Effects of Different G-Protein α-Subunits on Growth, Development and Secondary Metabolism of Monascus ruber M7

Strains of Monascus filamentous fungal species have been used to produce fermented foods in Asian countries, such as China, Japan, and The Korean Peninsula, for nearly 2,000 years. At present, their fermented products are widely used as food additives and nutraceutical supplements worldwide owing to their production of beneficial secondary metabolites. Heterotrimeric G-protein signaling pathways participate in regulating multiple biological processes in fungi. Previously, we identified three Monascus ruber M7 G-protein α subunits (Mga1–3) and demonstrated that Mga1 can regulate growth, reproduction and some secondary metabolites’ production. Here, we systematically analyzed and compared the roles of mga1–3 by combining single- and double-gene(s) knockouts and their transcriptomic data. First, mga2 and mga3 knock-out mutants and pairwise combinations of mga1–3 deletion strains were generated. Then the changes in growth, development and the main secondary metabolites, Monascus pigments and citrinin, in these mutants were systematically compared with M. ruber M7. Moreover, RNA-Seq analyses of these mutants were performed. All three Gα subunits worked together to regulate biological processes in M. ruber M7, with Mga1 playing a major role, while Mga2 and Mga3 playing supplemental roles. According to the existing literatures which we can find, gene knock-out mutants of the pairwise combination of mga1–3 and their transcriptome analysis are first reported in this study. The current results have clearly demonstrated the functional division of Mga1–3 in M. ruber M7, and could provide a deeper understanding of the effects of different Gα subunits on growth, development and secondary metabolism in other filamentous fungi.

The genes involved in SMs biosynthesis in filamentous fungi usually appear as gene clusters (Schwecke et al., 1995;Blin et al., 2015). In the past decade, the gene clusters of MPs, CIT and MK in Monascus spp. have been identified and their biosynthetic pathways have been fully illustrated (Chen et al., 2008;Li et al., 2015;He and Cox, 2016;Liu J. et al., 2016;Chen et al., 2017). The biosynthesis of these SMs cannot only be controlled by the intra-cluster regulating genes but can also be adjusted by the off-cluster global regulating genes, such as LaeA, VeA, and related genes in the G-protein signaling pathway (GPSP) (Fox and Howlett, 2008;Liu Q. et al., 2016;Lin et al., 2018). GPSPs, including the G-protein coupled receptor (GPCR), heterotrimeric G-protein (G-protein) and downstream effectors (Seo and Yu, 2006), play vital roles in growth, differentiation, SMs biosynthesis, pathogenicity and toxicity in filamentous fungi (Yu et al., 2008;Corrochano et al., 2016;Moretti et al., 2017;Liu et al., 2018;van den Hoogen et al., 2018).
In our previous study, the roles of the G protein α subunit gene mga1 (Gα 1 gene) in wild-type M. ruber M7 were analyzed, and mga1 can comprehensively regulate growth, reproduction, and MPs and CIT production (Li et al., 2010). Here, the other two Gα genes, mga2 (Gα 2 gene) and mga3 (Gα 3 gene), and the pairwise combinations of mga1-3 were independently deleted in M. ruber M7. The morphological observations and fermentation experiments of six Gα genes' mutants, mga1, mga2, mga3, mga1+2, mga1+3, and mga2+3, as well as their RNA-Seq analyses, were conducted to systematically investigate the functions of the Gα subunits in M. ruber M7. And we have found that all three Gα subunits work together to regulate extensive biological processes in M. ruber M7, Mga1 playing a major role and Mga2 and Mga3 as supplementary roles. In detail, during vegetative growth, Mga1 is the essential positive regulator, while Mga2 and Mga3 can enhance the regulatory process when either was double deleted with Mga1. However, a single deletion of Mga2 or Mga3 has little effect. Mga1 contributes the most to the regulation of sexual/asexual reproduction, and the regulation of asexual reproduction may occur prior to the central regulatory pathway. Different Gα subunits can be combined to negatively regulate secondary metabolism. Mga1 and Mga2 can negatively regulate MPs and CIT production individually or jointly, and Mga3 may work in combination with Mga1 to negatively enhance regulation of MPs production. These findings not only illuminate the functions of different Gα subunits in M. ruber M7 but could also provide a deeper understanding of the effects of different Gα subunits on growth, development and secondary metabolism in other filamentous fungi.

Strains and Media
Monascus ruber M7 (CCAM 070120, Culture Collection of State Key Laboratory of Agricultural Microbiology, China Center for Type Culture Collection, Wuhan, China) (Chen and Hu, 2005) was used to generate the gene knockout strains mga2 and mga3. The mga1 strain obtained by Li in our laboratory (Li et al., 2010) was used to generate the double-deletion strains mga1+2 and mga1+3. The mga2 strain obtained in this study was used to generate the double-deletion strain mga2+3.

Deletion of Gα Genes in M. ruber M7
The homologous gene recombination strategy was used to construct the deletion strains ( mga2, mga3, mga1+2, mga1+3, and mga2+3). The hygromycin resistence gene hph was used in the mga2-deletion cassette to construct the mga2 strain, while the G418 resistence gene neo was used in the mga3-deletion cassette to construct the mga3 strain. The mga1 strain had been constructed previously (Li et al., 2010). The mga1 strain and another mga2-deletion cassette with the neo gene were used to construct the double-deletion strain mga1+2. The mga1 strain and the mga3-deletion cassette were used to construct the double-deletion strain mga1+3. The mga2 strain and the mga3-deletion cassette were used to construct the double-deletion strain mga2+3. The gene deletion cassette was constructed by double-joint PCR, as shown in Supplementary Figures S1, S2, using the primers listed in Supplementary Table S1. The construction strategy for the complementary strains is also shown in Supplementary Figure S1. The mutants were generated using an Agrobacterium tumefaciens-mediated transformation method that was previously established in our laboratory (Li et al., 2010). The genotypes of deletion strains were confirmed using PCR amplification and Southern hybridization.

Southern Hybridization
Southern hybridization was performed according to a previously reported method (Liu et al., 2014) using a DIG-High Prime DNA Labeling and Detection Starter Kit I (Roche, Germany). Fragments of mga2 [open reading frame (ORF), probe 1], hph (selective marker gene, probe 2), mga3 (ORF, probe 3), and neo (selective marker gene, probe 4) were independently amplified to be used as probes. The single-deletion mutants' DNAs were digested by SacI and XhoI. The double-deletion mutants' DNAs were digested by KpnI. Primers are listed in Supplementary Table S1.

Phenotypic Analysis
Monascus ruber M7, mga1, mga2, mga3, mga1+2, mga1+3, and mga2+3 strains were cultivated on PDA, CYA, MA and G25N at 28 • C to observe their phenotypes. The colony sizes of these strains were measured after cultivated for 12 days, and the cleistothecia or conidia were observed and counted after cultivated for 5 days. Three replicates are for each strain.
Freshly harvested conidiospores (10 5 conidia mL −1 ) of M. ruber M7 and Gα-deleted strains were inoculated on PDA medium, covered with cellophane and incubated at 28 • C for 11 days. The mycelia and medium were sampled every other day from 3 to 11 days to analyze the intracellular and extracellular MPs and CIT production levels (Li et al., 2014).

RNA-Seq Analysis
Monascus ruber M7 and Gα mutant strains were independently inoculated on PDA medium, covered with cellophane and incubated at 28 • C. Two replicates were conducted for each strain. At 3 and 7 days, mycelia were collected for RNA extraction and then sequenced using the BGIseq-500RS platform (BGI, Wuhan, China 1 ). The expression levels of 11 randomly selected genes in M. ruber M7 and mga1+3 were assessed using quantitative real-time PCR (qRT-PCR) to confirm the reliability of the RNA-Seq results.
Monascus ruber M7 genome which contains 8,407 genes, was used as a reference genome  to calculate the BLAST rate of the genome, and clean data were aligned using Hierarchical Indexing for Spliced Alignment of Transcripts and bowtie2 (Langmead and Salzberg, 2012;Kim et al., 2015). Then, RNA-Seq by Expectation Maximization was used to calculate the expression level of each gene (Li and Dewey, 2011). The genes that possessed an expression differential multiple greater than 1, as well as a Q-value not greater than 0.001 (Benjamini and Hochberg, 1995;Storey and Tibshirani, 2003), were selected as differentially expressed genes (DEGs).
Gene ontology (GO) 2 and a KEGG pathway 3 functional analysis were used to investigate the functions of the DEGs between M. ruber M7 and Gα mutants. Moreover, the DEGs involved in fungal growth, sporulation and secondary metabolism were further analyzed to determine the roles of Gα subunits in development and secondary metabolism of M. ruber M7.

Targeted Deletion of Gα Genes in M. ruber M7
Single-deletion strains mga2 with hygromycin resistance and mga3 with G418 resistance, as well as double-deletion strains mga1+2, mga1+3, and mga2+3, were obtained. For mga2 strain, the PCR analysis confirmed the existence of the hph sequence as well as the absence of the mga2 ORF. Southern hybridization showed a single copy of the hph sequence in the mga2 strain. For mga3 strain, the PCR analysis confirmed the existence of the neo sequence as well as the absence of the mga3 ORF. Southern hybridization showed a single copy of the neo sequence in the mga3 strain. For mga1+2 strain, the PCR analysis confirmed the existence of the hph and neo sequence as well as the absence of the mga1 and mga2 ORF. Southern hybridization showed a single copy of the neo sequence in the mga1+2 strain. For mga1+3 strain, the PCR analysis confirmed the existence of the hph and neo sequence as well as the absence of the mga1 and mga3 ORF. Southern hybridization showed a single copy of the neo sequence in the mga1+3 strain. For mga2+3 strain, the PCR analysis confirmed the existence of the hph and neo sequence as well as the absence of the mga2 and mga3 ORF. Southern hybridization showed a single copy of the neo sequence in the mga2+3 strain. The results of PCR analysis and Southern hybridization were displayed in Supplementary Figures S1, S2. Additionally, the corresponding complementation strains were also obtained. The complementation strains possessed phenotypic characteristics similar to those of M. ruber M7 (Supplementary Figure S3).

Vegetative Growth and Reproduction
The phenotypes of the six Gα mutants, mga1 (prepared by Li et al. (2010)), mga2, mga3, mga1+2, mga1+3, and mga2+3, were compared with M. ruber M7. As shown in Figure 1, after cultivation on PDA medium for 12 days, the colony sizes of the mga1, mga2, mga3 and mga2+3 strains were similar to that of M. ruber M7, while those of mga1+2 and mga1+3 were about 45% and 80% smaller than M. ruber M7, respectively. We found that when a single Gα gene (mga1, mga2, or mga3) was deleted, the colony sizes did not significantly change. However, when the mga2 or mga3 gene was deleted in the mga1 strain, the colony sizes were smaller than other mutants.

DEG Analysis, Annotation and Functional Classification
The RNAs of M. ruber M7 and six deletion mutants ( mga1, mga2, mga3, mga1+2, mga1+3, and mga2+3) were independently extracted for a further transcriptomic analysis. The obtained clean sequence reads of the 14 samples were validated by qRT-PCR. In total, 11 genes in the M. ruber M7 genome were randomly selected for relative gene expression comparisons between M7 and mga 1+3 strain, the selected genes are listed in Supplementary Table S2. As shown in Supplementary Figure S4, the relative expression levels of these 11 random genes had the same trends as in the RNA-Seq, which indicated that the transcriptome sequencing was reliable.

DEGs Analyses and Transcriptome Classification
The genes that possessed an expression differential multiple greater than 1, as well as a Q-value not greater than 0.001, were selected as DEGs. Compared with M. ruber M7, different mutants at different time points have diverse trends in their numbers of DEGs. There were greater numbers of DEGs in double-deletion strains than in single-deletion strains. In particular, in the mga1+3 strain, 1,858 and 2,000 genes showed down-regulated expression levels at 3 rd day and 7 th day, respectively, which were much greater numbers than those in the corresponding single-deletion strains mga1 and mga3. This may explain the distinctive phenotype of the mga1+3 strain (Figure 2).
A GO enrichment analysis of DEGs was performed. GO has three ontologies: molecular biological function, cellular component and biological process. For each ontology, the functional enrichment was determined. Compared with M. ruber M7, metabolic process possessed the most DEGs in all the mutants at both 3 and 7 days, and most genes in this GO ontology were down-regulated, such as in cellular process, cell part and catalytic activity. Among the KEGG pathways, the metabolic pathway possessed the most DEGs in all the mutants at both 3 rd and 7 th day, and most genes in this KEGG pathway were down-regulated, including those involved in meiosis-yeast, SMs biosynthesis and carbon metabolism.

Gα Genes Positively Regulate Vegetative Growth
The DEGs of carbon and nitrogen source metabolism are listed in Supplementary Table S3. The regulation of carbon source metabolism mostly focuses on the tricarboxylic acid cycle (TCA cycle), meanwhile many major facilitator superfamily (MFS) transporters were down-regulated. RNA-Seq results revealed that the absence of Gα subunits generally depressed the TCA cycle, especially reducing the biosynthesis of citric acid and succinyl CoA. The absence of both Mga1 and Mga3 regulated most genes in the TCA cycle. Data on the DEGs related to the TCA cycle are presented in Supplementary Figure S5. On the basis of the GO and KEGG analyses, we also analyzed the influence of different Gα genes on nitrogen metabolism. The expression levels of genes related to nitrogen metabolism mostly decreased in the mutants, with mga1+2 and mga1+3 possessing the greatest numbers of DEGs related to nitrogen metabolism. This indicates that all the Gα genes positively regulated nitrogen metabolism. Data on DEGs related to nitrogen source metabolism are presented in Supplementary Figure S6. The decreased expression of vegetative growth-related genes corresponded to the repressed colony sizes of the mga1+2 and mga1+3 strains (Figure 1).

Gα Genes Play Different Roles in Sexual and Asexual Reproduction
In filamentous fungi, the central regulatory pathway of conidiospore formation generally consists of abaA, brlA and wetA genes (Yu, 2006). The most reported sexual reproductionrelated genes are mating type (MAT)-related genes (Varga et al., 2014). In addition, the velvet family genes are related to sexual/asexual reproduction (Yu et al., 2008;Liu Q. et al., 2016). The expression changes in all these genes as determined by the DEGs analysis are listed in Supplementary Figure S7.
Compared with M. ruber M7, the expression level of the cleistothecia-related gene MAT1-2 was decreased only in mga1deleted strains ( mga1, mga1+2, and mga1+3), while their expression levels increased in mga3 at 3 days. They were not changed in the mga2 and mga2+3 strains. This explained why cleistothecia were not found in mga1, mga1+2, and mga1+3 strains (Figure 2), and it suggested that Mga1 positively regulates sexual reproduction while Mga2 and Mga3 have slight effect in sexual reproduction. However, most genes involved in conidial production, including the conidiospore formation genes brlA and wetA, had increased expression levels in the Gα mutants, except in mga1+3. Only the expression levels of velvet regulators were decreased in almost all the mutants. This is different from the phenotypic analysis (Figure 2) that the conidia-forming ability was reduced in mga1related mutants ( mga1, mga1+2, and mga1+3) but not in the other mutants.

Gα Genes Negatively Regulate MPs and CIT Biosynthesis
RNA-Seq results revealed that the expression levels of MPs biosynthetic genes , except MpigL, were increased in all the mutants at 7 th day. However, these genes in mga1+2 and mga1+3 were up-regulated at 3 rd and 7 th day. In addition, in the mga1-deleted strains ( mga1, mga1+2, and mga1+3) more genes were up-regulated than those in the other mutants. The DEGs involved in MPs biosynthesis are shown in Supplementary Figure S8. This result matches the increased MPs yields in Gα mutants (Figure 3) and indicates that Gα negatively regulates MPs production by regulating the MPs biosynthetic gene cluster.
According to the RNA-Seq results, genes in the CIT gene cluster (He and Cox, 2016) showed different trends on different days. Gα subunits mainly regulated the expression of the CIT gene cluster at 3 rd day. Most genes in the CIT biosynthetic gene cluster were up-regulated in the single-deletion mutants and mga1+2, and CIT production also increased in these mutants (Figure 3). In mga1+3, although the pksCT gene was up-regulated, most other genes (MRR1-4 and MRR7-8) in the cluster were down-regulated, and the early stage (3-7 days) CIT production in the mga1+3 strain was lower than that in M. ruber M7. In mga2+3, only pksCT and MRL2 were up-regulated, and only in the later stage (9-11 days) the CIT production was greater than that in M. ruber M7 (Figure 3). Data on DEGs involved in CIT biosynthesis are provided in Supplementary Figure S9. This result indicates that Gα genes (mainly mga1 and mga2) negatively influenced CIT production by regulating the CIT biosynthetic gene cluster.

CONCLUSION AND DISCUSSION
G-protein signaling pathways play important roles in fungal reproduction and SMs production, and the functions of different Gα subunits (Gα1-3) have been analyzed in some fungi using single gene modification Yoda et al., 2015;Zhang et al., 2016). The positively regulatory function of the Gα1 subunit on colony growth and asexual reproduction, which is conserved and extensive in most reported fungi, has been extensively researched (Li et al., 2010;Yang et al., 2012;Hu et al., 2013;Wasil et al., 2013;Garcia-Rico et al., 2017). However, until now, there has been no literature regarding double deletions combined with RNA-Seq of Gα subunit genes. In the current study, single-and double-gene(s) deletion mutants of the three Gα subunits were first systematically analyzed to determine the effects of different Gα subunits on M. ruber M7 according to the phenotypic characteristics combined with RNA-Seq analyses. The results show that all three Gα subunits (Mga1-3) in M. ruber M7 work together to regulate biological processes. Briefly, Mga1 comprehensively regulates the growth, development and secondary metabolism, while Mga2 and Mga3 act as supplementary regulators on growth and secondary metabolism. These findings not only illuminate the functions of different Gα subunits in M. ruber M7, but also provide a deeper understanding of the functional connections among different Gα subunits that involve regulating growth, development and secondary metabolism in other filamentous fungi.
Different Gα subunits (Gα1-3) regulate different biological processes in fungi (Li et al., 2010;Yang et al., 2012;Hu et al., 2013;Wasil et al., 2013;Garcia-Rico et al., 2017). For vegetative growth, Gα1 positively regulate the related processes in fungi such as Penicillium camembertii and Fusarium oxysporum (Guo et al., 2016b;Garcia-Rico et al., 2017), and Gα2 has no significant influence on fungal vegetative growth in Valsa mali and F. oxysporum (Guo et al., 2016a;Song et al., 2017), while Gα3 possesses different regulatory functions in different fungi. For example, PGA3 (Gα3) in P. camembertii and Gvm3 (Gα3) in V. mali positively regulate vegetative growth (Hu et al., 2013;Song et al., 2017), while FGA3 (Gα3) in F. oxysporum has no influence on vegetative growth (Guo et al., 2016b). In the current study, we find that Mga1(Gα1) has slightly effects on the vegetative growth of M. ruber M7, while Mga2 and Mga3 have no significant effects, which is similar to the results in P. camembertii and F. oxysporum (Guo et al., 2016b;Garcia-Rico et al., 2017), and the colony sizes of mga1+2 and mga1+3 are much smaller than those of M. ruber M7 and mga1, which suggests that the Mga2 and Mga3 subunits enhance this regulatory process when either is deleted along with Mga1. In addition, a group of MFS transporters involved in carbon source metabolism are more down-regulated in the mga1+2 and mga1+3 strains than those in M. ruber M7 according to RNA-Seq analyses (Supplementary Table S3), which implies that the transportation of carbon sources may be essential for Monascus growth and that Gα subunits may directly regulate MFS transporters to affect Monascus vegetative growth. Thus, further investigations of these transporters could contribute to determining the key elements involved in Monascus and other fungi vegetative growth.
Asexual reproduction, in many filamentous fungi, is mainly positively regulated by the sporogenesis central regulatory genes, including abaA, brlA and wetA (de Vries et al., 2017;Wu et al., 2018). However, in this study, the increased expression levels of brlA and wetA (no abaA in Monascus genome) in mga1-related mutants ( mga1, mga1+2, and mga1+3) do not enhance conidial reproduction. This implies that a new asexual reproduction-related regulatory pathway might exist in M. ruber M7. Further studies on reproduction related regulatory pathways which we are doing, might find a new asexual reproduction regulatory pathway in Monascus spp.
The Gα regulation of SMs biosynthesis has been verified by single gene deletions, indicating that the negative regulation of Gα1 is conserved in most fungi (Yu et al., 2008;Guo et al., 2016a), and Gα2's regulatory roles are diverse. For example, Gvm2 (Gα2) in V. mali negatively regulates SMs biosynthesis (Song et al., 2017), while GanA (Gα2) in Aspergillus nidulans has no influence on SMs biosynthesis (Yu, 2006). Additionally, Gα3 has no significant influence on SMs biosynthesis (Chang et al., 2004;Guo et al., 2016b). In our study, the single gene deletions have revealed that Mga1 (Gα1) and Mga2 (Gα2) can negatively regulate MPs and CIT production and that Mga3 (Gα3) has no significant effect. These results are similar to those of studies in V. mali and F. oxysporum (Guo et al., 2016b;Song et al., 2017). Moreover, double-gene deletions of Gα1-3 subunits can jointly regulate SMs. For instance, Mga2 and Mga3 combined with Mga1 can negatively regulate MPs production,  Figure S8).
The RNA-Seq results ( Supplementary Table S4) show that, besides MPs and CIT polyketide synthase (PKS) genes, many other PKS and non-ribosomal peptide synthetase genes are also regulated by Gα subunits. This is especially true of the mga1+3 strain in which nearly all the PKS and non-ribosomal peptide synthetase genes are differentially expressed. The analyses of related SMs in mga1+3 may help to improve our understanding of Monascus SMs.
Based on the above findings, a Gα regulatory system in M. ruber M7 is proposed in Figure 4. First, vegetative growth is mainly positively regulated by Mga1, and Mga2 and Mga3 can improve this regulatory process when either is deleted along with Mga1. All the Gα subunits positively regulate carbon and nitrogen metabolism (Supplementary Table S3) to affect vegetative growth. Second, Mga1 contributes the most to the regulation of sexual/asexual reproduction compared with Mga2 and Mga3, and the regulation of asexual reproduction may occur prior to the central regulatory pathway. The regulation of sexual reproduction is reflected in the regulation of MAT1-2 gene, which is down-regulated in mga1-deleted strains (Supplementary Figure S7). Third, Gα subunits in M. ruber M7 negatively regulate the SMs. In detail, Mga1 and Mga2 can negatively regulate MPs and CIT production individually or jointly, while Mga3 may combine with Mga1 to only negatively regulate MPs yields.

DATA AVAILABILITY
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

AUTHOR CONTRIBUTIONS
FC managed the project. ML, JL, YS, and YF conducted the transformants construction, secondary metabolites analysis, and transcriptome results analysis in this work. LL constructed the mga1 strain. ML conducted the phenotypic characterization, and interpreted the analysis results and wrote the manuscript. J-HY and YF contributed to the revision of the manuscript. All authors reviewed the manuscript.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb. 2019.01555/full#supplementary-material FIGURE S1 | Deletion strategy and confirmation of mga2 and mga3 mutants.   mga2 strain and Rmga2 mutants were observed on PDA, MA, CYA, and G25N plates cultured at 28 • C for 5 days. The enlarged areas are indicated by arrows. Size bar = 100 µm. (C) Colony morphologies of M7, mga3 strain and Rmga3 mutants observed on PDA, MA, CYA, and G25N plates and cultured at 28 • C for 12 days. (D) Cl and Co morphologies among M7, mga3 strain and Rmga3 mutants were observed on PDA, MA, CYA, and G25N plates cultured at 28 • C for 5 days. The enlarged areas are indicated by arrows. Size bar = 50 µm. FIGURE S4 | Gene expression levels analyzed by RNA-Seq and qRT-PCR. The x-axis represents the selected 11 genes; the y-axis on the left side represents the gene expression levels as assessed by RNA-Seq; the y-axis on the right side represents the relative gene expression level as assessed by qRT-PCR.