Defining the involvement of HOCl or Cl2 as enzyme-generated intermediates in chloroperoxidase-catalyzed reactions.

Peroxidatic substrates, catechol (CAT) and 2,4,6-trimethylphenol (TMP) were used as probes of thechloride dependent reactions catalyzed by chloroperoxidase (CPO). TMP is consumed only in the presence of chloride. TMP is a competitive inhibitor versus CAT, but CAT is a noncompetitive inhibitor versus TMP in chloride-dependent CPO-catalyzed peroxidation reactions. The ratio of TMP versus CAT consumed by the chloride-dependent CPO reaction in direct competition studies increases as the chloride concentration is increased from 1.0 to 400 mM. Ratios of non-enzymatic HOCl reactions under conditions otherwise similar to those of the CPO reactions are relatively insensitive to changes in chloride concentration and are experimentally indistinguishable from the values attained by the enzyme system at high chloride concentrations. Comparison of enzymatic ratios with those of the HOCl reactions indicate that the proportion of the enzymatic reaction involving a freely dissociable, enzyme-generated, oxidized halogen species varies from 10% at low chloride concentrations to essentially 100% at high chloride concentrations. All data are consistent with a mechanism in which chloride competes with CAT for binding to both CPO compound I and the CPO chlorinating intermediate (EOCl). Chloride binding to CPO compound I leads to the formation of EOCl and initiates the CPO chloride-dependent pathway. When CAT binds to either compound I or EOCl, it is directly oxidized to product. When chloride binds to EOCl, it either induces release of HOCl or reacts with EOCl to produce Cl2, which is released from the enzyme. TMP and CAT compete for reaction with the free oxidized halogen species. This is the first direct evidence for kinetically significant involvement of a free oxidized halogen species as an intermediate in any CPO-catalyzed reaction.     

Compound I formation is a partially rate-limiting process in chloroperoxidase-catalyzed bromination reactions

The kinetics of chloroperoxidase-catalyzed bromination and chlorination reactions were studied at various halide and hydrogen peroxide concentrations. At very high concentrations, both chloride (KI = 370 mM) and bromide (KI = 150 mM) are competitive substrate inhibitors versus hydrogen peroxide. Results at subinhibitory halide concentrations for bromination reactions (kcat = 4 ms-1, kcat/KPeroxide = 1.6 microM-1 x s-1 and kcat/KBr = 4.0 microM-1 x s-1) and chlorination reactions (kcat = 1.5 ms-1, kcat/Kperoxide = 2.3 microM-1 x s-1, and kcat/KBr = 0.32 microM-1 x s-1) indicate that halide oxidation is rate-limiting in chlorination reactions. However, in bromination reactions, both compound I formation and bromide oxidation are partially rate-limiting. This is the first documented case where compound I formation participates in determining the overall rate of a peroxidase reaction.     

Compound X. An intermediate in enzymatic halogenation.

Previous studies have shown that chlorite serves as a halogenation substrate for horseradish peroxidase. In its substrate role, chlorite serves both as a halogen donor and as a source of oxidizing equivalents in the chlorination reaction. We now show that a new spectral intermediate, which we have termed Compound X, can be detected as the initial product of the reaction of chlorite with horseradish peroxidase. The reaction of chlorite with horseradish peroxidase to form Compound X is a relatively fast reaction especially at acidic pH values. The second order rate constant (Kf) for the formation of Compound X at pH 4.5 (optimum pH) is 0.9 X 10(6) M-1 S-1. Compound X, in the absence of a halogen acceptor, decomposes to Compound I and chloride ion. The first order rate constant (Kd) for the decay of Compound X to Compound I is 0.2 s-1 at pH 4.5. The pH optimum for enzymatic chlorination with chlorite compares favorably with the pH profile for the lifetime of Compound X (Kf/Kd). These observations indicate that Compound X is the halogenating intermediate in the chlorite reaction and that the rate of enzymatic chlorination is directly related to the stability of Compound X. We propose an -OCl ligand on a ferric heme as the most likely structure for Compound X.     

Crystal structures of chloroperoxidase with its bound substrates and complexed with formate, acetate, and nitrate

Chloroperoxidase (CPO) is a heme-thiolate enzyme that catalyzes hydrogen peroxide-dependent halogenation reactions. Structural data on substrate binding have not been available so far. CPO was therefore crystallized in the presence of iodide or bromide. One halide binding site was identified at the surface near a narrow channel that connects the surface with the heme. Two other halide binding sites were identified within and at the other end of this channel. Together, these sites suggest a pathway for access of halide anions to the active site. The structure of CPO complexed with its natural substrate cyclopentanedione was determined at a resolution of 1.8 A. This is the first example of a CPO structure with a bound organic substrate. In addition, structures of CPO bound with nitrate, acetate, and formate and of a ternary complex with dimethylsulfoxide (Me2SO) and cyanide were determined. These structures have implications for the mechanism of compound I formation. Before binding to the heme, the incoming hydrogen peroxide first interacts with Glu-183. The deprotonated Glu-183 abstracts a proton from hydrogen peroxide. The hydroperoxo-anion then binds at the heme, yielding compound 0. Glu-183 protonates the distal oxygen of compound 0, water is released, and compound I is formed.     

Determination of peroxidative halogenation in mixtures of chloride and bromide

A method for the differentiation of chlorinated and brominated products from peroxidative oxidation of mixtures of the halides is presented. Chlorination or bromination of monochlorodimedone (MCD) by fungal chloroperoxidase (CPO) was measured by loss of MCD absorbance. Although the Vmax was similar for both halides [approximately 0.08 mM (2 min)-1], the apparent Km for chlorination was 10 times greater than that for bromination (5.88 vs 0.67 mM). Chlorination was also quantitated as I3- produced from N-chlorotaurine and I-. The Vmax [0.076 mM (2 min)-1] and apparent Km (6.31 mM) determined by this method agreed with those determined with MCD. Selective reduction by H2O2 of the I-oxidizing potential of N-bromotaurine allowed determination of the brominated product from the difference between the amounts of halogenated MCD and N-chlorotaurine. The brominated product predominated at saturating and at physiologic halide levels. Hence, it is suggested that Br- plays a significant role in halogenation even though in vivo levels of Cl- are equal to or greater than 1000 times those Br-.     

The halide complexes of myeloperoxidase and the mechanism of the halogenation reactions

The spectral changes caused by the addition of halides to myeloperoxidase (donor:hydrogen-peroxide oxidoreductase, EC 1.11.1.7) have been investigated and the dissociation constants of the enzyme-halide complexes have been determined. The pH dependence of the dissociation constants suggests that halide binding is associated with a protonation step in myeloperoxidase. Myeloperoxidase catalyzes the peroxidative chlorination and bromination of monochlorodimedone. It is shown that at low pH, chloride acts as a competitive inhibitor with respect to H2O2, whereas at higher pH, H2O2 inhibits the chlorination reaction. The dissociation constant (Kd) of the spectroscopically detectable complex and the Km for chloride are considerably smaller than the inhibition constant (Ki) for chloride. These halogenation reactions are strongly pH dependent, the logarithm of the Km for chloride varies linearly with pH. The position of the pH optimum of the chlorination and bromination reaction is a linear function of the logarithm of the [halide]/[H2O2] ratio. A mechanism of the chlorination and bromination reaction is suggested with substrate inhibition for both hydrogen peroxide and the halide.     

Isomerization of (R)- and (S)-glutathiolactaldehydes by glyoxalase I: the case for dichotomous stereochemical behavior in a single active site.

The ability of glyoxalase I to isomerize both diastereomeric thiohemiacetals formed between glutathione and alpha-ketoaldehydes has been probed with stereochemically "locked" substrate analogues. Both (R)- and (S)-glutathiolactaldehyde (5 and 5') were unambiguously synthesized by employing the Sharpless epoxidation procedure as a key step. In the presence of human erythrocyte glyoxalase I, high-field 1H NMR analysis reveals that the R and S isomers (approximately 20 mM) are both converted to glutathiohydroxyacetone at rates of 0.8 and 0.4 s-1, respectively. This reaction is characterized by a nonstereospecific proton abstraction followed by a partially shielded proton transfer to the si face of the cis-enediol intermediate. Glyoxalase I catalyzes the exchange of the pro-S proton of glutathiohydroxyacetone with solvent deuterium. Glutathiohydroxyacetone was found to be a good competitive inhibitor of the normal glyoxalase I reaction (KI = 1.46 mM), suggesting that the slow processing rate of these compounds with respect to the normal thiohemiacetals is not due to poor binding. The results are consistent with a nonstereospecific proton abstraction and a stereospecific reprotonation at contiguous substrate carbons.     

Chlorinations catalyzed by chloroperoxidase occur via diffusible intermediate(s) and the reaction components play multiple roles in the overall process

The chlorination mechanism of the fungal enzyme chloroperoxidase (CPO) has been debated for (1) active site chlorination and (2) diffusible species mediated chlorination. Based upon the conversion of approximately 35 different substrates belonging to different reactive groups, it was found that substrate dimensions and topography had no pronounced effect on rates of CPO chlorination reaction. Epoxidation of indene was dependent on its concentration where as chlorination was not. Also, effective conversion was seen in the chlorination mixture for substrates that could not be epoxidized or sulfoxidized. Some insoluble substrates and certain molecules that exceeded the active site dimensions were chlorinated at rates comparable to the rates required for CPO's more natural substrate, monochlorodimedone. By terminating the enzymatic reaction with an active site ligand (azide), the amount of diffusible species was correlated to CPO in the reaction mixture. The preferential utilization of a substrate, earlier attributed to the active site, is found to be due to the specificity afforded by the reaction environment. It was found that the reaction medium components of peroxide, chloride and hydronium ions affected the reaction rates through varying roles in the enzymatic and non-enzymatic process. Besides these experimental evidences, key mechanistic and kinetic arguments are presented to infer that the final chlorine transfer occurs outside the active site via a diffusible species.     

The catalytic mechanism of Pseudomonas cytochrome c peroxidase

The catalytic mechanism of Pseudomonas cytochrome c peroxidase has been studied using rapid-scan spectrometry and stopped-flow measurements. The reaction of the totally ferric form of the enzyme with H₂O₂ was slow and the complex formed was inactive in the peroxidatic cycle, whereas partially reduced enzyme formed highly reactive intermediates with hydrogen peroxide. Rapid-scan spectrometry revealed two different spectral forms, one assignable to Compound I and the other to Compound II as found in the reaction cycle of other peroxidases. The formation of Compound I was rapid approaching that of diffusion control. The stoichiometry of the peroxidation reaction, deduced from the formation of oxidized electron donor, indicates that both the reduction of Compound I to Compound II and the conversion of Compound II to resting (partially reduced) enzyme are one-electron steps. It is concluded that the reaction mechanism generally accepted for peroxidases is applicable also to Pseudomonas cytochrome c peroxidase, the intramolecular source of one electron in Compound I formation, however, being reduced heme c.     

Atypical kinetic behavior of chloroperoxidase-mediated oxidative halogenation of polycyclic aromatic hydrocarbons

We have identified an atypical kinetic behavior for the oxidative halogenation of several polycyclic aromatic hydrocarbons (PAHs) by chloroperoxidase (CPO) from Caldariomyces fumago. This behavior resembles the capacity of some members of the P450 family to simultaneously recognize several substrate molecules at their active sites. Indeed, fluorometric studies showed that PAHs exist in solution as monomers and π-π dimers that interact to different extents with CPO. The dissociation constants of dimerization were evaluated for every single PAH by spectrofluorometry. Furthermore, docking studies also suggest that CPO might recognize either one or two substrate molecules in its active site. The atypical sigmoidal kinetic behavior of CPO in the oxidative halogenation of PAHs is explained in terms of different kinetic models for non-heteroatomic PAHs (naphthalene, anthracene and pyrene). The results suggest that the actual substrate for CPO in this study was the π-π dimer for all evaluated PAHs.     

Chlorination by a long-lived intermediate in the mechanism of flavin-dependent halogenases

The flavin-dependent halogenase RebH catalyzes the formation of 7-chlorotryptophan as the initial step in the biosynthesis of antitumor agent rebeccamycin. The reaction of FADH2, Cl-, and O2 in the active site generates the powerful oxidant HOCl, which was presumed to carry out the chlorination reaction. Herein, we demonstrate the formation of a long-lived chlorinating intermediate (t1/2 = 63 h at 4 degrees C) when RebH, FADH2, Cl-, and O2 react in the absence of substrate tryptophan. This intermediate remained on the enzyme after removal of FAD and transferred chlorine to tryptophan with kinetically competent rates. The identity of this intermediate is suggested by the X-ray crystal structure of RebH, which revealed an active site Lys79 located in a central position between flavin and tryptophan binding sites and just 4.1 A above C7 of tryptophan. The chlorinating species is proposed to be a Lys-epsilonNH-Cl (lysine chloramine) from reaction of enzyme-generated HOCl with the active site Lys79. This covalent enzyme chloramine likely plays a key role in directing regiospecific chlorination of substrate in this important class of biosynthetic enzymes.     

Chlorination of NADH: similarities of the HOCl-supported and chloroperoxidase-catalyzed reactions

The chloroperoxidase-catalyzed reactions of NAD(P)H with H2O2 in the presence of Cl- or Br- have been characterized. With 1 mol H2O2 per mol of NADH, one atom of 36Cl was incorporated into the 264-nm-absorbing intermediate product. This species was oxidized enzymatically by a second mole of H2O2 to a species distinct from NAD+, which retained one Cl atom. Spectroscopically identical species were also produced by reaction of NADH with one and two molar ratios of HOCl, respectively. These data indicate that, with respect to halogenation activities, chloroperoxidase functions similarly to myeloperoxidase, i.e., produces HOCl as the first product of Cl- oxidation by H2O2. Moreover, rapid chlorination of NAD(P)H followed by oxidation may be an important and highly lethal microbicidal effect of HOCl produced by myeloperoxidase in activated neutrophils.     

Catalysis of angiotensin I hydrolysis by human angiotensin-converting enzyme: effect of chloride and pH

The catalysis of the hydrolysis of angiotensin I, an important natural substrate, by human angiotensin-converting enzyme (ACE) was examined in detail as a function of chloride and hydrogen ion concentration. Chloride was found to be a nonessential activator over the pH range 5.0-10.0, with the chloride dependence increasing with increasing pH: the velocity enhancement at optimal [Cl-] increased from 1.6- to 42-fold; the chloride optimum and Ka' increased from 20 to 520 mM and from 0.22 to 120 mM, respectively, and activity in the absence of chloride decreased from 60.9 to 2.4% (relative to maximal activation). Kinetic analyses at pH 6.0, 7.5, and 9.0 confirmed the nonessential activator mechanism. At all pH values tested chloride was found to be inhibitory (relative to maximal activation) at supraoptimal chloride levels. Depending on the [Cl-] range, both apparent uncompetitive and competitive modes were demonstrated. From pH 6.0 to 9.0 Kis varied between 110 and 1140 mM (apparent). In all cases Ki' much greater than Ka'. We suggest that at high [Cl-] chloride binds to low-affinity inhibitory sites on the free enzyme and on the ES and EP complexes. The pH-rate profile demonstrated a chloride-dependent alkaline shift, with the pH optimum increasing from 7.1 at zero chloride to 7.6 at 400 mM NaCl. At [S] much greater than Km a plot of log nu vs pH revealed pKs of 5.9 and 9.4 in the ES complex in the absence of chloride, while at maximally activating [Cl-] only one ionization at pK = 6.3 was observed.(ABSTRACT TRUNCATED AT 250 WORDS)     

Interaction of Na+, K+, and Cl- with the binding of amphetamine, octopamine, and tyramine to the human dopamine transporter

Little information is available on the role of Na+, K+, and Cl- in the initial event of uptake of substrates by the dopamine transporter, i.e., the recognition step. In this study, substrate recognition was studied via the inhibition of binding of [3H]WIN 35,428 [2beta-carbomethoxy-3beta-(4-fluorophenyl)[3H]tropane], a cocaine analogue, to the human dopamine transporter in human embryonic kidney 293 cells. D-Amphetamine was the most potent inhibitor, followed by p-tyramine and, finally, dl-octopamine; respective affinities at 150 mM Na+ and 140 mM Cl- were 5.5, 26, and 220 microM. For each substrate, the decrease in the affinity with increasing [K+] could be fitted to a competitive model involving the same inhibitory cation site (site 1) overlapping with the substrate domain as reported by us previously for dopamine. K+ binds to this site with an apparent affinity, averaged across substrates, of 9, 24, 66, 99, and 134 mM at 2, 10, 60, 150, and 300 mM Na+, respectively. In general, increasing [Na+] attenuated the inhibitory effect of K+ in a manner that deviated from linearity, which could be modeled by a distal site for Na+, linked to site 1 by negative allosterism. The presence of Cl- did not affect the binding of K+ to site 1. Models assuming low binding of substrate in the absence of Na+ did not provide fits as good as models in which substrate binds in the absence of Na+ with appreciable affinity. The binding of dl-octopamine and p-tyramine was strongly inhibited by Na+, and stimulated by Cl- only at high [Na+] (300 mM), consonant with a stimulatory action of Cl- occurring through Na+ disinhibition.     

Chloroperoxidase, a janus enzyme

Chloroperoxidase is a versatile fungal heme-thiolate protein that catalyzes a variety of one-electron and two-electron oxidations. We report here that the alkylation of an essential histidine residue showed no effect on the one-electron peroxidations but inhibited two-electron oxidations. The pH profiles of different peroxidative substrates showed optimal activities at varying pH values for the same enzyme. 2-Allylphenol and substituted ortho-phenolics showed efficient peroxidations. Also, substrates excluded from the active site (or with no favorable positioning at the heme center or heme edge) were converted in the peroxidation reaction. While hydrogen peroxide serves as the superior activator in the two-electron oxidations, small alkylhydroperoxides give much better rates for peroxidation reactions. All the above observations indicate that one-electron oxidations are mechanistically quite different from the two-electron oxidations catalyzed by chloroperoxidase. We propose that the peroxidatic substrates interact predominantly outside the heme active site, presumably at the surface of the enzyme.     

Demonstration and chemical modification of a specific phosphate binding site in the phosphate-starvation-inducible outer membrane porin protein P of Pseudomonas aeruginosa

The interaction of phosphate ions with the Pseudomonas aeruginosa anion-specific protein P channel was probed. The single-channel conductance of protein P incorporated into planar lipid bilayer membranes in the presence of 0.3 M H2PO-4 was shown to be 6.0 pS, demonstrating that protein P channels allowed the permeation of phosphate. When large numbers of protein P channels were incorporated into lipid bilayer membranes in the presence of 40 mM Cl-, addition of small concentrations of phosphate resulted in reduction of macroscopic Cl- conductance in a dose- (and pH-) dependent fashion. This allowed calculation of an I50 value of e.g. 0.46 mM at pH 7.0, suggesting that the affinity of protein P for its normal substrate phosphate was at least 60-100-fold greater than the affinity of the channel for other ions such as chloride. Pyrophosphate and the phosphate analogue, arsenate, also inhibited macroscopic Cl- conductance through protein P with I50 values at pH 7.0 of 4.9 mM and 1.3 mM, respectively. To probe the nature of the phosphate binding site, the epsilon-amino groups of available lysine residues of protein P were chemically modified. Acetylation and carbamylation which produced uncharged, modified lysines destroyed both the anion (e.g. Cl-) binding site and the phosphate binding site as determined by single-channel experiments and macroscopic conductance inhibition experiments respectively. Nevertheless, the modified proteins still retained their trimeric configuration and their ability to reconstitute single channels in lipid bilayer membranes. Methylation, which allowed retention of the charge on the modified lysine residues, increased the Kd of the channel for Cl- 33-fold and the I50 for phosphate inhibition of macroscopic Cl- conductance 2.5-4-fold. A molecular model for the phosphate binding site of the protein P channel is presented.     

Regioselectivity of substrate hydroxylation versus halogenation by a nonheme iron(IV)–oxo complex: possibility of rearrangement pathways

Several nonheme iron enzymes and biomimetic model complexes catalyze a substrate halogenation reaction. Recent computational studies (Borowski et al. J Am Chem Soc 132:12887-12898, 2010) on α-ketoglutarate dependent halogenase proposed an initial isomerization reaction that is important to give halogenated products. We present here a series of density functional theory calculations on a biomimetic model complex-[Fe(IV)(O)(TPA)Cl](+), where TPA is tris(2-pyridylmethyl)amine-and investigate the mechanisms of substrate halogenation versus hydroxylation using the reactant and its isomer where the oxo and chloro groups have changed positions. We show here that the reactions occur on a dominant quintet spin state surface, although the reactants are in a triplet state. Despite the fact that the reactants can exist in two stable isomers with the oxo group either trans or cis to the axial ligand, they react differently with substrates, where one gives dominant hydroxylation and the other gives dominant chlorination of substrates. The ligand in the cis position of the oxo group is found to be active in the reaction mechanism and donated to the substrate during the reaction. A detailed thermochemical analysis of possible reaction mechanisms reveals that the strengths of the Fe-OH and Fe-Cl bonds in the radical intermediates are the key reasons for this regioselectivity switch of hydroxylation over halogenation. This study highlights the differences between enzymatic and biomimetic halogenases, where the former only react after an essential isomerization step, which is not necessary in model complexes.     

Formation of compound I by the reaction of catalase with peroxoacetic acid

1. The formation of Compound I by the reactions of bacterial and ox liver catalases with peroxoacetic acid was examined. In both cases the process occurs almost entirely by reaction of catalase with un-ionized peroxoacetic acid molecules. The result suggests an important role for the bound peroxidic proton in the enzyme-substrate interaction. 2. The peroxidatic properties of the Compounds I formed when peroxoacetic acid was used were examined by studying the oxidations of ethanol and formate; the results closely resemble those previously reported when H(2)O(2) and alkyl hydroperoxides were used. 3. Compound I formed with bacterial catalase and peroxoacetic acid is remarkably stable in the absence of added donor and the preparation has considerable potential for detailed studies of the nature of this intermediate.     

Hemes and hemoproteins. 5: Kinetics of the peroxidatic activity of microperoxidase-8: model for the peroxidase enzymes

The peroxidatic activity of the heme octapeptide from cytochrome c, microperoxidase-8 (MP-8), was assayed at 25 degrees C under conditions where formation of Compound I is rate limiting. In the pH range 6-9, the reaction rate increased linearly with a slope close to unity. The active form of the substrate is the hydroperoxide anion, HO2-, and an extrapolated second-order rate constant was obtained for the reaction of aquoMP-8 with HO2- of 3.7 X 10(8) M-1 sec-1, which is close to the second-order rate constants reported for reaction of the peroxidase enzymes with H2O2. Comparison with published data shows that the Fe3+ ion of MP-8 reacts as expected with simple anions, electrons, and HO2-, while the analogous reactions of the enzymes all show a requirement for one H+. We conclude that the peroxidase enzymes activate H2O2 under physiological conditions through a pH-independent, H+-coupled binding of the required H2O2-. The peroxidase activity of MP-8 can be increased more than tenfold by the presence of the guanidinium ion, which is ascribed to formation of the ion-pair GuaH+HO2-; this suggests a role for the invariant distal Arg in the enzymes.     

A study of the effect of ceruloplasmin and lactoferrin on the chlorination activity of leukocytic myeloperoxidase using the chemiluminescence method

The chlorination activity of free myeloperoxidase and myeloperoxidase bound with ceruloplasmin or with both ceruloplasmin and lactoferrin has been studied by luminal-dependent chemiluminescence. It was shown that the addition of hydrogen peroxide to the "myeloperoxidase + Cl- + luminal" system is accompanied by a fast flash of light emission. In the absence of myeloperoxidase or Cl-, the flash intensity was considerably reduced. The inhibitor of myeloperoxidase NaN3, the HOCl scavengers taurine and methionine, and guaiacol, a substrate for peroxidation cycle of myeloperoxidase, prevented luminescence. These results suggest that the generation of luminescence was due to the halogenating activity of myeloperoxidase, and hence, the flash light sum may serve as a measure of chlorination activity of myeloperoxidase. The activity of myeloperoxidase was suppressed by ceruloplasmin. Lactoferrin exhibited no significant influence on the myeloperoxidase activity, nor did it prevent the inhibitory effect of ceruloplasmin when they both were combined with myeloperoxidase. These data were confirmed using alternative approaches for evaluating the myeloperoxidase activity, namely, the assessment of peroxidation activity and the taurine chlorination assay. It is noteworthy that the inhibitory effect of ceruloplasmin on chlorination and peroxidation activities of myeloperoxidase is seen with the latter, traditional approaches only if ceruloplasmin is present in a large excess relative to myeloperoxidase, whereas the chemiluminescence method allows the detection of the inhibitory effect of ceruloplasmin using lower proportions of the protein with respect to myeloperoxidase, which are close to the stoichiometry of the myeloperoxidase/ceruloplasmin and the myeloperoxidase'ceruloplasmin'lactoferrin complexes.