Identification of a pathway for the utilization of the Amadori product fructoselysine in Escherichia coli.

Escherichia coli was found to grow on fructoselysine as an energetic substrate at a rate of about one-third of that observed with glucose. Extracts of cells grown on fructoselysine catalyzed in the presence of ATP the phosphorylation of fructoselysine and a delayed formation of glucose 6-phosphate from this substrate. Data base searches allowed us to identify an operon containing a putative kinase (YhfQ) belonging to the PfkB/ ribokinase family, a putative deglycase (YhfN), homologous to the isomerase domain of glucosamine-6-phosphate synthase, and a putative cationic amino acid transporter (YhfM). The proteins encoded by YhfQ and YhfN were overexpressed in E. coli, purified, and shown to catalyze the ATP-dependent phosphorylation of fructoselysine to a product identified as fructoselysine 6-phosphate by 31P NMR (YhfQ), and the reversible conversion of fructoselysine 6-phosphate and water to lysine and glucose 6-phosphate (YhfN). The K(m) of the kinase for fructoselysine amounted to 18 microm, and the K(m) of the deglycase for fructoselysine 6-phosphate, to 0.4 mm. A value of 0.15 m was found for the equilibrium constant of the deglycase reaction. The kinase and the deglycase were both induced when E. coli was grown on fructoselysine and then reached activities sufficient to account for the rate of fructoselysine utilization.

Fructosamines are the products of a non-enzymatic reaction of glucose with primary amines followed by an Amadori rearrangement. These reactions, known as glycation (to be distinguished from glycosylation, which is enzymatically catalyzed), typically modify the amino terminus and the lysine side-chains of proteins (reviewed in Refs. [1][2][3], as well as a variety of low molecular weight compounds including aminophospholipids (4). Their interest for human physio(patho)logy is 2-fold. A first aspect is that fructosamines form spontaneously and slowly in the body in proportion to the blood glucose concentration. The concentration of protein-bound fructosamine in the serum and the level of HbA1c (a form of hemoglobin with a fructosamine residue) are used to estimate the mean blood glucose concentration in the preceding weeks or months (5)(6)(7). Furthermore, fructosamines may participate in the development of diabetes complications (3). A second aspect is that low amounts of fructosamines are present in various foods (8). Fructoselysine, which is presumably released from glycated proteins in the course of digestion, is apparently partly metabolized by bacteria in the hind gut (8) (see also "Discussion").
The understanding of the metabolism of fructosamines has significantly progressed during the last few years. Various microorganisms (Pseudomonas sp., Corynebacterium sp., Aspergillus sp.) have been shown to produce amadoriases, which catalyze the oxidative cleavage of low molecular weight fructosamines (9 -13). In addition, a mammalian fructosamine 3-kinase acting on low molecular weight and protein-bound fructosamines has recently been identified, purified, and cloned. The role of this enzyme is most likely to initiate an intracellular protein "deglycation" process (14 -17).
While overexpressing mammalian fructosamine 3-kinase in Escherichia coli, we noted that control bacterial extracts contained low activities of an enzyme capable of phosphorylating fructoselysine. This stimulated us to study the metabolism of this compound in E. coli. In this paper, we show that fructoselysine can sustain growth of E. coli and that it is metabolized by a pathway involving a fructoselysine 6-kinase and an enzyme converting fructoselysine 6-phosphate to glucose 6-phosphate and lysine.
Non-radioactive fructoselysine and [ 14 C]fructoselysine (labeled on its deoxyfructose moiety) were synthesized as described elsewhere (17). For the synthesis of fructoselysine 3-phosphate, [ 14 C]fructoselysine (7.10 6 cpm) was incubated for 30 min at 30°C with 240 g of pure recombinant mouse fructosamine 3-kinase (15) in the presence of 25 mM Tris-HCl, pH 7.8, 5 mM ATP-Mg, 1 mM EGTA, and 0.1 mg/ml bovine serum albumin in a final volume of 2.5 ml. The sample was diluted with one volume of water and loaded onto a 10-ml column of AG 50W-X4 (Na ϩ ) equilibrated with 10 mM Hepes, pH 7.1. The column was washed with water to elute fructoselysine 3-phosphate, whereas fructoselysine remained bound to the column.
For the synthesis of fructoselysine 6-phosphate, 5 mM fructoselysine was incubated for 30 min at 30°C in a mixture (15 ml) containing 10 mM ATP-Mg, 50 mM Hepes, pH 7.1, and 5 units/ml of purified fructoselysine 6-kinase (YhfQ), which led to its complete phosphorylation. The reaction was stopped by the addition of 7.5 ml of ice-cold 10% (w/v) HClO 4 . After centrifugation, the supernatant was neutralized with 3 M KHCO 3 and diluted 5-fold with 20 mM sodium acetate, pH 5.0. The * This work was supported by the Concerted Research Action Program of the Communauté Française de Belgique, the Belgian Federal Service for Scientific, Technical, and Cultural Affairs, and by the Juvenile Diabetes Foundation International. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Conversion of [ 14 C]Fructoselysine in Bacterial Extracts-The incubation mixture (200 l) contained 25 mM Hepes, pH 7.1, 1 mM MgCl 2 , 1 mM dithiothreitol, 1 mM EGTA, 5 mM ATP-Mg, 100 M fructoselysine, and 20,000 cpm [ 14 C]fructoselysine. The reaction was initiated by addition of 2 l of bacterial extract (25 g of protein) and stopped after various times (0 -60 min) by mixing samples with 0.5 volume of cold 10% (w/v) HClO 4 . After a 10-min centrifugation at 2000 ϫ g, the supernatants were neutralized with 3 M KHCO 3 . After elimination of the KClO 4 precipitate by centrifugation, 250 l of the supernatant was mixed with 5.75 ml of 10 mM Tris-HCl, pH 8.0, and applied onto a 1-ml AG 1-X8 (Cl Ϫ ; in a Pasteur capillary pipette) column equilibrated with the same buffer. The column was washed with 4 volumes of 10 mM Tris-HCl, pH 8.0, and the retained radioactivity was eluted by applying 8 ml of 500 mM NaCl. The fractions were mixed with Optima Gold (Packard) scintillation fluid and counted for radioactivity.
Formation of Glucose 6-phosphate from Fructoselysine and ATP Catalyzed by E. coli Extracts-Reactions were carried out in glass cuvettes in a mixture (1 ml) containing 25 mM Hepes, pH 7.1, 0.25 mM NADP, 5 mM MgCl 2 , 1 mM ATP-Mg, 0.1 mM fructoselysine, 5 g of yeast glucose-6-phosphate dehydrogenase (grade I), and a bacterial extract (35 g of protein). A 340 was monitored to follow the appearance of NADPH.
Fructoselysine 6-kinase Assays-All enzymatic assays were carried out at 30°C. Fractions of the DEAE-Sepharose columns were assayed spectrophotometrically with a pyruvate kinase/lactate dehydrogenasecoupled assay. The assay mixture (1 ml) contained 25 mM Hepes, pH 7.1, 25 mM KCl, 1 mM MgCl 2 , 1 mM dithiothreitol, 0.25 mM phospho(enol)pyruvate, 0.15 mM NADH, 1 mM ATP-Mg, 0.5 mM fructoselysine, 10 g of rabbit muscle pyruvate kinase and 5 g of rabbit muscle lactate dehydrogenase. Fructoselysine 6-kinase (up to 4 milliunits) was added to initiate the reaction, which was followed by measuring A 340 . A deglycase-coupled assay was used to measure the enzymatic activity in crude extracts in the experiment shown in Fig. 9. The assay mixture (1 ml) contained 25 mM Hepes, pH 7.1, 25 mM KCl, 5 mM MgCl 2 , 1 mM dithiothreitol, 1 mM ATP-Mg, 0.1 mM EGTA, 0.25 mM NADP, 0.5 mM fructoselysine, 5 g of yeast glucose-6-phosphate dehydrogenase (grade I), and ϳ30 milliunits of purified deglycase. Fructoselysine 6-kinase (up to 2 milliunits) was added to initiate the reaction. One unit of enzyme is the amount that catalyzes the conversion of 1 mol of substrate per min under the indicated conditions.
Fructoselysine was assayed using fructoselysine 6-kinase and either pyruvate kinase and lactate dehydrogenase or deglycase and glucose-6-phosphate dehydrogenase. The assay mixtures were the same as described above for the assay of fructoselysine 6-kinase, except that fructoselysine was in limiting amount (up to 50 M). Coupling enzymes were added first and, when a plateau was reached, ϳ2 units/ml fructoselysine 6-kinase.
Fructoselysine 6-phosphate was measured spectrophotometrically using fructoselysine-6-phosphate deglycase and glucose-6-phosphate dehydrogenase. The assay mixture was the same as described above for the measurement of the deglycase, except that fructoselysine 6-phos-phate was in limiting amount (up to 50 M). Glucose-6-phosphate dehydrogenase was added first and then 20 milliunits of deglycase.
Overexpression and Purification of Fructoselysine 6-kinase-A 5Ј primer containing the putative ATG codon (GCATATGAAAACCCTG-GCGACAATCGG) in an NdeI site (in bold) and a 3Ј primer containing the putative stop codon (CGGATCCTACCAGGCACCGTGGTACTG) flanked by a BamHI site were used to PCR-amplify genomic DNA from E. coli BL21(DE3) with Pwo polymerase. An ϳ800-bp product was obtained, which was subcloned in pBlueScript and checked by sequencing. A NdeI-BamHI fragment was removed from the pBlueScript plasmid and ligated in pET-3a. This vector was used to transform BL21(DE3)pLysS (18). The resulting bacteria were grown in 1 liter M9 medium supplemented with 0.5 mg/liter biotin, 0.5 mg/liter thiamine, 2 g/liter glucose, 1 g/liter Casamino acids, 0.1 g/liter ampicillin, and 25 mg/liter chloramphenicol. The culture was grown at 37°C until A 600 reached 0.5-0.6. It was then cooled on ice for 20 min and the inducer isopropyl-1-thio-␤-D-galactopyranoside was added to a final concentration of 0.4 mM. After a further incubation at 18°C for 22-24 h, the cells were collected by centrifugation, resuspended in 50 ml of buffer A (20 mM Hepes, pH 7.4, 5 mM EDTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 5 g/ml antipain, 1 mg/ml lysozyme) and submitted to three cycles of freezing and thawing. The bacterial extract was incubated on ice for 1 h with 5 mg of DNaseI in the presence of 10 mM MgSO 4 and centrifuged for 30 min at 10,000 ϫ g. The resulting supernatant (45 ml) was made 22% (w/v) in poly(ethylene glycol) 6000 and centrifuged for 15 min at 10,000 ϫ g. The supernatant was diluted 2-fold with buffer B (25 mM Hepes, pH 7.1, 10 mM KCl, 1 mM dithiothreitol, 5 g/ml leupeptin, 5 g/ml antipain) and loaded onto a DEAE-Sepharose column (30 ml) equilibrated with 20 mM Hepes, pH 7.1. The column was washed with 150 ml of buffer B, and protein was eluted with a linear NaCl gradient (0 -0.5 M in 2 ϫ 125 ml of buffer B). Fractions of 3.5 ml were collected. Protein was assayed (19) using bovine ␥ globulin as a standard.
Overexpression and Purification of Fructoselysine-6-phosphate Deglycase-Two 5Ј primers containing each a potential initiator codon (GCATATGTTGGATATTGATAAAAGCACCGT and GCATATGGT-TCAGGAAGTGGAAAAAGTT) in a NdeI site (in bold) and a 3Ј primer containing the putative stop codon (CGGATCCTTAATATTCCACCA-GACCACCGTAA) flanked by a BamHI site were used to PCR-amplify genomic DNA from E. coli BL21(DE3) with Pwo polymerase. The obtained PCR products (ϳ1000 bp) were cloned in pET-3a as described above for fructoselysine 6-kinase. Expression of the protein was carried out as described above. The cells derived from a 1-liter culture were collected by centrifugation, resuspended in 40 ml of a buffer containing 20 mM Hepes, pH 7.4, 5 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, and 5 g/ml antipain and extracted in a French press. The bacterial extract was centrifuged for 15 min at 10,000 ϫ g. The resulting supernatant was diluted 3-fold with buffer C (20 mM Hepes, pH 7.1, 1 mM dithiothreitol, 1 mM EGTA, 5 g/ml leupeptin, and 5 g/ml antipain) and applied onto a DEAE-Sepharose column (30 ml) equilibrated with 20 mM Hepes, pH 7.1. The column was washed with 150 ml of buffer C, and protein was eluted with a linear NaCl gradient (0 -0.75 M in 2 ϫ 125 ml of buffer C). Fractions of 3.5 ml were collected.
Preparation and NMR Analysis of Phosphorylated Fructoselysine-Twenty five moles of fructoselysine were incubated in the presence of 10 mM ATP-Mg, 50 mM Hepes, pH 7.1, and 40 units of fructoselysine kinase during 15 min at 30°C. The incubation medium was mixed with 0.5 volume of cold 10% (w/v) HClO 4 , and the supernatant was neutralized with 3 M KHCO 3 . The sample was centrifuged for 10 min, and the supernatant was diluted 30-fold in 20 mM Tris-HCl, pH 8, and loaded on a 48-ml column of AG 1-X8 (Cl Ϫ ) equilibrated in the same buffer. The column was washed with 100 ml of the same buffer and fructoselysinephosphate was eluted with a linear NaCl gradient (0 -0.5 M in 2 ϫ 75 ml). Fractions containing fructoselysine-phosphate (assayed with the deglycase) were pooled, concentrated to a volume of 1.5 ml, and loaded on a Biogel P2 fine column (0.9 ϫ 50 cm) equilibrated with water to separate the phosphorylated compound from salts. Fractions containing fructoselysine-phosphate were pooled and freeze-dried for NMR analysis (15). The latter was run in a Bruker DRX500 spectrometer at 27°C. The pH value of the sample was 4.1.
Growth of E. coli on Fructoselysine-E. coli BL21(DE3)pLysS strain was precultured in LB medium (containing 25 mg/liter chloramphenicol) until A 600 was equal to ϳ1.8. The cells were centrifuged and resuspended in M9 medium (devoid of carbon source); 0.4 ml of this suspension was used to seed 16 ml of M9 medium supplemented with 25 mg/liter chloramphenicol and the appropriate substrate (fructoselysine, glucose, glucose ϩ lysine, lysine, all 20 mM, sterilized by filtration on a 0.22-m membrane). Samples (1 ml) were taken at different times to measure A 600 . They were then immediately centrifuged for 15 min at 5000 ϫ g. Perchloric extracts were prepared from the resulting supernatants to measure glucose and fructoselysine. The cell pellets were resuspended in 50 l of buffer A and extracted as described for fructoselysine 6-kinase. Proteins, fructoselysine 6-kinase, and deglycase activities were measured in the supernatants obtained after a 30-min centrifugation at 10,000 ϫ g. Fig. 1A, extracts of E. coli grown in the presence of fructoselysine catalyzed the ATP-dependent conversion of [ 14 C]fructoselysine (labeled on its deoxyfructose moiety) to anionic products suggesting the existence of a fructoselysinekinase activity. No such activity could be detected under these experimental conditions with an extract of cells that had been grown on glucose instead of fructoselysine. Analysis of the reaction products on cation-exchanger under acidic conditions (not shown) indicated that no conversion of fructoselysine oc-curred in the absence of ATP, suggesting that the phosphorylation did not take place after cleavage of [ 14 C]fructoselysine to lysine and a radioactive sugar.

Metabolism of Fructoselysine in Bacterial Extracts-As shown in
Because fructoselysine contains a carbohydrate moiety and appeared to be phosphorylated, we wondered if it would be converted to glucose 6-phosphate in bacterial extracts. In the experiment shown in Fig. 1B, we monitored the formation of glucose 6-phosphate in extracts of E. coli incubated with NADP and glucose-6-phosphate dehydrogenase. A progressive increase in the formation of NADPH took place after a lag period in extracts of induced cells incubated with fructoselysine and ATP, but not if one of these two substrates was omitted, or if an extract of cells grown on glucose was used. If exogenous glucose-6-phosphate dehydrogenase was omitted, a smaller increase was observed, presumably due to the presence of endogenous glucose-6-phosphate dehydrogenase. These results indicated therefore that the product of phosphorylation of fructoselysine was converted to glucose 6-phosphate in cell-free extracts of E. coli that had grown on this fructosamine.
Because the E. coli genome contains a sequence showing about 30% identity with mammalian fructosamine 3-kinase (15), extracts of cells grown with fructoselysine were incubated with radiolabeled [ 14 C]fructoselysine 3-phosphate with or without ATP. Analysis of the reaction mixtures by chromatography on anion-exchanger, under conditions allowing the separation of hexose-monophosphates from fructoselysine 3-phosphate did not indicate conversion of fructoselysine 3-phosphate in these extracts (not shown). Furthermore no evidence for the conversion of fructoselysine 3-phosphate (100 M) to glucose 6-phosphate could be obtained in experiments similar to those presented in Fig. 1B (not shown).
Taken together, these results indicated therefore that the metabolism of fructoselysine in E. coli proceeded through the phosphorylation of its deoxyfructose moiety on a carbon distinct from C3. Because sugar kinases commonly phosphorylate a hydroxyl group on a terminal carbon, this carbon was most likely C6. There is indeed no hydroxyl group on C1 in fructoselysine.
Identification of an Operon Containing the Enzymes Responsible for the Metabolism of Fructoselysine-We next searched the E. coli genome for a potential operon (in the present paper, we use this term to mean a series of neighboring ORFs, 1 all in the same orientation, without actual proof that they are transcribed as a single mRNA) responsible for the degradation of fructoselysine. Our hypothesis was that such an operon should contain sequences encoding: 1) a transporter for fructoselysine, possibly homologous to cationic amino acid transporters; 2) a kinase able to phosphorylate fructoselysine on C6 of its deoxyfructose moiety; this kinase would probably belong to the PfkB/ ribokinase protein family (20 -22), which comprises a series of enzymes phosphorylating a hydroxymethyl group bound to a furanose ring; and 3) a "deglycase", able to catalyze the conversion of fructoselysine 6-phosphate to lysine and glucose 6-phosphate; such a reaction is analogous to those catalyzed by glucosamine-6-phosphate synthase and glucosamine-6-phosphate isomerase (see "Discussion").
PSI-BLAST (23) searches were carried out to identify E. coli homologues of several of the proteins mentioned above (lysine and arginine transporters, PfkB, glucosamine-6-phosphate synthase, glucosamine-6-phosphate deaminase). For each hit, we tried to identify the function of the surrounding ORFs by performing additional BLASTp (24) or PSI-BLAST searches. 1 The abbreviations used are: ORF, open reading frame; HMQC, 1 H-detected heteronuclear multiple-quantum coherence via direct coupling. One operon with the expected features was identified (Fig. 2). It contains ORFs registered as YhfM, YhfN, YhfO, YhfP, YhfQ, and YhfR. YhfM (P45539) is a putative transporter sharing ϳ25% identity with eukaryotic and prokaryotic cationic amino acid transporters (not shown). YhfN (P45540) belongs to the same family of proteins as the isomerase domain of glucosamine-6-phosphate synthase (P17169) (Fig. 3). One of its closest homologues is MocD (T44930), a protein encoded by the Ti plasmid of Agrobacterium tumefaciens and thought to act, in the metabolism of mannopine, as a "deconjugase" that cleaves santhopine (fructoseglutamine) to glutamine and a sugar (25). Another close homologue of MocD and YhfN is YurP (CAB15251), an ORF present in the genome of Bacillus subtilis. YhfQ (P45543), which encodes a kinase of the PfkB family, is homologous to proteins encoded by ORFs (MocE, T44929 and YurL, O32153) belonging to the same operons as MocD and YurP (Fig. 2). YhfR (P45544) encodes a putative repressor homologous to YurK (O32152) belonging to the B. subtilis operon mentioned above.
Expression and Characterization of the Putative Fructoselysine Kinase-The ORF corresponding to YhfQ (P45543) was PCR-amplified and inserted into a pET-3a plasmid for expression in E. coli BL21(DE3)pLysS (18). Such cells expressed large amounts of a protein with the expected size and displayed an elevated fructoselysine kinase activity (14 mol/min/mg protein). Fructoselysine kinase was purified by fractionation with poly(ethylene glycol) and chromatography on DEAE-Sepharose (not shown). The enzyme eluted as a single peak with a specific activity of ϳ30 mol/min/mg protein and was nearly homoge-neous (Fig. 4A). About 175 mg of purified protein were obtained from a 1-liter culture. Gel filtration on Sephacryl S-200 showed that the native protein had a molecular mass of 28 kDa (Fig. 5), indicating that it is a monomer.
The kinetic properties were studied on the purified preparation. The K m for fructoselysine was 18 M in the presence of 5 mM ATP-Mg, and the K m for ATP was 50 M in the presence of 0.5 mM fructoselysine. Deoxymorpholinofructose (K m , 24 mM; V max , 6.5 mol/min/mg protein) and fructoseglycine (K m , 80 mM; V max , 0.8 mol/min/mg protein) were much poorer substrates than fructoselysine. Fructose phosphorylation was barely detectable, amounting to ϳ0.01 mol/min/mg protein at 50 mM fructose.
Identification of the Phosphorylation Product as Fructoselysine 6-phosphate-The product of fructoselysine phosphorylation was purified. Mass spectrometry analysis indicated the presence of a negative ion with the expected m/z ratio for fructoselysine phosphate (387.3). 31 P NMR spectroscopy (Fig.  6) showed triplet resonances, indicating that phosphate was esterified to a carbon bearing two hydrogen atoms, i.e. C6 of fructoselysine ( 3 J POCH ϭ 6.4 Hz). Three resonances were observed with chemical shifts of 1.32, 0.84, and 0.70 ppm and in a 0.1:1:0.9 intensity ratio at pH 4.1. These corresponded most likely to the keto-form, the ␤-furanose form and the ␣-furanose form of fructoselysine 6-phosphate. The coupling between phosphorus and methylene protons was further confirmed in a two-dimensional spectrum correlating 31 P and 1 H (HMQC spectrum, not shown). 1 H-NMR spectra (one dimensional and COSY) indicated the presence of several distinct resonances that could be unambiguously assigned to the hydrogens bound to carbon ␣ to ⑀ of the lysine moiety and to carbons 1 and 6 of the deoxyfructose moiety. Resonances corresponding to hydrogens bound to C3 to C5 of the deoxyfructose moiety could not be assigned unambiguously due to strong overlap of resonances.
Expression and Characterization of the Putative Deglycase-We prepared two expression vectors containing either the first or the second AUG codon of YhfN (P45540) as initiation codon. The first of the two constructs led to the production of a soluble protein of the expected size (39 kDa; Fig. 4B), whereas the second construct yielded an insoluble protein (not shown).
Extracts of cells expressing the first construct catalyzed the conversion of 100 M fructoselysine 6-phosphate to glucose 6-phosphate at a rate of 60 nmol/min/mg protein. No such activity was detected with extracts of control cells or of cells expressing the shorter form of YhfN. The deglycase was purified by chromatography on DEAE-Sepharose, from which it was eluted at a high salt concentration (400 mM NaCl, Fig. 7), in agreement with the calculated pI value (4.79) of the protein.
The enzyme was homogeneous after this purification step and displayed a specific activity of 0.14 mol/min/mg protein when measured with 100 M substrate. About 270 mg purified protein were obtained from a 1-liter culture. The enzyme was free from phosphoglucose isomerase, which allowed us to conclude that it produced glucose 6-phosphate and not fructose 6-phosphate. Gel-filtration on Sephacryl S-300 indicated that the protein had an apparent molecular mass of ϳ450 kDa, suggesting a dodecameric structure.
The enzyme was stimulated ϳ2-fold by EGTA, which had a maximal effect at a concentration of 10 M, and by EDTA, which had a maximal effect at higher concentrations (0.1 mM). It was strongly inhibited by ZnCl 2 , which, at 10 M, completely suppressed the activity; 0.1 mM EGTA was therefore included in all further assays. The K m for fructoselysine 6-phosphate amounted to 0.4 mM.
The reaction catalyzed by the deglycase could also be measured in the reverse direction through the conversion of [ 14 C]glucose 6-phosphate and lysine to a radioactive product that was retained on cation exchanger at acidic pH (Fig. 8).  7. Chromatography of fructoselysine-6-phosphate deglycase on DEAE-Sepharose. A bacterial extract derived from a 1-liter culture was applied onto a DEAE-Sepharose column and eluted with a NaCl gradient. The activity of the deglycase was measured spectrophotometrically at 0.1 mM fructoselysine 6-phosphate. Protein was measured according to (18). Fig. 9A illustrates that E. coli grows on fructoselysine at a rate of about one-third of that observed with glucose as a carbon source. Lysine itself did not support growth in the absence of other carbon source and did not affect the growth observed with glucose. Fig. 9B shows the rate of disappearance of glucose and fructoselysine (assayed using fructoselysine 6-kinase and the deglycase). The latter amounted to ϳ1.5 mol/ml/hour at mid logarithmic phase corresponding to ϳ70 nmol/min/mg soluble protein (at 37°C). The specific activity of fructoselysine 6-kinase (measured with a deglycase-coupled assay) in extracts of cells grown on fructoselysine amounted to ϳ30 nmol/min/mg protein (at 30°C) and that of fructoselysine-6-phosphate deglycase, to ϳ6 nmol/min/mg of protein (at 100 M fructoselysine 6-phosphate and 30°C). No activity was detectable in cells grown on glucose with or without lysine with the spectrophotometric assay (not shown). Assays of the kinase with radiolabeled fructoselysine (10 M) indicated an activity of ϳ0.02 nmol/min/mg protein.

DISCUSSION
Nature of the Catalyzed Reactions-We describe in this paper a novel pathway for the utilization of the Amadori product fructoselysine, involving a kinase and a deglycase. The first one phosphorylates fructoselysine on C6 of the deoxyfructose moiety, as indicated by NMR analysis. This conclusion is also consistent with the fact that the kinase belongs to the PfkBribokinase family (20 -22), which comprises enzymes that phosphorylate primary alcohols (e.g. C1 of fructose 6-phosphate or fructose, C5 of ribose, C6 of fructose 1-phosphate). The deglycase catalyzes a reaction that is easily reversible, at least in vitro. The fact that lysine and glucose 6-phosphate are used in the reverse reaction confirms that they are the products of the forward reaction. The equilibrium constant of the reaction (0.15 M) suggests, however, that the enzyme serves in vivo to produce glucose 6-phosphate and lysine from fructoselysine 6-phosphate rather than for the opposite conversion.
When considered in the non-physiological direction, the reaction catalyzed by the deglycase most likely involves the formation of a Schiff base with C1 of glucose 6-phosphate, followed by an isomerization step. It is therefore similar to the reaction catalyzed by glucosamine-6-phosphate synthase. This enzyme contains a glutaminase domain and a sugar isomerase domain, both of which have been crystallized (26 -28). Ammonia generated by the glutaminase domain serves to form a Schiff base with C2 of fructose 6-phosphate, and the Schiff base is then isomerized to glucosamine 6-phosphate. The role of C2 of glucose 6-phosphate in fructoselysine-6-phosphate deglycase is analogous to that played by C1 in glucosamine-6-phosphate synthase. Remarkably, fructoselysine-6-phosphate deglycase is homologous to the isomerase domain of glucosamine-6-phosphate synthase. Two of the residues thought to be involved in catalysis (His-504 and Glu-488, underlined in Fig. 3) in glucosamine-6-phosphate synthase are conserved in the fructoselysine-6-phosphate deglycase. The first one is thought to participate in ring opening, and the second, in a proton transfer reaction between C1 and C2 (26,27). Lys-603, known to form a Schiff base with fructose 6-phosphate in glucosamine-6-phosphate synthase (29), is not conserved in the deglycase, which suggests that Schiff base formation in the reverse deglycase reaction occurs directly with the epsilon amine of the substrate.
Putative Function of the Other ORFs-The fructoselysine operon also encodes a transporter, which most likely serves to transport fructoselysine, and a repressor, most likely responsible for the regulation of the expression of the operon by fructoselysine. The two other ORFs, O and P, appear to be one single ORF in E. coli 0157H7 (AAG58480). Their existence as two separate ORFs in E. coli K12 (P45541, P45542) may be due to a mutation or to a sequencing error. When this putative mutation is "corrected", ORF "OP" shows 21% sequence identity with tagatose 3-epimerase (30), a bacterial enzyme that interconverts D-tagatose and D-sorbose, as well as D-fructose and D-psicose (31). We speculate that Yhf OP catalyzes the reversible isomerization of psicoselysine to fructoselysine or of psicoselysine 6-phosphate to fructoselysine 6-phosphate.
tBLASTn searches with finished and unfinished bacterial genomes suggest that the occurrence of the fructoselysine operon is not common. One of the two closest protein homologues that we have found for the deglycase encodes an enzyme (MocD) involved in mannopine utilization. It has previously been suggested that MocD acts on santhopine (fructoseglutamine), the oxidation product of mannopine (25). The facts that the same operon contains a kinase homologous to YhfQ (MocE) and that MocD is similar to the isomerase domain of glucosamine-6-phosphate synthase suggest that santhopine is first phosphorylated to santhopine 6-phosphate before being converted to glutamine and glucose 6-phosphate in a reaction analogous to that catalyzed by the fructoselysine-6-phosphate deglycase. The role of the deglycase homologue (YurP) found in the B. subtilis genome remains to be established.
Physiological Role-The two enzymes that we have identified and purified are most likely responsible for the metabolism of fructoselysine by E. coli. This is not only indicated by the observation that both enzymes are induced by fructoselysine, but also by the fact that their activities in extracts of cells grown in the presence of fructoselysine can account for the rate of fructoselysine metabolism, if one takes into account the effect of temperature and the fact that the deglycase activity was measured at a subsaturating (0.1 mM as compared with a K m of 0.4 mM) concentration of substrate.
Balance studies have indicated that only a minor part (up to about 10%) of ingested fructoselysine (in the form of glycated casein) is excreted in the urine or in the feces in man, suggesting metabolism of this compound (8). Since fructoselysine is slowly taken up by the intestinal mucosa, by hepatocytes, and kidney cells (8), one possibility is that part of this metabolism takes place in tissues and involves phosphorylation by fructosamine 3-kinase and degradation of the latter to deoxyglucosone, lysine, and inorganic phosphate (15)(16)(17). Another possibility is that fructoselysine is utilized by bacteria present in the gut. The latter is consistent with the finding that fructoselysine (as glycated casein) is progressively consumed when incubated with feces in the absence but not in the presence of an antimicrobial agent (8). Our results indicate that E. coli may participate in this intraintestinal metabolism.