Production of Monacolin K in Monascus pilosus: Comparison between Industrial Strains and Analysis of Its Gene Clusters

Monascus pilosus strains are widely applied to yield a cholesterol synthesis inhibitor monacolin K (MK), also called lovastatin (LOV). However, the mechanism of MK production by M. pilosus strains is still unclear. In this study, we firstly confirmed four Monascus strains, MS-1, YDJ-1, YDJ-2, and K104061, isolated from commercial MK products as M. pilosus and compared their abilities to produce MK in solid-state and liquid-state cultures. Then, we sequenced and analyzed their genomes and MK biosynthetic gene clusters (BGCs). The results revealed that the MK yields of MS-1, YDJ-1, YDJ-2, and K104061 in solid-state cultures at 14 days were 6.13, 2.03, 1.72, and 0.76 mg/g, respectively; the intracellular and extracellular MK contents of MS-1, YDJ-1, YDJ-2, and K104061 in liquid-state cultures at 14 days reached 0.9 and 1.8 mg/g, 0.38 and 0.43 mg/g, 0.30 and 0.42 mg/g, and 0.31 and 0.76 mg/g, respectively. The genome sizes of the four M. pilosus strains were about 26 Mb, containing about 7000–8000 coding genes and one MK gene cluster. The MK BGCs of MS-1, YDJ-2, and K104061 contained 11 genes, and the MK BGC of YDJ-1 contained 9 genes. According to the literature search, there are few comparisons of gene clusters and related genes responsible for the synthesis of LOV and MK. We also compared the LOV BGC in A. terreus with the MK BGCs in different species of Monascus spp., and the results revealed that although LOV and MK were the same substance, the genes responsible for the synthesis of MK were much less than those for LOV synthesis, and the gene functions were quite different. The current results laid a foundation to explore the mechanism of MK produced by Monascus spp. and compare the synthesis of LOV and MK.


Introduction
Monascus spp. are important filamentous fungi for foods and medicines, whose fermented rice product, Hongqu, also known as red yeast rice, has been used for nearly two thousand years in China and other Asian countries [1][2][3]. Monascus spp. can produce abundant secondary metabolites (SMs), such as Monascus pigments (MPs), monacolin K (MK), and γ-aminobutyric acid (GABA), and a few strains of Monascus spp. can also produce citrinin (CIT), a kidney mycotoxin [4][5][6][7][8][9], which leads to the safety issue of Monascus products. At present, the species commonly used in the production of Hongqu mainly belong to M. pilosus, M. ruber, and M. purpureus [10][11][12][13]. Research has revealed that the strains of M. pilosus can produce a large number of MK without CIT; thus, they are considered ideal producers for functional Hongqu [14,15]. Research has also shown that the different strains of M. pilosus can produce MK at various concentrations [16]. However, the mechanism of MK produced by M. pilosus is still unclear.
Fungal SMs mainly include polyketides (PKs), nonribosomal peptides (NRPs), and terpenes (TEs) [17,18], whose biosynthetic genes usually appear in the clusters [19]. PKs and NRPs are synthesized by polyketide synthase (PKS) and nonribosomal peptide synthetase A 0.5 µL volume of the spore suspensions of the four strains was, respectively, placed on the center of Petri dishes (Φ = 9 cm), including 4 media commonly used for Monascus spp.: malt extract agar (MA), Czapek yeast extract agar (CYA), potato dextrose agar (PDA), and 25% glycerol nitrate agar (G25N) [33,35]. Then, the Petri dishes were incubated at 28°C for 7 days to observe the colonial morphologies of colony size and obverse and reverse colors of the colony and aerial hyphae. The colonial size was expressed as the average values of the colonial diameters in two vertical directions.

Microscopic Morphologies of Monascus spp. Strains
A 200 µL volume of the spore suspensions of the four strains was spread on MA, CYA, PDA, and G25N media plates, respectively. The sterilized coverslips were inserted into the media at an angle of 45 • . The plates were kept at 28 • C for 7 days, then the coverslips were placed under a microscope to observe the microscopic morphologies of mycelia, conidia, and cleistothecia. Molecular identification of the strains was performed by alignment of their internal transcribed spacer (ITS) sequences. The four ITS sequences of Monascus strains were obtained by genome sequencing and blasted on the NCBI database, then sequences with higher homology were selected to construct the phylogenetic tree. The phylogenetic tree was generated in MEGA X using the Neighbor-Joining method [36,37]. Bootstrap values in the bootstrap test (1000 replicates) were shown above the branches [38]. The tree was drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Kimura 2-parameter method [39] and were in the units of the number of base substitutions per site.

Preparation of Seed Liquid
The spore suspensions prepared in Section 2.2 were inoculated into the seed culture medium with 10% (v/v) inoculum and cultured on a shaker at 150 rpm and 30 • C for 30 h [33].

Solid-State and Liquid-State Cultures
Seed liquid (10% (v/v)) was inoculated in the solid-state media, cultured at 30 • C for 60 h, then transferred to 24 • C to continue to be cultured for 14 days [34]. The samples were taken on the 4th, 8th, 12th, and 14th days then dried at 45 • C and crushed.
Seed liquid (10% (v/v)) was inoculated in the liquid-state media, cultured at 30 • C and 110 rpm for 3 days, then transferred to 25 • C to continue to culture for 14 days. The samples were taken on the 4th, 8th, 12th, and 14th days [33].
2.5. MK and CIT Analysis 2.5.1. MK and CIT Extraction of Solid-State Samples MK extraction: 0.1 g of dried solid-state samples were dried to a constant weight at 40 • C, suspended in 10 mL of 75% (v/v) ethanol solution, and subjected to ultrasonication treatment (KQ-250B, Kunshan, China) for 60 min then centrifuged at 8000 rpm for 15 min. The supernatant was collected and filtered through a 0.22 µm microfiltration membrane.
CIT extraction: 0.3 g of dried solid-state samples were suspended in 3 mL of 80% (v/v) methanol solution and subjected to ultrasonication treatment for 40 min then centrifuged at 8000 rpm for 15 min, and the supernatant was gathered. Another 3 mL of 80% methanol was added to the precipitate. Both supernatants were combined, diluted to 10 mL, and filtered through a 0.22 µm microfiltration membrane after ultrasonic extraction for 20 min and centrifugation.

MK and CIT Extraction of Liquid-State Samples
Intracellular MK/CIT extraction: 0.1 g/0.3 g of freeze-dried mycelia were taken, and the extraction steps are the same as the MK/CIT extraction of the solid-state samples in Section 2.5.1.
Extracellular MK extraction: after the mycelia were filtered, absolute ethanol was added at the ratio of 1:3 (fermentation broth/absolute ethanol). After standing still for 30 min, the mixture was centrifuged at 10,000 rpm for 10 min. The supernatant was collected and filtered through a 0.22 µm microfiltration membrane. Extracellular CIT extraction: the mycelia were filtered to obtain the clarified fermentation broth, which was filtered through a 0.22 µm microfiltration membrane for further analysis.

MK and CIT Detection
MK and CIT contents were detected by high-performance liquid chromatography (HPLC, Shimadzu LC-20AT, Kyoto, Japan), equipped with a C 18 column (inertsil ODS-3 4.6 × 250 mm) by means of a diode array detector. A 20 µL volume of sample extract solution was injected into HPLC to detect MK and CIT. Both acid and lactone forms of MK were calculated as the MK yield. The detailed HPLC parameters were as follows.

Statistical Analyses
Statistical analyses were performed with SPSS (version 16.0) to calculate the means, standard errors, and standard deviations. The statistical significance was calculated by one-way analysis of variance (ANOVA), with significance levels set at p = 0.05.

DNA Sequencing and Assembly
DNA samples prepared in Section 2.6 were randomly broken into fragments of the length required to construct DNA libraries. NEBNext ® Ultra™ DNA Library Prep Kit for Illumina and SMRT bell TM Template kit 1.0 were used to construct the Illumina library and 20K SMRT Bell library, respectively. Illumina NovaSeq PE150 and PacBio Sequel platforms were applied for the whole-genome sequencing, then the raw data were valued by FastQC [41], and SMRT Link v5.1.0 software [42,43] was utilized to assemble genomes.

Annotation and Analysis of Monascus spp. Genomes
Based on the sequence information of the four strains in this research and other genomes of Monascus spp., which have been released in NCBI and JGI, prediction of the coding genes was performed with Augustus [44]. The SM BGCs were predicted by Anti-SMASH 5 for fungi [45]. The Pfam database (http://pfam.xfam.org/, accessed on 3 December 2019) and the conserved domain database (CDD, https://www.ncbi.nlm.nih.gov/cdd, accessed on 3 September 2020) were used to predict and analyze the conserved domains to re-predict the gene functions in the LOV and MK BGCs. The Geneious software [46] was used to analyze the sequence similarity of the genes in the LOV and MK BGCs by the Geneious alignment method, and the parameters were set as default.

Classification and Identification
All four strains of Monascus spp., MS-1, YDJ-1, YDJ-2, and K104061, used in the current study were from the production plants of Monascus products [33,34], so their taxonomic status was reidentified based on the morphological and molecular identification methods. As shown in Figure 1 and Table 1, on the seventh day, on MA media, the reverse surfaces of colonies were yellow at the margins and deep orange at the centers; on CYA media, colonies were irregular in shape; on PDA media, the edges of colonies were light yellow to golden yellow; on G25N media, colonies were floccose, mycelia were white, and the reverse was uncolored. The microscopic morphologies of the four strains showed that there were cleistothecia and conidia on MA and PDA media, while only conidia could be observed on CYA and G25N media. The morphological characteristics of the strains were similar to those described in the literature of M. pilosus [35,47].

Colonial Morphologies
The colonial diameters reached 40-48 mm. Their obverse and reverse surfaces were white and light orange to orange-red, respectively. Their mycelia were compact and dense.
The colonial diameters reached 50-65 mm. Their shapes were irregular. Their obverse and reverse surfaces were white and red to dark red with radial folds. Their mycelia were sparse and short.
The colonial diameters reached 40-55 mm. Their reverse surfaces were light orange or orange-red and the edges were light yellow to golden yellow. Their middle parts were raised, and their mycelia were dense and fluffy.
The colonial diameters reached 11-25 mm. Their obverse and reverse surfaces were white and colorless.

Microscopic Morphologies
Cleistothecia arose singly at the tips of stalk-like hyphae and walls were hyaline or pale brown. Conidia were hyaline and borne laterally on pedicels and terminally on hyphae, arising singly or occasionally in short chains, obpyriform to globose.
No cleistothecium; conidia were transparent or brown and obpyriform to globose.
Cleistothecia were globose and arose singly from distinct stalk-like hyphae. Conidia were spherical or upside-down pear-shaped with colorless or brown colors.

Molecular Identification of Monascus spp. Strains
The phylogenetic analysis in this study used ITS sequences (the four ITS sequences of Monascus strains were obtained by genome sequencing, and the other ITS sequences were from NCBI). The evolutionary history was inferred using the Neighbor-Joining method. Bootstrap values were above branches, and only those above 60% were indicated. The strains of Penicillium griseum and Aspergillus fischeri which were farther from the tested strains were used as the outgroup.
As shown in Figure  pedicels and terminally on hyphae, arising singly or occasionally in short chains, obpyriform to globose.
nidia were spherical or upside-down pear-shaped with colorless or brown colors.

Molecular Identification of Monascus spp. Strains
The phylogenetic analysis in this study used ITS sequences (the four ITS sequences of Monascus strains were obtained by genome sequencing, and the other ITS sequences were from NCBI). The evolutionary history was inferred using the Neighbor-Joining method. Bootstrap values were above branches, and only those above 60% were indicated. The strains of Penicillium griseum and Aspergillus fischeri which were farther from the tested strains were used as the outgroup.
As shown in Figure    Combined with the results of morphological and molecular identification, the Monascus strains, MS-1, YDJ-1, YDJ-2, and K10406, isolated from factories and used in this study were identified as M. pilosus.

MK and CIT Production in Solid-State and Liquid-State Cultures
According to the formula of solid and liquid culture media, stage-variable temperature culture was used and samples were taken on the 4th, 8th, 12th, and 14th days of the culture process.
As shown in Figure 3, MS-1 could produce the highest concentration of MK in both solid-state and liquid-state cultures. In solid-state cultures at 14 days, the MK yields of, MS-1, YDJ-1, YDJ-2, and K104061 reached 6.13, 2.03, 1.72, and 0.76 mg/g, respectively. In liquid-state cultures at 14 days, the intracellular and extracellular MK contents of MS-1, YDJ-1, YDJ-2, and K104061 were 0.9 and 1.8 mg/g, 0.38 and 0.43 mg/g, 0.30 and 0.42 mg/g, and 0.31 and 0.76 mg/g, respectively. CIT was detected neither in solid-state nor in liquid-state cultures for all tested Monascus strains (data not shown).

MK and CIT Production in Solid-State and Liquid-State Cultures
According to the formula of solid and liquid culture media, stage-variable temperature culture was used and samples were taken on the 4th, 8th, 12th, and 14th days of the culture process.
As shown in Figure 3, MS-1 could produce the highest concentration of MK in both solid-state and liquid-state cultures. In solid-state cultures at 14 days, the MK yields of, MS-1, YDJ-1, YDJ-2, and K104061 reached 6.13, 2.03, 1.72, and 0.76 mg/g, respectively. In liquid-state cultures at 14 days, the intracellular and extracellular MK contents of MS-1, YDJ-1, YDJ-2, and K104061 were 0.9 and 1.8 mg/g, 0.38 and 0.43 mg/g, 0.30 and 0.42 mg/g, and 0.31 and 0.76 mg/g, respectively. CIT was detected neither in solid-state nor in liquidstate cultures for all tested Monascus strains (data not shown).

Genome Sequencing and Prediction of SM Gene Clusters
The four Monascus strains were sequenced and analyzed. The results ( Table 2) showed that the genome sizes were roughly 26Mb, the G+C mole percentages were approximately 49%, and the coding genes varied from 7634 to 7771, respectively. AntiSMASH 5 was used to predict the SMs gene clusters in the four genomes, and the results (Table S1) showed that five types of SMs were predicted. Furthermore, the MK BGCs appeared in their genomes but no CIT BGCs.

Function Re-Prediction of the Genes in the LOV and MK BGCs Reported Previously
In 1999, Kennedy [48] reported a BGC related to LOV BGC in the genome of Aspergillus terreus ATCC 20542, which contained total 18 genes including 7 unknown functional ones at that time (Table 3), and in 2013, Xu et al. renamed one unknown gene, ORF5 as lovG [49]. In 2008, Chen and collaborators [50] reported a BGC related to

Genome Sequencing and Prediction of SM Gene Clusters
The four Monascus strains were sequenced and analyzed. The results ( Table 2) showed that the genome sizes were roughly 26 Mb, the G+C mole percentages were approximately 49%, and the coding genes varied from 7634 to 7771, respectively. AntiSMASH 5 was used to predict the SMs gene clusters in the four genomes, and the results (Table S1) showed that five types of SMs were predicted. Furthermore, the MK BGCs appeared in their genomes but no CIT BGCs.

Function Re-Prediction of the Genes in the LOV and MK BGCs Reported Previously
In 1999, Kennedy [48] reported a BGC related to LOV BGC in the genome of Aspergillus terreus ATCC 20542, which contained total 18 genes including 7 unknown functional ones at that time (Table 3) in M. pilosus BCRC38072, which only contained 9 genes (Table 3). In this research, we re-predicted functions of genes in LOV and MK BGCs by Pfam and CDD analysis, the results showed that in LOV BGC, all of ORF2, ORF10, and ORF16 were re-predicted as transporters, ORF14, ORF15, and ORF18 were re-predicted as mitochondrial carrier protein, dehydratase and glycosyl hydrolase, respectively, while ORF12 was still unknown, and lovG and mkD were re-predicted as α/β hydrolase (Table 3). Table 3. Re-prediction of functions of genes in lovastatin (LOV) and MK biosynthetic gene clusters (BGCs).

Comparison of MK BGCs
Based on the LOV and MK BGCs in Table 3, we compared twelve genomes of Monascus spp. including eight published genomes, M. purpureus YY-1, YJX-8, GB-01, HQ1, NRRL1596, and M. ruber FWB13, CBS127566, NRRL 1597 [29][30][31][32], and four genomes of M. pilosus sequenced in this study, and found that all genomes of four M. pilosus strains and three M. ruber strains contained the MK BGCs while there was no MK BGC in those of five M. purpureus strains. Total 9 or 10 genes named mkA-mkI were highly conserved in the LOV and MK BGCs (Table 4, Figure 4). However, there were some unique genes in MK BGCs of Monascus spp. For example, in YDJ-2, Both of mkF and mkG were combined into be one gene mkF-G, and mkC in M. pilosus BCRC38072 [50] was predicted to be one gene mkC in YDJ-1 or to be two independent genes mkC1 and mkC2 in other MK BGCs. In addition, there was an extra gene mkJ in the MK BGCs of MS-1, YDJ-1, YDJ-2, K104061, and NRRL1597, which neither existed in the LOV BGC of A. terreus ATCC 20542 [48] nor in the MK BGCs of FWB13, CBS127566, and BCRC38072. Based on KOG annotation result obtained from JGI, MKJ was one animal-type fatty acid synthase and related protein. It was worth noting that although lovF/mkB did not exist in the MK BGCs of NRRL1597, it was located elsewhere in its genome, and this situation also occurred in mkJ in the FWB13 and CBS127566 genomes (data not shown). The numbers with "%" are the similarity percentages between mkA-mkI from Monascus spp. investigated in this study and the corresponding genes in the LOV/MK BGCs reported previously [48,50].

Multiple sequence alignment of LOV and MK BGCs
A multiple sequence alignment of each gene in LOV and MK BGCs (Figure 4) revealed that genes in MK BGCs were quite different from those in LOV BGC, and the genes in MK BGCs from the same species of Monascus spp. showed higher homology ( Figures  S1-S10). Among MK BGCs of the strains of M. pilosus, the 1012th histidine of MKB in YDJ-2 was mutated to arginine (Figure 5a), and the 77th glycine of MKD in MS-1 was mutated to serine (Figure 5b). The arginine at position 259 of MKE was mutated to cysteine in YDJ-2 ( Figure 5c).

Multiple sequence alignment of LOV and MK BGCs
A multiple sequence alignment of each gene in LOV and MK BGCs (Figure 4) revealed that genes in MK BGCs were quite different from those in LOV BGC, and the genes in MK BGCs from the same species of Monascus spp. showed higher homology (Figures S1-S10). Among MK BGCs of the strains of M. pilosus, the 1012th histidine of MKB in YDJ-2 was mutated to arginine (Figure 5a), and the 77th glycine of MKD in MS-1 was mutated to serine (Figure 5b). The arginine at position 259 of MKE was mutated to cysteine in YDJ-2 ( Figure 5c). Further, we analyzed if these amino acid mutations occurred on the active or binding sites of MKB in YDJ-2, MKD in MS-1, and MKE in YDJ-2 and affected their functions. We found that the amino acid mutations in MKB, MKD, and MKE did not occur in the active or binding sites (Figures S11-S13, Table S2) and did not affect their functions.

Discussion
In 1979, Endo identified a substance from the fermentation broth of M. ruber and named it MK that could inhibit cholesterol synthesis [51]. In 1980, Albert [52] discovered a similar substance from A. terreus and named it mevinolin. Subsequent research has proven that mevinolin and MK are the same substance, and now, both of them are often referred to collectively as lovastatin (LOV) [53]. Although MK and LOV are the same substance, there were also some differences among the MK BGCs of Monascus spp. and LOV BGC. In addition, different species of Monascus spp. contained different MK BGCs, and there were also differences among genes related to MK synthesis in Monascus spp. In this research, the SM BGC prediction results of Monascus spp. showed that the strains of M. ruber and M. pilosus contained MK BGCs. There were 10 genes in each MK BGC of M. ruber FWB13, CBS127566, and NRRL 1597. Among M. pilosus strains, there were 11 genes in each MK BGC of MS-1, YDJ-1, and K104061 and 9 genes in the MK BGC of YDJ-2 (Table 4, Figure 4), while 18 genes were reported responsible for the LOV biosynthesis [48], in which there were 9 genes that may be essential and conserved genes for the biosynthesis of MK. In the MK BGC of M. ruber NRRL 1597, mkB was absent, but there was an extra gene, mkJ, which was related to the synthesis of animal-type fatty acid; mkC in M. pilosus BCRC38072 and YDJ-1 were predicted to be two independent genes mkC1 and mkC2 in other strains of Monascus spp., whose functions were the same as mkC; in YDJ-2, mkF and mkG were combined into one gene, mkF-G, with the functions of both of mkF and mkG.
The MK BGCs of the four strains of M. pilosus studied in this research contained the key genes related to the MK synthesis [50]; thus, theoretically, all four strains had the ability to produce MK. According to the results of solid-state and liquid-state cultures, all strains of MS-1, YDJ-1, YDJ-2, and K104061 could indeed produce MK at various concentrations, and MS-1 had the strongest ability to yield MK (Figure 3). The results of multiple sequence alignment revealed that there were amino acid mutations in MKB of YDJ-2, Further, we analyzed if these amino acid mutations occurred on the active or binding sites of MKB in YDJ-2, MKD in MS-1, and MKE in YDJ-2 and affected their functions. We found that the amino acid mutations in MKB, MKD, and MKE did not occur in the active or binding sites (Figures S11-S13, Table S2) and did not affect their functions.

Discussion
In 1979, Endo identified a substance from the fermentation broth of M. ruber and named it MK that could inhibit cholesterol synthesis [51]. In 1980, Albert [52] discovered a similar substance from A. terreus and named it mevinolin. Subsequent research has proven that mevinolin and MK are the same substance, and now, both of them are often referred to collectively as lovastatin (LOV) [53]. Although MK and LOV are the same substance, there were also some differences among the MK BGCs of Monascus spp. and LOV BGC. In addition, different species of Monascus spp. contained different MK BGCs, and there were also differences among genes related to MK synthesis in Monascus spp. In this research, the SM BGC prediction results of Monascus spp. showed that the strains of M. ruber and M. pilosus contained MK BGCs. There were 10 genes in each MK BGC of M. ruber FWB13, CBS127566, and NRRL 1597. Among M. pilosus strains, there were 11 genes in each MK BGC of MS-1, YDJ-1, and K104061 and 9 genes in the MK BGC of YDJ-2 (Table 4, Figure 4), while 18 genes were reported responsible for the LOV biosynthesis [48], in which there were 9 genes that may be essential and conserved genes for the biosynthesis of MK. In the MK BGC of M. ruber NRRL 1597, mkB was absent, but there was an extra gene, mkJ, which was related to the synthesis of animal-type fatty acid; mkC in M. pilosus BCRC38072 and YDJ-1 were predicted to be two independent genes mkC1 and mkC2 in other strains of Monascus spp., whose functions were the same as mkC; in YDJ-2, mkF and mkG were combined into one gene, mkF-G, with the functions of both of mkF and mkG.
The MK BGCs of the four strains of M. pilosus studied in this research contained the key genes related to the MK synthesis [50]; thus, theoretically, all four strains had the ability to produce MK. According to the results of solid-state and liquid-state cultures, all strains of MS-1, YDJ-1, YDJ-2, and K104061 could indeed produce MK at various concentrations, and MS-1 had the strongest ability to yield MK (Figure 3). The results of multiple sequence alignment revealed that there were amino acid mutations in MKB of YDJ-2, MKD of MS-1, and MKE of YDJ-2 ( Figure 5), but these mutations did not occur at the active or binding sites of these proteins (Figures S11-S13 , Table S2). We further analyzed the transcription of these genes in the four MK BGCs, and the results showed that each gene was expressed to varying degrees (data not shown), which could not figure out the reason for the difference to yield MK of the four strains, yet. However, the effect of a mutation also depends on the position of the amino acid in the 3D structure [54]. Later, the 3D structure of the mutant proteins can be simulated and compared by related experiments to further explore the effect of the mutations on the function of these proteins. In addition, transcriptomic analysis or other methods should be used to explore the differences in MK production of the four strains.
In conclusion, all Monascus strains, MS-1, YDJ-1, YDJ-2, and K104061, were identified as M. pilosus and had the ability to produce MK. The MK BGCs identified in the four strains involved 9-11 genes, in which 9 essential genes responsible for the MK biosynthesis were highly conserved in M. pilosus. Genes responsible for the synthesis of MK were much less than those of LOV, whose functions were also not the same. The results of this study may provide a theoretical basis to explore the mechanism of MK produced by Monascus spp. and compare the synthesis of LOV and MK.
Supplementary Materials: The following are available at https://www.mdpi.com/article/10.339 0/microorganisms9040747/s1, Table S1: Prediction of secondary metabolite gene clusters of four M. pilosus strains, Table S2: S2-1 Active sites in MKB of the MK biosynthetic gene cluster in MS-1, S2-2: Binding sites in MKB of the MK biosynthetic gene cluster in MS-1, S2-3: Binding sites in MKE of the MK biosynthetic gene cluster in MS-1. Figure S1: Multiple sequence alignment of gene mkA, Figure S2: Multiple sequence alignment of gene mkB, Figure S3: Multiple sequence alignment of gene mkC1, Figure S4: Multiple sequence alignment of gene mkC2, Figure S5: Multiple sequence alignment of gene mkD, Figure S6: Multiple sequence alignment of gene mkE, Figure S7: Multiple sequence alignment of gene mkF, Figure S8: Multiple sequence alignment of gene mkG, Figure S9: Multiple sequence alignment of gene mkH, Figure S10: Multiple sequence alignment of gene mkI. Figure S11: Prediction of active and binding sites in MKB, Figure S12: Prediction of active and binding sites in MKD, Figure S13: Prediction of active and binding sites in MKE.
Author Contributions: W.D. designed and carried out the present research work, conducted experiments, analyzed the data, and wrote the present manuscript. Y.S. provided the strains of Monascus spp. and supervised the experiment. F.C. provided a place in the laboratory and gave access to the lab facilities for experimentation and funds for the present work. All authors have read and agreed to the published version of the manuscript.