The Mannitol Utilization System of the Marine Bacterium Zobellia galactanivorans

Abstract

Mannitol is a polyol that occurs in a wide range of living organisms, where it fulfills different physiological roles. In particular, mannitol can account for as much as 20 to 30% of the dry weight of brown algae and is likely to be an important source of carbon for marine heterotrophic bacteria. Zobellia galactanivorans (Flavobacteriia) is a model for the study of pathways involved in the degradation of seaweed carbohydrates. Annotation of its genome revealed the presence of genes potentially involved in mannitol catabolism, and we describe here the biochemical characterization of a recombinant mannitol-2-dehydrogenase (M2DH) and a fructokinase (FK). Among the observations, the M2DH of Z. galactanivorans was active as a monomer, did not require metal ions for catalysis, and featured a narrow substrate specificity. The FK characterized was active on fructose and mannose in the presence of a monocation, preferentially K⁺. Furthermore, the genes coding for these two proteins were adjacent in the genome and were located directly downstream of three loci likely to encode an ATP binding cassette (ABC) transporter complex, suggesting organization into an operon. Gene expression analysis supported this hypothesis and showed the induction of these five genes after culture of Z. galactanivorans in the presence of mannitol as the sole source of carbon. This operon for mannitol catabolism was identified in only 6 genomes of Flavobacteriaceae among the 76 publicly available at the time of the analysis. It is not conserved in all Bacteroidetes; some species contain a predicted mannitol permease instead of a putative ABC transporter complex upstream of M2DH and FK ortholog genes.

INTRODUCTION

Brown algae (Phaeophyceae) are the dominant macroalgae in temperate and polar regions and thus play a crucial role in the primary production of coastal ecosystems (1). They contain large amounts of different structural and storage carbohydrates. For instance, their extracellular matrices are formed by the accumulation of cellulose, fucanes, and alginates (2–4), while they store carbon by accumulating laminarin and mannitol (5). The potential of this biomass resource for the production of liquid biofuels (6), including ethanol from alginate and mannitol (7, 8), and for the implementation of biorefineries (9) has been highlighted recently.

Depending on the species, mannitol can represent as much as 20 to 30% of the dry weight of brown seaweed (10). In the genomic and genetic model of brown algae Ectocarpus sp., formerly included in the species Ectocarpus siliculosus (11), it has been observed that the content of this polyol differs according to the diurnal cycle (12) and that it is likely to act as an osmoprotectant or a local compatible osmolyte (13). Mannitol is localized in the cytosol and is also present at the reducing ends of vacuolar laminarin molecules of the M series (in contrast to the G series, which contain only glucose residues) (14). Mannitol in brown algae is produced directly from the photoassimilate fructose-6-phosphate (F6P) by two steps: F6P is first reduced by mannitol-1-phosphate dehydrogenase (M1PDH) into mannitol-1-phosphate (M1P), which is then converted into mannitol by mannitol-1-phosphatase (M1Pase) (5, 15, 16). Mannitol is thought to be recycled by the successive actions of a mannitol-2 dehydrogenase (M2DH) and a hexokinase (HK) (5, 17), but little is known about the functioning of these enzymes in brown algae (18).

In contrast, mannitol metabolism has been intensively studied in Escherichia coli (19) and other bacteria, such as Bacillus subtilis (20, 21), Bacillus stearothermophilus (22), Clostridium acetobutylicum (23), and Streptococcus mutans (24). In these terrestrial bacteria, mannitol is taken up by a mannitol-specific phosphoenolpyruvate/carbohydrate phosphotransferase system (PTS) and is phosphorylated into M1P during its transport, and M1P is further oxidized to F6P by a M1P-specific dehydrogenase (25) before entering glycolysis. In the soil bacterium Acinetobacter baylyi (26), the M1P dehydrogenase is fused to a haloacid dehalogenase (HAD)-like phosphatase domain at the N terminus that was shown to catalyze M1Pase activity (27). In C. acetobutylicum (23) and B. stearothermophilus (22), the mannitol catabolic operon is regulated by two mechanisms: a glucose-mediated catabolite repression and a transcriptional activation mechanism controlled by MtlR using mannitol as an inducer. Other bacteria, such as Pseudomonas fluorescens, are known to contain a mannitol-2-dehydrogenase, an enzyme that oxidizes mannitol into fructose; mannitol is transported by ATP binding cassette (ABC) transporters (28). In this organism, fructose is believed to be phosphorylated into fructose-6-phosphate by a kinase coded for by mtlZ, a gene of the mannitol catabolic operon (28, 29). In the same way, Phaeobacter inhibens DSM 17395 imports mannitol via a specific ABC transporter whose corresponding genes are located next to the mtlK gene, coding for the mannitol-2-dehydrogenase; furthermore, the frk gene, encoding a fructokinase, colocalizes with genes corresponding to another ABC transporter (30).

Among the bacteria known to interact with algae, Flavobacteriia, and notably several species of Zobellia, have been found in association with macroalgae (31–33) and have been isolated from phytoplankton (34) and from seawater (35). The interactions between Zobellia strains and marine algae range from symbiosis, with some strains inducing the normal differentiation of green macroalgae (36), to algicidal behavior toward dinoflagellate blooms (34). The type species of the Zobellia genus, Zobellia galactanivorans (formerly known as “Cytophaga drobachiensis”), was initially isolated in Roscoff, France, from the red seaweed Delesseria sanguinea for its capacity to degrade carrageenans (31). This flavobacterium has been pivotal for the discovery and characterization of enzymes involved in the catabolism of red algal polysaccharides: kappa-carrageenase (37), iota-carrageenases (38), beta-agarases (39), beta-porphyranases (40), and 3,6-anhydro-l-galactosidases (41, 42). Based on these results, Zobellia galactanivorans is attracting interest as a model for the study of the bioconversion of macroalgal polysaccharides. Furthermore, although this bacterium has been isolated on a red alga, it can metabolize cell wall and storage polysaccharides from brown algae. Indeed, it possesses two alginolytic operons induced by the presence of alginate (43), and the first two alginate lyases of this complex system (AlyA1 and AlyA5) have recently been characterized at the biochemical and structural levels (44). This bacterium can also grow with brown algal laminarin as the sole carbon source, and among the five putative laminarinases identified in its genome (GenBank accession number FP476056), two GH16 laminarinases, Z. galactanivorans LamA (ZgLamA) and ZgLamC, have recently been analyzed biochemically and structurally (45). In addition to alginate and laminarin, mannitol is one of the most abundant carbohydrates in brown algae, and the annotation of the Z. galactanivorans genome has suggested the presence of proteins potentially involved in the use of this storage compound. Here we confirm this hypothesis and describe the biochemical characterization of the recombinant mannitol-2-dehydrogenase (ZgM2DH) and fructokinase (ZgFK1) involved in the catabolism of mannitol in Z. galactanivorans. This was completed by gene expression analysis indicating that the two genes encoding these enzymes are induced by mannitol and are organized as an operon.

MATERIALS AND METHODS

Bacterial strain and culture conditions.

The type strain of Z. galactanivorans, Dsij (31), was grown in Zobell medium 2216E (tryptone at 5 g liter⁻¹ and yeast extract at 1 g liter⁻¹ solubilized in seawater) at 20°C. Cells were transferred in 1 ml of Marine Mineral Medium supplemented with 4 ml of glucose or mannitol at 5 g l⁻¹. Briefly, one liter of Marine Mineral Medium is composed of 24.7 g NaCl, 6.3 g MgSO₄·7H₂O, 4.6 g MgCl₂·H₂O, 2 g NH₄Cl, 0.7 g KCl, 0.6 g CaCl₂, 200 mg NaHCO₃, 100 mg K₂HPO₄, 50 mg yeast extract, and 20 mg FeSO₄·7H₂O in Tris-HCl (50 mM; pH 8.0). Triplicate cultures were made for each substrate condition. Bacteria were grown at 20°C under agitation (180 rpm).

Nucleic acid extraction.

Genomic DNA was extracted as described previously (46). For RNA extraction, 2 ml of the growth media of cultures at the end of the exponential phase was added to 4 ml of RNAprotect reagent. RNAs were then isolated by using the RNA minikit (Qiagen) according to the manufacturer's instructions. Two steps of DNA digestion were performed using DNase I (Qiagen). RNAs were cleaned up on a minicolumn and were eluted in 30 μl of RNase-free water. Total elimination of genomic DNA was checked by PCR. For each sample, approximately 250 ng of RNAs was examined by 0.8% agarose gel electrophoresis to check their integrity. The concentrations of RNA samples were determined using a NanoDrop ND-1000 spectrophotometer. The OD₂₆₀ (optical density at 260 nm)/OD₂₈₀ and OD₂₆₀/OD₂₃₀ ratios were calculated to assess the purity of RNA extracts.

RT-PCR and quantitative PCR (qPCR).

cDNA were synthesized from 250 ng of total RNA using the Phusion reverse transcription-PCR (RT-PCR) kit (Finnzymes) with random hexamer primers according to the supplier's instructions, and cDNA samples were diluted to 1 ng μl⁻¹ and were stored at −80°C.

For the amplicons named Zg1489-1490, Zg1490-1491, Zg4259-4260, Zg4259-4261, Zg4260-4263 and Zg4262-4264, the PCRs were performed using the Advantage 2 PCR kit (Clontech) with 0.2 μM each primer (Table 1) and 20 ng of cDNA. The PCR program was as follows: 94°C for 5 min; 35 cycles of 94°C for 30 s, 50°C or 55°C for 30 s, and 72°C for 30 s or 3 min; and finally 5 min at 72°C. The reaction products were analyzed by 1% agarose gel electrophoresis.

TABLE 1.

Primers used in this study

Primers	Nucleotide sequence (5′–3′)a	Product length (bp)	Tm (°C)
qPCR primers
Zg1489-F	ACAAATTCGTCCCTACCCA	111	59
Zg1489-R	TCCTCCCTATTATAGCCATCC		58
Zg1490-F	GAAGACACCATCTATACCCGT	95	59
Zg1490-R	GTAGTATCCTGAATCGCCCA		59
Zg1491-F	CACACCTCATCCTATCCGA	79	58
Zg1491-R	GGTAGTGGTAATGGCTCCT		59
Zg4260-F	CTTATGCTGGAGGACTAAAGG	91	58
Zg4260-R	GAGGACACAAACATCGGAC		59
Zg4261-F	CATAGATGAGCCTACCCGA	112	58
Zg4261-R	GCATATCCGATGAGATGACC		58
Zg4262-F	CGTGACCTTTGCCATTCTC	108	59
Zg4262-R	GTAAGCCACCAAATAGCCC		59
Zg4263-F	CTTGCGACAACATTCAGGG	66	60
Zg4263-R	GCGGTGGAAACATAAGTGAG		59
Zg4264-F	TAGTACAGGTAAGCTTGTCCC	75	59
Zg4264-R	TTTATGTTGTCCCAAGCCC		58
cDNA PCR primers
Zg1489-1490-F	ACAAATTCGTCCCTACCCA	3,289	59
Zg1489-1490-R	GTAGTATCCTGAATCGCCCA		59
Zg1490-1491-F	GAAGACACCATCTATACCCGT	1,669	59
Zg1490-1491-R	GGTAGTGGTAATGGCTCCT		59
Zg4259-4260-F1	GCCATGTGGAATCCAGAAAAGGA	240	68
Zg4259-4260-R-	AAGGCGTTAAGTAAAATGGTAGATAC		72
Zg4259-4260-F2	TCAAATGCGTACAAAATTCCTCATAAC	1,240	72
Zg4259-4260-R	AAGGCGTTAAGTAAAATGGTAGATAC		72
Zg4259-4261-F1	GCCATGTGGAATCCAGAAAAGGA	1,260	68
Zg4259-4261-R	GCTTTTTGTATTTTGCATATCGTCCAT		72
Zg4260-4263-F	CTTATGCTGGAGGACTAAAGG	3,543	58
Zg4260-4263-R	GCGGTGGAAACATAAGTGAG		59
Zg4262-4264-F	CGTGACCTTTGCCATTCTC	2,530	59
Zg4262-4264-R	TTTATGTTGTCCCAAGCCC		58
Zg4260-4264-F	TACTTAACGCCTTAGGGTGTGCCGA	5,820	68
Zg4260-4264-R	CCTTCCCTACCGGCTACCATAGCTC		68
Cloning primers
Zg4263pFO4-F	GGGGGGGGATCCAAAAATTATAAATTAAACAGCACTAATCTT	1,491	70b
Zg4263pFO4-R	CCCCCCGAATTCTTAATTGTTGTTCTCCTGGTTACGGTTT		70b
Zg4264pFO4-F	GGGGGGGGATCCAAAAATATAGTCTGTTTTGGGGAAGTT	882	70b
Zg4264pFO4-R	CCCCCCGAATTCTTATTGGCGACTTTTGATAAAGGCTTGG		72b
Zg1491pFO4-F	GGGGGGGGATCCAAAACCGTATACTGCATTGGAGAATT	942	70b
Zg1491pFO4-R	CCCCCCATGCATTTATTCCCCGAAAACCATGCTGTTGTC		72b

^aRestriction sites are indicated in italics.

^bCalculation of T_m does not include the sequences corresponding to the restriction sites or the sequence upstream.

For quantitative PCR, the primers corresponding to the eight candidate genes were designed using the Perl Primer open-source software (47). They were 17 to 22 nucleotides (nt) long, with melting temperatures (T_m) between 58 and 60°C. A difference of 1°C between the forward and the reverse primer was accepted. The GC content of the primers was between 47 and 53%. The ΔG was greater than −10 kcal mol⁻¹ to avoid the pairing of the primer pairs. The PCR products ranged from 60 to 120 bp and had a GC content of >60%. The nucleotide sequences of all primers were compared to the genomic sequence of Z. galactanivorans (GenBank accession number FP476056) to check gene specificity. PCR of the genomic DNA of Z. galactanivorans using the different primer pairs was carried out to confirm specific amplification. Quantitative PCRs were performed in 96-well plates (Thermo Scientific) on a LightCycler system, model 480 (Roche). The composition of each reaction mixture was as follows: 2.5 ng of cDNA, 250 nM each primer (Table 1), 5 μl of 2× SYBR green, and water for a final volume of 10 μl. The reactions for each gene were carried out in technical triplicate. The program was as follows: 5 min at 95°C, followed by 45 cycles of 95°C (10 s), 51°C (15 s), and 72°C (15 s) with a single acquisition mode. Serial dilutions of genomic DNA ranging from 10 to 10⁵ copies were amplified by qPCR in the same run as the cDNA samples. The LightCycler480 software was used to obtain crossing point (Cp) values and PCR efficiencies. Relative expression was determined with REST 2009 software using glyA and icdA as reference genes (48). This software determines the ratio corresponding to the relative expression by the equation of Pfaffl: ratio = E_target^ΔC_T target (control − sample)/E_reference^ΔC_T reference (control − sample), where E refers to efficiency and C_T is the threshold cycle (49). The influence of mannitol on the expression of genes of interest was determined by comparing their expression levels in the presence of glucose.

Construction of plasmids for overexpression in Escherichia coli.

The Zg1491, Zg4263, and Zg4264 coding regions were cloned into the pFO4 expression vector as described previously (50). Briefly, the genes were amplified by PCR (the primers used are listed in Table 1) from Z. galactanivorans genomic DNA; the PCR fragments were digested by BamHI/MfeI for Zg1491 and by BamHI/EcoRI for Zg4263 and Zg4264 and were ligated into the pFO4 vector; and recombinant plasmids named pZg1491, pZg4263, and pZg4264, respectively, were transformed into E. coli BL21(DE3). The integrity of their sequences was verified by sequencing.

Nucleotide and protein sequence analyses.

The mannitol degradation-related genes of Z. galactanivorans were identified in its complete genome (GenBank accession number FP476056) using the GenDB program, version 2.4 (51). Signal peptide and transmembrane helices were predicted using SignalP version 2.0 (52) and TMHMM (53), respectively. The presence of orthologous genes in all available prokaryotic genomes was screened using the genomic BLASTP (54) at NCBI (http://www.ncbi.nlm.nih.gov/genome; accessed 14 January 2014). BLASTP analysis was also done against the UniProtKB/Swiss-Prot database in order to gain insight into the enzymatic characteristics of the Zobellia proteins. The Z. galactanivorans target proteins were aligned with their orthologs using MAFFT, applying the iterative refinement method and the scoring matrix Blosum62 (55). To identify putative promoters, intergenic regions were searched for the −33 (TTG)/−7 (TANNTTTG) motif with a spacer ranging from 10 to 30 bp. This motif has been identified as the consensus promoter in Flavobacterium species and is conserved in Bacteroides fragilis (56, 57). TransTermHP (58) software was used to predict putative Rho-independent transcriptional terminators.

Expression and purification of recombinant proteins.

E. coli strain BL21(DE3) (Novagen) transformed with plasmid pZg1491, pZg4263, or pZg4264 was grown in ZYP 5052 medium (59) at 20°C for 72 h, and protein expression was induced with 0.2% lactose supplemented with 200 μg/ml (final concentration) of ampicillin at 20°C and 200 rpm on an orbital shaker. Recombinant proteins were purified using the ÄKTA avant system (GE Healthcare) equipped with a HisPrep FF 16/10 column (GE Healthcare) equilibrated in 20 mM Tris-HCl (pH 7.5) buffer containing 200 mM NaCl and 15 mM imidazole. Proteins were eluted using a linear increasing gradient (from 0 to 100%) of 20 mM Tris-HCl (pH 7.5) containing 200 mM NaCl and 500 mM imidazole within 10 column volumes. Fractions were collected in a deep 96-well plate with 1 ml per well. Aliquots of fractions were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using 12% Criterion precast Bis-Tris gels (Bio-Rad). Protein concentrations were measured at 280 nm using a NanoDrop 2000 spectrophotometer (Thermo Fisher). A molar extinction coefficient of 58.705 M⁻¹ cm⁻¹ and a molecular mass of 55.61 kDa were used for the Zg4263 protein, and a molar extinction coefficient of 26.025 M⁻¹ cm⁻¹ and a molecular mass of 32.69 kDa were used for Zg4264, to calculate the concentration of pure proteins (ExPASy, ProtParam tool [60]). The molecular masses of the native recombinant proteins were estimated by size exclusion chromatography. Eluted fractions from Ni²⁺ affinity chromatography were loaded at a flow rate of 1 ml min⁻¹ onto a calibrated Sephacryl HiLoad Superdex 200 column (GE Healthcare) equilibrated with a buffer containing 20 mM Tris-HCl (pH 7.5) and 150 mM NaCl. Fractions were then eluted with the same buffer at a flow rate of 1 ml min⁻¹.

Determination of mannitol-2-dehydrogenase activity.

Unless noted otherwise, all the assays were carried out in triplicate at 25°C in 100-μl reaction volumes. All compounds used were ordered from Sigma-Aldrich. M2DH activity was assayed in both directions (i.e., fructose reduction, which produces mannitol, and mannitol oxidation, which produces fructose) by following changes in absorbance at 340 nm in a Safire2 UV spectrophotometer microplate reader (Tecan). The standard fructose reduction reaction mixture contained 0.1 M Tris-HCl buffer (pH 7), 1 mM fructose, 0.2 mM NADH, and 1 to 10 μg of purified enzyme. The standard mannitol oxidation reaction mixture contained 0.1 M Tris-HCl (pH 8.5), 1 mM mannitol, 0.5 mM NAD⁺, and 1 to 10 μg of recombinant enzyme. Both reactions were initiated by adding the substrate.

The dependences of enzyme activities on pH and temperature were determined over a pH range of 5.5 to 9.5 and a temperature range of 10°C to 50°C in steps of 10°C. The buffers used were 0.1 M morpholineethanesulfonic acid (MES) buffer (pH 5.5 to 6.5), 0.1 M Bis-Tris propane (pH 6.5 to pH 9.5), and 0.1 M Tris-HCl (pH 7 to pH 9). The effects of various metal ions and chemical reagents were examined at final concentrations of 1 and 10 mM. For enzymatic characterization, NAD(H) and NADP(H) were tested as potential cofactors, and different sugars and polyols (d-fructose, d-glucose, d-mannose, d-galactose, d-xylose, d-mannitol, d-sorbitol, d-arabitol) were tested as potential substrates. The influence of NaCl on M2DH activity was assessed by testing final concentrations ranging from 0 to 2 M. The kinetic parameters K_m and V_max were determined at 25°C for the reduction (at pH 7) and oxidation (at pH 8.5) reactions by varying the concentrations of d-fructose and NADH for the former and those of d-mannitol and NAD⁺ for the latter. For both activities, 1 U corresponded to 1 μmol of NAD(H) reduced or oxidized per min per mg of protein.

Determination of HK/FK activity.

Hexokinase (HK)/fructokinase (FK) activity was determined by an enzyme-coupled assay. The reaction mixture (total volume of 100 μl) contained the Zg4264 enzyme in solution in 50 mM Tris-HCl (pH 7.6), in the presence of 1 mM ATP, 100 mM KCl, 1.5 mM MgCl₂, 1 mM phosphoenolpyruvate, 0.5 mM freshly prepared NADH, and 0.2 μl of a mixture composed of lactate dehydrogenase (LDH) (900 to 1,400 U ml⁻¹) and pyruvate kinase (PK) (600 to 1,000 U/ml) (Sigma). The reaction was initiated with 1 mM fructose. Enzyme activity was measured by following changes in absorbance at 340 nm in a Safire2 UV spectrophotometer microplate reader (Tecan). The temperature and pH optimum were determined under the same conditions as those for M2DH activity. d-Fructose, d-glucose, d-mannose, d-sorbitol, d-mannitol, mannitol-1-phosphate, glucose-1-phosphate, fructose-1-phosphate, mannose-6-phosphate, glucose-6-phosphate, and fructose-6-phosphate were tested as potential substrates, and EDTA, mannose-6-phosphate, and fructose-6-phosphate were tested as potential inhibitors. The influence of NaCl on this enzyme was assessed by testing final concentrations ranging from 0 to 2 M. Kinetic parameters were determined at 25°C and pH 7.5 by varying the concentrations of d-fructose, d-mannose, and ATP. One unit corresponded to 1 μmol of NADH oxidized per min per mg of protein.

Thermostability analysis.

The thermostabilities of Zg4263 and Zg4264 were studied by dynamic light scattering (DLS). A solution of 50 μl of each enzyme at 10 mg ml⁻¹ was filtered on a 0.2-μm membrane. By using a Zetasizer Nano instrument (Malvern), the protein solution was heated from 10°C to 70°C in steps of 1°C for a total period of 12 h, and the radius of gyration (Rg) was measured at each degree. The denaturation temperature was determined as the point of sharp change in the radius of gyration.

RESULTS

Identification of genes coding for enzymes potentially involved in mannitol degradation.

In the genome of Z. galactanivorans, the Zg4263 gene (1,494 bp) is annotated as encoding a putative cytosolic d-mannonate oxidoreductase. The corresponding protein of 498 amino acids (aa), Zg4263, is 46% identical to the mannitol-2-dehydrogenase (M2DH) of Pseudomonas fluorescens DSM 50106 (61) and 42% identical to the M2DH of Corynebacterium glutamicum ATCC 13032 (62). These characterized M2DHs belong to the long-chain dehydrogenase/reductase family (LDR). Furthermore, the Zg4264 gene (885 bp; corresponding to a protein of 295 aa), which is located immediately downstream of Zg4263, encodes a putative cytosolic fructokinase (FK) belonging to the PfkB family of carbohydrate kinases. BLASTP analyses on the Z. galactanivorans genome with Zg4263 and Zg4264 sequences showed the absence of other genes coding for a putative M2DH but the presence of another putative cytosolic FK of 315 amino acids (also potentially belonging to the PfkB family of carbohydrate kinases) encoded by the Zg1491 gene (945 bp). There is 26% identity between the putative Zg4263 and Zg1491 FKs. Based on these observations, it was suggested that Zg4263, Zg4264, and Zg1491 could be implicated in the utilization of mannitol by Z. galactanivorans. To test this hypothesis, these three genes were cloned into E. coli. While both the Zg4263 and Zg4264 genes were overexpressed under the conditions tested, no overexpression was observed for Zg1491, despite several attempts to improve the conditions of induction. Therefore, we focus our attention on the Zg4263 and Zg4264 proteins.

Expression and purification of recombinant Zg4263 and Zg4264 proteins.

Recombinant Zg4263 and Zg4264 proteins were successfully expressed in high quantities in the E. coli BL21(DE3) expression strain as soluble forms and were purified to electrophoretic homogeneity (see Fig. S1A and B in the supplemental material). The final yield after one step of purification by affinity chromatography was about 30 mg for 200 ml of ZYP medium for Zg4263 and about 25 mg for Zg4264 under the same conditions. The Zg4263 and Zg4264 proteins eluted after the size exclusion chromatography step had apparent molecular masses of 56 kDa and 32 kDa, respectively. Comparison of these results with the theoretical masses of 55.6 kDa and 32.7 kDa indicated that both proteins are in the monomeric form in solution. The specific activities of the two recombinant enzymes were compared after Ni²⁺ affinity chromatography with and without subsequent gel filtration, and no significant difference was observed. This indicated that the Superdex 200 column did not improve the levels of purity of recombinant Zg4263 and Zg4264. These results were confirmed by dynamic light scattering (DLS) experiments. DLS was also used to study protein thermostability; above 40°C, a sharp increase in the radius of gyration (Rg) was observed, indicating the beginning of the denaturation of Zg4263 and Zg4264.

Biochemical characterization of Zg4263.

The specificity of the putative dehydrogenase/oxidoreductase Zg4263 was tested in the presence of different sugars and polyols and at several concentrations. Mannitol, arabitol, and sorbitol (at 5, 10, or 50 mM), with NAD⁺ or NADP⁺ as a cosubstrate, and fructose, mannose, glucose, galactose, and xylose (at 1, 10, or 50 mM), with NADH or NADPH as a cosubstrate, were examined as alternative substrates for Zg4263. Purified Zg4263 had a specific activity of 5.64 U/mg for fructose reduction with NADH at pH 6.5. For mannitol oxidation at pH 8.5, the specific activity was 7.18 U/mg. No activity on these substrates was observed when NAD(H) was replaced by NADP(H). This protein is thus a true M2DH and is referred to below as ZgM2DH. In the direction of polyol oxidation, only sorbitol at 50 mM exhibited activity, representing 14% of the level measured in the presence of mannitol with NAD⁺. In the reverse direction, no activity was observed in the presence of mannose, glucose, galactose, or xylose. These results show that the enzyme features narrow substrate specificity in both directions and may catalyze only the production of fructose or mannitol under physiological conditions.

The optimum pH for the mannitol oxidation catalyzed by ZgM2DH was 8.5, with 52% and 84% of maximum activity in Tris-HCl buffer at pH 8.0 and pH 9.0, respectively. The optimum pH for fructose reduction was 6.5, with 70% (Tris-HCl buffer) and 76% (MES buffer) of maximum activity at pH 6.0 and 7.0, respectively (Fig. 1A and B). These pH optimum values were in accordance with those measured for similar enzymes isolated from diverse bacteria (Table 2). The optimal temperature for mannitol oxidation was 40°C in Tris-HCl buffer at pH 8.5 (Fig. 1C), a value higher than the optimum temperatures determined for other organisms, except Thermotoga maritima (Table 2).

FIG 1.

Enzymatic properties of ZgM2DH. (A and B) Effects of various buffer pH values on the mannitol oxidation (A) and fructose reduction (B) activities of ZgM2DH. Symbols: □, 0.1 M Bis-Tris propane buffer; ◇, 0.1 M MES buffer; △, 0.1 M Tris-HCl buffer. (C) Effects of various temperatures on the mannitol oxidation activity of ZgM2DH. The enzyme activity observed at pH 8.5 (in panel A) or pH 6.5 (in panel B) and 40°C was considered 100%, and all other values are expressed as percentages of these activities. Each value represents means ± standard deviations calculated from three reaction assays for one round of purification.

TABLE 2.

Comparison of the biochemical properties of M2DHs from different organisms

Parametera	Z. galactanivoransb	Aspergillus fumigatus	P. fluorescens	Lactobacillus sanfranciscensis	R. sphaeroides	Leuconostoc pseudomesenteroides	T. maritima
Sp act (U/mg protein)
Fructose	5.64	94	55	15.48	60	450	54
Mannitol	7.18	220	31		30	106
pH optimum
Fructose	6.50	7.10	7.30	5.80	6.50	5.40	5.50
Mannitol	8.50	10	10	8	9	8.60	8.30
Temp optimum (°C)
Fructose	40	25	25	35			95
Mannitol	40		25
Km (mM)
Fructose	2.30	60	25	24	16.30	44	51
NADH	0.02	0.02	0.15		0.02	0.09	0.05
Mannitol	1.12	13	1.20	78	0.29	12	5.51
NAD+	0.51	0.15	0.15		0.18	0.20	0.14
kcat/Km (mM−1 s−1)
Fructose	2.30	1.40	2.20
NADH	263	4,700	436
Mannitol	6	17	18
NAD+	13	1,400	159
kcat (s−1)
Fructose	5.27	86	54
Mannitol	6.70	212	20
Mol mass (kDa) (no. of subunits)	56 (1)	58 (1)	54 (1)	53 (1)	52 (1)	36 (4)	34 (4)
Classificationc	LDR	LDR	LDR		LDR	MDR	MDR
Reference	This work	81	61, 86	87	80, 88	89	83

^aEach value represents the mean for triplicate measurements, and the standard errors for the parameters are less than 10%.

^bThe kinetic parameters of ZgM2DH were obtained at 25°C in 0.1 M Tris-HCl at pH 6.5 for the d-fructose reduction and at 25°C in 0.1 M Bis-Tris propane buffer at pH 8.5 for the d-mannitol oxidation.

^cLDR, long-chain dehydrogenase/reductase family; MDR, Zn²⁺-containing medium-chain dehydrogenase/reductase family.

M2DH oxidation and reduction activities were measured in the presence of different chemicals (Table 3). The results showed that the enzyme was only slightly sensitive to metal chelators such as EDTA (at final concentrations of 1 and 10 mM) and to mercaptoethanol. In contrast, Ca²⁺, Mg²⁺, K⁺, and Li⁺, inhibited M2DH activity in both directions by 20 to 30%, while NH₄⁺ decreased mannitol oxidation activity by 23% but fructose reduction activity by 54%. NaCl at 1 M slightly increased oxidation activity but significantly inhibited (by 45%) reduction activity. Similar effects were observed in the presence of 1.5 and 2 M NaCl (data not shown). Moreover, both activities of ZgM2DH were inhibited by their reaction products in a similar manner (about 50% at 50 mM). Interestingly, mannitol production was strongly activated by the addition of fructose-6-phosphate (20 mM), which was the product of FK activity; conversely, this reaction was inhibited (by 25%) by the addition of mannitol-1-phosphate (20 mM).

TABLE 3.

Effects of metal ions and chemical reagents on the activity of ZgM2DH^a

Ion or compound	Concn (mM)	Relative activity (%)b
		Mannitol oxidation	Fructose reduction
Ca2+	10	73	83
K+	10	79	80
Li+	10	83	86
NH4+	10	77	46
Mg2+	10	103	70
Mercaptoethanol	10	96	90
EDTA	10	115	96
Fructose	20	70	ND
Fructose	50	47	ND
Mannitol	20	ND	73
Mannitol	50	ND	46
Mannitol-1-phosphate	20	95	77
Fructose-6-phosphate	20	93	157
NaCl	100	100	100
NaCl	1,000	120	55

^aThe purified enzyme was assayed under standard assay conditions with ions in chloride forms and several chemicals.

^bPercentage of the activity remaining under the specific condition tested relative to the activity observed in the absence of any of these ions or compounds, taken as 100%. Each value represents the mean of triplicate measurements. The standard errors for activities are less than 10%. ND, not determined.

Initial velocities were determined in the standard assay mixture at pH 8.5 for mannitol oxidation and at pH 6.5 for fructose reduction. The substrates and cofactors had hyperbolic saturation curves, and the corresponding double-reciprocal plots were linear (see Fig. S2 in the supplemental material). The concentrations of mannitol and fructose substrates ranged from 0.1 to 50 mM, that of NAD⁺ ranged from 0.01 to 10 mM, and that of NADH ranged from 0.01 to 1 mM, since that was the highest cofactor concentration compatible with the spectrophotometric assay. The results showed that the ZgM2DH enzyme had equivalent K_m and k_cat values for mannitol and fructose (Table 2). Indeed, the specific activities and K_m values were of the same order of magnitude, i.e., 7.18 U/mg protein and 1.12 mM for mannitol and 5.64 U/mg protein and 2.30 mM for fructose.

Biochemical characterization of Zg4264.

The recombinant Zg4264 protein preferentially catalyzed the ATP-dependent phosphorylation of fructose into fructose-6-phosphate (specific activity, 2.55 U/mg at 5 mM fructose) but also used mannose as a substrate. Indeed, when mannose was added to the reaction medium at a final concentration of 10 or 50 mM, the specific activity corresponded to 5% or 16%, respectively, of that measured in the presence of fructose. The enzyme was not active on d-glucose, d-galactose, d-xylose, d-sorbitol, d-mannitol, mannitol-1-phosphate, glucose-1-phosphate, fructose-1-phosphate, mannose-6-phosphate, glucose-6-phosphate, and fructose-6-phosphate as phosphoryl acceptors. This protein represents a new FK and is referred to below as ZgFK1.

This fructokinase was active in the pH range of 5.5 to 9, with an optimum at pH 7.5 (Fig. 2A) (specific activity, 3.95 U/mg at 10 mM fructose), and its activity decreased sharply at more alkaline pH values (only 30% of maximum activity at pH 8.5), in contrast to what was observed for acidic pH values (73% activity at pH 6). The highest enzyme activity was observed at 40°C (Fig. 2B). For these assays, the reaction mixture contained KCl at a final concentration of 100 mM.

FIG 2.

Enzymatic properties of ZgFK1. (A) Effects of various Tris-HCl buffer pH values on the activity of ZgFK1. (B) Influence of various temperatures on the activity of ZgFK1. (C) Effects of various KCl concentrations on the activity of ZgFK1. The enzyme activity observed at pH 7.5 and 40°C in the presence of 100 mM KCl was considered 100%, and all other values are expressed as percentages of this activity. Each value represents means ± standard deviations calculated from three reaction assays for one round of purification.

Indeed, Zg4264 was dependent on a monovalent cation, since no activity was detectable in the absence of K⁺. Specific activity was highest in the presence of K⁺ at a final concentration of 100 mM (Fig. 2C). Na⁺ could partially replace K⁺ (about 20% of the activity at a final concentration of 0.05 to 1 M Na⁺), while no activity was observed in the presence of MgCl₂. This result contrasted with the Mg²⁺ requirement of several previously characterized bacterial FKs (Table 4). The influences of mannitol-6-phosphate and fructose-6-phosphate on hexokinase activity were determined with fructose or mannose as the substrate (final concentration in the reaction mixture, 10 or 50 mM, respectively). The addition of 20 mM (final concentration) mannitol-6-phosphate inhibited fructose phosphorylation by 85% and mannose phosphorylation by 80%, while the addition of 20 mM (final concentration) fructose-6-phosphate inhibited fructose phosphorylation by 45% and mannose phosphorylation by 70%.

TABLE 4.

Comparison of the biochemical properties of fructokinases from different organisms

Parametersa	Z. galactanivoransb	Bifidobacterium longum	Lactococcus lactis	Methylomicrobium alcaliphilum	Rhizobium leguminosarum	Thermococcus litoralis	Zymomonas mobilis
Sp act (U/mg protein) with fructose	2.55	0.84	112	141	10.80	730	250
Substrate specificity	Fructose, mannose	Fructose	Fructose, mannose	Fructose	Fructose	Fructose	Fructose, mannose
pH optimum	7.50	6	7	9	8	7.50	7.40
Temp optimum	40	50		60		80
Km (mM)
Fructose	10	0.74	0.31	0.26	0.31	2.30	0.7
ATP	0.06	0.76	0.59	1.30		0.81	0.45
Metal requirement	K+		Mg2+	Mg2+	Mg2+	Mg2+
Mol mass (kDa) (no. of subunits)	32 (1)	35 (?)	32 (2)	35 (1)	36 (1)	35 (2)	28 (2)
Reference	This work	90	76	91	92	93	94

^aEach value represents the mean for triplicate measurements, and the standard errors for parameters are less than 10%.

^bThe kinetic parameters of ZgFK1 were obtained at 25°C and pH 7.

The purified enzyme Zg4264 exhibited typical Michaelis-Menten kinetics when assayed with increasing concentrations of the substrates investigated (d-fructose, d-mannose, and ATP) (see Fig. S3 in the supplemental material). The apparent K_ms for d-fructose, d-mannose, and ATP are 10, 26.6, and 0.31 mM, respectively. Although the K_ms for fructose and mannose were in the same range, the k_cat/K_m value for fructose (0.136 mM⁻¹ s⁻¹) was 60 times higher than that for mannose (0.0023 mM⁻¹ s⁻¹). Similarly, the k_cat/K_m value for ATP in the presence of fructose was 23 mM⁻¹ s⁻¹, in contrast to 1 mM⁻¹ s⁻¹ in the presence of mannose (data not shown).

Organization of the gene clusters related to mannitol degradation in Z. galactanivorans.

In the genome, two gene clusters were predicted to be involved in the uptake/catabolism of mannitol (Fig. 3A). The first cluster, about 6,200 bp long, contained three genes, Zg1489, Zg1490, and Zg1491, annotated as encoding putative SusC-like TonB-dependent receptor (TBDR), SusD-like, and ScrK1 fructokinase proteins, respectively. The second cluster, about 6,900 bp long, featured six genes named Zg4259 to Zg4264. These two clusters were separated by 3,210 kbp in the bacterial chromosome. In this study, we have characterized the Zg4263 and Zg4264 proteins. The Zg4259 gene codes for an unknown conserved protein. The three genes Zg4260, Zg4261, and Zg4262 correspond to a putative transmembrane protein complex involved in the uptake of monosaccharides. The predicted amino acid sequences encoded by these open reading frames (ORFs) revealed a high similarity to different components of the ribose transport system in E. coli (63, 64). The first gene, Zg4260, of 1,062 nt, codes for a peptide of 354 aa with a calculated molecular mass of 32 kDa. The deduced protein showed an overall amino acid sequence identity of 31% with the E. coli ribose binding periplasmic protein RbsB. A putative peptide signal of 29 aa was located at the N terminus. The second gene, Zg4261, of 1,527 nt, encodes a peptide of 509 aa with a molecular mass of 56 kDa. The deduced protein exhibited an overall amino acid sequence identity of 41% with the E. coli ribose ATP-binding protein RbsA. Finally, the third gene, Zg4262, of 969 nt, codes for a peptide of 323 aa with a calculated molecular mass of 34 kDa. The deduced protein showed an overall amino acid sequence identity of 43% with the E. coli ribose permease protein RbsC, was predicted to be very hydrophobic, and contained nine transmembrane domains.

FIG 3.

Cluster organization and expression analysis of genes involved in the catabolism of mannitol in Z. galactanivorans. (A) Genetic organization of two putative mannitol utilization clusters. The numbers at the extremities of each operon indicate positions within the bacterial genome. (B) Analysis of the transcription patterns of these two clusters. RT-PCR experiments were performed on RNAs extracted from cultures with glucose (G) or mannitol (M). Agarose gels show the PCR products indicated in panel A by numbers 1 to 7. MW, molecular mass standard. (C) Differential expression of the genes of interest between cultures in the presence of mannitol and cultures in the presence of glucose. Bars indicate ratios of expression in the presence of mannitol to expression in the presence of glucose.

In order to test the hypothesis that some of these genes may be expressed on polycistronic messengers, semiquantitative RT-PCR experiments were carried out on cDNA prepared from RNA extracted from cultures for which the main source of carbon was glucose or mannitol. For the first gene cluster, the results showed that the TonB-dependent receptor and susD-like genes are expressed in the same polycistronic mRNAs, while this was not the case for the fructokinase scrK1 and susD-like genes (Fig. 3B, left). For the second gene cluster, three amplicons (Fig. 3B, middle) indicated that Zg4259 is expressed on polycistronic mRNAs containing Zg4260 and Zg4261. Then two long, overlapping PCR products were obtained: one showed that the three genes of the putative transport system (rbsB, rbsA, rbsC) were expressed on the same RNA messengers with the gene coding for the ZgM2DH, and the second corresponded to the three genes encoding RbsC, ZgM2DH, and ZgFK1(Fig. 3B, right). The occurrence of these both overlapping transcripts suggests that all the genes of this cluster are likely to be expressed on the same polycistronic RNA, even if no PCR product (expected length of 5,820 bp) was observed in the attempt to amplify a fragment covering the region from Zg4260 to Zg4264. Similar patterns of amplification were obtained after Z. galactanivorans was grown in the presence of glucose or mannitol as the main source of carbon.

To complete this gene expression analysis, qPCR experiments were carried out for individual genes to test the influence of mannitol on their transcription relative to transcription in the presence of glucose. The levels of expression were quite similar under culture in the presence of glucose or mannitol for the first cluster (ratio almost equal to 1). (Fig. 3C). In contrast, for the second cluster, genes coding for subunits of the ABC transporter and for proteins catalyzing the utilization of mannitol were induced simultaneously after culture in the presence of mannitol, supporting their organization as a true operon (Fig. 3C).

Taxonomic distribution and organization of genes orthologous to those encoding ZgM2DH and ZgFK1 in bacteria.

The Z. galactanivorans genome was analyzed to identify regions of sequence mediating transcription, such as promoters and Rho-independent terminators. No promoter sequence matching the consensus identified in other Bacteroidetes species (56, 57) was found between the Zg4259 and Zg4260 genes. However, directly upstream of the start codon of Zg4259, a gene encoding an unknown conserved protein, a putative promoter sequence presenting only one mismatch with the consensus sequence (TTG/TANNTTTG) was identified. The length of the spacer separating the −7 (TANNTTTG) and −33 (TTG) motifs was 19 bp. Furthermore, this region was relatively poor in GC (27% GC) compared with the mean value for the whole genome (43% GC). These parameters were consistent with studies of Flavobacterium species showing that promoter strength was influenced by spacer length (19 bp being the optimal value) and that the sequence of the promoter was enriched with A and T nucleotides (56). Sequence analyses also revealed the presence of a single putative terminator in this cluster of genes, downstream of the Zg4264 locus.

Orthologs of Zg4263 and Zg4264 genes were searched for in genomes of archaea (338 genomes) and bacteria (19,904 genomes) available in the NCBI genomic Blast server (14 January 2014). The genomic vicinity of each ortholog identified was then screened to determine the presence of genes coding for an ATP binding cassette (ABC) transporter complex in these microbes. Such clusters of genes involved in mannitol degradation, or mannitol utilization clusters, were absent in archaeal genomes and present in the Bacteroidetes phylum genomes, but only in 6 of the 76 complete genomes available for the Flavobacteriaceae family: Z. galactanivorans Dsij^T (31), Gillisia sp. strain CAL575 (65), Lacinutrix sp. strain 5H-3-7-4 (66), Cellulophaga lytica DSM 7489 (67), Cellulophaga algicola DSM 14237 (68), and Winogradskyella psychrotolerans RS-3^T (69). When the entire operon containing genes coding for the unknown conserved protein, RbsB, RbsA, RbsC, M2DH, and FK1 was considered, analysis of nucleotide sequences of Flavobacteriaceae showed very high percentages of identity, between 64% (minimum value, obtained for Z. galactanivorans and Winogradskyella psychrotolerans RS-3^T) and 75% (maximum value, obtained for Lacinutrix sp. strain 5H-3-7-4 and Winogradskyella psychrotolerans RS-3^T). In the same vein, analysis of amino acid sequences of the different proteins encoded by the mannitol utilization gene cluster showed the highest degree of conservation within the six Flavobacteriaceae species considered (Fig. 4) compared to orthologous genes identified in other bacteria. As an illustration, the percentages of identity of putative ABC transporter proteins ranged from 69% to 89% within these six species but decreased to a maximum of 49% with ABC transporter proteins of other bacteria. Interestingly, the organization of the operon observed in the species mentioned above was not conserved in Bacteroidetes. Indeed, in the genomes of Formosa agariphila KMM 3901^T (Flavobacteriaceae family) (70), Marinilabilia salmonicolor JCM 21150 (Marinilabiliaceae family) (71), and Cytophaga fermentans JCM 21142^T (Cytophagaceae family) (72), a gene coding for a predicted mannitol permease was identified upstream of orthologs of M2DH and FK1, while no gene for an ABC transporter complex was found.

FIG 4.

Organization of genes related to mannitol catabolism in selected bacteria. Each arrow represents a gene and its orientation. The genes code for an unknown conserved protein, a ribose binding periplasmic protein (RbsB), a ribose ATP-binding protein (RbsA), a ribose permease protein (RbsC), a mannitol-2-dehydrogenase (M2DH), and a hexokinase (FK1). The percentage of identity of each orthologous gene to the Z. galactanivorans gene is given inside the arrow. P, putative promoter; T, putative terminator. Z. galactanivorans, Zobellia galactanivorans Dsij^T; G. sp. CAL575, Gillisia sp. strain CAL575; L. sp. 5H-3-7-4, Lacinutrix sp. strain 5H-3-7-4; C. lytica, Cellulophaga lytica DSM 7489; C. algicola, Cellulophaga algicola DSM 14237; W. psychrotolerans, Winogradskyella psychrotolerans RS-3^T; F. agariphila, Formosa agariphila KMM 3901^T; M. salmonicolor, Marinilabilia salmonicolor JCM 21150; C. fermentans, Cytophaga fermentans JCM 21142^T.

DISCUSSION

Different metabolic pathways have been described for the use of mannitol by bacteria: the phosphoenolpyruvate-dependent phosphotransferase system (PTS) and a M2DH-based catabolic pathway, the latter involving different types of transporters for mannitol. The PTS consists of a phosphocarrier or histidine protein (HPr), a protein kinase enzyme I (EI), and a substrate-specific enzyme II (EII) (73). This system has been extensively studied in E. coli (19), Bacillus subtilis (20, 21), Bacillus stearothermophilus (22), Clostridium acetobutylicum (23), Streptococcus mutans (24), and Vibrio cholerae (74). In contrast, in the P. fluorescens mannitol operon, this polyol is transported by an ATP binding cassette transporter (MtlEFGK) and is oxidized by an M2DH (MtlD) to produce fructose, which is then converted into fructose-6-phophate by a kinase (MtlZ) (28). Lastly, the C. glutamicum mannitol operon contains two structural genes (mtlT and mtlD), encoding a major facilitator superfamily (MFS) transporter and an M2DH, respectively, and a regulator gene (mtlR) corresponding to a repressor (75).

In this report, we describe the identification of a cluster of genes involved in mannitol degradation based on the analysis of the genome of the Flavobacteriia member Z. galactanivorans. This cluster is approximately 6.9 kb long and comprises six genes in the following order: an unknown conserved protein (Zg4259), a putative mannitol ABC transporter complex (Zg4260, Zg4261, and Zg4262), a mannitol-2-dehydrogenase (Zg4263), and a fructokinase (Zg4264). The proteins encoded by the last two genes have been purified and successfully characterized: Zg4263 is a NAD-dependent M2DH (EC 1.1.1.67) capable of both mannitol synthesis and degradation, and Zg4264 is a fructokinase (EC 2.7.1.4), a member of the PfkB family of carbohydrate kinases, and is active on fructose and mannose in the presence of a monocation, preferentially K⁺. The latter observation is in line with the fact that bacterial fructokinases are usually mannofructokinases (76, 77).

Analysis of the primary sequence of ZgM2DH indicated that this protein is a member of the long-chain MDH family (78) and more specifically of the polyol-specific long-chain dehydrogenase/reductase subfamily (PSLDR) (79). Most members have been identified by similarity at the level of amino acid sequence, and only a few have been biochemically characterized (61, 80, 81). ZgM2DH was active as a monomer, did not require metal ions for catalysis, and featured narrow substrate specificity. These observations are in contrast to the results obtained for polyol dehydrogenases characterized in Saccharomyces cerevisiae (82), Rhodobacter sphaeroides (80), P. fluorescens (61), or Thermotoga maritima (83), because those M2DHs feature broader substrate specificity. In addition, most of the other bacterial M2DHs characterized have significantly higher K_ms and activity in the direction of fructose reduction to produce mannitol than for the oxidation of this polyol, while the Z. galactanivorans M2DH seems to have no preference between fructose and mannitol (Table 2). Moreover, none of the six cysteines present in the ZgM2DH protein are likely to be important for catalytic reactions, because mercaptoethanol has no effect on either activity of the enzyme. In contrast, oxidation and reduction are inhibited by their reaction products in a similar manner (about 50% at 50 mM). Such an effect has been already observed for fructose reduction in T. maritima TM0298 (83) and for both directions in S. cerevisiae (82).

Genes encoding ZgM2DH and ZgFK1 were adjacent in the genome and were located directly downstream of four loci coding for an unknown conserved protein and constituents of an ABC transporter complex, thus forming a cluster of six genes. This prompted us to investigate whether the six loci could be part of the same operon. Transcriptional analysis strongly suggests that these genes constitute a genuine operon in Z. galactanivorans. Analysis of publicly available prokaryotic genomes showed this operon for mannitol catabolism to be conserved only in five other species of Flavobacteriaceae. In addition, M2DH and FK1 genes were retrieved in three species of Bacteroidetes (F. agariphila, M. salmonicolor, and C. fermentans) harboring a gene predicted to encode a putative mannitol permease gene instead of genes corresponding to the different subunits of an ABC transporter.

The occurrence of two types of operons for the degradation of mannitol in a few members of the Bacteroidetes is intriguing and raises questions about the evolution of this metabolic pathway in this phylum. It is tempting to speculate that the M2DH and FK genes have been acquired recently by the last common ancestor of the nine strains of Bacteroidetes from which both genes have been retrieved, and this hypothesis is partly supported by the high sequence identity between these genes in these different species. This common ancestor might have been a bacterium associated with brown algae, which are known to produce and exude mannitol. Furthermore, the occurrence of such an operon may have conferred an adaptive advantage on this bacterium, because mannitol can be used as a carbon and energy source. The observation that extant flavobacterial species containing the mannitol utilization system are not strictly associated with macroalgae suggests that they have probably conserved this operon to degrade mannitol from other marine resources.

The occurrence of two types of mechanisms for the transport of mannitol (the ABC transporter is primary and the permease secondary) through the inner membrane in Bacteroidetes also suggests that the physiological role(s) and/or the regulation of mannitol catabolism may differ among members of this phylum. Meanwhile, it is difficult to relate these observations to any ecophysiological traits of the nine strains under consideration because there is no clear relationship between the type of mannitol utilization operon and the origin and metabolism of the strain. Another interesting subject for further analysis is the identification of the mechanism(s) used by these bacteria to sense and transport mannitol through the outer membrane.

For further understanding of mannitol catabolism and its physiological importance in Z. galactanivorans, it will be interesting to study its regulation. In general, changes in the expression of genes involved in carbohydrate catabolism occurs via one (or both) of two mechanisms: catabolic repression (suggesting the existence of a preferential carbon source) and specific induction by a substrate(s). Usually, transcription of catabolic genes in prokaryotes is regulated by either a repressor or an activator protein. In most cases, genes encoding this type of regulator are located near the targeted genes, likely in the same cluster/operon, although there are exceptions. For instance, in P. fluorescens DSM 50156, the gene coding for an activator protein (MtlR) belonging to the Xyl/AraC family and involved in the regulation of the mannitol catabolic pathway is not localized in the region containing all the genes involved in this catabolism (29). The genome of Z. galactanivorans contains several putative AraC-type transcriptional regulators, but none of them are located close to the mannitol utilization cluster. A reverse genetic approach will be very valuable for assessing the physiological role(s) of these potential regulators when this methodology becomes available for Z. galactanivorans.

Reconstruction of metabolic networks from genomic resources, including carbohydrate catabolic pathways, may contribute to better understanding of microbial ecophysiology and interactions between microorganisms and their hosts, in particular for alga-associated bacteria (84, 85). In this context, our results shed light on the evolution of mannitol catabolism in important environmental bacteria. They pave the way also to a better understanding of the recycling of algal biomass on the shore and to the exploitation of algae as a renewable resource.