Mannitol-1-Phosphate Dehydrogenase (MtlD) Is Required for Mannitol and Glucitol Assimilation in Bacillus subtilis: Possible Cooperation of mtl and gut Operons

Abstract

We found that mannitol-1-phosphate dehydrogenase (MtlD), a component of the mannitol-specific phosphotransferase system, is required for glucitol assimilation in addition to GutR, GutB, and GutP in Bacillus subtilis. Northern hybridization of total RNA and microarray studies of RNA from cells cultured on glucose, mannitol, and glucitol indicated that mannitol as the sole carbon source induced hyperexpression of the mtl operon, whereas glucitol induced both mtl and gut operons. The B. subtilis mtl operon consists of mtlA (encoding enzyme IICBA^mt1) and mtlD, and its transcriptional regulator gene, mtlR, is located 14.4 kb downstream from the mtl operon on the chromosome. The mtlA, mtlD, and mtlR mutants disrupted by the introduction of the pMUTin derivatives MTLAd, MTLDd, and MTLRd, respectively, could not grow normally on either mannitol or glucitol. However, the growth of MTLAd on glucitol was enhanced by IPTG (isopropyl-β-d-thiogalactopyranoside). This mutant has an IPTG-inducible promoter (Pspac promoter) located in mtlA, and this site corresponds to the upstream region of mtlD. Insertion mutants of mtlD harboring the chloramphenicol resistance gene also could not grow on either mannitol or glucitol. In contrast, an insertion mutant of mtlA could grow on glucitol but not on mannitol in the presence or absence of IPTG. MtlR bound to the promoter region of the mtl operon but not to a DNA fragment containing the gut promoter region.

Bacteria express many phosphoenolpyruvate-dependent phosphotransferase systems (PTS) consisting of sugar-specific transport proteins and phosphorylation mechanisms to utilize various sugars. The PTS are primarily specific to each sugar, but they can transport other sugars to a lesser extent (4). Furthermore, their substrate specificity can be changed by the amino acid replacement(s) (1, 2). Genome sequence analysis of Bacillus subtilis (15) has identified 21 complete PTS, 17 of which were characterized or estimated via the corresponding sugars (21; see also the B. subtilis genetic databases BSORF [http://bacillus.genome.ad.jp/] and SubtiList [http://genolist.pasteur.fr/SubtiList/]). The operon for the mannitol-specific PTS is located at 38.5° of the B. subtilis chromosome. Mannitol-specific PTS have been investigated in Escherichia coli (16), in Bacillus stearothermophilus (8), and in other organisms (3, 10). The mtl operon in E. coli consists of the mtlA, mtlR, and mtlD genes that encode the mannitol transporter (enzyme IICBA^mtl), a transcriptional regulator, and mannitol-1-phosphate dehydrogenase, respectively, and that in B. stearothermophilus comprises four genes: mtlA (enzyme IICB^mtl), mtlR, mtlF (enzyme IIA^mtl), and mtlD. In contrast, the B. subtilis mtl operon consists of only the mtlA (enzyme IICBA^mtl) and mtlD genes, and the operon does not contain a gene for a transcriptional regulator. However, the gene ydaA, which was recently renamed mtlR and which is homologous to the mtlR genes of E. coli and B. stearothermophilus, is located 14.4 kb down stream from the mtl operon (Fig. 1A) (15). Therefore, MtlR should also regulate the mtl operon of B. subtilis, although it is positioned at a great distance from the mtl operon. However, the function of MtlR in B. subtilis has not yet been biochemically verified.

FIG. 1.

(A) Organization of genes relating to mannitol-specific PTS consisting of mtlA, mtlD, and mtlR and the gene structure of mtlA and mtlD in the disrupted mutants MTLAd and MTLDd and the insertion mutants ΔmtlA and ΔmtlD. (B) Gene organization of gutR, gutB, and gutP for the glucitol transport system. A “P” with an arrowhead and stem-loop structure indicates a putative promoter for σ^A showing the direction of transcription and of the ρ-independent transcriptional terminator. Numbers in parentheses are predicted amino acid lengths of each gene product. “Kb” indicates the distance from the replication origin of the B. subtilis chromosome. Inserted DNA regions in ΔmtlA and ΔmtlD mutants are also indicated (▪).

Glucitol is found in various fruits in nature and is used as a sweetener in food. B. subtilis can utilize it as a sole carbon source via the gut operon (glucitol-utilizing transporter), consisting of gutP encoding a probable transporter (originally reported as gutA) (7) and gutB (glucitol dehydrogenase), which is located at ca. 57° on the chromosome. A transcriptional activator for the gut operon, gutR, is located immediately upstream of gutB and is transcribed in the opposite direction relative to gutB (Fig. 1B) (15, 20). The genes for glucitol-specific PTS have been identified in E. coli (25), but no homologous genes for the PTS have been found in B. subtilis.

Functional cooperation between the mtl and gut operons has not been reported. However, in the genome project of B. subtilis to construct disrupted mutants of whole genes, we found that disrupted mutants of mtlA, mtlD, and mtlR could not grow on either mannitol or glucitol. Therefore, we analyzed the relationships between the mtl and gut operons with respect to the expression of each gene in the two operons under conditions of mannitol and glucitol induction and examined the requirements for glucitol assimilation.

Here, we found MtlR to be a transcriptional activator of the mtl operon of B. subtilis. The mtl operon and mtlR were overexpressed in cells cultured on mannitol and on glucitol, and MtlD, in addition to GutB and GutP, was required to assimilate glucitol. We discuss the possible cooperation between the mtl and gut operons to utilize glucitol through MtlD.

MATERIALS AND METHODS

Bacterial strains and strain construction.

All B. subtilis strains were derivatives of B. subtilis 168 (trpC2) and are listed in Table 1. The host for gene manipulation was E. coli JM109. Mutants of E. coli and B. subtilis were maintained and isolated in Luria-Bertani (LB) broth containing ampicillin (50 μg/ml), chloramphenicol (5 μg/ml), or erythromycin (0.3 μg/ml) when necessary. Synthetic S6 medium (6) was used to isolate RNA for DNA microarray analysis and to analyze growth curves of wild-type and mutant strains.

TABLE 1.

B. subtilis strains used in this study

Strain	Relevant genotype and descriptiona	Reference or source
168	trpC2	Laboratory stock
MTLAd	trpC2 mtlA::pMUTin(Emr)	This study
MTLDd	trpC2 mtlD::pMUTin(Emr)	This study
MTLRd	trpC2 mtlR::pMUTin(Emr)	This study
GUTBd	trpC2 gutB::pMUTin(Emr)	JAFANb (K. Kobayashi)
GUTPd	trpC2 gutP::pMUTin(Emr)	JAFAN (K. Kobayashi)
GUTRd	trpC2 gutR::pMUTin(Emr)	JAFAN (K. Kobayashi)
LEVDd	trpC2 levD::pMUTin(Emr)	This study
LEVGd	trpC2 levG::pMUTin(Emr)	This study
RBSCd	trpC2 rbsC::pMUTin(Emr)	This study
FRURd	trpC2 fruR::pMUTin(Emr)	JAFAN (K. Kobayashi)
ΔmtlA	trpC2 mtlA::small part of pMUTin(Ems)	This study
ΔmtlD	trpC2 mtlD::cat(Cmr)	This study

^aEm^r, erythromycin resistance; Em^s, erythromycin sensitivity; Cm^r, chloramphenicol resistance.

^bJAFAN, http://bacillus.genome.ad.jp.

Construction of mtlA, mtlD, and mtlR disrupted mutants.

The N-terminal 317-bp DNA fragment corresponding to amino acid positions 7 to 112 of MtlA, a 330-bp fragment corresponding to amino acid positions 8 and 118 of MtlD, and a 184-bp fragment corresponding to amino acid positions 28 to 91 of MtlR were amplified from the B. subtilis chromosome by using the following primer sets: MTLA-F1 (5′-GCGCAAGCTTCAAGGCGGCATGAAAGTGAA-3′) and MTLA-R1 (5′-GCCGAGATCTCCAAGCGGACCCATAATCAT-3′), MTLD-F1 (5′-GCGCAAGCTTGGGAAATATCGGGAGAGGAT-3′) and MTLD-R1 (5′-GCCGAGATCTCGCAGGCAATGATATTCAGT-3′), and MTLR-F1 (5′-GCGCAAGCTTGCTGATGCAAGTCAGCAC-3′) and MTLR-R1 (5′-GCCGAGATCTGCAGTTTCCGTTCGTCGG-3′).

These primers contain linker DNA for HindIII or BglII sites (linker regions are underlined). The PCR products were digested with the two enzymes, cloned into the similarly digested E. coli vector pMUTin, which was constructed during the gene disruption project of Japan and the European Union (23). The constructed pMUTin derivatives were transferred into B. subtilis 168 by competent cell transformation (24). The plasmids were integrated into mtlA, mtlD, or mtlR by homologous recombination and the disrupted mutants of mtlA, mtlD, and mtlR were named MTLAd, MTLDd, and MTLRd, respectively.

Construction of insertion mutants of mtlA and mtlD (ΔmtlA and ΔmtlD).

The ΔmtlA mutant is an erythromycin-sensitive derivative of MTLAd generated by deletion of the major part of pMUTin, resulting in 278 bp of the plasmid DNA being inserted at amino acid position 7 of MtlA. This insertion induced a frameshift at the C-terminal region. The ΔmtlD mutant was constructed by inserting a chloramphenicol-resistant gene into mtlD. The upstream and downstream regions, 736 and 680 bp, respectively, of mtlD were amplified from the B. subtilis chromosome by using primer pair ΔmtlD-F1(5′-GGAGCTTGACATCAGTGTGA-3′) and ΔmtlD-R1 (5′-CGTCGTGACTGGGAAAACTTCACATCCGCAAACACCAC-3′) and primer pair ΔmtlD-F2 (5′-GTTATCCGCTCACAATTCCATTTCGGACGATGTGACCC-3′) and ΔmtlD-R2 (5′-TAATGGAAAGCTCAAGCTCC-3′). A chloramphenicol-resistant gene derived from plasmid pCBB331, a derivative of pC194 (5) was connected between the upstream and downstream DNA fragments, transferred into B. subtilis 168 to obtain the ΔmtlD mutant.

Construction of levD-, levG-, and rbsC-disrupted mutants.

We constructed LEVDd, LEVGd, and RBSCd mutants by the strategy described above and with the primer pairs LEVD-F1 (5′-GCGCAAGCTTGTCATGGAGATTTTCCCATAGCATT-3′) and LEVD-R1 (5′-GCCGAGATCTCTTCCGGGATATCCTTGAGCGCCTG-3′), LEVG-F1 (5′-GCGCAAGCTTACGAAGAAAGAAATTTTCAGCATGT3′) and LEVG-R1 (5′-GCCGAGATCTTGGTCCCATCAAACCGATTTTCATT-3′), and RBSC-F1 (5′-GCGCAAGCTTAACAGAACAAAAACGGATTCACTTC-3′) and RBSC-R1 (5′-GCCGAGATCTCAGCCAAGGATGATCGCGAGAACCG-3′).

RNA preparation, cDNA synthesis, hybridization, and DNA microarray analysis.

Total RNA was extracted from B. subtilis 168 cells cultured in 100 ml of synthetic S6 medium containing 1% glucose, 1% glucitol, or 1% mannitol at 37°C and harvested at the mid-log phase of growth (A₆₆₀ = 0.3). We used a two-step method of RNA isolation and for preparing the fluorescent cDNA probes used for hybridization to DNA microarrays as described previously (19, 29). Complementary DNAs for total RNA were aminoallyl labeled by reverse transcription with specific primers (0.5 pmol for each of 4,050 genes) in the presence of aminoallyl-dUTP, followed by fluorescence labeling of the resultant aminoallylated cDNA with N-hydroxysuccinimide-activated Cy3 or Cy5. The single microarray plate used in the present study contained two grids, each containing 4,055 genes encoding proteins and 39 calf thymus DNA spots as negative controls. We omitted from the grids 45 genes that could not be amplified by PCR. Hybridization and microarray analyses proceeded as described previously (19, 29). The mean values of the Cy3 and Cy5 fluorescence intensity for each gene on the two grids were calculated after subtraction of the background values obtained by determining the average values of the intensity of the 39 calf thymus DNA spots, and the ratios of gene expression were determined. The positive and internal controls were human transferring receptor mRNA and complementary human transferrin receptor primers.

Northern hybridization.

Total RNA for Northern hybridization was extracted from B. subtilis 168 cultured in synthetic S6 medium containing 1% glucose, 1% glucitol, or 1% mannitol at the mid-log phase of growth as described by Igo and Losick (12). The efficiency of extracting total RNA from cells grown in glucitol was lower than that from cells grown in mannitol and glucose. Therefore, approximately twice the amount of cells grown in glucitol was required to extract the same amount of total RNA as from the other cells. Northern hybridization proceeded according to a modification of the method described by Sambrook et al. (22) as modified by Kakeshita et al. (13). Total RNA (10 μg) was resolved by electrophoresis on 1.5% agarose gels containing 2.2 M formaldehyde and transferred to GeneScreen Plus nylon membranes (NEN Research Products). Prehybridization and hybridization proceeded at 65°C in hybridization buffer (0.9 M NaCl plus 0.09 M sodium citrate, 2× Denhardt reagent, 0.1% sodium dodecyl sulfate, and 100 μg of salmon sperm DNA/ml). To prepare DNA probes with which to detect mtlA, mtlD, and mtlR transcripts, 1,103-, 1,072-, and 1,020-bp DNA fragments at the N-terminal encoding regions of mtlA, mtlD, and mtlR, respectively, were amplified by PCR with the synthetic oligonucleotide primer sets (MTLA-F2 and MTLA-R2, MTLD-F2 and MTLD-R2, and MTLR-F2 and MTLR-R2 as described in the legend to Fig. 2). The chromosomal DNA of B. subtilis 168 grown on LB medium was the PCR template. After purification by agarose gel electrophoresis, amplified DNA fragments were labeled with ³²P by using a random primer DNA labeling kit (Takara-Shuzo Co., Ltd., Kyoto, Japan) and used as hybridization probes.

FIG. 2.

Northern hybridization of transcripts from mtlA, mtlD, and mtlR genes of B. subtilis 168 grown on glucose (Glc), mannitol (Mtl), or glucitol (Gut). Total RNA (10 μg) from cells grown on each sole carbon source were dissolved by agarose gel electrophoresis, and transcripts were detected by using mtlA probe for mtlA (A), mtlD probe for mtlD (B), and mtlR probe for mtlR (C), which are indicated by black bars in the lower portions of the schematic drawings of mtlA, mtlD, and mtlR. We prepared 1,103-, 1,072-, and 1,020-bp probes for mtlA, mtlD, and mtlR, respectively, by amplifying regions near the NH₂-terminal coding regions of each gene by PCR with the following specific primer sets: MTLA-F2 (5′-ATGAAAGTGAAAGTGCAACGCTTT-3′) and MTLA-R2 (5′-TGCTTTGGCTTGTTCCGCCTCTAA-3′), MTLD-F2 (5′-GCCTTACATTTCGGTGCGGGAAAT-3′) and MTLD-R2 (5′-GTTCATGGGACTGAATGCCGCACA-3′), and MTLR-F2 (5′-TTAAAGCACTTATTATTACAAAAC-3′) and MTLR-R2 (5′-TATGCGGCTGACGGCCGGCTCCAA-3′). Fragments were purified and labeled with [α-³²P]dCTP by using a random labeling kit from Takara-Shuzo, Ltd. Total RNA was extracted from cells at the mid-log phase of growth on 1% glucose (lanes 1, 4, and 7), mannitol (lanes 2, 5, and 8), or glucitol (lanes 3, 6, and 9). Arrows under the genes indicate the direction and sizes of the transcripts in kilonucleotides.

Preparation of purified MtlR.

The mtlR gene was amplified by PCR by using the primer set MTLR-F3 (5′-CGCTGGATATCATGTATATGACTGCCAGAGAACAAAA-3′) and MTLR-R3 (5′-CCGGCTCGAGCCGGCAGTATGTTTTTTTCTTCATCCA-3′), in which the underlined 5′-end linker regions for EcoRV and XhoI, respectively, are extended, and the B. subtilis 168 chromosome is the template. The PCR products were digested with the two enzymes, purified by agarose gel electrophoresis, and cloned into a similarly digested derivative of E. coli vector pET29(b) (Novagen), encoding a His₆ tag at the C terminus. An E. coli M15 (Qiagen) transformant harboring the constructed plasmid was cultured to an A₆₆₀ of 0.3 at 37°C and induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 4 h at 37°C. Thereafter, MtlR was purified from the cell lysate by passage through a Ni²⁺-nitrilotriacetic acid resin (Qiagen) column. MtlR-His₆ bound to the resin was eluted with 100 mM NaH₂PO₄-10 mM Tris HCl (pH 6.3) containing 8 M urea and 0.2 M imidazole. The eluted protein was sequentially dialyzed against 20 mM Tris HCl (pH 7.5), 0.5 M NaCl, 50% glycerol, and 0.1 mM phenylmethylsulfonyl fluoride containing 6, 4, 2, and 0 M urea. MtlR migrated as a single band in sodium dodecyl sulfate-polyacrylamide gels (see Fig. 6), and densitometry showed that the purity was >90%.

FIG. 6.

Purification of B. subtilis MtlR from the cell lysate of E. coli transformant. MtlR tagged with His₆ was purified from cell lysates of transformants (2-liter culture at 37°C) after 4 h of induction by 1 mM IPTG. Electrophoretic profiles of cell lysates after induction by IPTG each time and purified MtlR preparations are shown after being stained with Coomassie brilliant blue. MW, molecular size markers (given in kilodaltons).

Gel shift assay.

Specific DNA fragments for gel shift assay generated by PCR by using primer sets and the B. subtilis chromosome were purified by agarose gel electrophoresis and labeled with ³²P by using a random primer DNA labeling kit (Takara-Shuzo Co.). The boundary for these fragments is specified below (see the legend to Fig. 7). The DNA-binding reactions proceeded in a total volume of 30 μl in binding buffer containing 40 mM Tris-HCl (pH 8.0), 40 mM EDTA (pH 8.0), 150 μg of bovine serum albumin/ml, 1 μg of poly(dI-dC)/μl, 1 mM dithiothreitol, a 5-nmol DNA probe, and 5 nmol of purified MtlR. The reaction mixtures were incubated at 37°C for 15 min and then loaded onto 5% polyacrylamide gels in 0.5× TBS buffer (45 mM Tris-borate [pH 8.0], 1 mM EDTA). Samples were separated at 175V for 3 h and dried for autoradiography.

FIG. 7.

Gel mobility shift assay with purified MtlR. (A) Binding between MtlR and ³²P-labeled DNA fragment containing promoter and upstream region of mtlA and competition for the binding by nonlabeled fragment. A 101-bp DNA fragment containing the promoter and upstream region of mtlA was amplified by PCR with the primer pair MTLA-F3 (5′-AGGGACTGTAAGCGTTTTAACATAG-3′) and MTLA-R3 (5′-ATATAAACCCTCCCTGTTTTGTTTG-3′) and chromosomal DNA from 168 grown in LB medium. Lane 1 shows MtlR binding to a 101-bp DNA fragment without competitor. Lanes 2, 3, and 4 show competition for binding by 5, 50, and 500 nmol of nonlabeled competitor, respectively. (B) Binding between MtlR and a ³²P-labeled mtlA DNA fragment was not inhibited by nonlabeled 225-bp DNA fragment containing the promoter and upstream region of gutB and gutR. The 225-bp DNA fragment was amplified by using the primer pair GUTB-F (5′-CAAGTTCAGCCATAAGCCTGCACCT-3′) and GUTB-R (5′-GTGTGAGTCATTTGGCAAGTTCCTT-3′) and chromosomal DNA. Lane 5 shows the absence of competitor DNA. Lanes 6, 7, and 8 show that 5, 50, and 500 nmol of 225-bp DNA did not compete for binding. (C to E) MtlR did not bind to ³²P-labeled DNA fragments containing promoters and upstream regions of gut (lanes 9, 10, 11, and 12) (C), lev (lanes 13, 14, and 15) (D), and rbs (lanes 16, 17, and 18) (E) operons. Lanes 9, 13, and 16, absence of MtlR; lanes 10, 14, and 17, 5 nmol of MtlR; lane 11, 20 nmol of MtlR; and lanes 12, 15, and 18, 50 nmol of MtlR. Each reaction was mixed with ³²P-labeled 225-bp (5 nmol) fragments of the gut promoter region or ³²P-labeled 101-bp fragments (5 nmol each) of the promoters and upstream regions of lev or rbs. Two 101-bp fragments were amplified by using the primer set LEV-F (5′-ATCTATTGCTCCTTTCCTGT-3′) and LEV-R (5′-TGCTATTGGCTGAAATAACA-3′) or the primer set RBS-F (5′-CTGCTTTTGGGTATCATTAAAAAAC-3′) and RBS-R (5′-CTTAATCTTCCTTTCTTGTCGTCTT-3′), as well as chromosomal DNA. “C” and “DF” indicate the DNA-protein mobility shift complex and probe DNA, respectively.

RESULTS

Global analysis of upregulated genes in cells cultured on mannitol and glucitol as sole carbon sources.

To identify gene candidates involved in mannitol and glucitol utilization, total RNA was extracted from cells cultured on S6 medium containing glucose, mannitol, or glucitol and analyzed by using the DNA microarray. cDNA generated from RNA from cells cultured on glucose was labeled with Cy3, and cDNA generated from RNA from cells cultured on mannitol and glucitol was labeled with Cy5. Table 2 summarizes the Cy5/Cy3 ratio findings, indicating highly activated genes involved in the utilization of mannitol and glucitol.

TABLE 2.

Microarray analysis of activated genes in cells grown on mannitol and glucitol^a

Gene	Resultsb obtained with cells grown on:						Functionc	Description
	Mannitol			Glucitol
	Cy3_net	Cy5_net	Cy5/Cy3	Cy3_net	Cy5_net	Cy5/Cy3
mtlA	320.37	39,966.54	124.75	795.02	4,945.02	6.22	1.2.0	PTS mannitol-specific enzyme IIABC component
mtlD	85.41	2,400.19	28.1	88.24	330.03	3.74	2.1.1	Mannitol-1-phosphate dehydrogenase
mtlR	897.05	7,512.82	8.38	894.81	1,780.67	1.99	3.5.2	Transcriptional regulator of mannitol operon (mtlA/D)
levD	530.52	3,505.74	6.61	671.02	1,085.04	1.62	1.2.0	PTS fructose-specific enzyme IIA component
levG	915.35	5,923.68	6.45	1,422.37	1,865.44	1.31	1.2.0	PTS fructose-specific enzyme IID component
rbsC	328.67	2,087.09	6.35	438.29	2,191.79	5	1.2.0	Ribose ABC transporter (permease)
gutB	284.91	306.01	1.07	285.05	24,918.7	87.42	2.1.1	Glucitol dehydrogenase
gutP	690.59	863.82	1.25	1,527.2	10,369.8	15.96	1.2.0	Probable glucitol transporter
gutR	1,234.73	1,730.37	1.4	1539	2800.94	1.82	3.5.2	Transcriptional activator of the glucitol dehydrogenase
fruB	249.05	435.6	1.75	299.05	8,847.14	29.58	2.1.1	Fructose-1-kinase
fruR	287.05	443.25	1.54	230.4	6,239.74	27.08	3.5.2	Transcriptional repressor of the fructose operon
fruA	325.07	420.97	1.3	239.5	4,967.21	20.74	1.2.0	PTS fructose-specific enzyme IIBC component

^aThe gene names, functions and descriptions are basically as in the literature (15) and SubtiList (http://genolist.pasteur.fr/SubtiList).

^bThe activation ratios (Cy5/Cy3) in mannitol-grown cells or in glucitol-grown cells were calculated by dividing the spot intensities for cells grown on mannitol or glucitol (Cy5_net) by those for glucose (Cy3_net). The genes activated >6-fold versus glucose-grown cells are in boldface in the relating genes in the assimilation of mannitol and glucitol.

^cFunctional classification in SubtiList

A comparison of genes activated by mannitol from cells grown in mannitol or glucose showed that the Cy5/Cy3 ratios of the genes mtlA, mtlD, mtlR, levD, levG, and rbsC were >6.0. The activation ratios of mtlA and mtlD were 124.75 and 28.1, respectively. In contrast, the expression of gutR, gutB, and gutP that relate to glucitol assimilation was not stimulated in cells cultured in mannitol. The Cy5/Cy3 ratios were 1.07, 1.25, and 1.40, respectively. Therefore, the mtl but not the gut operon was highly activated when mannitol was the sole carbon source. The ratios of about 50 genes, including ptsG encoding the PTS glucose-specific enzyme II component, cotJC encoding the polypeptide component of the spore coat and other components, were <0.2.

On the other hand, the ratios of more than 70 genes were >6.0 in a comparison of cells grown in glucitol or glucose. The ratio was the highest for gutB (87.42), followed by fruB (29.58), fruR (27.08), fruA (20.74), gutP (15.96), and cotY (14.20). However, the ratio of gutR was 1.82. The ratio of mtlA was 6.22, and the ratios of mtlD and mtlR were 3.74 and 1.99, respectively. These findings indicate that the mtl operon was also activated by glucitol. Among the 70 genes activated by glucitol, 21 encoded proteins related to the spore coat. The ratios of nine genes, including ptsG, yvgL (which encodes an unknown protein that is similar to molybdate-binding protein), and others, were <0.2. The lev and fru operons were activated in cells grown in mannitol and glucitol, respectively (Table 2), although both operons encode fructose-specific PTS. These DNA microarray data are available online (http://www.genome.ad.jp/kegg/expression).

Northern blot analysis of the mtl operon.

To confirm the results of the DNA microarray experiments, we analyzed the expression of mtlA, mtlD, mtlR, levD, levG, and rbsC in cells cultured on 1% glucose, 1% mannitol, and 1% glucitol by Northern hybridization (Fig. 2). We detected clear 3.0-kb bands in RNA preparations from cells cultured in mannitol and glucitol by using the mtlA and mtlD probes (³²P-labeled 1,103- and 1,072-bp DNA fragments), respectively, as shown in Fig. 2A and B. The mtlR probe, which was a ³²P-labeled 1,020-bp DNA fragment, detected another clear 2.1-kb band of mtlR (Fig. 2C). These results indicated that mtlA and mtlD form an operon, that they are transcribed as a single mRNA in cells grown on both mannitol and glucitol, and that mtlR is transcribed as a single transcript. The DNA microarray experiments showed that the Cy5/Cy3 ratios of mtlA, mtlD, and mtlR in cells grown in mannitol and glucose (Table 2) were 124.75, 28.10, and 8.38, respectively. The ratios of mtlA, mtlD, and mtlR in cells grown in glucitol and in glucose were 6.22, 3.94, and 1.99, respectively. Therefore, expression of the three genes in cells grown in mannitol seemed much higher than that in glucitol. However, the intensity of the 3.0-kb bands detected by the mtlA and mtlD probe and of the 2.1-kb band detected by the mtlR probe in the RNA preparation from cells in glucitol was similar or rather higher than that of each corresponding band in RNA preparations from cells in mannitol (Fig. 2A, lanes 2 and 3, B, lanes 5 and 6, and C, lanes 8 and 9). The ratios of the intensity of bands between cells grown in mannitol and glucitol cells with the mtlA, mtlD, and mtlR probes were 1:1.3, 1:0.9, and 1:1, respectively, according to a densitometric analysis. In contrast, no bands were positive in the RNA preparation from cells grown on 1% glucose (Fig. 2, lanes 1, 4, and 7). The expression of the mtl operon will be inhibited by the catabolite repression of glucose because the cre consensus sequence (5′-WTGNAARCGNWWWCA-3′, where W, R, and N are A or T, G or A, and A, C, G, or T, respectively) (18) was identified at the promoter region of mtlA (5′-CTGTAAGCGTTTTAA-3′), as well as near the translation initiation site of mtlD (5′-ATGTGAACGAAACGA-3′).

On the other hand, Northern hybridization with each gene probe did not detect the expression of levD, levG, and rbsC in cells grown on glucose, mannitol, and glucitol (data not shown). Therefore, we concluded that these three genes were not expressed under these conditions. The apparently high expression of levD, levG, and rbsC according to the DNA microarray may be due to other reasons.

Figure 2 shows that the mtlA/D and mtlR mRNA was consistently smaller under conditions of glucitol rather than mannitol induction. We performed five Northern hybridization analyses to confirm this and obtained the same results each time. The reason for this is discussed below.

Northern hybridization of the gut operon.

Total RNA isolated from cells grown on mannitol and glucitol was Northern blotted to detect the expression levels of gutR, gutB, and gutP. A 2.8-kb band was detected by using the ³²P-labeled gutR probe, and a 2.5-kb band was detected by using the probes for gutP and gutB in the RNA prepared from cells grown in glucitol as shown by Ye et al. (27, 28). In contrast, no band was positive in cells grown on mannitol (data not shown). These results confirmed the findings of the DNA microarray analysis, indicating that the gutR, gutA, and gutB genes are expressed in cells when grown on glucitol but not when grown on mannitol.

Growth of gene-disrupted mutants on mannitol and glucitol.

To analyze the requirements for glucitol and mannitol assimilation, the mutants MTLAd, MTLDd, MTLRd, GUTRd, GUTBd, GUTPd, and others were cultured on glucose, mannitol, and glucitol. Figure 3 summarizes the growth curves of wild-type, MTLAd, MTLDd, and MTLRd on glucose, mannitol, and glucitol. Wild-type B. subtilis 168 can grow on glucose, mannitol, or glucitol as single carbon sources. However, MTLAd, MTLDd, and MTLRd could not grow on mannitol. Furthermore, the growth of MTLAd and MTLRd on glucitol was also extremely repressed, and MTLDd could not grow on glucitol. None of the erythromycin-resistant transformants of 168 mutated by MTLDd chromosomal DNA could grow on mannitol or glucitol.

FIG. 3.

Growth curves of B. subtilis 168 and the disrupted mutants MTLAd, MTLDd, and MTLRd, on glucose (A), mannitol (B), or glucitol (C). Overnight cultures of 168, MTLAd, MTLDd, and MTLRd in S6 minimal medium containing 1% glucose at 37°C were inoculated into fresh medium containing 1% glucose, 1% mannitol, or 1% glucitol with or without IPTG and then cultured at 37°C. Cell growth in each environment was monitored by absorbance at 660 nm. Symbols: •, strain 168 without IPTG; ○, MTLAd without IPTG; ▴, MTLAd with IPTG; ▵, MTLDd with or without IPTG; ▪, MTLRd without IPTG. The growth of each strain was tested more than four times, and the results were similar each time.

To analyze which gene product of mtlA or mtlD is required to utilize glucitol, we prepared the insertion mutants ΔmtlA and ΔmtlD and compared their growth with that of the wild type on glucose, mannitol, and glucitol (Fig. 4). The ΔmtlA mutant could grow as well as the wild type on glucose and glucitol but not on mannitol. In contrast, the ΔmtlD mutant could not grow on either mannitol or glucitol. We isolated five independent chloramphenicol-resistant mutants under the same conditions and found that none of them could grow on mannitol or glucitol.

FIG. 4.

Growth curves of B. subtilis 168 (A) and its insertion mutants, ΔmtlA (B) and ΔmtlD (C), on 1% glucose (•), 1% mannitol (▴), and 1% glucitol (○) at 37°C. Overnight cultures of 168 and both mutants in synthetic S6 medium containing 1% glucose were inoculated into fresh S6 medium containing 1% glucose, 1% mannitol, or 1% glucitol, and growth was monitored by determining the A₆₆₀.

Figure 5 shows the growth curves of wild type, GUTRd, GUTBd, and GUTPd on mannitol and glucitol. The three mutants can grow similarly on mannitol and glucose, but only GUTPd could grow on glucitol, although in a delayed manner. We also tested the other mutants carrying the disrupted genes levD, levG, rbsC, and fruR, which were activated in the DNA microarray studies. The four mutants proliferated on mannitol and glucitol as normally as the wild type. These results indicated that the mtl and gut operons are essential for mannitol and glucitol utilization, respectively, and that the mtlD gene product of the mtl operon is required to utilize glucitol in addition to the gut operon.

FIG. 5.

Growth curves of B. subtilis 168 and the disrupted mutants GUTRd, GUTBd, and GUTPd on mannitol (A) and glucitol (B). Overnight cultures of 168, GUTBd, GUTPd, and GUTRd in synthetic S6 medium containing 1% glucose were inoculated into S6 medium containing 1% mannitol or 1% glucitol with or without IPTG and cultured at 37°C. Growth was monitored as the absorbance at 660 nm. Symbols: •, strain 168 without IPTG; ▵, GUTBd without IPTG; ▴, GUTBd with IPTG; ○, GUTPd without IPTG; ▪, GUTRd without IPTG.

Effect of IPTG on growth of mutants on glucitol.

To confirm that MtlD is required to utilize glucitol, we analyzed the effect of IPTG on the growth of MTLAd. Figure 1 shows that the mtlA gene of the MTLAd mutant was disrupted by integration of the pMUTin derivative at the N-terminal coding region. This integration brought about localization of the IPTG-inducible spac-1 promoter (Pspac) (26) immediately upstream of the resultant truncated mtlA gene (′mtlA). Therefore, mtlD in the MTLAd mutant can be expressed by the addition of IPTG. The growth of MTLAd on glucitol was stimulated by 1 mM IPTG (Fig. 3C). In contrast, 1 mM IPTG did not stimulate the growth of MTLDd, which consists of integrated pMUTin at mtlD on glucitol, of MTLAd and MTLDd on mannitol, and of GUTBd and GUTPd on glucitol. These results indicated that MtlD is required for growth on glucitol. Since the growth rates of MTLAd and the wild type on glucitol in the presence of IPTG were not identical, expression of MtlD by the spac-1 promoter is not sufficient to support normal cell growth, or the mRNA from the promoter may be unstable.

Function of MtlR, a mannitol regulator of the mtl operon.

The function of GutR in expression of the gut operon has been identified as an activator (20). In contrast, the function of MtlR in B. subtilis has not yet been clarified because mtlR is located far from the mtl operon. We purified MtlR tagged with His₆ at the C terminus of the E. coli lysate. The molecular mass of the purified MtlR-tagged His₆ was 7.9 kDa (Fig. 6). A ³²P-labeled 101-bp DNA fragment containing the upstream region of mtlA and its translation initiation site was mixed with the purified MtlR protein, incubated, and analyzed by gel mobility shift assays (Fig. 7). A shift band is obvious in the mixture of the DNA fragment and MtlR on the autoradiograph (Fig. 7A, lane 1). A nonlabeled probe DNA fragment inhibited formation of the shift band (Fig. 7A, lanes 2, 3, and 4), and the band was completely abrogated by a 10-fold concentration of competitor DNA. In contrast, formation of the shift band was not competed by a 100-fold excess of another DNA fragment containing the gut promoter and the upstream region (Fig. 7B, lanes 5 to 8). We also analyzed the binding of MtlR to the DNA fragments containing the promoter regions of the gut, lev, and rbs operons. A gel mobility shift band was not evident in the DNA fragments containing the promoter regions of these operons (Fig. 7C, D, and E, lanes 9 to 18). These results indicated that MtlR specifically binds to the DNA fragment containing the promoter and the upstream region of the mtl operon. Since the lesion in mtlR prevents the expression of mtlA and of mtlD, MtlR will be an activator of the mtl operon. The amino acid identity between the two MtlR from B. subtilis and B. stearothermophilus is 40.34%, and they are similar in molecular size. At positions −85 to −59 bp from the transcription initiation site (+1) of B. subtilis mtlA, a 37-nucleotide sequence, 5′-TATTTTTAAAAAATTGTCACAGTATGTGCCAAAAGTC-3′, has 67.6% identity with the sequence 5′-TATTTTTAAAAAATGACAATAATGCATTGGACAAAATG-3′ of the MtlR-binding region in the B. stearothermophilus mtlA promoter region that was determined by footprinting (9). Therefore, MtlR might also bind at this upstream region to regulate the mtl promoter.

DISCUSSION

We found that the MtlD of mannitol-specific PTS is required to utilize glucitol as the sole carbon source. GutB and GutP that are independent of PTS were thought to transport glucitol into cells. The gene-disrupted mutants GUTBd, GUTPd, and GUTRd could grow on mannitol but not on glucitol. Glucitol-specific PTS has not been identified in B. subtilis, although its genome was sequenced in 1997 (15), and each gene has been functionally analyzed (14; see also the SubtiList Web Server and BSORF). On the contrary, MTLAd, MTLDd, and MTLRd could not grow in mannitol. Furthermore, growth of the mutants was repressed or prevented in glucitol. We concluded that MtlD is required to utilize glucitol for the following reasons. (i) The growth of MTLAd, MTLDd, and MTLRd was absent or repressed on glucitol. The lack or the repressed expression of mtlD caused the repression of MTLAd and MTLRd growth. (ii) The growth of MTLAd in glucitol was stimulated by 1 mM IPTG, by which mtlD expression was induced by the spac-1 promoter that is located immediately upstream of the truncated mtlA. (iii) Another mutant, ΔmtlA, carrying a small insertion at the N-terminal coding region of mtlA, could grow in glucitol like the wild type but not in mannitol. In contrast, the insertion mutant in mtlD, ΔmtlD, could not grow in either mannitol or glucitol. Moreover, Northern hybridization and DNA microarray studies revealed higher expression levels of mtlA, mtlD, and mtlR in mannitol and glucitol. These results agree with an old experiment reported by Horwitz and Kaplan (11). B. subtilis 168 synthesized 70- and 20-fold-increased levels of MtlD in the presence of mannitol and glucitol, respectively, than in glucose-grown cells. In contrast, only glucitol was capable of stimulating the production of glucitol dehydrogenase.

Figure 2 shows that the electrophoretic migration of mtlA/D and mtlR mRNA was consistently faster when isolated from cells grown in glucitol than when isolated from cells grown in mannitol. We repeated this Northern hybridization five times, and the results were identical each time. The extraction of RNA was less efficient from cells grown in glucitol than in mannitol and glucose, requiring twice the amount of cells to extract the same amount of RNA. Therefore, the RNA preparation from cells grown in glucitol might be rich in contaminants such as polysaccharides, which may have caused the faster RNA migration in agarose gel electrophoresis. In the DNA microarray analysis, the induction levels of genes relating to polysaccharide synthesis, such as sasA encoding spore coat polysaccharide synthesis protein, ytcC encoding lipopolysaccharide N-acetylglucosamine transferase, and yodU encoding capsular polysaccharide biosynthesis protein, and of other cells grown in glucitol were 10- to 30-fold higher than that of cells grown in mannitol. This low extraction efficiency from cells grown in glucitol might also have caused the discrepancy between the results of the Northern hybridization and DNA microarray analyses of mtlA, mtlD, and mtlR mRNA. However, it is possible that mtlD is expressed by another promoter locating in mtlA when cells are grown on glucitol. The mtlA induction levels in mannitol and glucitol were 125- and 6-fold higher, respectively, than that of cells grown in glucose in the DNA microarray analysis, whereas Northern hybridization revealed a similar intensity of mtlA/D bands between the two conditions. The reason for the difference between the two methods has not yet been determined.

Mannitol that is transported through the mannitol-specific PTS is phosphorylated for further metabolism. In contrast, glucitol is transported through the cooperation of GutB, GutP, and MtlD. The probable function of MtlD may be related to conversion to fructose or other sugars for further metabolism, although we have not yet determined the role of MtlD in the transport and utilization of glucitol in B. subtilis cells. In E. coli, glucitol is specifically transported and phosphorylated through the PTS system (17). Therefore, the metabolic pathways for glucitol in E. coli and B. subtilis will be different from one another. The molecular mechanism involved in the cooperation of GutB, GutP, and MtlD in glucitol transport is currently under investigation.

The DNA microarray studies revealed higher levels of lev and fru operon expression for fructose-specific PTS in cells grown in mannitol and glucitol, respectively. A small amount of glucitol and mannitol in the culture media might have been naturally converted into fructose or via an enzymatic reaction(s), although thin-layer chromatography did not demonstrate fructose in the culture medium. Such converted fructose may induce the lev or the fru operon to transport it into the cells, although Northern hybridization could not detect levD and levG transcripts. However, the growth rates of LEVDd, LEVGd, and FRURd, as well as of 168, on mannitol and glucitol were identical. Therefore, the lev and fru operons are not directly required to utilize the two sugar alcohols in B. subtilis.

In addition, we determined that MtlR is an activator of the mtl operon, although mtlR is located 14.4 kb downstream from the operon. Almost all of the genes for the regulator of mannitol-specific PTS in bacteria, except for B. subtilis, are included in the mtl operon as it is in E. coli, B. stearothermophilus, and Streptococcus mutans, and MtlR is essential for expression of the mtl operons. To confer an advantage in the proliferation of B. subtilis, the location of mtlR at a distance from the mtl operon may have been acquired by chromosomal rearrangement during evolution.