Characterization of Glycerol Dehydrogenase from Thermoanaerobacterium thermosaccharolyticum DSM 571 and GGG Motif Identification

Abstract

Glycerol dehydrogenases (GlyDHs) are essential for glycerol metabolism in vivo, catalyzing its reversible reduction to 1,3-dihydroxypropranone (DHA). The gldA gene encoding a putative GlyDH was cloned from Thermoanaerobacterium thermosaccharolyticum DSM 571 (TtGlyDH) and expressed in Escherichia coli. The presence of Mn²⁺ enhanced its enzymatic activity by 79.5%. Three highly conserved residues (Asp¹⁷¹, His²⁵⁴, and His²⁷¹) in TtGlyDH were associated with metal ion binding. Based on an investigation of glycerol oxidation and DHA reduction, TtGlyDH showed maximum activity towards glycerol at 60℃ and pH 8.0 and towards DHA at 60℃ and pH 6.0. DHA reduction was the dominant reaction, with a lower K_m(DHA) of 1.08 ± 0.13 mM and V_max of 0.0053 ± 0.0001 mM/s, compared with glycerol oxidation, with a K_m(glycerol) of 30.29 ± 3.42 mM and V_max of 0.042 ± 0.002 mM/s. TtGlyDH had an apparent activation energy of 312.94 kJ/mol. The recombinant TtGlyDH was thermostable, maintaining 65% of its activity after a 2-h incubation at 60℃. Molecular modeling and site-directed mutagenesis analyses demonstrated that TtGlyDH had an atypical dinucleotide binding motif (GGG motif) and a basic residue Arg⁴³, both related to dinucleotide binding.

Introduction

In vivo glycerol metabolism involves many different kinds of proteins, among which glycerol dehydrogenase (GlyDH, E.C.: 1.1.1.6) primarily catalyzes the conversion of glycerol to 1,3-dihydroxypropranone (DHA) coupled with the reduction of nicotinamide adenine dinucleotide (NAD+). GlyDHs effectively regulate physiological processes related to energy production, exchange, and consumption and have been isolated from a variety of prokaryotic and eukaryotic cells, including Escherichia coli [27], Bacillus megaterium [22], Bacillus stearothermophilus [26], Clostridium butyricum [18], and Schizosaccharomyces pombe [13]. Based on its PROSITE description, there are three known types of alcohol dehydrogenases (ADHs): zinc-containing “long-chain” alcohol dehydrogenases (Zn-ADHs), insect-type or “short-chain” alcohol dehydrogenases, and iron-containing alcohol dehydrogenases (Fe-ADHs). Of these, bacterial GlyDHs are closely related to Fe-ADHs. However, most reported GlyDHs are NAD+-linked and strictly zinc-dependent metalloenzymes. Moreover, most studies on GlyDH have been conducted in the thermophilic species B. stearothermophilus, due to the availability of its crystal structure [5,30]. However, biochemical assays in this species are conducted at 30℃, lower than its optimal growth temperature of 55℃ [21,26]. To date, biochemical characterization of GlyDHs in thermophiles has not been conducted at high temperatures.

The Rossmann fold in dehydrogenases is one of the most common structural features of supersecondary structures in many oxidoreductases that bind NAD+, nicotinamide adenine dinucleotide phosphate (NADP+), and related cofactors. Since Rossmann first described the dinucleotide binding fold in 1974 based on the structural alignment of four dehydrogenases (lactate, malate, alcohol, and glyceraldehyde-3-phosphate dehydrogenases) [20], a great deal of structural data on classical dinucleotide binding proteins have suggested that the initial βαβ fold is the most conserved unit in the Rossmann fold [8] and typically contains a phosphate binding motif (GXGXXG, where X is any amino acid) [3]. However, these traditional structural features provide insufficient information to interpret the interactions between the Rossmann fold and nicotinamide dinucleotides for all GlyDHs.

The lack of structural information on GlyDHs in thermophiles is the main bottleneck to determining the key residues within its active and dinucleotide binding sites. To date, only six delicate crystal structures of GlyDHs have been resolved and deposited in the Brookhaven Protein Data Bank (PDB), including three GlyDH structures from B. stearothermophilus (PDB entry: 1JQ5, 1JPU, and 1JQA) [5,21,30] and one structure each from Thermotoga maritima (PDB entry: 1KQ3) [11], Clostridium acetobutylicum (PDB entry: 3CE9), and Serratia plymuthica (PDB entry: 4MCA) [14]. Moreover, the detailed mechanism of GlyDHs has not been elucidated owing to the limitation of structural information and biochemical data. However, combining protein engineering techniques and protein structure prediction could provide an alternative approach to studying GlyDH function and mechanism. Rational design has been performed with varying degrees of success to identify potential active residues and improve the given properties of enzymes based on in silico prediction and modeling. For example, redesigning the coenzyme specificity of a dehydrogenase using protein engineering was conducted in 1990 [24], and the coenzyme specificities of an ADH from Rana perezi [19] and a xylitol dehydrogenase from Pichia stipitis [29] were reversed completely by substituting the key residues necessary for coenzyme binding, using the assistance of structural simulation.

Compared with mesophiles, thermophile-derived enzymes are relatively thermostable and tolerant to industrial production conditions such as high salt or solvent concentrations and high operation temperatures [4,25,28], and considerable research efforts in recent years have focused on thermostable enzymes. Thermoanaerobacterium thermosaccharolyticum is a thermophilic obligate anaerobic bacterium, and the genomic data from T. thermosaccharolyticum DSM 571 are accessible from the GenBank database [6,16]. However, there have been few reports on the biochemical or structural characterization of T. thermosaccharolyticum GlyDH.

In this study, we examined a putative GlyDH from T. thermosaccharolyticum DSM 571 (TtGlyDH) and described its cloning, expression, and biochemical characterization. We found that the GGG motif and a basic amino acid residue are critical for dinucleotide binding in TtGlyDH, based on rational design and site-directed mutagenesis.

 

Materials and Methods

Chemicals, Bacterial Strains, and Culture Conditions

All chemicals were purchased from Sangon Biotech (Shanghai, China), unless otherwise stated. All DNA restriction enzymes, T4 polynucleotide kinase, and ligase were purchased from TaKaRa (Dalian, China). Phusion High-Fidelity DNA Polymerase was purchased from New England Biolabs (Ipswich, MA, USA). The BIOMEGA PCR Purification Kit and Mini Plasmid Extraction Kit (Shanghai, China) were purchased for DNA purification and plasmid isolation. The genomic DNA from T. thermosaccharolyticum DSM 571 was purchased from the German Collection of Microorganisms and Cell Cultures (DSMZ; http://www.dsmz.de/). E. coli TOP10 was used for plasmid propagation. E. coli BL21 (DE3) was used as the host for heterologous expression. The E. coli strains were cultured at 37℃ in LB medium supplemented with ampicillin (100 μg/ml) when required for plasmid maintenance.

Construction of Plasmids and Strains

DNA manipulations were performed by following standard procedures. A pair of specific oligonucleotide primers (see below) for amplifying gldA was designed based on its reference DNA sequence (Gene ID: 9707383, Tthe_1821). The gldA coding sequence was amplified using genomic DNA from T. thermosaccharolyticum DSM 571 as a template:

5’-GGGAATTCCATATGACAAAAGCTATAATAGGCCCTTCG-3’ (forward)

5’-CCGCTCGAGTCTAGATCTCTTATTTTTGTACATTTTTCC-3’ (reverse)

The underlined letters represent the NdeI and XhoI restriction sites, respectively. The integrity and yield of the PCR products were assessed using agarose gel electrophoresis. The resulting PCR fragments were digested with NdeI and XhoI and ligated into the commercial vector pET-20b in frame with the His6 tag. The ligation mixtures were transformed into E. coli TOP10 by heat shocking the chemically competent cells for 90 sec at 42℃ without shaking. The transformants were screened on LB plates containing 100 μg/ml ampicillin. The transformed gldA was confirmed by DNA sequencing.

Expression and Purification of Recombinant TtGlyDH

Expression strain BL21 (DE3) cells harboring pET-20b-gldA were grown in 50 ml of LB (100 μg/ml ampicillin) at 37℃ and 180 rpm. IPTG was added to a final concentration of 0.5 mM to induce expression until the culture reached the stationary phase (OD600 = 0.6–0.8) and was incubated for an additional 3 h at 25℃ at 120 rpm. The cultured cells were harvested and centrifuged at 4℃ and 10,000 ×g for 5 min. The cell pellets were washed twice with ice-cold water to remove residual medium and suspended in 5 ml of binding buffer (5 mM imidazole, 0.5 M NaCl, and 20 mM Tris-HCl buffer, pH 7.9), followed by sonication on ice. The crude extract was heat-treated (60℃, 30min) and clarified by centrifugation (15,000 ×g, 4℃, 20 min) to remove cell debris and precipitates. The soluble fraction was collected and loaded onto a nickel affinity column (Novagen, San Diego, CA, USA) equilibrated with 50 mM nickel sulfate. The recombinant TtGlyDH was step-wise eluted using elution buffers containing imidazole (100, 200, 400, and 1,000 mM), 0.5 M NaCl, and 20 mM Tris-HCl buffer (pH 7.9). The molecular weight and purity of TtGlyDH were analyzed by SDS-PAGE using a prestained protein molecular weight ladder (Thermo, Waltham, MA, USA) as the marker. Protein concentration was measured using the Bradford assay with bovine serum albumin as the standard (Bio-Rad, Hercules, CA, USA) [15]. All protein purification procedures were performed at 4℃.

Activity Assays

One enzymatic unit was defined as the formation or consumption of 1 μmol/min NADPH under the tested assay conditions. Glycerol oxidation activity was assayed using purified enzyme (4.45 μg) at 60℃ in Tris-HCl buffer (50 mM, pH 8.0) containing 2.5 mM NADP+ and 137 mM glycerol. DHA reduction activity was assayed using purified enzyme (0.021 μg) at 60℃ in acetate buffer (50 mM, pH 6.0), supplemented with 0.1 mM NADPH and 2.0 mM DHA. The initial absorbance shift of NADPH at 340 nm (εNADPH = 6.22mM-1 cm-1) in 5 min was monitored. To eliminate the effect of autoxidation or the background rate shift of NADPH, the mixture was prepared without enzyme as a control for each assay. The pH of the assays was adjusted to the desired values at 60℃. All activity assays were performed in triplicates.

Biochemical Characterization

For glycerol oxidation, the optimal pH was determined in Tris-HCl buffers (50 mM) at various pH values (7.0–9.0) at 60℃. The optimum temperature was determined in Tris-HCl buffer (50 mM, pH 8.0) at temperatures ranging from 55℃ to 75℃.

For DHA reduction, the optimal pH was determined at 60℃ in two buffers (50 mM), acetate (pH 4.5–6.5) and imidazole (pH 5.5–7.5). The optimum temperature was determined in acetate buffer (50 mM, pH 6.0) at temperatures of 40-75℃. The desired pH values were adjusted according to temperature.

The effect of temperature on TtGlyDH stability was examined by measuring residual DHA reduction activity. Recombinant TtGlyDH (0.021 μg) was pre-incubated at various temperatures (55–70℃) without the addition of substrates or cofactors. Samples were collected every 30 min and quickly placed on ice for 10 min before conducting the activity assay. The relative activity of un-incubated TtGlyDH was set as 100%.

The effects of various additives on activity were examined by measuring residual DHA reduction activity. Recombinant TtGlyDH (0.021 μg) was incubated with various compounds at a 1.0 mM concentration, unless otherwise specified, including metal divalent cation salts (BaCl2, CaCl2, CoCl2, CuSO4, MgCl2, MnCl2, NiSO4, ZnSO4, and Fe(NH4)2(SO4)2 at a final concentration of 0.05 mM); chelating agents (ethylenediamine tetraacetic acid and 2,6-pyridinedicarboxylic acid); and surfactants (SDS, Triton X-100, and Tween 80 at a final concentration of 1.0% (w/v)). Enzyme activity in the absence of added chemical reagents was defined as 100%.

Determination of DHA Reduction Activation Energy

The activation energy of DHA reduction was measured at temperatures of 25-55℃. The Arrhenius curve was plotted as relative activity versus temperature (K), and logarithmic transformation was conducted to calculate the activation energy.

Determination of the Apparent Kinetic Parameters

For glycerol oxidation, the apparent Michaelis-Menten constants for glycerol were measured in Tris-HCl buffer (50 mM, pH 8.0) with a glycerol concentration gradient of 2.7–82.2 mM at 60℃.

For DHA reduction, the apparent kinetic parameters for DHA were determined in acetate buffer (50 mM, pH 6.0) containing gradient concentrations of DHA (0.017–1.5 mM) with a fixed concentration of NADPH (0.1 mM) at 60℃. The apparent kinetic parameters for NADPH were measured in acetate buffer (50 mM, pH 6.0) with various final concentrations of NADPH (0.006–0.3 mM) and a constant DHA concentration (4.0 mM) at 60℃.

All apparent kinetic parameters of TtGlyDH for glycerol, DHA, and NADPH were calculated by fitting the plots to the Michaelis-Menten equation.

Multiple Sequence Alignment, Phylogenetic Analysis, and Structure Simulation

All related protein sequences were retrieved from UniProtKB/Swiss-Prot (http://www.uniprot.org/) for the sequence alignment and phylogenetic analyses, using the TtGlyDH amino acid sequence as a BLAST query [1]. Iron-containing alcohol dehydrogenases were aligned using ClustalX 2.0 [9]. The alignment was displayed with ESPript (http://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi). The phylogenetic tree of metal-containing alcohol dehydrogenases was constructed using MEGA 6.0. The amino acid sequence of TtGlyDH was submitted to the server SWISS-MODEL (http://swissmodel.expasy.org/workspace/) for homology modeling [2,7,12,23]. The PyMOL Molecular Graphics System was programmed for structural visualization and superposition.

Site-Directed Mutagenesis of TtGlyDH

Mutants R43V (valine substituting arginine at position 43) and triG/A (AAA triplet substituting GGG triplet at positions 92–94) were created following the inverse PCR method using 5’ mutant-specific primers. The purified blunt-end PCR products were phosphorylated in the presence of ATP (1.0 mM) and T4 polynucleotide kinase (1 U) at 37℃. The phosphorylation reaction was deactivated at 70℃ for 5 min and cooled to room temperature. A self-ligation mixture was electroporated into the TOP10 host cells, and positive transformants were screened on LB plates with ampicillin (100 μg/ml). Mutated plasmids were extracted from single colonies and confirmed by DNA sequencing. The mutagenic oligonucleotide primers used are as follows:

R43V forward: 5’-GTGACAAAATCTATAATTGAAGAAAG-3’ (valine codon underlined)

R43V reverse: 5’-ATTACTACTAGCAATAACAAGAAAA-3’

triG/A forward: 5’-GCGGCGGCGAAAATATTTGATACTGTA-3’ (alanine codons underlined)

triG/A reverse: 5’- AATGCCAACTATGACATCAGAGTTTGT -3’

Nucleotide Sequence Accession Number

All DNA information was retrieved from the NCBI GenBank database. The Gene ID of gldA reported in this paper is 9707383 (Tthe_1821), and the genome accession number of T. thermosaccharolyticum DSM 571 is CP002171.1.

 

Results and Discussion

Cloning gldA

Full-length gldA was PCR-amplified from the genomic DNA of T. thermosaccharolyticum DSM 571; the PCR fragment (1,107 bp) was ligated into the commercial vector pET-20b in frame with the C-terminal hexahistidine (His6) tag after double digestion with NdeI and XhoI. The resulting construct was designated as pET-20b-gldA. The translated amino acid sequence of gldA indicated that TtGlyDH belonged to the dehydroquinate synthase-like and iron-containing alcohol dehydrogenase superfamily (DHQ_Fe-ADH superfamily), using the Protein BLAST program. It shared amino acid sequence similarities of 98% and 74% with GlyDHs from Thermoanaerobacterium saccharolyticum (GenBank No. WP_045408662.1) and Thermosediminibacter oceani (GenBank No. WP_041424025.1). SignalP 4.0 (http://www.cbs.dtu.dk/services/SignalP/) [17] failed to recognize a signal peptide within the amino acid sequence, suggesting that TtGlyDH participates in intracellular metabolic processes.

TtGlyDH Expression and Purification

The gldA from T. thermosaccharolyticum DSM 571 was heterologously expressed in E. coli BL21(DE3) after induction with isopropyl-β-D-thiogalactopyranoside (IPTG; 0.5 mM) for 3 h. After sonication and centrifugation, the soluble fraction was heat-treated at 60℃, followed by purification using a nickel affinity column. The flow-through eluted using elution buffer containing 400 mM imidazole was dialyzed against potassium phosphate buffer (10 mM, pH 7.4), and the cell-free preparation was subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis. One band corresponding to a size of approximately 40 kDa (Fig. 1) was observed clearly on the gel, consistent with its estimated molecular mass of 40.4 kDa in the monomer form.

Fig. 1.SDS-PAGE analysis of recombinant TtGlyDH expressed in E. coli BL21 (DE3). Lane M: prestained protein marker. Lane 1: crude extract from cell lysate. Lane 2: supernatant after treatment at 60℃ for 30 min. Lane 3: cell-free solution prepared by nickel affinity chromatography.

Biochemical Properties of Recombinant TtGlyDH

Substrate specificity was assessed using different substrates, including primary alcohols (methanol, ethanol, and propanol), diols (ethylene glycol, 1,2-propanediol, 1,3-propanediol, 2,3-butanediol, and 1,4-butanediol), a triol (glycerol), and DHA. The results showed that TtGlyDH preferentially catalyzed DHA reduction rather than alcohol compound oxidation. Glycerol oxidization activity was faintly detected in the presence of a high concentration of glycerol (137 mM). However, no activity was detected with primary alcohols or diols (data not shown). The highest glycerol oxidation activity was observed at the optimal growth temperature of 60℃ in Tris-HCl buffer (50 mM, pH 8.0). Maximum DHA reduction activity was also observed at 60℃, and TtGlyDH exhibited the highest activity in an acetate buffer, compared with 91% maximum activity in an imidazole buffer at the same pH of 6.0 (Figs. 2A and 2B).

Fig. 2.Biochemical properties of TtGlyDH. (A) The effects of pH on enzymatic activity. (B) The effect of temperature on enzymatic activity. Filled squares, glycerol oxidation activity; open circles, DHA reduction activity. (C) Thermal stability. (D) Effect of chemical reagents on TtGlyDH. The results are displayed as averages and standard deviations of three independent replicates.

The thermostability of TtGlyDH was investigated at four temperatures (55℃, 60℃, 65℃, and 70℃). At 70℃, TtGlyDH activity decreased to 60% maximum activity for the first 30 min and further decreased to 19% after 2 h. However, over 80% of activity was retained during the first 30 min at 65℃, 60℃, and 70℃. In addition, 65% of activity was maintained after incubating at the optimal temperature of 60℃ for 2 h, which was similar to the 70% activity maintained at 55℃ for 2 h (Fig. 2C). These results indicate that TtGlyDH is fairly thermostable at high temperatures from 55℃ to 65℃.

Furthermore, the effect of various additives on TtGlyDH activity was examined (Fig. 2D). Activity was significantly increased by ~79.5% in the presence of Mn2+. In addition, Co2+ and Cu2+ enhanced activity by 31.4% and 37.0%, respectively. Conversely, Ca2+, Zn2+, and two surfactants (SDS, Tween 80) inhibited enzyme activity to varying degrees. In particular, activity was inhibited by 24.0% by 1.0 mM Zn2+. The other chemical reagents had no significant influence on TtGlyDH.

Many dehydrogenases require metal divalent cations to stabilize intermediates and facilitate conversion. Bacterial polyol dehydrogenases, including GlyDHs, are grouped into the Fe-ADH family, which is not well annotated [21]. Biochemical evidence indicates that TtGlyDH differs from other ADHs and GlyDHs, as it is activated by Mn2+ and shows no preference for Fe2+, although sequence analysis suggests it is a member of the Fe-ADH family. To explain this, the atomic orbit arrangement of manganese and iron must be considered. The ionic radii of Mn2+ and Fe2+ are nearly identical, given that manganese and iron are located in the same period on the periodic table and have similar atomic numbers. In addition, Fe2+ ions are readily oxidized in the presence of oxygen and converted to Fe(OH)3, which precipitates at physiological pH and high temperatures. Finally, Fe(OH)3 precipitates have a detrimental impact on TtGlyDH. Because of this, 0.05 mM Fe(NH4)2(SO4)2 was added to investigate its effect on TtGlyDH activity.

Although zinc is widely reported to be an essential component for most ADHs and GlyDHs, the likely explanation for its inhibitory effect is that accommodation of Zn2+ may interfere with the formation or stability of the intermediate, resulting in decreased enzymatic activity.

The apparent activation energy of TtGlyDH DHA reduction was calculated as 312.94 kJ/mol after logarithmic transformation of the Arrhenius plot from 25℃ to 45℃ (Fig. 3).

Fig. 3.Activation energy of TtGlyDH. The results are displayed as averages and standard deviations.

The apparent kinetic constants for the reversible redox reaction of TtGlyDH were determined at the optimum temperature and pH using glycerol and DHA as substrates and NADP+/NADPH as cofactors. Table 1 lists the Michaelis-Menten parameters. In contrast to B. stearothermophilus GlyDH [26], DHA reduction was dominant over glycerol oxidization mediated by TtGlyDH, with a lower Km (DHA) of 1.08 ± 0.13 mM and a more effective turnover number (kcat) of 98.44 s-1.

Table 1.ND, not detected. a,dglycerol as a substrate. b,eDHA as a substrate. cNADPH as a cofactor.

Multiple Sequence Alignment and Phylogenetic Analysis

Multiple sequence alignment of TtGlyDH with representative Fe-ADHs revealed a glycine-rich consensus sequence (GGG motif) located in the N-terminal region. In addition, three polar residues (Asp171, His254, and His271) in TtGlyDH were highly conserved across five species (Fig. 4).

Fig. 4.Sequence alignment of TtGlyDH and Fe-ADH family members. The amino acid sequences are GlyDH from T. thermosaccharolyticum DSM 571 (TtGlyDH), alcohol dehydrogenase II from Zymomonas mobilis (ADH2, Accession No. F8DVL8), 1,3-propanediol dehydrogenase from Klebsiella pneumoniae (DHAT, Accession No. Q59477), lactaldehyde reductase from E. coli (FUCO, Accession No. P0A9S2), alcohol dehydrogenase IV from Saccharomyces cerevisiae (ADH4, Accession No. P10127), and alcohol dehydrogenase from T. maritima (ADH, Accession No. Q9X022). Residues involved in dinucleotide binding and metal divalent cation binding are labeled with triangles and stars, respectively.

To investigate the evolutionary relationship of metal-containing ADHs, a phylogenetic tree was generated using the Poisson substitution model with the neighbor-joining method, and the confidence of the tree was evaluated by bootstrapping with 1,000 replicates. In total, 30 candidate metal-containing ADHs grouped clearly into four clades: the zinc-containing ADH family (classes I, II, and III) and the iron-containing ADH family. TtGlyDH was well classified in the Fe-ADH family and exhibited a closer relationship with GlyDH from T. maritima (UniProtKB entry: Q9WYQ4) at an evolutionary level (Fig. 5). The GGG motif derived from the above alignment also existed in all Fe-ADH family members used for the phylogenetic tree construction (data not shown), but not in any of the Zn-ADH families. This suggests that the GGG motif could be used as a fingerprint for identifying Fe-ADH family members.

Fig. 5.Neighbor-joining tree constructed from 30 metal-containing ADHs. From top to bottom, the four monophyletic clades represent zinc-containing ADH classes I, II, and III and the iron-containing ADH family. The values adjacent to the nodes indicate the percentage bootstrap for 1,000 replicates.

Structure Prediction and Site-Directed Mutagenesis Analysis

We predicted the structure of TtGlyDH and performed site-directed mutagenesis to illustrate the relationship between sequence and structural features of TtGlyDH. The crystal structure of S. plymuthica GlyDH (PDB entry: 4MCA; resolution, 1.9 Å) was used as a template, with 57.2% sequence identity among 366 overlapping residues. The structure of TtGlyDH in its apoenzyme form was predicted using the SWISS-MODEL server. A partial structural superposition of TtGlyDH (residues 70–116) and T. maritima Fe-ADH (PDB entry: 1VHD; chain A; residues 72–136) was represented as a ribbon diagram with a root mean square deviation of 1.40 (Fig. 6). A classical βαβ unit was observed in the TtGlyDH structure, whereas an extra loop (residues 110–127) disrupted the βαβ fold in T. maritima Fe-ADH.

Fig. 6.Local superposition of the T. maritima ADH structure (grey ribbon) and TtGlyDH predicted model (magenta ribbon). The NADPH molecule and residues associated with dinucleotide binding and metal ion binding are depicted in the stick model. Carbon, oxygen, nitrogen, and phosphorus atoms are colored white, red, blue, and orange, respectively. The zinc ion is shown as a cyan sphere.

We introduced two mutations (R43V and triG/A) into TtGlyDH to investigate the residues critical for dinucleotide binding. The mutants R43V and triG/A only possessed 6.3% and 0.7% of the enzymatic activity of the wild type, respectively, indicating that both the basic residue Arg43 and GGG motif are essential for proper TtGlyDH function.

Nicotinamide dinucleotides contribute mainly to electron transfer as electron donors or acceptors in redox reactions. The GXGXXG (where X is any amino acid) motif generally occurs in most ADHs, serving to accommodate nicotinamide dinucleotides [10]. However, the GXGXXG motif is not present in TtGlyDH or other members of the Fe-ADH family. Structural alignment revealed an overall similarity between TtGlyDH and T. maritima Fe-ADH of only 17.5%, although the core structural units (βαβ) were almost the same. Interestingly, the GGG consensus sequence was highly conserved, located between the first β-strand and first α-helix of the initial βαβ unit. Herein, the GGG motif is purported to be involved in dinucleotide binding and accounts for the absence of the GXGXXG motif in the Fe-ADH family. Moreover, because of the missing glycine side chain, the GGG motif forms a more flexible turn and provides enough space to accommodate the pyrophosphate moiety of dinucleotides. The mutation analysis indicated that, although alanine is structurally similar to glycine with an extra methyl group, the mutant triG/A interfered with the formation of the flexible loop, due to steric hindrance caused by the side chain of alanine, and failed to recognize the pyrophosphate moiety, resulting in significant loss of enzymatic activity.

Arg43 is located in close proximity to the nicotinamide moiety, with a calculated distance of 3.8 Å. The hydrogen bond contacts are located between the arginine side chain and nicotinamide moiety, and their interactions facilitate and stabilize dinucleotide binding. Mutant R43V disrupted the hydrogen bond interaction owing to the strongly hydrophobic side chain of valine. This resulted in a weak orientation and recognition of dinucleotide molecules.

Despite TtGlyDH having been modeled in its apoenzyme form, three conserved residues (Asp171, His254, and His271) in TtGlyDH were clustered in the vicinity of Zn2+ in T. maritima Fe-ADH, with all side chains orientated towards the center of Zn2+. These orientations led us to assume that Asp171, His254, and His271 are associated with metal ion binding in TtGlyDH.