Isolation and Characterization of Tryptophan Transaminase and Indolepyruvate C-Methyltransferase ENZYMES INVOLVED IN INDOLMYCIN BIOSYNTHESIS IN STREPTOMYCES GRISEUS*

Two enzymes, tryptophan transaminase and indolepyruvate C-methyltransferase, which are active in the initial steps of the biosynthetic pathway of the antibiotic indolmycin, have been detected and partially purified from cell-free extracts of Streptomyces griseus. The transaminase has been purified S-fold by ammonium sulfate fractionation. At this stage of purification, it catalyzes the a-ketoglutarate and pyridoxal phosphate-dependent transamination of L-tryptophan, 3-methyltryptophan, L.-phenylalanine, and L-tyrosine. The C-methyltransferase catalyzes the transfer of a methyl group from S-adenosylmethionine to position 3 of the aliphatic side chain of indolepyruvate. No cofactors are required. The C-methyltransferase has been purified llO-fold by ammonium sulfate fractionation, Sephadex G-150 gel filtration, DEAE-Sephadex column chromatography, and Bio-Gel A-5m gel filtration. The enzyme has a broad PH optimum of 7.5 to 8.5. A molecular weight of 55,000 f 5,000 has been determined by Sephadex G-200 gel filtration with reference proteins and a molecular weight of 58,500 f 8,000 has been determined by sucrose density gradient centrifugation. The enzyme is relatively stable at temperatures of O-5” but is destroyed by freezing or by heating. The C-methyltransferase is inhibited strongly by the thiol reagents p-chloromercuribenzoate and N-ethylmaleimide. The Znz+ and Fez+ chelators l,lO-phenanthroline and 2,2’-bipyridine also inhibit the enzyme activity but EDTA does not. Michaelis-Menten constants have been determined for the IlO-fold purified enzyme as 1.2 x 10m5 M for S-adenosylmethionine and 4.8 x 1Om6 M for indolepyruvate. The enzyme activity in the crude extract is inhibited competitively by indolmycin (Ki = 2.3 mM) and L-tryptophan (Ki = 0.17 mM), but these effects are not observed after the enzyme has been passed through the Sephadex G-150 column during purification. The crude extract is capable of methylating phenylpyruvate and p-hydroxyphenylpyruvate but this capability is lost upon purification of the indolepyruvate C-methyltransferase activity. No methylation of L-tryptophan occurs under the conditions used.

Two enzymes, tryptophan transaminase and indolepyruvate C-methyltransferase, which are active in the initial steps of the biosynthetic pathway of the antibiotic indolmycin, have been detected and partially purified from cell-free extracts of Streptomyces griseus. The transaminase has been purified S-fold by ammonium sulfate fractionation. At this stage of purification, it catalyzes the a-ketoglutarate and pyridoxal phosphate-dependent transamination of L-tryptophan, 3-methyltryptophan, L.-phenylalanine, and L-tyrosine.
The C-methyltransferase catalyzes the transfer of a methyl group from S-adenosylmethionine to position 3 of the aliphatic side chain of indolepyruvate.
No cofactors are required. The C-methyltransferase has been purified llO-fold by ammonium sulfate fractionation, Sephadex G-150 gel filtration, DEAE-Sephadex column chromatography, and Bio-Gel A-5m gel filtration. The enzyme has a broad PH optimum of 7.5 to 8.5. A molecular weight of 55,000 f 5,000 has been determined by Sephadex G-200 gel filtration with reference proteins and a molecular weight of 58,500 f 8,000 has been determined by sucrose density gradient centrifugation.
The enzyme is relatively stable at temperatures of O-5" but is destroyed by freezing or by heating. The C-methyltransferase is inhibited strongly by the thiol reagents p-chloromercuribenzoate and N-ethylmaleimide. The Znz+ and Fez+ chelators l,lO-phenanthroline and 2,2'-bipyridine also inhibit the enzyme activity but EDTA does not. Michaelis-Menten constants have been determined for the IlO-fold purified enzyme as 1.2 x 10m5 M for S-adenosylmethionine and 4.8 x 1Om6 M for indolepyruvate.
The enzyme activity in the crude extract is inhibited competitively by indolmycin (Ki = 2.3 mM) and L-tryptophan (Ki = 0.17 mM), but these effects are not observed after the enzyme has been passed through the Sephadex G-150 column during purification.
The crude extract is capable of methylating phenylpyruvate and p-hydroxyphenylpyruvate but this capability is lost upon purification of the indolepyruvate C-methyltransferase activity. No methylation of L-tryptophan occurs under the conditions used.
The antibiotic indolmycin ( Fig. 1) is produced by Streptomyces griseus (ATCC 12648). It exhibits antimicrobial activity against gram-positive and gram-negative bacteria, including polyresistant staphylococci (l-3). The culture characteristics of the indolmycin-producing strain of S. griseus have been described (2), and a fermentation procedure for the production of indolmycin together with two other antibiotics of unknown structure has been patented (4). The structure (5), relative (3, 5) and absolute (6, 7) stereochemistry of indolmycin have been determined and syntheses of indolmycin and some of its degradation products and analogs have been reported (3,5). Previous studies have established that indolmycin is derived biosynthetically from tryptophan, the methyl group of methionine, and the amidino group of arginine (7) Specificity-Specificity studies carried out with tryptophan transaminase purified 3-fold by ammonium sulfate precipitation indicate that the enzyme catalyzes the pyridoxal phosphate-dependent transamination of L-phenylalanine, L-tyrosine, and 3-methyltryptophan, as well as L-tryptophan, but not of n-tryptophan.
Neither pyruvate nor oxalacetate were able to substitute efficiently for a-ketoglutarate as amino group acceptor in the reaction with L-tryptophan. These results are summarized in Table II. 3-Methyltryptophan, besides functioning as a substrate, was also observed to inhibit the tryptophan transaminase activity.
the solution was stirred for a further 40 min. Following 20 min of This could be measured using the spectral assay since neither 7821 centrifugation at 30,000 x g, ammonium sulfate was added to the supernatant in a similar manner to achieve 60% saturation, keeping the pH at 7.0. The protein that precipitated between 45 and 60% saturation contained 77% of the initial units of transaminase activity and 26% of the initial protein, and represented a 2.9-fold purification of the transaminase activity.
The protein was dissolved in 5 to 10 ml of 0.01 M phosphate buffer, pH 7.0, and dialyzed against three changes of 1 liter of the same buffer for 1 hour each. The transaminase was not purified further. C-Methyltransferase Step 1. Ammonium Sulfate Treatment-The initial step in the purification of the C-methyltransferase was an ammonilum sulfate precipitation performed in a manner similar to that described above for the transaminase, except that 0.1 mM Cleland's reagent was added to the crude enzyme extract and the protein precipitating between 35 and 55% saturation was collected. This precipitate was dissolved in 5 to 10 ml of 0.01 M phosphate buffer, pH 7.0, and dialyzed for 3 hours against two changes of 1 liter of the same buffer.
Step 2. Sephaden Chromatography-The dialyzed solution containing the C-methyltransferase was applied to a column (2.5 x 33 cm) of Sephadex G-150 which had been equilibrated with 0.01 M phosphate buffer, pH 7.0, and was eluted with the same buffer. Fractions of 5 ml each were collected.
The C-methyltransferase was recovered in Fractions 15 through 22. The total volume of the active fractions was reduced to approximately 4 ml using an Amicon pressure dialysis apparatus.
Step 3 Step 4. Bio-Gel A-5m Chromatography-The enzyme preparation from the preceding step was applied to a column (1.0 x 25 cm) of Bio-Gel A-5m which had been equilibrated with 0.01 M phosphate buffer, pH 7.0, and subsequently was eluted with the same buffer. One-milliliter fractions were collected. The C-methyltransferase activity was recovered in Fractions 9 through 12, coincident with a well defined protein peak. By the above procedure, the C-methyltransferase may be purified approximately llO-fold with an over-all yield of 40 to 45%. A typical purification is summarized in Table I  Addition of both isomers of 3-methyltryptophan resulted in 64% inhibition of r.-tryptophan transamination activity. These results are compatible with those obtained with the radioactive assay and suggest that L-tryptophan and 3-methyltryptophan are transaminated by the same enzyme. Apparently the enzyme is stereospecific only for position 2 of 3-methyltryptophan.
Identification of the unstable product of the reaction, indolepyruvate, was achieved upon reduction with sodium borohydride to indolelactic acid and thin layer chromatographic identification as the methyl ester in ligroin/l-octanol/ acetone (8/2/l). When radioactive tryptophan was used in the reaction and the final thin layer chromatography plate scanned in a radiochromatogram scanner all the radioactivity was contained in one peak which corresponded in R, value to a van Urk-positive spot of reference indolelactic acid methyl ester. Stability-The transaminase activity is unstable at O-10" and no activity is detectable after 12 hours at this temperature. The enzyme is best stored at -20" by freezing the crude cell-free extract with 10% glycerol or by freezing the 45 to 60% ammonium sulfate precipitate in a phosphate buffer solution. Inhibition Studies-Indolmycin added to the transaminase reaction mixture in concentrations of 5 x lo-' M and 1 x 10e3 M caused no inhibition of enzyme activity. The transaminase activity was unaffected by the addition of 10e6 M p-chloromercuribenzoate and, therefore, probably does not contain thiol groups essential to activity.

C-Methyltransferase
The C-methyltransferase enzyme catalyzes the transfer of the methyl group from S-adenosylmethionine to position 3 of indolepyruvate.
No cofactors have been found to be required for the reaction.
The formal name for the enzyme is Sadenosylmethionine:indolepyruvate 3-methyltransferase. The identification of the reaction product as 3-methylindolepyruvate is based on the following evidence. 1. The enzyme reaction product agrees in its chromatographic mobility with authentic 3-methylindolepyruvate prepared by the reaction of 3-methyltryptophan and L-amino acid oxidase in the presence of excess catalase.
2. Reduction of the reaction product obtained with radioactive S-adenosylmethionine with sodium borohydride yielded a compound that co-chromatographs and co-crystallizes with authentic indolmycenic acid. Kuhn-Roth oxidation yielded L-Tryptophan + oxalacetate 0 L-Tryptophan + pyruvate 0.3 "All substrates were present in 1 mM concentration. bThese values would be doubled if one considers that the enzyme probably can only react with the 2 S form.
acetic acid with a specific radioactivity identical with that of the indolmycenic acid. 3. Mild oxidation with dilute hydrogen peroxide yielded indoleisopropionic acid which could be co-crystallized with authentic carrier (R)-indoleisopropionic acid as the cinchonine salt.
The methods and results of the identifications are discussed in detail in a previous publication (24). The identification procedures were carried out initially using crude enzyme and Procedures 1 and 2 as indicated above were repeated using purified C-methyltransferase.
Purification steps carried out as described and summarized in Table I resulted in a llO-fold purification of the enzyme activity. Characterization of the properties of the enzyme as described in the following section was performed with enzyme at that stage of purification unless otherwise specified.
Specificity-When phenylpyruvate or p-hydroxyphenylpyruvate was substituted for indolepyruvate in the reaction mixture with crude enzyme preparation, the reaction proceeded at rates of 49 and 35%, respectively, of the reaction rate with indolepyruvate.
No reaction occurred when L-tryptophan was substituted.
However, ammonium sulfate precipitation of the indolepyruvate C-methyltransferase activity resulted in a 45% loss of initial p-hydroxyphenylpyruvate and phenylpyruvate C-methyltransferase activity and after purification on Sephadex G-150, only 14% of the initial activity remained. Curiously, no phenylpyruvate C-methyltransferase activity could be detected in any other ammonium sulfate fraction. This would indicate that either the enzyme is nonspecific in its natural form and undergoes a conformational or other change during purification, losing activity toward phenylpyruvate, or alternatively, that two enzymes are responsible for the dual activity and the phenylpyruvate-specific enzyme is denatured during the purification procedure. pH Dependence-The purified enzyme shows optimal activity between pH 7.5 and 8.5. There appears to be some dependency upon the buffer used since in phosphate buffer the enzyme is somewhat more active at pH 7.5 than at pH 8.0, while in other buffers, the optimum pH is slightly higher.
The enzyme is inactivated irreversibly at pH values of 5.5 and below. The longer it remains at pH 5.5 the less able it is to recover activity upon neutralization to pH 7.0. For example, neutralization of an enzyme solution that has been kept at pH 5.5 for 1 hour results in a 46% loss of activity compared to a control that has been kept at pH 7.0 continuously.
Storage for 2 days at pH 5.5 results in only 6% restoration of activity upon neutralization to pH 7.0.

Molecular
Weight-Two methods were used to determine the molecular weight of the enzyme. Using gel filtration on a calibrated Sephadex G-200 column (1.8 x 30 cm), C-methyltransferase activity eluted at a volume corresponding to a molecular weight of 55,000 i 5,000, assuming that the enzyme is a globular protein.
The molecular weight was also determined by ultracentrifugation in a 5 to 20% sucrose density gradient calibrated with yeast alcohol dehydrogenase (M, 150,000, s;'oTz = 7.4) and lysozyme (M, 17,200, s&$ = 2.1) according to the procedure described by Martin and Ames (25). Using the formula S&S', = (Mr,lMr,)2'3 a molecular weight of 58,500 f 8,000 was estimated for the C-methyltransferase.
Stability-Both the crude enzyme preparation and the more purified preparations were found to be stable at 2" for periods up to 1 month. Freezing the buffered enzyme preparations caused a partial loss of activity. The enzyme activity was not affected when the ammonium sulfate-precipitated pellet was frozen for a short period of time but after 1 year of storage the reconstituted protein had only 20% of its original activity. Heat was found to destroy enzyme activity. Heating the crude enzyme preparation to 55" for 5 min resulted in an 85% loss of activity.

Kinetic
Experiments-The enzyme was found to follow Michaelis-Menten kinetics. The double reciprocal plots gave K, values of 1.2 x 10m5 M for S-adenosylmethionine and 4.8 x 1Om6 M for indolepyruvate.
The K, value for indolepyruvate was also determined using a crude enzyme preparation and was found to be very close (4.0 x 1Om6 M) to the value determined with the purified enzyme.

Inhibition
Studies-Metal chelators were added to the reaction mixture prior to the addition of substrates. The results are summarized in Table III. As indicated  in this table,  o-phenanthroline and 2,2'-bipyridine in concentrations from 0.5 to 0.2 InM partially inhibited the enzyme activity. o-Phenanthroline and 2,2'-bipyridine are known to chelate Zn2+ and Fez+ and have been shown to inhibit enzymes requiring these metals (26). Diethyldithiocarbamate, which also chelates Zn2+ and Fez+ but more specifically Cu'+, does not inhibit C-methyltransferase activity at levels up to 2.0 mM and actually appears to slightly stimulate activity. The reason for the apparent inhibition by o-phenanthroline and 2,2'-bipyridine, but not by diethyldithiocarbamate, is not known. However, similar selectivity of inhibition has been observed in other enzyme systems such as a zinc-containing DNA-dependent  It was thought that perhaps o-phenanthroline and 2,2'-bipyridine were causing inhibition because of their aromatic character rather than their metal-chelating abilities and therefore were competing with indolepyruvate.
To test for competitive inhibition, varying concentrations of indolepyruvate (0. 15, 0.20, and 0.40 mM) and o-phenanthroline (0, 1.0, 1.5, and 2.0 mM) were added to the reaction mixture. Double reciprocal Dixon (28) plots of the results indicate that the inhibition is not competitive with indolepyruvate although it does not appear to be a case of pure noncompetitive inhibition either. EDTA added to the reaction mixture in concentrations as high as 5.0 mM did not inhibit the C-methyltransferase activity. Therefore divalent cations such as Mg2+, Mn*+, and Ca2+ are not essential for the activity of the enzyme.
The thiol group reagent p-chloromercuribenzoate strongly inhibits enzyme activity at 10 pM concentration (see Table IV). This inhibition is reversed by the addition of 1.0 mM cysteine. N-Ethylmaleimide, present at a concentration of 4 mM, causes 52% inhibition of C-methyltransferase activity, thus further supporting the conclusion that a -SH group is important to the enzyme action. However, the thiol group alkylating agents iodoacetate and iodoacetamide inhibit enzyme activity only slightly as indicated in Table IV. According to Dixon and Webb (29) these reagents are not as reactive nor as specific as p-chloromercuribenzoate.
Their failure to inhibit the Cmethyltransferase therefore does not rule out the involvement of a -SH group in the enzyme's action.
As seen in Table V, the indolepyruvate C-methyltransferase activity in the crude enzyme preparation is inhibited by several compounds including L-tryptophan, L-tryptophanylglycine, Ltryptophan amide, and indolmycin, and is activated by dimethylsulfoxide and possibly by N-acetyl-L-tryptophan. Dixon (28) plots of the inhibition at various concentrations of indolepyruvate indicate that indolmycin and L-tryptophan are competitive inhibitors for indolepyruvate with K, values of 2.3 and 0.17 mM, respectively.
It was found, however, that the C-methyltransferase is no longer inhibited by any of the above-mentioned effecters after purification through the Sephadex G-150 column step. The loss of inhibition does not appear to be due to aging since aging the enzyme in the crude extract by storage at 2" for 2 weeks does not result in loss of inhibition. As mentioned previously the K, value for indolepyruvate does not change as the enzyme is purified. Therefore, it is not likely that the loss of inhibition by the competitive inhibitors, L-tryptophan and indolmycin is due to a conformational change at the active site during purification.
A third possible  for the loss of inhibition is that a regulatory subunit dissociates during the gel filtration step. To examine this possibility, crude enzyme was placed on the Sephadex G-200 column that was used previously for molecular weight determination.
The indolepyruvate C-methyltransferase peak was eluted in tubes 50 through 62 (1 ml per tube) and was not inhibited, thus confirming that the Sephadex step is responsible for the loss of inhibition by tryptophan or indolmycin. These tubes were combined and designated Fraction 2. The tubes containing protein of a higher molecular weight (tubes 33 to 49) and the tubes containing lower molecular weight protein (tubes 63 to 85) were reduced in volume by pressure dialysis and designated Fractions 1 and 3, respectively. Fraction 2, containing the C-methyltransferase activity, was then assayed for inhibition by adding 1 and 5 mM L-tryptophan in the presence of Fraction 1 and Frrction 3. Neither fraction restored inhibition to Fraction 2. Therefore, if a regulatory subunit is present, it does not reassociate easily with the C-methyltransferase found in Fraction 2. Experiments are under way to further examine the reason for the loss of inhibition upon purification of the enzyme.

DISCUSSION
The presence of the tryptophan transaminase and the indolepyruvate C-methyltransferase in cell-free extracts of Streptomyces griseus supports the hypothetical pathway for indolmycin biosynthesis proposed by Hornemann et al. (7) in which tryptophan is transaminated first to indolepyruvate which is subsequently methylated at position 3 of the aliphatic side chain to form 3-methylindolepyruvate.
An alternative pathway could be postulated which would involve C-methyla-tion at the stage of tryptophan rather than indolepyruvate. Previous feeding experiments had shown that 3-methyltryptophan was incorporated efficiently into indolmycin, suggesting that this compound could be an intermediate in the pathway (7). However, in unpublished experiments 3-methyltryptophan formation could not be detected unequivocally in the culture by trapping experiments.
This fact combined with the finding that tryptophan is not methylated by the crude cell-free extract indicates that 3-methyltryptophan is not a natural intermediate in the biosynthesis. The finding that the transaminase is able to convert 3-methyltryptophan to 3-methylindolepyruvate explains the observed incorporation of exogenous 3-methyltryptophan, although this route is apparently not followed naturally.
Aromatic amino acid transaminases are fairly widespread in nature (30-32) and have been reported for a number of microorganisms including Escherichia coli (33,34), Rhizobium leguminosarium (35), Agrobacterium tumefaciens (36), and Clostridium sporogenes (32). Like the transaminase we have isolated, all of these utilize cu-ketoglutarate as the amino group acceptor. The transaminases from two of the organisms have been purified (32, 34) and single enzymes were found to catalyze the transamination of all three aromatic amino acids, although in C. sporogenes an additional phenylalanine transaminase was also found. Further purification of the tryptophan transaminase we have isolated from S. griseus would be necessary to determine whether a single enzyme is responsible in this organism for the observed activities.
Methyltransferases are fairly widespread in nature but most of those that have been purified and characterized transfer the methyl group of S-adenosylmethionine to oxygen, nitrogen, or sulfur atoms (37). The C-methyltransferase we have isolated resembles most other methyltransferases that utilize Sadenosylmethionine in that no cofactors are required for the reaction to occur. One exception is the AZ*-sterol methyltransferase reported by Moore and Gaylor (12) and by Bailey et al. (38) which requires glutathione for maximum activity. Most other S-adenosylmethionine methyltransferases so far reported also have no requirement for divalent metal ions. A few (11, 39) however, do require Mg '+ for maximum activity and at least one other methyltransferase, a homocysteine S-methyltransferase found in, Aerobacter aerogenes and E. coli, has been reported to be a metalloenzyme, most likely containing Zn2+ (40).
The observed inhibition of the indolepyruvate C-methyltransferase by p-chloromercuribenzoate and N-ethylmaleimide suggests that a sulfhydryl group is necessary for activity. Sulfhydryl groups are essential for the activity of a number of other methyltransferases (41, 42) but in no case is it known whether their role is in maintenance of enzyme conformation or in catalysis.
The K, value established for S-adenosylmethionine (1.3 x 10m5 M) is comparable to the values observed for many 0-, N-, and S-methyltransferases (43, 44) and for the DNA-C-methyltransferase from E. coli (45). The high affinity of the enzyme for indolepyruvate (K, = 4.8 x 1Om6 M) may reflect the need to prevent build-up of high intracellular concentrations of this unstable substrate.
The significance of the observed inhibition of C-methyltransferase activity by indolmycin and L-tryptophan in the crude enzyme extract is unclear, especially since this property is lost upon purification of the enzyme. The inhibition by indolmycin might be attributed to feedback inhibition by the