In Crystallo Posttranslational Modification Within a MauG/Pre–Methylamine Dehydrogenase Complex

See allHide authors and affiliations

Science  12 Mar 2010:
Vol. 327, Issue 5971, pp. 1392-1394
DOI: 10.1126/science.1182492


MauG is a diheme enzyme responsible for the posttranslational modification of two tryptophan residues to form the tryptophan tryptophylquinone (TTQ) cofactor of methylamine dehydrogenase (MADH). MauG converts preMADH, containing monohydroxylated βTrp57, to fully functional MADH by catalyzing the insertion of a second oxygen atom into the indole ring and covalently linking βTrp57 to βTrp108. We have solved the x-ray crystal structure of MauG complexed with preMADH to 2.1 angstroms. The c-type heme irons and the nascent TTQ site are separated by long distances over which electron transfer must occur to achieve catalysis. In addition, one of the hemes has an atypical His-Tyr axial ligation. The crystalline protein complex is catalytically competent; upon addition of hydrogen peroxide, MauG-dependent TTQ synthesis occurs.

Diversity of the cellular proteome is created by genetically encoded polypeptides and their subsequent chemical modification. Most posttranslational modifications (such as phosphorylation, glycosylation, and ubiquitination) occur at the protein surface, enabling direct contact between the processing enzyme and the site of modification. In contrast, protein cofactors essential for function often require in situ modification of amino acids buried within a protein. The formation of some of these is autocatalytic, such as the fluorophore in green fluorescent protein, but others require external enzymes (1). MauG is a diheme enzyme (2) that completes synthesis of the catalytic cofactor tryptophan tryptophylquinone (TTQ) (3) from two Trp residues in the β-polypeptide chain of methylamine dehydrogenase (MADH), a metabolic enzyme of methylotrophic and autotrophic bacteria (4). In MADH from Paracoccus denitrificans, two oxygen atoms are incorporated into the indole ring of βTrp57 and a covalent bond is formed between the indole rings of βTrp57 and βTrp108 (Scheme 1) (5).

Scheme 1

Overall reaction catalyzed by MauG.

The natural substrate for MauG is a 119-kD protein precursor of MADH (preMADH) with monohydroxylated βTrp57 and no cross-link (6, 7). PreMADH can be generated by expression of recombinant MADH in the background of a mauG deletion (6). MauG catalyzes a six-electron oxidation to complete TTQ biosynthesis (8). Oxidizing equivalents may be provided by three moles of either O2 (plus an electron donor) or H2O2 (8, 9). As such, the overall reaction can be viewed as three two-electron oxidations (of unknown order) to catalyze (i) insertion of an OH group at C6 of βTrp57, (ii) formation of the cross-link between βTrp57 and βTrp108, and (iii) oxidation of the quinol to the quinone. MauG contains two c-type hemes, the diferric form being the resting state. The hemes are covalently bound to the protein and have a His ligand to the iron. Typically, c-type hemes function as electron transfer mediators or catalytically in certain peroxidases, and indeed there is a sequence relationship between MauG and bacterial diheme cytochrome c peroxidases (diheme CCPs) (10). However, unlike diheme CCPs or other c-type hemes, MauG can activate molecular oxygen and forms an unprecedented diheme bis-Fe(IV) intermediate that is catalytically competent (11). This intermediate is composed of an Fe(IV)=O (ferryl) heme with the second oxidizing equivalent stored as Fe(IV) at the second heme. It is an intriguing alternative to Compound I, an Fe(IV)=O heme/porphyrin cation radical, which is the activated oxygen species required for cytochrome P450 and heme-dependent peroxidase–catalyzed chemistry (12, 13).

The x-ray crystal structure of the MauG-preMADH complex of P. denitrificans has been determined to 2.1 Å resolution (fig. S1 and table S1) (14). The asymmetric unit consists of two molecules of MauG and one of preMADH (Fig. 1A) forming a 203.6-kD complex. MADH is an α2β2 heterotetramer that contains two active sites and consequently two sites of posttranslational modification. MauG is monomeric and binds in an identical orientation to each αβ half of preMADH, with 2870 Å2 of surface area buried at each interface. The preMADH structure is essentially identical to mature MADH (5) except at βTrp57 and βTrp108, the residues that are modified by MauG. As predicted by 18O isotope labeling and mass spectrometry studies, βTrp57 of preMADH is monohydroxylated at the C7 carbon of the indole ring with no cross-link between the Trp residues (Fig. 2A) (7). The two Trps are in buried positions comparable to their positions in the mature enzyme. They have no direct contact to any part of MauG, the edge of the βTrp108 indole ring being closest to the interface (Fig. 1B). There is remarkable separation between the MauG hemes and the nascent TTQ site, with a distance of 40.1 Å between the most distant heme iron and βTrp108 (table S2).

Fig. 1

(A) Overall ribbon representation of the MauG-preMADH complex. (B) Spatial layout of potential redox groups. Color code: pink, MauG; blue, preMADH α; green, preMADH β. The hemes, Trp93, and Trp199 of MauG, as well as βTrp108 and monohydroxylated βTrp57 of preMADH, are drawn explicitly in stick representation. Figures were produced with PyMOL (

Fig. 2

Site of TTQ formation in MADH. (A) 2FobsFcalc electron density for the MauG-preMADH complex (resolution 2.1 Å). (B) The first 2FobsFcalc electron density calculated with MauG-preMADH + H2O2 structure factors (resolution 2.1 Å) and MauG-preMADH model phases with the preMADH βTrp57 and βTrp108 side chains omitted. Electron densities were contoured at 1σ. Carbon coloring: light green, preMADH; dark green, preMADH + H2O2.

The electron paramagnetic resonance (EPR) spectrum of diferric MauG identified two distinct c-type hemes: one high-spin and one low-spin (2). Addition of H2O2 to diferric MauG resulted in formation of the unusual diheme Fe(IV)=O/Fe(IV) reactive intermediate, as evidenced by Mössbauer and EPR spectroscopies (11). The Mössbauer spectrum of the intermediate was modeled as an Fe(IV)=O heme and a six-coordinate Fe(IV) heme. The crystal structure reveals that the MauG heme closest to preMADH is six-coordinate and exhibits a rare His-Tyr axial ligation (Fig. 3A). The axial Tyr294 ligand is likely responsible for the ability of this heme to stabilize Fe(IV) without requiring an exogenous ligand. Sequence alignments of known MauGs show that the Tyr is strictly conserved (fig. S2) (2). In contrast, the diheme CCPs, which do not stabilize the bis-Fe(IV) state, have either a Met or His in this position (fig. S2) (2, 10).

Fig. 3

MauG hemes. (A) Six-coordinate low-spin heme (HEC600). (B) Five-coordinate high-spin heme (HEC500). (C) Residues that line the distal pocket. 2FobsFcalc electron density was contoured at 1σ.

Spectroscopic data indicate that O2 binding and activation, or direct reaction with H2O2, occurs at the high-spin five-coordinate heme observed in the crystal structure (11). This is the heme farthest from preMADH in the complex, and thus demands a catalytic mechanism involving long-range interprotein electron and radical transfer (Fig. 1B). There is no evidence of solvent ligation at the coordination site trans to the proximal His35 at this heme (Fig. 3, B and C). It was previously shown that the initial two-electron oxidation of preMADH by MauG exhibits a random-binding kinetic mechanism in which the presence of preMADH neither stimulates nor impedes the reaction of MauG with H2O2, O2, or CO (15). The structure of the MauG-preMADH complex is consistent with that finding because it shows that the sites of H2O2/O2 and preMADH binding are well separated. This kinetic mechanism is in contrast to that of other enzymes that generate high-valent iron species as intermediates, such as the cytochrome P450 variants, which have ordered mechanisms (16). In the distal pocket of the five-coordinate high-spin heme, the ring of Pro107 is 4.3 Å from the heme iron (Fig. 3C). This is similar to the distance between the organic substrate and iron in ordered-mechanism oxygenases, such as cytochrome P450cam in complex with camphane (4.3 Å) (17). The rigidity of Pro107 may afford it the structural role of the substrate in an ordered mechanism by creating a stable O2 binding site and helping to keep the reactive oxygen species localized at the iron (16). As such, the distal pocket of the five-coordinate high-spin heme of MauG may be regarded as constitutively “on” with respect to its ability to activate O2.

The other characterized bis-Fe(IV) enzyme intermediate is the nonheme Fe(IV)2O2 (Intermediate Q) of methane monooxygenase (18). It has a diamond core structure with an Fe∙∙∙Fe interatomic distance of 2.46 Å. In contrast, the separation of the two heme irons in MauG is 21.1 Å. Despite this distance, the Fe(IV)=O/Fe(IV) intermediate forms very rapidly (>300 s−1) and has remarkable stability, decaying at a rate of 2 × 10−4 s−1 in the absence of preMADH (15). Trp93 of MauG is positioned between the two hemes such that the propionates of the high-spin and low-spin hemes are 3.3 Å and 3.8 Å, respectively, from the Trp indole ring (Fig. 1B). This Trp is conserved in both MauGs and diheme CCPs (2). It has been suggested to play a role in electron transfer between the hemes in the latter. Its position in MauG also suggests a role in mediating electron transfer between the hemes after H2O2/O2 binding and the subsequent generation of the bis-Fe(IV) species. MauG exhibits negative redox cooperativity between the c-type hemes in cycling between the diferric and diferrous forms upon successive one-electron redox events. The hemes have similar intrinsic oxidation-reduction midpoint potential (Em) values with facile electron transfer between them. Thus, although the determined Em values are distinct (−159 mV and −244 mV), MauG acts as a single diheme two-electron cofactor going through a valence-delocalized Fe(III)/Fe(II) state (19). In contrast, the Em values of the hemes of diheme CCPs are separated by more than 600 mV (20), and their reactive state is a mixed-valance Fe(III)/Fe(II) species (10). In comparing the structures of MauG and diheme CCP, the two hemes and intervening Trp overlay well (fig. S3). A Ca2+ ion is bound in an identical position in both enzymes (fig. S3B). Hence, a structural change altering the spatial relationship between the hemes is not responsible for the different redox and catalytic activities of the two enzymes. The key difference in the diheme unit of MauG is the distal Tyr294 ligand to the low-spin heme, which is a Met or His in diheme CCPs. Therefore, Tyr294 appears to be the major determinant in the redox properties and ability of MauG to stabilize the bis-Fe(IV) state.

The catalytic competence of the MauG-preMADH complex was examined in crystallo by solving the crystal structure after treatment with excess H2O2 (table S1) (14). The initial electron density (to 2.1 Å resolution) clearly showed that the second oxygen atom had been incorporated at C6 of βTrp57 and that a cross-link had been formed between βTrp57 and βTrp108 (Fig. 2B). It has not been possible to crystallize preMADH alone, so the catalytic requirement for MauG in H2O2-dependent TTQ synthesis was confirmed in solution under conditions that matched those of crystallization (fig. S4) (9). The structures before and after H2O2 addition are superimposable (root mean square deviation 0.158 Å), except for the posttranslational modifications. The TTQ cofactor of mature MADH purified from source (PDB code 2BBK) overlays well with that generated in the H2O2-treated MauG-preMADH crystals (5). Thus, the complex is catalytically competent and no major conformational rearrangements of the two proteins are required. At the center of the interface, halfway between the MauG low-spin heme and preMADH βTrp108, is MauG residue Trp199 (Fig. 1B). This residue may facilitate electron transfer across the interface through stabilization of a transient radical, although it is not conserved in other MauG sequences (fig. S2). The catalytic competence of the crystals also indicates a processive mechanism in which MauG and preMADH do not dissociate between each of the three two-electron oxidations. Furthermore, solvent must provide the second oxygen incorporated at βTrp57 C6, as this carbon is 43.7 Å from the Fe(IV)=O. The distal side of the high-spin heme of H2O2-treated crystals has residual electron density not present in the MauG-preMADH structure. The low occupancy and complexity of the electron density prevent it from being modeled, but it is consistent with a mixture of species that could include diatomics, such as hydroperoxo, and is evident in both copies in the asymmetric unit, thereby confirming that reaction with H2O2 occurs at this site (fig. S5).

This structure reveals that the purpose of the high-valent intermediate in MauG is to provide an oxidant with an extremely high reduction potential to extract electrons from the preMADH substrate, generating reactive radical intermediates that then acquire the oxygen atom from solvent and form the covalent cross-link between the Trp side chains. As such, the preMADH structure contributes substantially to the catalytic reaction through positioning the reacting portions of the Trps and creating the required chemical environment for TTQ synthesis. Thus, the posttranslational biosynthetic reaction requires both substrate-assisted and long-range catalysis.

Supporting Online Material

Materials and Methods

Figs. S1 to S5

Tables S1 and S2


References and Notes

  1. See supporting material on Science Online.
  2. Supported by NIH grants GM66569 (C.M.W.) and GM41574 (V.L.D.) and Minnesota Partnership for Biotechnology and Medical Genomics grant SPAP-05-0013-P-FY06. Computer resources were provided by the Basic Sciences Computing Laboratory of the University of Minnesota Supercomputing Institute, and we thank C. Ergenekan for his support. X-ray data were collected at the Kahlert Structural Biology Laboratory (KSBL) at the University of Minnesota and GM/CA-CAT at the Advanced Photon Source (APS), Argonne National Laboratory, Argonne, IL. GM/CA-CAT is funded by National Cancer Institute grant Y1-CO-1020 and National Institute of General Medical Sciences grant Y1-GM-1104. Use of the Advanced Photon Source was supported by the U.S. Department of Energy, Basic Energy Sciences, Office of Science, under contract DE-AC02-06CH11357. We thank E. Hoeffner for KSBL support and the staff at Sector 23, APS, for their support, especially M. Becker, S. Corcoran, V. Nagarajan, and D. Yoder. We thank S. Shin for performing the experiment shown in fig. S4. Coordinates and structure factors have been deposited in the Protein Data Bank with acquisition codes 3L4M (MauG-preMADH complex) and 3L4O (MauG-preMADH complex treated with H2O2).
View Abstract

Navigate This Article