Double knockout of two target genes via genome co-editing using a nitrate transporter gene nrtA and a putative thiamine transporter gene thiI as selectable markers in Aspergillus oryzae.

Genome co-editing technology is effective in breeding filamentous fungi for applications in the fermentation industry, achieving site-directed mutagenesis, the status of non-genetically modified organisms (non-GMOs), and wild-type-like growth phenotype. Prior to this study, thiI gene was found as a selectable marker for such genome co-editing in the filamentous fungus Aspergillus oryzae, while it cannot be reused via marker recycling. Therefore, we aimed to identify another marker gene to knock out another target gene via genome co-editing in A. oryzae. In this study, we focused on the membrane transporter gene nrtA (AO090012000623), which promotes uptake of nitrate (NO3-). It is known that, in nrtA knockout strain, chlorate (ClO3-), an analog of nitrate with antifungal activity, cannot be imported into the cytosol, which enables the mutant to grow in the presence of chlorate. Based on this information, knockout of the target gene wA was attempted using both nrtA- and wA-specific single-guide RNAs via genome co-editing with KClO3 supplementation in A. oryzae laboratory strain RIB40 and industrial strain KBN616. Resultantly, wA knockout mutant was generated, and nrtA was identified as a selectable marker. Moreover, this genome co-editing system using nrtA was compatible with that using thiI, and thus, a double knockout mutant of two target genes wA and yA was constructed in RIB40 while maintaining non-GMO status and wild-type-like growth. As nrtA homologs have been found in several industrial Aspergillus species, genome co-editing using homolog genes as selectable markers is plausible, which would contribute to the widespread breeding of industrial strains of Aspergilli.

Identification of cheese rancidity-related lipases in Aspergillus oryzae AHU 7139.

The adjunct product with enzymatic activity from Aspergillus oryzae is beneficial for flavor enrichment in the ripened cheese. However, an excessive lipolytic reaction leads to the release of volatile free fatty acids. Accordingly, a strong off-flavor (i.e., rancidity) has been detected when A. oryzae AHU 7139 is used. To identify the rancidity-related lipase from this strain, we evaluated the substrate specificity and lipase distribution using five mutants cultured on a whey-based solid medium under different initial pH conditions. The results showed a higher diacylglycerol lipase activity than triacylglycerol lipase activity. Moreover, an initial pH of 6.5 for the culture resulted in higher lipolytic activity than a pH of 4.0, and most of the activity was found in the extracellular fraction. Based on the gene expression analysis by real-time polymerase chain reaction and location and substrate specificity, five genes (No. 1, No. 19, mdlB, tglA, and cutL) were selected among 25 annotated lipase genes to identify the respective knockout strains. Because ΔtglA and ΔmdlB showed an outstanding involvement in the release of free fatty acids, these strains were applied to in vitro cheese curd experiments. In conclusion, we posit that triacylglycerol lipase (TglA) plays a key role as the trigger of rancidity and the resulting diglycerides have to be exposed to diacylglycerol lipase (MdlB) to stimulate rancidity in cheese made with A. oryzae AHU 7139. This finding could help screen suitable A.oryzae strains as cheese adjuncts to prevent the generation of the rancid-off flavor.

Overexpression of a predicted transketolase gene and disruption of an α-1,3-glucan synthase gene in Aspergillus oryzae DGLA3 strain enhances the yield of free dihomo-γ-linolenic acid.

Free dihomo-γ-linolenic acid (DGLA), a polyunsaturated free fatty acid (FFA), can potentially be used to produce eicosanoid pharmaceuticals, such as prostaglandin E1. Previously, we constructed an Aspergillus oryzae mutant strain, named DGLA3, which produced free DGLA at an increased yield by faaA gene disruption and cooverexpression of one elongase and two desaturase genes. In this study, we achieved a further increase. Since FFA production is increased by enhancing the pentose phosphate pathway, we overexpressed a predicted transketolase gene composing the pathway in DGLA3, which consequently increased the free DGLA yield by 1.9-fold to 403 mg/L. Additionally, we disrupted the α-1,3-glucan synthase gene agsB involved in cell-wall biosynthesis, which further increased it by 1.3-fold to 533 mg/L. Overall, the yield increased by 2.5-fold. Free DGLA productivity and biomass increased similarly, but residual glucose concentration decreased. Increased hyphal dispersion appeared to cause additional glucose consumption, resulting in an increase in biomass and yield.

Concomitant knockout of target and transporter genes in filamentous fungi by genome co-editing.

In most countries, genetically modified microorganisms are not approved for use for fermentation in the food industry. Therefore, random mutagenesis and subsequent screening are performed to improve the productivities of valuable metabolites and enzymes as well as other specific functions in an industrial microbial strain. In addition, targeted gene knockout is performed by genetic recombination using its enzyme genes as selectable markers to maintain self-cloning status. However, random mutagenesis has a drawback as it does not guarantee improvement of the targeted function. Conversely, self-cloning is rarely used to breed an industrial microbial strain. This is probably because a self-cloning strain is similar to a genetically modified strain, as both undergo homologous recombination, although exogenous genes are not introduced. In this article, I discuss the usefulness of genome editing technology as a substitute for conventional techniques to breed filamentous fungal strains. This article particularly focusses on "genome co-editing," a genome editing technology used for knocking out two genes concomitantly, as reported in Magnaporthe grisea and Aspergillus oryzae. Especially, when genome co-editing is applied to a target gene and a membrane transporter gene that aid the entry of toxic compounds into cells, the resulting clone can be categorized as an autotrophic and non-genetically modified clone. Such a clone should easily apply to industrial fermentation without being restricted by a genetically modified status. Genome co-editing will also be used to construct mutant strains with multiple target gene knockouts by eliminating multiple membrane transporter genes. This could substantially improve the productivities of valuable metabolites and enzymes in a stepwise manner. Thus, genome co-editing is considered a potentially powerful method to knock out single or multiple target genes that can contribute to the breeding of filamentous fungal strains in the food industry.

A study of lipolysis induced by adjuncts from edible Aspergillus sp. solid culture products on ripened semi-hard cheese.

BACKGROUND: Aspergillus sp. has been used in traditional Japanese fermented foods. Protease-containing culture products of A. oryzae have been applied as the adjunct enzyme source to enrich the flavor in ripened cheese. Although proteolysis was stimulated, the increase of free fatty acids (FFA) was recognized in some products. Since an excess amount of FFA accumulation can cause rancidity in cheese products, the assessment of lipase activity was considered to be essential for the cheese adjunct preparation. RESULTS: Although an equal lipase activity from the adjunct materials of A. kawachii NBRC 4308, A. luchuensis RIB 2604 and A. oryzae AHU 7139 was applied to semi-hard cheese, the FFA level was significantly higher in A. oryzae cheese than in the others. Furthermore, the profiles of volatile components were different in experimental cheeses. An in vitro study with experimental curds demonstrated that the high FFA might not depend on the lipase retainability on curds. On the contrary, the pronounced activation of the lipases occurred in A. oryzae after incubation with the curds. Moreover, incubation of the insoluble lipase that had been attached to the cells with skim milk curd extracts allowed the release of lipases from the cells into the medium with remarkable activation. CONCLUSION: A. oryzae AHU 7139 possessed a complex lipolytic system comprising extracellular and cell-binding lipases that were attributed to the increase in FFA in A. oryzae cheese. © 2022 Society of Chemical Industry.

Major involvement of two laccase genes in conidial pigment biosynthesis in Aspergillus oryzae.

Wild-type strains of Aspergillus oryzae develop yellow, yellow-green, green, or brown conidia. Previous reports suggested that the conidiation initiates with the biosynthesis of a yellow pigment YWA1 from acetyl-CoA by a polyketide synthase encoded by wA (AO090102000545). This is followed by the conversion to other pigment by a laccase encoded by yA (AO090011000755). Based on orthologous pathways in other Aspergilli, it is reasonable to hypothesize that in addition to yA, AO090102000546 encoding laccase and AO090005000332 encoding Ayg1-like hydrolase play a role in A. oryzae conidial pigment biosynthesis. However, the involvement of these two genes in conidial pigmentation remains unclear. In this study, we tested this hypothesis by assessing the conidial colors of both disruption and overexpression mutants to verify whether AO090102000546 and AO090005000332 were associated with the conidial pigmentation. Observation of single, double, and triple disruptants of these three genes suggested that conidial pigments were synthesized by two laccase genes, AO090011000755 and AO090102000546, whereas Ayg1-like hydrolase gene AO090005000332 was proven to have no obvious association with the synthesis. This was corroborated by observing the phenotype of each overexpression mutant. Interestingly, AO090005000332 overexpression mutant produced smoky yellow-green conidia, different from the wild-type strain. Thus, the AO090005000332-encoded protein is likely to maintain the enzymatic activity. However, the expression level was observed to be one-third of that of AO090102000546 and one-seventh of that of AO090011000755. Consequently, apparent lack of obvious contribution of AO090005000332 to conidial pigmentation could be attributed to its low expression level. Expression analysis indicated similar profiles in several wild-type strains. KEY POINTS: • Conidial pigment biosynthesis after YWA1 mainly involves two laccases in A. oryzae. • Ayg1-like hydrolase in A. oryzae is not obviously involved in conidial pigmentation. • Conidial color is deemed dependent on expression level of two laccases and hydrolase.

How to improve the production of peptidyl compounds in filamentous fungi.

Peptidyl compounds produced by filamentous fungi, which are nonribosomal peptides (NRPs) and ribosomally synthesized and post-translationally modified peptides (RiPPs), are rich sources of bioactive compounds with a wide variety of structures. Some of these peptidyl compounds are useful as pharmaceuticals and pesticides. However, for industrial use, their low production often becomes an obstacle, and various approaches have been challenged to overcome this weakness. In this article, we summarize the successful attempts to increase the production of NRPs and RiPPs in filamentous fungi and present our perspectives on how to improve it further.

Improvement of Free Fatty Acid Secretory Productivity in Aspergillus oryzae by Comprehensive Analysis on Time-Series Gene Expression.

Aspergillus oryzae is a filamentous fungus that has historically been utilized in the fermentation of food products. In recent times, it has also been introduced as a component in the industrial biosynthesis of consumable compounds, including free fatty acids (FFAs), which are valuable and versatile products that can be utilized as feedstocks in the production of other commodities, such as pharmaceuticals and dietary supplements. To improve the FFA secretory productivity of A. oryzae in the presence of Triton X-100, we analyzed the gene expression of a wild-type control strain and a disruptant strain of an acyl-CoA synthetase gene, faaA, in a time-series experiment. We employed a comprehensive analysis strategy using the baySeq, DESeq2, and edgeR algorithms to clarify the vital pathways for FFA secretory productivity and select genes for gene modification. We found that the transport and metabolism of inorganic ions are crucial in the initial stages of FFA production and revealed 16 candidate genes to be modified in conjunction with the faaA disruption. These genes were verified through the construction of overexpression strains, and showed that the manipulation of reactions closer to the FFA biosynthesis step led to a higher increase in FFA secretory productivity. This resulted in the most successful overexpression strains to have an FFA secretory productivity more than two folds higher than that of the original faaA disruptant. Our study provides guidance for further gene modification for FFA biosynthesis in A. oryzae and for enhancing the productivity of other metabolites in other microorganisms through metabolic engineering.

Enhancement of the productivity of free dihomo-γ-linolenic acid via co-overexpression of elongase and two desaturase genes in Aspergillus oryzae.

Free dihomo-γ-linolenic acid (DGLA), a polyunsaturated free fatty acid (FFA), is a precursor of the eicosanoid prostaglandin E1 and is expected to be a source material for artificial production. We previously constructed the Aspergillus oryzae mutant strain ARA1 that produced free DGLA from the disruptant of faaA, an acyl-CoA synthetase gene, where FFA productivity increased by 9.2-fold compared with that of the wild-type strain. Here, we aimed to achieve enhancement of free DGLA productivity. Because saturated FFAs, such as palmitic acid and stearic acid, accounted for about 45% and 25% of total FFAs produced by ARA1, respectively, we used a strategy to facilitate elongation and desaturation of these FFAs to oleic acid and linoleic acid by overexpressing genes encoding elongase, Δ9-desaturase, and Δ12-desaturase originally expressed in A. oryzae. Ten genes were predicted to encode desaturases, and their overexpression DNA constructs were introduced into ARA1. AO090001000224 and AO090011000488 facilitated Δ12-desaturation and Δ9-desaturation most, respectively, following overexpression. Next, ARA1 strain was modified to DGLA1cre strain for producing free DGLA as a final product, and co-overexpression of these two desaturase genes was then introduced to DGLA1cre. Moreover, single overexpression of two genes predicted to encode elongases was additionally introduced, and only AO090003000572 facilitated elongation. Consequently, in the co-overexpression mutant of AO090001000224, AO090011000488, and AO090003000572, free DGLA content ratio increased by 1.8-fold from ARA1 to 14.5%, and the productivity also increased by 1.8-fold to 0.086 mmol/g-dry-cell-weight. The yield was 284 mg/L. These findings provided insights into strategies for improving microbial production of polyunsaturated FFAs.

Ustiloxin biosynthetic machinery is not compatible between Aspergillus flavus and Ustilaginoidea virens.

Ustiloxins are ribosomally synthesized and post-translationally modified peptides (RiPPs) first reported in Ascomycetes. Originally identified as metabolites of the rice pathogenic fungus Ustilaginoidea virens, they were recently identified among the metabolites of the mold Aspergillus flavus, along with their corresponding biosynthetic gene cluster. Ustilaginoidea virens produces ustiloxins A and B, whereas A. flavus produces only ustiloxin B. Correspondingly, in U. virens, the ustiloxin precursor peptide, from which the compound backbone is cleaved and cyclized, contains the core peptides Tyr(Y)-Val(V)-Ile(I)-Gly(G) and Tyr(Y)-Ala(A)-Ile(I)-Gly(G) for ustiloxins A and B, respectively, whereas that of A. flavus contains only the YAIG motif for ustiloxin B. In this study, the gene that encodes the precursor peptide in A. flavus, ustA, was replaced with synthetic genes encoding the core peptides YVIG or FAIG, to investigate their compatibility with the ustiloxin biosynthetic machinery. We also examined the importance of the hydroxyl group on the aromatic ring of Tyr for cyclization of the YAIG core peptide. Against our expectation, the ustA variant possessing YVIG core peptides did not produce a detectable amount of ustiloxin A, even though the ustiloxin biosynthetic gene clusters of A. flavus and U. virens both contain 13 homologous genes. We confirmed that the lack of ustiloxin A production was not due to lack or insufficient expression of the substituted synthetic gene. This result, along with the differences between the primary sequences of UstYa and UstYb in A. flavus and U. virens, suggests that the ustiloxin biosynthetic machinery is optimized for the native core peptide sequences. The synthetic FAIG-encoding ustA did not yield any compounds specific to the FAIG core peptide, suggesting that the hydroxyl group on the aromatic ring of Tyr in the core peptide is indispensable for cyclization of the core peptide, even though it is not structurally involved in the cyclization.

Adjunctive application of solid-state culture products and its freeze-dried powder from Aspergillus sojae for semi-hard cheese.

BACKGROUND: Species belonging to the genus Aspergillus have been used in traditional Japanese fermented foods. Aspergillus sojae is a species responsible for strong proteolytic activity. Freeze-drying treatments followed by physical disruption enables the pulverization of the mycelia of A. sojae RIB 1045 grown in whey protein-base solid media. Intracellular proteases were extracted using this protocol to compare extracellular protease activity in terms of the reaction's pH dependence in the presence or absence of inhibitors. RESULT: With different sensitivities to inhibitors, intracellular and extracellular proteases showed the strongest activity under acidic conditions, which were considered suitable for cheese application. The raw culture product (CP) and its freeze-dried product (FDP) were mixed with cheese curds, prepared according to Gouda-type cheese-making methods, and were allowed to ripen for 3 months. Chemical analysis of the products showed 13.3% water-soluble nitrogen (WSN) in the control, which had received noncultured media, whereas 20.0% and 21.1% WSN was found in the CP and FDP experimental cheeses, respectively. Although these adjuncts significantly increased WSN, an insignificant difference was found between CP and FDP. Free fatty acids in all experimental cheeses were similar, showing that CP and FDP caused no rancid defects. CONCLUSION: The introduction of freeze-drying treatments accompanied by cell disruption resulted in a negligible effect in terms of WSN. However, the application of A. sojae can be beneficial when it comes to increasing the level of WSN compared with A. oryzae, as shown in our previous study. © 2020 Society of Chemical Industry.

Efficient heterologous production of atrochrysone carboxylic acid-related polyketides in an Aspergillus oryzae host with enhanced malonyl-coenzyme A supply.

We recently developed an Aspergillus oryzae strain in which malonyl-coenzyme A (CoA) supply is strengthened by the deletion of snfA and SCAP as an efficient host to produce a plant polyketide, curcumin. Here, we examined the effectiveness of this strain in producing another polyketide, atrochrysone carboxylic acid (ACA), which is synthesized from eight molecules of malonyl-CoA using an iterative type I polyketide synthase, ACA synthase (ACAS), and atrochrysone carboxyl ACP thioesterase (ACTE) in Aspergillus terreus. When ACAS and ACTE were introduced, the A. oryzae ΔsnfAΔSCAP strain produced approximately four times more ACA-related polyketides than did the control strain expressing both genes. This result further demonstrated the availability of the A. oryzae ΔsnfAΔSCAP strain for heterologous polyketide production.

Use of the kojA promoter, involved in kojic acid biosynthesis, for polyketide production in Aspergillus oryzae: implications for long-term production.

BACKGROUND: Aspergillus oryzae, a useful industrial filamentous fungus, produces limited varieties of secondary metabolites, such as kojic acid. Thus, for the production of valuable secondary metabolites by genetic engineering, the species is considered a clean host, enabling easy purification from cultured cells. A. oryzae has been evaluated for secondary metabolite production utilizing strong constitutive promoters of genes responsible for primary metabolism. However, secondary metabolites are typically produced by residual nutrition after microbial cells grow to the stationary phase and primary metabolism slows. We focused on a promoter of the secondary metabolism gene kojA, a component of the kojic acid biosynthetic gene cluster, for the production of other secondary metabolites by A. oryzae. RESULTS: A kojA disruptant that does not produce kojic acid was utilized as a host strain for production. Using this host strain, a mutant that expressed a polyketide synthase gene involved in polyketide secondary metabolite production under the kojA gene promoter was constructed. Then, polyketide production and polyketide synthase gene expression were observed every 24 h in liquid culture. From days 0 to 10 of culture, the polyketide was continuously produced, and the synthase gene expression was maintained. Therefore, the kojA promoter was activated, and it enabled the continuous production of polyketide for 10 days. CONCLUSIONS: The combined use of the kojA gene promoter and a kojA disruptant proved useful for the continuous production of a polyketide secondary metabolite in A. oryzae. These findings suggest that this combination can be applied to other secondary metabolites for long-term production.

Developing a codon optimization method for improved expression of recombinant proteins in actinobacteria.

Codon optimization by synonymous substitution is widely used for recombinant protein expression. Recent studies have investigated sequence features for codon optimization based on large-scale expression analyses. However, these studies have been limited to common host organisms such as Escherichia coli. Here, we develop a codon optimization method for Rhodococcus erythropolis, a gram-positive GC-rich actinobacterium attracting attention as an alternative host organism. We evaluate the recombinant protein expression of 204 genes in R. erythropolis with the same plasmid vector. The statistical analysis of these expression data reveals that the mRNA folding energy at 5' regions as well as the codon frequency are important sequence features for codon optimization. Intriguingly, other sequence features such as the codon repetition rate show a different tendency from the previous study on E. coli. We optimize the coding sequences of 12 genes regarding these sequence features, and confirm that 9 of them (75%) achieve increased expression levels compared with wild-type sequences. Especially, for 5 genes whose expression levels for wild-type sequences are small or not detectable, all of them are improved by optimized sequences. These results demonstrate the effectiveness of our codon optimization method in R. erythropolis, and possibly in other actinobacteria.

Production of the plant polyketide curcumin in Aspergillus oryzae: strengthening malonyl-CoA supply for yield improvement.

The filamentous fungus Aspergillus oryzae was recently used as a heterologous host for fungal secondary metabolite production. Here, we aimed to produce the plant polyketide curcumin in A. oryzae. Curcumin is synthesized from feruloyl-coenzyme A (CoA) and malonyl-CoA by curcuminoid synthase (CUS). A. oryzae expressing CUS produced curcumin (64 µg/plate) on an agar medium containing feruloyl-N-acetylcysteamine (a feruloyl-CoA analog). To increase curcumin yield, we attempted to strengthen the supply of malonyl-CoA using two approaches: enhancement of the reaction catalyzed by acetyl-CoA carboxylase (ACC), which produces malonyl-CoA from acetyl-CoA, and inactivation of the acetyl-CoA-consuming sterol biosynthesis pathway. Finally, we succeeded in increasing curcumin yield sixfold by the double disruption of snfA and SCAP; SnfA is a homolog of SNF1, which inhibits ACC activity by phosphorylation in Saccharomyces cerevisiae and SCAP is positively related to sterol biosynthesis in Aspergillus terreus. This study provided useful information for heterologous polyketide production in A. oryzae.

Search for transcription factors affecting productivity of the polyketide FR901512 in filamentous fungal sp. No. 14919 and identification of Drf1, a novel negative regulator of the biosynthetic gene cluster.

In order to increase secondary metabolite production in filamentous fungi, a transcription factor gene in the biosynthetic gene cluster and global regulator genes such as laeA are considered plausible as targets for overexpression by genetic modification. In this study, we examined these overexpression effect in fungal sp. No. 14919 that produces FR901512, an HMG-CoA reductase inhibitor. Resultantly, the productivity was improved at 1.7-1.8 fold by overexpressing frlE, a transcription factor gene in the biosynthetic gene cluster, whereas productivity did not change by overexpression of laeA and veA. Furthermore, we searched for extra transcription factors affecting the productivity by transcriptome analysis between wild-type strain and highly productive UV mutants. After verifying productivity decrease by overexpression, Drf1, a novel transcription factor encoded by drf1 was identified as the negative regulator. Because each frlE product (FrlE) and Drf1 worked on the same cluster in positive and negative regulatory manners, their network was analyzed.

Heterologous production of free dihomo-γ-linolenic acid by Aspergillus oryzae and its extracellular release via surfactant supplementation.

Free dihomo-γ-linolenic acid (DGLA) and its desaturated form, free arachidonic acid (ARA) are polyunsaturated free fatty acids (FFAs). They are useful raw materials to produce eicosanoid pharmaceuticals. In this study, we aimed at their production by the oleaginous filamentous fungus Aspergillus oryzae via metabolic engineering. Three genes encoding enzymes involved in the synthesis of DGLA and ARA, were isolated from the filamentous fungus Mortierella alpina that produces ARA in a triacylglycerol form. These genes were concatenated to promoters and terminators of highly expressed genes of A. oryzae, and the concatenated DNA fragments were further concatenated with each other to generate a single DNA fragment in the form of a biosynthetic gene cluster. By homologous recombination, the resulting DNA fragment was integrated to the chromosome of the A. oryzae acyl-CoA synthetase gene disruptant whose FFA productivity was enhanced at 9.2-fold more than the wild-type strain. The DNA-integrated disruptant produced free DGLA but did not produce free ARA. Thus, focusing on free DGLA, after removal of the gene for converting DGLA to ARA, the constructed strain produced free DGLA at 145 mg/l for 5 d. Also, by supplementing Triton X-100 surfactant at 1% to the culture, over 80% of free DGLA was released from cells without inhibiting the growth. Consequently, the constructed strain will be useful for attempting production of free DGLA-derived eicosanoids because it bypasses excision of free DGLA from triacylglycerols by lipase. To our knowledge, this is the first report on microbial production of free DGLA and its extracellular release.

Knockout of the SREBP system increases production of the polyketide FR901512 in filamentous fungal sp. No. 14919 and lovastatin in Aspergillus terreus ATCC20542.

In the production of useful microbial secondary metabolites, the breeding of strains is generally performed by random mutagenesis. However, because random mutagenesis introduces many mutations into genomic DNA, the causative mutations leading to increased productivity are mostly unknown. Therefore, although gene targeting is more efficient for breeding than random mutagenesis, it is difficult to apply. In this study, a wild-type strain and randomly mutagenized strains of fungal sp. No. 14919, a filamentous fungus producing the HMG-CoA reductase inhibitor polyketide FR901512, were subjected to point mutation analysis based on whole genome sequencing. Among the mutated genes found, mutation of the sterol regulatory element-binding protein (SREBP) cleavage-activating protein (SCAP) had a positive effect on increasing FR901512 productivity. By complementing the SCAP gene in the SCAP-mutated strain, productivity was decreased to the level of the SCAP-intact strain. Conversely, when either the SCAP or SREBP gene was deleted, the productivity was significantly increased. By genomic transcriptional analysis, the expression levels of three enzymes in the ergosterol biosynthesis pathway were shown to be decreased by SCAP mutation. These findings led to the hypothesis that raw materials of polyketides, such as acetyl-CoA and malonyl-CoA, became more available for FR901512 biosynthesis due to depression in sterol biosynthesis caused by knockout of the SREBP system. This mechanism was confirmed in Aspergillus terreus producing the polyketide lovastatin, which is structurally similar to FR901512. Thus, knockout of the SREBP system should be considered significant for increasing the productivities of polyketides, such as HMG-CoA reductase inhibitors, by filamentous fungi.