Gluconate formation and polyol metabolism in Aspergillus niger

The capacity of A.niger to accumulate metabolites is remarkable. Under all conditions polyols accumulate in the cell and when mycelium in later developmental stages is considered, depending on the carbon source, aeration and external pH, polyols and/or organic acids can be formed in a very efficient way. The aim of this thesis was to obtain a better understanding of the mechanisms governing the metabolism and formation of these metabolites. The first part of this thesis reports a study of gluconic acid formation and the second part involves polyol metabolism in A.niger.

The fungus has a general tendency to synthesize organic acids under conditions of good aeration and when sugars like D-glucose or sucrose are available. An important function of the organic acid synthesis might be the acidification of the medium. Combined with the removal of the sugars from the environment this may contribute to the competitiveness of the fungus. A.niger is tolerant to pH values as low as 1.5. When glucose is the carbon source and the culture is well aerated A.niger will produce at neutral or slightly acidic pH mainly gluconate, at very low pH values citrate formation will occur and at intermediate pH oxalate is formed. The result is that A.niger by the consecutive synthesis of a series of organic acids strongly acidifies the medium.

In this thesis the formation of gluconic acid was studied in more detail. In Chapter 2 evidence is provided for a cell wall localization of glucose oxidase. Furthermore, it was shown that two catalases are induced in parallel with glucose oxidase, one intracellular (CAT III) and one localized in the cell wall (CAT IV). Two other catalases, also one intracellular (CAT I) and one localized in the cell wall (CAT II), are constitutively present. About 50% of the lactonase activity was measured in the culture fluid. Therefore it was concluded that the whole glucose oxidase system is localized extracellularly, and is mainly localized in the cell wail. The induction of the intracellular CAT Ill may be seen as a second defense barrier, detoxifying hydrogen peroxide that diffuses into the cell.

The cell wall-localization of glucose oxidase combined with the easy detection of hydrogen peroxide produced in the enzyme reaction makes visualization of the enzyme in intact hyphae possible. This detection system was used to isolate a series of mutants with altered glucose oxidase induction. In Chapter 3 a phenotypic and genetic characterisation of these mutants is presented. The mutants were classified in 9 different complementation groups, 1 non-producing, 1 low producing and 7 overproducing mutants. From induction experiments with the wild type strain it was concluded that the carbon source and the dissolved oxygen level are main factors determining induction of glucose oxidase. One mutant was found which never synthesized glucose oxidase ( gox C). Only one of the mutant classes was no longer dependent on oxygen for induction ( gox B). Several mutant groups were found with a decreased glucose dependency of induction. Some of these were quite strong ( gox A and gox E) whereas others showed only a minor overproduction under conditions which are only weakly inducing in the wild type. The latter group might in fact influence glucose oxidase induction only indirectly. The genetic analysis provided the information necessary for the construction of recombinant strains containing different gox mutations and other genetic markers. This is essential for further analysis and also an important step in further strain breeding.

In Chapter 4 the induction mechanism of glucose oxidase, lactonase and the catalases was analyzed in more detail. For this we used beside the wild type strain gox B, gox C and gox E mutants. These mutants had a clear and pronounced phenotype. It was shown that in a wild type strain induction of all three activities is found only when glucose is present and the culture is well aerated. Induction of all three activities was effected by the gox B mutation. Neither glucose nor high oxygen levels were required for induction. The glucose dependency of glucose oxidase and lactonase induction was affected by the gox E mutation. In this mutant catalase was unaffected and high oxygen was still required. Thus with the gox B and gox E mutants the effects which oxygen and glucose have on induction could be partly separated. None of the activities was induced in the gox C mutant. This mutant could be transformed to a wild type phenotype using the structural gene of glucose oxidase, thus indicating that the mutation concerns the structural gene. The glucose oxidase structural gene was isolated via antibody screening of a cDNA expression library and subsequent homologous screening of a genomic library using the cDNA clone as a probe. Northern blotting showed that both the oxygen and carbon source effects were influencing induction at the transcriptional level. No glucose oxidase mRNA was observed in the gox C mutant. The absence of induction of all three activities in this mutant indicates that glucose oxidase activity is required for induction. It could be shown that hydrogen peroxide, besides gluconate a product of the glucose oxidase reaction, is inducing both catalase and lactonase in this mutant. It was concluded that hydrogen peroxide is the main factor required for the induction of the three activities and that the gox B gene product is involved in mediating this effect. It explained the requirement of high oxygen levels for induction because glucose oxidase has a high K _m for oxygen (K _m =0.48 mM at 27°C, Gibson etal., 1964). Even the glucose requirement for induction can be explained this way but the presence of the gox E mutant and the fact that the carbon source was still affecting the level of glucose oxidase in the gox B mutant, which is supposed to be involved in the transduction of the hydrogen peroxide signal, indicates that the induction process is more complicated. Using the plasmid plM503 which carries the structural gene for glucose oxidase multicopy transformants were made. Only a relatively small increase in glucose oxidase activity was observed (3 fold) even though more than 50 copies of the gene were integrated. Transformation of A.nidulans, a fungus that does not have the glucose oxidase gene itself, resulted in strains which produced glucose oxidase. In these strains glucose but not a high oxygen concentration was required for induction. This is indicating that a gox B-like gene is not present in A.nidulans but that the glucose requirement is presumably transduced by a more general system which is not specific for glucose oxidase.

Chapters 2,3 and 4 are contributions to a better understanding of how the glucose oxidation system functions and of the molecular mechanism of its induction. Thus far it is the only coordinately regulated enzyme system in A.niger of which regulatory mutants have been isolated and which has been worked out in some detail. However, the understanding of the mechanism is still incomplete, especially the way by which the carbon source is affecting the induction is still unclear. No explanation is yet available why on fructose or D-xylose a basal level of glucose oxidase is found, whereas on acetate, gluconate or glycerol no activity is detected. Somehow the system senses the presence of an easy metabolisable carbon source and low-level induction occurs. The mechanism behind this phenomenon is probably not glucose oxidase-specific since in the A.nidulans transformants still glucose is required for induction. The gox B system is probably more glucose oxidase specific and therefore no oxygen (H ₂ O ₂ ) effect is found in A.nidulans. For a better understanding of the factors involved in glucose oxidase expression, a detailed analysis of the gox A and gox E mutants and of the functioning of the promoter of glucose oxidase is required. An important consequence of the regulatory mechanism hypothesized in Chapter 4 is a build-in feedback control. Catalase is not only induced by hydrogen peroxide but degrades this inducer as well, thus diminishing the induction. Ever increasing amounts of especially oxygen will be required to cause induction to continue. This will not happen because oxygen is quite soon the limiting factor in gluconate fermentation processes. The feedback mechanism of preventing overinduction is absent in gox B mutants. Therefore these mutants might be valuable in industrial processes.

The function of the polyols is different from that of the organic acids. This is already clear from their presence in the fungus during all phases of the life cycle. Organic acids are formed only in late stages of development. Furthermore, the organic acids are excreted whereas the polyols beside being excreted also accumulate in large amounts in the mycelium. Information on carbon metabolism and more specifically polyol metabolism in A.niger was scarce at the start of this project. Therefore it was decided to analyze some metabolic pathways which directly relate to polyol metabolism.

In Chapter 5 the characterization of a glycerol kinase mutant is described. Glycerol is one of the main polyols accumulating in A.niger and it was shown that glycerol kinase is involved in the degradation of glycerol. It could be demonstrated that the degradation pathway of glycerol in A.niger is largely the same as in A.nidulans (Hondmann et al., 1990) and N.crassa (Courtright 1975). First phosphorylation to glycerol-3-phosphate occurs and this is followed by its oxidation to dihydroxyacetonephosphate by a mitochondrial FAD-dependent glycerol-3- phosphate dehydrogenase. However, there were some differences with the pathway in A.nidulans. Whereas in the latter fungus dihydroxyacetone was catabolized via glycerol, in A.niger a dihydroxyacetone kinase was present enabling growth of the glycerol kinase mutant on dihydroxyacetone. Combined with an NAD ⁺-dependent glycerol dehydrogenase converting glycerol into dihydroxyacetone this formed an escape route for glycerol catabolism in the glycerol kinase mutant. Growth on D-galacturonate was strongly affected in the glycerol kinase mutant thus demonstrating a D-galacturonate degradation pathway via glycerol.

Pentose metabolism is described in Chapter 6. The isolation of a D-xylulose kinase mutant played an essential role in this work. It could be demonstrated that L-arabinose and D-xylose are catabolized via a series of reduction and oxidation steps. L-arabinose is reduced to L-arabitol which is oxidized to L-xylulose. L-xylulose is reduced to xylitol which is oxidized to D-xylulose that is phosphorylated to D-xylulose-5-phosphate, an intermediate of the pentose phosphate pathway. D-xylose is reduced to xylitol and subsequently oxidized to D-xylulose-5-phosphate. All the reduction steps are NADPH-dependent and all the oxidation steps NAD ⁺-dependent The equilibrium of the reactions is far in the direction of the polyols, so several unfavourable steps that have to be taken which potentially can obstruct an efficient conversion of L-arabinose to D-xylulose-5- phosphate. The cofactor specificity of the dehydrogenases involved contributes to a higher efficiency since the anabolic reduction charge ([NADPH]/([NADPH]+[NADP ⁺])) is higher than the catabolic reduction charge ([NADH]/([NADH]+[NAD ⁺])) (Führer et al., 1980). A second mechanism for increasing the efficiency of this pathway was found by studying the two xylitol dehydrogenases of the L-arabinose pathway. This is described in Chapter 7. The NADPH-dependent L-xylulose reductase, catalyzing the reduction of L- xylulose to xylitol, was purified and the NAD ⁺-dependent xylitol dehydrogenase, catalyzing the oxidation of xylitol to D-xylulose, was partially purified. Comparison of the two enzymes, which catalyze similar reactions leading to the different stereoisomers, made clear that they differ in two major points. 1) When their affinity for xylitol was compared it was found that the NAD ⁺-dependent enzyme had a much higher affinity for xylitol than the NADPH-dependent enzyme. 2) Near the physiological pH (around 7) the ratio of the rductive relative to the oxidative catalytic activity was higher for the L-xylulose reductase than for the xylitol dehydrogenase. Both characteristics contribute to a more efficient catalysis of the reaction in the in vivo direction.

In Chapter 8 an attempt is made to obtain some information on the function of the various polyol pools in A.niger. It was found that glycerol was the main polyol involved in osmotic adjustment in the fungus. Furthermore, glycerol accumulation was observed to be related to fast growing hyphae, whereas mannitol and erythritol accumulated in older hyphae. Mannitol also was an important storage compound in conidiospores. This general scheme of polyol accumulation during different growth phases is modified by environmental parameters like aeration of the culture and the nitrogen source available. Changes in fluxes through metabolic pathways and as a result of that changes in the steady state concentrations of intermediary metabolites from which the polyols derive, presumably play a role in this. It also implies that the function of the polyols in the cell is not completely coupled to specific polyols but can, in part, be taken over by other polyols. It was observed that a considerable part of the accumulated polyols (>50% after 24 h) is found in the medium. Observations made with the glycerol kinase mutant suggested that polyol excretion is a way for the fungus to control the intracellular levels of the polyols, presumably for maintaining the osmotic balance of the cell. Long fermentations (5 days) showed that in late developmental stages the polyol excretion becomes more pronounced. Approximately 45% of the glucose taken up was converted into extracelluar polyols. The type of the polyols excreted was a reflection of the intracellular polyol pool composition.

Chapters 5 and 6 describe glycerol and pentose catabolism which has led to a better understanding of carbon metabolism in A.niger. Information on this subject is scarce in this fungus. The isolation of mutants in the degradation of such compounds is quite essential for metabolic studies. It turned out to be very difficult to isolate mutants in carbon metabolism in A.niger and still only a few of these mutants are available now. The reason for this is not known but the high intracellular polyol pools and the excretion of the polyols, resulting in crossfeeding during the filtration enrichment techniques, might play a role in this. The analysis of pentose metabolism and the xylitol dehydrogenases involved is also valuable in the light of understanding the mechanisms that play a role in extracellular enzyme production. The knowledge of the L-arabinose catabolic route has already proven useful in the analysis of the araban degrading system of A.niger. L-arabitol plays a major role in this (vd Veen et al., 1993). The analysis of polyol accumulation in Chapter 8 indicates that different polyols accumulate in different parts of the hyphae depending on their age. There still remains the question whether specific polyols accumulate in specific compartments of the cell, for example in the vacuoles, which are abundant in older hyphae. No information is available on this. It was shown that polyol excretion is a general phenomenon in A.niger. Although it has been observed before (Röhr et al., 1987), it was not considered to be such a common phenomenon in this fungus. The observation of large scale polyol accumulation in the culture fluid of 3-5 days old mycelial cultures grown under low oxygen conditions suggests similarities with organic acid fermentations. These are performed in strongly aerated cultures. An efficient flux to the TCA cycle is apparently not possible under the low aeration conditions used. This results in overflow metabolism in an earlier stage of the catabolic pathway leading to polyol formation. Cofactor regeneration might as well play a role in the polyol accumulation under these conditions.