Structural determinants of peroxidase activities

Horseradish peroxidase (HRP) is a robust enzyme with commercial applications as an immunodiagnostic reporter enzyme and in the catalysis of difficult chemical transformations. The commercial enzyme is still isolated from the roots of the horseradish plant Armoracia rusticana, and has been studied as a model haem enzyme system since the early 1940's. Following the development of methods to produce the active recombinant enzyme in E.coli (Smith et al., 1990) and completion of the crystallographic structure in 1997 (Gajhede et al., 1997) it has been possible to identify the structural requirements for activity and to extend these activities by protein engineering techniques. Three aspects relating to the enhancement of the ‘normal' and ‘designed' activities of selected variants have been explored at the structure function level in this work. Earlier work (Gajhede et al., 1997; White et al., 1997) highlighted residues that interact with aromatic substrates but that also potentially occlude access to the reactive haem edge by larger bulky substrates of potential commercial interest. The X-ray structure of the HRP-C* A140G/F179A variant was solved to 2.0Å. A larger engineered cavity at the haem edge was observed consistent with the ability of the variant to oxidise luminol directly, a property not seen in the wild-type enzyme. The structure factors for other residues in the haem access channel were not significantly affected. The structure of the S167M HRP variant was also solved, because of interest in its ability to form a novel sulphonium linkage to the haem vinyl group (K. Cali, DPhil thesis, University of Sussex). The sulphur to β vinyl distance was found to be 3.15Å, compared to 1.7Å in the natural sulphonium linkage of myeloperoxidase. This implies that significant thermal motion in the structure is required for the haem-protein crosslink to form, accounting for the relatively slow autocatalytic modification process observed in the presence of hydrogen peroxide. Previous work (Ngo and Smith, Int. Pat. No. WO/2007/020428) has shown that HRP engineered with a more open distal haem pocket (mimicking that of chloroperoxidase or cytochrome P450s) with a weak surrogate base (provided by a Glu residue as in chloroperoxidase) was capable of both entantioselective sulphoxidation and epoxidation. Building on this work, an alternative variant was designed in which the location of the weak base, provided by a Glu or Asp residue, has been varied within the active site. In particular, the HRP variant R38E:F41A:H42A (EAA) catalyses the production of the 1 and 2 naphthol from naphthalene at a rate of 124±4 min-1, suggesting the generation of an epoxide intermediate in the active site. The wild-type enzyme does not catalyse this reaction or the sulphoxidation reactions described for earlier variants. This is believed to be the first report of aromatic C-H bond activation by an engineered plant peroxidase and is unusual in that C-H bond activation of this type normally requires a P450-type thiolate ligated haem system. Equilibrium binding studies show that naphthalene binds to the engineered haem cavity with an estimated Kd of 30±2 μM. Unfortunately, crystals of HRP variants described by Ngo and Smith (2007) and of the new EAA variant described here could not be obtained, despite many crystallisation attempts under a wide range of conditions.