Glyoxal oxidase from Phanerochaete chrysosporium is a new radical-copper oxidase.

A free radical-coupled copper complex has been identified as the catalytic structure in the active site of glyoxal oxidase from Phanerochaete chrysosporium based on a combination of spectroscopic and biochemical studies. The native (inactive) enzyme is activated by oxidants leading to the elimination of the cupric EPR signal consistent with formation of an antiferromagnetically coupled radical-copper complex. Oxidation also leads to the appearance of a substoichiometric free radical EPR signal with an average g value (gav = 2.0055) characteristic of phenoxyl π-radicals arising from a minority apoenzyme fraction. Optical absorption, CD, and spectroelectrochemical measurements on the active enzyme reveal complex spectra extending into the near IR and define the redox potential for radical formation (E1/2 = 0.64 V versus NHE, pH 7.0). Resonance Raman spectra have identified the signature of a modified (cysteinyl-tyrosine) phenoxyl in the vibrational spectra of the active complex. This radical-copper motif has previously been found only in galactose oxidase, with which glyoxal oxidase shares many properties despite lacking obvious sequence identity, and catalyzing a distinct reaction. The enzymes thus represent members of a growing class of free radical metalloenzymes based on the radical-copper catalytic motif and appear to represent functional variants that have evolved to distinct catalytic roles.

Two of the substrates for glyoxal oxidase (glyoxal (OHCCHO) and methylglyoxal (CH 3 COCHO)) are found in extracellular fluid of ligninolytic cultures (10) and are likely to represent physiological substrates for the enzyme in a complex extracellular metabolic scheme (9,10). Purified glyoxal oxidase is catalytically inactive but can be activated by lignin peroxidase (11,12), suggesting a possible extracellular regulatory circuit for the control of H 2 O 2 production by glyoxal oxidase, and lignin peroxidase activity by H 2 O 2 .
P. chrysosporium glyoxal oxidase has been purified to homogeneity (11), cloned for high level expression in Aspergillus nidulans (13), and its biochemical properties have been determined (11). Glyoxal oxidase is an acidic monomeric glycoprotein with a naked molecular mass of 57 kDa (calculated from the nucleotide sequence; Ref. 14) requiring one g-atom of copper for full activity. In its catalytic reaction and copper requirement, glyoxal oxidase resembles another extracellular fungal enzyme, galactose oxidase (15,16) which efficiently catalyzes the conversion of alcohols to aldehydes with formation of hydrogen peroxide. Galactose oxidase (from Dactylium dendroides) contains copper and an essential protein free radical that is required for catalysis (16). The active site structure of galactose oxidase as revealed by crystallography (17) is illustrated in Scheme 1, showing the coordination of the copper ion by 2 histidine residues (His-496 and His-581), a simple tyrosinate (Tyr-495), and a covalently modified tyrosine (Tyr-272), cross-linked to a cysteinyl residue (Cys-228) to form a new, dimeric amino acid (cysteinyl-tyrosine). This feature has been identified by spectroscopic and modelling studies as the radical-forming site in galactose oxidase (18 -20).
Comparison of the nucleotide sequences for glyoxal oxidase and galactose oxidase shows no obvious homology in primary structure (14,21), with less than 20% amino acid sequence similarity based on standard sequence alignment algorithms. However, the present spectroscopic studies show a remarkable degree of similarity between the two enzymes at the level of active site structure and chemistry, identifying a free radicalcoupled copper catalytic motif in glyoxal oxidase. Spectroscopic and biochemical comparison of the two proteins suggests that glyoxal oxidase and galactose oxidase are functional variants catalyzing distinct reactions at nearly identical active sites.

MATERIALS AND METHODS
Recombinant P. chrysosporium glyoxal oxidase was isolated from growth medium of a stable transformant of A. nidulans (13) grown in Aspergillus transformant minimal medium containing 1% maltose as growth substrate, 70 mM NaNO 3 , 50 mM KH 2 PO 4 , 7 mM KCl, 2 mM MgSO 4 , 100 g/ml methionine, 1 g/ml biotin, and trace elements (22) at pH 4.6. Glyoxal oxidase was purified as described previously (11,14) except that the protein was first concentrated from culture media by ammonium sulfate precipitation (23). Enzyme activity was measured polarographically with an oxygen-sensitive electrode using methylglyoxal as substrate in the presence of the oxidant Na 2 IrCl 6 . For phosphate inhibition studies, glyoxal oxidase activity was measured using the oxygen electrode with dimethylsuccinate buffer (pH 4.5) or MES 1 buffer (pH 6.0). Redox-activated glyoxal oxidase was prepared by addition of a slight excess of Na 2 IrCl 6 to the enzyme solution. Protein concentration was determined by the method of Lowry (24).
Absorption spectra were recorded on a Varian Cary 5E UV-vis-NIR spectrometer interfaced to a microcomputer for data acquisition. EPR spectra were acquired on a ER300 EPR spectrometer equipped with X-band microwave bridge and a nitrogen flow system. Quantitative EPR measurements were performed as described previously (25). Circular dichroism (CD) spectra were recorded on an Aviv model 41DS UV-vis-NIR CD/MCD spectrometer (26). Copper analyses were performed on a Varian SpectrAA 20B atomic absorption spectrometer equipped with a GTA-96 graphite furnace. Spectrophotometric titrations were performed on a Hitachi U2000 UV-vis absorption spectrometer. Spectroelectrochemical studies on glyoxal oxidase used a special thin layer optical cell 2 based on modification of a commercial demountable IR cell (Spectra-Tech demountable liquid sampling cell) similar in plan to a cell recently described for IR spectroelectrochemistry applications (27). The cell is constructed by supporting a 1000 lines/inch gold grid working electrode (Buckbee-Mears Co., St. Paul, MN) between quartz disks separated by a 0.2 mm polyvinylchloride gasket, using silver/saturated AgCl reference electrode (0.22 V versus NHE; Ref. 28) and a platinum wire counter electrode. Electroanalytical measurements were performed with an EG&G model 263 potentiostat interfaced with a microcomputer for data acquisition and analysis. Resonance Raman spectra were collected with a custom McPherson 2061/207 spectrograph (0.67-m focal length, 600-groove grating, 7 cm Ϫ1 spectral resolution) using a Coherent Innova 302 krypton (647 nm) laser, a Kaiser optical super-notch filter, and a Princeton Instruments (LN-1100PB) liquid N 2 -cooled CCD detector. Spectra were obtained on samples in glass capillaries at 278 K using a 90°scattering geometry and a 10-min accumulation time. Sample integrity was verified by the observation of the same absorption spectrum (typical of active enzyme) before and after laser irradiation.

Redox Characterization
Purified native glyoxal oxidase is catalytically inactive, but can be oxidatively activated by lignin peroxidase (11) or by high potential inorganic oxidants including molybdicyanide (K 3 Mo(CN) 8 ) or hexachloroiridate (Na 2 IrCl 6 ), or Mn 3ϩ EDTA (which reacts relatively slowly). The redox-activated enzyme is unstable and reversibly loses activity on dialysis or exposure to chromatographic media, reverting to the reduced, inactive form. Native glyoxal oxidase exhibits a cupric EPR signal at low temperatures, which is completely eliminated by addition of strong oxidants (Fig. 1). Spin quantitation of this EPR signal shows that it represents 0.7-0.8 equivalents of Cu 2ϩ /enzyme, identical to the total copper content determined by atomic absorption spectroscopy. The disappearance of the EPR signal on oxidation is similar to the behavior previously reported for galactose oxidase (25,29) and is consistent with the formation of a free radical-coupled copper center (25), resulting in an EPR-silent oxidized complex with an antiferromagnetically coupled electronic ground state (16). Like galactose oxidase (25,30), glyoxal oxidase typically contains substoichiometric copper (0.7-0.8 Cu 2ϩ /mol) and the oxidized protein exhibits a minority EPR signal arising from the presence of a small amount of apoenzyme radical (Fig. 2). The radical signals for the two enzymes are very similar, each having an average g value near 2.005 characteristic of phenoxyl -radicals and revealing partly resolved hyperfine structure. In galactose oxidase the spectrum has been assigned by isotopic substitution (30) and ENDOR analysis (18) to the novel cysteinyl-tyrosinate phenoxyl (Cys-TyrO ⅐ ) with hyperfine interaction between the -electron of the phenoxyl radical and one ␤-methylene proton and a single ␣-ring proton of the tyrosyl residue. The strong spectroscopic correlations between the two enzymes are developed in more detail below.
Spectroelectrochemical measurements on glyoxal oxidase ( Fig. 3, left) provide an estimate of the redox potential of the radical-forming site. In these studies, the growth of intense absorption features associated with the radical-containing protein (see below) are used to monitor the progressive redox conversion of the sample, and the absorption/potential data are analyzed in terms of the Nernst equation (Equation 1).
A is the sample absorbance, A red and A ox are the limiting absorbances for fully reduced and fully oxidized enzyme, respectively, R is the gas constant (8.314 J/K⅐mol), T is the absolute temperature, n is the number of electrons involved in the redox couple, and F is the Faraday constant (96.48 kC/mol). A Nernst plot of the absorption intensities through the oxidation step (Fig. 3, right) averaged between increasing and decreasing potential scans is linear (r 2 ϭ 0.995) with a slope of 0.048 mV consistent with a single-electron step for the oxidation (n ϭ 1.2) and a midpoint potential (E 1/2 ϭ 0. 42 1. EPR spectra for glyoxal oxidase complexes. Spectra for native glyoxal oxidase (1.0 mM enzyme in 50 mM sodium phosphate buffer, pH 6.5) and redox-activated enzyme (same conditions, in presence of 1 equivalent of Na 2 IrCl 6 ). Instrumental conditions: microwave frequency, 9.39 GHz; microwave power, 10 milliwatts; modulation frequency, 100 kHz; modulation amplitude, 10 G; temperature 120 K. Inset, expansion of the m I ϭ ϩ1/2 copper hyperfine feature (solid line) and gaussian fit to five components having a 10.5 G separation (dotted line). diffusible high potential manganese chelates (illustrated here by Mn 3ϩ EDTA oxidation) may have a physiological role in regulation of the extracellular activity of glyoxal oxidase through the metabolic networking of the ligninolytic enzymes (1,2). The biological significance of this interaction is presently under further study. The stability of the redox-activated proteins (reflecting the stability of the free radical sites) differs between glyoxal oxidase and galactose oxidase. While the latter is stable for days at 4°C (25), the stability of the active glyoxal oxidase is markedly lower, with a half-life of 4 h for the radical under similar conditions (50 mM sodium phosphate buffer, pH 6.5).

Spectroscopic Correlations between Glyoxal Oxidase and Galactose Oxidase
Optical Absorption Spectra-The optical spectra of redoxactivated and native (inactive) glyoxal oxidase are shown in Fig. 4 (Table I), with the corresponding spectra previously obtained for galactose oxidase included for comparison. Blue, inactive glyoxal oxidase exhibits strong absorption in the visible spectrum at 451 and 678 nm (⑀ 451 ϭ 1875 M Ϫ1 cm Ϫ1 ; ⑀ 678 ϭ 1365 M Ϫ1 cm Ϫ1 ) that may be assigned as phenolate-to-Cu 2ϩ ligand-to-metal charge transfer (LMCT) and mixed LMCT and cupric ligand field (d 3 d) absorption, respectively, by analogy with the corresponding features in galactose oxidase (35). These assignments are further supported by resonance Raman spectroscopy (see below). Activation by oxidation of the protein results in conversion to a green form with the appearance of intense absorption bands in the blue (⑀ 448 ϭ 5700 M Ϫ1 cm Ϫ1 ) and red (⑀ 851 ϭ 4300 M Ϫ1 cm Ϫ1 ). The unusual red absorption band extending deep into the near IR region has been assigned in galactose oxidase as arising from ligand-to-ligand charge transfer associated with the presence of both a Cys-Tyr free radical ligand and an unmodified tyrosinate in the active site metal complex (36). The spectra for galactose oxidase and glyoxal oxidase are strikingly similar, suggesting nearly identical environments for the active site metal ions, in spite of the lack of obvious sequence homology between the two proteins (14).
The sensitivity of the active site of glyoxal oxidase to protons is reflected in spectrophotometric titrations. While the spectrum of the redox activated glyoxal oxidase is nearly unchanged over a wide pH range (4.5-8.5), more dramatic pHdependent changes occur in spectra of the inactive enzyme (Fig.  5). Titration of inactive glyoxal oxidase in 100 mM potassium phosphate buffer converts the enzyme to a purple form at low pH (⑀ 522 ϭ 1365 M Ϫ1 cm Ϫ1 ), maintaining two isosbestic points (at 504 and 568 nm) that identify the reaction as a simple two-state transition. More detailed analysis by Henderson-Hasselbach fitting of the titration data gives an estimate for the pK a of the protein group involved in the protonation step (pK a ϭ 7.7) (Fig. 5, inset). The spectral changes associated with the pH-dependent structure change (loss of 451 nm phenolateto-Cu 2ϩ LMCT absorption at low pH) suggest that the titratable group is a coordinated tyrosinate, ionizing more readily in the metal complex than is typical of the free amino acid (pK a ϭ 10.1; Ref. 37) and dissociating from the metal on protonation. The assignment to the unmodified tyrosinate (corresponding to Tyr-495 in the active site of galactose oxidase, Scheme 1) is supported by the observation that resonance Raman profiles at low pH reveal only vibrational modes of a Cys-Tyr group, whereas at high pH contributions from both tyrosinates are observed (see below).
In glyoxal oxidase, the acid titration of the coordinated tyrosinate reflects a perturbation of the phenolate basicity by approximately 2.5 pH units in the metal complex. The lack of sensitivity to protons for the corresponding low and high pH spectra of the redox activated enzyme implies a more dramatic perturbation of phenolate basicity in the radical-copper complex. This difference can be understood in terms of the decrease in overall charge of the ligand set associated with oxidation of the coordinated cysteinyl-tyrosinate anion to generate the neutral (Cys-TyrO ⅐ ) phenoxyl, making the copper a more strongly acidic metal center (38). The modulation of ligand acidity by metal interactions is an essential feature of ligand participation in catalysis in galactose oxidase (16) (see below). In that FIG. 2. Glyoxal oxidase free radical EPR signal. Spectrum of glyoxal oxidase (1 mM protein) oxidized with excess Na 2 IrCl 6 in 50 mM sodium phosphate buffer, pH 6.5. Instrumental conditions: microwave frequency, 9.55 GHz; microwave power, 2 milliwatts; modulation frequency, 100 kHz; modulation amplitude, 2 G; temperature 120 K. enzyme, the spectra of both active and inactive forms are relatively insensitive to pH, but there is evidence for a temperature-dependent proton transfer equilibrium in the active site and for proton-coupled anion binding (38).
Anion Interactions-Glyoxal oxidase is inhibited by phosphate in a pH-dependent fashion with K i values in the millimolar range (K i (pH 4.5) ϭ 5.1 mM; K i (pH 6.0) ϭ 1.0 mM) correlating with the ionization of the weak acid and preferential binding of the dibasic ion. Unlike azide, which blocks turnover by binding to copper (see below), phosphate has no significant effect on the optical spectrum of redox-activated enzyme, indicating that the inhibition is not a consequence of direct interaction with the active site metal ion and may result from steric effects, such as blocking substrate access to the active site. Inhibition of enzyme activity by phosphate may be physiologically important for regulation of glyoxal oxidase activity and can be expected to be a significant factor in environmental chemistry.
Azide acts as an analog of peroxide in oxygen-metabolizing enzymes, forming stable complexes that mimic reactive oxygen species. Addition of azide to both native and redox activated forms of glyoxal oxidase leads to the appearance of new spectral features (Fig. 6, Table I), that are again similar to those observed for the azide complexes of galactose oxidase and other copper proteins (39 -42). By analogy, the intense absorption near 380 nm is assigned to azide-to-Cu 2ϩ LMCT. The lower energy absorption bands have more metal-centered d 3 d character, as indicated by their relatively large CD anisotropy (see below). As previously observed for galactose oxidase (25,38), there is a significant difference in affinity for azide between active and inactive forms; active glyoxal oxidase binds azide much tighter (K D ϭ 0.5 mM) than the native enzyme (K D ϭ 25 mM). Although the trend in affinity parallels the results for galactose oxidase, the dissociation constants for both forms are approximately 25 times (active) and 130 times (native) larger than the corresponding values found for that enzyme.
The similarity in the shifts in absorption bands between the two enzymes on binding exogenous ligands is particularly interesting in view of the evidence for pseudorotation of the copper complex and an intrinsic proton transfer coordinate in the active site of galactose oxidase, relating to substrate activation in catalysis (16,38). Previous studies on galactose oxidase have demonstrated that coordination of an exogenous ligand results in a displacement of the unmodified tyrosinate (Tyr-495) through a fluxional interchange of strong and weak interactions in the metal complex that can be regarded as a pseudorotation distortion of the active site (16,38). Displacement of the tyrosinate ligand increases its basicity, allowing it to serve as a general base for a proton abstraction step in catalysis (38). In galactose oxidase, the pseudorotation mechanism leads to a thermodynamic linkage between anion binding and proton uptake (16). The close correspondence in the spec-  34, 8.15, 7.98, 7.86, 7.71, 7.50, 7.39, 7.16, 7.10, 6.92, 6.44, 4.53). Inset, absorption data fit to a protonation equilibrium model with pK a ϭ 7.7.  troscopic behavior of the two proteins under ligand perturbations indicates that not only are the two active sites structurally related but that these dynamical motifs are reproduced as well. Circular Dichroism Spectra-Because of the signed character of CD intensity, circular dichroism spectroscopy increases the resolution of absorption spectra. The CD spectrum of active glyoxal oxidase (Fig. 7A) resolves six visible absorption bands, closely matching features observed in CD for the aquo complex of galactose oxidase (25,38), but shifted to lower energy consistent with a weaker ligand field for the copper complex in the former enzyme. Even at this higher resolution, the similarity between the two enzymes is preserved in both intensity and sign of the polarization. For native glyoxal oxidase (Fig. 7B), the strongest polarization occurs in the ligand field absorption bands at 650 and 800 nm. In the redox-activated enzyme, the ligand field spectra, identified by their relatively large CD anisotropy (⌬⑀/⑀) are shifted to higher energy (600 nm) compared to the inactive form. Azide binding shifts the ligand field spectra of both enzyme forms to higher energy as expected for conversion to a more nearly tetragonal complex, an important feature of active site ligand interactions previously studied in galactose oxidase (38,43). The optical spectra of glyoxal oxidase and galactose oxidase are thus similar in both general features and essential details.
Resonance Raman Spectra-Resonance Raman spectroscopy is an important structural tool that provides detailed vibrational information for a specific chromophore through resonance enhancement of Raman intensities on excitation of electronic transitions. The resonance Raman spectra for glyoxal oxidase in its redox-activated form (Fig. 8A) and its azide adduct (Fig. 8C) are remarkably similar to those obtained for the corresponding forms of galactose oxidase (Fig. 8, B and D). In the case of active galactose oxidase (Fig. 8B), the features at 1170, 1246, 1499 (sh), and 1603 cm Ϫ1 were assigned to the phenolate modes of the Tyr-495 ligand (Scheme 1) on the basis of their similarity to phenolate vibrations in other metal-tyrosinate proteins and model complexes (36). The dominant inten-sity of the 1246 cm Ϫ1 C-O stretch is also typical of metaltyrosinates (44). The absence of these four modes in the Raman spectrum of the azide adduct (Fig. 8D) suggests that this Tyr is displaced from copper due to the increased ligand field strength arising from replacement of the coordinated water by azide. Glyoxal oxidase exhibits remarkably similar resonance Raman spectra. Excitation of the active enzyme at 647 nm (Fig. 8A) reveals vibrational modes at 1170, 1249, and 1604 cm Ϫ1 , which, by analogy, can be assigned to a coordinated phenolate. These modes become the dominant RR spectral features with 458 nm excitation (data not shown), revealing that they are all due to the same chromophore and that the absorption in the 460 nm region is due primarily to phenolate 3 Cu 2ϩ LMCT. Azide binding to glyoxal oxidase again causes the disappearance of these modes (Fig. 8C), suggesting a similar release of the axial phenolate.
The copper-coordinated phenoxyl radical (Cys-Tyr 272) in active galactose oxidase is associated with another set of vibrational modes at 1382, 1487, and 1595 cm Ϫ1 (Fig. 8B), which persist in the azide adduct ( Fig. 8D) (36,45). These three modes are characteristic of the Raman spectra of phenoxyl radicals, with the 1487 cm Ϫ1 feature being assigned to the C-O stretch on the basis of its isotope dependence in model compounds (36,46). The residual features at 1185, 1246, 1312, 1417, and 1439 cm Ϫ1 in the azide complex (Fig. 8D) are also assigned to vibrations of the phenoxyl radical. Active glyoxal oxidase exhibits a similar set of major features at 1375, 1486, and 1591 cm Ϫ1 with 647 nm excitation (Fig. 8A). Both these and the minor features at 1250, 1309, 1421, and 1444 cm Ϫ1 are again found in the azide adduct (Fig. 8C). All of these modes have diminished intensity with 458 nm excitation (data not shown), indicating that they are all due to the same chromophore. The similarity of the vibrational frequencies and intensities to those of galactose oxidase allows us to conclude that active glyoxal oxidase has a similar (Cys-Tyr) phenoxyl radical as a copper ligand.
The native (inactive) form of glyoxal oxidase exhibits distinctive resonance Raman features for both the unmodified tyrosinate and Cys-Tyr ligands (Table II) Similarly, in the presence of azide, only Cys-Tyr modes are observed at 1272, 1483, and 1599 cm Ϫ1 . Thus, as in the case of the oxidized enzyme, the resonance Raman profiles again indicate that the unmodified tyrosinate ligand no longer binds to copper in the presence of azide and also does not bind at low pH. The enhancement of the Cys-Tyr modes with 647 nm excitation demonstrates that the unusually intense absorption bands in the 600 -650 nm region of the reduced enzyme (Figs. 4 and 6) have substantial Cystyrosinate 3 Cu 2ϩ charge transfer character in addition to the Cu 2ϩ ligand field character indicated by CD spectroscopy (Fig. 7).
EPR Spectra-The EPR spectrum recorded for native glyoxal oxidase (Fig. 1) is characteristic of a cupric ion in a tetragonal ligand environment. The g values measured at the turning points of the EPR absorption spectrum (g z ϭ 2.28; g y ϭ 2.08; g x ϭ 2.03) and the copper nuclear hyperfine splitting (a z (copper) ϭ 172 G) closely reproduce the values for the inactive cupric galactose oxidase complex (25,38). Additional structure pres-ent on the copper hyperfine features can be resolved into five components with a uniform splitting of 10.5 G consistent with a small transferred ligand hyperfine interaction arising from a pair of equivalent I ϭ 1 nuclei (2nI L ϩ1 ϭ 5; n ϭ 2). An interpretation of the nearly identical structure observed in EPR spectra for galactose oxidase in terms of ligation by 2-3 nitrogen nuclei ( 14 N, I ϭ 1) (47) has been confirmed by the results of protein crystallography showing coordination of the active site copper by 2 histidines (17). Based on the extraordinary spectroscopic correlations between the two enzymes, it appears likely that this structural pattern also appears in the metal binding site of glyoxal oxidase.

Contrasts between the Two Enzymes
In spite of the spectroscopic similarities of glyoxal oxidase and galactose oxidase, significant differences exist that clearly distinguish the two enzymes, including the redox potential and stability of the free radical, anion affinity, and catalytic chemistry. The reactions catalyzed by the two enzymes are distinct and there is little overlap between their substrates. Galactose oxidase selectively oxidizes primary alcohols (15) (although slow oxidation of aldehydes has recently been detected; Ref. 48), while the reverse is true for glyoxal oxidase, with the best aldehyde substrate (methylglyoxal) supporting nearly 100 times higher catalytic rate (k cat /K m ) compared to dihydroxyacetone, the ␣-hydroxycarbonyl substrate that gives the best turnover (11). The substrates recognized by glyoxal oxidase may in fact be gem diols since simple aldehydes are extensively hydrated to acetals in aqueous solution (49,50), accounting for the elements of water in the mass balance equation for the reaction. Alternatively, solvent coordinated to copper in the active site may serve to hydrate carbonyl-containing substrates as they bind to the enzyme.
The distinct redox potentials for the radicals in the two enzymes may be important in the context of the catalytic reactions, the higher potential redox site being favorable for aldehyde oxidation chemistry. The tuning of the redox potentials of the metal and radical may be accomplished by varying the strength of ligand interactions in the active site complex. Inorganic modelling studies on a chelate that reproduces the complex functionality of the active site of galactose oxidase shows that metal interactions can shift the redox potential of a thioether substituted phenolate by more than 0.5 V from the midpoint potential of the free ligand (51).

Conclusions
Spectroscopic and biochemical characterization of the active sites of glyoxal oxidase and galactose oxidase reveals an amazing degree of similarity between two proteins essentially lack-SCHEME 1. ing any obvious sequence homology and catalyzing distinct reactions. As a result of the spectroscopic similarities between the two enzymes, much is already known about the active site of glyoxal oxidase by analogy with the crystallographically characterized galactose oxidase structure. Aside from the redox tuning of the free radical side chain, the basic features of the catalytic complexes appear to be nearly identical in the two proteins. Resonance Raman spectra provide strong evidence for a Cys-Tyr phenoxyl radical site in glyoxal oxidase, demonstrating that this unusual covalent modification occurs in contexts other than galactose oxidase. Crystallographic studies on glyoxal oxidase currently in progress 3 can be expected to provide the basis for more extensive structural comparisons of the two proteins.