New Complexes of Peroxidases with Hydroxamic Acids, Hydrazides, and Amides*

SUMMARY Horseradish peroxidase forms spectroscopically distinct, reversible complexes with hydroxamic acids (R-CO-NHOH), hydrazides (RCO-NHNH2), amides (RCONH2), and oc-hydroxyketones (RCO--CHsOH). Binding of these compounds to the enzyme depends on the polar and steric character of R and the hydrogen bonding capacity of-CO-X-Y (X-Y = NH-OH, NH-NH,, NH-H, CH,OH). Hydroxamate anions and hydrazide cations do not associate with the enzyme. The dissociation constants (IL) for the enzyme-RCOXY complexes span seven orders of magnitude (Kl - 0.3 to 2 x lop7 M), the greatest affinity being shown by compounds with a planar, aromatic R group. This is at-tributed to an interaction of the R moiety at an apoprotein hydrophobic crevice. Spectrophotometric, electron paramagnetic resonance, and ma.gnetic susceptibility measurements indicate that the association of horseradish peroxidase with hydroxamic acids entails a transition a mixed spin of the enzyme a high spin derivative. of

From, the Department of Biochemistry, University of Alberta Medical Xchool, Edmonton, Alberta, Canada, and the Johnson Research Foundation, University of Pennsylvania, Philadelphia, Pennsylvania 19104 SUMMARY Horseradish peroxidase forms spectroscopically distinct, reversible complexes with hydroxamic acids (R-CO-NHOH), hydrazides (RCO-NHNH2), amides (RCONH2), and oc-hydroxyketones (RCO--CHsOH). Binding of these compounds to the enzyme depends on the polar and steric character of R and the hydrogen bonding capacity of-CO-X-Y (X-Y = NH-OH, NH-NH,, NH-H, CH,OH). Hydroxamate anions and hydrazide cations do not associate with the enzyme.
The dissociation constants (IL) for the enzyme-RCOXY complexes span seven orders of magnitude (Kl -0.3 to 2 x lop7 M), the greatest affinity being shown by compounds with a planar, aromatic R group. This is attributed to an interaction of the R moiety at an apoprotein hydrophobic crevice. Spectrophotometric, electron paramagnetic resonance, and ma.gnetic susceptibility measurements indicate that the association of horseradish peroxidase with hydroxamic acids entails a transition from a mixed spin state of the enzyme to a high spin derivative.
The spectroscopic characteristics of enzyme-RCOXY complexes are similar, suggesting that X-Y substituents do not interact directly with the metal ion of the prosthetic group but perturb its environment.
Analogous conclusions were drawn from (a) the parallelism between RCOXY affinities for manganic and ferric peroxidases which does not pertain to ligands (e.g. F-) substituting in the first coordination sphere of the metal ion and (b) the lack of pronounced spectroscopic changes in RCOXY-ferroperoxidase complexes. The association of peroxidase with hydroxamic acids is competitively inhibited by specific enzyme substrates (hydrogen donors), permitting the evaluation of so far unknown enzyme- (donor) substrate binding parameters (KS). Such a competitive behavior also implies the proximity of RCOXY to the active site.
RCOXY compounds influence heme-linked ionizations and ligand interchange reactions, e.g. they inhibit the formation of alkaline peroxidase and peroxidase-cyanide complex. With cyanide and RCOXY, ferriperoxidase gives tertiary * Communicated in part at the 154th Meeting of the -4merican Chemical Societ,y, Chicago (September 1967 The current work originated with the observation that aromatic peracids are exceptionally efficient oxidants of horseradish peroxidase, the 1:l reaction resulting in the formation of compound I and release of the parent carboxylic acid (I).
Two interpretations are consonant with these observations: (a) aromatic peracids form a complex with the enzyme which rearranges to compound I concurrently with the scission of the O-O bond, or (b) in its complex with the enzyme, the peracid is rapidly hydrolyzed , giving the physiological oxidant H202 at, or near, the active site. In either case, peracids appear to have a greater affinity for the enzyme than hydrogen peroxide or its monoalkyl derivatives.
In view of these similarities, it seemed that Compounds ;I t,o C in Schcmr: 1 might serve as excellent probes of the active site, permitting the assessment of alternatives a and b. This proved to be the case. It has now been established that RCOXY compounds (where R is an aliphatic, alicyclic, or aromatic residue, and X-Y = NH-OH, T\'H-NHZ, CH-OH, etc.) are not hydrolyzed in the presence of H-peroxidasel but form stable, reversible complexes with the enzyme.
In this communication we delineate the spectroscopic and magnetic properties of such RCOXY-peroxidase complexes and present a hypothesis on the nature of enzyme-ligand interactions.
In another commm~icntion we shall outline the kinetics of RCOXY oxidation by compounds I and II, showing that hydroxamic acids are extraordinarily efficient reductants of enzymeperoxide derivatives.
IIgdroznmic dcids-Most of the compounds used in this study were synthesized in this laboratory, some by courtesy of Dr. J. Wodxk, by standard h!drox~aminolysis of the corresponding carboxylic eaters or acid chlorides (25). C, H, and N am&es of these derivatives, carried out by Galbraith Laboratories, Knoxville, Term., were within ~0.2% of the theoretical values. Some compounds wcrc checked against commercial preparations available from Hayncs Chrmical Research Corp., Lugoff, S. C. (o-fluorobcnz-and p-h-tlrosSbeilzhydrosamic acids) ; and Raylo Chemicals, Edmonton, .\lbertn, Canada (I-naphtho-, 2-naphtho-, m-iodobenz-, and cyclohexyl hydrosamic acids). 0-Benzopl hydrosylnminc was prepared by Jencks' method (26). Other highest purity reagents were obtained either from Fisher Scientific Co. or from Mdrich Chemica.1 Co. All solutions were made in water distilled twice, once from alkaline permangsnate.
Ethanol was the solvent for a-hydroxyacetophenonr.
Except for p-hydroxybenzhydroxamic acid, which is autosidizable, weakly acid solutions of hydroxamic acids and hydrazides ma.? be stored at 4" for several days without detectable decomposition (27). d[easwements of Equilibrium Constants-The formation of enzyme-ligand complexes (HRP .L) was followed spectrophotometrically at 410 mu. .W this wave length, the difference in molar nbsorptivities (le,lo) is greatest (Fig. 2), the ligands genera,lly do not absorb rind the measurements are most accurate, since 410 am is near t.he absorption maxima of H-peroxidase (403 nm) and its hydroznmic acid complexes (408 nm) (Fig. 1).
The dependence of the equilibrium constant on hydronium ion concentration was examined in 5 111~ acttate, pH 4 to 5; 10 mM 2,2'-dimethyl glutarate, pH 4 to 6; and 20 to 80 IXM phosphate, pH 5.8 to 8. Borate (30 mllr) was used only in titrations with benzamide, since it forms complexes with hydroxamic acids (28). Over the entire pH range, the equilibrium constants were nearly independent of ionic strength when p -0.05 or 0.03. Molar Absorp tivities of H-peroxidase-RCOXY Complexes-Molar absorptivities of H-peroxidase-RCOXY complexes were evaluated in different buffers at 25" with saturating amounts of the titrant.
Whenever this method was not applicable, either due to weak ligand-enzyme affinity (benzamide) or to limited titrant solubility (a-hydroxyacetophenone, cyclohexyl hydroxamic acid), the change in molar absorptivities (Ae,) was derived from the linearized form (Equation 1) of the standard equilibrium relationship (Equation 2). This procedure is permissible when the concentration of the enzyme-ligand complex (HRP,L) is relatively small compared to that of the ligand (AT) and when the enzyme-ligand association occurs in 1: 1 ratio.
(1) ). In the above equations, li, is the acid dissociation constant of the ligand; HRPT is the total concentration of all ionized forms of H-peroxidase; and Ailk represents t.he observed absorbance change, at. X am, when the solution light path is 1. From the plot of l/AA verBus l/L,, lil and AQ may be evaluated.
Specfrophoton2etric Ueasuremenfs-These measurements were carried out by using a Cary 14 recording spcctrophotometer equipped with a Universal transmission slidewire.
Hence, very small absorbance changes could bc accurnt.cly evaluated.
Magnetic Susceptibility-The experiment's were performed with an instrument designed by Tasaki et al. (29). Water was used as ca.libration st.andard. The measurements between 0 and -196' were first made by using H-peroxidase itself ( Fig. 4A) and then in t'he presence of equimolar amounts of benzhydroxamic acid (Fig. 4B). The sample volumes were the same in both cases (0.7 ml). 9 nearly complete conversion (-98%) of the enzyme into its benzhydroxamic acid complex was achieved by addit.ion of CBH&ONHOH (0.36 mg). Potentiometry-The pH values of the reaction media at 22" were measured by using a Radiometer model 25 pH meter equipped with a Radiometer t,ypc GK 2021 C combined calomel glass electrode.

RESULTS
Op/icnl nn.d Jlag?zetic Properties of II-peroxidase-RCONII Y Complexes (where Y = OH, NIIz, If)-The absorbance of horseradish ferriperoxidase between 220 and 1300 nm is drastically alt,ered on ligation t.o RCONHY. Fig. 1 illustrates this effect for benzlq-drosamic acid (R = phenyl, Y = OH), and a similar pattern is seen with hydrazides (Y = NIIZ) and amides (Y = H). In all cases, there is a typical hypcrchromic effect upon the Soret band, coupled to its shift from 403 nm in H-peroxidase to approximately 408 nm in the RCONHY derivatives.
In the visible absorpt,ion region, t,he complexes show bands at 503 to 505 nm and 637 to 639 nm, which may be compared to those of the  (Table VIII).
This distinct absorption pattern remains essentially unaltered from -196 to 40" and between pH 3 and 9.
In the ultraviolet region, the spectroscopic differences become less striking (Fig. 2) ; between 220 and 245 nm, the observed absorbance is nearly equivalent to the sum of absorba.nces contributed by the enzyme and its ligand.
The EPR spectra of H-peroxidase and its benzhydroxamic acid complex point to changes in the "heme-linked" interactions. Thus, at 77 and 4.2"K the unliganded enzyme absorbs near g = 6.2 and g = 5.0, respectively (Fig. 3, a and b). On the other hand, at 77°K the absorption derivative of the complex suggests coalescence of transitions giving, near g = 5.95, an intense but rather broad band (-220 Oe). The latter is resolved, 2 Spectroscopically analogous complexes are formed with wheat germ peroxidase but not with cytochrome c peroxidase, lactoperoxidase, chloroperoxidase, sperm whale metmyoglobin, or horse blood catalase. FIG. 3. EPR spectra of H-peroxidase (a and 5) and its benzhydroxamic acid complex (c and d) at 77 and 4.2'K; 0.6 mM H-peroxidase in 0.1 M potassium phosphate buffer (pH 6.8). EPR spectra were measured, under the same conditions, before and after addition of an equimolar amount of benzhydroxamic acid. Microwave frequency, 9.02 GHz; modulation amplitude, ~8 Oe. at 4.2"K, into a doublet with g -6.1 and g -5.4. The difference between the splitting factors is substantially smaller in the complex (Ag = 0.7) than in the free enzyme (Ag = 1.2) and implies a transition to a higher axial symmetry of the coenzyme (30) or reflects a change in the spin relaxations.
Magnetic susceptibility measurements complement the above data ( Fig. 4). For H-peroxidase, the effective magnetic moment (nerf) is 5.23 Bohr magnetons, which is in excellent agreement with the results of Theorell and Ehrenberg (31). In the presence of benzhydroxamic acid, the paramagnetism increases, giving n,fr of 5.97 Bohr magnetons. This is effectively the same as the theoretically computed value for an ion with five unpaired electrons (n,tf = 5.92 Bohr magnetons).
The magnetic susceptibilities of H-peroxidase and its benzhydroxamic acid complex follow Curie's Law, between -196 and 0'. Apparently, in contrast to several other hemoproteins (29,(32)(33)(34), both free and liganded forms of the H-peroxidase retain their high spin character over a wide temperature range.
Dissociation .1 X 1OW M H-peroxidase with benzhydroxamic acid; 0.02 %I pobassium phosphate buffer (pH 6.0) at 25". The apparent dissociation constants (Kl) were computed from the data at indicated points. Znsel, log-log plot, based on Equation 2; the slope of the line is 1.
The assay of the enzyme-RCOXY complexes was based on the spectroscopic measurements described in detail under "Experimental Procedure." In all cases, the equilibrium was established within the mixing time of the reagents.
Between pH 4 and 9 at 25", the complexes are stable and no secondary derivatives were detectable.
At all levels of enzyme saturation, the spectra pass through a single set of isosbestic points.
A typical experiment. for the measurement of benzhydroxamic acid binding is illustrated in Fig. 5. The data and the slope of 1, derived from the logarithmic plot of Equation 2, establish that the reactants combine in 1: 1 ratio. The same holds for Ohe other complexes listed in Table I.
In evaluating Zil, it is essential to take into account the acidbase properties of t'he ligands (2-4) (Equations 3 and 4).
RCONHOH + RCONHO-+ H+ (pK, = 8.8) (4) 505 For any given compound then, K1 calculated on the basis of the neutral form is virtually invariant from pH 3.5 to 9.0 (Table II). Amides, which do not ionize or protonate between pH 1 and 14, associate with the enzyme equally well in acid and alkaline solutions (Table II), providing that the conversion of neutral Hperoxidase into its alkaline derivative (characterized by pK, -10.9 (35)) is taken into account. This is expressed in Equation The intercept on the ordinate gives pK, -10.95, in excellent agreement with the value obtained through direct titration of the enzyme (35).
Three conclusions are allowed by the above data.
(a) The heme-linked effect leading to the formation of alkaline peroxidase is governed by a single ionization; (b) below pH 12, H-peroxidasebenzamide complex is not subject to such an ionization; and (c) if benzamide binds to the alkaline peroxidase, then its affinity is much weaker than that for the neutral enzyme.
Interaction of H-peroxidase with Cyanide in the Presence of Hydroxamic Acids and Hydrazides-If the enzyme-RCOXY interaction involves a direct coordination of -COXY to the ferric ion, such a binding should be competitively inhibited by typical hemoprotein ligands such as cyanide, fluoride, or aside. Of these, cyanide is the reagent of choice. It reacts readily with H-peroxidase (36, 37), giving a complex whose dissociation constant is nearly pH invariant from pH 4.2 to 7.5 (37) and whose spectrum (38, 39) differs radically from those of H-peroxidase-RCOXY derivatives (Fig. 1).
The results shown in Fig. 7a do not accord with this analysis. Instead, we note a progressive increase in t.he partition constant, KI/K3, with increasing concentration of benzhydroxamic acid (Table III) (12) Hy using Equation 12 and the data shown in Fig. 7b, Kq was found to be 1.5 =t 0.1 X 10W4 M and K1:K3 = 1.11 f 0.05, in agreement with the espccted K1 : Ks value.
The results of similar analyses, using l-and 2-naphthohydroxamic acid and benzhydrazide are summarized in Table IV. In all cases, the observed partition constant, K1:K3, closely agrees with Kl:Ksccalc) by using RI and Ks constants obtained in independent studies. Furthermore, according to Equation 11, m should converge to a limiting value with increasing ligand (I&) concentration and become equivalent to (~/HRPT) X (K1/K3) x (l/K,), when l/L, < l/Kd. Under such conditions, the extent but not the rate of formation of H-peroxidase-cyanide complexes (5 E. Cx + E. CN L) should be nearly independent of LT. And such is t,he case ( Table V).
The data presented in Table V indicates that the initial velocity (v;) of format,ion of the enzyme-cyanide complexes (E . CN -+-E. CN . BZH) is proportional to the concentrations of cyanide (HCNT) a.nd free enzyme (Ef).
The simplest scheme meeting  (Table V)  cm-l, it is evident that the increase induced by RCOXY compounds is relatively small. Nonetheless, this is adequate for an accurate measurement of RCOXY binding to the enzymecyanide complex.
Accordingly, from the data presented in Fig. 8a, we obtain K( = 4.4 x 10W4 M in good agreement with the value derived indirectly (cf. Table IV).
Interaction of H-peroxidase with Hydrogen Donors-Although phenols and aromatic amines have long been recognized as the specific peroxidase substrates (41), the association constants governing their interaction with the enzyme are unknown. Such parameters can be now readily calculated, exploiting the inhibition of RCOXY binding in the presence of hydrogen donor substrates. This is illustrated in Fig. 9A. From these results, we can derive the dissociation constants (KS) for H-peroxidasesubstrate (HRP.8) complexes, assuming that binding of the specific substrates (8) X HRP.L from which K, may be obtained.
Plots based on Equation 14 are shown in Fig. 9B for benzhydrazide, both in the absence of substrates and in the presence of 2.3 mM hydroquinone, 5.2 mM aniline, and 9.1 mM phenol.
In all cases, the lines intersect the abscissa at a point corresponding to the total enzyme concentration. Also, as would be expected for competitive binding, the calculated dissociation constants are independent of the initial hydrogen donor concentration or the nature of the titrant (Table VI).
Apparently peroxidases-The absorbance pattern of Mn(II1) protoporphyrin apoperoxidase is altered on formation of the enzyme-hydroxamic acid derivatives; the greatest differences being associated with r-z* transitions of the porphyrin macrocycle (Table VII). The dissociation constants, K1, derived in this manner are 2.1 f 0.2 X 10e6 M for the benzhydroxamic acid complex and 1.6 -or 0.2 X 1O-7 M for the 2-naphthohydrosamic acid derivative. For the corresponding complexes with the ferric enzyme, the dissociation constants are 2.4 X lop6 M and 2 X 10e7 M. Clearly, the affinities of peroxidases for the aromatic hydroxamic acids are not governed by the identity of the central metal ion. This is by no means typical of Mn(III)-and Fe(III)-enzyme reactivities (21,42).3 On the other hand, RCOXY compounds do not bind equally well to the reduced enzyme and, as shown in Table VII, do not  elicit large spectroscopic changes. Such weak heme-linked effects are analogous to those observed with peroxidase-phenol or peroxidase-cyanide-RCOXY complexes (Figs. 8 and 10) and indicate that the -COXY residue does not coordinate to the metal ion. The dissociation constant of ferroperoxidase-benshydroxamic acid complex is approximately K1 = 3 f 2 x 1OW M at pH 7.1 (in 5 mM phosphate) and 25"; i.e. it is of the same order of magnitude as Kd for the ferriperoxidase-cyanidebenzhydroxamic acid complex (Table IV). The uncertainty in K1 is largely due to the autoxidizability of the ferrous enzyme. Thus, the spectrophotometric measurements used in these assays are subject to errors which, so far, we have not been able to eliminate.

DISCUSSION
The RCOXY ligands may be divided into two categories.
In one subclass we have hydroxamic acids, hydraaides, amides, and cu-hydroxyacetophenone, all of which associate with H-peroxidase to give spectroscopically distinct derivatives.
The second group includes IV-and O-substituted hydrosamic acids, 0-benzoyl hydroxylamine, 4 N -hydroxybenzene sulfonamide, phenacyl halides, or benzaldehydc which, like phenols and aromatic amines, appear to bind to the enzyme but without markedly changing its characteristic spectrum.
In some cases, although not with a Unlike those for RCOXY derivatives, the dissociation constants for manganic and ferric fluoride complexes differ approximately 40-fold, the manganie enzyme showing a weaker affinity. 4 0-Benzoyl hydroxylamine reacts slowly with H-peroxidase forming, in the initial stages of the reaction, compounds which are spectroscopically similar to the enzyme-peroxide derivatives. Benzhydroxamic acid Benzhydrazide Benzhydrazide Benzhydrazide Benzhydrazide 6.8 6.8 6.7 6.7 6.7 Phenol Phenol Aniline Mesidine Hydroquinone evaluation of Kphenor by analysis of the difference spectrum between H-peroxidase and its phenol complex is shown in Fig. 1Oa. From the observed hyperbolic increase in absorbance with increasing concentration of phenol, Knhenol was computed to be 5.0 =I= 0.5 X lop3 M (Fig. 10~). Furthermore, the results in Fig.  lob demonstrate that phenol associates nearly as well with ferriperoxidase-cyanide complex. Thus, gauged by phenol binding the conversion of high spin ferric enzyme into a low spin cyanide derivative does not entail a pronounced change in the structure of the apoenzyme.  (44)) whereas its ligation to the enzyme is fast (ti/* < 10 s, when LT = 3 mM); (ii) imidic acids, enol forms of amides, are unknown (45) ; and (iii) in the 220 to 250 nm range the difference spectrum of the enzyme-"benzhydroxamic" acid complex and H-peroxidase (Fig. 2) is superimposable on the absolute spectrum of the ligand.
Contributions from the enol tautomer, benzhydroximic acid, would be expected to show an altered absorption pattern (cf. the differences in the spectra of 0-alkyl benzhydroxamic and benzhydroximic acids (46)). benzaldehyde and O-benzoyl hydroxylamine, this behavior may be due to inherent steric demands of a given substitutent and may also reflect conformational isomerism, dictated by the repulsion of nonbonding electrons on X-Y and the a-electrons of the carbonyl group (7,43).
In this context, the following points deserve emphasis: (a) Ionized ligands (benzhydrosamate anion, benzhydrazide cation) do not associate with H-peroxidase (Table II). This is consistent with, but does not establish, the nonpolar nature of the -COXY binding site. (b) The carbonyl group is not the sole determinant in the formation of spectroscopically active enzyme complexes. For example, benzaldehyde is an inert titrant.
(c) When X = NH, Y can be -OH, -NH2, or -H; and when Y = OH, X can be -NH or -CH2. Hence these groups per se cannot play a dominant role in coordination.
We infer that spectroscopically distinct complexes are only formed when the ligands contain an acidic residue (X), in a formally uncharged -C-X-Y constellation. This underscores II 0 the import,ance of the functional group as a whole rather than the individual components.
Mechanistically, this may imply either chelation of the hemin-iron or polyfunctional H-bonding. The chelation hypothesis is less likely on at least two counts. First, it is hardly plausible for amides; second, it is difficult to reconcile with the kinetics of RCOXY binding.
For instance, with benzhydroxamic acid, the apparent second order rate constant (k,,, -0.4 f 0.1 X lo* M-~ s-l at 25')s is greater than for any other H-peroxidase-ligand reaction. Polyfunctional hydrogen bonding, resembling the hydrogenbonded interactions of ureido groups with proteins (47, 48), is not open to these objections.
It is supported by a correlation between peroxidase-ligand affinity (expressed by K i) and acid dissociation constant of the ligands, K, (Fig. II).
Such relationships (49-52) rest on an intrinsic property of the hydrogen bond, relating its strength to the tendency of proton transfer (expressed by K,) from the donor-acid to the acceptor-base.
Hydrogen bonding, particularly of the monofunctional type,  If HRP.L' is the only enzyme-ligand derivative, then ligation of benzamide, benehydrazide, and benzhydroxamic acid are defined by A6410 values ranging from 28 x lo3 to 60 X lo3 M-I cm-1 (Table VIII), a gradation of perturbation effects dependent on the -COXY structure. Alternatively, if a spectroscopically inert HRP .L complex precedes the formation of HRP.L', then Ae will be contingent on K". When K" is Iargc, only small spectroscopic changes may be expected.
K' is then analogous to Krq, as defined for ternary enzyme-ligand complexes.
Such comparison strengthens the proposal that binding of the R group of RCOXY is the main driving force in the ligand-enzyme association.
It is this type of interaction, the hydrophobic binding, which must be responsible for the higher affinity of 2-naphthohydroxamic acid (K1 -2 x low7 M), compared to that of acetohydroxamic acid (K1 -620,000 x lo-' M) and formhydroxamic acid (K1 -0.3 M). The change in free energy of binding from 2-naphtho-t.o formhydroxamic acid amounts to -8 kcal per molt, i.e. within the range of binding energies (-5 to -10 kcal per mole) observed on association of nonpolar compounds with proteins (56, 57).
A plausible hypothesis is that both R and -CO-X-Y residues R-CNHY L ---Fe-L' e L--Fe--L'..MN associate at a heme-linked, extended, largely apolar region of H-peroxidase, in which the R-crevice is also the binding site for the hydrogen donor substrates.
The association of the -COXY moiety near the prosthetic group with a concomitant change in the relative position of trans.ligands, the central metal ion, and porphyrin would give a configuration more akin to that of metmyoglobin, where the porphyrin ring is not. planar but slightly concave towards the sixth coordination position (58). The lower rhombicity of the coenzyme (Fig. 3) and the higher paramagnetism of the H-peroxidase-benzhydroxamic acid complex (Fig. 4) and its metmyoglobill-like optical spectrum (Fig. 1) support this argument.
The inhibition of alkaline peroxidase format,ion by benzamide (Fig. 6) also accords with the proposed model, since both the heme-linked ionization of the Fe--L' . . 9 HX system and ligand interchange reactions are impaired in the ferriperosidase-benzamide complex.
Furthermore, the similar affinities of mangarlic and ferric peroxidases for hydroxamic acids cont,rast with those for fluoride and suggest again that the formation of RCOKHOH-enzyme complexes does not entail replacement of a ligand in the first coordination sphere of the metal ion. Rather, as outlined previously, CO-X-Y exerts its effect indirectly, through hydrogen bonding, most likely to water coordinated at the distal site of the prosthetic group (L'H=H*O). Such hydrogen bonding should be of lesser importance when cyanide occupies the distal site since the nitrile nitrogen is only a very weak proton acceptor (59). Indeed, the differences in affinities of ferriperoxida.se-cyanide for benzhydroxamic acid and benzhydrazide (Table IV) are much smaller than those for the corresponding complexes with ferriperoxidase itself (Table I). Similarly, if in ferroperoxidasc the sist,h coordination site of the prosthetic group is not occupied by water, as in ferromyoglobin (60), the dissociation constants for ferroperoxidase-RCOXY complexes should parallel lid rather than K1 values. The preliminary data on ferroperosidnse-benzhydrosamic acid complex, where K = 3 + 2 X 10-4~ supports this proposal.
These results do not preclude some int,eraction between the metal ion and the COXY moiety.
For thr observed shift of the ferroperoxidase Soret band from 437 to s-428 nm, which is caused by an amino reagent, benzhydrazide, but not by benzhydroxamic acid (Table VII), may indicate a, weak contribution from a hemochromogen type of complex. However, its formation cannot be extensive since the changes in the Soret region are not reflected in the pattern of cy-and P-bands.
The strongest evidence suggesting the proximity of the RCOXY site to the prosthetic group is the fully competitive binding between RCOXY ligands and t,he peroxidase H-donors, i.e. phenols and aromatic amines ( Asakura for their collaboration in EPR and magnetic susceptibility measurements, and to Mrs. L. Gomez-Rao for her excellent 31. assistance.
Thanks are due to Dr. P. Nicholls for helpful dis-32. cussions.