Characterization of a Novel Metal-Dependent D-Psicose 3-Epimerase from Clostridium scindens 35704

The noncharacterized protein CLOSCI_02528 from Clostridium scindens ATCC 35704 was characterized as D-psicose 3-epimerase. The enzyme showed maximum activity at pH 7.5 and 60°C. The half-life of the enzyme at 50°C was 108 min, suggesting the enzyme was relatively thermostable. It was strictly metal-dependent and required Mn2+ as optimum cofactor for activity. In addition, Mn2+ improved the structural stability during both heat- and urea-induced unfolding. Using circular dichroism measurements, the apparent melting temperature (T m) and the urea midtransition concentration (C m) of metal-free enzyme were 64.4°C and 2.68 M. By comparison, the Mn2+-bound enzyme showed higher T m and C m with 67.3°C and 5.09 M. The Michaelis-Menten constant (K m), turnover number (k cat), and catalytic efficiency (k cat/K m) values for substrate D-psicose were estimated to be 28.3 mM, 1826.8 s−1, and 64.5 mM−1 s−1, respectively. The enzyme could effectively produce D-psicose from D-fructose with the turnover ratio of 28%.


Introduction
D-Psicose (D-ribo-2-hexulose or D-allulose), the C3 epimer compound of D-fructose, is considered as a kind of rare sugar, which was defined by the International Society of Rare Sugars (ISIR) as monosaccharides and their derivatives existing in nature in extremely small quantities (The 1 st International Symposium of ISRS, Takamatsu, Japan, 2002). It has 70% relative sweetness but only 0.3% energy of sucrose [1], and is suggested as an ideal sucrose substitute for food products [2]. It has important physiological functions, such as blood glucose suppressive effect [3,4], reactive oxygen species scavenging activity [5,6], and neuroprotective effect [7]. In addition, it can improve the gelling behavior and produces good flavor during food process [8,9]. Importantly, it has been approved as ''generally regarded as safe'' (GRAS) by Food and Drug Administration (FDA) in Jun. 2012 (GRN No. 400), and has been allowed to be used as ingredient in a range of foods and dietary supplements. Commercial bioproduction of D-psicose may be realized through epimerization from Dfructose using D-tagatose 3-epimerase (DTEase) family enzymes, which have recently attracted attention of many researchers by its importance in application of rare sugar bioproduction [2].
The crystal structures of A. tumefaciens DPEase [17], P. cichorii DTEase [18], and C. cellulolyticum DPEase [19] have been identified. All the enzymes assemble into a tetramer and the subunits have around 295 amino acid residues with the molecular weight of approximately 32 kDa. They show the similar (b/a) 8 TIM barrel structure and display a divalent metal coordinating site in the enzyme's active center. Through multiple sequence alignment of reported DTEase family enzymes, they show relatively low similarity (20-60% identity) of amino acid sequence, but the amino acid residues responsible for metal coordination are strictly conserved in all the enzymes, suggesting that the divalent metal ion is possibly important to the structure and function of the enzyme as the cofactor. Interestingly, the abovementioned DTEase family enzymes show different requirement for metal cofactor to the activity and stability of the enzyme. P. cichorii DTEase does not require any cofactor for its activity, and even loses activity under some metal ions [20]. R. sphaeroides DTEase [14], A. tumefaciens DPEase [13], and Ruminococcus sp. DPEase [16] can display activity without metal ion, but their activity is remarkably enhanced by metal ions especially Mn 2+ . However, C. cellulolyticum DPEase is strictly metal-dependent and does not show activity in absence of metal ions [15]. C. cellulolyticum DPEase requires metal ion as essential cofactor to display activity and shows the maximal activity in the presence of Co 2+ . In addition, the Co 2+ can significantly improve the thermostability of C. cellulolyticum DPEase [15].
In this study, the gene encoding the hypothetical protein CLOSCI_02528 with protein ID ZP_02432283 from the Clostridium scindens ATCC 35704 was cloned and overexpressed in E. coli. The hypothetical protein CLOSCI_02528 was identified as a new member of DTEase family enzymes. The enzyme was strictly metal-dependent like C. cellulolyticum DPEase, and its optimum substrate was D-psicose. The enzyme properties and kinetic parameters were determined, and compared to those of other DTEase family enzymes. Especially, the effect of metal ion on the activity and stability of the enzyme was studied in detail.

Chemicals and Reagents
The resin for protein purification, the Chelating Sepharose Fast Flow, was obtained from GE (Uppsala, Sweden). The pET-22b(+) expression vector was obtained from Novagen (Darm-stadt, Germany). Electrophoresis reagents were purchased from Bio-Rad. Isopropyl b-D-1-thiogalactopyranoside (IPTG) and all chemicals used for enzyme assays and characterization were at least of analytical grade obtained from Sigma (St Louis, MO, USA) and Sinopharm Chemical Reagent (Shanghai, China).

Gene Cloning and Protein Expression
The full-length nucleotide sequence of the gene encoding the hypothetical protein CLOSCI_02528 from the C. scindens ATCC 35704 was synthesized and incorporated with NdeI and XhoI sites in the 59-and 39-terminal of the gene, respectively, and then cloned into pUC57 vector by Shinegene Molecular Biothnology Co., Ltd (Shanghai, China). Then the plasmid was sub-cloned into pET-22b(+) vector with NdeI and XhoI sites, therefore, an in-frame fusion 66histidine-tag sequence at the C-terminus was provided in the reconstructed plasmid, named by pET-cs-dpe. And the pETcs-dpe was transformed into E. coli BL21(DE3).
The recombinant E. coli for protein expression was cultivated with shaking (200 rpm) in 400 ml of LB medium containing 50 mg/ml ampicillin at 37uC. When the OD 600 reached 0.6, IPTG was added at 0.5 mM, and DPEase was induced and overexpressed at 30uC for 6 h, and harvested by centrifugation at 4uC for 10 min at 10,000 g. The enzyme was expressed as 66histidinetagged fusion protein, which was available for affinity chromatography.

Protein Purification
Four gram of wet cell pellets were resuspended in 20 mL of lysis buffer (50 mM Tris-HCl, 100 mM NaCl, pH 7.5) and disrupted by sonication at 4uC for 6 min (pulsations of 3 s, amplify 90) using a Vibra-Cell TM 72405 Sonicator. The lysate was centrifuged at 20,000 g for 30 min at 4uC. Then, the 66histidine-tagged fusion enzyme was purified by nickel-affinity chromatography (GE Healthcare) according to manufacturer's protocol (Instructions 71-5001-87 AE; GE Healthcare). The 15 mL of supernatant was loaded onto Chelating Sepharose Fast Flow resin column (1.0 cm610.0 cm), previously chelating NiSO 4 and equilibrated with a binding buffer (50 mM Tris-HCl, 500 mM NaCl, pH 7.5). After washing with the washing buffer (50 mM Tris-HCl, 500 mM NaCl, 50 mM imidazole, pH 7.5) to remove unbound fractions, the DPEase was eluted from the column with an elution buffer (50 mM 50 mM Tris-HCl, 500 mM NaCl, 500 mM imidazole, pH 7.5). The active fractions were pooled and dialyzed overnight against 50 mM Tris-HCl buffer (pH 7.5) containing 10 mM ethylenediamine tetraacetic acid (EDTA) at 4uC. Subsequently, the enzyme was dialyzed against 50 mM EDTA-free Tris-HCl buffer (pH 7.5) using a 10-kDa cut-off membrane (Millipore). Protein concentration was determined by the method of Bradford using bovine serum albumin as a standard. SDS-PAGE was carried out according to Laemmli. Gels (12% w/v polyacrylamide) were stained with Coomassie Brilliant Blue and destained with an aqueous mixture of 10% (v/v) methanol/10% (v/v) acetic acid.

Enzyme Assay
The activity was measured by the determination of the amount of produced D-psicose from D-fructose. The reaction mixture of 1 mL contained D-fructose (50 g/L), Tris-HCl buffer (50 mM, pH 7.5), 1 mM Mn 2+ , and 1 mM enzyme. The reaction was carried out at 60uC for 10 min, and was terminated by boiling for 5 min. The generated D-psicose was determined by the HPLC method. One unit of enzyme activity was defined as the amount of enzyme catalyzing the formation of 1 mmol D-psicose per min.

Effect of Temperature, pH, and Metal Ions
The optimum temperature of enzyme activity was measured by assaying the enzyme samples over the range of 35-75uC for 2 min. Two buffer systems, sodium phosphate (50 mM, pH 5.5-7.0) and Tris-HCl (50 mM, pH 7.5-9.0), were used for measuring the optimum pH of enzyme activity. The thermal stability of the enzyme was studied by incubating the enzyme in Tris-HCl buffer (50 mM, pH 7.5) at various temperatures. At given time intervals, samples were withdrawn and the residual activity was measured under standard assay conditions. To determine the pH stability, the enzyme was incubated at pH 5.5-9.0 at 4uC for up to 2 h, and the remaining enzyme activity was measured at time intervals under standard assay conditions. To determine the effects of cations, the samples were incubated in the presence of the different divalent metal ions at 1 mM: MnCl 2 , CoCl 2 , FeCl 2 , NiCl 2 , and MgCl 2 , respectively.

Determination of Kinetic Parameters
Kinetic parameters of C. scindens DPEase were determined in 50 mM Tris-HCl buffer (pH 7.5) containing 1 mM Mn 2+ and Dpricose, D-fructose, or D-tagatose as the substrate from 5 to 200 mM. Kinetic parameters, such as Michaelis-Menten constant (K m ) and turnover number (k cat ) values for substrates were obtained using the Lineweaver-Burk equation and quantification of enzyme concentration.

Circular Dichroism Measurements and Protein Unfolding Analysis
Circular dichroism (CD) measurements were carried out with a Jasco J-810 spectropolarimeter with a Peltier temperaturecontrolled cuvette holder. For thermal unfolding, protein concentration was set as 0.4 mg/ml in 50 mM Tris-HCl, and protein unfolding was monitored by CD ellipticity at 220 nm from 40 to 86uC at a heating rate of 1uC/min. Samples were overlaid with mineral oil and the lids of cuvette were sealed with paraffin tape to prevent evaporation. For urea-induced unfolding, protein samples (0.5 mg/ml in 50 mM Tris-HCl) were diluted in urea (0.32 mM increments up to a final urea concentration between 0-6 M), pre-equilibrated for 2 h at room temperature, and the change in the folded fraction was monitored by a CD signal at 220 nm at 25uC.
The ratio of denatured and total protein in the transition range, f d = (e N -e)/(e N -e U ), was calculated from their CD signal relative to the signal of the native and unfolded base line e N and e U . Melting concentration (T m ) and urea mid-transition concentration (C m ) were defined as the transition mid-point of thermal and ureainduced protein unfolding, respectively.

HPLC Measurements of Monosaccharides
The concentrations of D-fructose and D-psicose were analyzed by HPLC equipped with a refractive index detector and a Ca 2+carbohydrate column (Waters Sugar-Pak 1, Waters Corp., Milford, MA, USA). The monosaccharides were separated by Table 1. Comparison of biochemical properties and kinetic parameters of DTEase family enzymes.

Amino Acid Sequence Alignment of DTEase Family Enzymes
Compared to the reported microbial DTEase family enzymes, the DPEase from C. scindens 35704 (hypothetical protein CLOSCI_02528 with protein ID ZP_02432283) was found relatively homologous in amino acid sequence with the family enzymes from C. cellulolyticum H10, A. tumefaciens, Ruminococcus sp. 5_1_39BFAA, P. cichorii, and R. sphaeroides SK011, with 62%, 58%, 52%, 38%, and 29% identity, respectively. Although the homology among the family enzymes was not high (20-60%) in amino acid sequence, they had highly similarities in the key amino acid residues of active sites, metal coordinating sites and substrate combinant binding sites (Fig. 1). From the known structure information of A. tumefaciens DPEase [17], P. cichorii DTEase [18], and C. cellulolyticum DPEase [19], the metal coordinating sites were predicted consisting of Glu149, Asp182, His208, and Glu243 in C. scindens DPEase, and were strictly conserved in all DTEase family enzymes. The amino acid residues Glu156, His186, and Arg215, which were responsible for the interaction between the enzyme and O1, O2, and O3 of substrate, were also conserved in the family enzymes. However, interestingly, as for the residues providing a hydrophobic environment around the substrate, which consisted of Tyr6, Try14, Gly65, Ala107, and Phe245, C. scindens DPEase showed absolute identity with all reported DPEases, but was significantly different with the DTEases (Fig. 1). It was probably caused by the substrate binding sites composed by different amino acid residues in the structure.

Cloning, Expression, and Purification of C. scindens DPEase
The C. scindens 35704 genomic sequence was completed and deposited as GenBank accession number NZ_DS499706. According to the genome sequencing, the nucleotide sequence of the gene coding hypothetical protein CLOSCI_02528 (characterized as DPEase) was given as Gene ID 167760156. This DPEase-coding gene was obtained by the full-length gene synthesis, instead of the PCR amplification in vitro using genome as DNA template. The coding gene sequence was intentionally incorporated with NdeI and XhoI sites in the 59-and 39-terminal of the gene, respectively, and then was cloned into pET-22b(+) vector with NdeI and XhoI sites, therefore, an in-frame fusion 66histidine-tag sequence at the C-terminus was provided in the reconstructed plasmid, named by pET-cs-dpe. This allowed for a generic single step purification of the recombinant DPEase using nickel-affinity column chromatography.
The pET-cs-dpe was then introduced into the expression host, E. coli BL21(DE3), to overexpress DPEase by IPTG induction. SDS-PAGE analyses on the extracts of E. coli BL21(DE3) harboring pET-Csc-dpe induced by IPTG revealed the presence of large amounts of protein around 31 kDa, compared to that of the control E. coli BL21(DE3) cells harboring plasmid pET-22b(+); and the molecular weight was in agreement with the predicted molecular mass for the DPEase protein, suggesting the DPEase was overexpressed by IPTG induction (Fig. 2A). After IPTG induction, the E. coli cells were harvested by centrifugation and disrupted by sonification. Cell debris was removed by centrifugation at 4uC and 20,000 g for 20 min. The supernatant was subjected to the Ni 2+ -Chelating Sepharose Fast Flow affinity column, and the recombinant DPEase fused by C-terminal 66histidine-tag was purified to electrophoretic homogeneity with a single band on SDS-PAGE (Fig. 2B). The molecular weight of the purified C. scindens DPEase was estimated to be approximately 32 kDa by SDS-PAGE, however, the gel filtration experiment data showed the molecular weight of native enzyme was approximately 125 kDa, suggesting the enzyme was a tetramer with four identical subunits. The tetramer manner of subunits assembly was also characterized in other DTEase family enzymes, such as A. tumefaciens DPEase [13], C. cellulolyticum DPEase [15], and Ruminococcus sp. DPEase [19]; however, the R. sphaeroides SK011 DTEase [14] and P. cichorii DTEase were determined as dimers [21].

Effects of Temperature and pH on C. scindens DPEase
The temperature and pH profiles for C. scindens DPEase activity were shown in Fig. 3A and 3B, respectively. The enzyme activity was detected in presence of 1 mM Mn 2+ , because it displayed null activity without metal ions. The optimum temperature and pH for enzyme activity are 7.5 and 60uC. Comparison of biochemical properties of C. scindens DPEase and other reported DTEase family enzymes was shown in Table 1. The optimum temperature of the known DTEase family enzymes is the range of 40-60uC and the optimum pH is all weakly alkaline ranged from 7.5 to 9.0.
C. scindens DPEase retained 95 and 86% of its initial activity after 2 h of exposure at 40 and 45uC (Fig. 3C), which indicated that the enzyme was relatively stable below 50uC. The half-life of the enzyme at 50uC was approximately 108 min, which was longer than that of, A. tumefaciens DPEase (63.5 min) [13] and P. cichorii DTEase (approximately 60 min) [20], but shorter than that of C. cellulolyticum DPEase [15] and Ruminococcus sp. DPEase [16]. Like other wild DTEase family enzymes, C. scindens DPEase was easily deactivated under exposure at higher temperatures of more than 55uC. When incubated at 60uC, the half-life of C. scindens DPEase was around 50 min, which was shorter than that of Ruminococcus sp. DPEase (1.6 h) [16], but longer than other wild DTEase family enzymes. In addition, more than 80% activity of the purified C. scindens DPEase was retained pH 6.0-9.0 after incubation at 4uC for 2 h (Fig. 3D), indicating that the C. scindens DPEase was relatively stable under neutral and weakly alkaline conditions.
Thermostability is an important property of DTEase family enzymes to realize the industrial enzymatic production of Dpsicose. Some DTEase family enzymes could be more thermostable in presence of metal ion cofactor, such as C. cellulolyticum DPEase [15] and Ruminococcus sp. DPEase [16]. The thermostability also could be improved through the molecular modification. Choi et al. constructed the I33L S213C double-site variant of A. tumefaciens DPEase, and the variant enzyme showed significant increases in optimal temperature, half-life, and melting temperature, compared with the wild-type enzyme; and the variant enzyme's half-life at 50uC reached 1,853 min, which was 29.9-fold of that of the wild-type enzyme [22].
Effects of Metal Ions on C. scindens DPEase Activity C. scindens DPEase displayed null activity in absence of metal ion, and only had catalysis activity when adding divalent metal ion. Also, even in presence of metal ion, the enzyme would lose all activity when adding EDTA as metal ion chelating agent. Therefore, the C. scindens DPEase was a kind of strictly metaldependent enzyme, which is similar to C. cellulolyticum DPEase, but different with other DTEase family enzymes [15].
To investigate the effect of metal ions, the relative activity of C. scindens DPEase was detected in presence of different divalent metal ions, which were added at the final concentration of 1 mM. C. scindens DPEase showed a maximal activity in the presence of Mn 2+ ; however, when Mn 2+ was replaced with Co 2+ , Fe 2+ , Ni 2+ , and Mg 2+ , the enzyme activity was reduced to 92%, 70%, 69%, and 41% of that in presence of Mn 2+ , respectively. By comparison, C. cellulolyticum DPEase is also metal-dependent but with optimum metal ion of Co 2+ [15]; P. cichorii DTEase does not require any cofactor for its activity [20]; Ruminococcus sp. DPEase [16], A. tumefaciens DPEase [13], and R. sphaeroides DTEase [14] can display activity without metal ion, but their activity is significantly enhanced by metal ions especially Mn 2+ or Co 2+ .
In addition, the effect of metal cofactor concentration on catalytic activity of metal-dependent DPEase was also investigated. The Mn 2+ concentration titration on catalytic activity with 1 mM C. scindens DPEase was developed, and the overall results of the Mn 2+ concentration effect on activity was like S-curve, and different with general first order reaction model (Fig. 4A). The activity could not be detected when the Mn 2+ concentration was below 5 mM, probably because the enzyme needed a certain amount of Mn 2+ concentration to effectively bind the metal cofactor. Then, the relative activity increased linearly with the addition of metal ion with between 5 and 60 mM, which accorded with first order reaction (Fig. 4B). When the Mn 2+ concentration was more than 60 mM, the relative activity tended to be stationary, probably because the binding of enzyme to metal ion was close to saturation (Fig. 4A). The similar result was reported on another metal-dependent enzyme, Geobacillus stearothermophilus L-arabinose isomerase, which also displayed null activity without metal ions, but rapidly restored catalytic activity by adding trace amount of Mn 2+ [23]. Interestingly, for aldose isomerase or ketose epimerase having metal coordinating region in structure, some enzyme are metal-dependent and must require metal cofactor to display activity [15,23,24]; some not only do not require metal ions but also could not be activated by metal ions [20,25,26]; and other could display catalytic activity without metal ion, however, their activity could be remarkably enhanced by metal ions [13,14,16]. The mechanism of different metal independence of these enzymes having metal coordinating region may be fascinating topic worthy of deep investigation.

Effects of Mn 2+ on Structural Stability of C. scindens DPEase
To study the effect of Mn 2+ on thermostability based on the structural aspects, thermal unfoldings of both enzymes in the absence and the presence of Mn 2+ were performed using CD. Shown in Fig. 5A, metal-free C. scindens DPEase had the apparent melting temperature (T m ) of 64.4uC, whereas the metal-bound enzyme showed higher T m of 67.3uC. Therefore, it was suggested that the metal ion was important for C. scindens DPEase with respect to the unfolding and structural stability during thermal denaturation. Similar result about the improvement effect of metal ion on the thermostability was reported on the metal-dependent G. stearothermophilus L-arabinose isomerase, in which the T m of metalbound enzyme was 6uC higher than that of the metal-free enzyme [23].
In addition to the thermal unfolding, the urea-induced unfoldings of both metal-free and metal-bound C. scindens DPEases were also studied using CD. According to the urea-induced denaturation curves (Fig. 5B), the urea midtransition concentration (C m ) was derived, and the C m of metal-bound C. scindens DPEase (5.09 M) was much higher than that of metal-free enzyme (2.68 M), suggesting that the metal ion effectively improved the structural stability of enzyme against the protein denaturant urea. Therefore, the metal ion not only was the essential cofactor for C. scindens DPEase to have the catalysis activity, but also could improve the structural stability during both heat-and ureainduced unfolding.
The Michaelis-Menten constant (K m ), turnover number (k cat ), and catalytic efficiency (k cat /K m ) values of C. scindens DPEase for substrate D-psicose were estimated to be 28.3 mM, 1826.8 min 21 , and 64.5 mM 21 min 21 , respectively. And the k cat /K m for Dpsicose was much higher than that for other substrates, and was 7.4-fold and 258-fold of that for D-fructose and D-tagatose, respectively. These data are very similar to those of Ruminococcus sp. DPEase [16], C. cellulolyticum [15] and A. tumefaciens DPEases [13], showing these four DPEases had the same substrate specificity ( Table 1).
In conclusion, the noncharacterized gene CLOSCI_02528 from C. scindens ATCC 35704, encoding the hypothetical protein ZP_02432283, was cloned and expressed in E. coli, and the protein was characterized as a member of DTEase family enzymes with the optimum substrate of D-psicose. The enzyme showed maximum activity at pH 7.5 and 60uC. It was strictly metaldependent and required Mn 2+ as optimum cofactor for activity. In addition, Mn 2+ improved the structural stability during both thermal and urea-induced unfolding. The k cat /K m for D-psicose and D-fructose was estimated to be 64.5 and 8.72 mM 21 min 21 , respectively. The equilibrium ratio between D-psicose and Dfructose of the enzyme was 28:72.