Crystal structures of pristine and oxidatively processed lignin peroxidase expressed in Escherichia coli and of the W171F variant that eliminates the redox active tryptophan 171. Implications for the reaction mechanism.

The heme enzyme lignin peroxidase (LiP) from the white rot fungus Phanerochaete chrysosporium contains a solvent exposed redox active tryptophan residue (Trp171) that carries a unique hydroxy group stereo-specifically attached to its C(beta) atom. A Trp171Phe mutant has no activity at all towards the substrate veratryl alcohol. The mechanism of veratryl alcohol oxidation involving beta-hydroxy-Trp171 is largely unknown. Here, we present the first crystal structures of LiP isozyme H8 at high resolution in its pristine non-hydroxylated form, of the C(beta)-hydroxylated form, and of the Trp171Phe mutant using recombinantly expressed and refolded protein produced from Escherichia coli. As a consequence, all structures are unglycosylated. Structural changes in response to the mutation are marginal and allow us to attribute the complete lack of activity exclusively to the absence of the redox active indole side-chain. The origin of the stereospecificity of the Trp171 hydroxylation can be explained on structural grounds. A reaction mechanism involving Trp171 is proposed and the possible function of the modification is discussed. Another important result regarding the ongoing debate on the co-ordination state of the heme iron in the resting state is that the iron is six co-ordinate in all cases the data being collected at room temperature. The mean distance from the iron to the distal water ligand is 2.18(+/-0.08) A. The radical scavenger orcinol was found to decrease radiation damage to the crystals, during data collection at room temperature.

Two substrate interaction sites in lignin peroxidase revealed by site-directed mutagenesis.

It has been shown recently that Trp171 of lignin peroxidase (LiP) is hydroxylated at the Cbeta position [Blodig, W., Doyle, W. A., Smith, A. T., Winterhalter, K., Choinowski, T., and Piontek, K. (1998) Biochemistry 37, 8832-8838]. Comparative experiments, carried out on both wild-type fungal and recombinant LiP isoenzyme H8 (LiPH8), indicate that the process of hydroxylation is autocatalytic and that Trp171 may be implicated in catalysis. The role of this residue has therefore been examined using site-directed mutagenesis to obtain recombinant enzymes with Trp171 substituted by Phe or Ser (W171F and W171S LiPH8, respectively). The wild-type recombinant enzyme (LiPH8) was analyzed in solution using 1H NMR spectroscopy and its integrity confirmed prior to the kinetic and spectroscopic characterization of LiPH8 mutants. A charge neutralization mutation in the "classical heme edge" substrate access channel of LiP, in which Glu146 was substituted by Gly (E146G LiPH8), showed substantial activity with respect to veratryl alcohol (VA) oxidation and a marked (2.4 pH units) increase in pKa for the oxidation of a negatively charged difluoroazo dye. More surprisingly, the Trp171 LiPH8 mutants W171F and W171S LiPH8 were found to have lost all activity with VA as substrate, and compounds I and II were unable to react with VA. Both mutants, however, retained substantial activity with two dye substrates. These data provide the first direct evidence for the existence of two distinct substrate interaction sites in LiP, a heme-edge site typical of those encountered in other peroxidases and a second, novel site centered around Trp171 which is required for the oxidation of VA. Stopped-flow kinetic studies showed that all the mutants examined reacted normally with hydrogen peroxide to give a porphyrin cation radical (compound I). However, the rapid phase of spontaneous compound I reduction (2.3 s-1), typical of wild-type LiP, was absent in the Trp171 mutants, strongly suggesting that an electron-transfer pathway must exist within the protein leading from the heme to a surface site in close proximity to Trp171. The kinetic competence of such a pathway is dependent on interaction of the enzyme with VA, at or near Trp171.

Autocatalytic formation of a hydroxy group at C beta of trp171 in lignin peroxidase.

In the high-resolution crystal structures of two lignin peroxidase isozymes from the white rot fungus Phanerochaete chrysosporium a significant electron density at single bond distance from the C beta of Trp171 was observed and interpreted as a hydroxy group. To further clarify the nature of this feature, we carried out tryptic digestion of the enzyme and isolated the Trp171 containing peptide. Under ambient conditions, this peptide shows an absorbance spectrum typical of tryptophan. At elevated temperature, however, the formation of an unusual absorbance spectrum with lambda max = 333 nm can be followed that is identical to that of N-acetyl-alpha, beta-didehydrotryptophanamide, resulting upon water elimination from beta-hydroxy tryptophan. The Trp171 containing tryptic peptide isolated from the recombinant and refolded lignin peroxidase produced from Escherichia coli does not contain the characteristic 333 nm absorbance band at any temperature. However, treatment with 3 equiv of H2O2 leads to complete hydroxylation of Trp171. Reducing substrates compete with this process, e.g., in the presence of 0.5 mM veratryl alcohol, about 7 equiv of H2O2 is necessary for complete modification. We conclude that the hydroxylation at the C beta of Trp171 is an autocatalytic reaction which occurs readily under conditions of natural turnover, e.g., in the ligninolytic cultures of P. chrysosporium, which are known to contain an oxidase-based H2O2-generating system. No dependence on dioxygen was found for this oxidative process. Chemical modification of fungal lignin peroxidase with the tryptophan-specific agent N-bromo succinimide leads to a drastically reduced activity with respect to the substrate veratryl alcohol. This suggests that Trp171 is involved in catalysis and that electron transfer from this surface residue to the oxidized heme cofactor is possible under steady-state conditions.

Properties of xanthine dehydrogenase variants from rosy mutant strains of Drosophila melanogaster and their relevance to the enzyme's structure and mechanism.

Xanthine dehydrogenase, a molybdenum, iron-sulfur flavoenzyme encoded in the fruit fly Drosophila melanogaster by the rosy gene, has been characterised both from the wild-type and mutant files. Enzyme assays, using a variety of different oxidising and reducing substrates were supplemented by limited molecular characterisation. Four rosy strains showed no detectable activity in any enzyme assay tried, whereas from four wild-type and three rosy mutant strains, those for the [E89K], [L127F] and [L157P]xanthine dehydrogenases (in all of which the mutation is in the iron-sulfur domain), the enzyme molecules, although present at different levels, had extremely similar or identical properties. This was confirmed by purification of one wild-type and one mutant enzyme. [E89K]xanthine dehydrogenase. These both had ultraviolet-visible absorption spectra similar to milk xanthine oxidase. Both were found to be quite stable molecules, showing very high catalytic-centre activities and with little tendency to become degraded by proteolysis or modified by conversion to oxidase or desulfo forms. In three further rosy strains, giving [G353D]xanthine dehydrogenase and [S357F]xanthine dehydrogenase mutated in the flavin domain, and [G1011E]xanthine dehydrogenase mutated in the molybdenum domain, enzyme activities were selectively diminished in certain assays. For the G353D and S357F mutant enzymes activities to NAD+ as oxidising substrate were diminished, to zero for the latter. In addition for [G353D]xanthine dehydrogenase, there was an increase in apparent Km values both for NAD+ and NADH. These findings indicate involvement of this part of the sequence in the NAD(+)-binding site. The G1011E mutation has a profound effect on the enzyme. As isolated and as present in crude extracts of the files, this xanthine dehydrogenase variant lacks activity to xanthine or pterin as reducing substrate, indicating an impairment of the functioning of its molybdenum centre. However, it retains full activity to NADH with dyes as oxidising substrate. Mild oxidation of the enzyme converts it, apparently irreversibly, to a form showing full activity to xanthine and pterin. The nature of the group that is oxidised is discussed in the light of redox potential data. It is proposed that the process involves oxidation of the pterin of the molybdenum cofactor from the tetrahydro to a dihydro oxidation state. This conclusion is fully consistent with recent information [Romäo, M. J., Archer, M., Moura, I., Moura. J.J.G., LeGall, J., Engh, R., Schneider, M., Hof, P. & Huber, R. (1995) Science 270. 1170-1176) from X-ray crystallography on the structure of a closely related enzyme from Desulfovibrio gigas. It is proposed, that apparent irreversibility of the oxidative activating process for [G1011E]xanthine dehydrogenase, is due to conversion of its pterin to the tricyclic derivative detected by these workers. The data thus provide the strongest evidence available, that the oxidation state of the pterin can have a controlling influence on the activity of a molybdenum cofactor enzyme. Implications regarding pterin incorporation into xanthine dehydrogenase and in relation to other molybdenum enzymes are discussed.

Expression of lignin peroxidase H8 in Escherichia coli: folding and activation of the recombinant enzyme with Ca2+ and haem.

An engineered cDNA from Phanerochaete chrysosporium encoding both the mature and pro-sequence regions of Lip isoenzyme H8 (Lip) has been successfully overexpressed in Escherichia coli. The recombinant protein (LipP*) was sequestered in inclusion bodies. The reduced-denatured polypeptide has been purified by differential solubilization, and the active enzyme recovered after controlled in vitro refolding (albeit in low yield), by glutathione-mediated oxidation of disulphides, in a folding medium containing an intermediate concentration of urea, Ca2+, and haem. The procedure is analogous to that previously described for the production of active recombinant horseradish peroxidase (HRP-C*) from inclusion-body material. It is quite possible, therefore, that this type of procedure may be suitable for the recovery of most, if not all, active recombinant peroxidases. The resultant LipP* has spectral characteristics identical with that of the native enzyme as isolated from Phanerochaete chrysosporium. Its specific activity measured in the standard veratryl alcohol (VA) assay was 39 micromol of VA oxidized/min per mg of protein, a value which compares extremely favourably with that of the native enzyme (36 micromol of VA/min per mg). Although levels of active enzyme obtained are not yet as high as in the case of HRP-C* (1% conversion of crude inactive LipP* polypeptide into pure fully active Lip), it is envisaged that further refinement of the expression/folding/activation procedures will provide sufficient protein for biophysical characterization of both the wild-type and site-directed mutants.

Properties of rabbit liver aldehyde oxidase and the relationship of the enzyme to xanthine oxidase and dehydrogenase.

The properties of the molybdenum iron-sulfur flavoprotein, aldehyde oxidase from rabbit livers, have been further investigated in comparison with bovine milk xanthine oxidase. In agreement with earlier work, the ultraviolet/visible spectra indicate that the flavin and iron-sulfur centres of the enzymes are quite similar to one another. The molybdenum centres have been compared by EPR spectroscopy of molybdenum(V) and regarding re-insertion of the sulfido ligand of molybdenum into the desulfo enzyme forms. The pH optimum for sulfide insertion is approximately 2 lower for aldehyde oxidase than for xanthine oxidase. A detailed comparison of molybdenum(V) EPR signals has been made for the signals known as Arsenite, Slow and Rapid. Computer simulation of spectra in 1H2O and 2H2O, at 9 and 35 GHz was used. Slow signals from the two enzymes are scarcely distinguishable from one another. Under the conditions used, aldehyde oxidase yielded only the Rapid type 2 signal, whereas xanthine oxidase gives both the Rapid type 1 and 2 signals. The nature of the structural difference between the Rapid type 1 and type 2 signal-giving species is discussed. It is concluded that the molybdenum centres of xanthine oxidase and aldehyde oxidase are indeed similar to one another and that such differences as exist between their molybdenum(V) EPR signals and re-sulfuration properties are related to differences only in the substrate-binding sites. N-terminal amino acid analyses have been performed on peptides obtained by trypsin cleavage of aldehyde oxidase. Comparison with a sequence previously deduced [Wright, R. M., Vaitaitis, G. M., Wilson, C. M., Repine, T. B., Terada, L. S. & Repine, J. E. (1993) Proc. Natl Acad. Sci. USA 90, 10690-10694] makes it clear that the latter is not, as was assumed, that of a xanthine dehydrogenase but of an aldehyde oxidase. In contrast to the situation with xanthine oxidase, attempts to convert non-proteolysed aldehyde oxidase to a dehydrogenase form by treatment with dithiothreitol were unsuccessful. The reason for this is considered in the light of sequence data in the literature. The location of the NAD(+)-binding site is discussed, and the sequence data are also discussed in relation to the molybdenum, iron-sulfur and substrate-binding sites.

Use of rosy mutant strains of Drosophila melanogaster to probe the structure and function of xanthine dehydrogenase.

The usefulness in structure/function studies of molybdenum-containing hydroxylases in work with rosy mutant strains of Drosophila melanogaster has been investigated. At least 23 such strains are available, each corresponding to a single known amino acid change in the xanthine dehydrogenase sequence. Sequence comparisons permit identification, with some certainty, of regions associated with the iron-sulphur centres and the pterin molybdenum cofactor of the enzyme. Procedures have been developed and rigorously tested for the assay in gel-filtered extracts of the flies, of different catalytic activities of xanthine dehydrogenase by the use of various oxidizing and reducing substrates. These methods have been applied to 11 different rosy mutant strains that map to different regions of the sequence. All the mutations studied cause characteristic activity changes in the enzyme. In general these are consistent with the accepted assignment of the cofactors to the different domains and with the known reactivities of the molybdenum, flavin and iron-sulphur centres. Most results are interpretable in terms of the mutation affecting electron transfer to or from one redox centre only. The activity data provide evidence that FAD and the NAD+/NADH binding sites are retained in mutants mapping to the flavin domain. Therefore, despite some indications from sequence comparisons, it is concluded that the structure of this domain of xanthine dehydrogenase cannot be directly related to that of other flavoproteins for which structural data are available. The data also indicate that the artificial electron acceptor phenazine methosulphate acts at the iron-sulphur centres and suggest that these centres may not be essential for electron transfer between molybdenum and flavin. The work emphasizes the importance of combined genetic and biochemical study of rosy mutant xanthine dehydrogenase variants in probing the structure and function of enzymes of this class.