Development of Novel Sugar Isomerases by Optimization of Active Sites in Phosphosugar Isomerases for Monosaccharides

Abstract

Phosphosugar isomerases can catalyze the isomerization of not only phosphosugar but also of monosaccharides, suggesting that the phosphosugar isomerases can be used as sugar isomerases that do not exist in nature. Determination of active-site residues of phosphosugar isomerases, including ribose-5-phosphate isomerase from Clostridium difficile (CDRPI), mannose-6-phosphate isomerase from Bacillus subtilis (BSMPI), and glucose-6-phosphate isomerase from Pyrococcus furiosus (PFGPI), was accomplished by docking of monosaccharides onto the structure models of the isomerases. The determinant residues, including Arg133 of CDRPI, Arg192 of BSMPI, and Thr85 of PFGPI, were subjected to alanine substitutions and found to act as phosphate-binding sites. R133D of CDRPI, R192 of BSMPI, and T85Q of PFGPI displayed the highest catalytic efficiencies for monosaccharides at each position. These residues exhibited 1.8-, 3.5-, and 4.9-fold higher catalytic efficiencies, respectively, for the monosaccharides than the wild-type enzyme. However, the activities of these 3 variant enzymes for phosphosugars as the original substrates disappeared. Thus, R133D of CDRPI, R192 of BSMPI, and T85Q of PFGPI are no longer phosphosugar isomerases; instead, they are changed to a d-ribose isomerase, an l-ribose isomerase, and an l-talose isomerase, respectively. In this study, we used substrate-tailored optimization to develop novel sugar isomerases which are not found in nature based on phosphosugar isomerases.

INTRODUCTION

The development of new enzymes has long been a goal in the field of protein engineering, and many advances have been made regarding directed evolution and rational design (1). New enzymes with novel catalytic activities as biocatalysts can facilitate and simplify many chemical processes to produce a broad range of products (2). The protein engineering of enzymes has emerged as a powerful enabling technology for development of a new biocatalyst. Directed evolution does not require structural information but often results in various variants. Moreover, it requires a high-throughput screening system and can unpredictably alter enzyme properties. Rational design, employing site-directed mutagenesis, is relatively inexpensive and simple. However, detailed structural knowledge of a protein is often unavailable, and the effects of various mutations can be extremely difficult to predict (1). Substrate-tailored optimization is an easy way to create novel enzymes and combines the advantages of directed evolution and rational design while concurrently removing the aforementioned disadvantages. In substrate-tailored optimization, the target substrate is docked to an enzyme with different function using its determined structure or homology model, and residues of the active site that interact with the substrate are selected and optimized using site-directed mutagenesis.

Recently, carbohydrates have attracted attention as cell surface receptors of cells in glycobiology due to their effective functions. Synthesized carbohydrates that disrupt carbohydrate-dependent processes are emerging as important therapeutic agents (3). Among the carbohydrates, monosaccharides are the simplest carbohydrates and the most basic compounds in glycobiology. Currently, monosaccharides are synthesized using chemical or biological methods, but the chemical method has several disadvantages, including complex purification steps and the formation of by-products and chemical waste. To overcome these disadvantages, monosaccharides are synthesized through microbial and enzymatic reactions using various enzymes (4). Rare monosaccharides have a wide variety of applications, including their uses as low-calorie sweeteners, antioxidants, glycosidase inhibitors, nucleoside analogs, antiviral agents, anticancer agents, and immunosuppressants (5–11). However, natural biosynthetic enzymes are insufficient for the synthesis of various rare monosaccharides, and specific sugar isomerases have not yet been identified in nature. For example, some sugar isomerases, such as l-talose isomerase, d-ribose isomerase, d-talose isomerase, l-xylose isomerase, and l-lyxose isomerase, have not been identified, because organisms do not require such rare monosaccharides to survive. Thus, the discovery of new natural monosaccharide biosynthetic enzymes via screening is very difficult, and such enzymes may be obtained by modifying naturally occurring enzymes by using protein engineering techniques.

Three phosphosugar isomerases, namely, ribose-5-phosphate isomerase (RPI) (12), mannose-6-phosphate isomerase (MPI) (13), and glucose-6-phosphate isomerase (GPI) (14), participate in the pentose phosphate pathway and glycolysis metabolism (see Fig. S1 in the supplemental material). Because these isomerases are involved in the isomerization of phosphosugars, they can also catalyze the isomerization of various monosaccharides owing to their broad substrate specificity (15–19) (Fig. 1).

Fig 1.

Schematic diagrams of reactions catalyzed by phosphosugar isomerases, including CDRPI, BSMPI, and PFGPI. (A) Isomerization between ribose-5-phosphate and ribulose-5-phosphate and between d-ribose and d-ribulose catalyzed by ribose-5-phosphate isomerase from Clostridium difficile (CDRPI). (B) Isomerization between mannose-6-phosphate and fructose-6-phosphate and between l-ribose and l-ribulose catalyzed by mannose-6-phosphate isomerase from Bacillus subtilis (BSMPI). (C) Isomerization between glucose-6-phosphate and fructose-6-phosphate and between l-talose and l-tagatose catalyzed by glucose-6-phosphate isomerase from Pyrococcus furiosus (PFGPI).

In this study, we developed d-ribose isomerase, l-ribose isomerase, and l-talose isomerase, based on RPI from Clostridium difficile (CDRPI), MPI from Bacillus subtilis (BSMPI), and GPI from Pyrococcus furiosus (PFGPI), respectively, via substrate-tailored optimization.

MATERIALS AND METHODS

Materials.

Kits for PCR product purification, gel extraction, and plasmid preparation, as well as the DNA-modifying enzymes, were purchased from Promega. The phosphosugar and monosaccharide standards were purchased from Sigma and Carbosynth.

Bacterial strains, plasmids, and growth conditions.

C. difficile ATCC 43255, B. subtilis ATCC 23857, P. furiosus DSM 3638, Escherichia coli ER2566, and plasmid pET-28a(+) were used as the sources of genomic DNA for RPI, MPI, and GPI; as host cells; and as the expression vector, respectively. Recombinant E. coli cells for enzyme expression were cultivated in 500 ml of Luria-Bertani (LB) medium in a 2,000-ml flask containing 20 μg/ml kanamycin at 37°C with shaking at 250 rpm. When the optical density at 600 nm (OD₆₀₀) of the culture reached 0.6, 0.1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) was added to the culture medium, and the culture was incubated with shaking at 150 rpm at 16°C for 16 h to express the enzyme.

Cloning and site-directed mutagenesis of phosphosugar isomerases.

Primer sequences used for gene cloning were based on the DNA sequence of CDRPI (GenBank accession number AM180355). Forward (5′-TTTCATATGAAGATAGGATTAGGCT-3′) and reverse (5′-TTTCTCGAGTTATTTATTATGTTTTTCTTC-3′) primers were designed to introduce the NdeI and XhoI restriction sites, respectively, at the underlined sequences. Primer sequences used for gene cloning were based on the DNA sequence of BSMPI (GenBank accession number AF324506). Forward (5′-TTTCATATGACGCATCCTTTATT-3′) and reverse (5′-TTTCTCGAGTTAAGGATGAGATATCA-3′) primers were designed for introduction of the NdeI and EcoRI restriction sites, respectively, at the underlined sequences. The sequence of the primers used for gene cloning was based on the DNA sequence of the glucose-6-phosphate isomerase from P. furiosus (GenBank accession number AF381250). Forward (5′-TTTCATATGTATAAGGAACCTTTTGGAGTG-3′) and reverse (5′-TTTCTCGAGCTACTTTTTCCACCTGGGATTATC-3′) primers were designed to introduce the NdeI and XhoI restriction sites, respectively, at the underlined sequences.

Amplified DNA fragments were purified using a PCR purification kit (Promega). The purified sequences were ligated into individual restriction enzyme sites of pET-28a(+). The resulting plasmids were used to transform the E. coli ER2566 strain. Site-directed mutagenesis was performed using the QuikChange kit (Stratagene).

Purification of phosphosugar isomerases.

Washed recombinant cells were resuspended in 50 mM phosphate buffer containing 300 mM NaCl, 10 mM imidazole, and 0.1 mM phenylmethylsulfonyl fluoride (PMSF) as a protease inhibitor. The resuspended cells were disrupted using ultrasonication with the samples kept on ice. Cell debris was removed by centrifugation at 13,000 × g for 20 min at 4°C, and the supernatant was filtered through a 0.45-μm-pore-size filter. The filtrate was applied to a HisTrap HP chromatography column (GE Healthcare) equilibrated with 50 mM phosphate buffer. The column was washed extensively with the same buffer, and the bound protein was eluted with a linear gradient from 10 to 250 mM imidazole at a flow rate of 1 ml/min. The active fractions were collected and dialyzed at 4°C for 24 h against 50 mM piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES) buffer (pH 7.0). After dialysis, the resulting solution was used as the purified enzyme. Purification steps using a column were carried out using a fast protein liquid chromatography (FPLC) system (Bio-Rad Laboratories) in a cold room.

Comparative homology modeling.

Homology modeling of CDRPI was performed using MODELLER (20) and optimized using FoldX (21), and it was based on the X-ray structure model of RPI from Clostridium thermocellum (PDB code 3HEE) as a template. A homologous search and sequence alignment were conducted using sequence analysis and multiple-sequence alignment modules, respectively. Based on the optimized alignment, 5 comparative models of the target sequence were generated using MODELLER by applying the default building routine “model” with fast refinement. This procedure has an advantage in that the best model can be selected from several candidate models. Furthermore, variability among the models can be used to evaluate modeling reliability. Energy minimization was performed using the consistent valence force field and the Discover program using the steepest descent and conjugated gradient algorithms. The quality of these models was analyzed using PROCHECK (22).

Ligand docking.

Docking of ribose-5-phosphate/l-talose, mannose-6-phosphate/d-talose, and glucose-6-phosphate/l-talose initially was accomplished based on the predicted topological binding sites by several algorithms (23). The automated docking was carried out using the CDOCKER program (Accelrys) (24) based on the Merck molecular force field (MMFF) and AutoDock 4.0 program suite (25). The active site was defined as the collection of amino acid residues enclosed within a sphere with a radius of 4.5 Å from the center of the docked substrate. The Molecular Dynamics-simulated annealing process was performed using a rigid protein and flexible ligand. Ligand-protein interactions were computed from a full force field, and a final minimization step was applied to the ligand docking pose. The minimization consisted of 50 steps of the steepest descent followed by up to 200 steps of conjugated gradient using an energy tolerance of 0.001 kcal mol⁻¹. The substrate orientation giving the lowest interaction energy was chosen for additional docking studies.

Analytical methods.

The concentrations of phosphosugars and monosaccharides were determined by a Bio-LC system (Dionex ICS-3000) with an electrochemical detector using a CarboPac PAI column. To analyze phosphosugars, the column was eluted at 30°C with an Na-acetate gradient of 75 mM NaOH and 75 mM NaOH-500 mM Na-acetate. The gradient was increased to 100 mM between 0 and 35 min, to 150 mM between 35 and 38 min, to 350 mM between 38 and 65 min, and then to 500 mM for 75 min. The flow rate was 1 ml/min. To analyze monosaccharides, the column was eluted at 30°C with 200 mM sodium hydroxide at a flow rate of 1 ml/min.

RESULTS AND DISCUSSION

Substrate specificity of phosphosugar isomerases.

Three phosphosugar isomerases, including CDRPI, BSMPI, and PFGPI, were cloned and expressed in E. coli and purified as a single band using HisTrap HP affinity chromatography (15, 17, 18). These wild-type enzymes can catalyze the isomerization reactions not only for phosphosugars but also for monosaccharides. These properties allow these phosphosugar isomerases to be used as candidates for creating new sugar isomerases. The substrate specificity of these enzymes was investigated with the d and l forms of the pentoses and hexoses, including talose, allose, mannose, galactose, glucose, altrose, gulose, idose, xylose, arabinose, lyxose, and ribose. Among the monosaccharides, the specific activities of wild-type CDRPI, BSMPI, and PFGPI were the highest for d-ribose, l-ribose, and l-talose, respectively (15, 17, 18) (Table 1). Thus, these phosphosugar isomerases were used in the development of novel sugar isomerases.

Table 1.

Relative activities of the wild-type and variant enzymes of CDRPI, BSMPI, and PFGPI for monosaccharides

Substrate	Relative activity (%) of enzyme ofa:
	CDRPI		BSMPI		PFGPI
	Wild-type	R132D	Wild-type	R192N	Wild-type	T85Q
d-Talose	1 ± 0.1	2 ± 0.2	54 ± 1.3	95 ± 2.4	45 ± 1.5	152 ± 1.2
l-Talose	100 ± 2.5	156 ± 2.3	2 ± 0.1	5 ± 0.2	100 ± 0.7	456 ± 4.3
d-Allose	18 ± 0.2	31 ± 1.5	1 ± 0.1	2 ± 0.1	71 ± 1.5	260 ± 12
l-Allose	8 ± 0.2	15 ± 0.8	15 ± 0.9	35 ± 0.1	51 ± 1.4	192 ± 7.2
d-Mannose	NDb	ND	19 ± 0.5	42 ± 0.7	28 ± 0.2	128 ± 1.5
l-Mannose	ND	ND	3 ± 0.1	6 ± 0.1	31 ± 0.3	135 ± 6.8
d-Galactose	ND	ND	ND	ND	3 ± 0.1	15 ± 0.2
l-Galactose	ND	ND	ND	ND	4 ± 0.1	19 ± 0.3
d-Glucose	ND	ND	ND	ND	34 ± 0.8	142 ± 5.6
l-Glucose	ND	ND	ND	ND	41 ± 1.1	150 ± 9.5
d-Altrose	ND	ND	ND	ND	18 ± 0.2	75 ± 1.3
l-Altrose	ND	ND	ND	ND	11 ± 0.1	39 ± 0.5
d-Gulose	ND	ND	ND	ND	33 ± 0.2	139 ± 2.7
l-Gulose	ND	ND	ND	ND	26 ± 0.4	122 ± 4.5
d-Idose	ND	ND	ND	ND	33 ± 0.2	138 ± 4.6
l-Idose	ND	ND	ND	ND	33 ± 0.3	139 ± 8.7
d-Xylose	ND	ND	ND	ND	37 ± 0.1	148 ± 8.1
l-Xylose	ND	ND	ND	ND	39 ± 1.5	148 ± 3.6
d-Arabinose	ND	ND	ND	ND	24 ± 0.4	118 ± 1.8
l-Arabinose	ND	ND	ND	ND	14 ± 0.1	60 ± 1.7
d-Lyxose	ND	ND	62 ± 0.2	99 ± 3.7	28 ± 0.6	113 ± 2.1
l-Lyxose	ND	ND	1 ± 0.1	2 ± 0.1	31 ± 1.2	129 ± 3.8
d-Ribose	79 ± 0.9	116 ± 3.4	2 ± 0.1	3 ± 0.1	88 ± 0.5	290 ± 15
l-Ribose	4 ± 0.1	5 ± 0.2	100 ± 1.1	257 ± 6.0	63 ± 1.1	248 ± 14

^aThe relative activities of 100% for CTRPI for d-ribose, BSMPI for l-ribose, and PFGPI for l-talose were 7.4, 22.5, and 0.6 μmol min⁻¹ mg⁻¹, respectively. The data represent the means and standard deviations from three separate experiments.

^bND, not detected.

Determinant positions at active sites of phosphosugar isomerases for monosaccharides.

To identify the determinant residues responsible for developing novel sugar isomerases, we used the crystal structure models of BSMPI (PDB code 1QWR) and PFGPI (PDB code 2GC2) and the homology model of CDRPI. The monosaccharides d-ribose, l-ribose, and l-talose were docked onto the phosphosugar isomerases CDRPI, BSMPI, and PFGPI, respectively, using the Surflex docking program (24). Eleven residues of CDRPI, namely, Asp8, His9, Tyr43, Cys66, Thr68, His99, Asn100, Arg110, Arg133, His134, and Arg137; 15 residues of BSMPI, namely, Lys12, Arg14, Trp16, Leu86, Gln95, His97, Lys113, Glu154, Trp117, His172, Leu174, Glu182, Asp188, Tyr191, and Arg192; and 11 residues of PFGPI, namely, Tyr52, Thr71, Thr85, His88, His90, Glu97, Tyr99, His136, Tyr152, His158, and Tyr160, were shown to interact with the docked monosaccharides via hydrogen bonding. These residues were replaced one by one with alanine, and the wild-type and all variant enzymes were expressed and purified. The activities of the wild-type and variant enzymes were measured using phosphosugars and monosaccharides as substrates (Fig. 2). Three variants, CDRPI R133A, BSMPI R192A, and PFGPI T85A, showed the highest activities for d-ribose, l-ribose, and l-talose, respectively. However, the activities of these variants for phosphosugars were negligible. The different activity patterns observed for phosphosugars and monosaccharides indicate that Arg133 of CDRPI, Arg192 of BSMPI, and Thr85 of PFGPI are molecular determinants that can be used to develop novel sugar isomerases. These residues, in phosphosugar isomerases located near the phosphate group of phosphosugar, consist of several (more than two) positively charged or polar amino acids (Fig. 3A, C, and E), whereas residues located near the terminal (5 or 6)-OH of the monosaccharide are not typically positively charged or polar amino acids (Fig. 3B, D, and F). Thus, the phosphate-binding site of phosphosugar isomerases may contain crucial residues that, when replaced with other amino acids, result in the creation of a new sugar isomerase.

Fig 2.

Relative catalytic efficiencies of the wild-type and variant enzymes of CDRPI, BSMPI, and PFGPI for phosphosugars and monosaccharides. (A) Relative activities of the wild-type and variant enzymes of CDRPI for ribose-5-phosphate and d-ribose. The relative catalytic efficiencies of 100% for ribose-5-phosphate and d-ribose were 500 and 0.6 mM⁻¹ s⁻¹, respectively. (B) Relative activities of the wild-type and variant enzymes of BSMPI for mannose-6-phosphate and l-ribose. The relative catalytic efficiencies of 100% for mannose-6-phosphate and l-ribose were 2,014 and 13 mM⁻¹ s⁻¹, respectively. (C) Relative activities of the wild-type and variant enzymes of PFGPI for glucose-6-phosphate and l-talose. The relative catalytic efficiencies of 100% for glucose-6-phosphate and l-talose were 2,284 and 3.6 mM⁻¹ s⁻¹, respectively. The black and white bars represent relative activities for phosphosugar and monosaccharide, respectively. The data represent the means from three separate experiments, and the error bars represent standard deviations.

Fig 3.

Active-site structures of wild-type enzymes of CDRPI, BSMPI, and PFGPI with phosphosugars and monosaccharides. (A) Active site of CDRPI with ribose-5-phosphate. Arg133 (cyan) in CDRPI directly interacted with the phosphate group (red) in the phosphosugar. The dotted line indicates an interaction between the phosphate group of ribose-5-phosphate and the phosphate-binding site of CDPRI. (B) Active site of CDRPI with d-ribose as a substrate. Arg133 (cyan) and Asp132 (magenta) are visible at the bottom of the image. (C) Active site of BSMPI with mannose-6-phosphate. The charcoal sphere represents a metal ion. Arg192 (cyan) in BSMPI directly interacted with the phosphate group (red) in the phosphosugar. The dotted line indicates an interaction between the phosphate group of mannose-6-phosphate and the phosphate-binding site of BSMPI. (D) Active site of BSMPI with l-ribose as a substrate. Arg192 (cyan) and Asn192 (magenta) are visible at the bottom of the image. (E) Active site of PFGPI with glucose-6-phosphate. The charcoal sphere represents a metal ion. Thr85 (cyan) in PFGPI directly interacted with the phosphate group (red) in the phosphosugar. The dotted line indicates an interaction between the phosphate group of glucose-6-phosphate and the phosphate-binding site of PFGPI. (F) Active site of PFGPI with l-talose as a substrate. Thr85 (cyan) and Gln85 (magenta) are visible at the bottom of the image. The residue, metal ion, and distance are represented as a stick model, sphere, and dashed line, respectively. Docking of phosphosugars and monosaccharides was initially accomplished based on the predicted topological binding sites by several algorithms using a homology model of CDRPI and crystal structure of BSMPI and PFGPI. The automated docking was carried out using the CDOCKER program (Accelrys) based on the Merck molecular force field (MMFF) and AutoDock 4.0 program suite. PROCHECK examination of the mutant enzymes did not show any molecular clashes for the variant side chains.

Development of novel sugar isomerases using site-directed mutagenesis at determinant positions of phosphosugar isomerases.

The amino acid residues at determinant positions of phosphosugar isomerases were replaced with other amino acids, including Asp, Gln, Lys, Glu, Tyr, and Ile at position 133 of CDRPI; Glu, Lys, Leu, Asn, and Tyr at position 192 of BSMPI; and Ser, Gln, Asp, and Lys at position 85 of PFGPI. Expression of the wild-type and variant enzymes was confirmed by SDS-PAGE (data not shown). R133D of CDRPI, R192 of BSMPI, and T85Q of PFGPI displayed the highest catalytic efficiencies for monosaccharides as substrates among the wild-type and variant enzymes at position 133 of CDRPI, position 192 of BSMPI, and position 85 of PFGPI, respectively (Fig. 4). These enzymes exhibited 1.8-, 3.5-, and 4.9-fold higher catalytic efficiencies, respectively, than the corresponding wild-type enzymes (Table 2). However, the variants showed no activity for phosphosugars as original substrates. Indeed, the variants did not convert phosphosugars into their corresponding products. Monosaccharide production rates for the variant enzymes were higher than those obtained using the wild-type enzymes (Fig. 5).

Fig 4.

Relative catalytic efficiencies of the wild-type and variant enzymes of CDRPI, BSMPI, and PFGPI for phosphosugars and monosaccharides. (A) Relative activities of the wild-type and variant enzymes of CDRPI for ribose-5-phosphate and d-ribose. The relative catalytic efficiencies of 100% for ribose-5-phosphate and d-ribose were 500 and 0.6 mM⁻¹ s⁻¹, respectively. (B) Relative activities of the wild-type and variant enzymes of BSMPI for mannose-6-phosphate and l-ribose. The relative catalytic efficiencies of 100% for mannose-6-phosphate and l-ribose were 2,014 and 13 mM⁻¹ s⁻¹, respectively. (C) Relative activities of the wild-type and variant enzymes of PFGPI for glucose-6-phosphate and l-talose. The relative catalytic efficiencies of 100% for glucose-6-phosphate and l-talose were 2,284 and 3.6 mM⁻¹ s⁻¹, respectively. The black and white bars represent relative activities for phosphosugar and monosaccharide, respectively. The data represent the means from three separate experiments, and the error bars represent standard deviations.

Table 2.

Kinetic parameters of the wild-type and variant enzymes at position 132 of CDRPI for d-ribose, at position 192 of BSMPI for l-ribose, and at position 85 of PFGPI for l-talose^a

Enzyme	Km (mM)	kcat (s−1)	kcat/Km (mM−1 s−1)
CDRPI
Wild type	245 ± 10	139 ± 8	0.56 ± 0.04
R132A	217 ± 4	132 ± 3	0.61 ± 0.02
R132I	320 ± 31	71 ± 5	0.22 ± 0.03
R132Q	204 ± 11	106 ± 3	0.52 ± 0.03
R132K	265 ± 4	149 ± 3	0.56 ± 0.01
R132E	217 ± 0.4	148 ± 1	0.68 ± 0.005
R132Y	292 ± 5	161 ± 1	0.41 ± 0.005
R132D	216 ± 5	214 ± 3	0.99 ± 0.03
BSMPI
Wild type	688 ± 13	9,095 ± 91	13.2 ± 0.3
R192A	722 ± 43	4,653 ± 113	6.5 ± 0.4
R192N	569 ± 27	26,113 ± 886	45.9 ± 2.7
R192K	792 ± 4	7,348 ± 47	9.3 ± 0.08
R192E	789 ± 61	6,331 ± 259	17.6 ± 1.0
R192L	590 ± 12	6,293 ± 45	11.0 ± 0.2
R192Y	998 ± 44	17,595 ± 670	17.6 ± 1.0
PFGPI
Wild type	133 ± 4.9	475 ± 7	3.6 ± 0.1
T85A	186 ± 5.6	960 ± 24	5.2 ± 0.2
T85S	146 ± 3.6	381 ± 5	2.6 ± 0.1
T85Q	100 ± 2.5	1,756 ± 22	17.6 ± 0.5
T85D	185 ± 4.7	448 ± 7	2.4 ± 0.1
T85K	205 ± 6.1	396 ± 14	1.9 ± 0.1

^aThe data represent the means and standard deviations from three separate experiments.

Fig 5.

Production of phosphosugars and monosaccharides by the wild-type and variant enzymes of CDRPI, BSMPI, and PFGPI. (A) Production of d-ribulose (open symbol) from d-ribose and of ribulose-5-phosphate (closed symbol) from ribose-5-phosphate by the wild-type (circle) and R132D variant (square) CDRPIs. (B) Production of l-ribose (open symbol) from l-ribulose and of fructose 6-phosphate (closed symbol) from mannose 6-phosphate by the wild-type (circle) and R192N variant (square) BSMPIs. (C) Production of l-tagatose (open symbol) from l-talose and of fructose 6-phosphate (closed symbol) from glucose 6-phosphate by the wild-type (circle) and T85Q variant (square) PFGPIs. The data represent the means from three separate experiments, and the error bars represent standard deviations.

Specifically, authentic substrates of the phosphosugar isomerase variants R133D of CDRPI, R192 of BSMPI, and T85Q of PFGPI were converted from phosphosugars to monosaccharides. These variants are no longer an RPI, MPI, and GPI, respectively; instead, they have been changed into a d-ribose isomerase, an l-ribose isomerase, and an l-talose isomerase, respectively, which do not exist in nature. These novel enzymes can contribute rare monosaccharide production. Therefore, novel isomerases were developed based on phosphosugar isomerases via substrate-tailored optimization.

l-Ribose has been used as a starting material of l-nucleoside-based pharmaceuticals (26) and potent antiviral agents for hepatitis B virus and Epstein-Barr virus (27). Its chemical derivatives involve the inhibition of the viral nucleoside synthesis-replication process by exploiting the minor difference in the nucleoside synthesis process between a normal cell and a virus. l-Talofuranosyladenine, an l-talose nucleoside derivative, can be used as a slowly reacting substrate for calf intestinal adenosine deaminase and an inhibitor for the growth of leukemia cells in vitro (28). d-Ribose has been used as a precursor in the synthesis of nucleotide flavor enhancers and riboflavin (vitamin B₂) (29). Enzymes that can be used in the biosynthesis of these monosaccharides should be developed. While l-talose isomerase and d-ribose isomerase have not been reported, one l-ribose isomerase has been described (30). However, this l-ribose isomerase exhibited low activity and no extensive homology with MPI. Thus, the phosphate-binding site variant of MPI described above is a new type of efficient l-ribose isomerase. Recently, we applied the phosphate-binding site variant of MPI from Thermus thermophilus (TTMPI R142N) to produce l-ribose, and the enzyme exhibited the highest activity and productivity for l-ribose production ever reported (31). This enzymatic method is superior to the chemical synthetic method presently used in the manufacturing process due to a higher productivity.

The substrate specificity of TTMPI was similar to that of BSMPI. The catalytic efficiencies of TTMPI and its R142N variant (134 and 174 mM⁻¹ s⁻¹) for l-ribose were higher than those of BSMPI and its variant R192N (13 and 46 mM⁻¹ s⁻¹), whereas the increase of the catalytic efficiency by mutation of BSMPI was higher than that by mutation of TTMPI. TTMPI was used in a previous study for increasing l-ribose production (31), whereas BSMPI was used in this study for the investigation of the general role of phosphate-binding residues in the phosphosugar isomerases.

Structural analysis of novel sugar isomerases.

When ribose-5-phosphate, mannose-6-phosphate, and glucose-6-phosphate were docked to CDRPI, BSMPI, and PFGPI, respectively, Arg133, Arg192, and Thr85 interacted directly with the phosphate groups of the phosphosugars (Fig. 3A, C, and E). The phosphate group is located at the end of the monosaccharide moiety and may be critical for defining the substrate specificity of the corresponding phosphosugar isomerase. When monosaccharides were docked to modeled structures, phosphate-binding residues in the phosphosugar isomerases did not interact tightly with the terminal hydroxyl groups of the monosaccharides. Thus, the phosphate-binding sites of the phosphosugar isomerases were molecular determinants with different catalytic activities for phosphosugars and monosaccharides. Furthermore, optimization was accomplished by these sites with other amino acids to develop novel sugar isomerases.

When the monosaccharide substrates were docked to the active-site pockets of phosphosugar isomerases in a ligand docking study, distances from the terminal hydroxyl of the monosaccharides d-ribose, l-ribose, and l-talose to the side chains of CDRPI R133D (2.67 Å) (1 Å [0.1 nm]), BSMPI R192N (2.38 Å), and PFGPI T85Q (2.26 Å) variant enzymes were shorter than those of the respective wild-type enzymes (4.74, 3.85, and 4.83 Å, respectively) (Fig. 3B, D, and F). Therefore, we suggest that these shorter distances between the phosphate-binding sites and terminal hydroxyl groups of monosaccharides explain the enhanced k_cat/K_m values obtained for monosaccharide substrates compared to those of wild-type enzymes. The variant enzymes exhibited higher activities for other monosaccharides than the wild-type enzymes (Table 1). However, the actual structure of these wild-type and variant enzyme complexes with substrates must be obtained to provide further evidence for these identifications.

In summary, new sugar isomerases for the biosynthesis of monosaccharides were developed from phosphosugar isomerases by the substrate-tailored optimization method. Each of these new sugar isomerases dissipated the authentic function of phosphosugar isomerases and reinforced catalytic activity for monosaccharide biosynthesis. The crystal structures and homology models in complex with phosphosugars and monosaccharides should allow exploration of how altering the enzyme affects the catalytic properties of the protein at the molecular level. Our findings may be used for the enzymatic synthesis of chemicals not found in nature and may be applied to the establishment of new enzymes from naturally occurring enzymes.