The interactions of artificial coenzymes with alcohol dehydrogenase and other NAD(P)(H) dependent enzymes

The interactions of CL4, a biomimetic analogue of NAD⁺ comprising a nicotinamide functionality coupled via a triazine ring to a dibenzenesulphonic acid unit, and of a series of analogues, with HLADH and other dehydrogenases have been compared to those of the natural coenzymes NAD(P)⁺. CL4, together with one analogue with one of the sulphonic acid groups shifted by one position and another analogue with a single benzenedisulphonic acid unit, have been shown to be functional mimics of NAD⁺ in the oxidation of butan-1-ol by horse liver alcohol dehydrogenase (HLADH). A combination of discontinuous HPLC-based assays and continuous fluorescence based assays were used to deduce approximate kinetic constants for this reaction, using the artificial coenzymes, at pH 7.5 and 37°C. HLADH demonstrated a V_max with the most active analogue which was 4% of that with NAD⁺. The substrate specificity of HLADH using these coenzymes was found to change relative to that using the natural coenzyme. Activity was sought from a range of other dehydrogenases: Bacillus megaterium glucose dehydrogenase, Leuconostoc mesenteroides glucose-6-phosphate dehydrogenase and sheep liver sorbitol dehydrogenase; all displayed activity using a range of the biomimetic coenzymes.     

Artificial redox coenzymes: biomimetic analogues of NAD+

A range of biomimetic analogues of the nicotinamide nucleotide coenzymes NAD(P)(H) have been developed based on the structure of a triazine dye template. These biomimetic redox coenzymes are relatively straightforward and inexpensive to synthesise and display NAD⁺-like activity with different dehydrogenases, despite their apparently minimal structural similarity to the native coenzyme NAD⁺. Horse liver alcohol dehydrogenase oxidises butan-1-ol, using the most active biomimetic coenzyme (Nap 1), with a k _cat value an order of magnitude lower and a K _m for the coenzyme two orders of magnitude higher than those using native NAD⁺. The enzymatically reduced biomimetic coenzymes may be reoxidised by phenazine methosulfate. We believe that these coenzymes may find applications in biotransformations and biosensors, and in the development of biomimetic catalysts where the redox enzyme itself is replaced by a synthetic binding site. Received: 26 October 1998 / Received revision: 25 January 1999 / Accepted: 31 January 1999     

An artificial redox coenzyme based on a triazine dye template

A series of nicotinamide-containing compounds based on the structure of a triazine dye, C.I. Reactive Blue 2, which is known to interact at the coenzyme-binding sites of several NAD(P)(H)-dependent dehydrogenases,^1,2 were designed, synthesized, and characterized. The preparation of these compounds is described. Reduction of the coenzyme mimics with sodium borohydride led to an increase in absorption at 356 nm, analogous to the behavior of the natural coenzyme, NAD⁺.³ When incubated with horse liver alcohol dehydrogenase and ethanol at 25°C and pH 9.0, one of the mimics, Blue N-3, was converted into a new compound with an increased absorption at 356 nm and an R_f value on thin-layer chromatography identical to that of the reduced form produced by treatment with sodium borohydride. The oxidized and reduced forms of Blue N-3 could be separated by reverse-phase ion pair high-performance liquid chromatography. This technique could be used to measure the extent of Blue N-3 reduction: Approximately 90 turnovers were calculated for each enzyme active site over a 48-h period. Gas chromatography analysis suggested that ethanol was simultaneously converted to acetaldehyde. Blue N-3 represents the first example of a new generation of potentially inexpensive, stable, and active biomimetic redox coenzymes.     

The thermotropic properties of coenzyme Q10 and its lower homologues

The thermotropic properties of coenzymes Q₁₀, Q₉, Q₈, and Q₇ have been examined by differential scanning calorimetry and wide-angle X-ray diffraction. Typical scanning calorimetry cooling curves of coenzyme Q from the liquid state exhibit a single exothermic phase transition into a crystalline state at a temperature that decreases as the length of the polyisoprenoid side-chain substituent decreases. Upon subsequent heating, the molecules undergo a series of thermal events which precede the main crystalline-to-liquid endothermic phase transition. The temperature of these transitions increases with increasing chain length. The crystallization phase transition temperature depends markedly on the rate at which the sample is cooled and increases with decreasing scan rate; the temperature of the melting endotherm is not markedly affected by the scan rate. Detailed calorimetric studies of coenzyme Q₁₀ indicate that two crystalline states are formed, one at relatively high cooling rates to low temperatures and the other when preparations are cooled slowly from the liquid state to relatively high temperatures. Heating the crystalline phase formed by rapid cooling causes its transformation into the phase observed by cooling slowly. X-ray diffraction analysis confirmed the existence of these two crystal phases in coenzymes Q₉ and Q₁₀ and the transformation from the rapidly crystallized form to the more ordered form associated with slower cooling rates. At body temperature (310 K) under equilibrium conditions coenzyme Q₁₀ exists in an ordered crystalline phase; the implications of the thermotropic behavior of coenzyme Q₁₀ on mitochondrial functionin vitro andin vivo are discussed.     

Direct interaction of coenzyme M with the active-site Fe-S cluster of heterodisulfide reductase

Heterodisulfide reductase (HDR) catalyzes the formation of coenzyme M (CoM-SH) and coenzyme B (CoB-SH) by the reversible reduction of the heterodisulfide, CoM-S-S-CoB. This reaction recycles the two thiol coenzymes involved in the final step of microbial methanogenesis. Electron paramagnetic resonance (EPR) and variable-temperature magnetic circular dichroism spectroscopic experiments on oxidized HDR incubated with CoM-SH revealed a S=1/2 [4Fe-4S]3) cluster, the EPR spectrum of which is broadened in the presence of CoM-33SH [Duin, E.C., Madadi-Kahkesh, S., Hedderich, R., Clay, M.D. and Johnson, M.K. (2002) Heterodisulfide reductase from Methanothermobacter marburgensis contains an active-site [4Fe-4S] cluster that is directly involved in mediating heterodisulfide reduction. FEBS Lett. 512, 263-268; Duin, E.C., Bauer, C., Jaun, B. and Hedderich, R. (2003) Coenzyme M binds to a [4Fe-4S] cluster in the active site of heterodisulfide reductase as deduced from EPR studies with the [33S]coenzyme M-treated enzyme. FEBS Lett. 538, 81-84]. These results provide indirect evidence that the disulfide binds to the iron-sulfur cluster during reduction. We report here direct structural evidence for this interaction from Se X-ray absorption spectroscopic investigation of HDR treated with the selenium analog of coenzyme M (CoM-SeH). Se K edge extended X-ray absorption fine structure confirms a direct interaction of the Se in CoM-SeH-treated HDR with an Fe atom of the Fe-S cluster at an Fe-Se distance of 2.4A.     

Dynamic affinity between dissociable coenzyme and immobilized enzyme in an affinity chromatographic reactor with single enzyme

The affinity chromatographic reactor (ACR) is a bioreactor which utilizes the dynamic interaction or the dynamic affinity between a free coenzyme and immobilized enzymes for the highly efficient regeneration of dissociable coenzymes. Dynamic affinity between free NAD and immobilized alcohol dehydrogenase (ADH) in ACR was investigated by three different methods. ADH catalyzed both oxidation and reduction of NAD, consuming propionaldehyde and ethanol. The theoretical model under consideration elucidated a criterion for the expression of the dynamic affinity as a relationship among the affinity constants and the concentrations of a coenzyme and immobilized enzyme. This criterion was confirmed experimentally by the measurements of the retention time of NAD and the half-life period of the reactor activity after one-shot pulse injection of NAD to ACR. In the stability measurement of the immobilized enzyme, it became clear that ADH was more stable at the higher concentration in immobilization. Although the present case of coenzyme cycling by a single enzyme is very special, with limited chance for the direct application, the results obtained here provide a theoretical basis for ACR with multienzymes-which is of more general use.     

Dual nucleotide specificity of bovine glutamate dehydrogenase. The role of negative co-operativity.

The thionicotinamide analogues of NAD+ and NADP+ were shown to be good alternative coenzymes for bovine glutamate dehydrogenase, with similar affinity and approx. 40% of the maximum velocity obtained with the natural coenzymes. Both thionicotinamide analogues show non-linear Lineweaver-Burk plots, which with the natural coenzymes have been attributed to negative co-operativity. Since the reduced thionicotinamide analogues have an isosbestic point at 340nm and have an absorption maximum at 400nm, it is possible to monitor reduction of natural coenzyme and thionicotinamide analogue simultaneously by dual-wavelength spectroscopy. When glutamate dehydrogenase is presented with NADP+ and thio-NADP+ simultaneously, the enzyme oligomer senses saturation of its coenzyme-binding sites irrespective of the exact nature of the coenzyme and locks the oligomer into its highly saturated form even when low saturation of the monitored coenzyme is present. These experiments substantiate the suggestion that glutamate dehydrogenase shows negative co-operativity in its catalytically active form.     

Modular coenzyme specificity: A domain‐swopped chimera of glutamate dehydrogenase

Domain‐swopped chimeras of the glutamate dehydrogenases from Clostridium symbiosum (CsGDH) (NAD⁺‐specific) and Escherichia coli (EcGDH) (NADP⁺‐specific) have been produced, with the aim of testing the localization of determinants of coenzyme specificity. An active chimera consisting of the substrate‐binding domain (Domain I) of CsGDH and the coenzyme‐binding domain (Domain II) of EcGDH has been purified to homogeneity, and a thorough kinetic analysis has been carried out. Results indicate that selectivity for the phosphorylated coenzyme does indeed reside solely in Domain II; the chimera utilizes NAD⁺ at 0.8% of the rate observed with NADP⁺, similar to the 0.5% ratio for EcGDH. Positive cooperativity toward L ‐glutamate, characteristic of CsGDH, has been retained with Domain I. An unforeseen feature of this chimera, however, is that, although glutamate cooperativity occurs only at higher pH values in the parent CsGDH, the chimeric protein shows it over the full pH range explored. Also surprising is that the chimera is capable of catalysing severalfold higher reaction rates (V_max) in both directions than either of the parent enzymes from which it is constructed. Proteins 2009. © 2009 Wiley‐Liss, Inc.     

Quantification of coenzyme F420 analogues from methanogenic bacteria by HPLC and fluorimetry

Summary A quantitative assay for analogues of coenzyme F₄₂₀ is presented. The assay combines separation of the coenzymes by binary reversed-phase high performance liquid chromatography with detection by fluorescence, yielding high specificity and sensitivity. Quantification is by calibration with a coenzyme F₄₂₀ standard or by employing coenzyme F₄₂₀ fragments as internal standards.     

Effects of coenzyme analogues on the binding of p-aminobenzoyl-L-glutamate and 2,4-diaminopyrimidine to Lactobacillus casei dihydrofolate reductase

The binding of p-aminobenzoyl-L-glutamate and 2,4-diaminopyrimidine to dehydrofolate reductase from Lactobacillus casei MTX/R in the presence of a series of co-enzymes and coenzyme analogues has been measured fluorometrically. These two ligands, which can be regarded as "fragments" of the powerful inhibitor methotrexate, have been shown to bind cooperatively in the absence of coenzyme [Birdsall, B., Burgen, A. S. V., Rodrigues de Miranda, J., & Roberts, G. C. K. (1978) Biochemistry 17, 2102], p-amino-benzoyl-L-glutamate binding 58 times more tightly in the presence of 2,4-diaminopyrimidine than in its absence. In the presence of coenzymes, this cooperativity ranges from 1.8- to 428-fold. The effects of coenzymes on individual binding steps range from an 8-fold decrease in binding constant to a 23-fold increase. The structural specificity of these effects are discussed in terms of a model involving ligand-induced conformational changes and compared with the effects on trimethoprim and methotrexate binding described in the preceding paper [Birdsall, B., Burgen, A. S. V., & Roberts, G. C. K. (1980) Biochemistry (first paper of four in this issue)].     

Stable and specific expression of 4-coumarate:coenzyme A ligase gene (4CL1) driven by the xylem-specific Pto4CL1 promoter in the transgenic tobacco

The ability of 4-coumarate:coenzyme A ligase promoter from Populus tomentosa (Pto4CL1p) to drive expression of the GUS reporter gene and 4-coumarate:coenzyme A ligase gene in tobacco has been studied using transgenic plants produced by Agrobacterium-mediated transformation. Intense GUS histochemical staining was detected in the xylem of stem in transgenic tobacco plants carrying the 1140 bp Pto4CL1p promoter. To further investigate the regulation function of the tissue-specific expression promoter, Pto4CL1p, a binary vector containing Pto4CL1p promoter fused with 4CL1 gene was transferred into tobacco. The activity of the 4CL1 enzyme doubled in the stems of transgenic tobacco but did not increase in the leaves. The content of lignin was increased 25% in the stem but there was no increase in the leaves of transgenic tobacco.     

Pyridine coenzyme analogues. Synthesis and characterization of alpha- and beta-nicotinamide arabinoside adenine dinucleotides

The synthesis and characterization of a new pyridine coenzyme analogue containing a nicotinamide arabinonucleotide moiety are reported. The redox potentials are -339 mV for beta-oxidized nicotinamide arabinoside adenine dinucleotide and -319 mV for alpha-oxidized nicotinamide adenine dinucleotide, and the lambda max is 346 and 338 nm for beta- and alpha-reduced nicotinamide arabinoside adenine dinucleotides (araNADH), respectively. Anomerization of the reduced analogues leads to a 5:1 ratio of alpha-araNADH to beta-araNADH at 90 degrees C. These results establish that the relative configuration of the 2'-hydroxyl to the base is the primary determinant for the configuration-dependent changes in lambda max, the redox potential of the pyridine nucleotides, and the preferred anomeric configuration of the reduced coenzymes. Comparison of the 1H and 31P NMR spectral data of the analogues with those for the ribo coenzymes is reported and the conformational analysis discussed. The coenzyme properties of the arabino analogues have been evaluated with yeast and horse liver alcohol dehydrogenases. Both the alpha- and beta-anomers are found to serve as coenzymes, and the stereochemistry of hydride transfer is identical for both anomers.     

Changes in concentrations of coenzyme F420 analogs during batch growth of Methanosarcina barkeri and Methanosarcina mazei

Coenzyme F420 has been assayed by high-performance liquid chromatography with fluorimetric detection; this permits quantification of individual coenzyme F420 analogs whilst avoiding the inclusion of interfering material. The total intracellular coenzyme F420 content of Methanosarcina barkeri MS cultivated on methanol and on H2-CO2 and of Methanosarcina mazei S-6 cultured on methanol remained relatively constant during batch growth. The most abundant analogs in M. barkeri were coenzymes F420-2 and F420-4, whilst in M. mazei coenzymes F420-2 and F420-3 predominated. Significant changes in the relative proportions of the coenzyme F420 analogs were noted during batch growth, with coenzymes F420-2 and F420-4 showing opposite responses to each other and the same being also true for coenzymes F420-3 and F420-5. This suggests that an enzyme responsible for transferring pairs of glutamic acid residues may be active. The degradation fragment FO was also detected in cells in late exponential and stationary phase. Coenzyme F420 analogs were present in the culture supernatant of both methanogens, in similar proportions to that in the cells, except for FO which was principally located in the supernatant.     

Changes in concentrations of coenzyme F420 analogs during batch growth of Methanosarcina barkeri and Methanosarcina mazei.

Coenzyme F420 has been assayed by high-performance liquid chromatography with fluorimetric detection; this permits quantification of individual coenzyme F420 analogs whilst avoiding the inclusion of interfering material. The total intracellular coenzyme F420 content of Methanosarcina barkeri MS cultivated on methanol and on H2-CO2 and of Methanosarcina mazei S-6 cultured on methanol remained relatively constant during batch growth. The most abundant analogs in M. barkeri were coenzymes F420-2 and F420-4, whilst in M. mazei coenzymes F420-2 and F420-3 predominated. Significant changes in the relative proportions of the coenzyme F420 analogs were noted during batch growth, with coenzymes F420-2 and F420-4 showing opposite responses to each other and the same being also true for coenzymes F420-3 and F420-5. This suggests that an enzyme responsible for transferring pairs of glutamic acid residues may be active. The degradation fragment FO was also detected in cells in late exponential and stationary phase. Coenzyme F420 analogs were present in the culture supernatant of both methanogens, in similar proportions to that in the cells, except for FO which was principally located in the supernatant.     

4-Coumarate:coenzyme A ligase in black locust (<Emphasis Type="Italic">Robinia pseudoacacia</Emphasis>) catalyses the conversion of sinapate to sinapoyl-CoA

4-Coumarate:coenzyme A (CoA) ligase (4CL, EC 6.2.1.12) in crude enzyme preparation from the developing xylem of black locust (Robinia pseudoacacia) converted sinapate to sinapoyl CoA. The sinapate-converting activity was not inhibited by other cinnamate derivatives, such as p-coumarate, caffeate or ferulate, in the mixed-substrate assay. The crude extract prepared from the developing xylem was separated by anion-exchange chromatography into three different 4CL isoforms. The isoform 4CL1 had a strong substrate preference for p-coumarate, but lacked the activity for ferulate and sinapate. On the other hand, 4CL2 and 4CL3 displayed activity toward sinapate and also possessed high activity toward caffeate as well as p-coumarate. The crude extract from the shoots exhibited a very similar substrate preference to that of the developing xylem; therefore, 4CL2 may be a major isoform in both crude enzyme preparations. These results support the hypothesis that sinapate-converting 4CL isoform is constitutively expressed in lignin-forming cells.     

Regulation of coenzyme utilization by mitochondrial NAD(P)-dependent malic enzyme

1. Skeletal muscle mitochondrial NAD(P)-dependent malic enzyme [EC 1.1.1. 39, L-malate:NAD+ oxidoreductase (decarboxylating)] from herring could use both coenzymes, NAD and NADP, in a similar manner. 2. The coenzyme preference of mitochondrial NAD(P)-dependent malic enzyme was probed using dual wavelength spectroscopy and pairing the natural coenzymes, NAD or NADP with their respective thionicotinamide analogues, s-NADP or s-NAD, that have absorbance maxima in reduced forms at 400 nm. 3. s-NAD and s-NADP were found to be good alternate substrates for NAD(P)-dependent malic enzyme, the apparent Km values for the thioderivatives were similar to those of the corresponding natural coenzymes. 4. ATP produced greater inhibition of the NAD or s-NAD linked reactions than of the NADP or s-NADP-linked reactions of skeletal muscle mitochondrial NAD(P)-dependent malic enzyme. 5. At 5 mM malate concentration and in the presence of 2 mM ATP the NADP-linked reaction is favoured and the activity ratios, V(s-NADP)/V(NAD) or V(NADP)/V(s-NAD), are 6 and 26, respectively.     

Yeast aspartate aminotransferase: preparation of the apoenzyme and its interaction with coenzyme analogues

Affinity chromatography of yeast aspartate aminotransferase [l-aspartate: 2-oxoglutarate aminotransferase, EC 2.6.1.1] on N′(ω-aminohexyl) pyridoxamine-5-phosphate Sepharose 4B is reported. The specific activity of the enzyme obtained, fully activated with pyridoxal-5-phosphate, was higher than that of previous preparations but the yield of purified enzyme was poor. Purification using DEAE-cellulose gave a higher yield of enzyme with lower specific activity. This preparation contained an appreciable amount of the holoenzyme. Use of sodium borohydride permitted the preparation of apoenzyme containing only 1.4% of the holo-form. Four coenzyme analogues were synthesized. These were the N′-acetyl-, the N′-methyl- and the N′-benzyloxycarbonylglycyl-pyridoxamine-5-phosphate and the O-acetylpyridoxal-5-phosphate. The three N′-substituted pyridoxamine-5-phosphate derivatives were all effective inhibitors of the enzyme, while the O-acetylpyridoxal-5-phosphate bound to the apoenzyme and gave an active enzyme.     

The effect of thiamine pyrophosphate modification on its coenzyme function in a transketolase-catalyzed reaction

The coenzyme function of TPP analogues: 4'-NH-methyl-TPP,6'-methyl-TPP and 6'-methyl-4-nor-TPP has been studied in a transketolase-catalyzed reaction. Their dissociation constants have been found with the aid of the circular dichroism method, and coenzyme activity has been determined in a complete transketolase reaction, involving the substrate-donor and the substrate-acceptor, and also at the intermediate stage (by the alpha-carbanionic intermediate oxidation rate). The coenzyme activity values have been found different and largely dependend on the nature of the substrates used. A possibility of TPP functioning by the "two-center mechanism" in a transketolase-catalyzed reaction is discussed.     

Enhancement of coenzyme binding by a single point mutation at the coenzyme binding domain of E. coli lactaldehyde dehydrogenase

Phenylacetaldehyde dehydrogenase (PAD) and lactaldehyde dehydrogenase (ALD) share some structural and kinetic properties. One difference is that PAD can use NAD+ and NADP+, whereas ALD only uses NAD+. An acidic residue has been involved in the exclusion of NADP+ from the active site in pyridine nucleotide-dependent dehydrogenases. However, other factors may participate in NADP+ exclusion. In the present work, analysis of the sequence of the region involved in coenzyme binding showed that residue F180 of ALD might participate in coenzyme specificity. Interestingly, F180T mutation rendered an enzyme (ALD-F180T) with the ability to use NADP+. This enzyme showed an activity of 0.87 micromol/(min * mg) and K(m) for NADP+ of 78 microM. Furthermore, ALD-F180T exhibited a 16-fold increase in the V(m) /K(m) ratio with NAD+ as the coenzyme, from 12.8 to 211. This increase in catalytic efficiency was due to a diminution in K(m) for NAD+ from 47 to 7 microM and a higher V(m) from 0.51 to 1.48 micromol/(min * mg). In addition, an increased K(d) for NADH from 175 (wild-type) to 460 microM (mutant) indicates a faster product release and possibly a change in the rate-limiting step. For wild-type ALD it is described that the rate-limiting step is shared between deacylation and coenzyme dissociation. In contrast, in the present report the rate-limiting step in ALD-F180T was determined to be exclusively deacylation. In conclusion, residue F180 participates in the exclusion of NADP+ from the coenzyme binding site and disturbs the binding of NAD+.     

Antisense suppression of 4-coumarate:coenzyme A ligase activity in Arabidopsis leads to altered lignin subunit composition.

The phenylpropanoid enzyme 4-coumarate:coenzyme A ligase (4CL) is considered necessary to activate the hydroxycinnamic acids for the biosynthesis of the coniferyl and sinapyl alcohols subsequently polymerized into lignin. To clarify the role played by 4CL in the biosynthesis of the guaiacyl (G) and syringyl (S) units characteristic of angiosperm lignin, we generated 4CL antisense Arabidopsis lines having as low as 8% residual 4CL activity. The plants had decreases in thioglycolic acid-extractable lignin correlating with decreases in 4CL activity. Nitrobenzene oxidation of cell walls from bolting stems revealed a significant decrease in G units in 4CL-suppressed plants; however, levels of S lignin units were unchanged in even the most severely 4CL-suppressed plants. These effects led to a large decrease in the G/S ratio in these plants. Our results suggest that an uncharacterized metabolic route to sinapyl alcohol, which is independent of 4CL, may exist in Arabidopsis. They also demonstrate that repression of 4CL activity may provide an avenue to manipulate angiosperm lignin subunit composition in a predictable manner.