Enhancement of Monascus Azaphilone Pigments Production Without Citrinin Contamination by Targeting Overexpression of Histone Acetyltransferase MrEsa1 and Deletion of Polyketide Synthase PksCT

Abstract

Monascus spp. are renowned for producing valuable Monascus azaphilone pigments (MonAzPs), yet their biosynthesis is intrinsically linked to the co-production of the mycotoxin citrinin, posing a significant safety challenge and limiting industrial application. Conventional approaches to disrupt citrinin synthesis often inadvertently reduce MonAzPs yield. To circumvent this limitation, we employed a dual-targeting strategy in Monascus ruber. In this study, we selected the mresa1-overexpressed strain—which can produce more MonAzPs and citrinin—as wild strain to construct a pksCT-deleted strain and explore whether pksCT deletion can affect the enhancement of MonAzPs caused by MrEsa1 overexpression. The results showed that the growth, development, and production of MonAzPs in △pksCT-M7::PtrpC-mresa1 were comparable to those in M7::PtrpC-mresa1, showing accelerated growth and higher MonAzPs yields than in M7. In addition, the relative expression levels of genes involved in MonAzPs synthesis in △pksCT-M7::PtrpC-mresa1 and M7::PtrpC-mresa1 showed the same trend compared with M7, indicating that MrEsa1 overexpression can resist the reduction in MonAzPs caused by pksCT deletion. This study establishes a novel and effective paradigm for decoupling desirable metabolite production from toxin synthesis in fungi, providing a strategic framework for the safe and enhanced production of MonAzPs.

1. Introduction

Monascus azaphilone pigments (MonAzPs) are commercially significant natural colorants produced by filamentous fungi of the Monascus genus. They are widely employed as safe and stable color additives in the food industry (e.g., meat products, sauces, and baked goods) and are increasingly applied in textiles and cosmetics [1,2,3,4,5]. Beyond their coloring properties, MonAzPs also exhibit various physiological activities, such as antioxidant, anti-inflammatory, antimicrobial, and neuroprotective effects [6,7], which position them as promising candidates for functional foods and pharmaceuticals, highlighting their value far beyond that of mere colorants. However, some Monascus strains can produce citrinin, a kidney toxin [8], and high production of MonAzPs is often accompanied by a high production of citrinin. This co-production poses a significant public health risk and regulatory hurdle. Therefore, decoupling MonAzPs synthesis from citrinin contamination is a critical goal in fungal biotechnology and food safety.

Both MonAzPs and citrinin are polyketide compounds formed by the primary metabolites acetyl CoA and malonyl CoA as substrates, and their synthesis is catalyzed by multiple enzymes encoded by independent biosynthetic gene clusters [9,10], suggesting the possibility of pathway-specific intervention. Consequently, targeted disruption of the citrinin biosynthetic gene cluster has become a primary strategy. However, empirical results across different Monascus species reveal a spectrum of outcomes that challenge this straightforward prediction. Previous research showed that in Monascus purpureus, deletion of the citrinin biosynthetic gene cluster results in the elimination of citrinin and a 2-5% increase in red pigment production [11]. Additionally, ctnA knockout reduces citrinin by 78%, and MonAzP production remains stable [12]. In M. aurantiacus, when ctnF was knocked out, the content of citrinin decreased by 34% and MonAzP production decreased by 72% [13]. These inconsistencies highlight a critical biological reality: secondary metabolism in fungi is not merely the sum of its isolated biosynthetic gene clusters [14]. Instead, it is governed by a sophisticated regulatory network extending beyond isolated biosynthetic gene clusters, where cross-talk and competition for cellular resources can lead to unpredictable outcomes from single genetic modifications. Therefore, single genetic modifications, while rationally designed, frequently lead to unpredictable systemic outcomes, indicating the need for more integrated engineering strategies that account for the holistic cellular context to reliably achieve desired metabolic phenotypes.

In recent years, epigenetic regulation, particularly histone acetylation, has emerged as a key global regulator that globally coordinates the expression of biosynthetic gene clusters in fungal secondary metabolites [15]. In M. ruber, the genetic manipulation of histone acetyltransferases (HATs) and deacetylases (HDACs) has revealed a global regulatory role for histone acetylation in secondary metabolism. For example, deletion of histone deacetylases—including mrhda1, mrhos3, and mrhst4—and overexpression of histone acetyltransferase mresa1 increased the production of MonAzPs and citrinin, both of which positively regulated the production of MonAzPs and citrinin simultaneously [16,17,18,19]. These results confirm that improving the level of acetylation could act as a broad metabolic amplifier, likely by remodeling chromatin into a more transcriptionally permissive state that coordinately activates multiple biosynthetic gene clusters [20]. However, this strategy alone fails to address the concomitant issue of citrinin contamination because histone acetylation non-selectively enhances the output of both desired and undesired pathways. Therefore, integrating broad epigenetic activation with precise pathway elimination may be a logical strategy to achieve a desired metabolic outcome. This concept of combinatorial engineering remains underexplored in Monascus fungi.

In this study, we implemented and validated this dual-targeting strategy in M. ruber. We hypothesized that combining the global transcriptional activation conferred by mresa1 overexpression with the precise pathway blockage achieved by pksCT deletion would create a synergistic effect: retaining or enhancing the high MonAzPs yields driven by epigenetic remodeling while completely abolishing citrinin synthesis. The key gene pksCT, involved in the biosynthetic pathway of citrinin, was knocked out in the mresa1-overexpressed strain, and the growth, development, and MonAzPs production were analyzed. This work provides a new strategy for increasing the yield of MonAzPs and reducing or eliminating the yield of citrinin in Monascus fungi. Furthermore, it establishes a novel and effective paradigm for the precise rewiring of fungal metabolism.

2. Materials and Methods

2.1. Strains, Plasmids, and Culture Conditions

In a previous study, M7::PtrpC-mresa1 was constructed with M. ruber M7 as the wild type [19], whereas herein, △pksCT-M7::PtrpC-mresa1 was constructed with M7::PtrpC-mresa1 as the wild type. All strains were maintained under two preservation methods: For short-term use and phenotypic assays, all strains were maintained on PDA slants at 4 °C and routinely transferred to fresh PDA slants every 2–3 months. For long-term genetic stability and preservation, all strains were maintained in vacuum freeze-dried form as our primary archival method. Plasmids pSKH were kept in our laboratory for the construction of various vectors. To observe colonial morphology, we prepared the following materials: potato dextrose agar (PDA), whereby potato extract was prepared by boiling 200 g of peeled potatoes in distilled water, followed by filtration and the addition of 20 g dextrose, 15 g agar, and distilled water to a final volume of 1 L; G25N, comprising 25% glycerol nitrate agar, prepared with 3.0 g NaNO₃, 1.0 g K₂HPO₄, 0.5 g KCl, 0.5 g MgSO₄·7H₂O, 0.01 g FeSO₄·7H₂O, 30.0 g sucrose, 5.0 g yeast extract, 15 g agar, 250 g glycerol, and distilled water to a final volume of 1 L; Chapek Yeast Extract Agar (CYA), comprising 3.0 g NaNO₃, 1.0 g K₂HPO₄, 0.5 g KCl, 0.5 g MgSO₄·7H₂O, 0.01 g FeSO₄·7H₂O, 30.0 g sucrose, 5.0 g yeast extract, 15 g agar, and distilled water up to 1 L; and malt extract agar (MA), comprising 1 L of 15°Bx wort and 15 g agar.

2.2. Construction and Validation of pksCT-Deleted Strain

First, genomic DNA was extracted from M7 using a previously described method [19] and served as the template to amplify the 5′- and 3′-flanking regions of pksCT. Subsequently, selection makers (hph) were amplified from plasmid pSKH. Using these amplified DNA sequences, a deleted cassette was constructed via double-joint PCR [21]. The deleted cassette was digested by Xba I and Sac I and then ligated with pCAMBIA3300 vector, which had been digested with the same restricted enzymes to form recombinant vectors. The resulting plasmid was then transformed into Agrobacterium tumefaciens EHA105 cell, which was used to introduce the constructed cassette region into the hosts.

pksCT-deleted cassette consisted of the hygromycin resistance gene and the strains growing on media containing hygromycin were screened as candidates of △pksCT-M7::PtrpC-mresa1. A primer designed from the ORF region of pksCT was used to amplify the specific DNA fragments for screening the putative △pksCT-M7::PtrpC-mresa1. Then, putative △pksCT-M7::PtrpC-mresa1 was confirmed via Southern blot using the DIG-High Prime DNA Labeling and Detection Starter Kit I (Roche, Germany) with probe 1 (pksCT gene) and probe 2 (hph gene). The DNAs of M7::PtrpC-mresa1 and the putative △pksCT-M7::PtrpC-mresa1 strains were digested with Xba I and Sac I.

2.3. Analysis of Colonial Morphology and Biomass

M7, M7::PtrpC-mresa1, and △pksCT-M7::PtrpC-mresa1 were cultured for 10 d at 28 °C to observe colonial morphology on PDA, CYA, G25N, and MA. A spore suspension (10⁵ spores/mL) was prepared from a 10-day-old culture grown on a PDA slant. Then, 5 μL of the spore suspension was spot-inoculated at the center of 70 mm plates containing PDA, CYA, MA, and G25N media, respectively. The plates were incubated at 28 °C for 10 days to observe the colony morphologies. Meanwhile, the colony diameters were measured from digital images using ImageJ software (ImageJ 1.54g). Biomass was determined according to the published paper [19].

2.4. Detection of MonAzPs and Citrinin Production

Red yeast rice, also called Hongqu, can be used directly as a food colorant or material for MonAzPs extraction. Another effective means of extracting MonAzPs is liquid fermentation. Therefore, we detected the production of MonAzPs in Hongqu and mycelia fermented in PDB medium. To prepare a spore suspension, 10 mL of sterile water was added to a 10-day-old culture grown on a PDA slant. The mycelium was gently scraped off the agar surface with a sterile inoculating hook, and the resulting suspension was transferred to a 50 mL conical flask containing 5–10 sterile glass beads. The flask was then shaken at 120 rpm at 28 °C for 30 min. Subsequently, the homogenate was filtered through three layers of sterile lens cleaning tissue, and the spores were counted under a microscope. Then, 2 mL freshly harvested spore supernatant was sprayed onto 30 g of steamed rice and incubated at 28 °C. In parallel, 200 uL of freshly harvested spore supernatant was added into a 250 mL wide-neck Erlenmeyer flask containing 50 mL PDB, and then shaken at 150 rpm at 28 °C. We collected the fermented rice and mycelia every 2 days from the 3rd day until the 11th day. Then, the fermented rice was dried and ground into powder, while mycelia were harvested via filtration through a single layer of sterile gauze, washed three times with distilled water, and vacuum-freeze-dried for 36 h. A total of 20 mg of Hongqu and mycelia each was extracted using 1 mL of 80% methanol. The values corresponding to the maximum absorption peak at 380 nm, 470 nm, and 520 nm of extraction were detected using a microplate reader (SpectraMax iD5, Thermo Fisher Scientific, Waltham, MA, USA) to calculate the yield of yellow, orange, and red pigments; the results are expressed as OD units per gram of dry biomass (U/g). HPLC was conducted using a C18 column (Supersil AQ-C18, 250 mm × 4.6 mm, 5 μm) with isocratic elution to quantify citrinin content. The mobile phase included solvent A (acetonitrile) and solvent B (water adjusted to pH 2.5 with phosphoric acid) with a ratio of 75:25 (v/v), a 1 mL/min flow rate, and an injection volume of 10 μL.

2.5. Detection of the Relative Expression Levels of Specific Genes via RT-qPCR

The relative expression levels of genes responsible for polyketide synthesis and pksCT were detected in M7, M7::PtrpC-mresa1, and △pksCT-M7::PtrpC-mresa1. Firstly, total RNAs were extracted from each strain using EZ-10 DNA away RNA Mini-prep Kit (Sangon Biotech, Shanghai, China). They were used as a template to perform RT-qPCR using HiScript II 1st Strand cDNA Synthesis Kit and AceQ qPCR SYBR Green Master Mix (Vazyme, Nanjing, China). The results were expressed by relative expression level according to the formula described previously [19,22].

All primers used in this study are listed in Table S1.

2.6. Statistical Analysis

Experiments were performed in triplicate, and significance was assessed via one-way analysis of variance (ANOVA) using Statistical Package for the Social Sciences (SPSS) software (version 18.0, SPSS Inc., Chicago, IL, USA). Only P values less than 0.05 were regarded as statistically significant.

3. Results

3.1. Verification of pksCT-Deleted Vector and Strain

The constructed pksCT-deleted vector was confirmed via PCR and double digestion with Xba I and Sac I. PCR amplification yielded the expected size bands of about 5000 bp by amplifying the deletion cassette fragment. The vector was then digested by Xba I and Sac I to verify correctness, obtaining target fragments of about 5000 bp and 8000 bp, suggesting that the pksCT-deleted vector was successfully constructed (Figure 1A,B).

Then, the vector was transformed into Agrobacterium tumefaciens EHA105 cell, which was subsequently used to introduce the constructed cassette region into the hosts. The ORF region of pksCT was amplified to verify the candidates of the pksCT-deleted strain with M7::PtrpC-mresa1 as the control (WT). The PCR products of the screened △pksCT-M7::PtrpC-mresa1 strain showed no bands, and that of WT yielded the expected size band (Figure 1C). Following this, the DNAs of WT and putative △pksCT-M7::PtrpC-mresa1 strains were digested with Xba I and Sac I to conduct Southern blot analysis. The Southern blot analysis showed one band in putative △pksCT-M7::PtrpC-mresa1 using the hph probe, while no band was observed in WT (Figure 1D). This suggested that the screened pksCT-deleted strain (△pksCT-M7::PtrpC-mresa1) was correct.

3.2. Analysis of Growth and Development

Previous studies showed that mresa1 could accelerate the growth of M7 in different culture media and significantly increase the dry cell mass. Thus, we observed the growth change of △pksCT-M7::PtrpC-mresa1 on PDA, CYA, G25N, and MA. The results showed that the growth rates of △pksCT-M7::PtrpC-mresa1 and M7::PtrpC-mresa1 were similar, both of which were faster than that of M7, with no other notable differences observed (Figure 2A). Meanwhile, the colony diameter of △pksCT-M7::PtrpC-mresa1 was larger than that of M7 in PDA, CYA, G25N, and MA, and it showed significant increases on PDA (p < 0.05) (

Table 1. Comparison of colony diameters for different strains cultured on various agar media.

Colony Diameter (mm)	PDA	CYA	G25N	MA
M7	43.08	41.78	20.51	55.45
M7::PtrpC-mresa1	54.77	47.20	21.72	58.98
△pksCT-M7::PtrpC-mresa1	55.79	47.97	21.24	59.08

and Figure 2B). The dried weights of △pksCT-M7::PtrpC-mresa1 and M7::PtrpC-mresa1 significantly increased during the growth phase compared with M7 when inoculated in PDB (Figure 2C). Specifically, M7::PtrpC-mresa1 showed significant increases on days 7, 9, and 11 (all p < 0.05). Meanwhile, △pksCT-M7::PtrpC-mresa1 exhibited even more pronounced increases, with significant differences seen between days 5 and 7 (p < 0.01) and between days 9 and 11 (p < 0.05). This may be due to the accelerated growth rate caused by MrEsa1 overexpression.

3.3. Measurement of MonAzPs and Citrinin Production

MonAzPs are classified as either red, orange, or yellow pigments based on their maximum absorption wavelength [23]. Thus, we determined the content of red, orange, and yellow pigments in PDB fermentation and Hongqu. The results showed that the production of yellow, orange, and red pigments in the mycelia of △pksCT-M7::PtrpC-mresa1 growing in PDB was similar to that in M7::PtrpC-mresa1, both of which were significantly higher than that in M7 throughout the entire growth phase—approximately 2.3 times higher (Figure 3A–C). A similar trend was observed in Hongqu fermentation: the production of yellow, orange, and red pigments in Hongqu by △pksCT-M7::PtrpC-mresa1 was also similar to that by M7::PtrpC-mresa1, both of which were approximately 1.5 times higher than M7 (Figure 3D–F). In addition, HPLC analysis confirmed that the retention time of citrinin is about 5 min, and that no citrinin peak was detected in the extracts of △pksCT-M7::PtrpC-mresa1 at the same retention time (Figure 4), demonstrating the complete abolition of citrinin in the final dual-engineered strain (△pksCT-M7::PtrpC-mresa1).

3.4. The Expression Level of Genes Involved in the Biosynthesis of MonAzPs and Citrinin

To gain insight into the relative expression level of genes associated with MonAzPs and citrinin when mresa1 was overexpressed and pksCT was deleted, we performed RT-qPCR to quantify the relative expression level of these genes. The results showed that the trend of the relative expression levels of genes associated with MonAzPs in △pksCT-M7::PtrpC-mresa1 and M7::PtrpC-mresa1 were ultimately consistent. The relative expression levels of mrpigB, mrpigH, mrpigJ, and mrpigL in △pksCT-M7::PtrpC-mresa1 and M7::PtrpC-mresa1 were higher than in M7; in particular, the expression level of mrpigB was significantly upregulated. In addition, pksCT is the key gene involved in the biosynthesis of citrinin, and the results showed that when pksCT was deleted, citrinin biosynthesis was completely abolished [10]. Therefore, we detected the relative expression level of pksCT, and the results showed that this gene was not expressed in △pksCT-M7::PtrpC-mresa1 (Figure 5), indicating that △pksCT-M7::PtrpC-mresa1 does not produce citrinin. Furthermore, the relative expression level of pksCT was significantly upregulated in M7::PtrpC-mresa1 (Figure 5), which corresponds to an increase in the production of citrinin in M7::PtrpC-mresa1 [19].

4. Discussion

MonAzPs, prized as natural food colorants, share biosynthetic precursors acetyl-CoA and malonyl-CoA with the nephrotoxin citrinin in Monascus spp. [9,24]. Therefore, theoretically, the production yield of MonAzPs will increase when the yield of citrinin decreases by knocking out the key genes involved in the biosynthesis of citrinin. However, this straightforward metabolic trade-off is not always observed in practice. Contrary to theoretical predictions, in M. purpureus and M. aurantiacus, the deletion of genes involved in the biosynthesis of citrinin led to either the elimination of or decrease in citrinin and either a noticeable reduction or no increase in the production of MonAzPs [12,13]. This observed co-depletion contradicts straightforward metabolic balancing principles and has significantly hindered the development of commercially viable and safe fungal strains for MonAzPs production, as reducing citrinin content comes at the cost of decreasing the target product yield. Emerging evidence suggests that this paradox may stem from the pleiotropic roles of the involved gene clusters; the gene clusters of MonAzPs are involved in various biological processes such as naphthoquinone biosynthesis, reduction detoxification, starch and sucrose metabolism, and pksCT, which can also affect glycolysis, the tricarboxylic acid (TCA) cycle, and other metabolic processes [25,26,27,28]. This extensive metabolic interconnectivity likely disrupts simple precursor competition models, providing a rationale for the observed co-depletion and deviating from naive metabolic balancing principles.

Our study successfully demonstrates the efficacy of a dual-targeting metabolic engineering strategy in M. ruber, which synergistically combines broad epigenetic activation with precise pathway disruption. Our core hypothesis—that overexpressing MrEsa1 combined with deleting PksCT would enhance MonAzP production while eliminating citrinin—has been robustly validated. Critically, the resulting engineered strain exhibited the desired metabolic phenotype: the complete abolition of citrinin coupled with the retention of the high MonAzPs yield characteristic of the mresa1-overexpressing strain. Existing research demonstrated that strategies relying solely on the knockout of citrinin biosynthetic genes yield inconsistent and unpredictable effects on MonAzPs production, ranging from moderate increases to substantial decreases. In contrast, the dual-engineered strain constructed in this study (△pksCT-M7::PtrpC-mresa1) achieved the optimal outcome: it completely eliminated citrinin while simultaneously delivering the highest reported increase in MonAzPs production (

Table 2. Comparison of metabolic engineering strategies for MonAzP and citrinin production.

Strain	Engineering Strategy	Cultivation Conditions	MonAzPs Yield	Citrinin Yield
M. purpureus [11,12]	deletion of 15 kb citrinin biosynthetic gene cluster	PDB	2-5% increase	elimination
	deletion of ctnA	PDB	no increase	78% decrease
M. aurantiacus [13,29]	deletion of ctnF	PDB	72% decrease	34% decrease
	deletion of ctnE	PDB	40% increase	96% decrease
M. ruber (this study)	deletion of pksCT and overexpression of mresa1	PDB	130% increase	elimination
		Hongqu	50% increase	elimination

).

This outcome successfully addresses the persistent challenge of balancing safety and yield in Monascus engineering. The combined approach appears to function through complementary mechanisms. First, at the metabolic level, deleting pksCT—the gene responsible for the first step of citrinin biosynthesis—primarily abolishes citrinin synthesis [10]; meanwhile, it eliminates a major competitive sink for the shared acetyl-CoA and malonyl-CoA precursor pool and may remove potential cross-pathway inhibition [30]. Consequently, the cellular pool of acetyl-CoA, a key metabolic precursor, likely increases. Importantly, acetyl-CoA also serves as the essential donor substrate for histone acetylation [31]. Second, at the epigenetic level, mresa1 overexpression appears to compensate for these disruptions via epigenetic reprogramming, and the elevated availability of acetyl-CoA may thus further potentiate the histone acetyltransferase activity of MrEsa1, creating a synergistic loop that enhances the overall transcriptional activation required for metabolic compensation [32]. In addition, the growth and dried cell mass of △pksCT-M7::PtrpC-mresa1 showed a similar trend with M7::PtrpC-mresa1. It demonstrated that MrEsa1 still plays a positive regulatory role in growth and development under the synergistic effect of pksCT deletion, and both aspects are conducive to the production of secondary metabolites [33,34]. Moreover, compared with M7, the relative expression of genes associated with MonAzPs in △pksCT-M7::PtrpC-mresa1 and M7::PtrpC-mresa1 also showed similar changes. These results show that the underlying mechanism for this compensation is rooted in the fundamental role of MrEsa1.

Histone acetyltransferase Esa1, which can transform the chromatin structure from a tightly packed “heterochromatin” state to a loosely packed “euchromatin” state, greatly promotes gene transcription and expression [35,36]. Furthermore, overexpression of MrEsa1 causes hyperacetylation of K4 and K9 of the H3 subunit and H3Pan [19]. Consequently, overexpression of MrEsa1 establishes a highly active transcriptional environment at the epigenetic level, which may explain its ability to compensate for the decrease in MonAzP production caused by the deletion of pksCT. We hypothesize that this enhanced transcriptional activity serves a dual purpose: it directly or indirectly facilitates the MonAzPs’ gene cluster to utilize the redirected precursor pool, while also broadly enhancing cellular homeostasis and stress resilience to buffer metabolic disruption. Thus, it is a feasible strategy for the safe and efficient production of MonAzPs. In this study, we demonstrated that a strategy combining targeted gene knockout with global epigenetic regulation could effectively redirect metabolic flux toward desired compounds. The novelty of our study lies not only in the achieved decoupling of MonAzPs and citrinin, but also in providing a generalizable framework. This framework leverages epigenetic tools to confer robustness and redirect metabolic flux, thereby addressing the common pitfall of yield losses in microbial metabolic engineering.

5. Conclusions

MonAzPs, produced by Monascus species, have been used as food, food additives, folk medicine, as well as fermentation starters for many centuries in China and other Southeast Asian countries. However, citrinin production poses safety risks for MonAzPs. In many cases, a high production of MonAzPs is often accompanied by a high production of citrinin, which makes it difficult for Monascus spp. to produce MonAzPs safely and effectively solely through the synthetic gene cluster pathway. In this study, our experimental data revealed that the combination of mresa1 overexpression and pksCT deletion can produce more MonAzPs and eliminate the yield of citrinin. Our study provides a powerful and innovative strategy for the safe, efficient, and sustainable bioproduction of MonAzPs, with significant implications for food safety and fungal biotechnology. The engineered strain △pksCT-M7::PtrpC-mresa1 constructed in our study holds direct promise for applications in functional foods and nutraceuticals (as a safe source of natural pigments and bioactive compounds), as well as the cosmetic and pharmaceutical industries (as natural colorants or active ingredients). Beyond the specific case of MonAzPs, the strategic framework of synergizing targeted pathway disruption with global epigenetic activation offers a generalizable model. This “release inhibition and enhance drive” approach could be adapted to optimize the production of other valuable fungal secondary metabolites where yield is constrained by competing pathways or complex regulation, thereby contributing broadly to fungal biotechnology and metabolic engineering.