Presence of tightly bound NAD+ in urocanase of Pseudomonas putida.

mass spectral analysis. The presence of a nicotinic acidcontaining component was confirmed by incorporation of [‘%Jnicotinic acid into urocanase during growth of a nicotinic acid auxotroph of Pseudomonas putida. Treatment of this purified [ ‘Wurocanase with perchloric acid released a nicotinic acid-containing cofactor which was subsequently determined to be NAD+ by electrophoresis and enzymatic analysis. Incorporation of tritium upon NaB”H, reduction of urocanase, microbiological analysis for nicotinic acid, and the specific radioactivity of [“Clurocanase all supported a stoichiometry of 1 mol of NAD+ boundlmol of native enzyme. The necessity of using denaturing conditions to remove enzyme-bound NAD+ indicated that the NAD+ is very tightly bound. The principal criteria employed earlier for the identification of cu-ketobutyrate as a prosthetic group in urocanase (George, D. J., and Phillips, A. T. (1970)5. Biol. Chem. 245, 52%537), namely similar chromatographic and electrophoretic properties for cy-hydroxybutyrate and “H material isolated from urocanase after NaB:‘H, reduction and acid hydrolysis, were shown to be unable to distinguish between these compounds and nicotinic acid, thereby accounting for the incorrect conclusion that cu-ketobutyrate was an essential cofactor. Present results indicate that NAD+ is a functional coenzyme in urocanase and thus raise the possibility that catalysis proceeds through an internal oxidation-reduction mechanism.

The reduction of urocanase with NaB"H, plus NaB"H, produced an inactive enzyme upon incorporation of 1 g atom of hydrogen/m01 of enzyme. From acid hydrolysates of this reduced enzyme was isolated labeled nicotinic acid which was identified by chromatography, electrophoresis, and mass spectral analysis. The presence of a nicotinic acidcontaining component was confirmed by incorporation of ['%Jnicotinic acid into urocanase during growth of a nicotinic acid auxotroph of Pseudomonas putida. Treatment of this purified [ 'Wurocanase with perchloric acid released a nicotinic acid-containing cofactor which was subsequently determined to be NAD+ by electrophoresis and enzymatic analysis. Incorporation of tritium upon NaB"H, reduction of urocanase, microbiological analysis for nicotinic acid, and the specific radioactivity of ["Clurocanase all supported a stoichiometry of 1 mol of NAD+ boundlmol of native enzyme. The necessity of using denaturing conditions to remove enzyme-bound NAD+ indicated that the NAD+ is very tightly bound. The principal criteria employed earlier for the identification of cu-ketobutyrate as a prosthetic group in urocanase (George, D. J., and Phillips, A. T. (1970)5. Biol. Chem. 245, 52%537), namely similar chromatographic and electrophoretic properties for cy-hydroxybutyrate and "H material isolated from urocanase after NaB:'H, reduction and acid hydrolysis, were shown to be unable to distinguish between these compounds and nicotinic acid, thereby accounting for the incorrect conclusion that cu-ketobutyrate was an essential cofactor.
Present results indicate that NAD+ is a functional coenzyme in urocanase and thus raise the possibility that catalysis proceeds through an internal oxidation-reduction mechanism.
Urocanase (4'-imidazolone-5'-propionate hydro-lyase, EC 4.2.1.49) was purified and crystallized from Pseudomonas putia earlier in this laboratory (1). Its inhibition by such common carbonyl-attacking reagents as hydroxylamine, borohydride, and phenylhydrazine led to the belief that pyridoxal 5'-phos- phate would be a required cofactor, as had been reported for chicken liver urocanase by Gupta and Robinson (2), but no evidence could be gathered to support this expectation. Instead, based on chromatographic and electrophoretic properties of material isolated from acid-hydrolyzed, NaB:'H,-reduced urocanase, a-hydroxy[:'H]butyrate was identified as a reduction product, thereby suggesting that the parent material was a covalently bound cu-ketobutyrate (1). This finding was attractive because supplemental evidence indicated that the phenylhydrazine adduct of urocanase closely corresponded in spectral properties with a keto acid phenylhydrazone, stoichiometry of 'H incorporation upon NaB:'H, reduction was approximately 1 g atom of hydrogen/m01 of enzyme, the reduced enzyme was stable to dialysis but radioactivity could be removed by hydrolytic treatment with acid or proteases, and finally, several other enzymes had been found to contain the related cofactor, pyruvate, as an apparent substitute for pyridoxal phosphate (3)(4)(5).
In recent years, it has been difficult to reconcile the expected properties of a covalently bound a-ketobutyrate moiety with an accumulation of data on the properties of enzyme modified with NaB"H, and 0-['Y!]methy1hydroxy1amine. ' We herein report the finding that urocanase from P. putida contains a tightly bound NAD+ moiety, the properties of which account almost completely for the previous data leading to the conclusion that an ru-ketobutyrate group was present in urocanase. etry determinations on each sample revealed that 1.0 c 0.1 (S.D., n = 3) g atom of hydrogen was incorporated per mol (110,000 g) of enzyme, in agreement with earlier results (1).
The procedure previously employed by George and Phillips (1) for isolation of a-hydroxy["Hlbutyrat involved a proteolytic digestion and HCl hydrolysis of "H-labeled enzyme, with subsequent identification based largely on the chromatographic properties of tritiated material remaining after removal of HCl. No accurate accounting of radioactivity recovered at this stage was conducted, nor was further purification attempted. We now report ( Table I) that careful monitoring of recoveries of 3H after each step of the isolation procedure revealed that significant amounts of tritium were lost (presumably to water) at all steps subsequent to the initial separation of protein from unbound radioactivity.
Standard cu-hy-  amined and only the component of retention time 11.2 min was deuterated. Its mass spectral pattern (Fig. 1, top panel) agrees well with that of the nondeuterated sample. The deuterium content, calculated from the relative heights of m/e 180 and ml e 181 after correction for the contribution of 13C abundance and natural silicon isotopes, was 30%; this value leaves little doubt that this material was derived from the borodeuteride reduction of urocanase. The difference between the observed value (30% 'H) and the predicted 50% 2H content is discussed below.
A search for compounds exhibiting the chromatographic and mass spectral properties of the unknown revealed only one compound which fit all characteristics.
The monotrimethylsilyl derivative of nicotinic acid chromatographed in a fashion identical with the unknown, and, as seen in Fig. 1  Chromatographic and Electrophoretic Comparisons of a-Hydroxybutyrate and Nicotinic Acid -The obvious dissimilarity between the structures of a-hydroxybutyrate and nicotinic acid raises the question of whether these two compounds could have been distinguished from one another by the chromatographic systems originally used for the identification of 01hydroxybutyrate derived from urocanase. In Table II, the  migration of authentic nicotinic acid, the ["Hlnicotinic acid isolated from urocanase and a-hydroxybutrate are compared in the chromatographic systems described by George and Phillips (1). In all systems, the RF values of a-hydroxybutyrate, nicotinic acid, and the tritiated material were indistinguishable. Electrophoresis at pH 8.9 likewise failed to separate these compounds. Of the original systems, only electrophoresis at pH 3.3 gave evidence that separation could be achieved. Because nicotinic acid is isoelectric at pH 3.4 (17) and the PK,~ of a-hydroxybutyrate is 3.65 (18), a slight variation of pH in this region could significantly affect the migration of both compounds and thus might explain the identical migration at pH 3.3 of the "H material and cr-hydroxybutyrate seen by George and Phillips (1). Electrophoresis at pH 2.2, a condition not originally studied, was particularly effective in separating these compounds (Table II) and clearly indicated that the "H unknown derived from urocanase was not a-hydroxybutyrate.
Microbiological Assay for Nicotinic Acid -To confirm the presence of nicotinic acid in urocanase and to determine its stoichiometry with respect to the enzyme, the microbiological assay developed by Snell(9) was performed on native urocanase. Bovine serum albumin, molecular weight 66,000 (19), was used as a control and was found to contain no nicotinic acid when assayed at the same molar concentrations as urocanase. Based on a molecular weight of 110,000 (l), native urocanase contained 1.3 ? 0.1 mol of "nicotinic acid"/mol of enzyme.
Difference Spectra between Borohydride-reduced and Native Urocanase-A portion (3.2 mg) of the urocanase which had been treated with NaB:'HB was taken after the gel filtration step (see Table I) and its ultraviolet absorption spectrum compared to that of native urocanase at the same concentration (Fig. 2). A peak was observed at 335 nm having an absorbance difference of 0.08 cm-l. The wavelength maximum of enzyme-bound NADH is commonly offset from that of unbound NADH (20). Based on an ext,inction coefficient of 6.2 x 10" M cm-' at 339 nm for free NADH, 0.45 mol of NADH was present/m01 of enzyme. This is significantly lower than the stoichiometry of 1.0 g atom of hydrogen bound/m01 of enzyme calculated from "H incorporation upon NaB3H, reduction; however, Chaykin and Meissner (21) demonstrated that borohydride reduction of NAD+ yields not only 1,4-NADH but also 1,2-and 1,6-NADH, the latter of which have absorption maxima at 395 and 345 nm, respectively.
Values for extinction coefficients of 1,2-and 1,6-NADH were not presented by Chaykin and Meissner (211, but their data permit the calculation that roughly 60% (2 10%) of the total NADH (mixed isomers) formed would be accounted for by absorbance measurements made at 340 nm, using 6.2 x lo3 Mm' cm-' as extinction coefficient. This follows from the fact that 1,4-NADH comprised only 29% of the recoverable NADH and 34% was 1,6-NADH, the latter exhibiting a ratio of 0.9 for absorbance at 340 rim/A at 345 nm. While the proportion of these forms of NADH resulting from borohydride reduction of urocanase is not known, it is probable that this mixture of reduction products accounts for the discrepancy in stoichiometries. 5 Electrophoresis of NaB"H,-reduced Material from Urocanuse Prior to Hydrolysis -A portion of NaB"H,-reduced urocanase was treated with NaOH by the procedure described by Burch et al. (13) to remove NAD+ and NADH from tissues. Electrophoresis of this material and of labeled (reduced) NAD which had been reduced and treated under the same conditions is shown in Fig. 3. The compound isolated from the enzyme and authentic reduced NAD showed similar patterns. In both samples, the largest proportion of the tritium label migrated with NADH and the remainder with nicotinamide and NAD+. The presence of the latter compound is expected because NADH is known to be reoxidized in very dilute solutions (16). In addition, the decomposition of both NAD+ and NADH is accelerated in bicarbonate buffer, the buffer used during chromatography to purify both the reduced urocanase and reduced NAD prior to NaOH treatment.
Evidence for the instability of NADH at neutral pH (16) plus the above observations on decomposition of NADH to NAD+ and nicotinamide can account for the appearance of ["Hlnicotinic acid in acid hydrolysates of NaB3H,-reduced urocanase; these facts also provide an explanation for the loss of tritium during the isolation of ["Hlnicotinic acid (Table I) and for the low "H content of ['Hlnicotinic acid obtained from the reduction of urocanase with NaBlH,. It is probable that the 5min boiling treatment of reduced enzyme in bicarbonate buffer at pH 8.0 (after the Sephadex G-25 step in Table I) resulted in extensive oxidation of NADH to NAD+, with the accompanying release of one hydrogen (or tritium) atom nonstereospecifi- 9 In resnonse to reviewer comments reauesting auantification of the relatite amounts of 1,4-, 1,2-, and 1;6-NAl?H' resulting from borohydride reduction of NAD+, we offer the following data. Reduction of 10 wmol of NAD+ with NaB3H, at pH 7.5, followed by chromatography on Sephadex G-10, resulted in the isolation of a peak fraction containing 3. tralization along with standard nicotinic acid-containing com-2 " pounds revealed that 88% of the counts migrated with NAD+ (Fig. 5). The other 12% of the '"C migrated in the region I ["C]urocanase was treated with perchloric acid and the isolated neutralized coenzyme (73% recovery, 49,200 cpm) tested for its extent of reduction by yeast alcohol dehydrogenase and tally from the site of borohydride reduction. Acid hydrolysis of ethanol. Portions of this extract equivalent to between 10 and this NAD+ would produce nicotinic acid of approximately one-40 nmol of urocanase were examined and reduction was comhalf the exnected 'H or "H content. The 2H content of the deuterated nicotinic acid examined by mass spectrometry was, plete after 10 min. Under these conditions, reduction of similar amounts of NADP+ was less than 4% complete, thereby indiin fact, 30% and nonstereospecilic release according to the cating that the reducible material was not NADP+. The assays above explanation would predict a 25% deuterium content in showed that 107 ? 3 (SD., n = 4) nmol of NAD+ were renicotinic acid.
covered from acid-treated enzyme or after correction for the  5. Electrouhoresis of the 14C-labeled material isolated from ["Clurocanase by*perchloric acid treatment. The isolation procedure was as described under "Experimental Procedures" and the electrophoresis was as in Fig. 3.
GRADIENT VOLUME (ml) FIG. 6. Elution profiles of standard pyridine nucleotide coenzymes and '"C material isolated from [*4Clurocanase. Chromatography was performed at 25" on a column (1 x 25 cm) of DE52 (bicarbonate form). After sample application, the column was washed with 25 ml of water, then eluted with a linear gradient (800 ml) of 0 to 0.2 M NHIHCOI, plus 0.1 mM dithiothreitol; 5-ml fractions were collected. Radioactivity (0) is expressed as counts per min per ml. heart L-lactate dehydrogenase (100 pg) and pig heart alanine aminotransferase (100 pg) in a final volume of 1.0 ml, according to the procedure of Lowry and Passonneau (231. The reaction course was followed by A,,, increase and an absorbance change corresponding to the production of 86 nmol of NADH was measured (predicted value, 96 nmol). DISCUSSION The results described clearly indicate the presence of 1 mol of NAD+ firmly bound to each mole of urocanase and illustrate how nicotinic acid produced from the degradation of NAD+ was mistakenly identified as a-hydroxybutyrate in our earlier investigations.
The almost complete reliance upon chromatographic and electrophoretic similarities between a-hydroxybutyrate and the 3H unknown obtained from reduced and hydrolyzed urocanase can now be seen to have been inadequate for purposes of structural analysis of the urocanase coenzyme, in that the criteria did not differentiate between a-hydroxybutyrate and nicotinate. Two other chemical properties of the urocanase coenzyme or its reduction product were examined in the original study but have not been dealt with here. The first was a difference spectrum between native urocanase and urocanase modified with phenylhydrazine.
This spectrum revealed a difference peak with an absorption maximum at 312 nm, similar to that seen for keto acid phenylhydrazones. We now believe that this peak could also have been due to a phenylhydrazine-NAD adduct, analogous to the well known reaction product between NH,OH or hydrazine and NAD+ (24). Because of instability of such adducts at neutral pH unless bound to an enzyme (25), we have been unable to confirm or disprove the identity of the 312 nm absorbing species. A second property, namely formation of a p-bromophenacyl ester derivative of the 3H material isolated from reduced and hydrolyzed urocanase, has been re-examined by US.~ While such a derivative can be prepared, its chromatographic properties do not conform to those reported earlier. In fact, the bulk of the 3H derivative remains at the origin in the solvent system originally used (as does the p-bromophenacyl ester of nicotinic acid), while only a trace quantity (<lo%) of the radioactivity can be found in the area corresponding to the migration position of a-hydroxybutyric p-bromophenacyl ester. We have therefore concluded that this evidence can no longer be used to support an identification of a-ketobutyrate as the urocanase coenzyme.
The extensive data for the presence of NAD+ in urocanase, coupled with data on the stoichiometry of tritium incorporation upon NaB3H, reduction, the microbiological assay for nicotinic acid (or nicotinamide coenzymes) and the incorporation of l'*C]nicotinic acid into urocanase during biosynthesis, together provide strong evidence for a 1:l NAD+:urocanase ratio and point to the existence of one binding site for NAD+ with an extremely low dissociation constant at neutral pH. However, urocanase has been shown to consist of two apparently identical subunits (26) and the question can be posed as to whether a second NAD+ site of reduced affinity might exist. Preliminary experiments involving incubation of urocanase with concentrations of NAD+ up to 1 mM reveal that no increase in specific activity can be detected.' Thus the presence of another NAD+ site has not been experimentally verified. Furthermore, treatment of urocanase with charcoal leads to irreversible binding of the enzyme to charcoal, thereby precluding any attempts to prepare apoenzyme for direct titration with NAD+ and to show a direct catalytic requirement for NAD+.
A central consideration concerns the essentiality for NAD+ in the catalytic process. The strongest evidence presently available comes from the potent inhibition of enzymatic activity by so-called "carbonyl-attacking reagents" (l), most of which are also known to modify NAD+ (21,24,27). In the case of NaBH,, it has now been shown that this inhibition is accompanied by the formation of NADH on the enzyme. The recognized enzymatic function of NAD+ is believed to be as a hydride ion acceptor for the oxidation of substrates. In urocanase, the mechanism must also include the reoxidation of enzyme-bound NADH so that active enzyme is re-established upon product release. The mechanism proposed for this enzyme (1,281 requires that the hydration of urocanic acid result from proton abstractions and additions rather than an intramolecular hydride ion transfer as would be implied by the presence of NAD+. Indeed, it is difficult to envision a hydride ion transfer mechanism for urocanase since the newly added side chain hydrogen atoms of imidazolone propionate were shown to have a solvent origin (28). Work in our laboratory is currently directed toward clarifying this possibly novel NAD+mediated mechanism.