PQQ and quinoprotein research--the first decade

On the occasion of the first international symposium on pyrroloquinoline quinone (PQQ) and quinoproteins (Delft, September 1988), a review of this novel field in enzymology is presented. Quinoproteins (PQQ-containing enzymes) are widespread, from bacteria to mammalian organisms (including man), and occur in several classes of enzymes. Indications already exist that PQQ is a versatile cofactor, involved not only in oxidation but also in hydroxylation, transamination, decarboxylation and hydration reactions. The current list of quinoproteins shows that it was overlooked in several well-studied enzymes where the presence of a common cofactor had already been established. Up until now, all eukaryotic quinoproteins have covalently bound PQQ (or perhaps pro-PQQ), while free PQQ occurs exclusively in a number of (bacterial) dehydrogenases and in the culture fluid of certain Gram-negative bacteria. Biosynthesis of free PQQ in methylotrophic bacteria starts with tyrosine and glutamic acid as precursors while intermediates in the route have not been detected and the presence of free PQQ is not required for synthesis of the covalently bound form of the cofactor in glutamic acid decarboxylase from Escherichia coli. Therefore, the assembly of covalently bound cofactor might occur in situ, i.e. in the quinoproteins themselves. If the latter also applies to mammalian quinoproteins, this implies that PQQ is not a vitamin. On the other hand, positive effects have been reported upon administration of PQQ to test animals. Methods suited to detach and to detect PQQ with a derivatized o-quinone moiety may answer questions on the uptake and processing of the compound.     

Quinoproteins: structure, function, and biotechnological applications

A new class of oxidoreductase containing an amino acid-derived o-quinone cofactor, of which the most typical is pyrroloquinoline quinone (PQQ), is called quinoproteins, and has been recognized as the third redox enzyme following pyridine nucleotide- and flavin-dependent dehydrogenases. Some quinoproteins include a heme c moiety in addition to the quinone cofactor in the molecule and are called quinohemoproteins. PQQ-containing quinoproteins and quinohemoproteins have a common structural basis, in which PQQ is deeply embedded in the center of the unique superbarrel structure. Increased evidence for the structure and function of quinoproteins has revealed their unique position within the redox enzymes with respect to catalytic and electron transfer properties, and also to physiological and energetic function. The peculiarities of the quinoproteins, together with their unique substrate specificity, have encouraged their biotechnological application in the fields of biosensing and bioconversion of useful compounds, and also to environmental treatment.     

Evidence for PQQ as cofactor in 3,4-dihydroxyphenylalanine (dopa) decarboxylase of pig kidney

Pig kidney 3,4-dihydroxyphenylalanine (dopa) decarboxylase (EC 4.1.1.28) was purified to homogeneity. Treatment of the enzyme with phenylhydrazine (PH) according to a procedure developed for analysis of quinoproteins gave products which were identified as the hydrazone of pyridoxal phosphate (PLP) and the C(5)-hydrazone of pyrroloquinoline quinone (PQQ). This method failed, however, in quantifying the amounts of cofactor. Direct hydrolysis of the enzyme by refluxing with hexanol and concentrated HCl led to detachment of PQQ from the protein in a quantity of 1 PQQ per enzyme molecule. In view of the reactivity of PQQ towards amines and amino acids, we postulate that it participates as a covalently bound cofactor in the catalytic cycle of the enzyme, in interplay with PLP. Since several other enzymes have been reported to show the atypical behaviour of dopa decarboxylase, it seems that the PLP-containing group of enzymes can be subdivided into pyridoxoproteins and pyridoxo-quinoproteins.     

Pyrroloquinoline quinone as cofactor in galactose oxidase (EC 1.1.3.9)

Galactose oxidase from Dactylium dendroides was shown to contain one molecule of covalently bound pyrroloquinoline quinone (PQQ/enzyme molecule. From the spectroscopic characteristics reported for the enzyme forms, a mechanistic role for PQQ could be deduced. In analogy with other quinoproteins, the initial formation of a PQQ-substrate adduct is proposed. Following internal hydrogen transfer, leading to aldehyde product and reduced pyrroloquinoline quinone, reoxidation of the organic cofactor with molecular oxygen could be mediated by the PQQ-liganded copper ion with concomitant formation of hydrogen peroxide. With PQQ as an additional (two-electron) redox center the occurrence of a "superoxidized" enzyme form must be considered. Possible consequences of this view, in relation to a physiological function of the enzyme and interpretation of ESR data, are discussed.     

Quinoproteins: enzymes containing the quinonoid cofactor pyrroloquinoline quinone, topaquinone or tryptophan-tryptophan quinone.

The presently best known and largest group of quinoproteins consists of enzymes using the cofactor 2,7,9-tricarboxy-1H-pyrrolo[2,3-f]quinoline- 4,5-dione (PQQ), a compound having a pyrrole ring fused to a quinoline ring with an o-quinone group in it. Representatives of this group are found among the bacterial, NAD(P)-independent, periplasmic dehydrogenases. Despite their high midpoint redox potential, the overall behaviour of quinoprotein dehydrogenases is similar to that of their counterparts, those using a flavin cofactor or a nicotinamide coenzyme. Apart from an exceptional Gram-positive one, the sole organisms where the presence of PQQ has really been established are Gram-negative bacteria. Evidence for the occurrence of covalently bound PQQ is lacking since it has now been shown that several enzymes previously considered to contain this prosthetic group do not in fact do so. Another group of quinoproteins, consisting of amine oxidoreductases, has a protein chain containing one of the following quinonoid aromatic amino acids: 6-hydroxy-phenylalanine-3,4-dione (TPQ) or 4-(2'-tryptophyl)-tryptophan-6,7-dione (TTQ). There is no doubt that these o-quinones play a role as cofactor, in the case of TPQ in prokaryotic as well as eukaryotic amine oxidases. It appears, therefore, that a novel class of amino-acid-derived cofactors is emerging, ranging from the free radical form of tyrosine and tryptophan to those containing a dicarbonyl group (like the already known pyryvoyl group and the o-quinones here described.     

On the biosynthesis of free and covalently bound PQQ. Glutamic acid decarboxylase from Escherichia coli is a pyridoxo-quinoprotein

Analysis of glutamic acid decarboxylase (GDC) (EC 4.1.1.15) from Escherichia coli ATCC 11246 revealed the presence of six pyridoxal phosphates (PLPs) as well as six covalently bound pyrroloquinoline quinones (PQQs) per hexameric enzyme molecule. This is the second example of a pyridoxo-quinoprotein, suggesting that other atypical pyridoxoproteins (PLP-containing enzymes) have similar cofactor composition. Since the organism did not produce free PQQ and its quinoprotein glucose dehydrogenase was present in the apo form, free PQQ is not used in the assemblage of GDC. Most probably, biosynthesis of covalently bound cofactor occurs in situ via a route which is different from that of free PQQ. Thus, organisms previously believed to be unable to synthesize (free) PQQ could in fact be able to produce quinoproteins with covalently bound cofactor. Implications for the role of PQQ in eukaryotic cells are discussed.     

Adduct formation of pyrroloquinoline quinone and amino acid

When pyrroloquinoline quinone (PQQ) is mixed with an amino acid, a corresponding Schiff base PQQ adduct is readily formed between carbonyl groups of PQQ and the primary amino group. A potent growth stimulating effect for microorganisms was observed with the PQQ adduct when it was administered in a culture medium. Although PQQ itself shows a marked growth stimulating effect, PQQ adducts appeared to be more active than authentic PQQ when compared on a molar basis. Conversely, unlike authentic PQQ, PQQ adducts were shown to be less active (greater than or equal to 100-fold) as the prosthetic group for a quinoprotein apo-glucose dehydrogenase when examined by holoenzyme formation by exogenous addition of PQQ or PQQ adducts. These observations suggested that PQQ adduct formation readily occurs during isolation procedures for PQQ from biological materials or PQQ - chromophore from quinoproteins. Therefore, the presence of such adducts gives a PQQ estimation much lower than theoretically expected. As an example, formation, isolation and characterization of PQQ - serine are described.     

Escherichia coli PQQ-containing quinoprotein glucose dehydrogenase: its structure comparison with other quinoproteins

Membrane-bound glucose dehydrogenase (mGDH) in Escherichia coli is one of the pivotal pyrroloquinoline quinone (PQQ)-containing quinoproteins coupled with the respiratory chain in the periplasmic oxidation of alcohols and sugars in Gram-negative bacteria. We compared mGDH with other PQQ-dependent quinoproteins in molecular structure and attempted to trace their evolutionary process. We also review the role of residues crucial for the catalytic reaction or for interacting with PQQ and discuss the functions of two distinct domains, radical formation in PQQ, and the presumed existence of bound quinone in mGDH.     

Benzoyl-coenzyme A:glycine N-acyltransferase and phenylacetyl-coenzyme A:glycine N-acyltransferase from bovine liver mitochondria. Purification and characterization.

Two closely related acyl-CoA:amino acid N-acyl-transferases were purified to near-homogeneity from preparations of bovine liver mitochondria. Each enzyme consisted of a single polypeptide chain with a molecular weight near 33,000. One transferase was specific for benzoyl-CoA, salicyl-CoA, and certain short straight and branched chain fatty acyl-CoA esters as substrates while the other enzyme specifically used either phenylacetyl-CoA or indoleacetyl-CoA. Acyl-CoA substrates for one transferase inhibited the other. Glycine was the preferred acyl acceptor for both enzymes but either L-asparagine or L-glutamine also served. Peptide products for each transferase were identified by mass spectrometry. Enzymatic cleavage of acyl-CoA was stoichiometric with release of thiol and formation of peptide product. Apparent Km values were low for the preferred acyl-CoA substrates relative to the amino acid acceptors (10(-5) M range compared to greater than 10(-3) M). Both enzymes were inhibited by high nonphysiological concentrations of certain divalent cations (Mg2+, Ni2+, and Zn2+). In contrast to benzoyltransferase, phenylacetyltransferase was sensitive to inhibition by either 10(-4) M 5,5'-dithiobis(2-nitrobenzoate) or 10(-5) M p-chloromercuribenzoate; 10(-4) M phenylacetyl-CoA partially protected phenylacetyltransferase against 5,5'-dithiobis(2-nitrobenzoate) inactivation but 10(-1) M glycine did not. For activity, phenylacetyltransferase required addition of certain monovalent cations (K+, Rb+, Na+, Li+, Cs+, or (NH4)+) to the assay system but benzoyltransferase did not. Preliminary kinetic studies of both transferases were consistent with a sequential reaction mechanism in which the acyl-CoA substrate adds to the enzyme first, glycine adds before CoA leaves, and the peptide product dissociates last.     

Uncompetitive inhibition of Xenopus laevis aldehyde dehydrogenase 1A1 by divalent cations

Aldehyde dehydrogenases (ALDHs) convert aldehydes into their corresponding carboxylic acids. ALDH1A1, also known as ALDH class 1 (ALDH1) or retinaldehyde dehydrogenase (RALDH1), prefers retinal to acetaldehyde as a substrate. To investigate the effects of divalent cations on the dehydrogenase activity of Xenopus laevis ALDH1A1, the formation of acetate and retinoic acid from acetaldehyde and retinal, respectively, was investigated in the presence of Ca2+, Mg2+, Mn2+ or Zn2+. All divalent cations tested inhibited the oxidation of acetaldehyde and retinal by ALDH1A1. When acetaldehyde was used as a substrate, the 50% inhibitory concentrations (IC50) were 10, 24, 35 and 220 microM for Zn2+, Mn2+, Mg2+ and Ca2+, respectively. Kinetic studies of ALDH1A1 dehydrogenase activity in the presence or absence of each cation revealed that the inhibition mode by cations was uncompetitive against acetaldehyde, retinal, and NAD+, and that their inhibitory potencies were greater against acetaldehyde than retinal. It was concluded that the divalent cations inhibited X. laevis ALDH1A1 activity in a substrate-dependent manner by affecting a step of the dehydrogenase reaction that occurred after the formation of the ternary complex of the enzyme, substrate, and coenzyme.     

Coordination of divalent metal ions in the active site of poly(A)-specific ribonuclease

Poly(A)-specific ribonuclease (PARN) is a highly poly(A)-specific 3'-exoribonuclease that efficiently degrades mRNA poly(A) tails. PARN belongs to the DEDD family of nucleases, and four conserved residues are essential for PARN activity, i.e. Asp-28, Glu-30, Asp-292, and Asp-382. Here we have investigated how catalytically important divalent metal ions are coordinated in the active site of PARN. Each of the conserved amino acid residues was substituted with cysteines, and it was found that all four mutants were inactive in the presence of Mg2+. However, in the presence of Mn2+, Zn2+, Co2+, or Cd2+, PARN activity was rescued from the PARN(D28C), PARN(D292C), and PARN(D382C) variants, suggesting that these three amino acids interact with catalytically essential metal ions. It was found that the shortest sufficient substrate for PARN activity was adenosine trinucleotide (A3) in the presence of Mg2+ or Cd2+. Interestingly, adenosine dinucleotide (A) was efficiently hydrolyzed in the presence of Mn2+, Zn2+, or Co2+, suggesting that the substrate length requirement for PARN can be modulated by the identity of the divalent metal ion. Finally, introduction of phosphorothioate modifications into the A substrate demonstrated that the scissile bond non-bridging phosphate oxygen in the pro-R position plays an important role during cleavage, most likely by coordinating a catalytically important divalent metal ion. Based on our data we discuss binding and coordination of divalent metal ions in the active site of PARN.     

The interaction of aminogroups with pyrroloquinoline quinone as detected by the reduction of nitroblue tetrazolium

The interaction of pyrroloquinoline quinone (PQQ) with amino groups was followed by measuring the capacity of adducts to reduce nitroblue tetrazolium (NBT). Of the natural amino acids only glycine, ornithine, and lysine interacted strongly with PQQ. The reducing activity of other less reactive amino acids, but not of lysine, was increased by ammonia, primary or secondary amines. Divalent cations, in contrast inhibited development of NBT-reducing activity. PQQ also developed NBT-reactivity in the presence of serotonin and albumin. A reaction scheme is proposed which explains these findings. It is suggested that the NBT-reducing activity of plasma which is not caused by glycation of plasma proteins, arises from PQQ adducts inherent to plasma. This NBT-reducing activity corresponds to approximately 10 micrograms PQQ/ml plasma.     

Comparative kinetics of D-xylose and D-glucose isomerase activities of the D-xylose isomerase from Thermus aquaticus HB8

The D-xylose isomerase from T. aquaticus accepts, besides D-xylose, also D-glucose, and, with lower efficiency, D-ribose, and D-arabinose as alternative substrates. The activity of the enzyme is strictly dependent on divalent cations. Mn2+ is most effective in the D-xylose isomerase reaction and Co2+ in the D-glucose isomerization. Mg2+ is active in both reactions, Zn2+ only in the further one. The enzyme is strongly inhibited by Cu2+, and weakly by Ni2+, Fe2+, and Ca2+. A hyperbolic dependence of the reaction velocity of the D-xylose isomerase on the concentration of D-xylose xylose and of D-glucose was found, while biphasic saturation curves were obtained by variation of the metal ion concentrations. The D-glucose isomerization reaction shows normal behaviour with respect to the metal ions. A kinetic model was derived on the basis of the assumption of two binding sites for divalent cations, one cofactor site with higher affinity and a second, low affinity site, which modulates the activity of the enzyme.     

Characterization of enzymes of the branched-chain amino acid biosynthetic pathway in Methanococcus spp.

Methanococcus aeolicus, Methanococcus maripaludis, and Methanococcus voltae contain similar levels of four enzymes of branched-chain amino acid biosynthesis: acetohydroxy acid synthase, acetohydroxy acid isomeroreductase, dihydroxy acid dehydratase, and transaminase B. Following growth at low partial pressures of H2-CO2, the levels of these enzymes in extracts of M. voltae are reduced three- to fivefold, which suggests that their synthesis is regulated. The enzymes from M. aeolicus were found to be similar to the eubacterial and eucaryotic enzymes with respect to molecular weights, pH optima, kinetic properties, and sensitivities to O2. The acetohydroxy acid isomeroreductase has a specific requirement for Mg2+, and other divalent cations were inhibitory. It was stimulated threefold by K+ and NH4+ ions and was able to utilize NADH as well as NADPH. The partially purified enzyme was not sensitive to O2. The dihydroxy acid dehydratase is extremely sensitive to O2, and it has a half-life under 5% O2 of 6 min at 25 degrees C. Divalent cations were required for activity, and Mg2+, Mn2+, Ni2+, Co2+, and Fe2+ were nearly equally effective. In conclusion, the archaebacterial enzymes are functionally homologous to the eubacterial and eucaryotic enzymes, which implies that this pathway is very ancient.     

Autolysis of Clostridium acetobutylicum ATCC 824.

The optimum conditions for autolysis of Clostridium acetobutylicum ATCC 824 were determined. Autolysis was optimal at pH 6.3 and 55 degrees C in 0.1 M-sodium acetate/phosphate buffer. The ability of cells to autolyse decreased sharply at the end of the exponential phase of growth. Lysis was stimulated by monovalent cations and compounds that complex divalent cations, and inhibited by divalent cations. The autolysin of C. acetobutylicum, which was mainly cytoplasmic, was purified to homogeneity and characterized as a muramidase. The enzyme was identical to the extracellular muramidase in terms of M(r), isoelectric point and NH2-terminal amino acid sequence. The autolysin was inhibited by lipoteichoic acids and cardiolipin but not by phosphatidylethanolamine and phosphatidylglycerol. A mechanism of regulation and fixation involving lipoteichoic acid, cardiolipin and divalent cations is proposed.     

Ion regulation of homotypic vacuole fusion in Saccharomyces cerevisiae

Biological membrane fusion employs divalent cations as protein cofactors or as signaling ligands. For example, Mg2+ is a cofactor for the N-ethylmaleimide-sensitive factor (NSF) ATPase, and the Ca2+ signal from neuronal membrane depolarization is required for synaptotagmin activation. Divalent cations also regulate liposome fusion, but the role of such ion interactions with lipid bilayers in Rab- and soluble NSF attachment protein receptor (SNARE)-dependent biological membrane fusion is less clear. Yeast vacuole fusion requires Mg2+ for Sec18p ATPase activity, and vacuole docking triggers an efflux of luminal Ca2+. We now report distinct reaction conditions where divalent or monovalent ions interchangeably regulate Rab- and SNARE-dependent vacuole fusion. In reactions with 5 mm Mg2+, other free divalent ions are not needed. Reactions containing low Mg2+ concentrations are strongly inhibited by the rapid Ca2+ chelator BAPTA. However, addition of the soluble SNARE Vam7p relieves BAPTA inhibition as effectively as Ca2+ or Mg2+, suggesting that Ca2+ does not perform a unique signaling function. When the need for Mg2+, ATP, and Sec18p for fusion is bypassed through the addition of Vam7p, vacuole fusion does not require any appreciable free divalent cations and can even be stimulated by their chelators. The similarity of these findings to those with liposomes, and the higher ion specificity of the regulation of proteins, suggests a working model in which ion interactions with bilayer lipids permit Rab- and SNARE-dependent membrane fusion.     

Apparent oxygen-dependent inhibition by superoxide dismutase of the quinoprotein methanol dehydrogenase.

Methanol dehydrogenase activity, when assayed with phenazine ethosulfate (PES) as an electron acceptor, was inhibited by superoxide dismutase (SOD) and by Mn2+ only under aerobic conditions. Catalase, formate, and other divalent cations did not inhibit the enzyme. The enzyme also exhibited significantly higher levels of activity when assayed with PES under anaerobic conditions relative to aerobic conditions. The oxygen- and superoxide-dependent effects on methanol dehydrogenase were not observed when either Wurster's Blue or cytochrome c-55li was used as an electron acceptor. Another quinoprotein, methylamine dehydrogenase, which possesses tryptophan tryptophylquinone (TTQ) rather than pyrroloquinoline quinone (PQQ) as a prosthetic group, was not inhibited by SOD or Mn2+ when assayed with PES as an electron acceptor. Spectroscopic analysis of methanol dehydrogenase provided no evidence for any oxygen- or superoxide-dependent changes in the redox state of the enzyme-bound PQQ cofactor of methanol dehydrogenase. To explain these data, a model is presented in which this cofactor reacts reversibly with oxygen and superoxide, and in which oxygen is able to compete with PES as an electron acceptor for the reduced species.     

Regulation of connexin hemichannels by monovalent cations

Opening of connexin hemichannels in the plasma membrane is highly regulated. Generally, depolarization and reduced extracellular Ca2+ promote hemichannel opening. Here we show that hemichannels formed of Cx50, a principal lens connexin, exhibit a novel form of regulation characterized by extraordinary sensitivity to extracellular monovalent cations. Replacement of extracellular Na+ with K+, while maintaining extracellular Ca2+ constant, resulted in >10-fold potentiation of Cx50 hemichannel currents, which reversed upon returning to Na+. External Cs+, Rb+, NH4+, but not Li+, choline, or TEA, exhibited a similar effect. The magnitude of potentiation of Cx50 hemichannel currents depended on the concentration of extracellular Ca2+, progressively decreasing as external Ca2+ was reduced. The primary effect of K+ appears to be a reduction in the ability of Ca2+, as well as other divalent cations, to close Cx50 hemichannels. Cx46 hemichannels exhibited a modest increase upon substituting Na+ with K+. Analyses of reciprocal chimeric hemichannels that swap NH2- and COOH-terminal halves of Cx46 and Cx50 demonstrate that the difference in regulation by monovalent ions in these connexins resides in the NH2-terminal half. Connexin hemichannels have been implicated in physiological roles, e.g., release of ATP and NAD+ and in pathological roles, e.g., cell death through loss or entry of ions and signaling molecules. Our results demonstrate a new, robust means of regulating hemichannels through a combination of extracellular monovalent and divalent cations, principally Na+, K+, and Ca2+.     

Amino acid and divalent ion permeability of the pores formed by the Bacillus thuringiensis toxins Cry1Aa and Cry1Ac in insect midgut brush border membrane vesicles

The pores formed by Bacillus thuringiensis insecticidal toxins have been shown to allow the diffusion of a variety of monovalent cations and anions and neutral solutes. To further characterize their ion selectivity, membrane permeability induced by Cry1Aa and Cry1Ac to amino acids (Asp, Glu, Ser, Leu, His, Lys and Arg) and to divalent cations (Mg(2+), Ca(2+) and Ba(2+)) and anions (SO(4)(2-) and phosphate) was analyzed at pH 7.5 and 10.5 with midgut brush border membrane vesicles isolated from Manduca sexta and an osmotic swelling assay. Shifting pH from 7.5 to 10.5 increases the proportion of the more negatively charged species of amino acids and phosphate ions. All amino acids diffused well across the toxin-induced pores, but, except for aspartate and glutamate, amino acid permeability was lower at the higher pH. In the presence of either toxin, membrane permeability was higher for the chloride salts of divalent cations than for the potassium salts of divalent anions. These results clearly indicate that the pores are cation-selective.     

The quinohaemoprotein lupanine hydroxylase from Pseudomonas putida

Lupanine hydroxylase catalyses the first reaction in the catabolism of the alkaloid lupanine by Pseudomonas putida. It dehydrogenates the substrate, which can then be hydrated. It is a monomeric protein of M(r) 72,000 and contains a covalently bound haem and a molecule of PQQ. The gene for this enzyme has been cloned and sequenced and the derived protein sequence has a 26 amino acid signal sequence at the N-terminal for translocation of the protein to the periplasm. Many of the features seen in the sequence of lupanine hydroxylase are common with other quinoproteins including the W-motifs that are characteristic of the eight-bladed propeller structure of methanol dehydrogenase. However, the unusual disulfide bridge between adjacent cysteines that is present in some PQQ-containing enzymes is absent in lupanine hydroxylase. The C-terminal domain contains characteristics of a cytochrome c and overall the sequence shows similarities with that of the quinohaemoprotein, alcohol dehydrogenase from Comamonas testosteroni. The gene coding for lupanine hydroxylase has been successfully expressed in Escherichia coli and a procedure has been developed to renature and reactivate the enzyme, which was found to be associated with the inclusion bodies. Reactivation required addition of PQQ and was dependent on calcium ions.