibited a distinct preference for diastereomer II over diastereomer I. Thus, the introduction of a second bulky group at C₆ (the carboxyl and sulfonyl groups), results in a preference for one diastereomer over the other.

Figure 2.

500 MHz ¹H NMR spectrum of sulfonate adducts of OMP. A: Protio-OMP-SO₃ prepared in H₂O, spectrum recorded in D₂O. B: Deutero-OMP-SO₃ prepared in D₂O, spectrum recorded in D₂O. The two diastereomers of each adduct are indicated by I and II. Resonances of the C_1′ and C₅ protons are identified for each diastereomer. Note the two doublets in the protio adduct for the C_5A and C_5B protons for diastereomers II (with the large separation and marked with asterisks below the peaks), while the two doublets for the C_5A and C_5B protons for diastereomer I are closer together, and are within the resonances of diastereomers II. In the deutero adduct, the C_5A proton was replaced by deuterium in the course of the addition reaction.

The absolute axial/equatorial conformations of substituents in these sulfonate adducts remain to be determined. In dihydro derivatives of uracil and orotic acid derivatives, Katritzky et al. (20) found that C₆ carboxyl groups preferred the axial position even as their methyl esters, whereas C₆ methyl groups in dihydrouracils preferred the equatorial position. Those authors concluded that the shapes and sizes of substituent groups are important in determining conformational equilibria (20). The X-ray crystal structure of sodium 5,6-dihydrouracil-6-sulfonate, reported by Barnes and Hawkinson (21), shows the C₆ sulfonate group is attached axially to the saturated pyrimidine ring. It seems reasonable to conjecture that, in the present UMP-SO₃ adducts, the sulfonyl group tends to adopt an axial position.

Figure 3 shows that sulfite addition to UMP proceeds much more rapidly than sulfite addition to OMP. Sulfite addition to UMP proceeded nearly to completion at room temperature (Figure 3A), yielding the two diastereomers in constant proportions that remained unchanged after prolonged freezing. In the case of OMP, the ratio of diastereomers remained constant during the course of the reaction at 40^oC, but the proportion of diastereomer II increased considerably after frozen storage (Figure 3B).2, 3

Figure 3.

Figure 3A: Addition of sulfite to UMP monitored by ¹H NMR. A solution of UMP (0.052 M) was reacted with NaHSO₃ (0.6 M) at pH 7. The first three days of reaction were conducted at 25^oC, followed by 8 days at −20^oC. Note the constant proportion of the two diastereomers.

Figure 3B: Addition of sulfite to OMP monitored by ¹H NMR. A solution of OMP (0.028M) was reacted with NaHSO₃ (2.0 M) at pH 7. The first five days of reaction were conducted at 40^oC, followed by 17 days at −20^oC. Note the constant amount of diastereomer I and the increasing amount of diastereomer II observed upon frozen storage.

Equilibria of sulfite addition were very different for UMP and OMP. Apparent equilibrium constants observed for sulfite addition to OMP were 1.9 and 4.9 M⁻¹, respectively, for diastereomers I and II (Figure 4). Apparent equilibrium constants for sulfite addition to UMP were 26 and 27 M⁻¹ for diastereomers I and II respectively. Earlier, Triplett, et al. (17) used ¹³C NMR resonance intensities to estimate equilibrium constants of 6.9 and 7.2 M⁻¹ for addition of sulfite (1 M) to UMP (1 M) after 24 hrs at room temperature. The conditions employed by those investigators were different from those in the present experiments, in which the concentration of nucleotide was very much lower. Figure 4

Figure 4.

Mean apparent K_eq values for sulfite addition to UMP and OMP calculated from the ¹H NMR time course data in Figures 3A and 3B. UMP-SO₃ adduct diastereomers, present in approximately equal amounts as the reaction goes to completion, have nearly equivalent K_eq values. OMP-SO₃ shows a preference for diastereomer II, which is reflected in different K_eq values.

Rates of sulfite addition

To minimize the effect of sulfite absorption, UV changes were monitored at 276 nm. Whereas UMP reacted with sulfite nearly to completion in the presence of 0.25 M sulfite after 5 hours, the reaction with OMP had progressed only ~ 4% after 7 hrs (Figure 5). The difference between the reaction rates of UMP and OMP remained constant over the range of sulfite concentrations tested, UMP reacting approximately 37-fold more rapidly than OMP (Figure 6).

Figure 5.

UV absorbance changes measured at 276 nm during the reaction of NaHSO₃ (0.25 M) at pH 7 with UMP or OMP (1 x10⁻⁴ M) monitored over 5 and 7 hours respectively.

Figure 6.

Comparison of t_1/2 for the reaction of UMP or OMP over a range of NaHSO₃ concentrations at pH 7. OMP reacted with sulfite about 37 fold slower than UMP.

Inhibition by Unreacted Sulfite

Inorganic sulfite was evaluated as a potential inhibitor of ODCase over a concentration range of 0.02 to 0.10 M, well in excess of the level of unreacted sulfite that might remain as a contaminant in the assays from adduct solutions that were tested for nucleotide inhibition ( Figure 7). These experiments established an average K_i value of 2.4 mM for sulfite, several orders of magnitude higher than the K_i values of the nucleotide-sulfite adducts. Those findings render it unlikely that residual sulfite could have been responsible for the inhibition observed in the present experiments.

Figure 7. Sulfite Inhibition of ODCase.

Potential inhibition of ODCase by excess sulfite present in the adduct inhibitor solutions was evaluated at final in assay concentrations of 20, 50, and 100 mM. These levels are far higher than the ~3 mM present in the H-UMP-SO₃ and ~2.1 to ~6.2 mM present in the H-OMP-SO₃ inhibition assays. The weak inhibition observed from sulfite has an average K_i value of 2.4 ± 1.6mM.

Inhibition by Sulfonate Adducts of UMP and OMP

Before testing the sulfonate adducts of UMP and OMP for inhibitory activity, several potential pitfalls were considered.

1. The nucleotide-sulfonate adducts might be intrinsically unstable. However, solutions of nucleotide adducts at concentrations of 10⁻³ M or greater showed no signs of decomposition after 1 day at room temperature, while enzyme activity measurements were conducted over a period of minutes.

2. The nucleotide-sulfonate adducts might dissociate spontaneously at the high dilutions that were present in the inhibition experiments. That was shown not to be the case by performing control experiments in which adducts were diluted into an assay buffer at concentrations equivalent to the concentrations of substrate that would be used in assays of activity. When the absorbance at 279 nm was monitored over 10 minutes, exceeding the time for the enzyme assay, no change was observed. If the adducts had been dissociating with release of sulfite, an increase in absorbance at 279 nm would have been observed, due to formation of the C₅=C₆ double bond.

3. ODCase might use the sulfonate adduct of OMP as a substrate, with concomitant release of sulfite. That possibility was tested by incubating the enzyme with a reaction mixture containing buffer and the OMP-sulfonate adduct, but without the addition of substrate OMP. No change in absorbance was observed. If ODCase had acted on the sulfonate adduct of OMP (or catalyzed the release of sulfite from the adduct), followed by reaction of the enzyme with the OMP released, then one would have expected a decrease in absorbance associated with the breakdown of OMP. The total absence of such changes indicated that none of the sulfonate adducts is a substrate for ODCase.

After these control experiments had been performed, the diastereomeric mixtures of the sulfonate adducts formed from UMP and OMP were tested as inhibitors of ODCase (Table 3). In both cases, plots of [S]/v vs. [S] showed parallel lines in the presence of inhibitors, indicating competitive inhibition (Figures 8 and 9). K_i values for UMP-SO₃ adducts were about 7-fold lower than the K_i value for UMP (Table 3). The inhibition produced by sulfonate adducts of UMP was free from potential artifacts arising from unreacted UMP, since UMP is bound by the enzyme much less tightly, and any residual UMP would have been present at vanishingly low levels.

Table 3.

Inhibition of ODCase by Sulfonate Adducts of UMP and OMP

Derived from	Sulfite Adduct	Ki μM
UMP Ki 200 μM	UMP-SO3	29.2 ± 6.8
OMP Km 0.7μM	OMP-SO3	20.80 ± 5.7

Figure 8. Inhibition of ODCase by 200 μM H-UMP-S03.

ODCase hydrolysis of OMP in the absence and presence of 200 μM H-UMP-SO3. The average Ki was 29.2 ± 6.8μM.

Figure 9. Effect of Residual OMP in H-OMP-SO₃ on K_i.

¹H NMR spectra showed that this prep of H-OMP-SO₃ contained ~5% residual OMP. After acquiring the rate data, the S/V vs. S plots were constructed assuming “pure” (A) and “95% pure” (B) adduct was present in the incubation mixture. In the latter case the concentration of the adduct was reduced by 5% and that 5% was added to the OMP substrate added to the assay mixture. The average of the three K_i values in each situation was 20.8 ± 5.8 μM assuming pure adduct and 15.9 ± 3.5 μM for 95% pure adduct. Since the standard deviations of these two K_i values show considerable overlap, this would suggest there is no real effect from the small amount of OMP in the adduct. This is consistent with other tests performed as described in the Results section.

The sulfonate adducts of OMP-SO₃ were bound with higher affinity than those of UMP, but less tightly than OMP itself ⁽K_s = 7 x 10⁻⁷ M)(4, 18). For the sulfonate adducts of OMP, unreacted OMP might interfere with the assay, at least in principle, since its K_m value is lower than the K_i values of the adducts. However, as noted above, no change in absorbance at 279 nm was observed when OMP adducts were exposed to the enzyme in the absence of added OMP substrate. Further, when K_i values were recalculated to allow for residual OMP (~5%) that might be present in the adducts, the change in the calculated K_i value was less than the standard deviation (Figure 9B). Those findings suggest that the inhibition reported for the OMP-SO₃ adducts was not due to contaminating OMP.

Discussion

Comparison of binding affinities

The sulfonate adducts of both UMP (UMP-SO₃) and OMP (OMP-SO₃) are effective competitive inhibitors of ODCase, with K_i values that range between 2.0 to 2.9 x 10⁻⁵ M. UMP SO₃ can be considered an analogue of OMP, with a saturated C₅–C₆ bond and a sulfonyl group replacing the 6-carboxyl group. UMP-SO₃ is bound by yeast ODCase with an affinity (K_i = 2.9 x 10⁻⁵ M) that is higher than that of the reaction product UMP (K_i 2 x 10⁻⁴ M) but lower than that of 5,6-dihydro-OMP (K_i = 2.5 x 10⁻⁷ M). In view of the bulk and electrostatic charge of its two 6-substituents, we had not anticipated that OMP-SO₃ would be formed spontaneously, or that it would be bound with reasonably high affinity by the enzyme (K_i = 2.0 x 10⁻⁵ M). On the basis of the electrostatic stress hypothesis, OMP-SO₃ might have been expected to have been excluded from the active site. Instead, OMP-SO₃ is bound more tightly than UMP-SO₃, with only a single charged substituent.

Stereochemical preferences

Figure 10 compares the affinities of several pyrimidine nucleotide inhibitors of ODCase with the affinities of the substrate and product of decarboxylation, shown at the bottom. The inhibitors are shown in two groups, with unsaturated C₅=C₆ bonds at the top and with saturated C₅ – C₆bonds in the middle.

Figure 10.

Dissociation constants for UMP derivatives bound by ODCase shown in the known or predicted solution conformation, syn (black structures) or anti (blue structures) about the glycosidic bond (the vertical bar below N1 in each structure to C_1′ of the phosphoribosyl moiety, abbreviated as R₁). The nucleotides are in three groups: substrate and product (bottom row); 5,6 saturated or 5,6-dihydro nucleotides (middle row); and 5,6 unsaturated nucleotides (two top most rows). Two of the tightest binding nucleotides, BMP (K_D 9 x 10⁻¹² M⁻¹(45) and 6-thiocarboxamido-UMP (K_D 3.5 x 10⁻⁹ M⁻¹)(46) are not shown. Dissociation constants (M⁻¹) for the nucleotides in the figure are: 4-thioOMP, 2 x 10⁻⁶ (29); 6-methylaminoUMP, 3.3 x 10⁻⁶ (7); 6-carboxamidoUMP, 6x 10⁻⁴ (46); 6-azaUMP, 6 x 10⁻⁸ (4); 4-thioUMP, 1.5 x 10⁻⁶ (31); 6-cyanoUMP, 2.9 x 10⁻⁵ (42); 2-thioUMP, 2.3 x 10⁻⁵ (31); 2-thioOMP, >10⁻⁴(30); 5,6-hydroOMP 2.5 x 10⁻⁷ (47); 5,6-dihydroOMP-6-SO₃, 2 x 10⁻⁵ (this work); 5,6-dihydroUMP-6-SO₃, 2.9 x 10⁻⁵ (this work); OMP, 7 x 10⁻⁷ (4); UMP, 2 x 10⁻⁴ (4).

Each of the three 5,6-saturated inhibitors of ODCase (5,6-dihydro-OMP, 5,6-dihydro-UMP-SO₃ and 5,6-dihydro-OMP-SO₃) is a mixture of two diastereomers with a chiral center at C₆. It remains to be established whether one or both diastereomers of these three compounds are bound at the active site. In each of these molecules, the pyrimidine rings is somewhat puckered because of the presence of a saturated C₅–C₆ bond, and the C₆ substituents may be axially or equatorially oriented with respect to the dihydropyrimidine ring. Earlier analysis (20) of dihydroorotic acid showed that the carboxylate group tends to prefer an axial position; and the crystal structure of 5,6-dihydrouracil-6-sulfonate (21) shows the sulfonate group axially oriented at C₆. It seems reasonable to suppose that the same may be true of the sulfonyl group of UMP-6-SO₃ and of the carboxylate of 5,6-dihydro-OMP. In the case of OMP-SO₃, there seems to be no obvious basis for predicting the relative preferences of the –SO₃- or –CO₂^- groups.4

At the glycosidic C_1′-N₁ bond, pyrimidine nucleotides may adopt either the syn (with the C₂ carbonyl over the ribose ring) or anti (with C₆ over the ribose ring) conformation (22). The major structural features of pyrimidine nucleotides that are known to determine their intrinsic syn/anti preference include steric interactions of the C₂ and C₆ substituents on the pyrimidine ring, on the one hand, with the C_2′^, C_3′, C_4′ protons and C_5′ phosphoryl group of ribose, on the other (23–27). Thus, 5-methyluridine has been shown to adopt the anti conformation in which the C₂ carbonyl group is rotated away from substituent ribose. But in contrast, 6-methyluridine adopts the synconformation (with the C₂ carbonyl group situated over ribose) because of unfavorable interactions between the 6-methyl group and the ribose C_1′ proton (26). Most of the molecules shown in Figure 10 are equipped with C₆ substituents, and would therefore be expected to prefer the syn conformation. Moreover, ¹H NMR studies have shown that the substrate OMP (with a carboxylate group at C₆) adopts the syn conformation in solution, whereas the product UMP (with only a proton at C₆) adopts the anti conformation (25). In all three of the 5,6-dihydro inhibitors, the presence of one or more bulky 6-substituents would be expected to force the nucleotide into the syn conformation in solution.

In the crystal structures of ODCase complexes with UMP or OMP (in the latter case, using an inactive mutant version of the protein was used), both nucleotides are bound in the syn conformation (6). Thus, ODCase appears to bind OMP with its C₂ carbonyl group situated over the ribose ring, while the C₅=C₆ double bond carrying the scissile carboxylate group faces the catalytic quartet of charged residues. At the active site, decarboxylation presumably generates UMP as the immediate product, in the syn conformation that is thermodynamically disfavored in solution. With its release from the active site, UMP reverts to the anti conformation that is favored in solution.

The nucleotides in Figure 10 are shown in their known or predicted solution conformations (syn or anti). Of the inhibitors in Figure 10, all but 6-azaUMP carry a substituent at C₆ and would be expected to be bound syn conformation. Like UMP, 6-azaUMP prefers the anti conformation in solution (26). Interestingly, 6-aza-UMP is bound by ODCase in the syn conformation unless Ser-127 has been truncated to alanine. Pai and his associates have shown that the S127A mutant protein derived from the ODCase of M. thermoautotropicum appears to bind 6-azaUMP in the anti conformation (28). These authors suggest that the hydroxyl group of Ser-127 forms a strong H-bond to N₃ of 6-azaUMP, and that a backbone –NH group forms an H-bond to the 4-carbonyl oxygen atom, leading to a preference for the syn conformer.

Investigating the effects of sulfur substitution, Shostak and Jones (29) showed that 4-thioOMP exhibits a K_m value of 2 x 10⁻⁶ M and a k_cat value approaching that of OMP. In contrast, 2-thioOMP exhibits a K_i value of 2.9 x 10⁻⁵ M and is not decarboxylated by the yeast enzyme. More recent experiments indicate that 2-thioOMP is bound with a K_i value greater than 10⁻⁴ M (30). Smiley and Saleh have attributed the weak binding of 2-thioOMP to its binding in the “wrong” anti conformation (31). These investigators have also shown that both 2- and 4-thioUMP derivatives are competitive inhibitors of ODCase, with K_i values of 2.3 x 10⁻⁵ M and 1.5 x 10⁻⁶ M, respectively. The differences in binding affinity between 2-thio and 4-thio derivatives of UMP and OMP have been analyzed computationally by Phillips and Lee (32) in terms of the relative stabilities of the syn and anti conformers and the barriers to rotation around the glycosidic bond observed in tetrahydrofuryl derivatives. These authors suggest that 2-thioUMP, 4-thioUMP, and 4-thioOMP exhibit relatively high affinity for the active site because they can be bound in the preferred syn conformation. 2-ThioOMP, which which adopts the anti conformation in solution and exhibits the highest energy barrier to rotation about the glycosidic bond, cannot rotate easily to the syn conformation that may be needed for optimal binding and catalysis. 2-ThioOMP appears to be the one case in which the bulk of the carboxyl group at C₆, relative to that of the C=S group at C₂, is insufficient to favor the required syn conformation.

The three uridine nucleotides with 6-substituents of varying bulk — the 6-cyano, 6-carboxamido and 6–thiocarboxamido derivatives of UMP — presumably prefer the syn conformation in solution due to the bulk of their C₆ substituents. 6-CyanoUMP has been reported by Fujihashi et al. (33) to be “bound in the active site in a manner analogous to that of UMP, 6-azaUMP, and BMP”, i. e. the syn conformation. Although the structures of their enzyme complexes remain to be established, the 6-carboxamide and 6-thiocarboxamide nucleotides would also be expected to favor the syn conformation strongly. An additional pyrimidine nucleotide whose association with ODCase has been examined crystallographically is cytidine 5′-phosphate (CMP) and its 6-carboxylate derivative. CMP adopts the anti conformation in solution (26) and also within the active site of ODCase where it is very weakly bound (31). The 6-carboxylate derivative of CMP was found to be a very poor substrate (roughly 10⁵ less active than OMP) and was only weakly bound (31). The weak binding of CMP and CMP-6-carboxylate was ascribed to poor anchoring of the 2′-OH of ribose and to improper positioning of the pyrimidine ring, caused by the presence of the C₄-amino group, resulting in greater mobility within the active site.

In the enzyme from a thermophilic bacterium, Pai et al. (35) have shown that nucleotides containing xanthine or cytosine are bound in a different way from uracil-based ligands. As noted by the authors, those differences presumably arise from the inordinate size of the purine ring of xanthine and the inability of the cytosine ring to form the conserved H-bonding interactions between the enzyme and the 4-carbonyl group that stabilize the complexes formed by BMP, OMP and UMP in the published crystal structures of OMP decarboxylase. Instead, the cytosine ring with its 4-amino group is likely to be repelled. Presumably for these reasons, the rings of XMP and CMP are turned away from the binding site normally occupied by the uracil pyrimidine ring, so that the phosphate-binding loop does not close properly; the bases X and C find new H-bonding partners outside the site normally occupied by the substrate and product; and the void remaining at that site is occupied by a molecule of butanediol (in the case of XMP) or 5 water molecules (in the case of CMP).⁵

The most tightly bound pyrimidine nucleotide inhibitor of ODCase is barbituric acid ribofuranoside 5′-monophosphate (BMP), with an estimated K_i value of 9 x 10⁻¹² M⁻¹. This triketopyrimidine presents a complex conformational problem, in that the keto groups at C₂, C₄ and C₆ render this molecule nearly symmetrical along the N₁-C₄ axis. In principle each of those keto groups might exist at least to some extent as the hydroxyl tautomer. The accepted structure in solution is equipped with an ionized hydroxyl group (pK_a ~2.8) at C₆ and keto groups at C₂and C₄. BMP can exist in either the syn or anti conformation, and ¹H NMR data indicate that the anti conformer predominates in aqueous solution (26). In its preference for that anti conformation, BMP resembles β-cyanuric acid ribonucleoside, which is similar in structure to barbituric acid ribonucleoside, but with a third nitrogen atom replacing C₅ of the pyrimidine ring

Phillips and Lee’s (32) calculations indicate that the syn and anti conformers of BMP are similarly favorable energetically and are separated by a high barrier to rotation about the glycosidic bond. When bound at the active site of ODCase, BMP has been found to be in the syn conformation, which orients the C₆ oxyanion facing the catalytic tetrad of alternating aspartate and lysine residues (1). It is possible that ODCase selects the syn conformer of BMP from solution, or that the enzyme induces the anti conformer of BMP to rotate about the glycosidic bond to be bound in the syn conformation. If BMP were bound to any extent as the rare tri-keto species, then C₅ would carry two geminal protons (as in the three 5,6-dihydro inhibitors) and the true K_d value of that rare species would be even lower than the measured K_i value. The geminal C₅ protons of the triketo form would be difficult to observe by ¹H NMR spectra because they would presumably be subject to solvent exchange through keto-enol equilibration (39, 40).

Based on these considerations and on modeling several complexes of ODCase from M. thermoautotrophicum containing various uracil derivatives (Supporting Information), we present hypothetical structures of UMP-SO₃ and OMP-SO₃ bound within the active site of the enzyme (Figure 12). These structures are derived from the well characterized complex of ODCase with BMP and structural data for 5,6-dihydrouracil-6-sulfonate (21). The puckered dihydropyrimidine ring of each adduct is oriented such a way as to be superimposed on the pyrimidine ring of BMP as closely as possible. The phosphoribosyl groups are not altered. The sulfonate group is axial to the puckered dihydropyrimidine ring. Negatively charged sulfonate and carboxylate groups are located close to Lys-72 (within ~3Å) and are likely to experience a strong Coulombic interaction. When the protein surface was included in parts B and C of Figure 12, the sulfonate group projected slightly into the protein surface in the vicinity of Lys-72. The other protein side chain close to the sulfonate group is Met 126, but this sidechain is flexible. In view of the flexibility in the lysine and methionine side chains, there appears to be little steric impediment with the axial sulfonyl group. These tentative structural conjectures remain to be tested by exact structural methods.

Figure 12. Hypothetical binding of the UMP-SO₃ and OMP-SO₃ inhibitors in the active site of ODCase based on the binding of BMP.

(A) Shows BMP bound in ODCase from M. theromoautotrophophicum from reference 35. The protein surface around Lys-72 has been removed to expose the amino acid side chain. Met-126 is inside the surface above carbons 4 and 5 of the pyrimidine ring. (B) UMP-SO₃ modeled into the active site using BMP as the basic structure, with modification of the pyrimidine ring and the position of the sulfonate group based on the x-ray structure of sodium 5,6-dihydrouracil-6-sulfonate from reference 21. The sulfonate group is axially oriented coming up from the pyrimidine ring toward the viewer. Two of the three sulfonate oxygens are within 3 Å of the nitrogen of the NH₂ group of Lys-72. The C₆proton is shown in tan color. (C) OMP-SO₃ modeled into the active site using the structure in B with the carboxyl group replacing the proton on C₆ in the equatorial position and retaining the sulfonate group in the axial position. In both B and C the negatively charged sulfonate and carboxylate groups are in close proximity to the positively charged Lys-72. An alternative structure with the pyrimidine ring puckered in the opposite direction for both adducts caused unfavorable interactions between the sulfonate group and a proline residue in the bottom of the cavity.

Conflicting structural requirements of a commodious—yet discriminating—active site

In the absence of direct structural information, comparison of the behavior of the nucleotides discussed above suggests several inferences about the binding properties of the 6-sulfonate inhibitors described here. Based on their observed tight binding, it seems reasonable to suppose, that the sulfonate adducts of UMP and OMP are bound in the syn conformation, like OMP, UMP and other nucleotides that are tightly bound. Apparently the carboxylate binding pocket, which may be the initial location for the CO₂ molecule released from OMP, is large enough to accomodate a sulfonate group as well as a carboxylate group. The puckered 5,6-dihydropyrimidine ring may not position C₆ and its substituents as deeply in the carboxylate pocket as does the planar pyrimidine ring of OMP, with its coplanar carboxylate group. But 5,6-dihydro-OMP is bound with an affinity comparable with that of the substrate OMP, suggesting that the pocket can accommodate a non-coplanar 6-substituent. Saturation of the C₅ = C₆ bond of OMP converts the substrate into an inhibitor that does not undergo enzymatic decarboxylation. The sulfonate adduct of OMP is bound ~30 times less tightly than OMP itself, but the sulfonate adducts of UMP are bound about ~7-fold more tightly than product UMP. It seems evident that both negative charge at C₆ and C₅–C6 saturation of the pyrimidine are readily accommodated by the active site of ODCase.

The present observations, and an earlier report of the inhibition of ODCase by 6-methylaminoUMP (8), raise serious questions about the “electrostatic stress” mechanism as originally proposed (2). According to that mechanism, the carboxylate group of Asp-70 repels the 6-carboxylate group of OMP in such a way as to destabilize it, advancing the reaction toward a transition state from which CO₂ departs as the C₆ anion of UMP anion is generated. If electrostatic destabilization of the enzyme–substrate complex constituted an important part of such a mechanism, one might have expected that sulfonate adducts (especially OMP-SO₃) would tend to be excluded from the active site because of the greater size and charge density of the carboxylate and sulfonate substituents at C₆. Instead, these sulfonate adducts are bound almost as tightly as OMP and 5,6-dihydro-OMP. The behavior of the three 5,6-dihydro compounds indicates that ODCase can accommodate non-planar pyrimidine rings. Moreover, several of the pyrimidine nucleotides in Figure 10 are equipped with 6-substituents that vary in structure, number of atoms, and electrostatic charge, but are bound with similar affinities. Those are not the properties that one would expect of an active site with a rigid structure that imposes sufficient electrostatic stress on the substrate to produce a major advancement along the reaction coordinate.

In contrast to the overall similarity of the binding affinities summarized in Figure 10, the affinity of ODCase for the altered substrate in the transition state (which probably resembles a C₆ carbanion, in which an electron pair replaces the substrate’s C₆–COO⁻ group) is vastly higher than its affinity for OMP in the ground state. The remarkable magnitude of that increase in affinity (~ 10¹⁷-fold) requires that the enzyme-substrate interactions be extremely strong and intimate in the transition state. That unusual jump in affinity is partly reflected in the affinity of BMP (K_i = 10⁻¹¹ M), in which a C₆–CO⁻ group replaces the substrate’s C₆–COO⁻ group. What is most striking about this series (C⁻ > C–O⁻ > C–COO⁻) is that binding affinity increases very steeply with decreasing size, whereas electrostatic charge remains nominally unchanged. When the negative charge is separated from C₆ by 0, 1, or 2 bonds, the free energy of binding changes from –31.7 to –15.0 to –8.4 kcal./mole.

Perhaps these paradoxical properties of the active site (indiscriminate binding of substrates and their analogues in the ground state, but highly discriminating binding of the altered substrate in the transition state) can be reconciled if we suppose that in the transition state (to a much greater extent than in the ground state) the active site closes around the substrate in such a way as to provide the intimate and discriminating contacts that are required for catalysis (43). It seems reasonable to conjecture that these changes are reflected to some extent in the crystallographically observed ordering and closing of loop 207–217 of the yeast enzyme that accompanies the binding of BMP. After entry of the substrate, at a rate that approaches the diffusion limit (4), the closure of that loop permits new H-bonding interactions between the substrate phosphoryl group and the active site, notably Tyr-217. For such a loop movement to support catalysis, it is necessary that the cost of distorting the enzyme from its native “open” structure be less than the benefit of enhanced attractive forces between the enzyme and substrate in the transition state. That overriding benefit would seem to be possible if (a) loop closure involves the relative movement of parts of the active site that are intrinsically somewhat ordered (43), and (b) there is a high level of synergism between the forces of attraction that are work in the transition state, but not the ground state.

In apparent agreement with that sequence of events, the phosphoryl group of OMP has been shown by its presence to contribute a factor of 10¹¹ to the value of k_cat/K_m observed for OMP (5). Moreover, the phosphite ion has been shown to enhance, by a factor of 10⁸, the reactivity of a truncated form of the nucleoside orotidine (44). Just as ODCase produces one of the larger rate enhancements that is known to be produced by any enzyme, the substrate phosphoryl group (although spatially distant from the site of substrate decarboxylation) produces the largest effect of a simple substituent that appears to have been reported for any enzyme reaction. These effects, and the present findings, seem to be compatible with the view that much of the rate enhancement produced by this unusual catalyst arises mainly from its ability to stabilize the altered substrate in the transition state, rather than from electrostatic destabilization of the ground state enzyme-substrate complex.

Supplementary Material

1File003. SUPPORTING INFORMATION AVAILABLE.

Supporting information contains an analysis of the binding interactions between ODCase and five uracil based nucleotides as well as XMP and CMP to compare and contrast the interactions. This information supports the proposal that UMP-SO₃ and OMP-SO₃ are most likely bound like other uracil nucleotides, as illustrated in Figure 12. This material is available free of charge via the Internet at http://pubs.acs.org.

Figure 11.

Some of the conformers and ionic species of BMP. Although BMP is equally likely to be syn or anti in aqueous solution (Reference 31), the tri-keto species, shown on the left side of the figure, is only observed in non-aqueous solutions. The syn 6-anion of BMP (highlighted in red in the scheme below) is the only species bound within the active site of ODCase.

Table 4. Syn / Anti.

Conformation of UMP Derivatives in Solution and Bound within the Active Site of ODCase

Nucleotide	Solution Conformation	Dataa	Active Site Bound Conformation	Dataa
UMP	anti	NMR b,c,d Comp m	syn	X-ray e,f,g
6-carboxy-UMP (OMP)	syn	NMR c Compl	syn	X-ray g
6-carboxamido-UMP	syn	predicted	syn	predicted
6-cyano-UMP	syn	predicted	syn	X-rayk
6-aza-UMP	anti	NMR b Compm,o	syn/anti	X-ray f,g,k,n
6-thiocarboxamido-UMP	syn	predicted	syn	predicted
6-oxo-UMP or 6-hydroxy- UMP (BMP)	anti syn = anti	NMRb Compm	syn	X-ray g,h,i,j
5,6-dihydro-6-carboxy-UMP(5,6-dihydro-OMP)	syn	predicted	syn	predicted
5,6-dihydro-6-sulfonyl-UMP (UMP-SO3)	syn	predicted	syn	predicted
5,6-dihydro-6-sulfonyl-OMP(OMP-SO3)	syn	predicted	syn	predicted

^a”NMR” or “Comp” indicate that NMR or computational studies, from the references cited, have determined the solution conformation of the nucleotide. “X-ray” indicates that there is an x-ray structure of the nucleotide bound in the active site of ODCase, in the reference provided. “Predicted” indicates the solution conformation is anticipated based on analogy with 6-methyl uridine, a large group at C₆ would be expected to force the nucleotide into the syn configuration to avoid unfavorable interactions between the 6-substituent and the ribose protons and phosphoryl group.

^b– (26),

^c- (25),

^d- (23),

^e- (34),

^f- (2),

^g- (35),

^h- (36),

ⁱ- (37),

^j- (38),

^k- (6),

^l- (33),

^m- (32).

ⁿ6-aza-UMP has been shown to be bound by the S127A mutant ODCase in the anti-conformation.

^oNeutral 6-aza-UMP prefers the anti conformation, but its anion exhibits no preference according to computations (32).