Conversion of 5-S-Methyl-&thio-D-ribose to Methionine in Klebsiella pneumoniae STABLE ISOTOPE INCORPORATION STUDIES OF THE TERMINAL ENZYMATIC REACTIONS IN THE PATHWAY*

Extracts of Klebsiella pneumoniae convert 5-S-methyl-5-thio-D-ribose (methylthioribose) to methionine and formate. To probe the terminal steps of this biotransformation, [1-13C]methylthioribose has been synthesized and its metabolism examined. When supplemented with Mg2+, ATP, L-glutamine, and dioxygen, cell-free extracts of K. pneumoniae converted 50% of the [1-13C]methylthioribose to [13C]formate. The formation of [13C]formate was established by 13C and 1H NMR spectroscopy studies of the purified formate, and by 13C and 1H NMR spectroscopy and mass spectrometry studies of its p-phenylphenacyl derivative. By contrast, no incorporation of label from [1-13C]methylthioribose into the biosynthesized methionine was detected by either mass spectrometry or 13C and 1H NMR spectroscopy. The most reasonable interpretation of these results is that C-1 of methylthioribose is converted directly to formate concomitant with the conversion of carbon atoms 2-5 to methionine. The penultimate step in the conversion of methylthioribose to methionine and formate is an oxidative carbon-carbon bond cleavage reaction in which an equivalent of dioxygen is consumed. To investigate the fate of the dioxygen utilized in this reaction, the metabolism of [1-13C]methylthioribose in the presence of 18O2 was also examined. Mass spectrometry revealed the biosynthesis of substantial amounts of both [18O1]methionine and [13C, 18O1]formate under these conditions. These results suggest that the oxidative transformation in the conversion of methylthioribose to methionine and formate may be catalyzed by a novel intramolecular dioxygenase. A mechanism for this dioxygenase is proposed.

To The most reasonable interpretation of these results is that C-l of methylthioribose is converted directly to formate concomitant with the conversion of carbon atoms 2-5 to methionine. The penultimate step in the conversion of methylthioribose to methionine and formate is an oxidative carboncarbon bond cleavage reaction in which an equivalent of dioxygen is consumed.
To investigate the fate of the dioxygen utilized in this reaction, the metabolism of [l-'3C]methylthioribose in the presence of "OZ was also examined.
Mass spectrometry revealed the biosynthesis of substantial amounts of both ["Ol]methionine and [13C,'801]formate under these conditions. These results suggest that the oxidative transformation in the conversion of methylthioribose to methionine and formate may be catalyzed by a novel intramolecular dioxygenase.
A mechanism for this dioxygenase is proposed.
Phosphorylation of the a-oriented C-l hydroxyl group of methylthioribose by a specific kinase produces glycosyl phosphate 3 (14,17,18).
This compound is subsequently isomerized to ketose 4, which is then dehydrated to diketone 5 (19). In a partially purified system (19), diketone 5 is oxidatively cleaved to a-keto acid 6 and formate. Dioxygen is consumed during this process. In addition, variable amounts of 3-S-methyl-3-thiopropionate (not shown) are also produced from diketone 5. The observed stoichiometry is consistent with: 1) 1 mol of O2 consumed per mol of either a-keto acid 6 or 3-S-methyl-3-thiopropionate produced and 2) 1 equivalent of formate generated per mol of cu-keto acid 6 and 2 equivalents of formate generated per mol of 3-S-methyl-3thiopropionate synthesized. Preliminary evidence suggests the possibility of a nonphosphorylated intermediate between diketone 5 and cr-keto acid 6 and/or 3-S-methyl-3-thiopropionate (19). More recent studies suggest that the synthesis of cY-keto acid 6 and 3-S-methyl-3-thiopropionate from diketone 5 is effected by different enzymes.* In the last step of the salvage pathway, cY-keto acid 6 is transaminated to methionine (compound 7).
To understand the metabolic conversion of methylthioadenosine to methionine in K. pneumoniae, further characterization of the reaction(s) leading from diketone 5 to cY-keto acid 6 and formate was required. Unfortunately, studies of this transformation have been hampered by: 1) the difficulty of obtaining sufficient quantities of diketone 5, which is currently available only via a lengthy sequence of chemical and enzymatic steps; 2) the lability of certain enzymes in the pathway (including an enzyme required for the preparation of 5); and 3) the instability of diketone 5 (19). By contrast, methylthioribose is a readily accessible and relatively stable compound; furthermore, the mechanism of its conversion to diketone 5 is well understood.
Thus, in an attempt to gain further insight into the mechanism by which diketone 5 is converted into ol-keto acid 6 and formate, we decided to study this transformation in crude cell-free extracts of K. pneumoniae. These studies employed methylthioribose as the precursor and stable isotope labeling to follow the conversion.
formate and o-keto acid 6, C-l of diketone 5 (C-l of methylthioribose) is converted to formate, whereas carbon atoms 2-5 of diketone 5 yield carbon atoms l-4 of a-keto acid 6 (19). However, Volk and Bather (20) have recently provided evidence that the enzymatic conversion of ribulose 5-phosphate to 3,4-dihydroxybutanone-4-phosphate and formate is accompanied by a carbon chain rearrangement in which C-4 of the precursor is lost as formate and C-5 of the precursor becomes covalently bonded to C-3. Because the reactions between diketone 5 and o-keto acid 6 and formate (Scheme 1) were uncertain, we wished to provide direct evidence for the fate of the carbon atoms of methylthioribose during its enzymatic conversion to methionine and formate. Accordingly, [1-'"Clmethylthioribose was synthesized and its metabolism by cell-free extracts of K. pneumoniae studied. As indicated in Scheme 1, conversion of diketone 5 to 01keto acid 6 and formate is accompanied by the consumption of 1 mol of dioxygen/mol of diketone cleaved (19). The mechanism of this unusual oxidative carbon-carbon bond cleavage was unknown. We reasoned that valuable information on the mechanism of this reaction would be obtained if the fate of the dioxygen consumed during this transformation could be established.
To this end, the metabolism of [1-13C]methylthioribose in the presence of "02 was examined. Depending on the mechanism of the 4-electron oxidation of diketone 5 to a-keto acid 6 and formate, one would expect to observe: 1) no incorporation of "0 into either methionine or formate ("02 conversion to H2180J; 2) incorporation of one atom of la0 into either methionine or formate, the other atom of "0 being converted into H2180 (monooxygenase activity); or 3) incorporation of one atom of la0 into both methionine and formate (dioxygenase activity). exchange chromatography and reverse phase HPLC. Total formate was isolated by repetitive bulb-to-bulb distillation. A parallel incubation, lacking methylthioribose, was carried out in an identical fashion. The amounts of methionine and formate isolated from both incubations were determined, using the ninhydrin and formate dehydrogenase assays, respectively (Table I). Both products generated from [l-'"Cl methylthioribose were then examined by 'H and 13C NMR spectroscopy. In addition, an aliquot of the purified formate was derivatized to p-phenylphenacyl formate which was purified by normal phase silica HPLC. The p-phenylphenacyl formate was also examined by 'H and 13C NMR spectroscopy and both it and methionine were analyzed by direct inlet, chemical ionization mass spectrometry. NMR and MS analysis demonstrated that 13C enrichment was present in formate, but not in methionine, biosynthesized from [1-'3C]methylthioribose.
That ['3C]formate was indeed a product of [l-13C]methylthioribose metabolism was established by NMR in the following manner.
The 'H NMR spectrum of a 56 mM solution4 of sodium formate biosynthesized from [l-'3C]methylthioribose contained a singlet (6 = 8.24 ppm) and a doublet (6 = 8.24 ppm, 'Jcn = 194.8 Hz) corresponding to the fl*COO-Na' and R3COO-Na+ species, respectively. Integration of these signals revealed the ratio of H13C00-Na'/H'2C00-Na' to be l.O/l.O. The proton-decoupled 13C NMR spectrum of this sample contained an intense singlet (6 = 171.9 ppm) corresponding to H13COO-Na+. The area of this signal was -40 times greater than the signal arising from a 56 mM solution of natural abundance sodium formate (6 = 171.9 ppm). Use of the APT pulse sequence demonstrated that a single hydrogen atom was bonded to the carbon atom giving rise to the singlet at 6 = 171.9 ppm. Upon acidification of the sample, the expected upfield shifts were observed for the proton resonances of H"COOH (singlet, 6 = 8.03 ppm) and p3COOH (doublet, 6 = 8.03 ppm, 'JCH = 218.8 Hz), as well as for the carbon resonance of H13COOH (singlet, 6 = 166.6 ppm).
p-Phenylphenacyl formate, derived from the formate biosynthesized from [l-'3C]methylthioribose, was also examined by NMR. The 'H and 13C NMR spectra of this sample were superimposable upon those of authentic natural abundance p-phenylphenacyl formate except for: 1) the presence of an additional doublet in the 'H NMR spectrum ( 133 (MH' -NR) and 104 (MH+ -C02H2); the +l and +2 isotope satellite peaks were also observed for these signals, and their intensities agreed well with theory. The chemical ionization mass spectrum of methionine derived from [l-'3C]methylthioribose was identical in all respects to the spectrum obtained for natural abundance methionine (Table II). Similarly, the 'H and 13C NMR spectra of the biosynthesized methionine were identical in every respect to the spectra obtained for natural abundance methionine.
Thus, no enhancement was observed for any of the 13C signals of the biosynthesized methionine nor were any 13C-13C couplings detected by 13C NMR spectroscopy. 'H NMR spectroscopy failed to detect any 13C1H couplings in the biosynthetic sample. Based on these results, we conclude that C-l of methylthioribose is not incorporated into methionine as a result of the methionine salvage pathway. In view of the data presently available on this metabolic pathway (Scheme l), it seems highly unlikely that carbon chain rearrangements occur during the conversion of methylthioribose to methionine.
As seen from the data in Table I Thus, a major, if not the exclusive, metabolic product from C-l of methylthioribose is formate. Substantial amounts of methionine, lacking 13C enrichment, were also biosynthesized from [l-13C]methylthioribose during this experiment (Table I; cf. Table III).' The most reasonable interpretation of these results is that C-l of methylthioribose is converted directly to formate concomitant with the formation of methionine from carbon atoms 2-5 of methylthioribose. This interpretation is consistent with the results of studies performed with purified diketone 6 and partially purified enzymes from K. pneumoniue, in which 1 equivalent of formate/mol of a-keto acid 6 (the direct precursor of methionine) was produced (19). Our data do not rule out the possibility that some other l-carbon species, which is rapidly and quantitatively converted to formate, is produced from Cl of methylthioribose during methionine biosynthesis. However, formaldehyde was not detected as a product of the conversion of diketone 5 to a-keto acid 6 and formate in a partially purified system (19). Taken together, these results suggest that C-l of diketone 5 is converted directly to formate concomitant with the formation of cY-keto acid 6 from carbon atoms 2-5.
The data in Table I also demonstrate the biosynthesis of substantial amounts of natural abundance formate from [l-'3C]methylthioribose, which presumably results from alter-' We estimate that -14 pmol of methionine was biosynthesized from [l-'3C]methylthioribose in the experiment presented in Table I. This estimate is based on the results of previous studies in which we established that the recovery of methionine, when purified as described for the experiment presented in Table I (Tables I and III).' These data suggest that the conversion of methylthioribose to 3-S-methyl-3-thiopropionate results in the formation of formate from C-l and apparently also from C-2, whereas 3-S-methyl-3-thiopropionate originates from carbon atoms 3-5 of methylthioribose. This suggestion is consistent with the stoichiometry observed in a partially purified system of 2 mol of formate generated per mol of 3-S-methyl-3-thiopropionate produced from diketone 5 (19). scheme employed was specifically designed to minimize the exposure of methionine and formate to acidic conditions, to avoid the loss of enzymatically incorporated "0 in the resulting carboxylic acids. However, the use of acidic conditions during purification could not be completely avoided and therefore loss of "0 from the biosynthesized methionine and formate was anticipated. A parallel incubation, lacking methylthioribose, was carried out in an identical fashion. The amounts of methionine and formate produced in both incubations were determined, respectively, by radiochemical dilution and the formate dehydrogenase assay (Table III). A portion of the isolated formate was then derivatized to p-phenylphenacyl formate which was purified by normal phase silica HPLC. This sample, as well as the isolated methionine sample, were then analyzed by chemical ionization mass spectrometry. and "OZ is demonstrated by the data in Table IV. For example, whereas the relative ratio of the signal intensities for the peaks at m/e 152 versus 150 ((MH + 2)+ and MH', respectively) and 135 versus 133 ([(MH -NH,) + 2]+ and (MH -NH3)+, respectively) were both 0.06/ 1.0 for natural abundance methionine, these ratios were increased to 1.4/1.0 and 1.2/1.0 in the methionine sample isolated from the incorporation experiment.5 By contrast, the ratio of the signal intensities for the peaks at m/e 106 and 104 ([(MH -C02H2) + 21' and (MH -COzH2)', respectively) were essentially identical for both the natural abundance and biosynthetic methionine samples. Such data demonstrated the incorporation of one atom of "0 from '*Oz into a portion of the methionine isolated from the incubation containing [l-'"Clmethylthioribose.
Based on the MS data, the ratio of ['801]/['60]methionine isolated from the incorporation experiment was 1.3/l. Thus, 56% of the isolated methionine contained one atom of "0 from 1802 in its carboxyl group. No evidence for the formation of ["OJmethionine was found. We attribute the observed lower than stoichiometric incorporation of "0 in the isolated methionine primarily to chemically induced washout of the biosynthesized [1801]methionine during its purification. This washout most likely occurred as a result of the strongly acidic conditions encountered during the cation exchange chromatography step. Strongly acidic conditions are well known to catalyze washout of "O-labeled carboxylic acids, including amino acids (21,22). Based on these results, it seems most likely that the conversion of methylthioribose to methionine results in the incorporation of one atom of oxygen from dioxygen into the carboxylate group of the resultant methionine.
Similarly, the results in Table IV  and 90 versus 87 ((HCO&H,CO + 3)+ and HCO&H&O+, respectively) were both 0.00/l for natural abundance p-phenylphenacyl formate, these ratios were increased to 0.05/l in the p-phenylphenacyl formate sample derived from the formate isolated from the incorporation experiment. loss of "0 is due to further chemical washout caused by the increased exposure to acidic conditions. Other experiments by Corina (24) have demonstrated 76% washout of "0 from ['"Olformate upon incubation in 1 N aqueous hydrochloric acid for 45 min at 25 "C. Based on these data, it seems most likely that the conversion of C-l of methylthioribose into formate results in the incorporation of one atom of oxygen from dioxygen into the resultant formate.
Quantitative analysis (Table III)  One mol of dioxygen is consumed per mol of diketone 5 converted to cu-keto acid 6 and formate (19). No other 02consuming reaction is involved in the conversion of methylthioribose to methionine and formate. We have now demonstrated the incorporation of one atom of "0 from 1802 into the carboxyl groups of both methionine and ['%]formate derived by the metabolism of [ 1-i3C]methylthioribose, and provided evidence suggesting that the "0 incorporation is stoichiometric.
That the incorporation of "0 into formate and methionine is not the result of some process unrelated to methylthioribose metabolism is suggested by: most reasonable interpretation of these results is that the conversion of diketone 5 to a-keto acid 6 and formate results in the incorporation of one atom of oxygen from dioxygen into the carboxylate group of both products.
Based on these arguments, we suggest that the oxidative carbon-carbon bond cleavage in the conversion of methylthioribose into methionine and formate is effected in a single reaction catalyzed by a novel intramolecular dioxygenase. A speculative mechanism, consistent with the presently available data, for the dioxygenase-mediated conversion of diketone 5 into cu-keto acid 6 and formate is presented in Scheme 2. According to this mechanism, an enzyme generated, phosphorylated enediol 8 serves as the nucleophile for attack on dioxygen. By analogy, a similar carbanion mechanism has been proposed for substrate oxygenation by ribulose-1,5-bisphosphate oxygenase (25). Dioxygen may be enzymatically activated for nucleophilic attack by, for example, metal catalysis. The nucleophilic oxygen atom of the C-l peroxide group of the resulting intermediate (9) then attacks the electrophilic carbonyl carbon at C-2, yielding intermediate 10. Decomposition of dioxetane 10 into cY-keto acid 6 and formate is then initiated by hydrolysis of the C-l phosphate ester. Alternatively, decomposition of dioxetane 10 might be initiated by deprotonation of the C-2 hydroxyl group, in which case (Yketo acid 6 and formyl phosphate would be produced (not shown). Due to its hydrolytic instability (26), any formyl phosphate formed in our incubations would be converted to formate during work-up. Clearly, other variations of the mechanism shown in Scheme 2 exist. Notable among these is the possibility that oxygenation of the enediol-phosphate 8 occurs at C-2, C-2 carbanion formation being initiated by hydrolysis of the C-l phosphate ester; in this case, the intermediate dioxetane 10 would not be phosphorylated.
It should be noted