Structure of glycerol dehydrogenase (GldA) from Escherichia coli

The crystal structure of the heat-stable glycerol dehydrogenase from E. coli was solved at 2.79 Å resolution and compared with those of glycerol dehydrogenases from other organisms.

Abstract

Escherichia coli (strain K-12, substrain MG1655) glycerol dehydrogenase (GldA) is required to catalyze the first step in fermentative glycerol metabolism. The protein was expressed and purified to homogeneity using a simple combination of heat-shock and chromatographic methods. The high yield of the protein (∼250 mg per litre of culture) allows large-scale production for potential industrial applications. Purified GldA exhibited a homogeneous tetrameric state (∼161 kDa) in solution and relatively high thermostability (T _m = 65.6°C). Sitting-drop sparse-matrix screens were used for protein crystallization. An optimized condition with ammonium sulfate (2 M) provided crystals suitable for diffraction, and a binary structure containing glycerol in the active site was solved at 2.8 Å resolution. Each GldA monomer consists of nine β-strands, thirteen α-helices, two 3₁₀-helices and several loops organized into two domains, the N- and C-terminal domains; the active site is located in a deep cleft between the two domains. The N-terminal domain contains a classic Rossmann fold for NAD⁺ binding. The O₁ and O₂ atoms of glycerol serve as ligands for the tetrahedrally coordinated Zn²⁺ ion. The orientation of the glycerol within the active site is mainly stabilized by van der Waals and electrostatic interactions with the benzyl ring of Phe245. Computer modeling suggests that the glycerol molecule is sandwiched by the Zn²⁺ and NAD⁺ ions. Based on this, the mechanism for the relaxed substrate specificity of this enzyme is also discussed.

1. Introduction  

Microorganisms are able to grow by anaerobically utilizing glycerol as the sole carbon and energy source (da Silva et al., 2009 ▸). Glycerol fermentation is initiated by glycerol dehydrogenase (GldA; glycerol:NAD⁺ 2-oxidoreductase; EC 1.1.1.6), forming dihydroxyacetone (DHA), with concomitant reduction of the NAD⁺ nucleotide cofactor. The product is then phosphorylated by dihydroxyacetone kinase (EC 2.7.1.29) and funneled into the glycolytic pathway for further degradation (May & Sloan, 1981 ▸). GldA belongs to the metal-containing family III polyol dehydrogenases, which often require NAD⁺ and zinc (Zn²⁺) ions in the active site for catalytic activity (Tang et al., 1979 ▸; Ruzheinikov et al., 2001 ▸). GldAs from different organisms exhibit substrate promiscuity and are capable of oxidizing a broad range of diols, including propan-1,2-diol, butan-2,3-diol, ethan-1,2-diol and 3-mercapto-1,2-dihydroxy­propane (Wang et al., 2014 ▸; Spencer et al., 1989 ▸; Ruzheinikov et al., 2001 ▸; Subedi et al., 2008 ▸). Enzymes with such relaxed substrate specificities help organisms to gain an evolutionary advantage with better environmental adaptation and provide opportunities in a broad range of industrial applications (Gatti-Lafranconi & Hollfelder, 2013 ▸). Detailed structural studies of such enzymes are pivotal to understanding their mechanism, which will provides a basis for further directed evolution for industrial purposes. Following the first GldA structure, that from Bacillus stearothermo­philus, which was reported in 2001 (Ruzheinikov et al., 2001 ▸), a number of other GldA structures from different microbes, including the Gram-negative Thermotoga maritima (Lesley et al., 2002 ▸), Entero­bacteriaceae (Hatti et al., 2017 ▸), Serratia (Musille & Ortlund, 2014 ▸) and Sinorhizobium meliloti (PDB entry 3uhj; New York Structural Genomics Research Consortium, unpublished work), the Gram-positive Clostridium acetobutylicum (PDB entry 3ce9; Joint Center for Structural Genomics, unpublished work) and the fission yeast Schizosaccharomyces pombe (PDB entry 1ta9; A. M. Mulichak, unpublished work), have been published or deposited in the Protein Data Bank (PDB).

Over the past decade, the industrial production of biodiesel has emerged as a new trend as, in contrast to conventional petroleum-derived fuels, it provides an eco-friendly platform to manage the CO₂ balance in the atmosphere (Vasudevan & Briggs, 2008 ▸). Glycerol, which is the major byproduct of the biodiesel industry, can be used for the production of a variety of useful chemicals such as DHA, 1,3-propanediol, ethanol and organic acids by microbial fermentation (Piattoni et al., 2013 ▸). Escherichia coli, a Gram-negative facultative anaerobic bacterium, has a long history of laboratory culture. In addition, owing to genome plasticity, it has also served as the most commonly used tool in modern biological engineering and industrial fermentation (Lee, 1996 ▸; Liu & Khosla, 2010 ▸). E. coli GldA has been overexpressed for the increased fermentative metabolism of crude glycerol. Recombinant purified GldA protein is also clinically used as a coupling enzyme for lipoprotein lipase-based assays in the quantitation of triglycerides (Mamoru et al., 1977 ▸). Its kinetic and biochemical properties have been well characterized (Piattoni et al., 2013 ▸). However, despite the determination of detailed GldA structures from various microbes, the structure of E. coli GldA is not available. We envisaged that determining the crystal structure of E. coli GldA would be of particular interest for protein engineering, directed evolution and industrial optimization. Towards this end, this work presents the three-dimensional structure of E. coli GldA and compares it with known structures of GldAs from other organisms. The mechanism behind the substrate promiscuity is also discussed.

2. Materials and methods  

2.1. Materials  

Luria–Bertani/Lennox (LB) medium was prepared with yeast extract and tryptone obtained from Oxoid, Hampshire, England. Q5 DNA polymerase, restriction enzymes (NcoI and XhoI) and the Gibson assembly cloning kit were obtained from New England Biolabs, Massachusetts, USA. Centrifugal filters were purchased from Sartorius, Göttingen, Germany. All other chemicals and reagents used in this study were of analytical grade and were obtained from local firms in Tianjin, People’s Republic of China.

2.2. Cloning, expression and purification of GldA  

The gldA gene was amplified by E. coli (strain K-12, substrain MG1655) colony PCR using the forward and reverse primers 5′-GTTTAACTTTAAGAAGGAGATATACCATGGACCGCATTATTCAATCAC-3′ and 5′-ATCTCAGTGGTGGTGGTGGTGGTGCTCGAGTTATTCCCACTCTTGCAGGAAA-3′, respectively. The amplified gldA fragment was inserted into the NcoI/XhoI double restriction enzyme-digested pET-28a(+) vector by Gibson assembly. The recombinant plasmid pET28a-gldA was verified by sequencing and was transformed into E. coli BL21 (DE3) strain (Invitrogen). The cells were grown at 37°C with shaking at 240 rev min⁻¹ using LB medium (1 l) supplemented with 50 µg ml⁻¹ kanamycin in a 2.6 l flask. When the OD_600 nm reached ∼0.6, the temperature was decreased to 24°C and protein expression was induced by the addition of isopropyl β-d-1-thiogalactopyranoside (IPTG) to a final concentration of 0.2 mM. The cells were cultured for a further 18 h before harvesting by centrifugation (6000g for 10 min at 4°C).

The cell pellet was resuspended in buffer A consisting of 20 mM Tris–HCl pH 8.0, 0.2 mM phenylmethanesulfonyl fluoride (PMSF), 5% glycerol and was subjected to lysis by passage twice through a French press at 97 MPa (PandaPLUS, Niro Soavi, Italy). Cell debris was removed by centrifugation (14 000g for 20 min at 4°C) and the supernatant was briefly heat-shocked in a water bath at 60°C for 10 min (Asnis & Brodie, 1953 ▸; Piattoni et al., 2013 ▸). The sediment was removed by centrifugation (20 000g for 30 min at 4°C) and the soluble fraction was loaded onto a 5 ml Q Sepharose anion-exchange column (GE Healthcare). The column was washed with 50 ml buffer A and the protein was eluted with a linear gradient of 200 ml 0–1 M NaCl in buffer A. Fractions containing GldA were collected, pooled and dialyzed with buffer B (20 mM Tris–HCl pH 7.4, 50 mM sodium chloride, 150 mM ammonium acetate, 5% glycerol). The dialyzed protein was concentrated to 12 mg ml⁻¹ using a centrifugal filter (Sartorius, 30 000 molecular-weight cutoff) and examined on a 12% SDS–PAGE gel (Laemmli, 1970 ▸). Aliquots of the protein were then flash-frozen with liquid nitrogen and stored at −80°C until further use.

2.3. Determination of the oligomeric state of GldA  

2 ml of 2.5 mg ml⁻¹ GldA solution was injected into a Superdex 200 gel-filtration column (300 ml) and eluted with buffer A. The same conditions were used to analyze a solution of molecular-weight markers (Sigma) including bovine thyro­globulin (669 kDa), horse apoferritin (443 kDa), sweet potato β-amylase (200 kDa), yeast alcohol dehydrogenase (150 kDa), bovine serum albumin (66 kDa) and bovine carbonic anhydrase (29 kDa). The molecular weight of GldA was calculated from the elution volume using a second-degree polynomial for the relationship between log(molecular weight) and retention time.

2.4. Crystallization, data collection and structure determination  

Initial crystal screening was performed by the sitting-drop vapor-diffusion method at 295 K using various crystallization kits from Hampton Research and Molecular Dimensions. The crystal screening was carried out in a 96-well format with a 1:1(v:v) ratio of protein solution and reservoir solution using an automated liquid-handling robotic system (Griffin, Art Robbins). Several crystallization conditions resulted in plate-shaped crystals. The crystallization conditions were further optimized by the hanging-drop vapor-diffusion method in 24-well plates. The best condition, which yielded large plate-shaped crystals, was observed with 2.0 M ammonium sulfate. Crystals were then mounted on a CryoLoop (Hampton Research) and flash-cooled in liquid nitrogen using reservoir solution containing 20% glycerol as a cryoprotectant. Diffraction data from a single crystal were collected to a resolution of 2.8 Å on BL17U1 at Shanghai Synchrotron Radiation Facility (SSRF; Wang et al., 2018 ▸). The data set was indexed, integrated and scaled using the HKL-3000 suite (Minor et al., 2006 ▸). Molecular replacement was performed with PHENIX (Adams et al., 2010 ▸) using Serratia GldA as a search model (PDB entry 4mca; Musille & Ortlund, 2014 ▸). After running Phaser-MR, the model sequence was replaced with the object sequence. The structure was further manually built into the modified experimental electron density using Coot (Emsley et al., 2010 ▸) and refined in PHENIX (Adams et al., 2010 ▸) in iterative cycles. Data-collection and final refinement statistics are presented in Table 1 ▸.

Table 1. Data-collection and refinement statistics for the E. coli GldA crystal.

Values in parentheses are for the highest resolution shell.

Data collection
Wavelength (Å)	0.9
Space group	P43212
a, b, c (Å)	162.4, 162.4, 293.1
α, β, γ (°)	90.0, 90.0, 90.0
Resolution (Å)	39.1–2.7 (2.8–2.7)
R merge†	0.09 (0.7)
Average I/σ(I)	14.8 (1.8)
Completeness (%)	99.5 (98.2)
Multiplicity	6.7 (6.5)
Z	4
Refinement
Resolution (Å)	2.8
No. of reflections	97076 (9400)
R factor/R free (10% data)	0.17/0.20
R.m.s.d., bond lengths (Å)	0.008
R.m.s.d., angles (°)	0.9
No. of atoms
Protein	1468
Ligands	12
Water	137
Ramachandran plot (%)
Most favored	97.7
Additionally allowed	2.1
MolProbity clashscore	5.0

2.5. Determination of the melting temperature of GldA  

The melting temperature of GldA was determined using a fluorescence-based thermal shift assay (Nettleship et al., 2008 ▸). 0.2 mg ml⁻¹ GldA and 1× SYPRO Orange protein gel stain solution (Sigma) in a total volume of 50 µl were used in the assay. Thermal melts were obtained using a QuantStudio 6 Flex real-time PCR machine (Life) with the SYBR green filter option. The temperature was increased from 10 to 95°C with an increment rate of 0.033°C s⁻¹ and the melting temperature was obtained by taking the midpoint of each transition.

2.6. Miscellaneous  

All structure figures were generated using UCSF Chimera (Pettersen et al., 2004 ▸). The coordinates and structure factors of E. coli GldA have been deposited in the PDB (PDB entry 5zxl).

3. Results and discussion  

3.1. Protein purification  

A combination of heat-shock and chromatographic methods resulted in highly pure GldA protein that migrated as a single band at ∼39 kDa on a 12% SDS–PAGE gel (Fig. 1 ▸ a), which is consistent with the predicted molecular weight of 38.7 kDa. The oligomeric state of purified GldA was determined using a sizing column. The elution profile shows that E. coli GldA exists as a homogeneous tetramer (∼161 kDa) in a solution containing 5% glycerol (Fig. 1 ▸ b). Interestingly, in the absence of glycerol GldA has been reported to elute as a mixture of a dimer and an octamer (Tang et al., 1979 ▸; Piattoni et al., 2013 ▸), indicating the important role of glycerol in maintaining the quaternary structure of GldA. Typically, the yield of GldA from 1 l of cell culture was 250 mg.

Figure 1.

Electrophoretic and chromatographic analyses of the purified GldA used for crystallization. (a) 12% SDS–PAGE of purified GldA. Lane A, protein markers (labelled in kDa); lane B, purified E. coli GldA protein. (b) Elution profiles of GldA (solid line) and molecular-weight markers (dotted line) using Superdex 200 gel-filtration chromatography. P1, bovine thyroglobulin (669 kDa); P2, horse apoferritin (443 kDa); P3, sweet potato β-amylase (200 kDa); P4, yeast alcohol dehydrogenase (150 kDa); P5, bovine serum albumin (66 kDa); P6, bovine carbonic anhydrase (29 kDa). The estimated molecular weight of GldA is 161 kDa. (c) E. coli GldA protein crystals produced by the vapor-diffusion method.

3.2. Crystallization and structure determination  

During preliminary crystal screening trials, crystals were observed in several conditions. The best crystallization condition consisted of 2.0 M ammonium sulfate (Fig. 1 ▸ c). A single crystal diffracted to 2.8 Å resolution on beamline BL17U1 at SSRF. The crystal belonged to space group P4₃2₁2, with unit-cell parameters a = 162.4, b = 162.4, c = 293.1 Å, α = 90.0, β = 90.0, γ = 90.0°. There are four monomers in the asymmetric unit, with 56% solvent content. Serratia GldA (PDB entry 4mca; Musille & Ortlund, 2014 ▸), which exhibits 58.9% sequence identity, was used as a search model for molecular replacement. The structure was further refined in PHENIX (Adams et al., 2010 ▸), with a final R factor and R _free of 0.17 and 0.20, respectively.

3.3. Overall structural details  

We solved the crystal structure of E. coli GldA in complex with glycerol at 2.8 Å resolution. The asymmetric unit contained four chains of GldA. Interactions among the four monomers suggest a homotetrameric solution state for the protein (Fig. 2 ▸ a). Each monomer contains 367 amino-acid residues forming nine β-strands, 13 α-helices, two 3₁₀-helices and a number of loops, which fold into two domains (the N-terminal domain and the C-terminal domain) separated by a deep cleft (Fig. 2 ▸ b). Compared with other GldA structures, E. coli GldA lacks one 3₁₀-helix between helices 9 and 10, which might be caused by different crystal packing. The N-terminal domain (NTD; residues 1–161) consists of six parallel β-sheets flanked by the first four α-helices. This arrangement belongs to the Rossmann fold that is commonly observed in dinucleotide-binding proteins (Hanukoglu, 2015 ▸; Rao & Rossmann, 1973 ▸). β7 and β8 of the NTD form a β-hairpin linked by a short loop of five amino acids. This loop is expected to undergo a conformational change to form a channel for the entry of glycerol and to stabilize NAD⁺ binding (Fig. 2 ▸ b). The C-terminal domain (CTD; residues 162–367) is exclusively constituted by a number of helices, which can be further divided into two subdomains: a long antiparallel helix bundle formed by α6–α8 and an antiparallel Greek-key helix bundle (Richardson, 1981 ▸) formed by α9–α13 and two 3₁₀-helices (Fig. 2 ▸ b).

Figure 2.

Structural features of E. coli GldA. (a) Tetrameric state of E. coli GldA. Glycerol (cyan sticks), Zn²⁺ (gray sphere) and active-site residues (thick sticks) are shown for each monomer. (b) Secondary-structural details of the GldA monomer, represented by a ribbon diagram, in which α-helices are colored cyan, β-strands are colored plum and loops are colored white. The Rossmann fold, long antiparallel helix-bundle and Greek-key helix-bundle domains/subdomains are shown in boxes.

The Greek-key helix-bundle subdomain of the CTD and the Rossmann fold of the NTD form two opposing faces of the active site, while the long antiparallel helix-bundle subdomain (α6–α8) forms the floor (Musille & Ortlund, 2014 ▸). A Zn²⁺ ion is coordinated by His254, His271 and Asp171 deep in the cleft between the NTD and CTD. The Zn²⁺ ion directly interacts with the substrate glycerol. The distances between the Zn²⁺ ion and O₁ and O₂ of glycerol are 2.6 and 3.1 Å, respectively. O₁ and O₂ are coordinated by His254 and His271, respectively. The orientation of the glycerol molecule in the active site is stabilized by van der Waals interaction between the glycerol C atoms and the benzyl ring of Phe245 and by electrostatic inter­actions between the negatively charged π-electron cloud of the Phe245 benzyl ring and the partially positively charged C atoms of glycerol (Fig. 3 ▸). NAD⁺ is not present in our structure. Therefore, we modeled NAD⁺ by superposing a GldA structure containing NAD⁺ (PDB entry 1jq5; Ruzheini­kov et al., 2001 ▸) onto our E. coli GldA structure. This reveals that the NAD⁺-coordinating residues are conserved. In addition to the Rossmann fold, the β-hairpin formed by β7 and β8 and a short loop linking the two also contribute to the interaction with NAD⁺ (Fig. 3 ▸). From our structure, it seems that the main restraint of the substrate glycerol is the interaction between the O₁/O₂ atoms and Zn²⁺, which might explain the relaxed substrate specificity of GldA towards a wide range of 1,2-diols (Wang et al., 2014 ▸; Spencer et al., 1989 ▸; Ruzheinikov et al., 2001 ▸; Subedi et al., 2008 ▸).

Figure 3.

The active site of E. coli GldA. Glycerol (cyan sticks) is sandwiched between Zn²⁺ (gray sphere) and NAD⁺ (plum sticks) superimposed from PDB entry 1jq5. An enlarged view of the active site is shown with bound glycerol, Zn²⁺ and NAD⁺. Zinc is coordinated by two conserved histidines (His254 and His271) and an aspartate (Asp171). The O₁ and O₂ atoms of glycerol (stabilized by Phe245) serve as ligands of the Zn²⁺ ion.

3.4. Comparison of E. coli GldA with other GldAs  

There are seven other GldA structures that have been deposited in the PDB. Phylogenetic analyses of the sequences of all structurally resolved GldAs were performed (Fig. 4 ▸). MatchMaker from the UCSF Chimera software package (Pettersen et al., 2004 ▸) was used to compare the structure of E. coli GldA with each of the seven previously solved GldA structures (Table 2 ▸). The degrees of structural similarity are in general agreement with the phylogenetic analyses. PDB entries 5xn8 and 4mca, which are both structures of GldAs from the Proteobacteria phylum and the Enterobacteriaceae family, showed the highest structural similarity to E. coli GldA (Table 2 ▸). PDB entries 1kq3 and 1jqa are the structures of GldAs from different phyla, the Thermotogae and Firmicutes, respectively, and thus when compared with E. coli GldA their r.m.s.d.s are relatively higher (Table 2 ▸). S. meliloti is an alphaproteobacterium, which also belongs to the Proteobacteria phylum, as does E. coli. Regardless of the slightly lower sequence identity and higher overall r.m.s.d., when the N- and C-terminal domains were compared separately even higher structural similarities were observed (Table 2 ▸). The GldAs from the eukaryotic yeast S. pombe and the Gram-positive C. acetobutylicum (belonging to the Firmicutes phylum) are the least structurally similar to E. coli GldA among the seven, and also have the lowest sequence identities (Table 2 ▸).

Figure 4.

Phylogenetic analyses of E. coli GldA (PDB entry 5zxl). MUltiple Sequence Comparison by Log-Expectation (MUSCLE) was used for sequence alignment, PhyML was used to construct the tree and Interactive Tree Of Life (iTOL) was used to plot the tree. Each entry is named starting with the PDB code and its source.

Table 2. Comparison of E. coli GldA with other GldA structures in the PDB.

Organism	Enterobacteriaceae sp.	Serratia sp.	T. maritima	B. stearothermophilus	S. meliloti	S. pombe	C. acetobutylicum
PDB code	5xn8	4mca	1kq3	1jqa	3uhj	1ta9	3ce9
Identity (%)	58.9	58.9	53.2	49.7	47.1	42.6	20.4
R.m.s.d. (Å)
Overall	0.67	1.41	1.56	1.59	2.61	1.58	4.21
NTD	0.80	1.90	1.89	1.73	1.40	1.70	2.96
CTD	0.49	0.71	1.00	1.07	1.01	1.21	4.90

An interesting observation is that the CTDs are generally more conserved than the NTDs in all reported GldA structures except C. acetobutylicum GldA (Table 2 ▸). The β7–β8 hairpin of E. coli GldA exhibits a similar conformation to those in PDB entries 5xn8 and 1kq3. By contrast, in the other structures (PDB entries 4mca, 1jqa, 3uhj, 1ta9 and 3ce9) this hairpin loop is shifted towards the active-site cleft (Fig. 5 ▸ a). Consistent with the lowest sequence identity (only 20.4%), obvious structural differences are found between E. coli GldA (PDB entry 5zxl) and C. acetobutylicum GldA (PDB entry 3ce9). A large conformational difference between their NTDs, including the orientation of the N-terminal tail (β1) and the β7–β8 hairpin loop, is observed when the CTDs are used for superposition. On the other hand, a large structural change in the CTD such as the orientation of the C-terminal tail helix and the α9–α10 loop is noticed when the NTDs are used for superposition (Fig. 5 ▸ b). Comparison of the glycerol-binding sites in all reported co-crystal structures revealed two different binding modes (Fig. 5 ▸ c). PDB entries 1jqa, 1ta9, 4mca and E. coli GldA (PDB entry 5zxl) exhibit a glycerol-binding site in proximity to the Zn²⁺ ion, whereas the glycerol molecules in PDB entries 5xn8 and 3uhj are distant from the Zn²⁺ ion and interestingly are in distinct orientations (Figs. 5 ▸ c and 5 ▸ d). Thus, we hypothesize that the observed unique secondary binding site of glycerol in PDB entries 5xn8 and 3uhj may indicate the entry path of glycerol into a channel. Furthermore, a comparison of the orientations of glycerol(s) within the active site(s) of PDB entries 1jqa, 1ta9, 4mca and 5zxl reveals that the position of O₁ is more conserved compared with those of O₂ and O₃. It appears that during substrate binding O₁ is a dominant atom that binds first to Zn²⁺, which then is followed by O₂ binding. Since O₃ is not directly involved either in substrate coordination or in catalysis, O₃ positioning is the least conserved (Fig. 5 ▸ d). For example, the O₃ in PDB entry 1jqa is shifted 1.8 Å compared with that in the E. coli GldA structure (PDB entry 5zxl).

Figure 5.

Structural comparison of E. coli GldA with other GldAs. (a) Superposition of the E. coli GldA crystal structure (orchid) with other structures reported in the PDB (dark khaki, PDB entry 5xn8; dark gray, PDB entry 4mca; hot pink, PDB entry 1kq3; cornflower blue, PDB entry 1jqa; tan, PDB entry 3uhj; cyan, PDB entry 1ta9; white, PDB entry 3ce9). (b) Superposition of the E. coli GldA crystal structure (orchid) with PDB entry 3ce9 (white). (c) Comparision of the glycerol-binding sites of different GldAs. (d) Enlarged view showing the orientiation of glycerol within the active-site pocket.

3.5. Thermal stability  

We measured the thermal stability of E. coli GldA using a fluorescence-based thermal shift (Thermofluor) assay. This showed a relatively high melting temperature of 65.6°C (Fig. 6 ▸ a). This is surprising since E. coli is not a thermophilic microorganism like T. maritima and B. stearothermophilus (Lesley et al., 2002 ▸; Ruzheinikov et al., 2001 ▸). We speculate that the observed high thermal stability is contributed by the high internal hydrophobicity observed in the interfaces between the secondary-structure elements within the domain (Fig. 6 ▸ b). Such a high thermal stability of E. coli GldA allows protein production in large quantities via the simple and cost-effective heat-shock method, and therefore makes it an ideal candidate for industrial and clinical applications.

Figure 6.

Thermal stability of E. coli GldA. (a) Thermal melt curves (four repeats). (b) Surface view of the monomeric unit. Negatively charged, positively charged and noncharged surfaces are colored red, blue and white, respectively.

3.6. Proposed mechanism  

In the ternary complex of E. coli GldA containing Zn²⁺ and modeled NAD⁺ (Fig. 3 ▸), the Zn²⁺ ion was found to be buried deep in the cleft between the two domains and it appears that NAD⁺ blocks the interaction of Zn²⁺ ion with solvent molecules, as described previously for B. stearo­thermophilus GldA (Ruzheinikov et al., 2001 ▸). As glycerol binds to the active-site pocket with coordinated interactions between two adjacent hydroxyl groups and the neighboring Zn²⁺ ion, the pK _a of the C₂ hydroxyl O atom of glycerol may be lowered considerably, allowing the hydroxyl proton to be removed to form an alkoxide ion. The subsequent conversion of the Zn²⁺ ion-stabilized alkoxide ion intermediate to the ketone might proceed with the transfer of a hydride from C₂ to the C₄ of the nicotinamide ring located at a distance of 3.0 Å (Fig. 3 ▸) in accordance with a previous report reporting hydride transfer (Wilkie & Williams, 1995 ▸). As a consequence, the H⁺ ion produced during the reaction is released as a proton into the surrounding solution, followed by the release of the products dihydroxyacetone and NADH (Hammes-Schiffer & Benkovic, 2006 ▸).

In summary, we report the structure of GldA from E. coli, the most commonly used organism in the laboratory and in the fermenting industry. Furthermore, we made a systematic comparison of all known GldA structures. Regardless of the high similarities between the structure of E. coli GldA and those from other organisms, detailed structural information would be appreciated for protein engineering and directed evolution, considering the popularity of this organism.

Supplementary Material

PDB reference: glycerol dehydrogenase from Escherichia coli, 5zxl

Funding Statement

This work was funded by National Natural Science Foundation of China grant 31570060 to Yan Zhang. State Administration of Foreign Experts Affairs grant . National Key Research and Development Program of China grant 2017YFD0201400, 2017YFD0201403 to Zhiguang Yuchi.