If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* This work was supported, in whole or in part, by National Institutes of Health Grants T32GM075762 (a Chemistry-Biochemistry-Biology Interface Program Training Grant, which funded fellowships for T. A. W. and R. E. F.) and GM20877 (to D. P. B.). The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1–S13, Scheme S1, and Tables S1–S5.
Many siderophores used for the uptake and intracellular storage of essential iron contain hydroxamate chelating groups. Their biosyntheses are typically initiated by hydroxylation of the primary amine side chains of l-ornithine or l-lysine. This reaction is catalyzed by members of a widespread family of FAD-dependent monooxygenases. Here the kinetic mechanism for a representative family member has been extensively characterized by steady state and transient kinetic methods, using heterologously expressed N5-l-ornithine monooxygenase from the pathogenic fungus Aspergillus fumigatus. Spectroscopic data and kinetic analyses suggest a model in which a molecule of hydroxylatable substrate serves as an activator for the reaction of the reduced flavin and O2. The rate acceleration is only ∼5-fold, a mild effect of substrate on formation of the C4a-hydroperoxide that does not influence the overall rate of turnover. The effect is also observed with the bacterial ornithine monooxygenase PvdA. The C4a-hydroperoxide is stabilized in the absence of hydroxylatable substrate by the presence of bound NADP+ (t½ = 33 min, 25 °C, pH 8). NADP+ therefore is a likely regulator of O2 and substrate reactivity in the siderophore-associated monooxygenases. Aside from the activating effect of the hydroxylatable substrate, the siderophore-associated monooxygenases share a kinetic mechanism with the hepatic microsomal flavin monooxygenases and bacterial Baeyer-Villiger monooxygenases, with which they share only moderate sequence homology and from which they are distinguished by their acute substrate specificity. The remarkable specificity of the N5-l-ornithine monooxygenase-catalyzed reaction suggests added means of reaction control beyond those documented in related well characterized flavoenzymes.
). Central to iron trafficking in many microbes are siderophores, small molecular weight secondary metabolites that competitively chelate and repackage host iron for uptake and/or intracellular storage by the pathogen (
). The iron-ligating portions themselves, however, are far less diverse. Siderophores most often contain one of three bidentate chelating moieties: hydroxamate, catecholate, or α-hydroxycarboxylate.
Hydroxamate-dependent siderophores have special significance in pathogenic contexts. First, they are not bound by siderocalins, mammalian immunoproteins that selectively trap and thereby remove catecholate siderophores from circulating blood (
). Second, hydroxamate siderophores are acutely important to pathogenic fungi, which are not known to make catecholate or other types of chelating moieties. Many fungi, lacking ferritin, use variants of the same hydroxamate siderophore for both uptake and subsequent intracellular trafficking and storage functions (
). For viability and virulence, Aspergillus species, including deadly strains of Aspergillus fumigatus, have been shown to depend on functional pathways for biosynthesis of ferricrocin and fusarinine, siderophores involved in extracellular iron uptake and intracellular storage. Significantly, reductive pathways for mobilizing host iron, identified in some fungi, cannot compensate for the lost siderophore in many Aspergillus species (
Hydroxamate biosynthesis is therefore an attractive target for medicinally controlling A. fumigatus and possibly other pathogenic species. Antifungal targets are critically needed in light of recent sharp increases in hospital-associated infections, a paucity of effective treatments, and the consequent high mortality associated with invasive forms of Aspergillus (
). Many pathogens, for example, have multiple iron uptake pathways; if one is inhibited, another may or may not compensate. However, recent work demonstrated that Mycobacterium tuberculosis cultures are susceptible to picomolar levels of chemical inhibitors of MbtA, an enzyme catalyzing biosynthesis of the aryl cap of mycobactin siderophores (
). Like Aspergillus sp., mycobacteria have been predicted to depend on variants of the same siderophore for multiple steps in the uptake, internalization, and storage of iron and hence may be especially susceptible to chemical inhibition of mycobactin biosynthesis.
The biosynthesis of siderophores by A. fumigatus begins with the hydroxylation of l-ornithine (l-Orn),
which is the first committed step (Scheme 1). This reaction is catalyzed by l-ornithine monooxygenase (OMO), a flavin-adenine dinucleotide (FAD)-dependent enzyme. l-Lysine monooxygenase, the first described siderophore-associated monooxygenase (SMO), was initially identified by McDougal and Nielands (
The SMOs have only moderate global homology with the well characterized families of flavin-dependent monooxygenases, such as the hepatic microsomal flavin-containing monooxygenases (FMOs), nucleophilic Baeyer-Villiger monooxygenases (BVMOs), and aromatic hydroxylases (
Massey V. Williams C.H. Flavins and Flavoproteins: Proceedings of the Seventh International Symposium on Flavins and Flavoproteins, Ann Arbor, Michigan, June 21–26, 1981. Elsevier,
New York1982: 301-310
). These all have FAD- and NADPH-binding sequence motifs (associated with Rossmann folds) near the N termini and the centers of their sequences as well as a hydrophobic region near their C termini. Phylogenetic analysis places the SMOs nearest the FMO and BVMO families but within a novel subclass (
). However, among the flavin monooxygenases, the SMOs serve chemically distinct roles, being involved in biosynthetic rather than degradative processes. They almost certainly require unique ways of promoting highly substrate-specific hydroxylations. Early work on aerobactin biosynthesis established that SMOs have remarkable specificity for their substrates (e.g. they can discriminate between l-lysine and l-ornithine, molecules that differ only by one methylene unit) (
) led to the conclusion that PvdA might have mechanistic differences from the cohort of known FAD-dependent monooxygenases. This report provides the most detailed kinetic mechanistic studies, including both steady state and transient kinetic data, for any enzyme in the widespread SMO family. These data and information about the chemical mechanism that were derived spectroscopically allow us to suggest how this class of enzyme controls access of hydroxylatable substrates and O2 to the active site and to begin to devise inhibition strategies that may have medicinal value.
Flavin monooxygenases that catalyze the N-hydroxylation of the side chain of ornithine, lysine, or other primary amines are essential for the biosynthesis of hydroxamic acids. These acids act as bidentate chelating moieties in the siderophores used by hundreds of organisms (
). We consequently chose to examine the OMO enzyme from A. fumigatus as a representative of its class. The larger family of N-hydroxylating monooxygenases also includes enzymes involved in the biosynthesis of the plant hormone, auxin. Low overall sequence conservation between these enzymes and other flavin monooxygenases, their roles in biosynthetic rather than degradative processes, and their unusually acute substrate specificity suggest that they may be mechanistically or structurally distinct from either of the known families of flavin monooxygenases. We therefore undertook an extensive mechanistic study of OMO to define the place of NMOs in the larger scheme of flavin enzymes and to begin to address the feasibility of using the enzyme as an antifungal target.
All known flavin monooxygenases activate dioxygen to produce the substrate-hydroxylating species, a flavin C4a-hydroperoxide (or peroxide) adduct. The reactive species must then be directed toward the desired substrate and at the same time protected from releasing H2O2 to form oxidized flavin. The two major classes of well characterized flavin monooxygenases differ according to how they achieve the regulatory functions that avoid releasing H2O2. In the aromatic hydroxylases, of which para-hydroxybenzoate hydroxylase is the best studied example (
), regulation occurs at the level of flavin reduction. Rapid reduction occurs only following the binding and deprotonation of the substrate, an event that triggers a protein conformational change that brings FAD and NADPH into an optimal position for hydride transfer. Thus, unless a substrate is present, the flavin is not effectively reduced, and reaction with O2 cannot occur. Additional movements of the flavin enclose it and the substrate into a solvent-protected pocket, where the C4a-hydroperoxide can safely form and react with substrate. By contrast, the hepatic FMOs and bacterial BVMOs regulate their reactivity at the level of the C4a-hydroperoxide itself, which when formed is quite stable and releases H2O2 only very slowly in the protein's interior (
). Such a mechanism suits the biological role of the FMOs in hydroxylating and thereby increasing the aqueous solubility of xenobiotic compounds. Consistent with this role, these enzymes hydroxylate a broad variety of nitrogen- and sulfur-containing nucleophiles and even halides (
) and the FMO/BVMO enzymes, showed no substrate-dependent enhancement in the rate of FAD reduction, ruling out a possible substrate-gated regulatory mechanism of the kind identified in para-hydroxybenzoate hydroxylase. Furthermore, retention of NADP+ on the enzyme following reduction stabilizes the C4a-hydroperoxide intermediate in the hepatic FMOs and the BVMOs (
) and, as shown here, also in OMO. A quasi-stable hydroperoxide was observed only in the presence of NADP+ (Fig. 2). In the absence of hydroxylatable substrate, decomposition of this species to form oxidized FAD and H2O2 is rate-limiting and slow in all of these enzymes. Exposure of dithionite-reduced FMO (
) or OMO to air resulted in the rapid production of oxidized FAD without observable intermediates (supplemental Fig. S13). Consistent with retention of NADP+ on the enzyme throughout the oxidative half-reaction, NADP+ showed a pattern of pure competitive inhibition with NADPH, indicating that the NADP+-releasing and NADPH-binding enzyme forms are the same. The same pattern was observed by Poulsen and Ziegler (
Both steady state and transient studies of OMO also indicated important differences in the way it and the FMO or BVMO enzymes interact with their hydroxylatable substrates. Steady state kinetic data for OMO showed a pattern of substrate interactions suggesting ordered binding of NADPH, l-Orn, and O2. This is different from what was observed or would be predicted for either the FMO-like or aromatic hydroxylases and initially appeared to indicate a unique mechanism. Transient kinetic and product inhibition data subsequently shed light on these results. First, as in PvdA, it was noted that the rate of formation of the C4a-hydroperoxide is regulated by the presence of substrate (
). Studies of the reaction between E-FADred(NADP+) and O2 showed that the second order rate constant increased roughly 5-fold in the presence of saturating l-Orn. However, because formation of the C4a-hydroperoxide is not the rate-limiting step in catalysis (see Scheme 2), enhancing the rate of this step has no effect on the rate of turnover. It is therefore unclear whether the observed rate acceleration serves a biological role. The dependence of the rate constant for formation of the C4a-hydroperoxide on [l-Orn] was hyperbolic with an apparent Kd = 0.6 mm, very similar to the measured KI (0.5 mm) for l-Orn/l-Orn-OH inhibition (see below) as well as Km,l-Orn. (0.5 mm). This similarity suggests that each measured quantity is due to the same interaction between protein and l-Orn. In the case of Km, identity with the apparent Kd and KI could be observed if l-Orn achieves equilibrium with E-FADred(NADP+) rapidly. A rapid equilibrium was indeed identified via double mixing kinetic experiments.
Evidence for the interaction of a second enzyme form with l-Orn likewise came from both transient kinetics and product inhibition. The reaction of a preformed C4a-hydroperoxide with l-Orn resulted in generation of hydroxylated product, demonstrating that the intermediate formed in the absence of l-Orn is catalytically competent. Moreover, the intermediate reacted with l-Orn with identical kinetics whether l-Orn was present prior to C4a-hydroperoxide formation or was added to the preformed intermediate. In either case, the reaction demonstrated a saturable dependence on [l-Orn] (Kd(app) = 2.1 mm). The conversion of the C4a-hydroperoxide to the hydroxide and then to oxidized flavin occurred via a series of single exponential processes that were identical in each type of experiment. These observations, coupled with the fast off-rate for l-Orn from the FADred(NADP+) complex, suggest that with low concentrations of l-Orn present, the molecule that is hydroxylated most likely adds to the enzyme in its C4a-hydroperoxide form.
Further evidence for the interaction of l-Orn with both FADred(NADP+) and a second enzyme form came from product inhibition. The N5-hydroxyornithine product of the reaction was shown to act as a competitive inhibitor with l-Orn, indicating that the two vie for the same reduced, NADP+-bound enzyme form. By contrast, at higher but non-saturating concentrations of l-Orn, l-Orn-OH is uncompetitive with NADPH, suggesting that a second enzyme form can also bind l-Orn-OH. This form is presumably the C4a-hydroperoxide. The KI for l-Orn-OH and this enzyme form is 1.6 mm, very similar to the measured Kd(app) of 2.1 mm cited above, suggesting that these two constants indeed describe the interaction of the same enzyme form with l-Orn/l-Orn-OH. Notably, the affinity of the substrate/product for this form is somewhat less than for the reduced enzyme.
A simple analysis of the relative rates of the l-Orn-promoted and -unpromoted pathways for C4a-hydroperoxide formation suggests that the former pathway is preferred at [l-Orn] ≥ 140 μm, which includes all of the concentration range probed in this study. The ambient concentrations of l-Orn inside the cytosol of A. fumigatus are not known, although ornithine is produced exclusively in the mitochondrion and actively transported into the cytosol, where siderophore biosynthesis takes place. It is possible that concentrations in this range could be reached, particularly if ornithine transport occurred in response to iron stress. It is unclear whether the l-Orn binding sites on the FADred(NADP+) and C4a-hydroperoxide species are physically in the same or different locations. The kinetic and spectroscopic data are consistent with either two physically independent binding sites for l-Orn, one allosteric and the other catalytic, or a single binding site that has differing l-Orn affinities depending on whether the enzyme is in its FADred(NADP+) or C4a-hydroperoxide (NADP+) form. If the former description is true, then it is possible that other effectors that could stimulate C4a-hydroperoxide formation are available in the cell. If the latter is more accurate, then the reaction with l-Orn and O2 is best described as random. The observed dependences of both the rates of C4a-hydroperoxide formation and conversion to FAD on [l-Orn] are consistent with either description. Deciphering these two possibilities will be the subject of future work.
By the same token, the physical mechanism by which l-Orn accelerates C4a-hydroperoxide formation is unknown. It is possible that binding of the positively charged l-Orn could lower the barrier for the formation of the flavin semiquinone/superoxide radical pair (
). This would most likely require the regulatory and catalytic binding sites to be physically one and the same. It is also possible that l-Orn could dynamically induce a structural change that would promote access of the reduced flavin to O2 or that would bring the C4a position closer to an enzyme or NADP+-supplied positive charge. Interestingly, the turnover rates of FMOs have long been known to be accelerated by allosteric effectors, including octylamine (
). The nature of the effector interaction with substrate is unknown, but because it influences the turnover rate, it would need to have its effect on a rate-limiting step (i.e. either hydroxyflavin dehydration or NADP+ release at the end of each turnover). Studies of recombinant, chimeric forms of FMOs indicated that the substrate- and octylamine-binding portions of the FMO enzymes are probably remote from one another (
The ability of the C4a-hydroperoxide to form a measurable complex with its substrate is unusual. It has been demonstrated that the FMOs in their C4a-hydroperoxide forms do not reversibly form complexes with their hydroxylation substrates but instead react by a second order mechanism (
). For OMO, saturation in the plot of kobsversus [l-Orn], measured under pseudo-first order conditions for the reaction of the C4a-hydroperoxide and l-Orn, indicates that a complex does indeed form. Moreover, this enzyme form is subject to inhibition via the binding of l-Orn-OH. The difference in pKa for the amine substrate (∼10) and hydroxylamine product (∼5) indicates that the latter has a neutral side chain, whereas the former is positively charged at neutral pH. The fact that l-Orn-OH is able to serve as an inhibitor with a KI similar to the substrate Kd (for either l-Orn binding interaction) suggests either that the substrate is deprotonated in the active site or that the side chain charge is not relevant to the binding interaction. Future work will focus on understanding the interaction of substrate-like molecules with the two l-Orn-binding enzyme forms. Such an understanding is essential for exploring possible inhibition strategies as well as for understanding the remarkable specificity of OMO and its homologs for their substrates.
The OMO from A. fumigatus is a representative N-hydroxylating monooxygenase involved in siderophore biosynthesis. We have shown that it forms a quasi-stable C4a-hydroperoxide solely in the presence of bound NADP+. The formation of this species is modestly accelerated by the interaction of the reduced enzyme with a molecule of substrate. Although the effect is easily detected, it is probably insufficient to explain how OMO achieves its remarkable substrate selectivity or how it regulates the reaction with O2. The effect could be more pronounced in other NMOs. On the other hand, in contrast to FMOs, the C4a-hydroperoxide of OMO appears to bind l-Orn reversibly, potentially offering a means of screening for appropriate substrates.
We thank Prof. Marvin J. Miller for intellectual and material contributions toward the synthesis and characterization of l-Orn-OH. We thank Dr. Alexander Gehrke of the Hans Knöll Institute (Jena, Germany) for the kind gift of A. fumigatus from which genomic DNA was extracted. Amy Zercher gave technical assistance in cloning the sidA gene. We thank Garrett Moraski for technical assistance and for helpful discussions and Barrie Entsch (University of New England, New South Wales, Australia) for critical reading of the manuscript. NMR facilities were provided by the Lizzadro Magnetic Resonance Research Center at the University of Notre Dame, and mass spectrometry was provided by the University of Notre Dame Mass Spectrometry and Proteomics Facility (N. Sevova, Dr. W. Boggess, and Dr. M. V. Joyce; supported by National Science Foundation Grant CHE-0741793).
Massey V. Williams C.H. Flavins and Flavoproteins: Proceedings of the Seventh International Symposium on Flavins and Flavoproteins, Ann Arbor, Michigan, June 21–26, 1981. Elsevier,
New York1982: 301-310
Nature of the 4a-Flavinhydroperoxide of Microsomal Flavin-containing Monooxygenase (Flavoprotein, Hydroperoxide, Oxygen, Oxygenation, Flavinhydroxide). University of Michigan,
Ann Arbor, MI1985 (Ph.D. dissertation)