Advanced
Characterization and Cofactor Binding Mechanism of a Novel NAD(P)H-Dependent Aldehyde Reductase from Klebsiella pneumoniae DSM2026
Characterization and Cofactor Binding Mechanism of a Novel NAD(P)H-Dependent Aldehyde Reductase from Klebsiella pneumoniae DSM2026
Journal of Microbiology and Biotechnology. 2013. Dec, 23(12): 1699-1707
Copyright © 2013, The Korean Society For Microbiology And Biotechnology
  • Received : July 10, 2013
  • Accepted : August 30, 2013
  • Published : December 28, 2013
Download
PDF
e-PUB
PubReader
PPT
Export by style
Share
Article
Author
Metrics
Cited by
TagCloud
About the Authors
Cheng-Wei Ma
Le Zhang
Jian-Ying Dai
jydai@dlut.edu.cn
Zhi-Long Xiu

Abstract
During the fermentative production of 1,3-propanediol under high substrate concentrations, accumulation of intracellular 3-hydroxypropionaldehyde will cause premature cessation of cell growth and glycerol consumption. Discovery of oxidoreductases that can convert 3hydroxypropionaldehyde to 1,3-propanediol using NADPH as cofactor could serve as a solution to this problem. In this paper, the yqh D gene from Klebsiella pneumoniae DSM2026, which was found encoding an aldehyde reductase ( Kp AR), was cloned and characterized. Kp AR showed broad substrate specificity under physiological direction, whereas no catalytic activity was detected in the oxidation direction, and both NADPH and NADH can be utilized as cofactors. The cofactor binding mechanism was then investigated employing homology modeling and molecular dynamics simulations. Hydrogen-bond analysis showed that the hydrogen-bond interactions between Kp AR and NADPH are much stronger than that for NADH. Free-energy decomposition dedicated that residues Gly37 to Val41 contribute most to the cofactor preference through polar interactions. In conclusion, this work provides a novel aldehyde reductase that has potential applications in the development of novel genetically engineered strains in the 1,3-propanediol industry, and gives a better understanding of the mechanisms involved in cofactor binding.
Keywords
Introduction
1,3-Propanediol (1,3-PD) is widely used in the modern polyester industry, especially in the synthesis of polytrimethylene terephthalate (PTT) [13] . So far, 1,3-PD is manufactured mainly by chemical synthesis started from acrolein or ethylene oxide. However, the yield of 1,3-PD is relatively low (<40% for acrolein, about 80% for ethylene oxide), and a toxic intermediate is released during the production process. As a result, the production of 1,3-PD was shifted from chemical to biological synthesis [20] . Fermentation of glycerol into 1,3-PD has been widely studied with the genera Klebsiella [23] and Clostridium [5] . Klebsiella pneumoniae is mostly used because of its high productivity of 1,3-PD.
The anaerobic metabolism of glycerol is a process of dismutation [1 , 4] . In the oxidative pathway, glycerol is first converted to dihydroxyacetone (DHA) by an NAD + -dependent glycerol dehydrogenase, and DHA is phosphorylated by dihydroxyacetone kinase. The generated dihydroxyacetone phosphate is further oxidized in the subsequent metabolism. In the reductive pathway, glycerol is first converted to 3-hydroxypropionaldehyde (3-HPA) by glycerol dehydrase, and then 3-HPA is reduced to 1,3-PD by an NADH-dependent 1,3-propanediol oxidoreductase (1,3-PDOR).
It is observed that 3-HPA is intracellularly accumulated in batch cultures at a high initial glycerol concentration [2 , 22] . 3-HPA has been proved to be a highly toxic compound, which could cause a cessation of cell growth and glycerol consumption [2] . A straightforward approach to overcome this problem is to construct genetically engineered strains that overexpress 1,3-PDOR. However, no evident increase of production yield is achieved as expected, although the concentration of 3-HPA is successfully kept at a low level [25] . On the other hand, metabolic flux analysis of glycerol utilization suggests that the production yield of 1,3-PD could be increased by utilizing not only NADH from the glycolytic pathway (EMP pathway) but also NADPH from the pentose phosphate pathway (PP pathway) [26] . These results indicate that the production yield of 1,3-PD is limited by the amount of NADH rather than the activity of 1,3-PDOR. Therefore, it is necessary to obtain novel oxidoreductases that are able to convert 3-HPA to 1,3-PD using both NADH and NADPH as cofactors.
In proteomics study, a hypothetical oxidoreductase (HOR) is detected as the concentration of 3-HPA increases in K. pneumoniae [22] . HOR can catalyze the conversion of 3-HPA to 1,3-PD in order to protect cells against the toxic condition. In the meantime, an NADPH-dependent aldehyde reductase (YqhD) from E. coli has recently been reported to serve as an antidotal enzyme to physiologically protect cells against the toxic effect of aldehydes derived from lipid oxidation [17] . Inspired by these observations, we expected to discover a novel YqhD-type enzyme in K. pneumoniae that is able to convert 3-HPA to 1,3-PD using NADPH as cofactor. As shown in this work, a novel aldehyde reductase was successfully cloned and characterized, and its cofactor binding mechanism was investigated through hydrogen-bond analysis and free-energy decomposition.
Materials and Methods
- Materials
The bacterial genomic DNA extraction kit, ExTaq DNA polymerase, 250 bp DNA Ladder, plasmid pMD 19-T, and IPTG were purchased from TaKaRa (TaKaRa Biotechnology (Dalian) Co., Ltd., China(University of New Mexico, USA). Restriction enzymes and T4 DNA ligase were purchased from New England Biolabs (Beijing, Ltd, China). The primers used in this work were custom-synthesized at TaKaRa ( Table 1 ).
Primers used in this work.
PPT Slide
Lager Image
Primers used in this work.
1,3-Propanediol was purchased from Fluka (Germany). 3-HPA was synthesized according to the published method [8] . NADPH, NADH, and NADP + were products of Roche (F. Hoffmann-La Roche, Ltd, Switzerland). All other reagents used were commercial products of the highest grade available.
- Cloning, Expression, and Purification
The genomic DNA of K. pneumoniae DSM2026 was extracted according to the method provided by the kit. The gene yqh D was amplified using primers Kp AR- F and Kp AR- R . Then the PCR product was linked to plasmid pMD19-T. The recombinant pMD18-T- yqh D was transferred into E. coli DH5α using the methods of CaCl 2 and hot attack. The plasmid from E. coli DH5α was isolated with a TIANprep Mini plasmid isolation kit (Tiangen Biotechnology (Beijing) Co., Ltd., China). The gene sequencing was carried out at TaKaRa. The sequencing result was submitted to the GenBank database and the accession number is EU740388.
Because of the existence of an NdeI site in the original sequence, which would influence the subsequent gene expression, synonymous substitution was carried out employing the overlap extension PCR technique. First, primers Kp AR- NF and Kp AR- ssR were used to amplify the 5’ end of gene yqh D; Kp AR- ssF and Kp AR- BR for the 3’ end. Second, the full-size fragment was amplified using the two overlapping fragments as template and Kp AR- NF , and Kp AR- BR as primers. Then, the amplified DNA fragment was digested and ligated to the Nde I- Bam HI-digested pET23a(+). The produced recombinant plasmid pET23a(+)- yqh D was transferred into E. coli BL21(DE3).
E. coli BL21(DE3) containing pET23a(+)- yqh D was harvested by centrifugation. The suspension was disrupted in an ice bath by sonication, and the supernatant was loaded onto a Q Sepharose FF ion-exchange column. The column was first equilibrated and then washed with buffer A (50 mM Tris-HCl buffer (pH 7.4) containing 0.1 mM MnCl 2 and 2 mM DTT(buffer A supplemented with 1 M KCl) using a linear gradient of 0 to 1 M. Active fractions were applied on a column of Sephacryl S300, which was eluted with buffer A.
- Enzyme Activity and Characterization
Enzyme activity in the physiological direction was determined at 340 nm by the initial rate of substrate-dependent NAD(P)H decrease. The assay mixture contained 27 mM 3-HPA, 0.37 mM NAD(P)H, and 1 µM ZnCl 2 in 100 mM Tris-HCl buffer (pH7.4). For the reverse reaction, the assay mixture contained 100 mM 1,3-PD, 2 mM NAD(P) + , 1 µM ZnCl 2 in 100 mM Tris-HCl buffer (pH7.4), 30 mM ammonium sulfate, and 100 mM potassium carbonate buffer (pH9.0). One unit of enzyme activity is the amount catalyzing the formation of 1 µmol product per minute. Protein concentration was determined according to the Coomassie Brilliant Blue method.
To determine the kinetic parameters for the cofactors, the initial rate of NAD(P)H consumption was measured in reaction solutions containing 87 nM Kp AR, 27 mM 3-HPA, 1 µM ZnCl 2 , and varied concentrations of NAD(P)H.
- Homology Modeling
Sequence similarity search was performed with the BLAST program [21] within the protein data bank (PDB). Coordinates of the X-ray crystal structure that shared the highest identity of amino acid sequence with Kp AR were used as the template. Sequence alignment was performed and the 3D models were generated using Modeller 9v2 [19] . To increase the accuracy of the modeled structures, loop modeling was performed with the enumeration algorithm [9] . The homology-modeled structures were evaluated by Procheck [12] .
- Molecular Dynamics (MD) Simulations
Complexes were neutralized by adding sodium counter-ions randomly and solvated in a rectangular box of TIP3P water molecules with a solute-wall distance of 10 Å. The solvated systems were energy-minimized prior to the MD simulations. Each system was minimized by six consecutive rounds of 1,500 steps. Harmonic constraints were applied to all non hydrogen atoms, with the strength of 500, 400, 300, 200, 100, and 0 kcal mol -2 at each round. After that, the systems were slowly heated from 0 to 300 K in 120 ps.
The MD simulations were performed with Amber 9.0 using the force field developed by Cornell et al . [6] with a periodic boundary condition in the NPT ensemble at 300 K. Langevin dynamics was used with the collision frequency of 1.0 ps -1 and constant pressure of 1 atm. The Shake algorithm was applied to fix all covalent bonds. The Particle Mesh Ewald method was used to treat long-range electrostatic interactions. A residue-based cutoff of 10 Å was applied to the noncovalent interactions. No constraint was applied during MD simulations that lasted 1,270 ps for each complex, to get an equilibrium state.
- Free-Energy Decomposition
The MM-GBSA approach [11] , which is implemented in the Amber program, was applied to compute the binding free energy (G binding ):
PPT Slide
Lager Image
PPT Slide
Lager Image
PPT Slide
Lager Image
where G polar is the polar interaction energy; G nonpolar is the nonpolar interaction energy; G ele is the electrostatic interaction energy; G GB is the electrostatic solvation free energy; G vdW is the van der Waals interaction energy; and G SA is the nonpolar solvation free energy. The binding free energy was calculated for each residue.
Results
- Substrate Specificity of KpAR
The study of substrate specificity showed that Kp AR was capable of catalyzing a number of substrates to corresponding products in the physiological direction, whereas no activity was detected when using 1,3-PD as substrate. Among the studied substrates, the highest activity was obtained with butyraldehyde, followed by 3-HPA, acrolein, propionaldehyde, acetaldehyde, and acetoin ( Table 2 ). In contrast, 1,3-PDOR showed catalytic activity both in the oxidation and reduction reactions. 1,3-PDOR is most active with 3-HPA and considerably less active with propionaldehyde, acetaldehyde, acrolein, and butyraldehyde [7 , 15] . Based on the study of YqhD [17] , it was deduced that Kp AR could protect cells against the toxic effects of aldehydes and maintain the normal growth of cells by catalyzing aldehydes to the corresponding alcohols.
Substrate specificity determination ofKpAR.
PPT Slide
Lager Image
aActivities are given relative to that obtained with 3-HPA as substrate; bND: not detected.
- Cofactor Specificity of KpAR
The steady-state kinetics constants of Kp AR with NAD(P)H were determined using 3-HPA as substrate. The K m for NADPH was 0.07 ± 0.01 mM, V max was 6.13 ± 0.12 mM/min, k cat was 7.85 ± 0.22 s -1 , and k cat /K m was 114 ± 18 s -1 ·mM -1 , whereas that for NADH was 1.42 ± 0.01 mM, 1.20 ± 0.08 mM/min, 1.54 ± 0.11 s -1 , and 1.1 ± 0.1 s -1 ·mM -1 , respectively. The K m value for NADPH was 20-fold less than that for NADH, indicating its preference for NADPH. The turnover number (k cat ) was 5-fold higher for NADPH than that for NADH. Thus, the catalytic efficiency (k cat /K m ) was over 100-fold larger for NADPH than that for NADH. In contrast to Kp AR, 1,3-PDOR utilizes specially NADH as cofactor, whereas no activity was determined when using NADPH as cofactor [7 , 10] . To get a deeper insight into the mechanisms involved in the cofactor recognition, a 3D structure of Kp AR was constructed followed by hydrogen-bond and free-energy decomposition.
- Homology Modeling of KpAR
According to the results of template search, Kp AR showed the highest sequence identity with YqhD. Thus, the X-ray crystal structure of YqhD (PDB ID: 1OJ7) was used as the template to generate 3D structures of Kp AR. The Ramachandran plot illustrated that the best modeled structure has 315 (94.3%) residues in the most allowed region, 19 (5.7%) in the additional allowed region, 0 in the generously allowed region, and 0 in the disallowed region. The overall structure of Kp AR ( Fig. 1 A) showed high similarity to that of 1,3-PDOR ( Fig. 1 B). The molecule folds into two structural domains. The N-terminal domain (residues 1 to 176) is formed by an α/β region containing the dinucleotide-binding fold. The C-terminal part (residues 177 to 387), in contrast, is an all-helical domain. The cofactor resides in the deep cleft formed by the two structural domains. Like 1,3-PDOR, there is a conserved βαβαβ dinucleotide-binding Rossmann-like fold at the N-terminal domain of Kp AR ( Figs. 1 C and 1 D), which means that Kp AR belongs to the same family as 1,3-PDOR; namely, the NAD(P) + -dependent alcohol dehydrogenase group III of “iron-activated” dehydrogenase family [16 , 26] .
PPT Slide
Lager Image
3D-modeling and wiring structures of KpAR and 1,3-PDOR. Three-dimensional structure of KpAR with NADPH (A) and 1,3-PDOR with NADH (B). The helix is colored in cyan; sheet in magenta; loop in pink; NADPH or NADH in green. Wiring diagram of the secondary structure of KpAR (C) and 1,3-PDOR (D). Helixes are labeled in H1, H2, etc.; strands by their sheet A, B; turns by β, γ. Red dots illustrate residues interacting with NADPH or NADH.
- Comparison of the Binding Pocket with 1,3-PDOR
It has been known that there is a conserved, negatively charged residue (Asp or Glu) at the end of the second strand of the dinucleotide-binding fold in NAD(H)-dependent oxidoreductases [3] . The space that would be taken up by the extra phosphate of NADP(H) is occupied by the conserved hydrogen bond between the Asp or Glu in the binding site and the adenosine ribose of the NAD(H). As shown in Fig. 2 , the conserved Asp41 exists in the dinucleotide-binding fold of 1,3-PDOR, whereas the corresponding residue is Gly38 in the case of Kp AR. Thus, the cofactor binding space of Kp AR is large enough to accommodate either NADPH or NADH ( Figs. 2 A and 2 B). Owing to the large size and strong negative charge of the side chain of Asp41, there are strong electrostatic repulsion and stereospecific blockage between Asp41 and the phosphate group esterified to the 2’-hydroxyl group of the ribose at the adenine end of NADPH ( Fig. 2 C). Therefore, the binding pocket of 1,3-PDOR is not big enough to accommodate NADPH, although there is no such problem with NADH ( Fig. 2 D). In contrast, the repulsion and blockage are weakened when Asp41 is replaced by Gly38 in Kp AR ( Fig. 2 A).
PPT Slide
Lager Image
Analysis of the cofactor binding pockets. Cofactor binding pocket of KpAR with NADPH (A) or NADH (B). 1,3-PDOR with NADPH (C) or NADH (D) was produced by transferring cofactor from (A) and (B) through structural superimposition. Cofactors are shown as sticks; amino acids of the binding pocket shown as surface; others shown as lines. The red circle in (C) illustrates the space that would be taken up by the extra phosphate group esterified to the 2’-hydroxyl group of the ribose at the adenine end of NADPH.
- Hydrogen-Bond Analysis of KpAR with Cofactor
Since the hydrogen bond plays a vital role in protein-ligand interactions, hydrogen-bond interactions between NAD(P)H and the cofactor binding pocket of Kp AR were analyzed based on the MD simulations ( Fig. 3 ). It was evident that there are more hydrogen-bond interactions between Kp AR and NADPH than that for NADH. It is noteworthy that Gly38, Gly39, and Ser40 form strong hydrogen bonds with the additional phosphate group esterified to the 2’-hydroxyl group of the ribose at the adenine end of NADPH ( Fig. 3 A, upper panel), whereas they contribute no hydrogen-bond interaction in the case of NADH ( Fig. 3 B, upper panel). Ser96 and Asp99 make contributions to the hydrogen-bond interactions in both cases. It is also noteworthy that residues that form hydrogen bonds with cofactors are chiefly glycine (Gly38, Gly39, and Gly149 with NADPH), serine (Ser40, Ser96 with NADPH and NADH), and threonines (Thr44 with NADPH; Thr138 with NADH). Additionally, there is a hydrogen bond between the δNG of Asn147 and the oxygen at the nicotinamide end of NADPH. Furthermore, the hydrogen-bond interactions between Kp AR and NADPH are much stronger than that for NADH, as shown in the lower panels of Fig. 3 , which describe the existence of hydrogen-bond interactions along the MD simulations. It can be seen that all of the hydrogen-bond interactions exist along the entire simulation in the case of NADPH ( Fig. 3 A, lower panel). As for NADH, only the hydrogen bonds formed between δO of Asp99 and NADH (#1and3), one of the hydrogen bonds between γO of Ser96 and NADH (#2), and one of the hydrogen bonds between the main-chain nitrogen atom of Ser96 and NADH (#4) existed along almost the entire simulation ( Fig. 3 B, lower panel). However, another hydrogen bond formed between γO of Ser96 and NADH (#6) disappeared at around 600 ps when the main-chain nitrogen atom of Ser96 forms the other hydrogen-bond with NADH (#5); the hydrogen bond between Ser40 and NADH (#8) appeared at around 750 ps when the hydrogen bond between Thr138 and NADH (#7) breaks.
PPT Slide
Lager Image
Hydrogen-bond interactions between the cofactor binding pocket of KpAR and NADPH (A) or NADH (B). Amino acid residues involved are shown as a line model; cofactors as a ball-and-stick model. Hydrogen bonds are represented as dashed green lines. Lower panels describe the existence of hydrogen-bond interactions along the MD simulations in each case.
- Free-Energy Decomposition of KpAR
The binding free energy, which was decomposed at residue level, was employed to find key residues that are responsible for the cofactor preference. Fig. 4 A shows the differences between the binding free energy of Kp AR with NADPH and that with NADH for each residue. It can be seen that about 20 residues contribute to the cofactor preference. Among these residues, Gly95, Glu145, Lys160, Asp194, and Glu201 prefer the binding of NADH; whereas the others prefer NADPH, where residues Gly37 to Val41 make the greatest contributions. When displaying these residues in the 3D structure of Kp AR ( Fig. 4 B), we can see that they all reside in the cofactor binding pocket. Residues Gly37 to Val41 are around the phosphate group esterified to the 2’-hydroxyl group of the ribose at the adenine end of NADPH. Furthermore, the differences in the binding free energy are mainly caused by the polar interactions rather than the nonpolar interactions ( Figs. 4 C and 4 D), indicating the significance of electrostatic interactions and polar solvation interactions for the cofactor recognition. The exceptions are Val151, Lys160, and His281, whose free-energy differences are chiefly decided by the nonpolar interactions. In addition, Asn68 does not contribute to the cofactor preference owing to its equally opposite effects in nonpolar and polar interactions.
PPT Slide
Lager Image
Free-energy decomposition of KpAR. (A) Differences between the binding free energy of KpAR with NADPH and that with NADH for each residue. (B) 3D structure of the amino acids responsible for the cofactor preference. Residues are shown as lines. Differences between the binding free energy of KpAR with NADPH (C) and that with NADH (D) in the aspect of nonpolar and polar interactions for each residue.
Discussion
Aldehydes are formed during anaerobic fermentation or oxidation of cell components. Because of their high reactive activities, intracellular accumulation of aldehydes will disturb normal cell growth. This has become a common problem in the field of industrial biotechnology. To overcome this problem, unspecific oxidoreductases are employed to catalyze aldehydes to the corresponding alcohols. The broad substrate specificity of Kp AR indicates its potential applications in biotechnology to protect cells against the toxic effects caused by aldehydes. Moreover, the conversion of produced alcohols back to aldehydes is difficult owing to its catalytic irreversibility. In contrast, 1,3-PDOR cannot play such a protective role owing to its substrate specificity, cofactor binding specificity, and its catalytic reversibility. The role of 1,3-PDOR is to reduce the excess NADH generated by the oxidative process to keep the balance of intracellular oxidoreductive potentials. As regard to the biological production of 1,3-PD, its production could be increased if both the NADPH from the PP pathway and NADH from the EMP pathway could be utilized in the conversion of 3HPA to 1,3-PD, according to the metabolic flux analysis [24] . The unspecific cofactor binding feature of Kp AR means that it can serve as a good candidate in the construction of genetically engineered strains in the 1,3-PD industry.
According to Reid and Fewson [18] , the oxidoreductases that catalyze the interconversion of alcohols, aldehydes, and ketones can be divided into three major categories: (ⅰ) NAD(P) + -dependent alcohol dehydrogenases, (ⅱ) NAD(P) + -independent alcohol dehydrogenases, and (ⅲ) flavin adenine dinucleotide-dependent alcohol oxidases. The first category can, in turn, be divided into the following three groups on the basis of the coenzyme binding site: group Ⅰ, medium-chain zinc-dependent dehydrogenases; group Ⅱ, short-chain zinc-independent dehydrogenases; and group Ⅲ, “iron-activated” dehydrogenases. The conserved βαβαβ dinucleotide-binding Rossmann-like fold illustrates that Kp AR belongs to the NAD(P) + -dependent alcohol dehydrogenase group III of “iron-activated” dehydrogenase family. However, it is named NAD(P)H-dependent aldehyde reductase instead of alcohol dehydrogenase or oxidoreductase in this paper, because no enzyme activity could be detected in the oxidation direction. In the investigation of cofactor binding mechanism, Kp AR has no negatively charged residue (Asp or Glu) at the end of the second strand of the dinucleotide-binding fold. Instead, short side-chain residues Gly37-Val41 form a large space to accommodate the additional phosphate group esterified to the 2’-hydroxyl group of the ribose at the adenine end of NADPH with strong hydrogen-bond interactions. This structural feature of Kp AR results in its capability of utilizing either NADPH or NADH as cofactor. In our previous work [14] , relaxation of the coenzyme specificity of 1,3-PDOR was successfully achieved by weakening the effects of the conserved negatively charged residue. The discovery of Kp AR provides another good example to dedicate the mechanisms involved in cofactor binding.
Acknowledgements
This work was supported by the Major State Basic Research Development Program of China (No. 2007CB714306), the Teaching and Research Award Program for Outstanding Young Teachers in High Education Institutions of Ministry of Education of China, and the Specialized Research Fund for the Doctoral Program of Higher Education (No. 200801411113).
View Fulltext  
Ahrens K , Menzel K , Zeng AP , Deckwer WD 1998 Kinetic, dynamic, and pathway studies of glycerol metabolism by Klebsiella pneumoniae in anaerobic continuous culture: III. Enzymes and fluxes of glycerol dissimilation and 1,3-propanediol formation. Biotechnol. Bioeng. 59 544 - 552
Barbirato F , Grivet JP , Soucaille P , Bories A 1996 3-Hydroxypropionaldehyde, an inhibitory metabolite of the glycerol fermentation by Enterobacterial species. Appl. Environ. Microbiol. 62 1448 - 1451
Bellamacina CR 1996 The nicotinamide dinucleotide binding motif: a comparison of nucleotide binding proteins. FASEB J. 10 1257 - 1269
Biebl H , Menzel K , Zeng AP , Deckwer WD 1999 Microbial production of 1,3-propanediol. Appl. Microbiol. Biotechnol. 52 289 - 297    DOI : 10.1007/s002530051523
Chatzifragkou A , Papanikolaou S , Dietz D , Doulgeraki AI , Nychas GJ , Zeng AP 2011 Production of 1,3-propanediol by Clostridium butyricum growing on biodiesel-derived crude glycerol through a non-sterilized fermentation process. Appl. Microbiol. Biotechnol. 91 101 - 112    DOI : 10.1007/s00253-011-3247-x
Cornell WD , Cieplak P , Bayly CI , Gould IR , Merz KM , Ferguson DM 1995 A second generation force field for the simulation of proteins, nucleic acids, and organic molecules. J. Am. Chem. Soc. 117 5179 - 5197    DOI : 10.1021/ja00124a002
Daniel R , Boenigk R , Gottschalk G 1995 Purification of 1,3-propanediol dehydrogenase from Citrobacter freundii and cloning, sequencing, and overexpression of the corresponding gene in Escherichia coli. J. Bacteriol. 177 2151 - 2156
Durrwachter JR , Drueckhammer DG , Nozaki K , Sweers HM , Wong CH 1986 Enzymatic aldol condensation isomerization as a route to unusual sugar derivates. J. Am. Chem. Soc. 108 7812 - 7818    DOI : 10.1021/ja00284a053
Fiser A , Do RKG , Sali A 2000 Modeling of loops in protein structures. Protein Sci. 9 1753 - 1773    DOI : 10.1110/ps.9.9.1753
Johnson EA , Lin EC 1987 Klebsiella pneumoniae 1,3-propanediol: NAD+oxidoreductase. J. Bacteriol. 169 2050 - 2054
Kollman PA , Massova I , Reyes C , Kuhn B , Huo S , Chong L 2000 Calculating structures and free energies of complex molecules: combining molecular mechanics and continuum models. Acc. Chem. Res. 33 889 - 897    DOI : 10.1021/ar000033j
Laskowski RA , MacArthur MW , Moss DS , Thornton JM 1993 PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26 283 - 291    DOI : 10.1107/S0021889892009944
Liu H , Xu Y , Zheng Z , Liu D 2011 1,3-Propanediol and its copolymers: research, development and industrialization. Biotechnol. J. 5 1137 - 1148    DOI : 10.1002/biot.201000140
Ma C , Zhang L , Dai J , Xiu Z 2010 Relaxing the coenzyme specificity of 1,3-propanediol oxidoreductase from Klebsiella pneumoniae by rational design. J. Biotechnol. 146 173 - 178    DOI : 10.1016/j.jbiotec.2010.02.005
Malaoui H , Marczak R 2000 Purification and characterization of the 1,3-propanediol dehydrogenase of Clostridium butyricum E5. Enzyme Microb. Technol. 27 399 - 405    DOI : 10.1016/S0141-0229(00)00219-2
Marçal D , Rêgo AT , Carrondo MA , Enguita FJ 2009 1,3-Propanediol dehydrogenase from Klebsiella pneumoniae: decameric quaternary structure and possible subunit cooperativity. J. Bacteriol. 191 1143 - 1151    DOI : 10.1128/JB.01077-08
Pérez JM , Arenas FA , Pradenas GA , Sandoval JM , Vásquez CC 2008 Escherichia coli YqhD exhibits aldehyde reductase activity and protects from the harmful effect of lipid peroxidation-derived aldehydes. J. Biol. Chem. 283 7346 - 7353    DOI : 10.1074/jbc.M708846200
Reid MF , Fewson CA 1994 Molecular characterization of microbial alcohol dehydrogenases. Crit. Rev. Microbiol. 20 13 - 56    DOI : 10.3109/10408419409113545
Sali A , Blundell TL 1993 Comparative protein modelling by satisfaction of spatial restraints. J. Mol. Biol. 234 779 - 815    DOI : 10.1006/jmbi.1993.1626
Saxena RK , Anand P , Saran S , Isar J 2009 Microbial production of 1,3-propanediol: recent developments and emerging opportunities. Biotechnol. Adv. 27 895 - 913    DOI : 10.1016/j.biotechadv.2009.07.003
Schäffer AA , Aravind L , Madden TL , Shavirin S , Spouge JL , Wolf YI 2001 Improving the accuracy of PSI-BLAST protein database searches with composition-based statistics and other refinements. Nucleic Acids Res. 29 2994 - 3005    DOI : 10.1093/nar/29.14.2994
Wang W , Sun J , Hartlep M , Deckwer WD , Zeng AP 2003 Combined use of proteomic analysis and enzyme activity assays for metabolic pathway analysis of glycerol fermentation by Klebsiella pneumoniae. Biotechnol. Bioeng. 83 525 - 536    DOI : 10.1002/bit.10701
Wang Y , Teng H , Xiu Z 2011 Effect of aeration strategy on the metabolic flux of Klebsiella pneumoniae producing 1,3-propanediol in continuous cultures at different glycerol concentrations. J. Ind. Microbiol. Biotechnol. 38 705 - 715    DOI : 10.1007/s10295-010-0851-1
Zhang Q , Teng H , Sun Y , Xiu Z , Zeng AP 2007 Metabolic flux and robustness analysis of glycerol metabolism in Klebsiella pneumoniae. Bioprocess Biosyst. Eng. 31 127 - 135    DOI : 10.1007/s00449-007-0155-7
Zhao L , Zheng Y , Ma X , Wei D 2009 Effects of over-expression of glycerol dehydrogenase and 1,3-propanediol oxidoreductase on bioconversion of glycerol into 1,3-propandediol by Klebsiella pneumoniae under micro-aerobic conditions. Bioprocess Biosyst. Eng. 32 313 - 320    DOI : 10.1007/s00449-008-0250-4
Zheng Y , Cao Y , Fang B 2004 Cloning and sequence analysis of the dhaT gene of the 1,3-propanediol regulon from Klebsiella pneumoniae. Biotechnol. Lett. 26 251 - 255    DOI : 10.1023/B:BILE.0000013715.04456.0a