An Integrated Approach to Determine the Boundaries of the Azaphilone Pigment Biosynthetic Gene Cluster of Monascus ruber M7 Grown on Potato Dextrose Agar

Abstract

Monascus-type azaphilone pigments (MonAzPs) are produced in multi-thousand ton quantities each year and used as food colorants and nutraceuticals in East Asia. Several groups, including ours, described MonAzPs biosynthesis as a highly complex pathway with many branch points, affording more than 110 MonAzP congeners in a small group of fungi in the Eurotiales order. MonAzPs biosynthetic gene clusters (BGCs) are also very complex and mosaic-like, with some genes involved in more than one pathway, while other genes playing no apparent role in MonAzPs production. Due to this complexity, MonAzPs BGCs have been delimited differently in various fungi. Since most of these predictions rely primarily on bioinformatic analyses, it is possible that genes immediately outside the currently predicted BGC borders are also involved, especially those whose function cannot be predicted from sequence similarities alone. Conversely, some peripheral genes presumed to be part of the BGC may in fact lay outside the boundaries. This study uses a combination of computational and transcriptional analyses to predict the extent of the MonAzPs BGC in Monascus ruber M7. Gene knockouts and analysis of MonAzPs production of the mutants are then used to validate the prediction, revealing that the BGC consists of 16 genes, extending from mrpigA to mrpigP. We further predict that two strains of Talaromyces marneffei, ATCC 18224 and PM1, encode an orthologous but non-syntenic MonAzPs BGC with 14 genes. This work highlights the need to use comprehensive, integrated approaches for the more precise determination of secondary metabolite BGC boundaries.

Keywords: Monascus azaphilone pigment; comparative genomics; gene cluster boundary; gene knockout; transcription analysis.

FIGURE 1

Motif-independent comparative genomic prediction of the MonAzPs BGC boundaries. (A) Homology search against the deduced proteome of T. marneffei ATCC 18224 using the MonAzPs biosynthesis-related proteins of M. ruber M7 as the bait. Dashed lines, gene pairs encoding homologous proteins (e < 1.0e^–10 for the encoded proteins). Due to mis-annotation, XP_002149763 is described in NCBI as a separate gene; however, this nucleotide sequence is in fact part of the XP_002149764 gene. (B) Local protein sequence alignment using the Smith-Waterman algorithm. Pairs of contiguous genes encoding MrPigG to MrPigK in M. ruber M7 and from XP_002149762 to XP_002149767 in T. marneffei ATCC 18224 form the seed region (R₀) for predicting the MonAzPs BGC. SW scores shown were calculated as described (Takeda et al., 2014). (C) Extension of the gene cluster. The seed region (R₀) was extended to include a total of 35 genes (Takeda et al., 2014). The symbols I_x and I_y represent the stretches of genes added to the seed region in the M. ruber M7 and the T. marneffei ATCC 18224 genomes, respectively. (D) Trimming of the BGC boundaries. i_begin and i_end, the locations of the genes at the beginning and end, respectively, of the MonAzPs gene cluster in M. ruber M7; j_begin and j_end, the corresponding gene locations in T. marneffei ATCC 18224. CB values are the maximum cumulative SW scores of the predicted BGCs with the upstream and the downstream boundaries indicated (Takeda et al., 2014).

FIGURE 2

Bioinformatic and transcriptomic analysis of the MonAzPs BGC. (A) Gene map of the MonAzPs locus and the extent of the MonAzPs BGC in different fungi according to different authors and different bioinformatic or experimental assignment methods. The predicted functions of the proteins encoded in the MonAzPs locus are also listed. (B) RT-qPCR analysis of the genes in the MonAzPs locus of the wild-type M. ruber M7 strain and its ΔmrpigB derivative, measured under identical cultication conditions. Gene expression levels in the wild-type M7 strain are taken as the basis of comparison, with the means and standard deviation calculated from measurements in three biological replicates for each strain. Statistically significant differences (p < 0.05) in expression levels are indicated by stars.

FIGURE 3

Knockout of the mrpigPdown2 gene in M. ruber M7. (A) Schematic representation of the gene knockout strategy yielding the ΔmrpigPdown2 strain. The primers, and the sizes of the corresponding PCR amplicons used to verify the gene knockout event are indicated. (B) Confirmation of the gene deletion event using PCR. Lane 1, a representative isolate of the ΔmrpigPdown2 strain; Lane 2, the wild-type strain M. ruber M7.

FIGURE 4

Colony morphology and MonAzPs production of M. ruber strains. (A) Representative colonies of the indicated M. ruber strains, grown at 28°C for 10 d on PDA plates. (B) HPLC traces (PDA, 380 nm) of crude extracts of the cultures of the indicated strains. Unlabeled peaks are unidentified M. ruber metabolites not related to MonAzPs. (C) Structures of MonAzPs.