myo-Inositol Catabolism in Bacillus subtilis*

The iolABCDEFGHIJ operon of Bacillus subtilis is responsible for myo-inositol catabolism involving multiple and stepwise reactions. Previous studies demonstrated that IolG and IolE are the enzymes for the first and second reactions, namely dehydrogenation of myo-inositol to give 2-keto-myo-inositol and the subsequent dehydration to 3D-(3,5/4)-trihydroxycyclohexane-1,2-dione. In the present studies the third reaction was shown to be the hydrolysis of 3D-(3,5/4)-trihydroxycyclohexane-1,2-dione catalyzed by IolD to yield 5-deoxy-d-glucuronic acid. The fourth reaction was the isomerization of 5-deoxy-d-glucuronic acid by IolB to produce 2-deoxy-5-keto-d-gluconic acid. Next, in the fifth reaction 2-deoxy-5-keto-d-gluconic acid was phosphorylated by IolC kinase to yield 2-deoxy-5-keto-d-gluconic acid 6-phosphate. IolR is known as the repressor controlling transcription of the iol operon. In this reaction 2-deoxy-5-keto-d-gluconic acid 6-phosphate appeared to be the intermediate acting as inducer by antagonizing DNA binding of IolR. Finally, IolJ turned out to be the specific aldolase for the sixth reaction, the cleavage of 2-deoxy-5-keto-d-gluconic acid 6-phosphate into dihydroxyacetone phosphate and malonic semialdehyde. The former is a known glycolytic intermediate, and the latter was previously shown to be converted to acetyl-CoA and CO2 by a reaction catalyzed by IolA. The net result of the inositol catabolic pathway in B. subtilis is, thus, the conversion of myo-inositol to an equimolar mixture of dihydroxyacetone phosphate, acetyl-CoA, and CO2.

genes/Klebsiella mobilis) (3), Rhizobium leguminosarum bv. viciae (4), Sinorhizobium meliloti (5), Sinorhizobium fredii (6), Corynebacterium glutamicum (7), and Lactobacillus casei (8) can grow on MI as the sole carbon source. MI catabolism in A. aerogenes was studied biochemically, and a pathway of the catabolism finally yielding acetyl-CoA and dihydroxyacetone phosphate (DHAP) was proposed (9). However, our knowledge of the molecular genetics of bacterial MI catabolism has been restricted to B. subtilis (1, 10 -12). In B. subtilis, the iol divergon, comprising the operons iolABCDEFGHIJ and iolRS (1), and the iolT gene (12) were shown to be required for inositol catabolism (Fig. 1). Nowadays, a large number of bacteria have genes annotated iol in their genome sequence, but the annotation is only based on sequence similarity to B. subtilis iol genes, as relatively few studies have been done to demonstrate the participation of the deduced iol genes in MI catabolism.
In B. subtilis, a repressor encoded by iolR is responsible for the regulation of all the iol genes (11,12). In the absence of MI in the growth medium, the IolR repressor binds to the operator site within the promoter regions to repress the transcription. In its presence, however, MI is taken up by the cell and converted to a catabolic intermediate that acts as an inducer by antagonizing IolR, thereby inducing the iol divergon and iolT (11,12). Consequently, inactivation of iolR makes the transcription of the iol divergon and the iolT gene constitutive (1).
Some enzymes involved in MI catabolism have been characterized. Inositol dehydrogenase encoded by iolG is responsible for the first step of the degradation cascade by converting MI to 2-keto-MI (2KMI), conversion of compound [1] to [3] see ( Fig.  1) (10). Recently IolG was also shown to be able to act on D-chiro-inositol (DCI, compound [2]) to yield 1-keto-DCI (compound [4]), and IolI interconverts 1-keto-DCI and 2KMI, indicating that not only MI but also DCI is metabolized through the MI catabolic pathway (Fig. 1) (14). In addition, pinitol (3-Omethyl-DCI) contained in soybean appeared to be an alternative substrate of IolG, and this substrate was degraded depending on the presence of functional iol genes (15). 2KMI dehydratase, encoded by iolE, is responsible for the second step to produce 3D-(3,5/4)-trihydroxycyclohexane-1,2-dione (THcHDO, compound [5], formerly called D-2,3-diketo-4-deoxy-epi-inositol) (13). The iolT and iolF genes encode the primary and secondary inositol transporters, respectively ( Fig. 1) (12). The iolA gene was shown to encode malonic semialdehyde (MSA) dehydrogenase (16), supposed to be involved in MI catabolism based on the assumption that the catabolic pathway is similar to that proposed by Anderson and Magasanik for A. aerogenes (9). But the reaction step producing MSA has not been identified. Homology searches indicated possible functions for the products of the iolC and iolJ genes (13), but none has been identified experimentally. In the present studies we tried to identify the functions of iolD, iolB, iolC, and iolJ, aiming to define the whole pathway of MI catabolism and to understand how it is linked to the transcriptional regulation of the iol genes.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Plasmids, and Growth Conditions-Bacterial strains and plasmids used in this study are listed in Table 1. B. subtilis 60015 (trpC2 metC7) is our standard laboratory strain. B. subtilis strains YF244, YF258, YF259, and YF260 are 60015 derivatives. YF244 possesses the iolR::cat mutation inactivating the IolR repressor to allow constitutive expression of the iol divergon (1). In a previous study, some iol mutants defective in MI utilization were obtained by ethyl methanesulfonate mutagenesis, and subsequent detection of colonies were unable to ferment inositol on tetrazolium-containing plates (17). The complementation test through transformation using various PCR fragments (18) covering each of the iol genes demonstrated that some appeared to possess mutations localized within iolB and iolC. Subsequent nucleotide sequencing revealed three mutations of iolB52 (Gln-137 (CAA) to stop (TAA)), iolB58 (Glu-162 (GAA) to Lys (AAA)), and iolC62 (Ala-269 (GCC) to Thr (ACC)) as described previously (13). Strains YF258, YF259, and YF260 have iolB52, iolB58, and iolC62, respectively, besides iolR::cat, which allows the constitutive iol expression (1). B. subtilis cells were grown on tryptose blood agar base (Difco) supplemented with 0.18% glucose at 30°C and incubated in S6 medium (19) containing 0.5% casamino acids (Difco), 50 g/ml each of Trp and Met, and supplemented with or without 10 mM MI at 37°C with shaking.
Cloning and Expression of the iol Genes in E. coli and Purification of the Gene Product-The native iolB, iolC, iolD, and iolE genes were cloned into the pUC18 vector and expressed in E. coli cells grown in TGA medium as described previously (13). To express iolB and iolJ and produce the corresponding proteins as C-terminal His 6 -tag fusions, pIOLB-His and pIOLJ-His, respectively, were constructed as follows. DNA fragments covering the iolB or iolJ coding regions, associated with their Shine-Dalgarno sequence and a short linker enabling the C-terminal His 6 tag fusion at the head and tail, respectively, were amplified from DNA of strain 60015 by PCR using the specific primers. The primers for the iolB PCR fragment were iolBE (CCCGAATTCTAGAAAGAGGAGGGCTTCTC, an EcoRI site is indicated in bold) and iolBHisC (CCATCGATGCGT-TCTAAAATCCATTCATG, a ClaI site is in bold), and those for the iolJ fragment were iolJE (CCGAATTCTTCAGCATG-TAATGGATAAAGG, an EcoRI site is in bold) and iolJHisE (CCGGAATTCGCCGACCTGCTGCTTAGCTG, in an EcoRI site is in bold). The iolB fragment was trimmed with EcoRI and  ClaI, whereas the iolJ fragment was cut with EcoRI alone, and each of these fragments was ligated to the arm of pDGHisC that had been cleaved with the respective enzyme(s). The ligated DNA mixtures were used for the transformation of JM109 to ampicillin resistance to result in plasmids pIOLB-His and pIOJE-His, which carried the respective iol genes as C-terminal His 6 tag fusion placed under the control of the pDGHisC-borne T7 promoter. To express iolA as a C-terminal His 6 tag fusion, pIOLA-His was constructed similarly as above but using pET30a (Novagen) as the cloning vector. DNA fragments covering the iolA coding region and a short linker enabling the C-terminal His 6 tag fusion at the tail was amplified from DNA of strain 60015 by PCR using the specific primers, iolA-NdeI (GGA-ATTCCATATGATGGCAGAAATCAGAAAATTAAA, an NdeI site is indicated in bold) and iolA-NotI (ATAAGAATG-CGGCCGCGTTGAAATCAGGTGCCGGAT, an NotI site in bold). The PCR fragment was trimmed with NdeI and NotI and ligated to the arm of pET30a that had been cleaved with the same enzymes. The ligated DNA mixtures were used for the transformation of JM109 to kanamycin resistance to result in the plasmid pIOLA-His, which carried iolA as C-terminal His 6 tag fusion placed under the control of the pDGHisC-borne T7 promoter. Correct construction of the plasmids was confirmed by nucleotide sequencing. To express the His 6 tag fusion of iolB, iolJ, and iolA genes, pIOLB-His, pIOLJ-His, and pIOLA-His, respectively, were introduced into BL21(DE3) cells together with pG-KJE8. BL21(DE3) carrying both plasmids was inoculated into TGA medium containing ampicillin, arabinose, chloramphenicol, and tetracycline and grown at 37°C with shaking. At an optical density of the cultures of 0.35 at 600 nm, production of the C-terminal His 6 tag fusion proteins was induced for 2 h by the addition of IPTG. The cells were harvested, treated with lysozyme, and disrupted by brief sonication, and the soluble fraction was obtained after centrifugation. The His 6 tag fusion proteins were purified from the soluble fraction using the His-Bond system (Novagen), employing the standard procedure as recommended by the supplier. The purified proteins were subjected to Sephadex G-25 (Amersham Biosciences) gel filtration to remove imidazole and salts. The purity of the His 6 tag fusion proteins was verified by SDS-PAGE and subsequent immunoblot analyses using anti-His tag antibody (Nacalai Tesque). Enzyme Assays-THcHDO hydrolase activity was measured as follows. The reaction substrate, THcHDO, was prepared by the enzymatic conversion of 2KMI and purified as described (13). JM109 cells carrying pUC18, pIOLB, pIOLC, or pIOLD were grown with IPTG, and the protein extracts were prepared as described (13). The hydrolase assay was performed by the procedures modified from those previously reported (3). 0.98 ml of assay mixture was prepared to contain 50 mM Tris-Cl (pH 8.0), 0.1 mM glutathione, 0.05 mM CoCl 2 , and sufficient THcHDO (giving an absorbance of 0.6 at 260 nm; ϳ0.1 mM). After adding 20 l of the protein extract containing ϳ60 g of protein to the mixture, the rate of decrease in the absorbance at 260 nm was measured, and consumption of the substrate THcHDO was calculated from its estimated molar absorption coefficient of 6000 (3).
Phosphorylation triggered by the addition of 5-deoxy-glucuronic acid (5DG) was monitored as follows. The reaction substrate, 5DG, was prepared by the enzymatic conversion of 2KMI and purified as described below. JM109 cells carrying pUC18, pIOLB, pIOLC, or pIOLBC were grown with IPTG, and the protein extracts were prepared as described (13). B. subtilis cells were grown with and without inositol, and the extracts were prepared in the same way as described for the inositol dehydrogenase assay (1). Phosphorylation was monitored by the procedures modified from those reported previously (9). 0.9 ml of assay mixture was prepared to contain 50 mM Tris-Cl (pH 8.0), 2.5 mM MgCl 2 , 1.25 mM phosphoenolpyruvate, 0.5 mM ATP, 3.0 mM KCl, 10 mM 2-mercaptoethanol, 10 g/ml of pyruvate kinase, 5 g/ml of lactate dehydrogenase, 0.25 mM NADH, and 100 l of protein extract (ϳ200 -300 g of protein). 0.1 ml of 20 mM 5DG or H 2 O was added to the mixture, and the rate of decrease in the absorbance of NADH at 340 nm was measured.
IolJ aldolase activity was measured as follows. The substrate was the reaction product of the IolBC reaction prepared as described below. IolJ-His was purified as indicated above and used for the enzyme assay. The IolJ reaction was supposed to produce DHAP, and the production of DHAP was monitored by the procedures modified from those reported previously (9). The purified IolJ-His protein was serially diluted in test tubes, each containing the reaction mixture (100 mM Tris-Cl (pH 7.6), 3.3 mM DKGP, 1.43 mM NADH, and glycerol-1-phosphate dehydrogenase (0.1 unit/l)), and incubated at 24°C for 30 min. Each of the reaction mixtures was diluted 25 times immediately after the reaction to measure the absorbance of NADH at 340 nm.
The IolJ reaction was supposed to produce MSA besides DHAP, and MSA would be the substrate of IolA MSA dehydrogenase, which was shown to convert MSA to acetyl-CoA and CO 2 with the specific reduction of NAD ϩ to NADH (16). Thus, MSA was detected as follows by the procedures modified from those reported previously (9). The purified IolJ-His protein was serially diluted in test tubes, each containing the reaction mixture (100 mM Tris-Cl (pH 8.0), 3.3 mM DKGP, IolA-His (5.7 g), 3.0 mM CoA, and 3.0 mM NAD ϩ ), and incubated at 24°C for 60 min. Each of the reaction mixtures was diluted 25 times immediately after the reaction to measure the absorbance of NADH at 340 nm.
Purification and Structural Analysis of the Reaction Products-The product of the IolD reaction was synthesized and characterized as follows. JM109 cells carrying pIOLDE (13) were grown in 100 ml of LB containing ampicillin and IPTG for 16 h, harvested, suspended in 2.2 ml of water, treated with lysozyme, disrupted by sonication, and centrifuged to obtain the supernatant solution containing both the IolD and IolE enzymes. The enzyme solution (ϳ3 ml) was mixed with 250 mg of 2KMI in 9 ml of 33 mM phosphate buffer (pH 8.0) and incubated at 25°C, allowing the reaction to proceed. During the reaction the pH values were maintained at 8.0 by adding NaOH when required. After 20 h appropriate amounts of 6 N H 2 SO 4 were added to drop the pH down to 3.0, and the reaction was terminated. The mixture was heated at 85°C for 10 min to denature the proteins and then centrifuged for 10 min at 5000 rpm. The supernatant solution was saved, concentrated, and dried. The dried solute was dissolved in 20 ml of methanol/ethyl acetate (4:1, v/v), and the solution was passed through a column filled with 12 ml silica gel (Merck) to remove the remaining substrate and intermediates. The eluent was concentrated once again and dried to obtain 119 mg of pale yellow compound. The 1 H NMR spectrum of this compound was obtained in The IolBC reaction product was identified as follows. JM109 cells carrying pIOLBC (13) were grown in 100 ml of LB containing ampicillin and IPTG for 16 h, harvested, suspended in 2.2 ml of water, treated with lysozyme, disrupted by sonication, and centrifuged to obtain the supernatant solution containing both the IolB and IolC enzymes. The enzyme solution (ϳ3 ml) and 222 mg of 5DG were mixed to make 13 ml of reaction mixture (pH 8.0, adjusted with NaOH) containing 90 mM ATP, 3 mM KCl, and 3 mM MgCl 2 . The mixture was incubated at 26°C, allowing the reaction to proceed. 24 h later the reaction was terminated by the addition of HCl to lower the pH to 2-3. After being concentrated to 10 ml and mixed with 40 ml of methanol, the mixture was centrifuged to recover the supernatant solution. The solution was concentrated to 10 ml, neutralized with NaOH to adjust the pH to 7.0, and combined and mixed well with 2.5 ml of BaCl 2 . The mixture was centrifuged to obtain the white sediment, which was suspended in 10 ml of water, mixed with 10 ml of Duolite C20 resin (H ϩ type; Sumitomo Chemical), and stirred for 2 h at room temperature. The resin was removed by filtration and washed with 20 ml of water. All of the filtrate (ϳ30 ml) was centrifuged to obtain the clear supernatant. After adjusting its pH to 3.0 with HCl, the supernatant was concentrated and dried to obtain the solute. The dried solute was dissolved in 4 ml of methanol/water (3:1, v/v), and subjected to column chromatography through 25 ml of silica gel with a mobile eluent of methanol/water (9:1, v/v) to collect the eluent in fractions. The desired fractions were combined, concentrated, and then dried to obtain 60 mg of a pale brown compound. The 1 H NMR spectrum of this compound was obtained in D 2 O.
The IolB-His reaction product was characterized as follows. The purified IolB-His protein (ϳ150 g) and 10 mg of 5DG were combined in 0.14 ml of reaction mixture (pH 8.0, adjusted with NaOH). The reaction mixture was incubated at 26°C for 16 h. The reaction was terminated by the addition of H 2 SO 4 to lower the pH to 2-3 and centrifuged to save the clear supernatant solution, which was concentrated and dried to obtain the solute. The dried solute was suspended in 5 ml of chloroform/ methanol (1:1, v/v) and centrifuged to save the supernatant solution. The solution was passed through a column filled with 2 ml of silica gel, and the eluent was concentrated once again and dried to obtain 9 mg of pale yellow compound. The 1 H NMR spectrum of this compound was obtained in D 2 O.
Mass spectra of the IolD and IolBC reaction products were obtained by ESI-TOF/MS (Mariner Biospectrometry Work station; Applied Biosystems). The products were dissolved in 50% acetonitrile containing 1% triethylamine for anion analysis.
DNase I Footprinting Experiment-DNase I footprinting experiments were performed as described (11). The IolR pro-tein was produced in JM109 carrying pIOLR3, and the crude extract was used for the experiments as described (11). The probe DNA for the gel retardation and DNase I footprinting experiments was prepared by PCR amplification of a DNA stretch containing the iol promoter region using plasmid DNA of pPIOLf (11) as a template and a pair of specific primers of M4 and RV (Takara). The 5Ј-terminus of the RV primer had been labeled with a Megalabel kit and [␥-32 P]ATP before the PCR amplification. DNase I footprinting reactions were done in the presence and absence of intermediates of the MI catabolic pathway.

The Third
Step Reaction of the MI Catabolic Pathway Is Catalyzed by IolD Hydrolyzing THcHDO to 5DG-Previous studies revealed that IolB, IolC, IolD, IolE, and IolG were all required for the initial reaction steps of MI catabolism to produce an intermediate acting as the inducer in the cell to antagonize DNA binding of IolR to induce transcription of the iol operon (13). Because IolG and IolE catalyze the first and second reactions, respectively (10, 13), either IolB, IolC, IolD, or combinations of these must be required for the third reaction.
The substrate of the third reaction is THcHDO, which is the product of the IolE reaction. In A. aerogenes, the third reaction was demonstrated to be ring scission (hydrolysis) of the substrate to yield 2-deoxy-5-keto-D-gluconic acid (DKG, compound [7]) in Fig. 1) as the product (25). Assuming that the third reaction in B. subtilis could be similar to that in A. aerogenes, THcHDO hydrolyzing activity in the protein extracts of the E. coli cells expressing iolB, iolC, or iolD was measured ( Table  2). The results clearly indicated that only the extract prepared from cells expressing iolD contained a large amount of the THcHDO hydrolyzing activity. From these results, we concluded that iolD encoded the B. subtilis THcHDO hydrolase. The ESI-TOF mass spectrum of the isolated product gave a major peak of an anion at m/z 177.0484 (data not shown). It was likely that this anion could correspond to the molecular ion of [C 6 H 10 O 6 -H] ؊ , suggesting that it might be DKG. However, its 1 H NMR spectra revealed that the product of the B. subtilis enzyme reaction was not DKG but 5DG (compound [6]) in Fig.  1). Actually, the product was an ϳ1:1 mixture of compounds [12] and [13], ␣and ␤-anomers of the five-membered ring compounds (Fig. 2), indicating that these were produced after spontaneous formation of a hemiacetal linkage between the C1 aldehyde and the C4 hydroxyl groups of 5DG (compound [6]). Therefore, the IolD THcHDO hydrolase of B. subtilis could cleave the C2-C3 bond of THcHDO to yield 5DG (Fig. 1).
Kinase Activity Triggered by 5DG Addition Requires Both IolB and IolC-In A. aerogenes, the next step is phosphorylation of the product of the third reaction, DKG, and the reaction is mediated by a kinase requiring ATP (9). In B. subtilis, the product of the third step appeared to be 5DG, and we were not sure if such phosphorylation would occur on 5DG itself. Nevertheless, we tried to detect the possible kinase activity, which might be triggered by 5DG addition, in the protein extracts of the E. coli cells expressing iolB and/or iolC. In this detection system, the activity was judged spectrophotometrically as described (9) [2] [9] [10]  ; [11], acetyl-CoA. Carbon numbering is defined for MI, 2KMI, THcHDO, 5DG, DKG, and DKGP; for the former three the definition from previous studies is applied (3,25).
absorbance of NADH at 340 nm in the presence and absence of 5DG in the reaction mixture. The kinase reaction converts ATP to ADP, which will result in the oxidation of NADH through the coupling reactions involving pyruvate kinase and lactate dehydrogenase (9). Therefore, a higher kinase activity triggered by 5DG addition will result in a more pronounced decrease in absorbance at 340 nm. This was indeed observed. As shown (Table 3), every protein extract prepared from E. coli cells gave a decrease in the absorbance of NADH, and this did not depend on 5DG. However, only the extract of the cells with pIOLBC expressing both iolB and iolC exhibited an activity almost 2.5 times higher in the presence of 5DG than in its absence (Table  3). Extracts from B. subtilis cells were also prepared and subjected to the measurement of the kinase activity similarly as above. The activity found in the extracts of B. subtilis was lower than that in E. coli. When the cells were grown in the presence of MI, the wild-type strain 60015 was judged to contain the higher activity triggered by 5DG, whereas YF244, an iolR::cat mutant allowing constitutive expression of the iol operon, always exhibited the higher activities triggered by 5DG regardless of growth conditions. These results suggested that the kinase could be encoded within the iol operon and induced in the presence of MI (Table 3). On the other hand, none of the other three iolR::cat mutants, each of which had an additional defect either in iolB or iolC, exhibited an activity as high as YF244 did ( Table 3), suggesting that the mutations in iolB and iolC might diminish and abolish the kinase activity, respectively. Taken together, we concluded that both IolB and IolC were required for the kinase activity triggered by 5DG addition.
The phosphorylated product of the kinase reaction involving both IolB and IolC was extracted as a barium salt form and found to give a very complicated ESI-TOF mass spectrum, probably due to its lower purity as well as differential ionization (data not shown). However, its 1 H NMR spectra clearly revealed that the product was DKGP (compound [8]  IolD reaction products turned out to be an equimolar mixture of compound [12] and [13] (with carbon and proton numbering). These compounds are ␣and ␤-anomers of 5DG forming a five-membered ring made by a hemiacetal linkage between the C1 aldehyde and the C4 hydroxyl groups (Fig. 1). Summary tables of their 1 H NMR spectrum analysis (right) and defined structural formulas (left) are given.  The IolBC reaction product was identified as compound [8], DKGP (with carbon and proton numbering). A summary table of its 1 H NMR spectrum analysis (top) and defined structural formula (bottom) are given.  a Kinase activity triggered by 5DG addition was investigated in the protein extracts of the cells. Values shown are the rates of decrease in absorbance of NADH at 340 nm (⌬A 340 /min)/mg of protein in the presence of 5DG (ϩ5DG) and in its absence (Ϫ5DG). A higher kinase activity triggered by 5DG addition resulted in a more pronounced decrease in absorbance at 340 nm. Only a set of representative data is shown, but at least three independent measurements were repeated with similar results. b E. coli JM109 cells carrying the indicated plasmid were grown in TGA medium containing ampicillin and 1 mM IPTG. c B. subtilis cells were grown in S6 medium with (ϩ) or without (Ϫ) 10 mM MI as indicated.
NMR summary shown in Fig. 3), which turned out to be the same compound as in A. aerogenes (9). IolB Isomerizes 5DG to DKG, and IolC Phosphorylates DKG to DKGP-The kinase reaction involving both IolB and IolC produced DKGP from 5DG. On the other hand, a homology search revealed that IolB exhibited similarities to a number of 4-deoxy-L-threo-5-hexulose uronate isomerases, and its amino acid sequence contains the motif of 5-keto-4-deoxyuronate isomerase as suggested in the BSORF data base. These facts prompted us to test whether IolB might be an isomerase converting 5DG to DKG. IolB was produced in E. coli as a C-terminal His 6 tag fusion protein and purified as IolB-His. The purified IolB-His was incubated with 5DG, and the reaction product was obtained. Although its ESI-TOF mass analysis was not performed, its 1 H NMR spectrum was clear enough to indicate that it contained an equimolar mixture of 5DG and DKG (compound [7]) in Figs. 1 and 4; 1 H NMR summary shown in Fig. 4). These results allowed us to conclude that IolB encodes the isomerase responsible for the fourth-step reaction, thereby converting 5DG to DKG reversibly. The amino acid sequence of IolC exhibits significant similarity to fructo-kinases and in its amino acid sequence contains a carbohydrate kinase motif of the pfkB family as listed in the BSORF data base, suggesting that IolC could be a kinase phosphorylating the C6 hydroxyl group of DKG (Fig. 4). Thus, the conversion of 5DG to DKGP appeared to involve two successive reactions, namely the fourth reaction of the pathway catalyzed by IolB isomerizing 5DG to DKG and then the fifth reaction by IolC, thereby phosphorylating DKG to DKGP.

The Product of the IolC Reaction, DKGP, Antagonizes IolR
Binding to DNA-IolG, IolE, IolD, IolB, and IolC appeared to catalyze the initial five reaction steps of the MI catabolic pathway in this order (Fig. 1), and previously all of these proteins were also shown to be required for the production of an intermediate acting as the inducer in cell (13). These findings suggested that the product of the IolC reaction might be the inducer that antagonizes DNA binding of IolR. It is known that the IolR repressor binds to DNA containing the iol promoter region, thereby altering the DNase I digestion pattern (11). Therefore, each of the intermediates appearing in the catabolic pathway was tested for its ability to interfere with DNA binding of IolR (Fig. 5). None of the compounds MI, 2KMI, THcHDO, the mixture of ␣-and ␤-anomers of 5DG, or DKG altered the  IolB reaction product was identified as compound [7], DKG (with carbon and proton numbering). A summary table of its 1 H NMR spectrum analysis (top) and defined structural formula (bottom) are given. DNase I footprinting analysis of the interaction of IolR with the iol promoter region was performed as described in the text. DNase I footprints of the 5Ј-endlabeled coding strand of the probe DNA were prepared as described in the text. Lanes 1-10 contained 0.04 pmol of the 32 P-labeled probe DNA in the reaction mixture (50 l). Lanes 1-9 contained 4.0 g of the protein extract of JM109 cells bearing pIOLR3, whereas lane 10 contained that of the cells bearing pUC19. Lanes 1, 5, 9, and 10 contained no additional chemicals, whereas lanes 2, 3, 4, 6, 7, and 8 did contain MI, 2KMI, THcHDO, 5DG (ϳ1:1 mixture of ␣and ␤-anomers), DKG/5DG (ϳ1:1 mixture), and DKGP, respectively (each approximately at 10 mM). Lanes G, A, T, and C contained the products of the respective sequencing reactions performed with the same primers as used for the probe preparation. The enhanced cleavage sites and protected regions are indicated by arrowheads and vertical lines on the right side, and nucleotide numbers are shown on the left (ϩ1 is the transcription start nucleotide of the iol operon).
DNase I digestion pattern upon IolR binding (11), whereas only DKGP rendered the digestion pattern almost the same as that without IolR binding. The results clearly indicated that the reaction product of the IolC reaction, DKGP, antagonized IolR for its DNA binding, revealing DKGP as the intermediate that functions as the inducer in the cell.
IolJ Is an Aldolase Acting on DKGP to Produce DHAP and MSA-In A. aerogenes it was shown that an aldolase converted DKGP into two C3 compounds, DHAP (compound [9]) in Fig. 1) and MSA (compound [10]) in Fig. 1). The amino acid sequence of IolJ shares high similarities with those of many aldolases (1). IolJ was produced in E. coli as a C-terminal His 6 tag fusion protein and purified as IolJ-His. IolJ-His was incubated with DKGP to test its aldolase activity, employing the assay system previously described (9). The aldolase reaction would produce DHAP, which is known to serve as the specific substrate for glycerol-1-phosphate dehydrogenase involving the oxidation of NADH. Under the conditions we used, the more IolJ-His was contained in the reaction mixture, the more DHAP was produced (Fig. 6A), clearly indicating that IolJ-His had the aldolase activity producing DHAP from DKGP. When fructose bisphosphate was used as another substrate in the same system as above, no production of DHAP was observed (data not shown), suggesting that IolJ might not function as fructose-bisphosphate aldolase as annotated in databases.
The iolA gene product was previously characterized as MSA dehydrogenase (16), which acts on MSA to convert it into acetyl-CoA and CO 2 with reduction of NAD ϩ to NADH. Because MSA is very unstable and its authentic standard for NMR analysis is not available, production of MSA was also shown in the enzymatic way employing IolA. IolA was produced in E. coli and purified as a C-terminal His 6 tag fusion protein, IolA-His. When the IolJ aldolase reaction was performed in the presence of IolA-His together with CoA and NAD ϩ , the specific reduction of NAD ϩ to NADH was observed, suggesting that MSA produced by the aldolase reaction could serve as the substrate for the successive IolA reaction. The more IolJ-His was contained in the reaction mixture, the more MSA was produced (Fig. 6B), although the production was as low as 10% that of DHAP. Probably under the conditions we used, most of MSA was lost due to its instability and possible side reactions. Nevertheless the results clearly indicated that IolJ-His functioned to produce MSA from DKGP. We, thus, concluded that IolJ was the aldolase acting on DKGP, thereby converting it into DHAP and MSA. DHAP is a known glycolytic intermediate, and MSA is converted into acetyl-CoA and CO 2 with reduction of NAD ϩ by the MSA dehydrogenase of the iolA gene product.

DISCUSSION
The work described here together with previously published works revealed that MI catabolism of B. subtilis comprises seven stepwise reactions (Fig. 1). The first step is the dehydrogenation of MI to 2KMI, which is catalyzed by IolG, and the second step is the successive dehydration by IolE to give THcHDO. The third step is ring scission of THcHDO, catalyzed by IolD, to yield 5DG, and the fourth step is the isomer-ization of 5DG by IolB to produce DKG. The fifth step is the phosphorylation of DKG by IolC, yielding DKGP, and the sixth step is the aldolase reaction by IolJ, cleaving DKGP into DHAP and MSA. The last, seventh, step is catalyzed by IolA, thereby converting MSA to acetyl-CoA and CO 2 . The net result of the inositol catabolic pathway in B. subtilis is the conversion of one MI molecule to yield one DHAP, one acetyl-CoA, and one CO 2 molecule, with acquisition of two NADH molecules and the consumption of an ATP molecule.
In A. aerogenes, the ring scission of THcHDO, was initially reported to give 4-deoxy-5-ketohexonic acid as the product (25). This was reinvestigated 5 years later, however, with the result that the product should be DKG (26). This suggests IolJ-His (µg) FIGURE 6. Aldolase activity of IolJ-His producing DHAP (A) and MSA (B) from DKGP. A, DHAP production. The experiment was performed as described under "Experimental Procedures." Amounts of DHAP produced in the reaction tube within 30 min (nmol) were calculated from the decrease in the absorbance of NADH at 340 nm and plotted against the amounts of IolJ-His (g). B, MSA production. The experiment was performed as described under "Experimental Procedures." Amounts of MSA produced in the reaction tube within 60 min (nmol) were calculated from the increase in the absorbance of NADH at 340 nm and plotted against the amounts of IolJ-His (g). Only the sets of representative data are shown, but at least three independent measurements were repeated with similar results.