Cross-utilization of β-galactosides and cellobiose in Geobacillus stearothermophilus

Abstract

Strains of the Gram-positive, thermophilic bacterium Geobacillus stearothermophilus possess elaborate systems for the utilization of hemicellulolytic polysaccharides, including xylan, arabinan, and galactan. These systems have been studied extensively in strains T-1 and T-6, representing microbial models for the utilization of soil polysaccharides, and many of their components have been characterized both biochemically and structurally. Here, we characterized routes by which G. stearothermophilus utilizes mono- and disaccharides such as galactose, cellobiose, lactose, and galactosyl-glycerol. The G. stearothermophilus genome encodes a phosphoenolpyruvate carbohydrate phosphotransferase system (PTS) for cellobiose. We found that the cellobiose-PTS system is induced by cellobiose and characterized the corresponding GH1 6-phospho-β-glucosidase, Cel1A. The bacterium also possesses two transport systems for galactose, a galactose-PTS system and an ABC galactose transporter. The ABC galactose transport system is regulated by a three-component sensing system. We observed that both systems, the sensor and the transporter, utilize galactose-binding proteins that also bind glucose with the same affinity. We hypothesize that this allows the cell to control the flux of galactose into the cell in the presence of glucose. Unexpectedly, we discovered that G. stearothermophilus T-1 can also utilize lactose and galactosyl-glycerol via the cellobiose-PTS system together with a bifunctional 6-phospho-β-gal/glucosidase, Gan1D. Growth curves of strain T-1 growing in the presence of cellobiose, with either lactose or galactosyl-glycerol, revealed initially logarithmic growth on cellobiose and then linear growth supported by the additional sugars. We conclude that Gan1D allows the cell to utilize residual galactose-containing disaccharides, taking advantage of the promiscuity of the cellobiose-PTS system.

Geobacillus stearothermophilus T-1 is a thermophilic, Gram-positive, soil bacterium, which is capable of utilizing plant cell wall–derived polysaccharides, including xylan, arabinan, and galactan (1–3). Utilization of these polysaccharides includes extracellular and intracellular hemicellulolytic enzymes, ABC sugar-transport systems, carbohydrate-sensing systems, sugar metabolism enzymes, and regulatory proteins (4–8). A major challenge for soil bacteria in the natural environment is to sense the scarce carbon sources and compete for these sources with nearby microorganisms. Indeed, we have demonstrated that G. stearothermophilus employs a unique strategy for the efficient utilization of polysaccharides in its immediate environment. First, it utilizes two- or three-component systems to sense minute amounts of mono- or disaccharides in the surroundings, which signal the presence of the corresponding polysaccharides (2, 9). The two- or three-component sensing systems then activate dedicated ABC sugar transporters that transfer the sugars into the cell and induce the corresponding systems for expressing extracellular endo-type glycoside hydrolases that partially degrade the high-molecular-weight polysaccharides into short (usually decorated) oligosaccharides. Additional ABC transporters for oligosaccharides transfer the large oligosaccharides into the cell, and those are further hydrolyzed into sugar monomers by a battery of specific intracellular glycoside hydrolases (10–16). This utilization strategy allows the bacterium to (a) react rapidly to presence of potential polysaccharides such as xylan, arabinan, and galactan in the immediate environment; (b) transfer efficiently the degradation products into the bacterium cell; and (c) almost exclusively utilize the degraded decorated oligosaccharides because these are rarely imported by other organisms. Essentially the same strategy was recently demonstrated for the utilization of yeast mannan by the human gut Bacteroidetes and was coined a “selfish mechanism” (17).

In addition to the complete utilization machinery for xylan, arabinan, and galactan, G. stearothermophilus also possesses scavenging mechanisms for the utilization of mono- or disaccharides that are often found in the surroundings, resulting from the extensive degradation of the corresponding polysaccharides by other soil microorganisms. In this case, the sugars are imported by the bacterium into the cell via a different type of transporters, the phosphoenolpyruvate-dependent phosphotransferase systems (PTS). The PTS systems usually use phosphoenolpyruvate as the phosphoryl donor for sugar phosphorylation, together with three essential catalytic entities, termed enzyme I, enzyme II, and HPr (heat-stable, histidine-phosphorylatable protein) (18). During their import via the PTS systems, sugars are simultaneously phosphorylated at the C6 hydroxyl group of the terminal sugar unit at the nonreducing end and are further cleaved (inside the cell) by dedicated 6-phospho-β-glycosidases/galactosidases (19).

In the present study, we identified in G. stearothermophilus strain T-1 such PTS systems for cellobiose and galactose and biochemically characterized the corresponding enzymes involved. These are the 6-phospho-β-gal Gan1D, belonging to glycoside hydrolase (GH) family GH1, and the 6-phospho-β-glycosidase Cel1A, belonging also to family GH1. In addition, we demonstrate that the bacterium can utilize the β-galactosides lactose and galactosyl-glycerol, using the cellobiose-PTS system together with the bifunctional 6-phospho-β-gal, Gan1D.

Results

Identifying of the cellobiose utilization gene cluster

As part of a systematic search for hemicellulolytic utilization systems in G. stearothermophilus, we identified in strain T-1 a 4.3-kb gene cluster, which appears to be involved in the utilization of cellobiose (Fig. 1A). The cluster was identified via bioinformatics analysis of the genome sequence of strains T-1 and includes an operon of four genes and a transcriptional regulator gene. Based on sequence homology, the celBCD genes encode for a complete PTS system. The CelD protein is homologous to proteins from the lactose/cellobiose EIIC family that bind cellobiose or lactose. CelC and CelB are the EIIA and EIIB domains, respectively, which transfer the phosphoryl group from HPr to the transported sugar. The fourth gene in this operon, celA, encodes for a 6-phospho-β-glucosidase, Cel1A, which shows sequence homology to a number of GH1 6-phospho-glucosidases, including those from Streptococcus pyogenes (51% sequence identity), Lactobacillus plantarum (32%), and Streptococcus mutans (32%). Adjacent to the operon lies the celR gene encoding for a transcriptional regulator belonging to the GntR family. The genes for the EI and HPr proteins are located elsewhere on the chromosome.

Figure 1.

Schematic representation of the cellobiose-PTS system and mRNA expression levels of the related genes. A, the gene cluster encodes for the cellobiose-PTS enzyme system, celBCD, a GH1 6-phospho-glucosidase, celA, and a GntR family transcriptional regulator, celR. The letter P indicates the proposed promoter site, and Ω indicates the position of the putative ρ-independent transcription termination site. B, real-time RT-PCR analysis of the cellobiose utilization gene cluster. Total RNA was extracted from mid-logarithmic phase cultures of G. stearothermophilus grown in minimal medium supplemented with either 0.5% cellobiose, xylose, or glucose as the sole carbon source. Normalization was performed using the housekeeping gene ict, which encodes for isocitrate dehydrogenase. Error bars, S.D.

mRNA expression of cellobiose-utilizing genes

G. stearothermophilus T-1 is unable to grow on cellulose but grows very well on cellobiose. To test whether the celBCD, celA and celR genes are involved in cellobiose utilization, we measured the corresponding mRNA levels in cultures grown on cellobiose, xylose, or glucose. Total mRNA was extracted from mid-exponential phase cultures, and the cDNA was amplified with primers specific to celA, celB, celC, celD, and celR as well as for the isocitrate dehydrogenase gene, ict, that was used for normalization. Relative expression was measured by real-time RT-PCR, as presented in Fig. 1B. The expression levels of the corresponding genes were 7–15-fold higher in the cellobiose-grown culture, as compared with cultures grown on xylose or glucose. The relatively high levels of celA mRNA compared with the celBCD genes may reflect differences in mRNA stability. These results suggest that the celBCDA operon is induced by cellobiose and thus highly likely to be involved in cellobiose utilization.

Biochemical characterization of Cel1A

The biochemical activity of Cel1A was determined using different chromogenic and natural substrates (Table 1). Cel1A exhibited a significant catalytic activity on substrates with glucose 6-phosphate at the glycon moiety and had no detectable activity toward unphosphorylated substrates. These results suggest that Cel1A is a 6-phospho-β-glucosidase, with high specificity toward glucose 6-phosphate as the glycon moiety. The effect of temperature on Cel1A activity was determined at pH 7.0, using the chromogenic substrate oNP-β-d-glucopyranoside 6-phosphate (oNPβGlc6P). The optimal temperature in a 20-min reaction was 65 °C (Fig. 2A). The thermal stability was determined after incubating Cel1A at temperatures between 30 and 85 °C for 10 min The enzyme was stable at temperatures below 65 °C and lost over 90% of its activity at 75 °C (Fig. 2B). The activity of Cel1A was determined at different pH values in the range of 3.5–9.5 (Fig. 2C). The pH profile is a typical bell-shaped curve, which may reflect the ionization of the two catalytic carboxylates. The enzyme was most active at the pH range of 6.5–8 and lost about 90% of its activity at pH 3.5 and 9.5. Based on sequence alignment of Cel1A with other retaining GH1 enzymes, the putative acid/base and nucleophile catalytic residues are Glu-174 and Glu-373, respectively. These residues were substituted with alanine, and the effect of the mutations on the activity of the enzyme was determined using oNPβGlc6P as the substrate. The Michaelis–Menten catalytic constants of the Cel1A-E174A catalytic mutant toward oNPβGlc6P were 0.01 mm, 2.1 s⁻¹, and 2.1 × 10⁵ s⁻¹ m⁻¹ for K_m, k_cat, and k_cat/K_m, respectively. The k_cat value was about 50-fold lower, compared with the WT, and the K_m was 10-fold lower, compared with the WT, suggesting the accumulation of a glycosyl-enzyme intermediate. Such results are expected for an acid-base mutant because the relatively high reactivity of the o-nitrophenol leaving group (pK_a 7.22) elevates the rate of the first glycosylation step even without a proton assistant. The second deglycosylation step remains considerably slow due to the loss of the base required for the activation of the catalytic water molecule (20). The alanine replacement of the Glu-373 nucleophile resulted in a nondetectable activity toward oNPβGlc6P, reflecting the inability of this mutant to bind the substrate and stabilize the first transition state. In similar retaining glycoside hydrolases, the nucleophile is usually involved in a strong interaction with the C2 hydroxyl group of the glycone sugar (21).

Table 1.

Michaelis–Menten catalytic constants for the hydrolysis of synthetic and natural substrates by Cel1A. The reactions were performed at 40 °C, in 100 mm citric acid-Na₂HPO₄ buffer, pH 7

Substrate	kcat (s−1)	Km (mm)	kcat/Km (s−1·M−1)
oNPβGlc6P	106	0.22	4.8 × 105
oNPβGal6P	NDa	ND	ND
Cellobiose 6-phosphate	60	5.1	1.2 × 104
Lactose 6-phosphate	ND	ND	ND
oNPβGlc	ND	ND	ND

^aND, nondetectable.

Figure 2.

Effect of temperature and pH in a 20 min reaction on Cel1A activity toward oNPβGlc6P. A, relative activity at pH 7 at different temperatures. B, thermostability of Cel1A at pH 7. Residual activity was measured after a 10-min incubation at the indicated temperatures. C, initial rate activity at different pH levels. The reactions were performed at 40 °C in either sodium phosphate-citric acid buffer (pH 3–7.8) or Atkins and Pantin buffer (pH 7.5–9.5).

Identification of the sensing and transport systems for galactose

We have demonstrated previously that G. stearothermophilus T-1 consumes galactan efficiently, utilizing the galactan utilization gene cluster, ganREFGBA (3). Following bioinformatics analysis of the bacterium genome sequence, we identified a new cluster that appears to be involved in galactose utilization. This new 12.5-kb cluster is composed of genes encoding for a putative three-component regulatory sensing system (GalPST2), an ABC sugar transport system (GalE2F2G2), a regulatory protein (GanR2), and two 6-phospho-glycosidases (Gan4C and Gan1D) (Fig. 3).

Figure 3.

Schematic representation of the 12.5-kb cluster containing galactose utilization genes. The cluster contains a Tn31 transposon between gan4C and galP. The letter P indicates the proposed promoter site, and Ω indicates the position of the putative ρ-independent transcription termination site.

GalP exhibits significant sequence similarity to periplasmic sugar-binding proteins with a 27-amino acid signal peptide at its N terminus. GalS exhibits characteristic features of bacterial histidine kinase proteins, including two transmembrane (TM) helices (TM1, residues 7–25; TM2, residues 177–197) flanking an extracellular domain (residues 26–176) and a conserved C-terminal cytoplasmic region containing the ATP-binding kinase domain. Downstream from the galS gene lies the galT2 gene, which encodes for a protein with high sequence similarity to response regulators. As in the case for many response regulators, GalT2 has a predicted two-domain architecture, with an N terminus signal receiver domain (REC), linked to a C terminus effector domain (22). The N-terminal signal receiver domain (positioned at residues 5–118) shares a significant homology with the Rec superfamily of the REC domains, whereas the C-terminal domain (residues 157–225) contains a putative helix-turn-helix motif, resembling the AraC-type DNA-binding domain (23). Taken together, these observations suggest that the galPST2 and galE2F2G2 clusters encode for a three-component sensing system and an ABC galactose transport system, respectively.

mRNA expression levels of the galactose-related genes

Based on genome sequence analysis, strain T-1 encodes for four sugar-specific PTS systems, two for the disaccharides, cellobiose and trehalose, and two for the monosugars, mannitol and galactose (Fig. 4A). To determine whether the galPST2 and galE2F2G2 gene clusters are involved in galactose utilization and to identify the PTS system for galactose, the mRNA levels of the galPST2, galE2F2G2, and the four sugar-PTS systems were measured in cultures grown in the presence of galactose and related sugars (Fig. 4). The expression of galE2 was 5-fold higher on galactose-grown cultures, compared with cultures grown on either glucose, xylose, or galactan (Fig. 4B). These results support our original suggestion that the galE2F2G2 operon constitutes an ABC galactose transporter. The expression of the galP gene, which is part of a three-component sensing system, appears to be relatively low and constant, regardless of the carbon source used (Fig. 4B). Although these results cannot correlate the system to galactose, it is expected that the sensing systems will be expressed constitutively. The mRNA levels of the four PTS systems were also measured in cultures grown on either galactose or glucose (Fig. 4C). The expression level of the ptsA gene in galactose-grown culture was about 9-fold higher than for a culture grown on glucose, suggesting that ptsA is part of the galactose-PTS transporter, dedicated to galactose import. Taken together, these results suggest that G. stearothermophilus T-1 has two transport systems for galactose.

Figure 4.

Presentation of four putative PTS systems in strain T-1 and the expression of galactose-related genes in the presence of galactose. A, schematic gene organization of the PTS systems for cellobiose, galactose, trehalose, and mannitol. Cellobiose-PTS enzyme II complex is encoded by celB, celC, and celD; ptsABC ORFs encode for a PTS transporter for galactose; ptsD encodes for trehalose-PTS EIIBC; and mltE encodes for manitol PTS EIIBC. The celR and gntR genes encode for transcriptional regulators; glcT and blgG encode for anti-terminator proteins; celA encodes for 6-phospho-β-glucosidase; amyA encodes for trehalose-6-phosphate hydrolase; and mltD encodes for mannitol-1-phosphate 5-dehydrogenase. B, real-time RT-PCR analysis of the galactose-related gene cluster in G. stearothermophilus T-1. C, real-time RT-PCR analysis of the four putative PTS systems in strain T-1. Total RNA was extracted from mid-logarithmic phase cultures of G. stearothermophilus grown in minimal medium supplemented with different sugars (0.5%) as the sole carbon source. Normalization was performed using the housekeeping gene for isocitrate dehydrogenase, ict. Error bars, S.D.

GalP and GalE2 bind galactose

The putative three-component system for galactose, GalPST2, and the galactose ABC transport system, GalE2F2G2, both have dedicated sugar-binding proteins (GalP and GalE2, respectively) that are tethered to the local membrane. The ability of these proteins to bind galactose was confirmed by isothermal titration calorimetry (ITC). The calorimetric titration curves are shown in Fig. 5, and the thermodynamic binding parameters are summarized in Table 2. These results demonstrate that GalE2 and GalP are able to bind galactose quite tightly, with dissociation constants in the micromolar range, K_D = 1.4 µm and K_D = 6.1 µm, respectively. Similar titrations of GalP with cellobiose and lactose did not result in a significant enthalpy change. Surprisingly, however, both proteins bind glucose in similar affinities to galactose. Nevertheless, although glucose seems to bind these proteins quite tightly, it is rather unlikely that it can activate the corresponding sensing systems, because the expression of the operons of these systems is practically unaffected by glucose (Fig. 4B).

Figure 5.

Calorimetric titration of GalE2 and GalP with galactose or glucose. The top half of each experiment shows the raw data for calorimetric titration of the protein with the ligand, and the bottom half displays the integrated injection heats from the top panel. The solid line is the curve of best fit to a single binding site (n = 1) model that was used to derive the binding parameters. Experiments were performed at 40 °C.

Table 2.

Thermodynamic parameters for the binding of galactose or glucose to GalE2 and GalP

Protein	Ligand	n	KD (µm)	ΔHB (kcal mol−1)	TΔSB (kcal mol−1)	ΔGB (kcal mol−1)
GalE2	Galactose	1.12	1.4	−5.8 ± 0.3	0.3	−6.1
	Glucose	1.04	3.1	−7.0 ± 0.4	0.8	−7.8
GalP	Galactose	0.65	6.1	−10 ± 0.6	−2.6	−12.6
	Glucose	0.71	4.2	−9.0 ± 0.2	−1.3	−10.2

GalT2 binds to the galE2F2G2 promoter region

The fact that the potential three-component sensing system for galactose (galPST2) is located adjacent to a putative ABC transport system (galE2F2G2) suggests that the sensing system is functionally linked to this transport system and that GalT2 is therefore a response regulator that regulates the expression of the transporter. This type of adjacent arrangement of gene clusters was found in G. stearothermophilus also for the arabinose and xylotriose utilization systems (2, 9). Sequence analysis of the promoter region of galE2 revealed a putative −35 sequence (TTGATA), which is a relatively close match to the σ^A consensus sequence, TTGACA. This −35 sequence is separated by 18 bp from the potential −10 region (CAACAT), which differs from the Bacillus subtilis consensus, TATAAT, by three nucleotides (24). The upstream region of the −35 of the galE2 promoter contains two direct repeats, CAAAAAAGT, separated by 11 bp, which may function as the recognition sequences for the response regulator GanlT2 (Fig. 6A). This putative binding site, upstream of the –35 region, can allow direct interaction of the activator with the C-terminal domain of the α subunit of RNA polymerase (25, 26). To test whether GalT2 can bind the galE2 promoter region, we utilized gel mobility shift assays. These studies indicated that the GalT2 protein (using N-His₆-GalT2) can bind significantly to a 113-bp DNA fragment containing the two direct repeats discussed above, and a nearly complete shift was obtained in the presence of 0.3 µm GalT2 (Fig. 6B). This relatively high concentration of GalT2 (reflecting a relatively weak protein-DNA binding) may originate from the fact that the protein was not fully phosphorylated, as often observed for other response regulators (27). To test whether the phosphorylation of GalT2 increases significantly its binding to DNA, we phosphorylated GalT2 in vitro, using Mg²⁺ and acetyl phosphate as the phosphate donor. As shown by size-exclusion chromatography, such phosphorylation changed the oligomeric state of GalT2 in solution, transforming it from a monomer (not phosphorylated) to a dimer (phosphorylated) (Fig. 6C), as usually observed for related response regulators. Indeed, the fully phosphorylated GalT2 protein (in its dimeric form) gave a complete shift at 0.07 µm (Fig. 6D), demonstrating a significant increase in its DNA-binding capabilities. These results further support the identification of GalT2 as a response regulator, as such regulators usually change their conformation upon phosphorylation, allowing them to enhance their binding to their target DNA segments (28, 29).

Figure 6.

Gel shift assays demonstrating binding of unphosphorylated and phosphorylated His₆-GalT2 to the galE2 promoter. A, promoter region of galE2. The horizontal arrows indicate repeat binding sites for the response regulator GalT2. The −35 and −10 regions, the ribosome-binding site (RBS), and the ATG-initiating codon are in boldface type. The sequences of the primers used to synthesize a PCR product of 113-bp DNA fragment the for the gel shift assay are underlined. B, fluorescently labeled 113-bp DNA fragment containing the galE2 promoter was incubated with increasing amounts of His₆-GalT2. C, size-exclusion chromatograms (overlaid) of unphosphorylated His₆-GalT2 (solid black line) and phosphorylated His₆-GalT2 (solid gray line). The inset shows the linear regression of the protein standards used for M_r determination (filled circles), and the positions of the dimeric and monomeric forms of GalT2 on this curve are marked with arrows. D, the fluorescently labeled 113-bp DNA fragment was incubated with increasing amounts of phosphorylated His₆-GalT2.

Biochemical characterization of Gan1D

The 12.5-kb galactose utilization cluster also contains the gan1D gene, which is expressed in cultures grown on galactose or galactan (Fig. 4B). We have previously described the 3D crystal structure of Gan1D, as well as its catalytic mutants complexed with substrates and products (30). These Gan1D structures revealed the structural features that allow the enzyme to accommodate both 6-phospho-galactose and 6-phospho-glucose at the glycone-binding site (−1) (30). To further characterize the specificity of Gan1D, its catalytic activity was studied using different chromogenic and natural substrates (Table 3). Gan1D readily hydrolyzed substrates with a glycoside 6-phosphate as the glycone moiety, such as lactose 6-phosphate. The enzyme had no detectable activity toward nonphosphorylated substrates, in a good correlation with similar studies conducted on other GH1 6-phospho-β-glycosidases (20, 31). Interestingly, however, the enzyme exhibited bifunctional specificity, showing similar catalytic activities on substrates containing glucose 6-phosphate or galactose 6-phosphate as their glycone moiety (Table 3).

Table 3.

Michaelis–Menten catalytic constants for the hydrolysis of synthetic and natural substrates by Gan1D. The reactions were performed at 40 °C in 100 mm citric acid-Na₂HPO₄ buffer, pH 6.5. All measurements were performed in duplicate, and the relative error was less than 5%

Substrate	kcat (s−1)	Km (mm)	kcat/Km (s−1 m−1)
oNPβGal-6P	148	0.32	4.9 × 105
oNPβGlc-6P	143	0.17	8.4 × 105
pNPβGlc-6P	114	0.12	9.3 × 105
Cel-6P	74	0.86	8.6 × 104
Lac-6P	25	1.1	2.3 × 104
Gen-6P	12	2.9	4.1 × 103
oNPβGal	NDa	ND	ND
oNPβGlc	ND	ND	ND

^aND, nondetectable.

Growth of strain T-1 on β-galactosides substrates in the presence of cellobiose

The physiological role of Gan1D in G. stearothermophilus T-1 was initially puzzling because we failed to identify a dedicated PTS system for galacto-disaccharides. Indeed, neither lactose nor galactosyl-glycerol was found to support the growth of G. stearothermophilus T-1. Lactose is not expected to be abundant in soil, the natural niche of the bacterium; however, galactosyl-glycerol, the degradation product of galactolipids, should be available there. The most abundant galactolipids in plants are those of monogalactosyl-diacyl-glycerol and digalactosyl-diacyl-glycerol, mainly found in plant chloroplast membranes (32, 33). The degradation of both of these molecules by lipases releases galactosyl-glycerol, a β-galactoside bound to glycerol (1-0-β-galactosyl-glycerol). Interestingly, unlike the genetic context of many phospho-glycosidases, the gan1D gene in strain T-1 as well as its homologous enzymes in many other bacteria do not lie adjacent to the genes of the corresponding PTS system for disaccharides. Considering these observations, we hypothesized that the utilization of β-galactosides (lactose or galactosyl-glycerol) in strain T-1 is linked to an alternative PTS system. Somewhat supporting evidence for that is the observation that Lactococcus lactis IL1403, can utilize lactose only in the presence of cellobiose (34). In analogy, it was tempting to speculate that the utilization of β-galactosides (lactose or galactosyl-glycerol) in G. stearothermophilus T-1 is actually dependent on its cellobiose-PTS system (celBCD). To test this hypothesis, strain T-1 was grown in a defined modified basic salt medium (mBSM), containing either 0.4% lactose or galactosyl-glycerol in the presence of various concentrations of cellobiose. The resulting growth curves were characterized with an initial exponential growth on cellobiose, followed by a linear growth on lactose or galalctosyl-glycerol (Fig. 7, A and B). The extent of the logarithmic growth was proportional to the initial cellobiose concentration, suggesting that in this phase, it is cellobiose that is mostly consumed. Following the logarithmic phase, the cultures exhibited linear growth, this time likely consuming lactose or galactosyl-glycerol. The slopes of the linear growth were proportional to the turbidity (cell mass) at the end of the logarithmic phase (see insets in Fig. 7, A and B). These unexpected results can be explained by the following behavior of the bacterium. At the end of the logarithmic phase, it seems that all of the cellobiose is already consumed, and at that stage, the cells do not express the now noninduced cellobiose-PTS genes (celBCD). Hence, at the end of the exponential growth, the number of PTS systems per cell seems to remain constant, neglecting obvious natural turnover. At that point, the limiting factor of the growth rate is likely to be the transport of lactose or galactosyl-glycerol into the cell, probably via the already fixed population of the cellobiose-PTS systems. This hypothesis seems to be in good correlation with the linear cell growth observed from this point on, which is now proportional to the (now unchanged) overall number of cellobiose-PTS systems per cell.

Figure 7.

Growth of G. stearothermophilus strain T-1 and T-6 on lactose or galactosyl-glycerol in the presence of cellobiose. A, cellobiose-induced logarithmic culture at 0.4 OD₆₀₀ was washed twice with mBSM and then diluted 10-fold in a 96-well plate containing mBSM supplemented with 0.4% lactose and various concentrations of cellobiose. Cellobiose concentrations were 0.01% (▵), 0.02% (☐), 0.04% (⧫), 0.05% (■), 0.1% (●), and 0.2% (▲) cellobiose. mBSM supplemented only with 0.1% cellobiose (○) or 0.4% lactose (◊) as a sole carbon source is shown. B, mBSM supplemented with 0.4% galactosyl-glycerol and various concentrations of cellobiose. Cellobiose concentrations were 0.05% (■), 0.1% (▲), and 0.2% (●). C, growth of G. stearothermophilus T-6 (lacks the gan1D gene) on mBSM supplemented with 0.4% lactose and various concentrations of cellobiose. Concentrations of cellobiose were 0.01% (▵), 0.02% (☐), 0.04% (⧫), 0.1% (●), and 0.2% (▲). The path length in the wells was 0.6 cm. The insets show the growth rate (dx/dt) on lactose or galactosyl-glycerol as a function of final cell mass grown on cellobiose.

Taking these observations together, it appears as if G. stearothermophilus T-1 is capable of utilizing galacto-disaccharides using the alternative cellobiose-PTS system for import and the Gan1D enzyme for processing. In support of this hypothesis, it is noted that G. stearothermophilus T-6, a close relative of strain T-1, possesses a similar cellobiose-PTS system but lacks a 1825-bp segment containing the gan1D homologous gene. In turn, this strain grows very well on cellobiose yet fails to grow on lactose or galactosyl-glycerol in the presence of cellobiose (Fig. 7C).

Discussion

G. stearothermophilus possesses cellobiose- and galactose-PTS systems

G. stearothermophilus encodes for four putative PTS systems, dedicated to the import of cellobiose, galactose, mannitol, and trehalose. The cellobiose-PTS system was identified based on the observation that the celBCDA operon is expressed in the presence of cellobiose (Fig. 1B) and that Cel1A rapidly hydrolyzes 6-phospho-cellobiose (Table 1). The cellobiose-PTS system is most likely regulated by CelR. A sequence resembling an operator site for CelR includes an inverted repeat of 8 bp (5′-TTTTTATT-N₁₃N-AATAAAAA-3′), and it is located within the promoter region of the celBCDA operon. G. stearothermophilus can readily utilize cellobiose, although it does not encode for cellulolytic genes. Thus, this bacterium seems to possess the ability to scavenge from the surroundings residual cellobiose, which is the main product of cellulose degradation by cellulolytic microorganisms.

The PTS system for galactose was identified in G. stearothermophilus based on the measurements of mRNA levels of the ptsA gene in the presence of galactose. The system is encoded by a four-gene operon, consisting of a multidomain EII protein, with the corresponding domains of EIIA, EIIB, and EIIC fused together (ptsA), an HPr-like protein (ptsB), and a phosphoenolpyruvate-protein kinase (the EI component) (ptsC). The first gene in the operon is glcT, encoding for an anti-terminator protein of the BlgG family. Both genes (glcT and ptsA) are separated by a palindromic sequence that functions as a transcriptional terminator. GlcT possesses an N-terminal RNA-binding domain, followed by two PTS regulation domains, which are likely to be phosphorylated and thereby modulate the RNA binding activity (35, 36). Based on these observations, the galactose-PTS system of G. stearothermophilus seems to be regulated by the anti-terminator GlcT.

In G. stearothermophilus strain T-1, galactose can also be imported by a non-PTS ABC galactose transporter (galE2G2F2) (Fig. 8). The operon is up-regulated in the presence of galactose, and the corresponding substrate-binding protein (SBP), GalE2, was shown to have a significant galactose binding (Table 2). In L. lactis, the import of galactose occurs via a high-affinity non-PTS system (K_d in the micromolar range), whereas the galactose-PTS system is a low-affinity transporter (K_d in the millimolar range). The specific binding affinity of the strain T-1 galactose-PTS system is yet to be determined; however, our current data indicate that GalE2 binds galactose with a relatively high affinity, with K_d values in the micromolar range (Table 2).

Figure 8.

Proposed pathways for the utilization of galactan, galactose, cellobiose, and β-galactosides in G. stearothermophilus T-1. The extracellular β-1,4-galactanase Gan53A cleaves galactan and generates short galacto-oligosaccharides that enter the cell via a specific ABC transport system (GanEFG) and are further degraded by the intracellular β-galactosidase Gan42B into galactose (3). Galactose is metabolized via the Leloir pathway involving the galKETR operon encoding for galactokinase (galK), UDP-glucose 4-epimerase (galE), galactose-1-phosphate uridyltransferase (galT), and a transcriptional regulator (galR). Extracellular galactose binds the high-affinity sugar-binding protein GalP, which presents the sugar to the sensor histidine kinase (GalS). GalS phosphorylates the response regulator (GalT2), which in turn binds to the galE2 promoter and activates the expression of the ABC transport system for galactose (GalE2F2G2). In addition, galactose can be imported by a specific PTS for galactose (PtsABC). Inside the cell, the phosphorylated galactose (galactose 6-phosphate) is further metabolized via the tagatose 6-phosphate pathway to dihydroxyacetone phosphate and glyceraldehyde 3-phosphate or, alternatively, hydrolyzed by a phosphatase to galactose and P_i. Strain T-1 encodes for homologous tagatose pathway genes, including galactose-6-phosphate isomerase (galA), tagatose-6P kinase (galB), and tagatose 1,6-bisphosphate aldolase (galC). Cellobiose is imported to the cell by a dedicated PTS (celBCD) and is further hydrolyzed inside the cell by 6-phospho-β-glucosidase (Cel1A) to glucose 6-phosphate and glucose. Other galacto-moiety disaccharides, including lactose and galactosyl-glycerol, can be also imported by the PTS for cellobiose. Lactose and galactosyl-glycerol are further hydrolyzed by Gan1D to galactose 6-phosphate and glucose or glycerol, respectively. The letter P on the genetic map indicates proposed promoter sites, and Ω indicates the position of hypothetical ρ-independent transcription termination sites.

A three-component sensing system regulates the expression of the galactose ABC transporter

The galP, galS, and galT2 gene products constitute a three-component sensing system, in which GalP is a galactose-binding lipoprotein, GalS is a class-I histidine kinase, and GalT2 is the response regulator, which seems to be phosphorylated by GalS (Fig. 8). In good correlation with similar sensing systems, GalP binds extracellular galactose and is likely to activate the sensor histidine kinase, either by direct protein-protein interaction or by presenting the sugar to the histidine kinase sensor domain. In turn, the autophosphorylated GalS histidine kinase phosphorylates the response regulator, GalT2, thereby passing the sensing signal into the cell. Thus, GalP seems to be a typical SBP, functioning here as an auxiliary component that assists this three-component system in sensing small amounts of extracellular galactose.

GalT2 seems to function as a typical response regulator and accordingly possesses a phospho-acceptor REC, a common module in a variety of response regulators (27). Interestingly, many response regulators are able to catalyze their own phosphorylation in vitro in the presence of Mg²⁺ and a suitable phosphoryl donor and in the absence of any auxiliary proteins (37). Phosphorylation of the REC domains usually induces dimerization of the response regulator, commonly associated with conformational changes, thereby facilitating its binding to the target DNA segment (usually palindromic or direct repeat sequences) (38, 39). Indeed, we were able to show that in vitro autophosphorylation of GalT2 promoted its dimer formation in solution and facilitated its ability to bind DNA. The potential binding sites for GalT2 are located upstream of the −35 region of galE2 (the first gene in the galactose ABC transporter operon), which is made of two direct repeat sequences, CAAAAAAGT, separated by 11 nucleotides. A binding upstream of the −35 region usually enables a direct interaction of the bound activator with the α C-terminal domain of the corresponding RNA polymerase (26). Direct repeats are common binding sites for activators (38, 39), and in G. stearothermophilus such sequences are found in the upstream region of xynE (a part of the xylotriose ABC transporter) and araE (a part of the arabinose ABC transporter) (2, 9).

Three-component sensing systems for monosugars are likely to provide the bacterium with a highly sensitive and rapidly responding mechanism for detecting and utilizing potential high-molecular-weight polysaccharides in the extracellular environment. Such a sensing system was recently characterized in Clostridium beijerinckii, dedicated to the utilization of xylose (40). A somewhat different mechanism for sensing was found in Clostridium thermocellum for the regulation of cellulosomal genes. This mechanism is based on a transmembranal RsgI-like protein, which is made of an extracellular carbohydrate-binding module and an intracellular anti-σ peptide. In such systems, the binding of the target saccharide to the carbohydrate-binding module sensing domain leads to the release of the corresponding alternative σ factor, thereby enabling the expression of the corresponding genes (41, 42).

The 6-phospho-galactose metabolism in G. stearothermophilus T-1

In principle, 6-phospho-galactose and 6-phospho-lactose can be metabolized by the tagatose 6-phosphate (Tag6P) pathway, whereas galactose can be also metabolized via the Leloir pathway (43). In G. stearothermophilus T-1, d-galactose 6-phosphate is formed after the transport of lactose or galactose via the corresponding PTS systems (Fig. 8). d-Galactose 6-phosphate is catabolized by three enzymes of the Tag6P pathway, resulting in dihydroxyacetone phosphate and d-glyceraldehyde 3-phosphate (Fig. 8). Using sequence alignment, we identified in G. stearothermophilus T-1 homologous genes for these three enzymes, tagatose isomerase, tagatose-6-phosphate kinase, and d-tagatose-1,6-bisphosphate aldolase. Interestingly, these three genes are not clustered together, as common in other Gram-positive bacteria (44, 45). The putative tagatose aldolase gene from strain T-1 shares 40% identity with the tagatose-bisphosphate aldolase gene (gatY), which in Bacillus licheniformis is part of a five-gene cluster involved in the d-tagatose pathway (46). Although such bioinformatics analysis suggests a similar tagatose pathway in Geobacillus, we cannot rule out the involvement of a phosphatase, which, in principle, could alternatively catalyze the hydrolysis of Gal-6P into galactose and phosphate (47).

Regulation of the galactose-related genes

In G. stearothermophilus, regulation of the galactose utilization-related genes appears to involve several regulatory mechanisms. As noted above, the galactose-PTS system is regulated by the anti-terminator transcriptional protein GlcT, and the expression of the ABC galactose transporter (galE2G2F2) is positively regulated by the response regulator GalT2. Surprisingly, the sugar-binding protein of the galactose ABC transport system (GalE2) binds with similar affinities to both galactose and glucose, as was also observed for GalP, the SBP component of the galactose-sensing system. These observations could suggest that glucose competes with galactose over the binding to both GalP and GalE2, thereby preventing the expression of galactose-related genes when the glucose concentration is relatively high. Such glucose-galactose competition will ensure that the activation of the galactose utilization system occurs only when the glucose levels are relatively low, following a mechanism similar to “inducer exclusion” (48). In Escherichia coli, the nonphosphorylated form of the EIIA protein, a part of the glucose-PTS system, is known to inhibit sugar transporters, exerting the known “glucose effect” (48). In support of this hypothesis is the absence of catabolic responsive elements in the ganE2 promoter, which usually involve the binding of the catabolite control protein A (CcpA). CcpA is known to be a major regulator of carbon metabolism in Gram-positive bacteria (49, 50).

The regulation of the gan1D gene probably involves the negative regulator GanR. We have previously characterized the galactan utilization cluster (ganEFGBA). This operon is regulated by GanR, which binds to an invert repeat upstream of the −35 site (3). Galacto-oligosaccharides, although not galactose, function as the molecular inducers. Very similar binding sequence is located in the gan1D promoter region, suggesting that gan1D may also be regulated by GanR. Interestingly, almost the same sequence appears in the promoter region of the galactose metabolism operon, galKETR (the Leloir pathway), which is most likely regulated by a different negative regulator, GalR.

Gan1D is a bifunctional 6-phospho-β-gal/glucosidase

In the GH1 family, two catalytic specificities toward phosphorylated substrates have been reported so far, those of 6-phospho-glucosidase and 6-phospho-galactosidase. Glucose 6-phosphate (Glc-6P) and galactose 6-phosphate (Gal-6P) differ only by the exact configuration of their C4 atom, where the O4 hydroxyl group adopts an axial position in Gal-6P and an equatorial position in Glc-6P. Surprisingly, Gan1D exhibited relatively similar catalytic activities toward sugar substrates with Glc-6P or Gal-6P in their glycone moiety. From the genomic context, however, it was expected that Gan1D would function only as a 6-phospho-β-gal. These results appear to be somewhat less surprising in view of the catalytic specificity data reported for homologous GH1 6-phospho-β-galactosidases, many of which were also shown to be less specific with regard to the glycone sugar of their substrate and specifically with regard to the configuration of the OH group at the C4 position of this sugar (e.g. glucose versus galactose) (21, 23).

The natural biological role of Gan1D

6-Phospho-β-galactosidases are relatively common in lactic bacteria, used for the utilization of lactose, which is imported into the cell via a lactose-PTS system (51). In the case of G. stearothermophilus, however, the fact that strain T-1 encodes for an enzyme (Gan1D) with 6-phospho-β-gal activity but lacks a dedicated lactose-PTS system seemed to be enigmatic. Nevertheless, we were surprised to find out that strain T-1 can utilize lactose and galactosyl-glycerol in the presence of cellobiose. These unexpected observations suggested that the cellobiose-PTS system could also transfer lactose and galactosyl-glycerol. Growth curves of strain T-1, growing in the presence of cellobiose together with either lactose or galactosyl-glycerol showed an initial logarithmic growth followed by a linear growth (Fig. 7). The slope curves of the linear growth were proportional to the turbidity (cell mass) at the end of the logarithmic phase. These results suggest that cellobiose is utilized first, supporting logarithmic growth, and when depleted, either lactose or galactosyl-glycerol can enter the cell via the same cellobiose-PTS system, although at a somewhat lower affinity. The linear growth is likely to reflect the fact that the transfer of lactose (or galactosyl-glycerol) is a rate-limiting step and that in the absence of cellobiose, the cellobiose-PTS operon is not expressed. With this in mind, it seems that Gan1D could be involved in two important functions in G. stearothermophilus T-1. First, it allows the cell to utilize residual galactose-containing disaccharides, especially the more abundant galactosyl-glycerol, taking advantage of the relative promiscuity of the cellobiose-PTS system. Considering the competition for the scarce carbon sources in the soil, this could represent a significant growth advantage. Second, it is also possible that the promiscuous cellobiose-PTS system allows the entrance of other nonmetabolizable galactose-based compounds, which may accumulate in the cell and inhibit growth. In such cases, Gan1D can potentially hydrolyze these compounds to facilitate their removal. In this respect, the E. coli lacA gene product, thiogalactoside transacetylase, could have a somewhat similar function (52) as suggested here for Gan1D.

Experimental procedures

Bacterial growth conditions

Growth medium for G. stearothermophilus was mBSM (53) supplemented with 1% of carbon source. Liquid mBSM contained the following per liter: 0.4 g of KH₂PO₄, 0.1 g of MgSO₄·7H₂O, 2 g of (NH₄)₂SO₄, and 10 g of MOPS-sodium salt, adjusted to pH 7.0. After autoclave, mBSM was supplemented with (per liter) 0.2 g of nitrilotriacetic acid, 0.125 g of NaOH, 1 ml of amino acid stock solution (0.05 g/liter of 19 amino acids except tyrosine), and 1 ml of three different trace element solutions. Solution I contained 0.145 g/liter MgSO₄·7H₂O, solution II contained 0.132 g/liter CaCl₂·2H₂O, and solution III contained per liter 0.998 g of FeSO₄·7H₂O, 0.592 g of MnSO₄·4H₂O, 0.42 g of ZnCl₂, and 0.624 g of CuSO₄·5H₂O. The pH of the latter solution was adjusted to 2.0 with sulfuric acid.

DNA and RNA isolation and manipulation

G. stearothermophilus T-1 genomic DNA was isolated according to the procedure published by Marmur (54) as outlined by Johnson (55). Plasmid DNA was purified using the DNA Clean-Up System (Promega). DNA was manipulated by standard procedures (56). Total RNA was isolated with the RNeasy kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's protocol.

Real-time RT-PCR analysis

Total RNA was isolated from mid-exponential culture of G. stearothermophilus T-1 grown on mBSM supplemented with 0.5% (w/v) of different sugars as the sole carbon source (53). Cell pellets of about 10⁹ cells were suspended in 1 ml of Tri reagent (Sigma) and sonicated (Sonicator model W-375, Heat System-Ultrasonics Inc., Plainview, NY, USA), and total RNA was extracted according to the Tri reagent protocol. Contaminating genomic DNA was removed by DNase I treatment (Qiagen) followed by a clean-up protocol with the RNeasy kit (Qiagen). RT of RNA (1 µg of total RNA) was performed with the qScript cDNA synthesis kit (Quanta BioSciences, Beverly, MD, USA) following the manufacturer's instructions. The primers were designed to generate amplicons ranging in size from 90 to 200 bp, and control reactions were performed in the absence of reverse transcriptase. The reactions were performed using the Applied Biosystems 7300 real-time PCR system and contained (20-µl volume) template cDNA, reverse and forward primers (10 µm each), and PerfeCTa SYBR Green Fastmix (Quanta BioSciences). All measurements were performed in triplicate, and data analysis was performed with the Applied Biosystems 7300 system software by using the housekeeping gene ict (encoding for isocitrate dehydrogenase) for normalization.

Synthesis of 6-phospho-β-glycosides

The synthetic substrate oNP-β-d-galactopyranoside 6-phosphate (oNPβGal-6P) was purchased from Carbosynth (Berkshire, UK). The phosphorylated forms of pNP-β-d-glucopyranoside, oNP-β-d-glucopyranoside, cellobiose, lactose, and gentiobiose were made enzymatically using β-glucosidase kinase (BglK) from Klebsiella pneumonia as described previously with minor modifications (57). The phosphorylation reaction contained 12.5 mm HEPES, pH 7.5, 1 mm MgSO₄, 50 mm ATP (adjusted to pH 7.5), 50 mm substrate, and 25 μg/ml BglK. During the 2-h incubation at room temperature, the pH of the reaction was maintained at 7.5 by the periodic addition of 40 μl of 3 m NH₄OH solution. Following the incubation, the pH of the solution was adjusted to 8.2, and barium acetate was gradually added to a final concentration of 0.6 m. The sample was centrifuged, and the soluble fraction (containing the phosphorylated sugar) was filtrated, chilled on ice, and incubated at 4 °C overnight with 4 volumes of absolute ethanol. Following centrifugation, the pellet was dried at 37 °C, resuspended in a 4-ml aqueous solution of AG 50W-X2 resin (Bio-Rad), filtrated via 0.45-µm-pore-size filters (Merck Millipore, Darmstadt, Germany), and lyophilized. Typically, a total yield of 100–150 mg of phosphorylated substrate was obtained in these reactions. The structure and purity of oNPβGlu 6-phosphate (oNPβGlu-6P), pNPβGlu 6-phosphate (pNPβGlu-6P), cellobiose 6-phosphate (Cel-6P), lactose 6-phosphate (Lac-6P), and gentiobiose 6-phosphate (Gen-6P) were confirmed by NMR and MS. ¹H NMR and ¹³C NMR spectroscopy was performed in D₂O on Bruker Avance AV-III 400 and AV-III 500 spectrometers. Signal assignments were confirmed by correlation spectroscopy, total correlation spectroscopy, and heteronuclear correlation experiments. The mass spectrum was obtained on a Bruker Daltonix Apex 3 mass spectrometer under electrospray ionization.

Synthesis of galactosyl-glycerol

The target glycerol derivative of galactose (compound 5) was synthesized from the commercial penta-O-acetyl-β-d-galactopyranoside (1) and dl-α,β-isopropylideneglycerol (3) in four chemical steps as illustrated in Scheme 1 (58–60). First, treatment of 1 with HBr (33% solution in acetic acid) resulted in the α-bromide (2) (60), which then reacted with the compound 3 in the presence of Ag₂CO₃ to afford selectively the β-glycoside (4). Sequential two-step deprotection of 4, removal of acetate esters by treatment with sodium methoxide followed by acid treatment to remove the isopropylidene ketal, resulted in the target β-galactoside (5). The product was characterized by NMR and MS spectral analysis, and the data were identical to the data reported previously (58, 59).

Scheme 1.

Reagents and conditions. a, HBr (33% AcOH), CH₂Cl₂, 0–20 °C; b, 3, Ag₂CO₃, Drierite, iodine, CH₂Cl₂, room temperature; c, NaOMe, MeOH, 0–20 °C; d, AcOH, H₂O, 60 °C.

The ¹H NMR spectra (including COSY, DEPT, HSQC, and HMBC) were recorded on a Bruker Avance^TM 400 spectrometer or 600 spectrometer, and chemical shifts reported (in ppm) are relative to internal Me₄Si (Δ = 0.0) with CDCl₃ as the solvent and to HOD (Δ = 4.63) with D₂O as the solvent. ¹³C NMR spectra were recorded on a Bruker Avance^TM 500 spectrometer or at 125.8 MHz, and the chemical shifts reported (in ppm) are relative to the residual solvent signal for CDCl₃ (Δ = 77.00) or to the solvent signal for MeOD (Δ = 49.0 ppm). Mass spectra were obtained either on a Bruker Daltonix Apex 3 mass spectrometer under electron spray ionization, or by a TSQ-70B mass spectrometer (Finnigan Mat). Reactions were monitored by TLC on Silica Gel 60 F254 (0.25 mm; Merck), and spots were visualized by charring with a yellow solution containing (NH₄)MO₇O₂₄·4H₂O (120 g) and (NH₄)₂Ce(NO₃)₆ (5 g) in 10% H₂SO₄ (800 ml). Column chromatography was performed on Silica Gel 60 (70–230 mesh). All reactions were carried out under an argon atmosphere with anhydrous solvents, unless otherwise noted. Penta-O-acetyl-β-d-galactopyranoside (1) was purchased from Tokyo Chemical Industry Co., Ltd. (TCI, Japan), and dl-α,β-isopropylideneglycerol (3) was from Sigma–Aldrich.

Penta-O-acetyl-β-d-galactopyranoside bromide (2)

To a solution of compound 1 (15.0 g, 0.038 mol) in 80 ml of anhydrous dichloromethane was added at 0 °C to a solution of HBr (9.33 g, 0.12 mol, 33% solution in acetic acid). The reaction was stirred at 0 °C for 1 h and then an additional 3 h at 20 °C. Reaction progress was monitored by TLC (ethyl acetate/hexanes, 1:1). When the reaction was complete, the mixture was cooled to 0 °C and diluted with cold dichloromethane, followed by the addition of cold solution of NaHCO₃ (5% in water). The aqueous layer was washed with three portions of cold dichloromethane; the combined extracts were dried with anhydrous MgSO₄ and concentrated at reduced pressure. Compound 2 was rapidly decomposed upon the storage and/or on column chromatography on silica gel. Therefore, it was used in the following step without additional purification. R_f 0.67 (ethyl acetate/hexanes, 1:1). MALDI-TOF MS: calculated for C₁₄H₁₉BrO₉ m/e 412.20; measured 413.0.

1,2-Isopropylidene-3-O-(β-2′,3′,4′,6′-tetra-O-acetyl-galactopyranosyl)-rac-glycerol (4)

The mixture of dl-α,β-isopropylideneglycerol (3) (2.25 g, 0.017 mol) in anhydrous dichloromethane (40 ml), Ag₂CO₃ (4.69 g, 0.017 mol), and fine Drierite (8.5 g) was stirred in the dark under argon. After being stirred for 40 min, iodine (0.43 g, 0.0017 mol) was added, followed by the dropwise addition of the bromide (2) (7.0 g, 0.017 mol) in anhydrous dichloromethane (30 ml) for ∼1 h at room temperature. The reaction mixture was stirred overnight at room temperature and then filtered. The filtrate was evaporated, the residue (orange oil) was washed twice with anhydrous dichloromethane, and the combined fractions were evaporated to dryness. The crude was purified by column chromatography on silica gel (EtOAc/hexanes 3:2) to afford 4 (3.75 g, 47%) as a mixture (1:1) of two diastereomers. R_f 0.60 (ethyl acetate/hexanes, 7:3).

¹H NMR (400 MHz, CDCl₃): δ_H 1.31 (s, 3H, isopropylidene-CH₃), 1.37 (s, 3H, isopropylidene-CH₃), 1.92 (s, 3H, acetate-CH₃), 1.98 (s, 3H, acetate-CH₃), 2.00 (s, 3H, acetate-CH₃), 2.09 (s, 3H, acetate-CH₃), 3.55–4.25 m (2 × 8H), 4.48 (d, 1H, J = 7.8 Hz), 4.51 (d, 1H, J = 7.8 Hz), 4.95 (d, 2 × 1H, J = 10.2 Hz), 5.13 (t, 2 × 1H, J = 8.2 Hz), 5.32 (s, 2 × 1H). MALDI-TOF MS calculated for C₂₀H₃₀O₁₂Na ([M + Na]⁺) m/e 485.45; measured 485.02.

3-O-(β-galactopyranosyl)-rac-glycerol (5)

To a solution of compound 4 (4.06 g, 0.009 mol) in anhydrous methanol (50 ml) at 0 °C was added sodium methylate (1.9 g, 0.035 mol) and the reaction progress was monitored by TLC (methanol/chloroform, 1:4). After about 1 h at 0 °C, the reaction was warmed to room temperature, and stirring was continued until its completion. The reaction mixture was loaded onto a silica gel column and eluted with methanol. Yield: 1.83 g, 71%, R_f 0.52 (methanol/chloroform, 1:4). ¹H NMR (400 MHz, CDCl₃): δ_H 1.33 (s, 3H, isopropylidene-CH₃), 1.38 (s, 3H, isopropylidene-CH₃), 3.56 (m, 1H, H-5′), 3.59 (m, 1H, H-3′), 3.62 (m, 1H, H-2′), 3.82 (br s, 1H, H-6′-methylene), 4.02 (br s, 1H, H-4′), 4.24 (d, 1H, H-1′, J = 7.8 Hz), 4.12–4.25 (m, 4H, 2 methylene groups). MALDI-TOF MS calculated for C₁₂H₂₂O₈ m/e 294.30; measured 295.10.

The observed material from the previous step (1.75 g, 0.006 mol) was stirred in a mixture of acetic acid-water (60 ml, 4:1) at 60 °C during 4.5 h. The reaction progress was monitored by TLC (methanol/chloroform, 1:4). After completion, the reaction mixture was evaporated; the residue was washed four times with methanol, and the crude was purified by column chromatography on silica gel (chloroform/methanol, 1:1) to afford compound 5 as colorless oil. Yield: 1.16 g, 77%, R_f 0.53 (chloroform/methanol, 1:1). ¹H NMR (600 MHz, MeOD-d₄): δ_H 3.52 (dd, 1H, H-1a, J_1a,2 = 3.6 Hz, J_1a,1b = 12.0 Hz), 3.77 (dd, 1H, H-1b, J_1b,2 = 3.5 Hz, J_1b,1a = 12.0 Hz), 3.58 (m, 1H, H-2), 3.47 (dd, 1H, H-3a, J_3a,2 = 5.2 Hz, J_3a,3b = 11.0 Hz), 3.69 (dd, 1H, H-3b, J_3b,2 = 5.1 Hz, J_3b,3a =11.0 Hz), 4.23 (d, 1H, H-1′, J_1′,2′ = 7.6 Hz), 3.76 (t, 1H, H-2′, J_2′,3′ = 6.4 Hz), 3.55 (dd, H-3′, J_3′,2′ = 6.4 Hz, J_3′,4′ = 5.5 Hz), 3.68 (dd, 1H, H-4′, J_4′,3′ = 5.5 Hz, J_4′,5′ = 4.2 Hz), 3.72 (ddd, 1H, H-5′, J_5′,4′ = 4.2 Hz, J_5′,6a = 4.7 Hz, J_5′,6b = 5.5 Hz), 3.81 (m, 2H, H-6′). ¹³C NMR (600 MHz, MeOD-d₄): δ_C 62.51 (C-6′), 64.07 (C-1), 70.31 (C-3), 72.19 (C-2), 72.42 (C-2′), 72.58 (C-4′), 72.64 (C-3′), 76.75 (C-5′), 105.21 (C-1′). MALDI-TOF MS calculated for C₉H₁₈O₈Na ([M + Na]⁺) m/e 277.20; measured 277.10.

Production and purification of His₆-tagged GalP, GanE2, Gan1D, Cel1A, and GalT2

The galP, ganE2, gan1D, celA, and galT2 ORF were all cloned into pET9d vector (Novagen). The lipoproteins GalP and GalE2 were cloned without their 27 and 30 N-terminal lipoprotein-coding sequences, respectively. All of the primers were designed to allow in-frame cloning of the genes into the T7 polymerase expression vector using restriction sites at the 5′ terminus and at the 3′ terminus (Table 4). The primers contained six histidine codons to provide His₆-fused products (Table 4). Mutagenesis was performed using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA). The mutated genes were sequenced to confirm that only the desired mutations were inserted. For protein production of Gan1D, Cel1A GalP, and GalE2, E. coli BL21(DE3) cultures containing the appropriate vector (pET9d-Gan1D, -Cel1A GalP, or -GalE2) were grown overnight in Terrific Broth (56) with kanamycin (25 g/ml) 0.5 liter in 2-liter baffled shake flasks shaken at 230 rpm at 37 °C to a final turbidity at 600 nm of 15 units. The cultures were harvested, resuspended in 30 ml of buffer (20 mm imidazole, 20 mm phosphate buffer, 500 mm NaCl, pH 7.0), disrupted by two passages through a homogenizer (Avestin, Emulsiflex), and centrifuged (14,000 × g for 15 min) to obtain soluble extracts. The His-tagged, fused proteins were purified using a 5-ml His-trap column mounted on an ÄKTA-avant FPLC system (GE Healthcare). In the case of the purified response regulator GalT2, E. coli BL21(DE3) (pET9d-galT2) culture was grown in Terrific Broth with kanamycin (25 μg ml⁻¹) (1 × 500 ml in 2-liter shake flasks) at 37 °C without induction until turbidity reached 3–4 OD₆₀₀. To improve solubility, growth was then carried out at 18 °C in the presence of 0.4 mm isopropyl 1-thio-β-d-galactopyranoside for 16 h to a final turbidity at 600 nm of about 8–11 units. The cultures were harvested, resuspended in 30 ml of buffer (20 mm imidazole, 20 mm phosphate buffer, 500 mm NaCl, pH 7.0), disrupted by two passages through homogenizer (Avestin, Emulsiflex), and centrifuged (14,000 × g for 15 min) to obtain soluble extracts. The His-tagged GalT2 was mounted on an ÄKTA-avant FPLC system (GE Healthcare) according to the manufacturer's instructions and isolated using a 5-ml His-trap column (GE Healthcare). The purified response regulator, GalT2, was dialyzed overnight against 2 liters of buffer containing 50 mm Tris-HCl (pH 7.0) and 100 mm KCl, followed by the addition of EDTA and glycerol to final concentrations of 1 mm and 10%, respectively.

Table 4.

Oligonucleotides used in this study

Primer	Sequence (5′–3′)a,b	Application
gan1D N-ter	5′-TATAGTCATGATACATCACCACCACCACCATGAGCATCGTCATCTTAAAC-3′	Cloning into pET9d
gan1D C-ter	5′-TATACGGATCCCTACAGCTCGGCACCGTTC-3′
celA N-ter	5′-TATACCCATGGGGCATCATCATCACCACCACAAACGATTGAAAATG-3′
celA C-ter	5′-ATGCGAGATCTTTAGTTTCCAGCTACTGTTTTG-3′
galP N-ter	5′-TATACCCATGGGGCACCATCATCATCATCATGCTTGGAACGTGTATGCGTATCC-3′
galP C-ter	5′-TATCAGGATCC TCATGGCGTGTTGACCTCATAATTGC-3′
galT2 N-ter	5′-ATATATCCATGGTGCATCATCATCATCATCATAAAGTGGCGATCGTCGAT-3′
galT2 °C-ter	5′-TATATAGGATCCTCAGCGCGCCTGTACTTTTTT-3′
galE2 N-ter	5′-TATCGCCATGGGGCATCATCATCATCATCAT GATGACGCCAAACAA-3′
galE2 °C-ter	5′-ATGCGAGATCTTACTTCTTCAGTTCTG-3′
Fwr gan1D	5′-GTTCCCGCCTGAGTTTTT-3′	Real-time RT-PCR
Rev gan1D	5′-CCTTTTCCATCTTCGTTCCA-3′
Fwr galE2	5'-CTCTTCTTCGTCCGATGA-3′
Rev galE2	5'-GTCCACCAGCTGAAAATCT-3′
Fwr galP	5′-CGGCAGGCGAATATTGATG-3′
Rev galP	5′-CAAATTCCGCTTCCGTCAAC-3′
Fwr ptsA	5′-GTTTGGCAACGCCTTGAAA-3′
Rev ptsA	5′-GCCTGCTTGCTCCATCACAT-3′
Fwr ptsD	5′-CGCTCAAAGAGGAAGGAAA-3′
Rev ptsD	5′-CCGATGACGACTTGAAACTG-3′
Fwr mltE	5′-ACGAGACGTTTGCGAAGCT-3′
Fwr mltE	5′-ATTTTGCCGCCCGTATAG-3′
Fwr citC	5′-GACTTGCGGCCGTGTTC-3′
Rev citC	5'-GAAGACATCTACGCTGGCATT-3′
Fwr celB	5′-ACAAGCTTGCTAGTGACGAA-3′
Rev celB	5′-GTGGGCCAATAAGAATGAC-3′
Fwr celC	5′-GATTCAAGAAGCAGAGCAAG-3′
Rev celC	5′-TGACATGACATGGTCTTCC-3′
Fwr celD	5′-CTAAGCAAAGTTCTCGTTCC-3′
Rev celD	5′-TTTCACTCATCACCTTGTCC-3′
Fwr celA	5′-GGTTTACCCACAATGAACC-3′
Rev celA	5′-CATCGTATGGTAAGCCACTT-3′
Fwr celR	5′-CACCGGTGTACCACTTAATC-3′
Rev celR	5′-GAGTTCGTGAGTGATGTAACG-3′
Fwr galE2	5′-CATAAAAGCGAAATGGCAGTC-3′	DNA-binding assays
Cy5 Rev galE2	5′-CTTATCAATCGGCGGCTTG-3′

^aBoldface bases indicate engineered restriction sites.

^bCy5, cyanine 5′-end–labeled primers.

Biochemical characterization of Gan1D

The activity of Gan1D toward aryl-phospho-β-d-glycosides was determined spectroscopically at 405 nm, following the release of the p-nitrophenyl or o-nitrophenyl. The reactions were performed at 40 °C in 100 mm citric acid-Na₂HPO₄ buffer, pH 6.5, 1 mg/ml BSA, and 0.13 µg/ml Gan1D. Substrate concentrations were in the range of 0–5 mm. The reactions were initialized by adding 20 µl of appropriately diluted prewarmed substrate to 180 µl of prewarmed enzyme solution in a 96-well plate. The absorbance at 405 nm was measured at 30-s intervals for 20 min using a Synergy HT Multi-Mode Microplate Reader (Bio-Tek Instruments, Winooski, VT, USA). The molar extinction coefficients of o-nitrophenol and p-nitrophenol as determined under the described conditions were Δε = 0.63 mm⁻¹ cm⁻¹ and Δε = 3.03 mm⁻¹ cm⁻¹, respectively. All measurements were performed in duplicates, and the relative error was <5%. The kinetic studies with natural substrates were performed by following the release of glucose. The reactions were performed at 40 °C in 100 mm citric acid-Na₂HPO₄ buffer, pH 6.5, 1 mg/ml BSA, and 0.5 µg/ml Gan1D. Substrate concentrations were in the range of 0–24 mm. The reactions were initialized by adding 20 µl of appropriately diluted prewarmed substrate to 80 µl of prewarmed enzyme solution in 1.7-ml tubes. After 10 min, the reactions were terminated by the addition of 25 µl of 0.2 m Na₂CO₃. The amount of glucose released was measured using a high-performance anion-exchange chromatography system equipped with a PA1 column (Dionex, Sunnyvale, CA, USA). The elution was carried out using two buffer eluents, eluent A (150 mm NaOH) and eluent B (150 mm NaOH and 500 mm sodium acetate). The gradient was programmed as follows: 0–1 min, isocratic elution using 100% A and 0% B; 1–50 min, linear gradient to 0% A and 100% B. The elution rate was 1 ml/min, where glucose and galactose (of known concentrations) were used as standards.

ITC

ITC measurements were performed at 40 °C with a MicroCal iTC200 titration calorimeter (Malvern Institute, Worcestershire, UK). Protein solutions were dialyzed overnight against McIlvaine buffer (100 mm citric acid-Na₂HPO₄, pH 6.5). Ligand solutions of galactose, glucose, cellobiose, and lactose were prepared by dilution with the protein dialysis buffer. Ligand solution aliquots (2–4 μl), at least 10 times the molar concentration of the protein, were added by a 40-μl rotating stirrer-syringe to a reaction cell containing 200 μl of 0.04 mm protein solution. Separate titrations of the ligand into the buffer solution determined that the heat of dilution was negligible. Calorimetric data analysis was carried out with the Origin 7.0 software (OriginLab).

Mobility-shift DNA-binding assays

To map the DNA region to which GalT2 is bound, fluorescently labeled dsDNA probes were generated via PCR, using Cy5-labeled primers (Table 1). Binding was performed in binding reaction mixture (30-µl total volume) containing 20 µl of solution comprised of 50 mm Tris-Cl (pH 7.5), 100 mm KCl, 10% glycerol, 1 mm EDTA, 2 µg of salmon sperm DNA, 0.66 mm DTT, 33 µg of BSA, 0.1 pmol of labeled probe and N terminus His₆-GalT2. The binding mixture was incubated for 30 min at 45 °C and then separated on a 6.6% nondenaturing polyacrylamide gel prepared in Tris borate-EDTA buffer that was run for 1 h. Fluorescence was detected directly on gel using a Fusion FX system (Vilber).

Phosphorylation of His₆-GalT2

Phosphorylation of His₆-GalT2 was accomplished by following the protocol of Lukat et al. (37). The GalT2 protein (250 µg) was incubated for 30 min at room temperature in 0.2 ml of a buffer containing 100 mm Tris-HCl (pH 7.0), 8.5 mm MgCl₂, and 50 mm acetyl phosphate (Sigma). The phosphorylated form of GalT2 (dimer in solution) was separated by size exclusion using an ÄKTA-avant system (GE Healthcare) equipped with a Superose 12 HR gel-filtration column (GE Healthcare) of 24-ml total column volume. Protein samples (0.2 µl) were applied onto the column and eluted at room temperature with a solution consisting of 50 mm Tris-HCl buffer, pH 7.0, 100 mm NaCl, and 0.02% sodium azide at a flow rate of 0.5 ml min⁻¹. Molecular weights were determined from regression analysis of the log relative molecular weight (M_r) of protein standards. The protein standards used were thyroglobulin (M_r 669,000), ferritin (M_r 440,000), aldolase (M_r 158,000), conalbumin (M_r 75,000), and ovalbumin (M_r 44,000) (GE Healthcare).

Growth of G. stearothermophilus T-1 on lactose and galactosyl-glycerol

G. stearothermophilus T-1 was grown on mBSM (53) supplemented with 0.4% (w/v) cellobiose as the sole carbon source. The cellobiose-induced logarithmic culture that reached 0.4 OD₆₀₀ was washed twice with mBSM and then diluted 10-fold in a 96-well plate containing mBSM supplemented, for example, with 0.05% cellobiose and 0.4% lactose. The plate (200-µl total volume/well) was incubated with shaking at 60 °C and monitored for cell growth (OD₆₀₀) in a microplate spectrophotometer (Epoch, BioTek).

Data availability

All data presented are contained in the article. The 14,067-bp sequence containing the galactose utilization region, the cellobiose-PTS system, and the galactose-PTS system from G. stearothermophilus T-1 have been deposited in the GenBank^TM under accession numbers KF840174.1, MT424749, and MT441723, respectively.