β-Glucoside Kinase (BglK) fromKlebsiella pneumoniae

ATP-dependent β-glucoside kinase (BglK) has been purified from cellobiose-grown cells of Klebsiella pneumoniae. In solution, the enzyme (EC2.7.1.85) exists as a homotetramer composed of non-covalently linked subunits of M r ∼33,000. Determination of the first 28 residues from the N terminus of the protein allowed the identification and cloning of bglK from genomic DNA ofK. pneumoniae. The open reading frame (ORF) ofbglK encodes a 297-residue polypeptide of calculatedM r 32,697. A motif of 7 amino acids (AFD7IG9GT) near the N terminus may comprise the ATP-binding site, and residue changes D7G and G9A yielded catalytically inactive proteins. BglK was progressively inactivated (t 1 2 ∼ 19 min) by N-ethylmaleimide, but ATP afforded considerable protection against the inhibitor. By the presence of a centrally located signature sequence, BglK can be assigned to the ROK (Repressor, ORF,Kinase) family of proteins. Preparation ofHis6BglK by nickel-nitrilotriacetic acid-agarose chromatography provided high purity enzyme in quantity sufficient for the preparative synthesis (200–500 mg) of ten 6-phospho-β-d-glucosides, including cellobiose-6′-P, gentiobiose-6′-P, cellobiitol-6-P, salicin-6-P, and arbutin-6-P. These (and other) derivatives are substrates for phospho-β-glucosidase(s) belonging to Families 1 and 4 of the glycosylhydrolase superfamily. The structures, physicochemical properties, and phosphorylation site(s) of the 6-phospho-β-d-glucosides have been determined by fast atom bombardment-negative ion spectrometry, thin-layer chromatography, and 1H and 13C NMR spectroscopy. The recently sequenced genomes of two Listeria species, L. monocytogenes EGD-e and L. innocua CLIP 11262, contain homologous genes (lmo2764 and lin2907, respectively) that encode a 294-residue polypeptide (M r ∼ 32,200) that exhibits ∼58% amino acid identity with BglK. The protein encoded by the two genes exhibits β-glucoside kinase activity and cross-reacts with polyclonal antibody to His6BglK from K. pneumoniae. The location oflmo2764 and lin2907 within a β-glucoside (cellobiose):phosphotransferase system operon, may presage both enzymatic (kinase) and regulatory functions for the BglK homolog inListeria species.

phoglycosylhydrolase activities. The former chromogenic substrates (and structurally similar alkyl and aryl glycoside phosphates) have usually been prepared by phosphorylation of the parent glycosides at the single primary -OH group with 2-cyanoethyl phosphate or phosphorous oxychloride (17)(18)(19). Unfortunately, these non-selective phosphorylating agents cannot be used for synthesis of the disaccharide-6Ј-P products of the P-enolpyruvate:PTS, because the presence of two primary -OH groups yields a tripartite mixture of the -6-P, -6Ј-P, and -6,6Ј-P 2 derivatives. Regiospecific chemical syntheses of disaccharide 6Ј-phosphates would entail the addition and subsequent removal of protecting groups, together with potentially laborious methods for purification of the derivatives. It is perhaps in light of these difficulties that compounds such as cellobiose-6Ј-phosphate and maltose-6Ј-phosphate are not commercially available, and it was for these reasons that we sought enzymatic routes for the biosynthesis of the ␣and ␤-conformers of disaccharide monophosphates. Fortuitously, during studies of PTS functions in permeabilized cells of K. pneumoniae, we discovered that, although low pH caused inactivation of intracellular phospho-␣-glucosidase, the ␣-glucoside-specific PTS remained operative under the same conditions. This serendipitous finding permitted the synthesis and facile isolation of a wide variety of phospho-␣-glucoside products of the PTS, including maltose-6Ј-P and the 6-phospho-derivatives of sucrose and its five linkage isomers (20).
As a potential route for the biosynthesis of phospho-␤-glucosides, we turned our attention to the report some 30 years ago by Palmer and Anderson (21) of an ATP-dependent ␤-glucoside kinase present in K. pneumoniae. Although only partially purified, the enzyme preparation of Palmer and Anderson catalyzed the in vitro phosphorylation of several ␤-glucosides and, indeed, a small quantity of cellobiose-6Ј-phosphate was prepared by these investigators.
In the present communication we describe the purification, substrate specificity, and kinetic parameters of ␤-glucoside kinase (BglK, EC 2.7.1.85) from K. pneumoniae. The chromosomal gene bglK has been cloned, and catalytically active His6 BglK has been purified from a high expression system by Ni 2ϩ -NTA-agarose chromatography. The availability of His6 BglK in quantity (and high purity) has allowed the first preparative synthesis of 10 phospho-␤-glucosides. The biosynthesis, method of isolation, and some of the physicochemical properties of these novel phospho-␤-glucosylhydrolase substrates are presented herein. By sequence-based alignment we show that BglK of K. pneumoniae can be assigned to the ROK (Repressor, ORF, Kinase) family of proteins (22) 3 and that a homolog of BglK is encoded within the cel-PTS operon in two species of Listeria.
Growth of Organisms K. pneumoniae ATCC 23357 was grown in the medium described by Sapico et al. (23), supplemented with 0.4% (w/v) of appropriate sugar. The medium (800 ml) was contained in 1-liter bottles, and cells were grown at 37°C to the stationary phase. Cells were harvested by centrifugation (13,000 ϫ g for 10 min at 5°C) and washed twice by resuspension and centrifugation from 25 mM Tris-HCl buffer (pH 7.5) containing 1 mM MgCl 2 (designated TM buffer). The yield of cells was approximately 2 g of wet weight/liter. Cells of E. coli TOP 10 (pTrcHis-bglK) were grown with vigorous aeration at 37°C in LB medium containing ampicillin (150 g/ml). Isopropyl-1-thio-␤-D-galactopyranoside (1 mM) was added to the culture at A 600 nm ϳ 0.4, and 4 h later, the cells were harvested and washed with TM buffer. The yield was approximately 3.7 g of wet weight of cells/liter.

Electrophoresis Procedures
The Novex X-Cell mini system (Invitrogen) was used for SDS-PAGE procedures. Denatured samples of cell extracts and Novex Mark 12 protein standards were electrophoresed in Novex NuPage (4 -12%) Bis-Tris gels with MES-SDS (pH 7.3) as the running buffer. Proteins were visualized by staining with Coomassie Brilliant Blue R-250. The pI of BglK was determined by analytical electrofocusing (at 10°C) in an Amersham Biosciences Multiphor flat-bed electrophoresis unit, using precast Ampholine PAG layers (pH range, 3.5-9.5) and broad range pI standards.

Immunodetection of Native and Mutant Forms of BglK
Cell extracts and SeeBlue (Invitrogen)-prestained markers were electrophoresed by SDS-PAGE, and proteins were electrophoretically transferred to nitrocellulose membranes in NuPage transfer buffer. Immunodetection of native and mutated forms of BglK was achieved by sequential incubation of the membranes with polyclonal antibody to His6 BglK and goat anti-rabbit horseradish peroxidase-conjugated antibody, as described previously (9).

Physicochemical Methods
Negative-ion FAB spectra of the P-␤-D-glucosides were obtained on a JEOL SX102 mass spectrometer. The compounds were desorbed from a glycerol matrix with 6 KeV xenon atoms, and mass measurements [M-H] Ϫ1 in FAB mode were performed at 10,000 times resolution using electric field scans and matrix ions as reference material. Low resolution analysis provided integer-mass information, and high resolving power was used for determination (and confirmation) of molecular formulae. Thin-layer chromatographic analyses were performed using 0.1-mm-thick layers of microcrystalline cellulose and a solvent containing n-butanol/acetic acid/water (5:2:3, v/v). The phosphate-containing derivatives were visualized by sequential dipping of the air-dried layers in solutions containing: (i) 50 mg of ferric chloride and 1 ml of 1 N HCl, dissolved in 100-ml of acetone, and (ii) 1.25 g of sulfosalicylic acid in 100 ml of acetone (24). 1 H and 13 C NMR spectra of the phosphates and their parent sugars were recorded on a Bruker AVANCE 500 spectrometer. Signal assignments were confirmed by COSY, heteronuclear correlation, and total correlation spectroscopy experiments. Chemical shifts (listed below in Tables IV and V) are reported in D 2 O relative to sodium 2,2,3,3-tetradeutero-3-trimethylsilyl propionate as internal standard.

Biosynthesis of P-␤-glucosides
The procedure outlined for cellobiose-6Ј-P (with adjustment for the limited quantity of some substrates) was used for the preparation of other phosphorylated compounds. Cellobiose (2 mmol) was dissolved in 10 ml of 25 mM HEPES buffer (pH 7.5) containing 2 mM MgSO 4 and immediately added to 10 ml of water containing 1 mmol of ATP (adjusted to pH 7.5 with ϳ0.6 ml of 3 M NH 4 OH). His6 BglK was added (ϳ40 units), and, throughout a 2-h incubation period at room temperature, the pH of the reaction mixture was maintained at 7.5 by addition of a 3 M NH 4 OH solution. Thereafter, the pH was adjusted to 8.2 with NH 4 OH, and 4 ml of barium acetate solution (3 mmol) was added with stirring. The heavy white precipitate of the barium salts of ADP and residual ATP was removed by centrifugation, and the supernatant fluid was clarified by filtration through a Millex (0.22-M pore size) membrane. The filtrate (approximately 22 ml) was chilled on ice, 4 volumes of absolute ethanol (0°C) was added, and the mixture was transferred to a cold room overnight. The flocculent precipitate of the Ba 2ϩ salt of cellobiose-6Ј-P (together with trace amounts of nucleotide salts) was collected by centrifugation. The white pellet was dried at 37°C for about 30 min, and Ba 2ϩ ions were exchanged for H ϩ by addition of 2-3 ml of an aqueous suspension of Bio-Rad AG 50Wx2 (H ϩ form) resin.
Resin beads were removed by filtration, and the filtrate was adjusted to pH 7.2 by NH 4 OH addition. The solution was frozen and lyophilized, to yield 300 -500 mg of the white, crystalline ammonium salt of cellobiose-6Ј-P. Except for thiocellobiose-6Ј-P, all P-␤-glucosides were quantitatively determined by enzymatic assay of Glc6P released by acid hydrolysis (1 N HCl for 2 h at 100°C). Prior to mass spectrometry and NMR spectroscopy, trace contaminants of ADP and ATP were removed from ϳ50 mg of each derivative by paper chromatography (20).

Cloning of the bglK Region of K. pneumoniae ATCC 23357
Using sequence information from the unfinished genome project of K. pneumoniae strain MGH 78578 (Washington University Genome sequencing Center, St. Louis, MO), two primers, KPC808 (5Ј-TTGC-CCCTGCGGAAAAATAC-3Ј) and KPC2116 (5Ј-TACAGTCTGGTGCTT-GCCCTCTACG-3Ј), were designed to amplify the DNA fragment encoding bglK and a portion of a putative phospho-␤-glucosidase (pbgA) gene of K. pneumoniae ATCC 23357. PCR amplification was carried out in a thermal cycler (PerkinElmer Life Sciences Model 9600) in a reaction mixture (100 l) containing 100 ng of K. pneumoniae ATCC 23357 chromosomal DNA, 10 l of 10ϫ reaction buffer, 20 mM each of the four DNTPs, 250 ng of each primer, 5 units of Pfu DNA polymerase (Stratagene, La Jolla, CA), and 1% (v/v) Me 2 SO. After an initial 2-min denaturation at 95°C, the mixture was subjected to 30 cycles of amplification. Each cycle consisted of 1 min of denaturation at 95°C, 1 min of annealing at 50°C, and 2 min and 36 s of extension at 72°C. These were followed by a 10-min runoff at 72°C. The PCR product was purified (QIAquick PCR purification kit, Qiagen) and ligated into the pCR-Blunt vector (Invitrogen, Carlsbad, CA). The recombinant plasmid was transformed into E. coli TOP 10-competent cells, and colonies were selected on Luria-Bertani agar plates containing 50 g/ml kanamycin.

DNA Sequence Analysis
Sequencing was accomplished by the dideoxynucleotide chain-termination method using the Sequenase 7-deaza-dGTP sequencing kit (U.S. Biochemicals) and [␣-35 S]dATP for labeling. Both strands of the DNA insert were sequenced. Sequences were assembled, edited, and analyzed with MacVector sequence analysis package (version 7.0, Genetics Computer Group, Madison, WI).

Cloning and High Expression of His6 BglK in E. coli TOP 10
For the amplification of bglK, two primers were synthesized from the sequence data presented in Fig. 2. Forward primer F1, 5Ј-CCCCG-GATCCCATGAAGATTGCGGCATTTGATATCGG-3Ј (the bglK sequence is in boldface, and the BamHI site is underlined); reverse primer R1, 5Ј-GGGGGTAAGCTTCTACTAATGTCGATCGTCGTCTGGC-G-3Ј (the sequence complimentary to the downstream region of bglK is in boldface, and the HindIII site is underlined). After amplification with high fidelity Pfu DNA polymerase, the DNA fragment was digested with restriction endonucleases (BamHI and HindIII), electrophoresed (1% agarose), and purified (QIAquick gel extraction kit). The purified 0.9-kb DNA fragment was ligated into the similarly digested and purified high expression vector pTrcHisB. The recombinant plasmid (pTrcHisBbglK) was transformed into competent cells of E. coli TOP 10, and transformants were selected in LB agar plates containing 150 g/ml ampicillin. A high level of expression of the histidine-tagged enzyme His6 BglK was initiated by addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside to logarithmic phase cultures (A 600 nm ϳ 0.3) of E. coli TOP 10 (pTrcHisBbglK).
Cloning and High Expression of the "BglK" Genes from L. monocytogenes EGD-e and L. innocua CLIP 11262 in E. coli TOP 10 From the complete genome sequence(s) of L. monocytogenes EGD-e and L. innocua CLIP 11262 (25), the following pairs of primers were designed for amplification of genes lmo2764 and lin2907, respectively. For amplification of lmo2764, the forward primer EGD-eF1 was 5Ј-CCCCGGATCCCATGAAAATTGCAGCTTTTGATATCGG-3Ј (the  lmo2764 sequence is in boldface, and the BamHI site is underlined) and reverse primer EGD-eR1 was 5Ј-GGGGGTAAGCTTCATCATTCAT-GTCTATTTTCCTCC-3Ј (the sequence complimentary to the downstream region of lmo2764 is in boldface, and the HindIII site is underlined). For amplification of lin2907, the forward primer CLIP-F1 was 5Ј-CCCCGGATCCCATGAAAATTGCAGCATTTGATATTGG-3Ј (the lin2907 sequence is in boldface, and the BamHI site is underlined), and reverse primer CLIP-R1 was identical to EGD-eR1. After amplification, digestion, and purification, the two ϳ0.9-kb DNA fragments were ligated (as described previously) into the high expression vector pTrcHis2B to yield plasmids pTrcHisBlmo2764 and pTrcHisBlin2907, respectively. The two plasmids were transformed into E. coli TOP 10-competent cells for expression of the Listeria gene products. (Note: incorporation of the two stop codons in our reverse primers prevents expression and incorporation of the C-terminal His 6 fusion peptides normally generated by use of pTrcHis2 vectors.)

Site-directed Mutagenesis of BglK
The method required the use of PfuTurbo DNA polymerase, a temperature cycler, and reagents that were obtained as a kit (QuikChange site-directed mutagenesis kit, Stratagene). E. coli TOP 10-competent cells were transformed with plasmids of pTrcHisBbglK containing the desired mutation. Base changes effected by site-directed mutagenesis were confirmed by sequence analysis.

Purification of Native BglK from K. pneumoniae ATCC 23357
Washed cells (ϳ16 g of wet wt) grown previously on cellobiose as the energy source were resuspended in 24 ml of TM buffer. Cells were disrupted at 0°C, by 2ϫ 1.5-min periods of sonic oscillation in a Branson (Model 350) instrument operated at ϳ75% of maximum power. BglK was purified in four stages by low pressure chromatography. Column flow rates were maintained by a P-1 peristaltic pump inter-FIG. 2. Nucleotide sequence of the BglK region of K. pneumoniae. This ϳ1.3-kilobase DNA fragment contains genes bglK and pbgA that encode the ATP-dependent ␤-glucoside kinase (BglK) and the incomplete coding sequence of a putative phospho-␤-glucosidase (PbgA), respectively. The nucleotide sequence is numbered on the right, and the deduced amino acid sequences are shown below in single-letter code. A potential ribosomal binding site (RBS) preceding bglK is underlined (fine line). The N-terminal amino acid sequence (28 residues) obtained by microsequence analysis is underlined (bold line). Residues shown boxed represent the putative ATP-binding site. The double underline indicates those residues that comprise the signature sequence for proteins of the ROK family.
faced with a Frac-100 collector. Proteins present in column eluents were monitored at 280 nm by a UV-1 optical control unit connected to a single-channel chart recorder. (All instruments were from Amersham Biosciences.) Step 1: Preparation of Dialyzed High Speed Supernatant Fluid-The sonicated extract was clarified by ultracentrifugation (180,000 ϫ g for 2 h at 5°C). The HSS was transferred to sacs and dialyzed overnight against 4 liters of TM buffer (at 4°C) Step 2: DEAE-TrisAcryl M (Anion Exchange) Chromatography-The dialyzed HSS (ϳ35 ml) was transferred at a flow rate of 0.8 ml/min to a column of DEAE-TrisAcryl M (2.6 ϫ 10 cm) previously equilibrated with TM buffer. Non-adsorbed materials were removed by washing with TM buffer, and BglK was eluted with 400 ml of a linear, increasing concentration gradient of NaCl (0 -300 mM) in TM buffer. Fractions of 6 ml were collected, and samples (30 l) were assayed for enzymatic activity. Fractions (33)(34)(35)(36)(37)(38) were pooled and concentrated to 6 ml in a pressure filtration unit (Amicon PM-10 membrane, 40 psi).
Step 3: Ultrogel AcA-54 (Molecular Sieve) Chromatography-5.5 ml of concentrate from step 2 was applied (flow rate of 0.3 ml/min) to a column of Ultrogel AcA-54 (2.6 ϫ 94 cm) equilibrated with TM buffer containing 0.1 M NaCl. Fractions of 4 ml were collected, and 20-l samples were assayed for BglK activity. Fractions 49 -53 were pooled, concentrated to ϳ7 ml, and dialyzed against 2 liters of TM buffer.
Step 4: ATP-linked Agarose (Affinity) Chromatography-The dialysate from step 3 was transferred (0.12 ml/min) to a small column (1 ϫ 7 cm) of ATP-linked agarose that had been washed extensively with TM buffer. Fractions (1 ml) were collected, and a large protein peak (devoid of BglK activity) was eluted by TM buffer (fractions 9 -25). An attempt to elute BglK with a gradient of ATP (0 -20 mM) was unsuccessful. However, the enzyme was readily eluted with 20 ml of TM buffer containing 1 M NaCl. The eluate was dialyzed against 2 liters of TM buffer, and concentrated to ϳ2 ml. This preparation contained 0.5 mg of electrophoretically pure BglK (specific activity, 24 units/mg).

Purification of His6 BglK from E. coli TOP 10 (pTrcHisBbglK)
Washed cells (ϳ22 g of wet weight) were resuspended with 45 ml of TM buffer, cells were sonicated, and HSS was prepared. His6 BglK was partially purified by DEAE-TrisAcryl M chromatography as described (above), and the concentrate (ϳ14 ml) was dialyzed against 2 liters of 20 mM sodium phosphate buffer (pH 7.2) containing 500 mM NaCl (designated binding buffer). This preparation was transferred to a column of Ni 2ϩ -NTA-agarose (1.5 ϫ 8 cm) equilibrated with binding buffer. Nonadsorbed proteins were eluted first with binding buffer and, subsequently, with a solution of 20 mM sodium phosphate buffer (pH 6.0) containing 500 mM NaCl (wash buffer). His6 BglK was eluted with 180 ml of wash buffer of increasing concentration of imidazole (0 -150 mM). Appropriate fractions were pooled and concentrated to ϳ3 ml, and the few contaminating proteins were removed by molecular filtration through a column of Ultrogel AcA-44 (1.6 ϫ 94 cm). These procedures yielded 12 mg of high purity His6 BglK (specific activity, ϳ87 units/mg). In these experiments, the standard (Control) incubation mixture was 100 l of 0.1 M HEPES buffer (pH 7.5) that contained 7.5 g of His6 BglK. The mixture was maintained at room temperature (ϳ23°C) and at times indicated, 10-l samples were removed and enzyme activity was measured in the PK/L-LDH coupled assay with cellobiose as the substrate. As required, the various compounds were included in the (100 l) reaction mixture at the following concentrations: iodoacetate (IAA), NEM, and cellobiose, all 10 mM; and ATP, 2.5 mM. The inset shows the first-order rate of inactivation of the enzyme by NEM.

Analytical Methods
Two procedures were used for the assay of enzyme activity. During purification of native and His6 BglK, enzyme activity was detected by continuous spectrophotometric measurement of NADPH formation in a phospho-␤-glucosidase (CelF)/G6PDH-coupled reaction. (The purification of NAD ϩ /Mn 2ϩ -dependent CelF (EC 3.2.1.86) has been reported previously (14).) In this assay, cellobiose is first phosphorylated by ATP-dependent BglK to yield cellobiose-6Ј-P, which is hydrolyzed (by CelF) to Glc6P and glucose. Oxidation of Glc6P is coupled to the reduction of NADP ϩ by G6PDH, and the increase in A 340 nm is recorded in a Beckman DU 640 spectrophotometer. The standard 1-ml reaction contained 0.1 M HEPES buffer (pH 7.5), 1 mM NADP ϩ , 1 mM NAD ϩ , 1 mM MgCl 2 , 1 mM MnCl 2 , 1 mM ATP, 2 mM cellobiose, 2 units of G6PDH, and non-limiting amounts of CelF. Reactions were initiated by addition of BglK preparation.
A separate assay was used to determine the substrate specificity and kinetic parameters of purified His6 BglK. In this assay, the ATP-mediated phosphorylation of ␤-glucosides (by BglK) yields ADP that, in the presence of P-enolpyruvate and pyruvate kinase, forms ATP and pyruvic acid. Reduction of pyruvate to lactate by LDH is coupled to the oxidation of NADH. The 1-ml reaction mixture contained 0.1 M HEPES buffer (pH 7.5), 5 mM ATP, 0.15 mM NADH, 10 mM MgCl 2 , 2 mM P-enolpyruvate, ϳ2 units each of PK/L-LDH, and 0.75 g of purified His6 BglK. In the two assays, initial rates of either NADP ϩ reduction or NADH oxidation (equivalent to rates of formation of ␤-glucoside-6-P) were determined using the kinetics program installed in the instrument. A molar extinction coefficient for the reduced forms of dinucleotides, ⑀ ϭ 6220 M Ϫ1 cm Ϫ1 was assumed in all calculations. One unit of BglK is the amount of enzyme that catalyzes the formation of 1 mol of ␤-glucoside-6-P/min. For most kinetic studies, the concentration range of the ␤-glucoside substrate exceeded the experimentally determined K m by 5-to 10-fold. Kinetic parameters were determined from Eadie-Hofstee plots generated by the dogStar software kinetics program, version 1.0c. Protein concentrations were measured either by the BCA protein assay (Pierce) or from the predicted molar absorption coefficients (26) for BglK (⑀ ϭ 30,940 M (subunit) Ϫ1 cm Ϫ1 ) and His6 BglK (⑀ ϭ 32,430 M (subunit) Ϫ1 cm Ϫ1 ). The sequence of residues from the N terminus of BglK was determined with an ABI 477A protein sequencer (Applied Biosystems Inc.) with an on-line ABI phenylthiohydantoin analyzer.
Purification of BglK-The enzyme was purified from cellobiose-grown cells of K. pneumoniae as described under "Experimental Procedures." The four-step process yielded 0.5 mg of BglK (specific activity, ϳ24 units/mg) from 16 g of wet weight of cells (Table I). SDS-PAGE of the preparation of BglK eluted from ATP-agarose revealed a single polypeptide of M r ϳ 33,000 (Fig. 1A), whereas the molecular weight of the enzyme estimated from AcA-44 gel filtration chromatography was ϳ130,000 (data not shown). These results provide evidence that BglK is oligomeric and, in solution, most likely comprises four identical subunits. Microsequence analysis of BglK identified the first 28 residues from the N terminus: MKIAAFDIG-GTALKMGVMARDGRLLETA. The finding of a single polypeptide by SDS-PAGE and the unambiguous nature of the microsequence data attested to the homogeneity of BglK. Interestingly, analytical electrofocusing of the enzyme preparation revealed two polypeptides (Fig. 1C, lane 2) whose esti-mated pI values of ϳ5.3 and 5.6, were significantly lower than the theoretical pI of 6.11.
Cloning and Sequence Analysis of bglK-With the N-terminal sequence of BglK as a probe, BLAST (27) search of the unfinished genome of K. pneumoniae led to the tentative identification of the gene (bglK) that encodes ␤-glucoside kinase. By methods described under "Experimental Procedures," bglK, together with a portion of the adjacent downstream gene pbgA, FIG. 6. Demonstration by thin-layer chromatography of purity and relative migration of enzymatically synthesized P-␤-glucosides. The phosphorylated compounds (ϳ0.05 mol) were applied to a 0.1-mm-thick layer of microcrystalline cellulose, and chromatography (6 h) was achieved in a solvent containing n-butanol/acetic acid/water (5:2:3, v/v). Phosphate-containing compounds were revealed by the staining procedure of Wade and Morgan (24) as white spots against a uniformly pink background: 1, arbutin-6-P; 2, cellobiitol-6-P; 3, cellobiose-6Ј-P; 4, gentiobiose-6Ј-P; 5, isopropyl 1-thio-␤-glucoside-6-P; 6, methyl ␤-glucoside-6-P; 7, 4-methylumbelliferyl ␤-glucoside-6-P; 8, phenyl ␤-glucoside-6-P; 9, salicin-6-P; and 10, thiocellobiose-6Ј-P. (R F values are listed in Table III.) Fig. 6). ␤-Glucoside Kinase from K. pneumoniae was cloned and sequenced (Fig. 2). The gene (bglK) comprises a coding sequence of 894 nucleotides that begins with an ATG codon at position 117 and terminates with a TAG stop codon at position 1010. A potential ribosome-binding site (AGGAG) is centered ϳ11 bases upstream of the Met start codon. The open reading frame (ORF) of bglK encodes a 297-residue polypeptide, whose calculated M r of 32,697 agrees well with that determined experimentally for purified BglK (M r ϳ 33,000). Importantly, the sequence of residues deduced by translation of the first 28 codons agreed perfectly with that obtained from microsequence analysis. Additional features of the sequence include: (i) a putative ATP-binding site (AFDIGGT) located at the N terminus, which is also found in yeast hexokinase (28), rat liver glucokinase (29), and bacterial glucokinases (30,31) and (ii) a glycine-rich central domain, which is present in all members of the ROK family of bacterial proteins (22). 3 This . Because of this well-defined consensus sequence (Fig. 2, double underline) and the presence of other conserved motifs (31), BglK can readily be assigned to the ROK family of proteins. Translation of the ORF encoded by pbgA provided the N-terminal sequence of the first 98 amino acids of an incomplete polypeptide. A homology search of the protein data base(s) tentatively identifies PbgA as a phospho-␤-glucosidase belonging to Family 1 of glycosylhydrolases. 2 Expression and Purification of His6 BglK-The primary aim of our work was the preparation of ␤-glucoside kinase in quantity sufficient for the biosynthesis of substrates for P-␤-glucosylhydrolases of Family 4. For this purpose, the histidine-tagged fusion protein His6 BglK was produced in a high expression system E. coli TOP 10 (pTrcHisBbglK). Single-step affinity chromatography on Ni 2ϩ -NTA-agarose provided highly purified enzyme, but this preparation retained the considerable activity of the constitutively expressed P-␤-glucosidase of the E. coli TOP 10 host. To achieve complete removal of this enzyme (BglA (19); Swiss-Prot identifier Q46829), the cell extract was successively passed through DEAE-TrisAcryl, Ni 2ϩ -NTAagarose and AcA-44 columns, and 12 mg of highly purified His6 BglK was obtained (Fig. 3). Fusion of the His 6 tag, together with other residues derived from the multicloning site of the vector (pTrcHisB), results in the addition of 33 amino acids at the N terminus of BglK. Although the fusion contains an enterokinase cleavage recognition site (DDDDKD) close to the Met start residue of His6 BglK, removal of the polypeptide was unnecessary, because its presence had no adverse effect on catalytic activity (ϳ87 units/mg, with cellobiose as substrate). The presence of these extra residues (MGGSHHHHHHG-MASMTGGQQMGRDLYDDDDKDP) was reflected in the increased size of the monomer, whose estimated molecular mass  3-7) in the extracts. The cell extract applied in lane 1 (control) was prepared from the (plasmid-free) host E. coli TOP 10. The ϩ and Ϫ signs indicate the presence or absence, respectively, of enzymatic activity in the extracts. ␤-Glucoside Kinase from K. pneumoniae (ϳ36 kDa) was similar to that calculated from the amino acid composition of His6 BglK (M r ϭ 36,338). Like the native enzyme, His6 BglK is also tetrameric, and from studies with dithiothreitol (Fig. 3, lanes 1 and 2, respectively) formation of intersubunit disulfide linkage(s) appears not to be a prerequisite for oligomerization. A polyclonal antibody against His6 BglK was prepared for use in site-directed mutagenesis studies (see below). As shown in the immunoblot (Fig. 1B), this antibody also cross-reacts specifically with native BglK. Substrate Specificity and Kinetic Properties of His6 BglK-Using various buffer systems, Palmer and Anderson (21) reported a pH optimum of 7-8 for activity of BglK. For our kinetics experiments, Tris-HCl (pH 7.5) was the buffer of choice in the PK/L-LDH coupled assay described under "Experimental Procedures." A variety of ␤-glucosides were phosphorylated by His6 BglK, and the pertinent kinetic parameters are presented in Table II. The importance of both the glucopyranosyl moiety and ␤-O-linkage for enzyme recognition is evident from the findings that glucose per se is an extremely poor substrate (K m (Glc) ϳ 40 mM) and at 50 mM final concentration, there was no detectable phosphorylation of the glucose epimers (mannose C-2, allose C-3, or galactose C-4) or of glucose analogs, including 2-deoxy-D-glucose, 2-deoxy-2-fluoro-D-glucose, and 5-thioglucose. Of two cellobiose analogs tested (at 25 mM), N, NЈdiacetylchitobiose was a substrate for BglK (ϳ62 mol phosphorylated min Ϫ1 mg Ϫ1 ), but there was no detectable phosphorylation of thiodiglucoside (D-glucopyranosyl-␤-D-1thioglucopyranoside). Eadie-Hofstee transformation of kinetic data (obtained in assays containing cellobiose and Mg 2ϩ at 5 and 10 mM, respectively) provided an estimated K m (ATP) of 0.24 Ϯ 0.02 for BglK. The deduced amino acid sequence for BglK predicts 5 cysteine residues (per monomer), and the effects of sulfhydryl-reactive agents were accordingly investigated (Fig. 4). Exposure of His6 BglK to N-ethylmaleimide (10 mM) resulted in a gradual inactivation of the enzyme (t1 ⁄2 ϳ 19 min), but surprisingly other -SH reagents, including iodoacetate and iodoacetamide, caused little or no inactivation. The presence of ATP (and to a lesser extent the substrate cellobiose) resulted in partial protection of His6 BglK from the inhibitory effect of NEM. The effect of ADP was comparable to that of ATP, but AMP afforded no protection from the inhibitor.
Biosynthesis and Proof of Structure of the 6-P-␤-D-Glucosides-The availability of His6 BglK in milligram quantities, allowed the preparative synthesis of the ten phosphorylated compounds illustrated in Fig. 5. Except for thiocellobiose-6Ј-P and isopropyl-␤-D-thioglucopyranoside-6-P, all P-␤-glucosides were hydrolyzed by one or more P-␤-glucosidases. The molecular formula and predicted [M-H] Ϫ1 values for each derivative were confirmed by negative-ion FAB mass spectrometry (Table  III). Analytical thin-layer chromatography provided evidence for purity of the 10 derivatives, by revealing a single phosphate-containing spot for each preparation (Fig. 6). The estimated R F values of the derivatives, under the prescribed conditions, are listed in Table III. Proof of structure and verification of the exclusive phosphorylation at the primary glucosyl-O-6 were provided by comparison of the 1 H and 13 C NMR data of the mono-and disaccharide phosphates with their respective parent (non-phosphorylated) glycosides. It is well FIG. 8. Multiple alignment of the deduced sequences of polypeptides encoded by bglK of K. pneumoniae (GenBank TM accession number AY035305) and genes lmo2764 and lin2907 from L. monocytogenes and L. innocua (EMBL accession numbers AL591984 and AL596174, respectively). Amino acid sequences were aligned by the Clustal W algorithm (34). Fully conserved residues are shown as boldface letters against gray background, and the numbers at right denote residue positions. Amino acids indicated by asterisks are highly conserved in bacterial glucokinases of the ROK family of proteins, and in BglK, these residues were changed by site-directed mutagenesis (see Fig. 7). established (20,32,33) that phosphorylation of a hydroxyl group in a mono-or disaccharide usually results in a small but distinctive downfield shift of 0.2-0.5 ppm for the protons attached to the phosphate-carrying carbon (here the 6-CH 2 ), whereas hydrogens situated vicinal thereto (H-5) or in the ␤-position (H-4) exhibit only minor deshielding, with all other signals being essentially identical to those of the parent sugar. As borne out by the juxtaposition of 1 H NMR data for the parent sugar versus monophosphate in Table IV, all ten glycosides (Fig. 5) were phosphorylated at O-6, because there is a uniform 0.2-to 0.3-ppm downfield shift of the respective 6-CH 2 protons relative to their parent sugars (cf. column of data under "6-H 2 " in Table IV). Similarly, the 0.15-to 0.20-ppm shift of all H-4 signals to lower magnetic field can be attributed to the spatial proximity of the phosphate ester group at O-6 and the axially oriented H-4 proton. All other resonances, including that of H-5 vicinal to the phosphate group, remain essentially unchanged.
Substantiation of these assignments was provided by comparative data obtained from compounds that contain two glucosyl residues (cellobiose, thiocellobiose, and gentiobiose). Of the two 6-CH 2 groups present in these disaccharides, only the terminal one (situated in the non-reducing portion of the molecule) was shifted in the phosphorylated derivative, whereas signals from CH 2 in the reducing moiety remained unchanged. Finally, the signals from the aglycon moieties of salicin (primary benzylic OH group) and arbutin (phenolic OH group) also remained constant and indicative of the absence of a phosphate ester at these positions. The 13 C NMR data presented in Table  V confirmed the findings from 1 H spectroscopy. In accord with previous analyses of various sugar phosphates (20,32,33), the most pronounced 13 C downfield shift was observed for the carbon carrying the phosphate ester. The marked downfield shift of 1.7-2.4 ppm recorded in Table V (column of data under "C-6") confirmed the terminal C-6 as the site of phosphorylation in all derivatives. The differences found for the ␤and ␥-carbons (C-5 and C-4, respectively) were comparatively small and have not been relied upon for structural assignments. Additional evidence for phosphorylation at O-6 in cellobiitol and the six ␤-D-glucosides is provided by the doublet splitting of the C-6 resonance with a coupling constant, 3 J(C,P), of about 4 -5 Hz that is, however, not detectable in the more complex 13 C data of the disaccharides.
Site-directed Mutagenesis of BglK-Several residues of BglK, including Asp 7 , Gly 9 , Asp 103 , Gly 131 , and Gly 133 , are positionally conserved in microbial glucokinases that belong to the ROK family of proteins (24). Although it is reasonable to assume that these particular amino acids are structurally or catalytically important, the issue was addressed by site-directed mutagenesis of bglK. Residue changes effected were D7G, G9A, D103G, G131A, and G133A. After transformation of E. coli TOP 10 with plasmids (pTrcHisbglK) containing the desired mutation, the cells were grown and extracts were prepared. High level expression of the five mutant forms of His6 BglK was confirmed by Western blot analyses using polyclonal antibody against this protein (Fig. 7). All extracts contained a single immunoreactive polypeptide of the size (ϳ36 kDa) expected for His6 BglK, but remarkably, none of the extracts contained measurable activity of ␤-glucoside kinase.
Homologs of BglK-Presently, BglK activity has been described only in K. pneumoniae, and (except for two species of Listeria) a BLAST (26) search of non-redundant protein data bases revealed no entries with significant homology to this protein. Surprisingly, the recently sequenced genomes (25) of L. monocytogenes EGD-e (serovar 1/2a) and L. innocua CLIP 11262 (serovar 6a) contain a gene (lmo2764 and lin2907, respectively) whose ORF encodes a predicted polypeptide of 294 amino acids (M r ϳ 32,200) that exhibits ϳ59% residue identity (ϳ72% similarity) with BglK (Fig. 8). Both lin2907 and lmo2764 were cloned into pTrcHis2B, and the gene products were expressed in E. coli TOP 10. SDS-PAGE and Western blot analyses of cell extracts of E. coli TOP 10 (pTrcHis2B lmo2764) and E. coli TOP 10 (pTrcHis2Blin2907) revealed high level expression of a polypeptide of M r ϳ 32 kDa, which also crossreacted with antibody prepared against His6 BglK from K. pneumoniae (Fig. 9, A and B, respectively). In both cases, the expressed protein catalyzed the ATP-dependent phosphorylation of cellobiose (specific activity (avg.) ϳ 2.8 mol of cellobiose-6Ј-P formed min Ϫ1 mg Ϫ1 ) and was thus identified as a homolog of BglK.
Properties of BglK-The enzyme is exacting with respect to ␤-glucoside substrates, and NMR data establish that phosphorylation occurs exclusively at the C-6 position of the glucopyr- ␤-Glucoside Kinase from K. pneumoniae anosyl moiety. BglK tolerates wide variation in both size and structure of the aglycon component, and substrates include molecules such as hexose linked via C-2, C-3, C-4, or C-6 positions (sophorose, laminaribiose, cellobiose, and gentiobiose, respectively); hexitol (cellobiitol); aliphatic group (methyl ␤-glucoside); and aryl groups (salicin, arbutin, phenyl ␤-glucoside, and 4-methylumbelliferyl ␤-glucoside). The importance of the ␤-O-linked aglycon in substrate recognition is evidenced by the fact that BglK does not phosphorylate maltose (␣-isomer of cellobiose) and has remarkably little affinity for glucose itself. Indeed, the monosaccharide is phosphorylated only when present at high concentration (K m ϳ 40 mM). These findings affirm the classification of BglK as a ␤-glucoside kinase (EC 2.7.1.85) that is separate and distinct from glucokinase (EC 2.7.1.2) found in many species of bacteria, including K. pneumoniae. Interestingly, the two ATP-dependent kinases contain the common signature sequence by which they are assigned to the ROK family of proteins. Site-directed mutagenesis of the individual residues Asp 7 , Gly 9 , Asp 103 , Gly 131 , and Gly 133 of BglK, yields proteins that are catalytically inactive. It is noteworthy that residues Asp 7 and Gly 9 reside in a motif XD 7 XG 9 GT that is conserved in many microbial glucokinases and that may represent the ATP-binding site in these enzymes (31). The loss of activity of BglK, attendant upon the amino acid changes D7G and G9A, would be consistent with an inability of the two mutant proteins to bind the phosphoryl donor. Whether residues Asp 103 , Gly 131 , and Gly 133 fulfill catalytic or structural functions cannot be discerned from our results.
BglK contains 5 cysteine residues, but the resistance of the enzyme to high concentrations of sulfhydryl-reactive agents (iodoacetate and iodoacetamide) suggests that Cys residues are not prerequisite for catalysis. Although the time-dependent inactivation of BglK by NEM appears to contradict this statement, an earlier discussion of NEM reactivity by Means and Feeney (38) may provide a satisfactory accommodation of the data. Means and Feeney point out that, although NEM is fairly specific in its interaction with -SH groups at pH Յ 7, at higher pH a reaction may also occur (albeit more slowly) between the olefinic bond of NEM and amine groups of proteins. Our inhibitor experiments were performed at pH 7.5, and the comparatively slow first-order rate of inactivation of the enzyme (t1 ⁄2 ϳ 19 min) may be indicative of a reaction between NEM and a catalytically functional lysine or histidine residue in BglK. A locus for the inhibitor-sensitive residue(s) close to the FDIGGT motif is consistent with the protection that ATP affords against inactivation by NEM.
Function(s) of BglK-Palmer and Anderson (21) described the expression of BglK during growth of K. pneumoniae on cellobiose, and here we report that other ␤-glucosides, including gentiobiose, cellobiitol, and methyl-␤-glucoside may also elicit synthesis of this enzyme. The bglK gene lies adjacent to the partially sequenced pbgA gene (Fig. 2) that encodes a P-␤-glucosidase belonging to Family 1 of glycosylhydrolases, 2 and it is likely that the phosphorylated products of BglK activity are substrates for the P-␤-glucosidase. Although BglK may participate in a metabolic pathway initiated by ATP-dependent phosphorylation of cellobiose, data obtained from the (incomplete) genome sequence of K. pneumoniae strain MGH 78578 4 reveal at least one cel-operon that may promote the P-enolpyruvate-dependent phosphorylation and hydrolysis of cellobiose and related ␤-glucosides, such as arbutin, salicin, esculin, and phenyl-␤-glucoside. To date BglK has been described and purified only from K. pneumoniae, but as we now report, a gene present in two species of Listeria (25) also encodes a polypep-tide with extensive homology to this enzyme. The products of lmo2764 and lin2907 (in L. monocytogenes and L. innocua, respectively) exhibit ␤-glucoside kinase activity, and both proteins cross-react with antibody to His6 Bglk from K. pneumoniae (Fig. 9). What role the "BglK" homolog plays in the metabolism of ␤-glucosides by Listeria is unknown, but two facts are worthy of consideration. First, the gene bglK resides in a cel:PTS operon (Fig. 10), and BglK (a member of the ROK family) may serve the dual roles of kinase and transcriptional regulator of this operon. Second, inspection of Fig. 10 reveals two genes whose products may catalyze the phosphorylation of ␤-glucosides using different phosphoryl donors: (i) ATP-dependent phosphorylation catalyzed by BglK and (ii) P-enolpyruvate-dependent phosphorylation mediated via EIIB cel of the PTS. Whether BglK functions in conjunction with (or can substitute for) EIIB cel of the PTS cel remains to be determined. However, it is pertinent that Erni and his colleagues (39) have investigated the potential of ATP-dependent glucokinase (Glk) to serve as substitute for EIIB glc in the PTS glc of E. coli. In this elegant study, recombinant proteins were created by linkage of the EIIC glc domain to glucokinase. Unfortunately, the EIIC glc :Glk fusion proteins failed to catalyze the phosphorylative translocation of glucose with ATP as the phosphoryl donor.