David B. Archer

The Molecular Biology of Secreted Enzyme Production by Fungi

Enzymes from filamentous fungi are already widely exploited, but new applications for known enzymes and new enzymic activities continue to be found. In addition, enzymes from less amenable non-fungal sources require heterologous production and fungi are being used as the production hosts. In each case there is a need to improve production and to ensure quality of product. While conventional, mutagenesis-based, strain improvement methods will continue to be applied to enzyme production from filamentous fungi the application of recombinant DNA techniques is beginning to reveal important information on the molecular basis of fungal enzyme production and this knowledge is now being applied both in the laboratory and commercially. We review the current state of knowledge on the molecular basis of enzyme production by filamentous fungi. We focus on transcriptional and post-transcriptional regulation of protein production, the transit of proteins through the secretory pathway and the structure of the proteins produced including glycosylation.

Solution structure of the granular starch binding domain of Aspergillus niger glucoamylase bound to β-cyclodextrin

BACKGROUNDnCarbohydrate-binding domains are usually small and physically separate from the catalytic domains of hydrolytic enzymes. Glucoamylase 1 (G1) from Aspergillus niger, an enzyme used widely in the food and brewing industries, contains a granular starch binding domain (SBD) which is separated from the catalytic domain by a semi-rigid linker. The aim of this study was to determine how the SBD binds to starch, and thereby more generally to throw light on the role of carbohydrate-binding domains in the hydrolysis of insoluble polysaccharides.nnnRESULTSnThe solution structure of the SBD of A. niger G1 bound to beta-cyclodextrin (betaCD), a cyclic starch analogue, shows that the well-defined beta-sheet structure seen in the free SBD is maintained in the SBD-betaCD complex. The main differences between the free and bound states of the SBD are observed in loop regions, in or near the two starch-binding sites. The two binding sites, each of which binds one molecule of betaCD, are structurally different. Binding site 1 is small and accessible, and its structure changes very little upon ligand binding. Site 2 is longer and undergoes a significant structural change on binding. Part of this site comprises a flexible loop, which appears to allow the SBD to bind to starch strands in a range of orientations.nnnCONCLUSIONSnThe two starch-binding sites of the SBD probably differ functionally as well as structurally; site 1 probably acts as the initial starch recognition site, whereas site 2 is involved in specific recognition of appropriate regions of starch. The two starch strands are bound at approximately 90 degrees to each other. This may be functionally important, as it may force starch strands apart thus increasing the hydrolyzable surface, or alternatively it may localize the enzyme to noncrystalline (more hydrolyzable) areas of starch. The region of the SBD where the linker to the catalytic domain is attached is flexible, allowing the catalytic site to access a large surface area of the starch granules.

Glucoamylase : green fluorescent protein fusions to monitor protein secretion in Aspergillus niger

A glucoamylase::green fluorescent protein fusion (GLA::sGFP) was constructed which allows the green fluorescent protein to be used as an in vivo reporter of protein secretion in Aspergillus niger. Two secretory fusions were designed for secretion of GLA::sGFP which employed slightly different lengths of the glucoamylase protein (GLA499 and GLA514). Expression of GLA::sGFP revealed that fluorescence was localized in the hyphal cell walls and septa, and that fluorescence was most intense at hyphal apices. Extracellular GLA::sGFP was detectable by Western blotting only in the supernatant of young cultures grown in soya milk medium. In older cultures, acidification of the medium and induction of proteases were probably responsible for the loss of extracellular and cell wall fluorescence and the inability to detect GLA::sGFP by Western analysis. A strain containing the GLA::sGFP construct was subjected to UV mutagenesis and survivors screened for mutations in the general secretory pathway. Three mutants were isolated that were unable to form a halo on either starch or gelatin medium. All three mutants grew poorly compared to the parental strain. Fluorescence microscopy revealed that for two of the mutants, GLA::sGFP accumulated intracellularly with no evidence of wall fluorescence, whereas for the third mutant, wall fluorescence was observed with no evidence of intracellular accumulation. These results indicate that the GLA::sGFP fusion constructs can be used as convenient fluorescent markers to study the dynamics of protein secretion in vivo and as a tool in the isolation of mutants in the general secretory pathway.

A glucoamylase::GFP gene fusion to study protein secretion by individual hyphae of Aspergillus niger.

Although Aspergillus niger is used as a host for heterologous protein production, yields are generally lower than those obtained for homologous proteins. Mechanisms of protein secretion and the secretory pathway in filamentous fungi are poorly characterised, although there is evidence to suggest that secretion occurs by a mechanism similar to that in other eukaryotes, but with proteins destined for secretion being directed to the hyphal tip. We report on a method using a glucoamylase: GFP gene fusion which allows us for the first time to monitor, in vivo, protein secretion in A. niger at the single hyphal level. A synthetic green fluorescent protein (sGFP(S65T)) was fused to truncated A. niger glucoamylase (GLA:499). Southern blot analysis of transformants confirmed that the gene fusion had successfully integrated into the A. niger genome. Confocal and fluorescence microscopy revealed that the GLA::GFP fusion protein is fluorescent in A. niger and appears to be directed to the hyphal tip. In young mycelia, hyphal cell wall fluorescence is apparent and immunogold labelling of GFP confirmed that GFP was partially localised within the hyphal cell wall. Using Western blotting, extracellular GLA::GFP was detected only in culture filtrates of young mycelia grown in a soya milk medium. The actin inhibitor latrunculin B was used to disrupt the secretion process, and its effects on the distribution of GLA::GFP were monitored.

Plant low-molecular-weight phospholipase A2s (PLA2s) are structurally related to the animal secretory PLA2s and are present as a family of isoforms in rice (Oryza sativa)

Recently, we purified to homogeneity and characterized a low-molecular-weight calcium-dependent phospholipase A2 (PLA2) from developing elm seed endosperm. This represented the first purified and characterized PLA2 from a plant tissue. The full sequences of two distinct but homologous rice (Oryza sativa) cDNAs are given here. These encode mature proteins of 119 amino acids (PLA2-I, preceded by a 19 amino acid signal peptide) and 128 amino acids (PLA2-II, preceded by a 25 amino acid signal peptide), and were derived from four expressed sequence tag (EST) clones. Both proteins were homologous to the N-terminal amino acid sequence of the elm PLA2. They contained twelve conserved cysteine residues and sequences that are likely to represent the Ca2+-binding loop and active-site motif, which are characteristic of animal secretory PLA2s. A soluble PLA2 activity was purified 145xa0000-fold from green rice shoots. This had the same biochemical characteristics as the elm and animal secretory PLA2s. The purified rice PLA2 consisted of two proteins, with a molecular weight of 12xa0440 and 12xa0920, that had identical N-terminal amino acid sequences. This sequence was different from but homologous to the PLA2-I and PLA2-II sequences. Taken together, the results suggest that at least three different low-molecular-weight PLA2s are expressed in green rice shoots. Southern blot analysis suggested that multiple copies of such genes are likely to occur in the rice and in other plant genomes.

A defined level of protein disulfide isomerase expression is required for optimal secretion of thaumatin by Aspergillus awamori

Abstract. Thaumatin, a 22-kDa protein containing eight disulfide bonds, is secreted by the filamentous fungus Aspergillus awamori at levels which are dependent upon the extent of overexpression of protein disulfide isomerase (PDIA). Additional copies of the PDIA-encoding gene pdiA were introduced into a strain of A. awamori that expresses a cassette encoding thaumatin. Transformants with different levels of pdiA mRNA and measured PDIA levels were chosen for examination of the impact that PDIA levels had on thaumatin secretion. The secretion of two native proteins, α-amylase and acid phosphatase, was also examined in relation to varying levels of PDIA. Over a range of PDIA levels of 1–8, relative to the native level in strains with just one copy of the pdiA gene, the fraction of α-amylase and acid phosphatase in the total secreted protein was unaffected. In contrast, a peak level of thaumatin, about 5-fold higher than in the strain with one copy of pdiA, was found in strains with a relative PDIA level of between two and four. Improved thaumatin production was confirmed in 5-l fermenters using a strain of A. awamori with six pdiA gene copies, containing 3.2-fold higher levels of PDIA than wild-type strains.

Two fatty acid 9-desaturase genes, ole1 and ole2, from Mortierella alpina complement the yeast ole1 mutation

Genes encoding two distinct fatty acid delta9-desaturases were isolated from strains of the oleaginous fungus Mortierella alpina. Two genomic sequences, delta9-1 and delta9-2, each containing a single intron, were cloned from strain CBS 528.72 while one cDNA clone, LM9, was isolated from strain CBS 210.32. The delta9-1 gene encoded a protein of 445 aa which shared 99% identity with the LM9 gene product. These proteins also showed 40-60% identity to the delta9-desaturases (Ole1p) of other fungi and contained the three conserved histidine boxes, C-terminal cytochrome b5 fusion and transmembrane domains characteristic of endoplasmic reticulum membrane-bound delta9-desaturases. LM9 and delta9-1 are therefore considered to represent the same gene (ole1). The ole1 gene was transcriptionally active in all M. alpina strains tested and its function was confirmed by complementation of the Saccharomyces cerevisiae ole1 mutation. Fatty acid analysis of yeast transformants expressing the CBS 210.32 ole1 gene showed an elevated level of oleic acid (18:1) compared to palmitoleic acid (16:1), the major fatty acid component of wild-type S. cerevisiae. This indicated that the M. alpina delta9-desaturase had a substrate preference for stearic acid (18:0) rather than palmitic acid (16:0). Genomic clone delta9-2 (ole2) also encoded a protein of 445 aa which had 86% identity to the delta9-1 and LM9 proteins and whose ORF also complemented the yeast ole1 mutation. The transcript from this gene could only be detected in one of the six M. alpina strains tested, suggesting that its expression may be strain-specific or induced under certain physiological conditions.

Effect of pH on hen egg white lysozyme production and evolution of a recombinant strain of Aspergillus niger.

An Aspergillus niger strain (B1) transformed to produce mature hen egg white lysozyme (HEWL) from a glucoamylase fusion protein under control of the A. niger glucoamylase promoter was grown in glucose-limited chemostat culture at a dilution rate of 0.07 h-1 at various pH values. Maximum HEWL production (9.3 mg g-1; specific production rate = 0.65 mg g-1 per h) was obtained at pH 4.5. However, in chemostat culture, HEWL production was not stable at any pH tested. After 240 h in steady state, specific production decreased to only 0.03 +/- 0.01 and 0.24 +/- 0.02 mg g-1 per h at pH 6.5 and 4.5, respectively. Some isolates removed from the chemostat cultures had lost copies of the HEWL gene and when grown in shake flask cultures all of the isolates produced less HEWL than the parental strain. Morphological mutants with similar phenotypes were isolated at all pHs, but their rate of increase in the population was pH dependent, with cultures at low pH (< 4.5) being more morphologically stable than cultures at high (> 4.5) pH. The selective advantage of these mutants was also generally dependent on pH. Both yellow pigment producing mutants and brown sporulation mutants had higher selective advantages over the parental strain at high than at low pH, regardless of the pH at which they were isolated. However, the selective advantage of densely sporulating mutants was independent of pH.

Aberrant processing of wild-type and mutant bovine pancreatic trypsin inhibitor secreted by Aspergillus niger.

Bovine pancreatic trypsin inhibitor (BPTI) was secreted by Aspergillus niger at yields of up to 23 mg l-1 using a protein fusion strategy. BPTI was linked to part of the fungal glucoamylase protein (GAM) with a dibasic amino acid (KEX2) processing site at the fusion junction. Electrospray ionisation mass spectrometry and N-terminal protein sequencing revealed that, although biologically active in vitro, the purified products from a number of independent transformants consisted of a mixture of BPTI molecules differing at the N-terminus. Approximately 35-60% of this mixture was processed correctly. Aberrant processing of the GAM-BPTI fusion protein by the A. niger KEX2-like endoprotease was the most likely cause of this variation although the involvement of other fungal endoproteases could not be ruled out. In vitro studies have highlighted a weak interaction between BPTI and the Saccharomyces cerevisiae KEX2 endoprotease, suggesting that BPTI is not a potent inhibitor of KEX2p. A small proportion of the recombinant BPTI (10%) showed nicking of the K15-A16 bond, indicating an interaction with a fungal trypsin-like enzyme. Mutant BPTI homologues designed to have anti-elastase activity, BPTI(K15V), BPTI(K15V,P13I) and BPTI(K15V,G12A), have also been expressed and secreted by A. niger. They also showed a similar spectrum of aberrant N-terminal processing but no nicking of the K15-V16 bond was observed. Comparison of A. niger with other expression systems showed that it is an effective system for producing BPTI and its homologues, although not all molecules were correctly processed. This variation in processing efficiency may be useful in understanding the important determinants of protein processing in this fungus.

Molecular basis of glucoamylase overproduction by a mutagenised industrial strain of Aspergillus niger.

We have compared a mutagenized strain of Aspergillus niger (S1), used industrially for glucoamylase production, and a related low glucoamylase-producing strain (S2) with a laboratory strain of A. niger (AB4.1). Our aim was to assess the properties of S1 in relation to the laboratory strain and to account at the molecular level for the basis of its glucoamylase overproduction. Both S1 and S2 have similar multiple copies of the glucoamylase-encoding gene (glaA) but only S1 has enhanced glaA transcript and glucoamylase levels compared to AB4.1 that has a single copy of the glaA gene. Glucoamylase production by S1 and AB4.1 was repressed by xylose and induced by starch but, in S2, remained unaffected by carbon source. S1 also secreted elevated levels of alpha-amylase relative to both S2 and AB4.1 but the production of alpha-glucosidase was low in all three strains. The gene encoding aspergillopepsin (pepA), an abundant secreted aspartyl protease, was present as a single copy in all strains but no aspergillopepsin could be detected by Western blotting in either S1 or S2 culture supernatants. We conclude that A. niger strain improvement by mutagenesis and screening for glucoamylase overproduction has led to glaA gene multiplication and an expression defect in the pepA gene.