Cofactor Biosynthesis through Protein Post-Translational Modification

Abstract

Post-translational modifications of amino acids can be used to generate novel cofactors capable of chemistries inaccessible to conventional amino acid side chains. The biosynthesis of these sites often requires one or more enzyme or protein accessory factors, the functions of which are quite diverse and often difficult to isolate in cases where multiple enzymes are involved. Herein is described the current knowledge of the biosynthesis of urease and nitrile hydratase metal centers, pyrroloquinoline quinone, hypusine, and tryptophan tryptophylquinone cofactors along with the most recent work elucidating the functions of individual accessory factors in these systems. These examples showcase the breadth and diversity of this continually expanding field.

Introduction

The role of post-translational modifications as regulatory or cell localization strategies has long been recognized. However, the observation that amino acids can be modified to generate cofactors with novel functions is relatively recent [1]. In some cases, the biosynthesis of these cofactors is autocatalytic, requiring only the proper protein fold and perhaps a second cofactor such as a heme or metal ion to initiate amino acid modification and complete cofactor formation. A well known example of such autocatalytic synthesis comes from the copper amine oxidases, where generation of the topaquinone (TPQ) cofactor from a Tyr residue requires only copper and oxygen [2]. In other cases, one or more accessory proteins are required for cofactor maturation. This review discusses some interesting examples of such systems and the recent advances in understanding the enzymes which generate these protein-derived cofactors.

Modified Amino Acid Ligands to Metal Cofactors: Urease and Nitrile Hydratase

Urease and nitrile hydratase are metalloproteins with post-translationally modified amino acid ligands to the metal(s) at the active site (Figure 1). In urease, a dinuclear Ni center is coordinated by a bridging carbamylated Lys residue [3]. Active site maturation in urease requires typically four accessory proteins (UreDEFG) as well as Ni²⁺, CO₂ (derived from carbonate), and GTP. Despite intensive study over a number of years, the specific roles of these proteins in urease activation remain elusive. UreE is thought to function as a metallochaperone and UreG as a GTPase. GTP hydrolysis by UreG has been proposed to either cause a conformational change allowing access of Ni and CO₂ to the urease active site, or to generate carboxyphosphate as a CO₂ donor to the active site lysine residue. Less is known about the other accessory proteins. However, successful methods for the soluble expression of UreD (as a maltose binding protein fusion) [4] and a truncated UreF [5] were only recently achieved and may lead to significant advances in this field. The activation of Rubisco provides some interesting parallels to the urease system in that it also requires carbamylation of a lysine residue [6] in order to bind Mg²⁺ in the active site [7]. Here again an “activase” protein is involved, although it is not believed to participate directly in Lys carbamylation [8], which appears to occur spontaneously.

Figure 1.

Proposed pathways of cofactor biogenesis for Ni-urease, Co-nitrile hydratase, pyrroloquinoline quinone (PQQ), hypusine, and tryptophan tryptophylquinone (TTQ). Post-translational modifications made at each step are shown in red.

Low-molecular-mass nitrile hydratase (L-NHase) utilizes a mononuclear Co site coordinated by two oxidized Cys residues, one Cys-sulfenic acid (-SOH) and one Cys-sulfinic acid (-SO₂H) [9]. Expression of the structural genes for the L-NHase α₂β₂ heterotetramer (nhlAB), in the absence of the downstream activator gene nhlE resulted in a protein with very little activity, low Co content, and no Cys-sulfinic acid (-SO₂H) in the α-subunit [10]. Co-expression of nhlA with nhlE yields a trimeric complex (holo-αe₂) which contains Co and modified Cys residues in the α-subunit. It was further demonstrated in vitro that NhlE was responsible for Co insertion and Cys oxidation [11]. The holo-αe₂ complex is able to activate apo-α₂β₂ by a novel mechanism dubbed “self-subunit swapping” (Figure 2), where two holo-αe₂ complexes exchange α-subunits with apo-α₂β₂ forming active holo-α₂β₂ and apo-αe₂ [10,11]: this mechanism also holds true for the high-molecular-mass nitrile hydratase (H-NHase) [12]. The driving force for the exchange appears to be the formation of a salt bridge between two conserved Arg residues of the β-subunit and the negatively charged Cys-SO₂⁻ and Cys-SO⁻ residues of the holo-α-subunit [10]. NhlE was consequently designated as the first self-subunit swapping chaperone as well as a metallochaperone. However, the sequence of NhlE has no known metal binding motifs and metal coordination by NhlE has never been shown directly. Thus, the question of how NhlE facilitates Co insertion and Cys oxidation in the α-subunit remains unclear.

Figure 2.

Proposed mechanism of nitrile hydratase active site maturation by “self-subunit swapping”.

The Pyrroloquinoline Quinone Cofactor

Pyrroloquinoline quinone (PQQ) biosynthesis can be viewed as an unusual case of posttranslational modification that generates a stand-alone cofactor primarily utilized by bacterial alcohol and sugar dehydrogenases. The cofactor is synthesized from Tyr and Glu residues within a peptide encoded by pqqA (Figure 1). In Klebsiella pneumonia five additional genes (pqqB–F) are required beyond pqqA [13]. Of these gene products, PqqB is homologous to β-lactamases, and a crystal structure is available (PDB ID: 3JXP): it is thought to be a PQQ carrier, as some PQQ synthesis occurs in its absence [13]. The protein contains a conserved zinc-finger motif, but has undergone no in vitro functional studies at this time, and so the role of this and other conserved residues waits to be elucidated. The protein PqqF is thought to be a protease that may be involved in excising a cross-linked Tyr-Glu intermediate from the PqqA peptide.

Earlier work primarily focused on PqqC; the final enzyme in the pathway [14]. PqqC is a cofactorless oxidase and catalyzes the final ring closure and an 8-electron oxidation to generate PQQ. Using a combination of kinetics, site-directed mutagenesis and crystal structures, current studies have broadly begun to define the roles of individual amino acids, and an observed open and closed conformation [15]. Three equivalents of O₂ are required for a single turnover producing two equivalents of H₂O₂, indicating that one equivalent of H₂O₂ produced by the enzyme is used as an electron acceptor during PQQ synthesis [14]. The activation of O₂ in the absence of a metal or cofactor is of particular interest. Diatomic electron density was observed bound by a group of conserved amino acids [16] at a distance of 3.4 Å from a carbon of a trapped quinoid or quinol reaction intermediate [15].

The enzyme PqqD has engendered the most interest as it has no homology to any characterized protein, and its crystal structure was recently determined (PDB ID: 3G2B) [17]. Initially the focus was on its assumed interaction with PqqC, as it is found as a pqqCD fusion in some organisms [18]. However, recently, an interaction has been determined with PqqE [19], a radical S-adenosyl-L-methionine (SAM) enzyme containing two [4Fe4S]²⁺ clusters [20]. PqqE is highly sensitive to O₂, with catalytic SAM cleavage to generate the 5'-deoxyadenosine radical only occurring in PqqE that has been expressed and purified under anaerobic conditions [20]. The addition of PqqD to anaerobic PqqE in the presence of SAM leads to an increase in α-helical content compared to the individual proteins, and changes in the electronic environment of the [4Fe4S]²⁺ clusters indicating a specific interaction between the two proteins [19]. It has been proposed that PqqE may be the first enzyme in the pathway catalyzing C-C bond formation, but thus far no reactivity with PqqA has been reported. Perhaps one of the most intriguing mysteries is the incongruity that PqqD binds to both a strictly anaerobic PqqE and an O₂-dependent PqqC.

The Hypusine Residue of eIF5A

Eukaryotic translational initiation factor 5A (eIF5A) is conserved among eukaryotes and is involved in cell proliferation. Although initially described as a translation initiation factor, eIF5A is now considered to be primarily involved in translation elongation [21]. However, the transcripts which rely upon it for efficient translation and its exact role in the cell cycle have remained elusive. It has been established that eIF5A contains a unique and essential post-translational modification of a specific lysine residue to form hypusine in two enzymatic steps (Figure 1) [22, 23]. The first step involves the formation of deoxyhypusine from spermidine and the eIF5A precursor lysine and is catalyzed by deoxyhypusine synthase (DHS). The second step is catalyzed by deoxyhypusine hydroxylase (DOHH) and involves hydroxylation of the deoxyhypusine intermediate to give hypusine. DHS has been structurally characterized and the mechanism of deoxyhypusine formation is well understood. In contrast, little is known about DOHH. Homology modeling and mutational analysis suggest a structure with two arms separated by a flexible linker, with each arm composed of 4 consecutive α-hairpin repeats known as Huntingtin Elongation factor 3 protein phosphatase 2A TOR1 (HEAT) motifs: a motif consisting of two α-helices separated by a nonhelical region. Conserved His and Glu residues on each arm coordinate a bridging diiron center [22]. Consistent with this, native gel electrophoresis and SAXS show a more compact conformation for Fe-bound versus apo-DOHH [24]. Although many of the bacterial diiron monoxygenases also utilize a His/Glu coordinated diiron center [25], DOHH appears to be structurally and sequentially unrelated to these enzymes.

Recombinant expression and purification of human DOHH yields a blue solution. Extensive spectroscopic characterization including EPR, Mössbauer, resonance Raman, and X-ray absorption spectroscopy confirmed the identity of the blue species as a (μ-1,2-peroxo)diiron(III) complex (Figure 3) [26]. The complex can be chemically reduced by dithionite and regenerated by exposure to air. Most importantly, incubation of the peroxo species with deoxyhypusine-containing eIF5A leads to nearly stoichiometric formation of mature, hypusine-containing eIF5A.

Figure 3.

Proposed structure of the diiron-peroxo species of deoxyhypusine hydroxylase. The R reflects that the single atom bridging ligand is not expected to be an oxo bridge, but its exact identity is unknown.

DOHH is the first example of a eukaryotic hydroxylase capable of oxygen activation at a non-heme diiron site. Of particular interest is the remarkable stability of the peroxo species of DOHH, which persists for days whereas other diiron(III)-peroxo complexes decay with half-lives on the order of milliseconds to minutes [27]. EXAFS results suggest that the source of this enhanced stability may be a single atom bridging ligand between the Fe atoms [26]. Substrate addition to peroxo-DOHH enhances its decay rate and results in substrate hydroxylation, suggesting that substrate binding causes a conformational change at the active site, facilitating O-O bond cleavage and substrate oxidation. Tantalizingly, the slow rate of activity observed in vitro may imply that there are other accessory factors required to maximize the reaction rate.

Tryptophan Tryptophylquinone

Tryptophan tryptophylquinone (TTQ) is an in situ cofactor generated from two tryptophans within the polypeptide (Figure 1) [1,28]. It is found in the active site of the bacterial enzymes methylamine dehydrogenase (MADH) and aromatic amine dehydrogenase. The synthesis of TTQ has been studied in the Paracoccus denitrificans MADH system, and involves the βTrp57 and βTrp108 of the α₂β₂ MADH tetramer. Four gene products are required for maturation of MADH, and thus far only the final enzyme responsible for TTQ biosynthesis, MauG, has been studied. MauG is a di-heme c enzyme that catalyzes a 6-electron oxidation to complete formation of TTQ in an O₂ + reducing equivalents or H₂O₂-dependent reaction [1,28]. The MADH substrate for MauG (preMADH) has a partially assembled cofactor (preTTQ) consisting of a monohydroxylated βTrp57 at C7 of the indole ring. The enzyme responsible for the addition of this initial −OH group is currently unknown. The reaction consists of three MauG-dependent 2-electron events: (1) cross-link formation between βTrp57 and βTrp108, (2) hydroxylation of βTrp57 C6 to form quinol, and (3) oxidation of quinol to quinone. The order of these events remains to be determined. MauG catalysis is highly unusual for a number of reasons: (1) it contains a c-type heme that can bind and activate O₂, (2) the two hemes act in concert as a single redox cofactor, and (3) the catalytic MauG oxidant is a unique bis-Fe(IV) species consisting of ferryl heme (Fe(IV)=O) and the second heme as six-coordinate Fe(IV) [29,30].

The crystal structure of MauG in complex with preMADH (PDB ID: 3L4M), and the finding that the crystals were catalytically active, has been the cornerstone of recent work (Figure 4) [31]. This structure demonstrated that MauG did not need to dissociate between the three 2-electron events, and that no significant structural rearrangements were required. The fold of preMADH was essentially that of the mature enzyme, and the βTrp108 and monohydroxylated βTrp57 of preTTQ were at the same site as mature TTQ. Neither residue was in direct contact with either heme of MauG: the closest atom of preTTQ being 15.5 and 37.5 Å from each of the porphyrin macrocycles. Therefore, oxidation of preTTQ to TTQ involves long-range electron transfer. The kinetics of the first 2-electron cycle [32] and the MauG-dependent oxidation of dithionite-reduced MADH (quinol to quinone) [33] have recently been reported. One of the kinetic surprises was that the reaction was random order: the potent bis-Fe(IV) species could form in the absence of preMADH [29,32] and is extremely long-lived for one that has the potential to cause significant non-specific oxidative damage (decay to diferric MauG being 0.0002 s⁻¹) [32]. Indeed, cycles of formation and decay of the bis-Fe(IV) species in the absence of preMADH leads to suicide inactivation [34]. The site of H₂O₂ and O₂ binding was shown to be the heme furthest from preMADH [31,35,36]. The heme closest to preMADH has an unusual His/Tyr axial ligation required for bis-Fe(IV) formation [37]. MauG Trp199 at the interface with preMADH was shown to be critical for TTQ synthesis, although formation of the bis-Fe(IV) was unimpaired [38]. W199F and W199K MauG mutants did not support initiation of TTQ biosynthesis, implicating a MauG-dependent hopping mechanism that requires formation of a Trp199 radical.

Figure 4.

Position of hemes and residues relevant to TTQ synthesis in the crystal structure of MauG-preMADH. MauG is shown in pink cartoon, the α and β subunits of preMADH are shown in blue and green, respectively. β-preMADH Trp57 and Trp108 and MauG Trp199, Trp93, hemes and heme ligands are drawn in stick form. Figure produced using Pymol (http://www.pymol.org).

Concluding Remarks

The number and diversity of cofactors derived from post-translational modification of protein residues is considerable and continually expanding. Here we have touched upon a few examples that showcase recent advances in our understanding of their biosynthesis. As the field matures, new cofactors and systems will surely come to light, highlighting new ways in which nature has expanded the chemistry available to the 20 conventional amino acids.

Highlights

• Maturation of metal centers with modified amino acid ligands; urease and nitrile hydratase

• Biosynthesis of pyrroloquinoline quinone from a peptide precursor

• Spectroscopic characterization of an enzyme required for hypusine synthesis

• Long range electron transfer is required for tryptophan tryptophylquinone biosynthesis

ABBREVIATIONS

TPQ

topaquinone

Rubisco

Ribulose-1,5-bisphosphate carboxylase oxygenase

L-NHase

low-molecular-mass nitrile hydratase

H-NHase

high-molecular-mass nitrile hydratase

PQQ

pyrroloquinoline quinone

SAM

S-adenosyl methionine

eIF5A

eukaryotic translation initiation factor 5A

DHS

deoxyhypusine synthase

DOHH

deoxyhypusine hydroxylase

HEAT

(Huntingtin Elongation factor 3 protein phosphatase 2A TOR1) a motif consisting of two α-helices separated by a non-helical region

SAXS

small angle x-ray scattering

EPR

electron paramagnetic resonance

EXAFS

extended x-ray absorption fine structure

TTQ

tryptophan tryptophyl quinone

MADH

methylamine dehydrogenase

preMADH

biosynthetic precursor protein of MADH with incompletely synthesized TTQ (preTTQ)