8α-Substituted flavins of biological importance

Abstract

ion from the N(5) position. 6. 8GGcysteinylfIavin thiohemiacetals On chemical grounds thiohemiacetals are generally regarded as being unstable. Yet, all available evidence suggests that the peptide chain of Chromxztium cytochrome c552 is linked by a thiohemiacetal bridge to the 8a position of FAD [24,27], although the flavin is very tightly bound to the enzyme, and that even flavin peptides are stable enough for isolation in the pure state. This raises the question as to the additional bonds or interactions responsible for the stable binding of the FAD to the apoenzyme in this protein. It was reported many years ago [43] that the flavin component of this flavocytochrome is not released by acid ammonium sulfate or trichloroacetic acid treatment but is cleaved from the protein by tryptic digestion, prolonged incubation with saturated urea solution, exposure to pH > 9, and treatment with organic mercurials [44]. A related enzyme, flavocytochrome cs 5 3 from Chlorobium thiosulfatophilum has also been reported to contain a covalently bound flavin, the release of which required prolonged incubation with saturated urea solution [44,45] . While the nature of the flavin component of the Chlorobium enzyme has not been further investigated, the Chromatium flavin has been extensively studied in two laboratories. Release of the flavin from the protein on prolonged incubation with 8 M urea has been confirmed [46] , but it appears likely [47] that the agent Volume 42, number 1 FEBS LETTERS May 1974 responsible is the cyanate impurity in urea preparations. On the other hand, release of the flavin on exposure to alkaline pH or organic mercurials could not be confirmed by the present authors. The product released by digestion with urea is an FAD derivative, substituted at the 8~ position, as shown by the hypsochromic shift of the absorption spectrum [ 19,461 and the hyperfine EPR spectrum of the radical cation [ 191. Its properties strongly suggest that is 8-carboxy FAD. Oxidation of the denatured cytochrome with performic acid also releases 8 carboxyriboflavin in excellent yield [ 191, showing that the peptide is indeed substituted on the g&-carbon. Digestion with trypsin-chymotrypsin or with pepsin releases FAD peptides, which have been isolated in homogeneous form [24] . These exhibit a hypsochromic shift of the second fluorescence excitation maximum from 372 to 365 mn and extensive quenching of fluorescence, independent of pH between 3.2 and 7.0 (5% of the quantum yield, compared with riboflavin, in the case of trypticchymotryptic peptide and 1% in the peptic peptide, both at the FMN level). On performic acid oxidation the fluorescence increases, with further hypsochromic shift of the second excitation band [27] . The properties are strongly reminiscent of the flavin thioether isolated from monoamine oxidase [ 171 and suggest that the substituent is cysteine. The presence of cysteine was confirmed by the positive chloroplatinic reaction and by analysis for cysteine after acid hydrolysis of the two flavin peptides [24,27] . The properties of the flavin in the flavocytochrome, however, differ in important respects from those in monoamine oxidase, so that a thioether linkage is highly unlikely. (1) Performic acid oxidation of the cytochrome cs52, but not of monoamine oxidase, releases 8 carboxyriboflavin. (2) Reduction by Zn releases cysteine in the case of flavin peptides from monoamine oxidase (or cysteinyl riboflavin thioether) but not from the Chromatium flavin peptides. (3) The flavin peptide from monoamine oxidase, after performic acid oxidation, shows 80-85% of the fluorescence of riboflavin, while those from Chromatium only SC%. (As discussed below, with the peptic peptide this degree of fluorescence is obtained only on oxidation at 4O’C) [6,24] . Alternate structures compatible with the negative 12-azide test of the Chromatium peptides are: disulfide, thioester, and thiohemiacetal. The first two are ruled out since neither dithionite nor hydroxylamine liberate the flavin from the protein. This leaves a flavin thiohemiacetal as the only possibility. Direct evidence for this structure was obtained by acid hydrolysis of the peptic peptide (6 N HCl, 95” C, N,) and demonstration of the release of 8-formylriboflavin [6]. Supporting evidence came from the observation [24] that digestion of the peptic peptide with aminopeptidase M yields two flavin components (electrophoresis, pH 6.5). One has 0.95 of the mobility of FAD, as expected for an aminoacyl FAD derivative, the other a much greater anionic mobility than FAD, suggesting that the amino group of the cysteine is blocked. This is compatible with cyclization to a thiazolidine derivative during aminopeptidase digestion (fig. 3).