Inactivation of the oxidase gene mppG results in the selective loss of orange azaphilone pigments in Monascus purpureus

Abstract

Monascus species are filamentous ascomycetes fungi and produce azaphilone (Az) pigment that is a well-known food colorant. Az is a class of fungal polyketides that bears a highly oxygenated pyranoquinone bicyclic core and is produced by a nonreducing fungal polyketide synthase with a reductive release domain (NR-fPKS-R). MpPKS5 encodes an NR-fPKS-R for Monascus Az (MAz) and is clustered with four oxidoreductase genes including mppG; mpp designates Monascus pigment production. MAz pigments are classified as yellow and orange MAz, and their structures differ in two hydride reductions with yellow MAz as the reduced type. The biosynthesis of yellow MAz (monascin, Y-1 and ankaflavin, Y-2) is completed by a reductive pathway involving a reductase gene mppE. This reductive pathway is diverged from a common MAz pathway involving two other reductase genes of mppA and mppC. This suggests that the biosynthesis of orange MAz (rubropunctatin, O-1 and monascorubrin, O-2) is completed by an oxidative branch pathway and the cognate oxidative role of mppG is genetically characterized in the present study. A targeted gene inactivation mutant of ΔmppG displayed a severe impairment in the production of orange MAz with no significant alteration in the level of yellow MAz. The feeding experiment with Y-1 in ΔMpPKS5 indicated that Y-1 could not be converted into O-1, which excludes the possibility that mppG mediates the conversion of yellow into orange MAz. This study supports the existence of divergent pathways in MAz biosynthesis and creates a recombinant strain for the selective production of yellow MAz.

Introduction

Azaphilone (Az) consists of a group of fungal aromatic polyketides featuring the oxidized pyranoquinone bicyclic structure having a tertiary alcohol that is generally acylated. Az polyketides display diverse biological activities, which apparently involve interfering with specific protein–protein interactions [1]. The food fermentation fungus Monascus is well known for its high production of Az pigments, and Monascus ethanol extract has been used as a food colorant [2, 3]. The main pigment components of Monascus Az (MAz) are yellow (monascin, Y-1 and ankaflavin, Y-2), orange (rubropunctatin, O-1 and monascorubrin, O-2) and red MAz (rubropunctamine, R-1 and monascorubramine, R-2) (Fig. 1). Yellow MAz contains a reduced pyranoquinone core and has a visible absorption peak at 390 nm. Orange MAz, bearing a typical Az pyranoquinone core, has a maximum absorption at 470 nm and is readily converted into red MAz in the presence of amine [4]. The color of orange MAz is pH sensitive, shifting to red and purple at pH values greater than 6.0 in aqueous ethanol [5]. The term Az was coined from the tendency of incorporating nitrogen atom to generate a vinylogous γ-pyridone moiety, as exemplified in red MAz. It is thus not surprising to find diverse amino acid derivatives of red MAz in some Monascus culture extracts [6]. Many Az members, including yellow MAz, have no affinity for nitrogen, however. The MAz is not unique to Monascus, but it can also be found in other related filamentous fungi [7].

Fig. 1

Structures of MAz compounds with their proposed biosynthetic pathway, which is deduced by targeted gene inactivation studies [4, 24, 25, 27, 28]. The oxidative role of mppG in orange MAz production is demonstrated in this study. To emphasize the biosynthetic origin, the acetate units and S-adenosyl-l-methionine-derived carbons are denoted with bars and black dots, respectively. The compound names are in bold

The health benefits of Monascus-fermented products are generally attributed to monacolin K (lovastatin), a well-known polyketide compound used for treating hypercholesterolemia [8, 9]. Recent studies have demonstrated that yellow MAz is also effective in blood lipid control [10], modulating lipoprotein metabolism in a health-benefitting manner [11]. In addition, several notable biological activities have been reported for yellow MAz, including anti-atherosclerosis [12], anti-diabetic [13, 14], anti-inflammatory [15, 16] and anti-obesity activities [17]. These support the notion that yellow MAz contributes to the health benefits associated with Monascus fermentation products [18]. The culture methods for promoting yellow MAz production have been developed. A high level of yellow MAz was obtained via a pH-static batch fermentation in a defined medium [19]. Nonionic surfactant-extractive fermentation afforded the accumulation of yellow MAz in culture supernatant with a high overall production [20]. An edible oil version of this method was also reported [21].

Biosynthetic genes for a microbial secondary metabolite are often clustered in chromosome, and their overall nucleotide sequence information provides an excellent opportunity for systematic study of the related biosynthetic mechanism [22, 23]. In an effort to delineate MAz biosynthetic pathway, we have identified MAz biosynthetic gene cluster from Monascus pilosus [4] and M. purpureus genome sequence [24] (Fig. 2A). The subsequent gene knockout studies in M. purpureus led us to consider a divergent pathway scenario for simultaneous production of yellow and orange MAz [24, 25]. All MAz compounds are produced from an aromatic polyketide pathway with nonreducing fungal polyketide synthase with a reductive release domain (NR-fPKS-R) [4, 26]. The MAz NR-fPKS-R (MpPKS5) pathway involves azanigerone E, which is further processed to be divergently converted into yellow or orange MAz (Fig. 1). The involvement of azanigerone E in MAz biosynthesis is supported by these two findings, the hydroxylation-mediated conversion of FK17-P2a into azanigerone E, demonstrated in azanigerone biosynthesis [26], and accumulation of FK17-P2a in M. purpureus knockout mutant of the hydroxylase gene mppF [25]. Azanigerone E is proposed to be decorated with a 3-oxo-acyl moiety that is generated by MpFAS2 with the assistance of the acyltransferase MppB in MAz biosynthesis [4, 27]. It is also proposed that the subsequent modifications by Mpp7 and MppC generate a hypothetical intermediate that then diverge into reductive (yellow MAz) and oxidative (orange MAz) pathways [24, 25] (Fig. 1).

Fig. 2

Gene inactivation of mppG in M. purpureus. (A) Genetic organization of MAz biosynthetic gene cluster in M. purpureus. MpPKS5 encodes NR-fPKS-R for MAz biosynthesis [4]. The products of MpFasA2 and MpFasB2 are deduced to constitute a fatty acid synthase MpFAS2, which generates short-chain oxo-acyl moieties for MAz production [27]. Four oxidoreductase genes are highlighted with filled arrows. (B) A NheI restriction map of the DNA region flanking mppG in WT, pBIJ19 and the resulting ΔmppG mutant. The size of each DNA fragment is shown in kb above the corresponding arrow. (C) Analytical PCR results of total DNA from the WT strain (lane 1) and the ΔmppG mutant (lane 2) with the primer pair of MppEup-F and MppEdw-R. The resulting PCR products were digested with NheI. Lane M indicates a DNA molecular weight marker, with sizes of 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.8, 1.0, 1.5, 2.0, 3.0, 4.0, 5.0, 6.0, 8.0 and 10.0 kb (from bottom to top). Highlighted with the bars are 1.0- and 3.0-kb fragments

It was previously shown that inactivation of mppE in M. purpureus resulted in a significant reduction in yellow MAz with a concomitant increase of orange MAz [28]. This indicated that mppE is involved in the biosynthesis of yellow MAz, implying a role for MppE in the reduction of the pyranoquinone core during yellow MAz biosynthesis (Fig. 1). The expression of an extra copy of mppE in M. purpureus failed to promote the production of yellow MAz in potato dextrose media but was effective in doing so in a chemically defined medium culture [28]. Considering the structural differences between yellow and orange MAz, we hypothesized that an oxidative modification is involved at a late stage of orange MAz biosynthesis. The MAz biosynthetic gene cluster encodes only one oxidase gene candidate, mppG (the gene model of CE4855_5133 and protein ID of 4856 in the M. purpureus genome sequence browser in the Joint Genome Institute portal) [24]. The mppG region was initially annotated in a part of mppD to encode an amine oxidase domain [4] and was lately annotated as a discrete oxidase gene [24] (Fig. 2A). MppG is predicted to be a flavin-containing oxidase and is not found in any other Az biosynthetic gene clusters so far reported.

In this study, we demonstrated that an mppG knockout mutant resulted in a dramatic reduction of orange MAz while retaining the production of yellow MAz. These findings support the divergent pathway proposal and provide a genetic engineering strategy for the preparation of yellow MAz with a minimal production of orange MAz.

Materials and methods

Strains, culture conditions and extraction methods

Monascus purpureus KACC (Korean Agricultural Culture Collection) 42430 and Agrobacterium tumefaciens AGL1 were used in this study. M. purpureus and its derivatives were maintained on potato dextrose agar (PDA) for 7 days at 30 °C [29]. The PDA volume in each plate was approximately 50 mL. Five pieces of 1 cm³ agar blocks from the PDA culture were used to initiate a potato dextrose broth culture (PDB; Gellix™, Ventech Bio, Seoul, Republic of Korea). This PDB brand affords a high level of orange MAz production from the M. purpureus strain used in this study. The culture condition and pigment extraction method were previously described [28]. The organic extract from each 50 mL culture was dissolved in 1 mL methanol and was used in high-performance liquid chromatography (HPLC) analyses. For the feeding experiment of exogenous Y-1, 5 mg of Y-1 was added to a 50-mL PDB culture of the ΔMpPKS5 mutant at 3 days after the initiation of the culture [4]. After 4 days of incubation, the culture was harvested for extraction.

Gene inactivation

The primers used in this study are listed in Table 1. Polymerase chain reaction (PCR) was performed with Herculase II Fusion DNA Polymerase (Agilent, Santa Clara, CA, USA). To prepare the mppG inactivation construct, 1856- and 1819-bp DNA fragments upstream and downstream of mppG were PCR amplified with the primer pairs of MppGup-F/MppGup-R and MppGdw-F/MppGdw-R, respectively. These two fragments were cloned such that they flanked a 3.9-kb hygromycin resistance cassette (hyg) in pCAMBIA1300 (GenBank accession no. AF234296) through a three-fragment ligation using an In-Fusion cloning method (Clontech, Mountain View, CA, USA), generating pBIJ19 (Fig. 2B). A primer pair of hyg-F/hyg-mppG-R was used to amplify hyg from pUR5750, which is a pBIN19 derivative containing hyg from pAN7.1 [30]. The inactivation construct pBIJ19 was introduced into M. purpureus through an Agrobacterium-mediated transformation as previously described [31].

Table 1 Primers used in this study. The nucleotide sequences introduced for In-Fusion cloning are italicized and underlined

Chemical analysis methods

Thin-layer chromatography (TLC) was performed on silica gel 60 F₂₅₄ TLC plates (Merck), which were developed with a mixture of n-hexane, ethyl acetate and formic acid (24:18:1). Ultraviolet (UV)–visible absorption spectra were collected with a Cary50 spectrophotometer (Varian, Palo Alto, CA, USA). HPLC analysis was performed on a ProStar system (Varian) with a Gemini C18 column (150 × 3.0 mm, 5.0 μm; Phenomenex, Torrance, CA, USA), and the elution was monitored at 330 nm. The mobile phase consisted of 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B) [32]. The flow rate was maintained at 0.5 mL/min. The system was run with the following gradient program: 100% A for 5 min, from 100% A to 100% B over 20 min, then maintained at 100% B for 10 min. For the analysis of the PDB extracts, the gradient elution started from 50% A; after 5 min, the eluent composition was changed to 100% B over 20 min. For large-scale injections, a semi-preparative ODS-A C-18 column (250 × 10 mm, particle size of 5 μm, pore size of 12 nm; YMC, Kyoto, Japan) was used with a flow rate of 1.5 mL/min. The system was run with the following gradient program: 100% A for 5 min, from 100% A to 90% A over 25 min, from 90% A and 100% B over 20 min, then maintained at 100% B for 20 min.

Results and discussion

Targeted inactivation of mppG in M. purpureus

An mppG knockout mutant (Δ mppG) was generated by deleting a 951-bp internal region and replacing it with hyg (Fig. 2B). A PCR using the primer pair MppGup-F/MppGdw-R amplified ~5-kb-sized fragment from the wild-type (WT) DNA. The expected size for this WT fragment is 4.6 kb, and its identity was verified by a NheI digestion that generated expected fragments of 3.8 and 0.8 kb (Fig. 2C). The 4.6-kb WT band was absent in an identical PCR of the Δ mppG DNA, with ~8-kb-sized fragment being evident instead. The successful marker replacement in the chromosome was expected to yield a 7.7-kb band in this PCR amplification. This Δ mppG band was also subjected to a NheI digestion, which resulted in 4.1-, 2.6- and 0.8-kb-sized fragments, confirming the genotype of the Δ mppG strain (Fig. 2).

Dramatic reduction of orange MAz production in the ΔmppG strain

In a PDA culture, the Δ mppG mutant mycelia displayed a yellow hue and were easily distinguishable from the orange color of the WT strain (Fig. 3A). In UV–visible absorption measurements of the organic extracts, the Δ mppG mutant displayed an overall reduction in the visible absorption (Fig. 3B). It needs noting that the orange and red MAz compounds have considerable absorption near 400 nm, as well as at 470–530 nm. Thus, the visible absorption spectrum itself is incapable of deciphering the MAz profile in detail. In a TLC analysis of the extracts, there is evident dominance of orange MAz in the WT strain, and this orange MAz could not be found in the Δ mppG mutant (Fig. 3C). Yellow MAz Y-1 and Y-2 in the Δ mppG mutant appeared prominent due to the absence of orange MAz. Y-1 and O-2 overlapped in this TLC analysis, where the four MAz compounds were validated with the isolated compounds [28]. We need to mention that the TLC positions of O-1 and O-2 were mistakenly reversed in the previous report [28].

Fig. 3

The production of orange MAz was abolished in the ΔmppG mutant when cultured on PDA. (A) Approximately 3 cm² blocks of PDA cultures were placed on new PDA media and maintained for 5 days. (B) UV–visible spectra of the organic extracts were collected with proper dilutions in methanol, and the data were converted to the corresponding values in the original culture volume. These values are indicated as OD in the y-axis. The WT and ΔmppG strains are represented by a bold line and unfilled circles, respectively. The wavelengths characteristic of MAz is indicated as Y-axis reference lines at 390, 410, 470 and 530 nm. (C) TLC traces of the organic extracts on silica gel

HPLC analyses also indicated that orange MAz compounds were absent in the Δ mppG mutant, while approximately 5 and 8 mg of O-1 and O-2, respectively, were found in the WT strain (Fig. 4A). Notably, the level of Y-1 was comparable between the WT and Δ mppG strain, and the content of Y-1 was calculated to be approximately 6 mg from each 50 mL culture. Further analysis using a semi-preparative column also confirmed that yellow MAz compounds were absent in the Δ mppG mutant (Fig. 4B). This pattern was consistently observed in three independent trials. These experiments indicate that the biosynthesis of orange MAz involves mppG, and that the inactivation of mppG impaired orange MAz production, at least when grown on PDA plate.

Fig. 4

HPLC analysis of the organic extracts from the PDA cultures of the WT and ΔmppG strains on analytical (A) and semi-preparative (B) reverse-phase columns. For analytical column analysis in (A), 2 μL of the samples was applied from 50 mL culture extracts dissolved in 1 mL methanol. HPLC traces were monitored at 330 nm and are drawn to the same scale

We further investigated the MAz profile of the Δ mppG mutant in a PDB submerged culture. The visible spectra of the extracts showed a pattern similar to that of the PDA cultures, which was an overall reduction in the Δ mppG mutant (Fig. 5A). A TLC analysis of the WT extract indicated that orange MAz O-1 and O-2 were predominant and the yellow MAz bands were too faint to be reliably assigned (Fig. 5B). The production of orange MAz was severely impaired in the Δ mppG mutant, although O-2 was apparently present. The presence of O-2 was a distinguishable feature in the PDB experiment when compared with the PDA experiment results in Figs. 3 and 4. HPLC analyses revealed several minor peaks in the PDB extracts, and thus, the HPLC gradient elution method was slightly modified to resolve peaks relevant to this experiment. This HPLC analysis verified a severe impairment in the production of orange MAz in the extract from the Δ mppG extract (Fig. 5C). An HPLC analysis on a semi-preparative column displayed two distinctive peaks near the elution time of O-2 (Fig. 5D). These two peaks were collected from a repeat HPLC elution and were subjected to a ¹H-NMR measurement, which confirmed the identities of O-2 and MC-4 (data not shown). MC-4 is a MAz derivative that accumulates in a Δ mppC mutant [25] and was isolated as a mixture with O-2 in this experiment. MC stands for mppC, and the four MAz compounds from the Δ mppC mutant are designated MC-1 to -4 according to their elution order in a reverse-phase HPLC [25]. It better be noted that MC-4 could be produced in WT at a trace level [25]. A residual production of O-2 in the Δ mppG mutant implies that there exists another pathway that does not involve mppG, though the mppG pathway serves as the main pathway for orange MAz production. We also generated an mppG expression strain by introducing mppG under the control of trpC promoter in WT. This strain showed no significant difference from WT in the pigment production in several culture conditions tested (data not shown).

Fig. 5

The production of orange MAz was severely impaired in the ΔmppG mutant in the PDB submerged culture. (A) UV–visible spectra of the organic extracts that were collected with proper dilutions in methanol. The data were converted to represent the values in the original culture volume and are indicated as OD on the y-axis. The WT and ΔmppG strains are represented by a bold line and unfilled circles, respectively. The wavelengths characteristic of MAz is indicated as Y-axis reference lines at 390, 410, 470 and 530 nm. TLC (B) and HPLC (C, D) analyses of the WT and ΔmppG extracts. For HPLC analyses, both analytical (C) and semi-preparative (D) columns were used. For the analytical column analysis in (C), 2 μL of the samples was applied from 50 mL culture extracts dissolved in 1 mL methanol. HPLC traces that were monitored at 330 nm are drawn to the same scale

No conversion of Y-1 to O-1 occurred in a ΔMpPKS5 mutant of M. purpureus

The MAz polyketide pathway produces two distinctive structural types of yellow MAz and orange MAz. It has been assumed that one group is the precursor to the other; orange MAz is converted into yellow MAz or vice versa. Bioorganic postulation based on an aromatic polyketide biosynthetic mechanism predicted that the MAz pathway generates orange MAz that is then reduced into yellow MAz [4]. The involvement of the pyranoquinone intermediate azanigerone E supports this idea (Fig. 1). Hypothesis that yellow MAz is converted into orange MAz has also been proposed. The continuous extraction of yellow MAz resulted in its selective production over orange MAz [20]. Based on this observation, the authors suggested that yellow MAz serves as a precursor for orange MAz, and that the extraction of yellow MAz into the culture medium limits the production of orange MAz. A biochemical mechanism for this proposal was not detailed, however.

The present experiments demonstrated that mppG is involved in orange MAz biosynthesis. We propose that mppG is specifically involved in orange MAz production but not in yellow MAz. This proposal is based on a divergent pathway hypothesis (Fig. 1). However, we could not exclude the possibility that Y-1 (or Y-2) was converted into O-1 (or O-2) by the oxidase MppG, and this step is blocked in the ΔmppG mutant. To clarify this issue, we performed a bioconversion experiment of Y-1 in M. purpureus by supplying Y-1 in a PDB submerged culture of ΔMpPKS5 and then examining whether O-1 is generated or not. The ΔMpPKS5 mutant is deficient at the first stage of MAz biosynthesis and is incapable of producing any MAz compounds [4]. An HPLC analysis confirmed that a substantial amount of Y-1 was recovered from the cells, indicating that Y-1 can be taken in by M. purpureus cells (Fig. 6). A high production of citrinin is a characteristic feature of the ΔMpPKS5 mutant, with citrinin generally being found in the culture supernatant. O-1 was not found in the extract, indicating that the ΔMpPKS5 strain is incapable of converting Y-1 into O-1. We therefore conclude that there is no biochemical activity that converts Y-1 into O-1 in M. purpureus and that mppG is specifically involved in orange MAz biosynthesis.

Fig. 6

Addition of Y-1 to the M. purpureus ΔMpPKS5 PDB culture. SN denotes supernatant. Citrinin peaks are marked with asterisks. HPLC traces that were monitored at 330 nm are drawn to the same scale

This study is the first report to describe the genetic factor that selectively contributes to the biosynthesis of orange MAz in Monascus. A characteristic feature of the ΔmppG mutant strain is a loss in the productivity of orange MAz, with no significant change in the production of yellow MAZ. This metabolic feature could provide an advantage in the preparation of yellow MAZ from Monascus products, and MppG is proposed to be the oxidase that completes orange MAz biosynthesis.