Enzymatic release of iron from sideramines in fungi. NADH: Sideramine oxidoreductase in Neurospora crassa

Young mycelia of the fungus Neurospora crassa contain a soluble NADH-linked sideramine reductase, which may be responsible for liberating iron in vivo from accumulated sideramines during iron-deficient cultivation. The enzymes can be assayed using a soluble supernatant fraction, EDTA, and an atmosphere of pure nitrogen. The enzyme is stable without loss of activity up to 45°C and has an optimum of activity at pH 7.0. Besides coprogen (K_m = 100 μM, V = 2.8 nmol/min. per mg protein), some other ferrichrome-type compounds are reduced. However, ferrichrome, ferrirubin, coprogen B and ferrioxamine are poor substrates. When the mucelia were grown in a medium containing 10^?5 M ferric iron, the activity of the reductase was found to be only 30% of that found under low iron conditions. The enzyme is inhibited by oxygen, SH-alkylating agents and partly by some detergents. Unlike the reductase of N. crassa, the corresponding enzyme from Aspergillus fumigatus revealed low reduction of coprogen and high reduction of ferrichrome, indicating genus-dependent specificities of sideramine reduction enzymes in fungi. The participation of acids of the citric acid cycle as natural iron acceptors during strong iron deficiency is studied and confirmed by iron uptake measurements on isolated mitochondria.     

Evidence for a common siderophore transport system but different siderophore receptors in Neurospora crassa.

Uptake and competition experiments were performed with Neurospora crassa and Penicillium parvum by using 14C-labeled coprogen and 55Fe-labeled ferrichrome-type siderophores. Several siderophores of the ferrichrome family, such as ferrichrome, ferricrocin, ferrichrysin, and tetraglycyl-ferrichrome as well as the semisynthetic ferricrocin derivatives O-(phenyl-carbamoyl)-ferricrocin and O-(sulfanilyl-carbamoyl)-ferricrocin were taken up by N. crassa. The ferrichrome-type siderophores used vary in the structure of the peptide backbone but possess a common lambda-cis configuration about the iron center and three identical ornithyl-delta-N-acetyl groups as surrounding residues. This suggests that these ferrichrome-type siderophores are recognized by a common ferrichrome receptor. We also concluded that the ferrichrome receptor is lambda-cis specific from the inability to take up the synthetic enantiomers, enantio-ferrichrome and enantio-ferricrocin, possessing a delta-cis configuration about the iron center. On the other hand, we found that coprogen, possessing a delta-absolute configuration and two trans-anhydromevalonic acid residues around the metal center, was also taken up by N. crassa and was competitively inhibited by the ferrichrome-type siderophores. We therefore propose the existence of a common siderophore transport system but the presence of different siderophore receptors in N. crassa. In addition, ferrirubin, which is very slowly transported by N. crassa, inhibited both coprogen and ferrichrome-type siderophore transport. Contrary to the findings with N. crassa, transport experiments with P. parvum revealed the presence of a ferrichrome receptor but the absence of a coprogen receptor; coprogen was neither transported nor did it inhibit the ferrichrome transport.     

Soluble and membrane-bound ferrisiderophore reductases of Escherichia coli K-12

After uptake of microbial ferrisiderophores, iron is assumed to be released by reduction. Two ferrisiderophore-reductase activities were identified in Escherichia coli K-12. They differed in cellular location, susceptibility to amytal, and competition between oxygen and ferrichrome-iron(III) reduction. The ferrisiderophore reductase associated with the 40,000×g sediment (membrane-bound enzyme) was inhibited by 10 mM amytal in contrast to the ferrisiderophore reductase present in the 100,000×g supernatant (soluble enzyme). Reduction by the membrane-bound enzyme followed sigmoid kinetics, but was biphasic in the case of the soluble enzyme. The soluble reductase could be assigned to a protein consisting of a single polypeptide of M _r 26000. Reduction of iron(III) by the purified enzyme depended on the addition of NADH or NADPH which were equally active reductants. The cofactor FMN and to a lesser degree FAD stimulated the reaction. Substrate specificity of the soluble reductase was low. In addition to the hydroxamate siderophores arthrobactin, schizokinen, fusigen, aerobactin, ferrichrome, ferrioxamine B, coprogen, and ferrichrome A, the iron(III) complexes of synthetic catecholates, dihydroxy benzoic acid, and dicitrate, as well as carrier-free iron(III) were accepted as substrates. Both ferrisiderophore reductases were not controlled by the fur regulatory system and were not suppressed by anaerobic growth.Abbreviations DHB dihydroxybenzoic acid - MECAM 1,3,5-N,N,N-tris-(2,3-dihydroxybenzoyl)-triamino-methylbenzene - MECAMS 2,3-dihydroxy-5-sulfonyl-derivative of MECAM     

FhuF, part of a siderophore-reductase system

FhuF is a cytoplasmic 2Fe-2S protein of Escherichia coli loosely associated with the cytoplasmic membrane. E. coli fhuF mutants showed reduced growth on plates with ferrioxamine B as the sole iron source, although siderophore uptake was not defective in transport experiments. Removal of iron from coprogen, ferrichrome, and ferrioxamine B was significantly lower in fhuF mutants compared to the corresponding parental strains, which suggested that FhuF is involved in iron removal from these hydroxamate-type siderophores. A redox potential E(1/2) of -310 +/- 25 mV relative to the normal hydrogen electrode was determined for FhuF by EPR redox titration; this redox potential is sufficient to reduce the siderophores coprogen and ferrichrome. M?ssbauer spectra revealed that FhuF in its [Fe(2+)-Fe(3+)] state is also capable of direct reduction of ferrioxamine B-bound ferric iron, thus proving its reductase function. This is the first report on a bacterial siderophore-iron reductase which in vivo seems to be specific for a certain group of hydroxamates.     

Molecular recognition of siderophores in fungi: role of iron-surrounding N-acyl residues and the peptide backbone during membrane transport in Neurospora crassa.

Recognition of ferric siderophores in Neurospora crassa was found to depend on the number and kind of N-acyl residues that surrounded the iron coordination center. In the coprogen series, uptake decreased in the order of coprogen, neocoprogen I, and neocoprogen II, indicating that gradual replacement of the N-transanhydromevalonyl groups by N-acetyl groups had an adverse effect on uptake. The reverse effect was observed in the ferrichrome series, where uptake decreased in the order of ferrichrysin, asperchrome D1, asperchrome B1, and ferrirubin. Configuration of the anhydromevalonyl group (cis or trans) in ferrichromes was also an important determinant in the recognition process. On the basis of uptake and inhibition studies, it is proposed that in ferrichromes part of the molecule (iron configuration and the N-acyl groups) is responsible for binding, whereas another (cyclic peptide ring) is involved in the subsequent process of transport.     

Isolation and identification of the conidial germination factor of Neurospora crassa.

The germination-essential substance (germination factor [GF]) that is lost from conidia of Neurospora crassa on exposure to solutions of low water activity has been isolated and identified as a group of iron-transport compounds, or siderochromes. The principal siderochrome of conidia is ferricrocin, a cyclic hexapeptide. A closely related substance, ferrichrome C, is tentatively identified as a minor constituent. The same substances are also present in extracts of mycelium along with small amounts of a third siderochrome, which has not been identified. The GF activity of culture filtrates is due to coprogen, the only siderochrome previously identified with N. crassa.     

Kinetic studies on the specificity of chelate-iron uptake in Aspergillus.

Three strains of the fungus Aspergillus, Aspergillus quadricinctus (E. Yuill), A. fumigatus (Fresenius), and A. melleus (Yukawa), each producing different iron-chelating compounds during iron-deficient cultivation, were used for 55Fe3+ uptake measurements. Iron from chelates of the ferrichrome-type family was taken up by young mycelia of all strains tested, irrespective of the ferrichrome-type compound these strains predominantly produce in low-iron cultures. Ferrichrysin-producing strains, however, seem to favor ferrichrysin iron uptake, whereas ferrichrome, ferricrocin, and even ferrirubin showed similar iron transport properties in all of these strains. Compared to iron uptake from ferrichrome-type compounds (Km approximately 4 uM) iron uptake from fusigen revealed completely different kinetic values (Km approximately 50 to 80 muM). Iron from exogenous chelates, e.g., from coprogen produced by Neurospora crassa for ferrioxamine B produced by Streptomyces pilosus, can obviously not be taken up by Aspergillus, confirming the pronounced specificity of chelate-iron transport in fungi.     

Cellular and extracellular siderophores of Aspergillus nidulans and Penicillium chrysogenum.

Aspergillus nidulans and Penicillium chrysogenum produce specific cellular siderophores in addition to the well-known siderophores of the culture medium. Since this was found previously in Neurospora crassa, it is probably generally true for filamentous ascomycetes. The cellular siderophore of A. nidulans is ferricrocin; that of P. chrysogenum is ferrichrome. A. nidulans also contains triacetylfusigen, a siderophore without apparent biological activity. Conidia of both species lose siderophores at high salt concentrations and become siderophore dependent. This has also been found in N. crassa, where lowering of the water activity has been shown to be the causal factor. We used an assay procedure based on this dependency to reexamine the extracellular siderophores of these species. During rapid mycelial growth, both A. nidulans and P. chrysogenum produced two highly active, unidentified siderophores which were later replaced by a less active or inactive product--coprogen in the case of P. chrysogenum and triacetylfusigen in the case of A. nidulans. N. crassa secreted coprogen only. Fungal siderophore metabolism is varied and complex.     

The role of ferrichrome reductase in iron metabolism of Ustilago sphaerogena.

Ferrichrome, the ferric ionophore for Ustilago sphaerogena, can serve as a source of iron for the enzyme ferrochelatase (protoheme ferrolyase, EC 4.99.1.1) in this organism, but only after enzymatic removal of the iron from its carrier. U. sphaerogena contains a specific ferrichrome reductase (NADH:ferrichrome oxidoreductase) which catalyzes cellular dissociation of the complex by reduction of the metal to the ferrous state. A spectrophotometric assay was developed based on trapping of the ferrous ion produced by ferrozine. There is an apparent inhibition by oxygen which is thought to be due to re-oxidation of the metal under the assay conditions. The close structural analogue, ferrichrome A, is not a substrate, nor is the ester type siderochrome ferric hexahydro-N,N',N"-triacetylfusarinine C. Aluminum desferriferrichrome is inhibitory. The importance of this enzyme for the metabolism of iron in this organism is discussed.     

Metabolic utilization of 57Fe-labeled coprogen in Neurospora crassa. An in vivo M?ssbauer study

M?ssbauer spectra of whole cells of Neurospora crassa arg-5 ota aga (a siderophore-free mutant) show that the siderophore coprogen is accumulated inside the cell as an entity. 57Fe from 57Fe-labeled coprogen is slowly removed from the complex (45% in 27 h). The rate of removal depends on the degree of iron starvation of the cells. The distribution of 55Fe from [55Fe]coprogen in vacuoles, membranes, and cytoplasm has been also determined. From this it is clear that coprogen is accumulated in the cytoplasm. In addition to its role as a siderophore, coprogen serves as an iron-storage compound. No holoferritins could be detected. We therefore conclude that this type of iron-storage protein is lacking in N. crassa. Metabolized iron was found predominantly to exist as an envelope of Fe(II) high-spin (delta = 1.2-1.3 mm s-1; delta EQ = 3.0-3.1 mm s-1 at 4.2 K) and fast-relaxing Fe(III) high-spin species (delta approximately equal to 0.25 mm s-1 and 0.45 mm s-1; delta EQ approximately equal to 0.6 mm s-1 and 0.55 mm s-1, respectively, at 4.2 K). An assignment of these major iron metabolites is difficult. The M?ssbauer data of the Fe(II) species do not fit those reported for heme, cytochromes and ferredoxins. We therefore assume that this iron metabolite represents a novel internal iron compound. One of the Fe(III) species becomes the dominant component of the cell spectra after 65 h of metabolization and might correspond to an iron-storage compound with iron oxide cores similar to bacterioferritin. After 27 h of growth in mycelia supplied with 57Fe-labeled coprogen, the siderophore ferricrocin was observed in the cell spectra. This is unexpected, since N. crassa arg-5 ota aga is unable to synthesize ornithine. We assume that ferricrocin is synthesized by the use of coprogen degradation products.     

Ferricrocin functions as the main intracellular iron-storage compound in mycelia ofNeurospora crassa

Summary Neurospora crassa produces several structurally distinct siderophores: coprogen, ferricrocin, ferrichrome C and some minor unknown compounds. Under conditions of iron starvation, desferricoprogen is the major extracellular siderophore whereas desferriferricrocin and desferriferrichrome C are predominantly found intracellularly. Mössbauer spectroscopic analyses revealed that coprogen-bound iron is rapidly released after uptake in mycelia of the wild-typeN.crassa 74A. The major intracellular target of iron distribution is desferriferricrocin. No ferritin-like iron pools could be detected. Ferricrocin functions as the main intracellular iron-storage peptide in mycelia ofN. crassa. After uptake of ferricrocin in both the wild-typeN. crassa 74A and the siderophore-free mutantN. crassa arg-5 ota aga, surprisingly little metabolization (11%) could be observed. Since ferricrocin is the main iron-storage compound in spores ofN. crassa, we suggest that ferricrocin is stored in mycelia for inclusion into conidiospores.     

Ferrichrome transport in Escherichia coli K-12: altered substrate specificity of mutated periplasmic FhuD and interaction of FhuD with the integral membrane protein FhuB.

FhuD is the periplasmic binding protein of the ferric hydroxamate transport system of Escherichia coli. FhuD was isolated and purified as a His-tag-labeled derivative on a Ni-chelate resin. The dissociation constants for ferric hydroxamates were estimated from the concentration-dependent decrease in the intrinsic fluorescence intensity of His-tag-FhuD and were found to be 0.4 microM for ferric aerobactin, 1.0 microM for ferrichrome, 0.3 microM for ferric coprogen, and 5.4 microM for the antibiotic albomycin. Ferrichrome A, ferrioxamine B, and ferrioxamine E, which are poorly taken up via the Fhu system, displayed dissociation constants of 79, 36, and 42 microM, respectively. These are the first estimated dissociation constants reported for a binding protein of a microbial iron transport system. Mutants impaired in the interaction of ferric hydroxamates with FhuD were isolated. One mutated FhuD, with a W-to-L mutation at position 68 [FhuD(W68L)], differed from wild-type FhuD in transport activity in that ferric coprogen supported promotion of growth of the mutant on iron-limited medium, while ferrichrome was nearly inactive. The dissociation constants of ferric hydroxamates were higher for FhuD(W68L) than for wild-type FhuD and lower for ferric coprogen (2.2 microM) than for ferrichrome (156 microM). Another mutated FhuD, FhuD(A150S, P175L), showed a weak response to ferrichrome and albomycin and exhibited dissociation constants two- to threefold higher than that of wild-type FhuD. Interaction of FhuD with the cytoplasmic membrane transport protein FhuB was studied by determining protection of FhuB degradation by trypsin and proteinase K and by cross-linking experiments. His-tag-FhuD and His-tag-FhuD loaded with aerobactin specifically prevented degradation of FhuB and were cross-linked to FhuB. FhuD loaded with substrate and also FhuD free of substrate were able to interact with FhuB.     

Siroheme: a prosthetic group of the Neurospora crassa assimilatory nitrite reductase.

The Neurospora crassa assimilatory nitrite reductase (EC 1.6.6.4) catalyzes the NADPH-dependent reduction of nitrite to ammonia, a 6-electron transfer reaction. Highly purified preparations of this enzyme exhibit absorption spectra which suggest the presence of a heme component (wavelength maxima for oxidized senzyme: 390 and 578 nm). There is a close correspondence between nitrite reductase activity and absorbance at 400 nm when partially purified nitrite reductase preparations are subjected to sucrose gradient centrifugation. In addition, a role for an iron component in the formation of active nitrite reductase is indicated by the fact that nitrate-induced production of nitrite reductase activity in Neurospora mycelia in vivo requires the presence of iron in the induction medium. The heme chromophore present in Neurospora nitrite reductase preparations is reducible by NADPH. Complete reduction, however, requires the presence of added FAD. The NADPH-nitrite reductase activity of the enzyme is also dependent upon addition of FAD. A spectrally unique complex is formed between the heme chromophore and nitrite (or a reduction product thereof) when nitrite is added to NADPH-reducted enzyme. Carbon monoxide forms a complex with the heme chromophore of nitrite reductase with an intense alpha-band maximum at 590 nm and a beta-band of lower intensity at 550 nm. CO is an inhibitor of NADPH-nitrite reductase activity. Spectrophotometrically detectable CO complex formation and Co inhibition of enzyme activity share the following properties...     

Neurospora crassa NAD(P)H-nitrite reductase. Studies on its composition and structure

Neurospora crassa nitrite reductase (Mr = 290,000) catalyzes the NAD(P)H-dependent 6-electron reduction of nitrite to ammonia via flavin and siroheme prosthetic groups. Homogeneous N. crassa nitrite reductase has been prepared employing conventional purification methods followed by affinity chromatography on blue dextran-Sepharose 4B. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of homogeneous nitrite reductase reveals a single subunit band of Mr = 140,000. Isoelectric focusing of dissociated enzyme followed by sodium dodecyl sulfate-gel electrophoresis in the second dimension yields a single subunit spot with an isoelectric point at pH 6.8-6.9. Two-dimensional thin layer chromatography of acid-hydrolyzed nitrite reductase treated with 5-dimethylaminoaphthalene-1-sulfonyl chloride yields a single reactive NH2-terminal corresponding to glycine. An investigation of the prosthetic groups of nitrite reductase reveals little or no flavin associated with the purified protein, although exogenously added FAD is required for activity in vitro. An iron content of 9-10 Fe eq/mol suggests the presence of nonheme iron in addition to the siroheme moieties. Amino acid analysis yields 43 cysteinyl residues and sulfhydryl reagents react with 50 thiol eq/mol of nitrite reductase. The non-cysteinyl sulfur content, determined as 8.1 acid-labile sulfide eq/mol, is presumably associated with nonheme iron to form iron-sulfur centers. We conclude that N. crassa nitrite reductase is a homodimer of large molecular weight subunits housing an electron transfer complex of FAD, iron-sulfur centers, and siroheme to mediate the reduced pyridine nucleotide-dependent reduction of nitrite to ammonia.     

Identification of an iron uptake system specific for coprogen and rhodotorulic acid in Escherichia coli K12

With the lac operon fusion technique, mutants were isolated in two genes that specify two outer membrane proteins designated FhuE (76 K) and Fiu (83 K). The synthesis of both proteins was increased under low iron growth conditions. The FhuE-protein was shown to be necessary for iron uptake via coprogen, an iron chelator produced by certain fungi, e.g. Neurospora crassa. In addition to fhueE the genes fhuCDB, tonB and exbB were necessary for iron coprogen uptake. The gene fhuE was mapped between kdp and gltA near 16 min on the genetic map of E. coli K12, while gene fiu was mapped near 18 min between chlA and chlE. Nor iron transport system could be assigned as yet to the Fiu protein.     

Iron release from ferrisiderophores. A multi-step mechanism involving a NADH/FMN oxidoreductase and a chemical reduction by FMNH2.

Release of iron from various ferrisiderophores (ferripyoverdines, ferrioxamines B and E, ferricrocin, ferrichrome A, ferrienterobactin and its analog ferric N,N',N'-tri(1,3,5-Tris) 2,3-dihydroxybenzoylaminomethylbenzene) was obtained through an enzymic reduction of iron, involving NADH, FMN and the ferripyoverdine reductase of Pseudomonas aeruginosa PAO1. The iron released from the same complexes was also obtained through chemical reduction of iron involving FMNH2. Evidence is given that the enzymic process acts through a FMNH2 reduction; the P. aeruginosa enzyme, purified according to its ferripyoverdine-reductase activity [Hallé, F. & Meyer, J. M., Eur. J. Biochem. 209, 613-620], functions as a NADH:FMN oxidoreductase, the FMNH2 produced being able to chemically reduce the iron complexed by siderophores. The general occurrence of such a multi-step mechanism, which denies the existence of specific ferrisiderophore reductases, is discussed.     

Iron acquisition by plants: the reduction of ferrisiderophores by higher plant NADH: Nitrate reductase

Siderophores are avid Fe³⁺-chelators of microbial origin. Plant roots are colonized by fungi and bacteria which synthesize siderophores, and plants have been shown to metabolize these substances to obtain iron. We have previously shown that nitrate reductase from squash catalyzed the reduction of the ferrisiderophore ferrioxamine B with the subsequent loss of Fe²⁺. Using a spectrophotometric assay which traps Fe²⁺ in a ferrozine complex, we have noted that the substrate diversity of nitrate reductase as a ferrisiderophore reductase includes ferrichrome A, ferrichrome, ferrirhodotorulic acid, ferrischizokinen, and the novel siderophore ferri-‘AAHS’. These reductions were inhibited by polyclonal antibodies against nitrate reductase, but ferrisiderophore reductase activity, as evidenced with ferrirhodotorulic acid, was unaffected by low concentrations of azide. In addition, maximal activity occurred between pH 4 and 5, and appaarent K_m values were approx. 100 μmolar. Thus, we suggest that plant nitrate reductases might be involved in iron assimilation as well as nitrate reduction.     

3-Hydroxy-3-methylglutaryl CoA reductase and mevalonate kinase of Neurospora crassa

Two enzymes of polyisoprenoid synthesis, 3-hydroxy-3-methylglutaryl coenzyme A (HMG CoA) reductase (mevalonate:NADP oxidoreductase [acylating CoA], EC 1.1.1.34) and mevalonate kinase (ATP:mevalonate 5-phosphotransferase, EC 2.7.1.36), are present in the microsomal and soluble fractions of Neurospora crassa, respectively. HMG CoA reductase specifically uses NADPH as reductant and has a K(m) for dl-HMG CoA of 30 micro M. The activities of HMG CoA reductase and mevalonate kinase are low in conidia and increase threefold during the first 12 hr of stationary growth. Maximum specific activities of both enzymes occur when aerial hyphae and conidia first appear (2 days), but total activities peak later (3-4 days). Addition to the growth media of ergosterol or beta-carotene, alone or in combination, does not affect the specific or total activity of either enzyme. The mevalonate kinase of N. crassa, purified 200-fold to a specific activity of 5 micro moles/min/mg, is free from HMG CoA reductase, phosphomevalonate kinase, ATPase, adenylate kinase, and NADH oxidase activities. Mevalonate kinase specifically requires ATP as cosubstrate and exhibits a marked preference for Mg(2+) over Mn(2+), especially at high ratios of divalent metal ion to ATP. Kinase activity is inhibited by p-hydroxymercuribenzoate, and this inhibition is partially prevented by mevalonate or MgATP. Optimum activity occurs at pH 8.0-8.5 and at about 55 degrees C. The Neurospora kinase, like that of hog liver, has a sequential mechanism for substrate addition. The Michaelis constants obtained were 2.8 mM for dl-mevalonate and 1.8 mM for MgATP(-2). Geranyl pyrophosphate is an inhibitor competitive with MgATP (K(i) = 0.11 mM).     

Iron reductases from Pseudomonas aeruginosa

Cell-free extracts of Pseudomonas aeruginosa contain enzyme activities which reduce Fe(III) to Fe(II) when iron is provided in certain chelates, but not when the iron is uncomplexed. Iron reductase activities for two substrates, ferripyochelin and ferric citrate, appear to be separate enzymes because of differences in heat stabilities, in locations in fractions of cell-free extracts, in reductant specificity, and in apparent sizes during gel filtration chromatography. Ferric citrate iron reductase is an extremely labile activity found in the cytoplasmic fraction, and ferripyochelin iron reductase is a more stable activity found in the periplasmic as well as cytoplasmic fraction of extracts. A small amount of activity detectable in the membrane fraction seemed to be loosely associated with the membranes. Although both enzymes have highest activity reduced nicotinamide adenine dinucleotide, reduced glutathione also worked with ferripyochelin iron reductase. In addition, oxygen caused an irreversible loss of a percentage of the ferripyochelin iron reductase following sparge of reaction mixtures, whereas the reductase for ferric citrate was not appreciably affected by oxygen.     

Identification of siderophores and siderophore-mediated uptake of iron in Stemphylium botryosum

Under iron-deficient conditions Stemphylium botryosum f. sp. lycopersici produces three major siderophores; dimerum acid, coprogen B and an unidentified monohydroxamate siderophore designated as A. The system of siderophores mediating uptake of iron was characterized. It exhibits active transport, saturation kinetics and an optimum at pH 6 and 30°. The rate of iron uptake via dimerum acid and coprogen B was four times higher than siderophore A. S. botryosum was capable of taking up iron from hydroxamate siderophores produced by other fungi, e.g. ferrichrome, fusigen, rhodotorulic acid but not ferrioxamine B. Double labelling experiments suggest that ferric coprogen B accumulates in mycelial cells as an intact chelate.