Purification and properties of creatinine iminohydrolase from Flavobacterium filamentosum.

Creatinine iminohydrolase (EC 3.5.4.21), which catalyzes hydrolysis of creatinine to N-methylhydantoin and ammonia, was purified from Flavobacterium filamentosum. The average molecular weight of the purified enzyme was 272,480, and the subunit molecular weight was 44,300. Extensive specificity studies indicated that the enzyme utilized cytosine (Km, 0.62 mM; Vm, 20.1 units/mg) as well as creatinine (Km, 5.00 mM; Vm, 40.4 units/mg) as a substrate. Each was a competitive inhibitor toward hydrolysis of the other compound. Dialysis of creatinine iminohydrolase in the presence of 0.01 M Tris phosphate buffer, pH 7.5, containing 1,10-phenanthroline decreased activity by 98%. Reactivation was accomplished by incubating the apoenzyme in the presence of certain divalent metal chlorides, listed in decreasing order of effectiveness: iron(II), zinc, cobalt(II), cadmium, and nickel. The extent of reactivation depended on the substrate and on which metal ion was added to the apoenzyme. Creatinine to cytosine activity ratios varied from 1:3.75 (iron(II) chloride), to 1:0.9 (zinc chloride), to 1:0.06 (nickel chloride). For different preparations of the holoenzyme that ratio ranged from 1:0.45 to 1:1.10. Variable but significant quantities of zinc and iron were present in all preparations of the purified enzyme.

There are two major pathways for the microbial degradation of creatinine. In one, creatinine is first hydrolyzed to creatine in a reversible reaction catalyzed by creatinine amidohydrolase (1). Proteins catalyzing that reaction have been isolated from several microbial sources, and their properties have been described (2-6). Further metabolism via that pathway ultimately leads to production of urea and ammonia through sarcosine and glycine intermediates (7). In a second pathway, creatinine is hydrolyzed to N-methylhydantoin and ammonia in a reaction catalyzed by creatinine iminohydrolase (EC 3.5.4.21) (8). Although a creatinine iminohydrolase was first isolated by Szulmajster in 1958 (9) from the anaerobic bacterium Clostridium paraputrificum, these enzymes, as a class, have been little studied. This may be due, at least in part, to the difficulty of isolating organisms that metabolize creatinine by this pathway. However, one other report has described the purification of creatinine iminohydrolase from Corynebacterium lilium (10). The physical and kinetic properties of that enzyme including possible cofactor requirements have not * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
To whom correspondence should be addressed.
been described in detail. In this second pathway N-methylhydantoin accumulated in the culture medium of c. paraputrificum, indicating that creatinine was serving primarily as a nitrogen source. In 1981, Goodhue and Masurekar (11) described isolation of a soil bacterium, Flavobacterium filamentosum, which produced high levels of creatinine iminohydrolase when cultured in a medium containing creatinine as a nitrogen source. Similar to the anaerobe discussed above, N-methylhydantoin was not further metabolized. Also, when the bacterium was cultured in the presence of other nitrogen sources, the creatinine iminohydrolase activity was negligible, although excellent growth was observed. We describe here a procedure for purification of the F. filamentosum creatinine iminohydrolase, studies on the physical properties of the enzyme, and kinetic studies that show that this metal-containing enzyme displays dual specificity using either creatinine or cytosine as a substrate.

RESULTS
Purification and Molecular Properties of Creatinine Iminohydrolase-Details of the purification of creatinine iminohydrolase, proof of purity, and molecular properties are presented in the Miniprint Section. Briefly, the specific activity of the purified enzyme was 39.8 units/mg. Also, the average molecular weight was 272,480, and the subunit molecular weight was 44,300, consistent with the native enzyme being composed of six subunits.
Substrate Specificity of Creatinine Zminohydrohe-Several amine-containing compounds were evaluated as potential substrates for creatinine iminohydrolase (Table 111). In each case, the substrate concentration given was the highest value tested and represents a saturating concentration for all active compounds. Creatine, arginine, and guanidine, all of which contain the same imino nitrogen functionality as creatinine, were completely inactive. The free base cytosine was utilized efficiently as a substrate (57% creatinine value), whereas the nucleoside, cytidine, and CMP were inactive. Also, 5-fluorocytosine was utilized as a substrate (18% of creatinine value), but negligible activity was observed with either the 3-methyl or 5-methyl derivative (4.6 and 2.3% of creatinine value).
Portions of this paper (including "Experimental Procedures," Fig.   1, and Table I and 11) are presented in miniprint at the end of this paper. The abbreviation used is: SDS, sodium dodecyl sulfate. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document NO. 84M-2730, cite the authors, and include a check or money order for $4.40 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press. Substrate specificity of creatinine iminohydrolase Various compounds at indicated concentrations were substituted for creatinine in the standard reaction mixture. Reaction was initiated by addition of 0.25 pg of creatinine iminohydrolase, and the rate of ammonia production was measured in the standard reaction mixture as described under "Experimental Procedures." Coincident migration of activity against creatinine and cytosine during disc gel electrophoresis. Purified creatinine iminohydrolase (200 pg) was applied to a 7% buffer gel, electrophoresis was carried out, and the gel was sliced into 34 equal fractions. Protein was eluted into 50 p l of 0.01 M Tris phosphate buffer, pH 7.5, and activity was measured with creatinine (0-0) or cytosine (0--0) as substrate in the standard reaction mixture as described under "Experimental Procedures." Although this dual specificity for creatinine and cytosine was observed with an essentially pure enzyme preparation, it was possible that a very minor component in the preparation was responsible for activity with cytosine. Therefore, electrophoresis was carried out in a 7% buffer gel at pH 8.5, the proteins were eluted from the gel, and enzyme activity was measured. The results in Fig. 2 show that both activities displayed identical migration patterns. Also, a constant creatinine to cytosine activity ratio of k0.45 was observed in the three active fractions.

3915
Substrate Saturation Curves- Fig. 3 shows substrate saturation curves for creatinine and cytosine. Normal hyperbolic kinetics were observed with both substrates, and Lineweaver-Burk treatment of the data yielded K , = 5.00 mM for creatinine and 0.62 mM for cytosine. For this preparation, the V,,, for creatinine was 203 units/ml, whereas that for cytosine was 100 units/ml. Thus, the creatinine to cytosine activity ratio was 1:0.50; for other purified preparations this value was 1:045-1:1.10. K , for 5-fluorocytosine (not shown) was 0.42 mM and the V,,, 42 units/ml. The data in Fig. 4 show the results of testing each substrate as an inhibitor toward activity with the other substrate. With creatinine. With cytosine as substrate, activity was monitored at 280 nm, and with creatinine as substrate, activity was monitored by a colorimetric procedure, as described under "Experimental Procedures." cytosine as the substrate, activity was determined by monitoring the decrease in cytosine concentration at 280 nm, and with creatinine as the substrate, activity was measured by the colorimetric procedure based on the Jaffe method (12). Preliminary experiments showed that in each case the inhibitor had no effect on the assay procedure. Creatinine displayed competitive inhibition when cytosine was the substrate, and cytosine displayed competitive inhibition when creatinine was the substrate. These data are consistent with both activities resulting from catalysis at one active site.
Creatinine Iminohydroluse Activity at Various pH Values-Enzyme activity with creatinine or cytosine as substrate was measured at various pH values. The enzyme displayed a broad pH optimum at pH 7.0-8.5 with a constant ratio of activity with the two substrates throughout the optimum (Fig. 5). However, at lower or higher pH, activity with cytosine decreased more rapidly than that for creatinine. The reason for this is not known.

Metal Ion Dependence of Creatinine Iminohydroluse-
When creatinine iminohydrolase was dialyzed in 0.01 M Tris phosphate buffer, pH 7.5, containing 10 mM 1,lO-phenanthroline for 48 h at 5 "C, only 2% of the initial activity remained.
In other experiments, as much as 20% of the original activity remained; the extent of inactivation was dependent on several factors, including the duration of dialysis and the enzyme concentration. As shown in Table IV, the salts of certain divalent metal ions restored activity, when added to the apoenzyme. The indicated concentration of each metal ion was the optimum concentration for reactivation of the enzyme and was the result of evaluating each over a wide range of concentrations. Zinc chloride and cobalt(I1) chloride restored activity with either creatinine or cytosine as substrate to values similar to those observed with the holoenzyme. The

IV Activation of apocreatinine iminohydrolase by salts of divalent metal
ions The apoenzyme was prepared by dialysis against 0.01 M Tris phosphate buffer, pH 7.5, containing 10 mM 1,lO-phenanthroline for 48 h at 5 "C. Each effector was incubated with 0.25 pg of the apoenzyme in 1.0 ml of 0.01 M Tris phosphate buffer, pH 7.5, for 5 min at room temperature. The concentration of iron(I1) chloride and zinc chloride was 0.1 mM; the concentration of the other effectors was 10.0 mM. Then a 10-pl aliquot of that solution was used for measurement of enzyme activity in the standard reaction mixture with various concentrations of either creatinine or cytosine as substrate, as described under "Experimental Procedures." The kinetic constants were determined by Lineweaver-Burk treatment of the data. creatinine to cytosine activity ratio in the presence of zinc chloride was 1:O.g. Iron(I1) chloride greatly enhanced activity with cytosine (485% of activity with holoenzyme), whereas activity with creatinine was again similar to that of the holoenzyme. The creatinine to cytosine activity ratio was 1:3.75. Nickel chloride, on the other hand, only slightly reactivated the enzyme with cytosine (4% of activity with holoenzyme), whereas with creatinine, significant reactivation was observed (39% of activity of holoenzyme). The creatinine to cytosine activity ratio was 1:0.06. Thus, the relative activity toward a particular substrate is a function of the metal ion present and can be varied by a factor of at least 50, as judged by creatinine to cytosine activity ratios. Table IV also shows K,,, values determined for each substrate using creatinine iminohydrolase activated by the various divalent metal ion salts. Minor variations in the K,,, values were observed, but the major effect was on V,,, values.
Metal Ion Binding to Creatinine Iminohydrolase-Nickel chloride and zinc chloride were chosen to determine the effect of the order of addition of metal ions on reactivation of the apoenzyme. When the apoenzyme was preincubated first in the presence of 0.1 mM nickel chloride and then in the presence of 0.1 mM zinc chloride in addition to the nickel chloride, the activity was the same as that obtained in the presence of nickel chloride alone (Table V). Also, the reverse experiment was carried out; when the enzyme was first preincubated in the presence of 0.1 mM zinc chloride, nickel chloride had no effect on the extent of reactivation. These data are consistent with the interpretation that the metal ions compete for the same or similar sites on the enzyme, and once a particular metal ion is bound to the enzyme, it is not easily displaced.
These experiments were carried out with creatinine as substrate, but identical results were obtained with cytosine as substrate and different divalent metal ion salt combinations.

Effect of order of metal ion addition on activation of apocreatinine iminohydrolase
The apoenzyme (15 pglml), prepared as described in Table 111, was incubated in a solution containing 0.1 mM metal ion at room temperature for 5 min. The numbers (1) and (2) refer to order of metal ion additions. The time between (1) and (2) was 5 min. Five min after the last metal ion addition, 10 pl of that solution was used for measurement of enzyme activity in the standard reaction mixture with creatinine as substrate as described under "Experimental Procedures."

TABLE VI Presence of metals in creatinine iminohydrolase
The metal content of creatinine iminohydrolase was determined by atomic absorption analysis as described under "Experimental Procedures." Two other points should be noted with regard to binding of metal ions to the apoenzyme. First, activation of the apoenzyme was rapid. The normal protocol for such experiments was to add creatinine iminohydrolase to 0.01 M Tris phosphate buffer, pH 7.5, containing the appropriate divalent metal ion salt; then at various times an aliquot was removed for measurement of enzyme activity in the standard reaction mixture.
No exogenous metal ion was added to the reaction mixture.
When the apoenzyme was incubated in the presence of 0.1 mM zinc chloride, full activation was observed immediately after addition of the enzyme to that solution. Second, after reactivation of the apoenzyme in the presence of 0.1 mM zinc chloride, dialysis against 0.01 M Tris phosphate buffer, pH 7.5, for 24 h at 5 "C caused no loss of enzyme activity. Hence, sufficient metal ion remained bound to the enzyme to retain activity.
Metal Content of Creatinine Iminohydrolase-Metals present in five purified preparations of creatinine iminohydrolase were determined by atomic absorption analysis. Table VI shows the average values for metals that reactivate the apoenzyme and the range of values. Cobalt and cadmium were not found in any of the five preparations. Zinc was consistently found in the highest molar ratio, followed by iron and then nickel. There was considerable variation among the different preparations, as shown by the range of values in Table VI. Indeed, in two preparations there was no observable nickel present. These data, then, cannot reflect stoichiometry, but they do reflect the fact that significant proportions of catalytically important metals, especially zinc and iron, were present in the purified creatinine iminohydrolase.

DISCUSSION
Creatinine iminohydrolase was purified from extracts of F. filamentosum, a microorganism which, when grown in the presence of creatinine, produces large amounts of that enzyme. The specific activity of the purified preparation (39.8 units/mg) was similar to that reported for two other creatinine iminohydrolase preparations, one from C. lilium (34.0 units/ mg) (10) and one from the anaerobe C. paraputrificum (36.1 units/mg) (9). However, the extent of purity of the latter enzyme is not known. The K , of the F. filamentosum enzyme (5.00 mM) for creatinine was higher than that value for the C. lilium enzyme (1.27 mM), whereas the iminohydrolase from the anaerobe had a significantly higher K,,, of  Extensive substrate specificity studies revealed that the purified F. filamentosum creatinine iminohydrolase also displayed cytosine deaminase activity. The enzyme from C. lilium, on the other hand, did not utilize cytosine as substrate. All kinetic data reported herein are consistent with both activities being catalyzed by the same protein and indeed at the same active site. At first the structural relationship between creatinine and cytosine is not immediately obvious, but as shown by examination of the possible tautomeric forms of cytosine, similarity does exist with three positions on the ring, including the same reactive imino nitrogen functionality in both molecules.
cytosine creatinine SCHEME I The 5-position on the cytosine ring has a marked effect on activity. Substitution of the hydrogen at that position by fluorine (5-fluorocytosine) decreases the activity to 33% of the cytosine value with no apparent affect on the K,, and substitution with a methyl group in that position virtually eliminates activity. Indeed, 5-methylcytosine inhibits reaction with either cytosine or creatinine substrate. At the same time, the analogous position on the creatinine molecule does contain a methyl group (N-methyl).
Cytosine deaminases have been isolated from several microbial sources and purified preparations obtained from Serratia marcescens (25), Pseudomonas aureofaciens (26), and bakers' yeast (27). The bacterial enzymes are very highmolecular-weight multisubunit proteins of 580,000 and 630,000 respectively, whereas the yeast enzyme has a molecular weight of 34,400. Kinetic data were available only for the yeast enzyme, which showed a K,,, of 5.0 mM for cytosine.
This should be compared with the value of 0.62 mM obtained for the F. filamentosum enzyme. In terms of substrate specificity, none of these cytosine deaminases were tested with creatinine. Based upon the results reported herein, it is possible that one or more of these proteins might utilize that compound as a substrate.
The F. filamentosum creatinine iminohydrolase is a metalion-activated and metal-containing enzyme. The apoenzyme was prepared by dialysis against 1,lO-phenanthroline, and rapid reactivation was obtained by addition of specific divalent metal ion salts to the apoenzyme. The relative activity with creatinine or cytosine was dependent on the metal ion present, such that creatinine to cytosine activity ratios could be varied from 1:3.75 (iron(I1) chloride) to 1:O.g (zinc chloride) to 1:0.06 (nickel(I1) chloride). We are aware of no reports on metal ion involvement in catalysis by known creatinine iminohydrolases or cytosine deaminases. However, the specificity of several well-characterized metalloenzymes is also dependent on the metal used to reactivate the apoenzyme. For example, the protease to esterase ratio of carboxypeptidase A is a function of the metal ion present (281, and the same is true for the phosphate ester to o-phosphorothioate ester activity ratio of alkaline phosphatase (29). Both of these enzymes are zinc metalloenzymes.
Studies on the metal content of purified preparations of creatinine iminohydrolase show that zinc and iron were present in all preparations. However, several lines of evidence suggest that the metal content in the purified enzyme does not reflect the stoichiometry in the native enzyme. First, the amounts of each metal varied considerably from preparation to preparation. Variation in metal content as a function of the purification procedure has been observed with other metalloenzymes (30-32). Second, of the five preparations, two were metal-ion deficient; addition of zinc chloride to the preparation increased the activity by 45%. The third point relies on the fact that the ratio of activities with creatinine or cytosine as substrate is a function of the metal ion present. When activity in the crude extract was measured with both substrates, the creatinine to cytosine ratio was 1:3. As noted above, in the purified preparations that ratio was 1:0.45-1:1.10. Thus, this purified enzyme does not reflect what is observed in the crude extract. When the apoenzyme was reactivated by various metal ions, a creatinine to cytosine ratio of 1:3.75 was observed with the iron-activated enzyme ( Table IV). Determination of the metal content of the native enzyme requires modification of the purification procedure to lessen the possibility of dissociation of the metal ion(s) from the enzyme during the purification, and the ratio of the activities with the two substrates can serve as a guide for those studies. For inhibJtlon experiments. creatinine Lminohydrolass deaminase was determined by the standard method above by substitutina 5 "mol of cvtoaine for creatinine as aubstrate. Purification of Creatinine Iminohvdrolase-- Table I creatinine imlnohydrolase. The propanol fractionation served two purposes. First, the solvent stabilized the enzyme; in crude extracts 50% activity losses were observed in 12 h at 5 'C, whereas in a 50% propanol solution no loss was observed. Second, that step efficiently concentrated the enzyme in good yield and with an 18-fold increase in specific activity.

Measurement of Cytosine Deaminase Activity--Cytosine
of a Solvent solution, we speculated that hydrophobic interaction chromatography might allow significant purification. After we evaluated several Sepharoae 4B acylamine materials as chromatography media, we chose the hexylamine derivative, based on yield and increase in specific activity. Generally, yields directly from that column were 70-90%. The 45% yield shown in Table I   in general the final product was at least 95% pure, as judged of the purification procedure was identified 8s creatinine after scanning the stained gels. The protein band in Step IV eluting the protein, and measuring enzyme activzty.
iminohydrolase by slicing an unstained portion of the gel, occurred.
addition to 7% gels, no dissociation of the protein band When electrophoresis was carried out In 5% or 9% gels in molecular wexght of creatinine imlnohydrolase was determined by several techniqlles (Table 11). Values ranged from 245.000, 288,400, as estimated from gel filtration data.
a8 estimated after electrophoresis in gradient gels, to The subunit molecular weight was estimated after electrophoresis in 7% sodium dodecyl sulfate gels. A value of 44,300 was obtained by comparing the mobzllty of Creatinlne was vlslhle when electrophoresis was carried Out under these iminohydrolase with that of standards. Since only one band conditions, creatinine iminohydrolase dissociated into subunits of similar molecular welghts upon treatment with sodium dodecyl sulfate.
If we assume that the partial specific volume of creatinine iminohydrolase is similar to that of standard globular proteins, the average molecular weight is 272,480. The subunit molecular weight data, then, show that the enzyme is Composed of six subunits of equal molecular welght.