AmericanChemicalSociety.com

As described in the previous progress report, we have focused our efforts on the studies of the hydroquinone (HQ) ring-cleaving dioxygenases and Fe(II)-containing model complexes designed to mimic these enzymes. The results described below were used in an NSF-RUI grant that was submitted in July, 2009.

1.      Studies of 2,6-dichlorohydroquinone 1,2-dioxygenase and related enzymes

Several Fe(II)-containing ring-cleaving dioxygenases are similar to extradiol dioxygenases (EDOs), but operate on very different substrates and are poorly characterized. One is 2,6-dichlorohydroquinone 1,2-dioxygenase from Sphingobium chlorophenolicum (PcpA). Unlike the EDOs, where few enzymes can cleave chlorinated substrates successfully, for PcpA, 2,6-dichlorohydroquinone (2,6-diCl-HQ) is the native substrate. 

Determination of the active site of PcpA

In an earlier progress report, we reported using a homology-based approach to construct a structural model of PcpA based upon a homologue of known structure, 1ZSW. We tested the location of the active site that was predicted by the structural model by site-directed mutagenesis. Two Whitman undergraduate students constructed mutants of PcpA in which each of the putative amino acids that could bind the iron (H11, H159, H227, and E276) was mutated to alanine. A fifth site-directed mutant was constructed in which a nearby amino acid that is conserved among all of these enzymes and is known to be critical for function in the EDOs was changed. The H11A, H227A, and E276A mutants exhibited less than 6% of the activity of the wild-type, thus showing that these are the residues that bind the Fe(II). In contrast, the H159A mutant retained about 67% of activity of the wild-type, which shows that H159 does not bind the Fe(II). The Y266F mutant showed 6% of the activity of the wild-type, showing that Y266 is also important for the function of the enzyme. The mutagenesis data confirmed our computer-generated structural model. A manuscript of this work was submitted to Journal of Biological Inorganic Chemistry and was accepted with minor revisions.

We are now attempting to obtain a crystal structure for PcpA. We sent protein to a high-throughput crystallization facility and obtained several crystals. We are now attempting to reproduce those results in-house in order to obtain crystals of sufficient size for X-ray diffraction.

Substrate specificity of PcpA

Another undergraduate student has performed steady-state kinetic experiments on PcpA. The native substrate, 2,6-diCl-HQ, has a K_m of 3.±0.3 mM. 2,6-dibromo-HQ (2,6-diBr-HQ) and 2,6-dimethyl-HQ (2,6-diMe-HQ) are also substrates. The k_cat and K_m for 2,6-diBr-HQ are both somewhat lower than for 2,6-diCl-HQ, while for 2,6-diMe-HQ, k_cat is ~7-fold lower and K_m is ~13-fold higher. In contrast, 2-Cl-, 2-Br-, and 2-Me-HQ exhibit no evidence of ring cleavage, but are instead competitive inhibitors of PcpA. The K_Is of monosubstituted HQs are in the following order: 2-Br-HQ < 2-Cl-HQ < 2-Me-HQ. HQ and 2,5-dichloro-HQ show no evidence of behaving has substrates or inhibitors.

In addition, 2-Br- and 2-Cl-HQ also act as mechanism-based inactivators. When 2,6-diCl-HQ is used as the substrate, the plot of product formation versus time is strongly curved in the presence of either 2-Br- or 2-Cl-HQ. This curvature arises from progressive inactivation of the enzyme over time, and is known to occur in the EDOs for certain catechols that lead to mechanism-based inactivation. We are working to analyze these data for a clearer picture of this mechanism-based inactivation. Catechol, 3-methyl- and 4-methylcatechol, gentisate and pyrogallol are neither substrates nor inhibitors. 2,6-dichloro-, 2,6-dibromo-, and 4-amino-2,6-dichloro-phenol are weak inhibitors but do not show evidence of inactivating the enzyme.

To summarize: PcpA exhibits an absolute requirement for a hydroxyl groups at the 1- and 4-positions and for substituents at both the 2- and 6-positions in order for ring-cleavage to occur. Halogens are greatly preferred at these positions. Monosubstituted HQs are inhibitors and inactivators, and disubstituted phenols are weak inhibitors. These observations have lead us to hypothesize that a weak chloroarene-iron interaction could contribute to substrate binding and specificity. We plan to test this hypothesis in several ways.

Progress on LinE

Another known hydroquinone dioxygenase, LinE from Sphingomonas paucimobilis UT26, has been reported to be active towards 2-chloro-HQ.  In our last progress report we had cloned the LinE gene from Sphingobium indicum CCM 7286 (which is identical to that of S. paucimobilis UT26) and cloned it into a pET vector. This protein has proven to be difficult to express in a soluble form, but we remain hopeful that we will find a successful strategy. Our current approach is to co-express bacterial chaperone proteins.

2.      Fe(II)-orthchlorophenolate model complexes

To probe the possible existence of Fe(II)-haloarene interactions, we collaborated with Prof. Patrick Holland (University of Rochester) to synthesize Fe(II) ortho-chlorophenolate model complexes. A Holland Lab graduate student synthesized several a novel chelating triimine ligands based on the cis,cis-1,3,5-triaminocyclohexane (TACH) backbone, TACH-o-tol, and showed that it yielded the desired 1:1 Fe(II)(TACH-o-tol) complex upon addition of Fe(II).

The ¹H NMR spectra showed the formation of the desired phenolate complexes upon addition of several different substituted phenols as well as 2-Me-HQ to the Fe(II)(TACH-o-tol) complex. Crystal structures were obtained for the [Fe(II)(TACH-o-tol)(2-chlorophenolate)]OTf and [Fe(II)(TACH-o-tol)(2,6-dichloro-phenolate)]OTf complexes. These structures show a distorted trigonal bipyramidal coordination geometry, in which the chloro group and one of the nitrogens are in the axial positions, with Cl-Fe-N bond angles of 168° and 167°/164° for the 2-chlorophenolate and 2,6-di-chlorophenolate complexes, respectively (the latter has two nonequivalent molecules per unit cell). The Fe-Cl distances are 2.942 Å and 2.988/2.980 Å, respectively, for the two complexes. For comparison, the average Fe(II)-chloride bond distances are 2.301 Å and 2.348 Å for four- and six-coordinate complexes, respectively, and the sum of the van der Waals radii for Fe and Cl is 3.75 Å. These Fe-Cl distances are suggestive of formation of an Fe(II)-Cl secondary bond. Very recently, the crystal structure was obtained for the [Fe(II)(TACH-o-tol)(2-methylphenolate)]OTf, which showed the methyl group pointed away from the Fe(II), in contrast to the 2-chloro- and 2,6-dichlorphenolate complexes.  The structures of these complexes will serve as a useful basis for comparison when we investigate the nature of the substrate binding mode in PcpA and LinE. A draft of a manuscript of this work has been completed.