Introduction
Microbes degrade aromatic compounds aerobically using well-established pathways. Most reported aerobic degradation pathways proceed through the dioxygenation of the aromatic ring forming catechol-like intermediates (Mason & Cammack, 1992). Subsequently, the thermodynamically stable aromatic ring is cleaved to generate products that can enter the Krebs cycle directly for complete aerobic degradation. The dihydroxylated catechol ring is either cleaved in an intradiol manner (Hayaishi et al., Reference Hayaishi, Katagiri and Rothberg1955) or an extradiol fashion (Kojima et al., Reference Kojima, Itada and Hayaishi1961). The extradiol dioxygenase class of enzymes has three distinct evolutionary family trees based on diverse substrate preferences (Vilchez-Vargas et al., Reference Vilchez-Vargas, Junca and Pieper2010). Since the first structure of an extradiol dioxygenase 2,3-dihydroxybiphenyl 1,2-dioxygenase (DHBD) from Pseudomonas sp. LB-400 (Han et al., Reference Han, Eltis, Timmis, Muchmore and Bolint1995), numerous studies have elucidated the mechanism for these classes of enzymes (Kovaleva & Lipscomb, Reference Kovaleva and Lipscomb2008). The structures typically contain two repeating βαβββ motifs in a domain forming a funnel-shape to accommodate the 2-His-1-carboxylate active site residues that are bound to Fe(II) or Mn(II) metal ions as a facial triad (Kita et al., Reference Kita, Kita, Fujisawa, Inaka, Ishida, Horiike, Nozaki and Miki1999). Mostly, no cofactor other than the metal ion is present in these reported structures, and the electron transfer occurs primarily to the bound substrate via the metal to the bound oxygen, activating them both for the reaction. Our study enriches the structural aspect of a catechol 2,3-dioxygenase of a Diaphorobacter sp. strain DS2 (Singh & Ramanathan, Reference Singh and Ramanathan2013), which is a non-motile gram-negative bacterium that specifically degrades 3-nitrotoluene and can use it as the sole nitrogen and energy source. We report the crystal structures of recombinant C23O64 catechol dioxygenase enzyme in three forms: ligand-free, 4-methylcatechol-bound (4-MC), and 3-fluorocatechol-bound (3FA).
Objective
Elucidate the structure of C23O64 (catechol dioxygenase) from Diaphorobacter sp. strain DS2 using X-ray crystallography. The structure was then compared with other extradiol enzyme structures to infer a reasonable reaction mechanism for C23O64.
Methods
The supporting information contains all the specific methods used in this manuscript.
Results and discussion
The diamond-shaped protein crystals appeared after 3 to 4 days at 18 °C (Fig. S2). The crystals formed in 0.2 M MgCl2, 0.1 M HEPES sodium salt pH-7.5, and 30% v/v PEG-400 was soaked in substrate solution. Aerobic conditions were used for crystallization and soaking experiments. Therefore, the proteins in the crystal were enzymatically inactive, due to oxidation of Fe (II) to Fe (III) at the active site.
The crystal structure was solved using molecular replacement methods. Table 1 shows the results generated after iterative refinements. An asymmetric unit contains a 314 amino acid monomer, whereas based on results from size exclusion chromatography, the protein is tetrameric in solution (Fig. S4).
Table 1. Structure solution and refinement values in parentheses are for the outer shell.
Crystal contacts form the biologically active tetramer. Like the other reported catechol dioxygenases, the C23O64 protein is also a two-domain protein. It has a repeating βαβββ motif forming an antiparallel β barrel structure in each N- and C-terminal domains. The active site is present in the C-terminal domain within the barrel-shaped structure (Fig 1). The electron density map of the substrate-free crystal did not show a good density for 294–314 amino acid residues, due to disorder in the absence of the substrate.
Fig. 1. (a) A monomer of C23O64 in 3-fluorocatechol bound crystal. The active site is present in the C-terminal domain (b) Diagram for the active site of the 3-fluorocatechol bound protein, the ligand was omitted from the structure, and a simulated annealed omit map (to remove bias) was calculated in the region of the ligand using Phenix. The grey map is the (2Fo-Fc) map at 1*RMS around the active site. In Magenta is the (Fo-Fc) at 1*Sigma. The position of the refined 3FA fits the difference electron density nicely. The electron density for the water molecule bound to Fe was also omitted and is visible in the structure. The figure was made with Pymol (c) Active site of the protein in different crystals showing different coordination geometries around the metal ion and two-dimensional representation of substrate interaction in the active site pocket using LIGPLOT (i) In the 4-methylcatechol bound protein, the Fe is having distorted square pyramidal geometry (ii) In the substrate-free structure, the active site Fe is bound to two water molecules while exhibiting a distorted trigonal bipyramidal geometry (iii) In the 3-fluorocatechol bound protein, the Fe is coordinated to the substrate and one water molecule, displaying a distorted octahedral geometry.
The active-site contains His150, His220, and Glu271 bound to the metal ion in a facial triad manner. In the active-site of the substrate-free C23O64, the penta-coordinated Fe was bound to two water molecules in a trigonal bipyramidal coordination geometry. In the 4-methylcatechol-bound form, the Fe is penta-coordinated with a square pyramidal geometry. In contrast, in 3-fluorocatechol-bound structure, the active-site iron is hexa-coordinate with the substrate and a water molecule in a distorted octahedral geometry (Fig. 1). A comparison of the active-site pocket in these three forms of C23O64, revealed changes in His252, Ile254, and Thr255 side chains with an RMSD of 0.393 Å (Fig. S3).
A comparison of the active sites of C23O64 with LapB from Pseudomonas sp KL28, having approximately 42% sequence identity, revealed three differences. The residues F198, I298, and I254 in C23O64 replace W193, L293, and V250 from LapB, respectively. These substitutions provide a better binding pocket for 4-substituted catechols in C23O64 through hydrophobic interactions with I254 and I298. Another significant comparison with HPCD from Brevibacterium fuscum whose sequence identity is 28% (RMSD 1.98 Å) suggests that residues W192, R292 from HPCD replace F198 and I298 in the active site pocket of C23O64. These amino acid sidechains play a crucial role in accommodating catechol and not homoprotocatechuate in the case of C23O64. The amino acid side chain substitutions explain the inactivity of C23O64 to 2,3-dihydroxybenzoic acid, and 3,4-dihydroxybenzoic acid as the smaller active site pocket cannot accommodate –COO− group. The C23O64 active-site also lacks ionic interactions provided by R292 in HPCD. The sequence and structure comparisons present strong evidence that explains the difference in binding sites for the substrate accommodation. Despite having less than 28% sequence identity, the similarity in the active site pocket residues (Fig. 2) suggests a similar reaction mechanism (Kovaleva & Lipscomb, Reference Kovaleva and Lipscomb2007).
Fig. 2. Comparison of the active site pocket of catechol 2,3-dioxygenase (C23O64) from Diaphorobacter sp. strain DS2 (5ZNH) with (a) LapB from Pseudomonas sp. KL28 (3HPY) and (b) HPCD from Brevibacterium fuscum (1Q0C). The metal binding facial triad residues (His His Glu) has been hidden in the active site pocket for clarity.
Sequence comparison of C23O64 with other reported extradiol dioxygenases (Fig. 3) shows that it is indeed a type I extradiol dioxygenase. All these have 22 strictly conserved residues to play essential structural and functional roles. Where the metal binding H150, H220, E271 and H206 (Kovaleva et al., Reference Kovaleva, Rogers and Lipscomb2015), Y261 (Kovaleva & Lipscomb, Reference Kovaleva and Lipscomb2012) play key functional role, others determine substrate specificity and maintain structural integrity of the enzyme.
Fig. 3 Structure-based alignment of C23O64 from Diaphorobacter sp. strain DS2 5ZSZ and selected type 1 extradiol dioxygenases that are reported. 1MPY from Pseudomonas sp. MT2 (43% sequence identity), 3HPV is LapB from Pseudomonas sp. KL28 (42%), 1F1X is HPCD from Brevibacterium fuscum, 1F1R is HPCD from A. globiformis, and 3LM4 is DHBN from Rhodococcus has less than 28% sequence identity
Acknowledgements
KM thanks CSIR India for Junior and Senior research fellowship. We would like to thank Dr. Deepak Singh from IITK for isolating and characterizing the Diaphorobacter sp strain DS2, and Dr. Vinod Nayak for helping with Phenix Suite.
Conflicts of Interest
None.
Author Contributions
K.M, R.G., S.R. designed the experiment. K.M. and C.A. did the experiments. K.M. analyzed the data and wrote the paper with support from R.G. and S.R.
Funding Information
The work has been supported by Indo-Swedish grant (S.R., BT/IN/SWEDEN/41/SR/2013); DBT (S.R., BT/PR5081/INF/156/2012, BT/PR12422/MED/31/287/214, BT/INF/22/SP22660/2017) and Council of Scientific and Industrial Research (K.M., 09/092(0869)/2013-EMR-I).
Supplementary Materials
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/exp.2020.50.
Open access
Comments
Comments to the Author: This article is acceptable for publication after minor revisions, because it is scientifically sound, has the appropriate methodology, is statistically valid, and enriches our understanding of the structure-function relationships of extradiol dioxygenases. Necessary revisions: The key aspects of crystallization, structure determination and refinement should be moved from the Supplementary Information to the Main Text (Methods). The roles of the conserved residue motifs from Fig. 3 in structure-function relationships should be briefly discussed. In Fig. 1c, all structures should be shown in the same orientation (same projection), so they can be directly compared. There is a problem with residue numbering in Fig. 2, e.g. His151 should probably be His150. Check all residue numbering for consistency throughout the paper. The labels His151, His220 and Glu271 should be removed from Fig. 2, as the clarification in the figure caption is sufficient, UNLESS these residues change their conformations in the different structures, in which case also these residues and their side chains should be displayed. In Fig. S2 each monomer should be colored with a unique color. Displaying local and crystallographic symmetry elements that produce the tetramer in Fig. S2, would enhance our understanding of the assembly of the complete molecule.