E. L. Raven, Department of Chemistry, Henry Wellcome Building, University of Leicester, University Road, Leicester LE1 9HN, UK Fax: +44 116 2523789 Tel: +44 116 2297097 E-mail: firstname.lastname@example.org P. Moody, Department of Biochemistry and Henry Wellcome Laboratories for Structural Biology, University of Leicester, Leicester LE1 9HN, UK Fax: +44 116 252 7084 Tel: +44 116 2297097 E-mail: email@example.com
Guaiacol is a universal substrate for all peroxidases, and its use in a simple colorimetric assay has wide applications. However, its exact binding location has never been defined. Here we report the crystal structures of guaiacol bound to cytochrome c peroxidase (CcP). A related structure with phenol bound is also presented. The CcP–guaiacol and CcP–phenol crystal structures show that both guaiacol and phenol bind at sites distinct from the cytochrome c binding site and from the δ-heme edge, which is known to be the binding site for other substrates. Although neither guaiacol nor phenol is seen bound at the δ-heme edge in the crystal structures, inhibition data and mutagenesis strongly suggest that the catalytic binding site for aromatic compounds is the δ-heme edge in CcP. The functional implications of these observations are discussed in terms of our existing understanding of substrate binding in peroxidases [Gumiero A et al. (2010) Arch Biochem Biophys500, 13–20].
The family of heme peroxidases catalyse the hydrogen peroxidase-dependent oxidation of a wide variety of different substrates. With a few exceptions, the substrate is a small organic molecule. Our knowledge of how these organic substrates bind to peroxidases has developed relatively recently, and is based on crystallographic information for a number of different heme peroxidase enzymes. The first structures to appear [1–3] showed binding of aromatic substrates close to the so-called δ-heme edge (reviewed recently in ). These structures were consistent with other observations, e.g. from earlier chemical modification work [5–7], in which substrate binding at the δ-heme edge was implicated. Subsequent structures for the ascorbate peroxidase–ascorbate complex  and for nitric oxide synthase in complex with its cofactor tetrahydrobiopterin [9,10], together with that previously known for the manganese peroxidase–Mn(II) complex , revealed binding at the γ-heme edge, and thus it became clear that substrate binding at the δ-heme edge was not the only means by which the enzyme and substrate might productively associate with one another. This is presumably helpful for the protein, because it means that oxidation of different types of substrate can be accommodated within the same protein framework. In ascorbate peroxidase, for example, (hydrophobic) aromatic substrates and (hydrophilic) ascorbate are oxidized at different sites [8,12].
Guaiacol is a universal substrate for peroxidases, and is oxidized by all known peroxidases. The end product, tetraguaiacol (formed from polymerization of the radical product), is intensely coloured, which makes it very simple to detect. So widespread is guaiacol activity that this simple assay has become an easy way of identifying and quantifying activity across different peroxidases. However, the actual binding site for guaiacol has never been unambiguously identified. Here we present structures of cytochrome c peroxidase (CcP) in complex with guaiacol, and we test the sites identified in crystal structures (by competition assays and site-directed mutagenesis) for their involvement in the enzymatic reaction.
Crystal structure of guaiacol and phenol-bound CcP
The overall structure of CcP in complex with guaiacol and with phenol is shown in Fig. 1A. No major structural rearrangements of the protein occur on binding of the substrate or phenol as compared with the crystal structure of CcP in the absence of guaiacol (data not shown). Two molecules of guaiacol are bound to the protein, and phenol is found to bind to occupy the same sites. The first site is in a hydrophobic pocket defined by Phe89 and Phe108, shown in Fig. 1B and labeled 1 in Fig. 1A. This site is ∼ 23 Å from the heme. The second site is ∼ 15.5 Å from the heme iron, and is shown in Fig. 1C and labeled 2 in Fig. 1A. The hydroxyl group of guaiacol hydrogen bonds to the backbone carbonyl of Ile40 (2.74 Å), and a hydrogen-bonding network links guaiacol to the 6-propionate of the heme via Gly41 and three water molecules. For comparison, an overlay of the structure of CcP in complex with isoniazid (INH)  [Protein Data Bank (PDB) entry 2V2E] with the CcP–guaiacol complex in the region of the δ-heme edge is shown in Fig. 1D. INH is known  to bind close to the δ-heme edge, and is thus a useful comparative structure for the guaiacol structure (the Km for guaiacol is 53 mm ). In the case of the INH complex, the key interactions are hydrogen bonds between the NH group of INH and the backbone carbonyl of Pro145 (2.95 Å), the pyridine nitrogen of INH and Ser185 (via a H2O molecule), and the carbonyl group of INH and the NH group of Arg48 (2.84 Å).
It was not clear from the data above whether or not these sites identified crystallographically are catalytically relevant, as they are distant from the catalytic sites used in other peroxidases . Competitive inhibition studies were therefore carried out, as guaiacol oxidation would be expected to be unaffected by INH binding at the δ-heme edge if there were two binding sites at different places.
Steady-state assays for guaiacol oxidation were carried out in the presence of various concentrations of INH (1–10 mm). Values of Vmax remained unchanged at each INH concentration (Table 1). However, as the concentration of INH was increased, the apparent Km value also increased. Figure 2A shows Lineweaver–Burk plots for the oxidation of guaiacol by CcP in the presence of INH; the data are consistent with competitive inhibition of guaiacol oxidation by INH. A replot of the slope of the reciprocal plots versus [INH] was linear (Fig. 2A, inset), which was also consistent with competitive inhibition, and could be used to obtain the dissociation constant, Ki, for the binding of INH to CcP from the x-axis intercept (Ki(INH) = 2.9 ± 0.2 mm). In contrast, INH did not inhibit cytochrome c oxidation (kcat and Km were unaffected) (Fig. 2B). Hence, there was no change in the slope or y-axis intercept of the Lineweaver–Burk plots for the oxidation of cytochrome c in the presence of INH, consistent with the two substrates binding at different locations.
Table 1. Variation of steady-state oxidation of guaiacol in the presence of INH (pH 6.0, 100 mm potassium phosphate, 25.0 °C). The apparent increase in kcat for wild-type CcP as INH concentration is increased and the increasing error in Km are attributable to the inability to produce accurate Michaelis–Menten plots in the presence of INH, as the concentrations of guaiacol needed to reach Vmax at high INH concentrations are not accessible.
4.1 ± 0.1
53 ± 6
5.8 ± 0.2
28 ± 2
1.2 ± 0.1
19 ± 1
1 mm INH
7.5 ± 0.4
150 ± 10
5.8 ± 0.3
45 ± 4
1.4 ± 0.1
43 ± 1
5 mm INH
11 ± 3
440 ± 160
5.2 ± 0.4
48 ± 7
1.1 ± 0.1
120 ± 10
10 mm INH
12 ± 7
1060 ± 620
5.3 ± 0.5
100 ± 10
1.1 ± 0.1
170 ± 20
Crystal structure of the CcP M119W–guaiacol complex and enzyme inhibition
The inhibition studies suggested that INH and guaiacol compete for the same binding site (presumably the δ-heme edge, as established crystallographically for INH ), which was in contrast to the crystallographic observations above showing that guaiacol bound at sites distinct from the δ-heme edge. In order to probe these discrepancies further, a tryptophan was introduced to replace Met119. Met119 is positioned close to Ile40 (Fig. 1C), and the substitution of a tryptophan at this position was proposed to sterically block guaiacol binding, allowing determination of whether this site is catalytic.
Figure 3A shows the overall crystal structure of the CcP M119W–guaiacol complex. Only one molecule of guaiacol is bound in this structure, and it is positioned at the site close to Phe89 and Phe108 (site 1 in Fig. 1A), in the same way as in the CcP–guaiacol complex (Fig. 1A). The new tryptophan at position 119 has blocked the binding of guaiacol close to Ile40. In addition, there is a structural change in the loop composed of Ala193–Phe198 (Fig. 3B); this slight structural rearrangement accommodates the bulkier tryptophan side chain, but does not allow guaiacol binding at this site.
As with the CcP–guaiacol structure, guaiacol is not observed at the δ-heme edge, so if the binding site close to Ile40 was a catalytically active site in CcP, then CcP M119W would be expected to show a decreased ability to oxidize guaiacol (as there is no binding at this site for CcP M119W). However, steady-state kinetic data did not support this, because they showed that values of kcat and Km for guaiacol oxidation were essentially unchanged (Table 2). Assuming that the site close to Phe89/Phe108 is also not catalytically active (because it is too distant from the heme), we conclude that oxidation must be occurring only at the δ-heme edge, but that we do not observe binding at this site under the conditions of our crystallographic experiments. Inhibition studies with CcP M119W confirmed this (Fig. 2C), as INH competitively inhibited guaiacol oxidation [Ki(INH) = 3.2 ± 0.6 mm; Fig. 2C, inset], indicating that guaiacol and INH are binding at the same site.
Table 2. Steady-state kinetic data for guaiacol oxidation by CcP (pH 6.0, 100 mm potassium phosphate, 25.0 °C).
4.1 ± 0.3
53 ± 6
0.08 ± 0.01
1.2 ± 0.1
20 ± 1
0.06 ± 0.01
5.9 ± 0.2
26 ± 2
0.23 ± 0.01
In a further mutagenesis experiment, we attempted to completely block aromatic binding at the δ-heme edge with the S81W mutation (intended to occupy the INH binding site), but this was not successful, because steady-state assays showed guaiacol oxidation by CcP S81W to be unchanged as compared with the wild type (Table 2), and because competitive inhibition by INH was still observed [Ki(INH) = 1.8 ± 0.2 mm; Fig. 2D], showing that both guaiacol and INH were still able to bind, and at the same site. The crystal structure (Fig. S1) helps to rationalize this, because it shows that the side chain of Trp81 swings outwards to form a hydrogen bond with Asp146 instead of blocking the δ-heme edge, as predicted. This probably allows substrate binding at the δ-site.
A picture of substrate binding for small, aromatic substrates at the δ-heme edge emerged largely from chemical modification work (reviewed in [5,6,14,15]), although NMR also played an important role [16–23]. A few structures appeared at the end of the 1990s for horseradish peroxidase and Arthromyces ramosus peroxidase in complex with various small organic molecules [1–3], showing binding at the δ-heme edge, and there is now a collection of other crystal structures in which binding of various organic compounds has been observed at the same site (not all of which are genuine substrates) [1–3,12,13,24–27]. For CcP, there was a general consensus of opinion that small substrates were bound close to the δ-site [7,28–30]. The actual binding site for guaiacol has never been identified, although there is a report of phenol binding to an artificially created cavity in CcP .
In our structure, we observed guaiacol binding at two sites, close to Phe68 and Ile40. Guaiacol was not seen bound at the δ-heme edge in the crystal structure, but the inhibition and mutagenesis data strongly suggest that it binds there, because INH and guaiacol compete for the same site. We conclude that the catalytic binding site for guaiacol is at the δ-heme edge, as for other substrates in other peroxidases , but that the binding interaction is weak at this site and not seen crystallographically. This would be consistent with conclusions drawn from a much more extensive survey  of substrate binding, in which it has been noted that neutral substrates bind rather weakly. We thus conclude that the binding sites observed in our structure are a consequence of nonspecific binding of guaiacol, and that these aromatic binding sites can also be occupied by phenol. [The approximate binding affinities for INH (judged from the Ki of 2.70 mm) and guaiacol (judged from the Km of 53 mm ) give an indication of the relative strengths of the interactions, and explain why we were able to observe INH at the δ-site but not guaiacol]. The observation that the binding of these molecules is relatively nonspecific (and of low affinity) may account for the widespread activity of peroxidase enzymes for guaiacol.
Guaiacol (Aldrich Chemical Co., Milwaukee, WI, USA), phenol and buffers (Fisher) were all of the highest analytical grade (> 99% purity) and used without further purification. INH (minimum 99% purity) was purchased from Sigma. Water was purified with an Elga PURELAB purification system, and all buffers were filtered (0.2 μm) prior to use. Hydrogen peroxide solutions were freshly prepared by dilution of a 30% (v/v) solution (BDH): exact concentrations were determined from the published absorption coefficient (ε240 nm = 39.4 m−1·cm−1) . All molecular biology kits and enzymes were used according to the manufacturer’s protocols.
Mutagenesis, protein expression, and purification
Site-directed mutagenesis of CcP was performed according to the QuikChange protocol (Stratagene, Cambridge, UK). For the M119W mutation, the primers were 5′-GCTGTGCAGGAATGGCAGGGTCCC-3′ (forward primer) and 5′-GGGACCCTGCCATTCCTGCACAGC-3′ (reverse primer) (mutations in bold). For the S81W mutation, the primers were 5′-GAGTTTAACGATCCATGGAATGCGGGC-3′ (forward primer) and 5′-GCCCGCATTCCATGGATCGTTAAACTC-3′ (reverse primer) (mutations in bold). Recombinant CcP (a Y39A/N184R mutant that has been optimized for crystallization, referred to as CcP in this article), CcP M119W and CcP S81W were prepared and isolated with modifications to published procedures . Enzyme purity was assessed by examination of the Asoret/A280 nm value: in all cases, an Asoret/A280 nm value of > 1.2 for CcP was considered to indicate purity. Enzyme purity was also assessed with SDS/PAGE, and the preparations were judged to be homogeneous by the observation of a single band on a Coomassie Blue-stained reducing SDS/polyacrylamide gel. Absorption coefficients were determined, with the pyridine–haemochromogen method , to be 108 mm−1·cm−1 and 102 mm−1·cm−1 for CcP S81W and CcP M119W, respectively.
Electronic absorption spectroscopy
Spectra were collected with a Perkin-Elmer Lambda 35 or 40 spectrophotometer, linked to a PC workstation running uv-winlab software.
Crystals were prepared by microdialysis with 100 μL of a 10–30 mg·mL−1 solution of enzyme in 500 mm potassium phosphate (pH 6.0) against 10 mL of 50 mm potassium phosphate (pH 6.0) containing 30% 2-methyl-2,4-pentanediol by volume. The crystals were grown at 4 °C, and once grown the crystals were soaked either in mother liquor containing 100 mm phenol for 5 min or in 150 mm guaiacol solution for 10 min; guaiacol was solubilized in 20% methanol before being added to the mother liquor. After soaking, crystals were rapidly cooled to 100 K.
Data collection and refinement
Diffraction data were collected in-house with a Rigaku RU2HB X-ray generator equipped with a copper anode and Xenocs multilayer optics and an R-Axis IV detector. All data were collected at 100 K. Data were indexed, integrated and scaled with mosflm  and scala . Data collection statistics are shown in Table 3; 5% of the data were flagged for the calculation of Rfree, and excluded from subsequent refinement. The structures were refined from the 1.70-Å CcP structure  (PDB entry 2CYP). refmac5  from the ccp4 suite  was used for all refinement. Calculation of difference Fourier maps showed clear and unambiguous electron densities for bound guaiacol and phenol molecules in all of the structures. Guaiacol and phenol were incorporated into the last cycles of refinement. coot  was used throughout for manual adjustment, ligand fitting, and interpretation of the water structure. All crystal structure figures were created with pymol .
Table 3. Data collection and refinement statistics for the CcP–guaiacol complex (PDB code 4a6z), CcP–phenol complex (4a71), CcP M119W–guaiacol complex (4a78), and CcP S81W (4a7m). Values in parentheses refer to the outer resolution bin.
Unit cell (Å)
192 254 (2692)
68 320 (9587)
62 544 (7656)
105 421 (15 609)
46 241 (1991)
21 669 (3096)
24 226 (3082)
43 966 (6177)
Rmsd from ideal
Steady-state kinetic experiments
Steady-state oxidations of guaiacol (2-methoxyphenol) in 100 mm potassium phosphate (pH 6.0, 25 °C) were carried out according to published protocols , both in the absence of INH and in the presence of INH (1–10 mm). Steady-state data were fitted to the Michaelis–Menten equation as described previously .
We thank G. Mauk for the gift of the CcP expression vector and I. Efimov for helpful discussions.