Effects on Gene Transcription Profile and Fatty Acid Composition by Genetic Modification of Mevalonate Diphosphate Decarboxylase MVD/Erg19 in Aspergillus Oryzae

Mevalonate diphosphate decarboxylase MVD/Erg19 is required for ergosterol biosynthesis, growth, sporulation, and stress tolerance in Aspergillus oryzae. In this study, RNA-seq was used to analyze the gene transcription profile in AoErg19 overexpression (OE) and RNAi strains. There were 256 and 74 differentially expressed genes (DEGs) in AoErg19 OE and RNAi strains, respectively, compared with the control strain (CK). The most common DEGs were transport- and metabolism-related genes. Only 22 DEGs were obtained that were regulated in both OE and RNAi strains. The transcriptomic comparison between CK and AoErg19 overexpression strain (CK vs. OE), and between CK and AoErg19 RNAi strain (CK vs. RNAi) revealed that the greatest difference existed in the number of genes belonging to the cytochrome P450 family; 12 were found in CK vs. OE, whereas 1 was found in CK vs. RNAi. The expression patterns of lipid biosynthesis and metabolism related genes were altered in OE and RNAi strains, either by gene induction or suppression. Moreover, the total fatty acid content in the RNAi strain was 12.1% greater than the control strain, but no difference in total acid content was found between the overexpression strain and the control strain. Therefore, this study highlights the gene expression regulation within mevalonate (MVA), ergosterol biosynthesis, and fatty acid biosynthesis pathways.


Transcriptome Overview
The numbers of total reads for all strains were greater than 38,391,610. Among those reads, the mapped reads ranged from 86.95% to 91.69%, and more than 87% were unique mapped reads. The %≥Q30 was greater than 91% and the GC content for all treatments was approximately 52%. The Pearson's correlation coefficient between two repeats of each strain was greater than 0.767, which indicates that the RNA-seq results of each strain were repeatable and reliable. A summary of the RNA-seq results is shown in Table 1.  Tables S1-S3. For these DEGs, only five DEGs were commonly shared among the three groups, and the other commonly shared DEGs between every two groups were 22, 31, and 60, respectively ( Figure 2A). The transcription levels of AoErg19 in RNA-seq results confirmed that AoErg19 was overexpressed and knocked down in OE and RNAi strains, respectively, compared with CK ( Figure S1), which demonstrate the reliability of the RNA-seq results. DEGs were analyzed by Cluster of Orthologous Groups of proteins (COG) functional classification. The COGs mapped to DEG gene function were divided into seven classes: energy production and conversion, transport and metabolism, cell wall/membrane/envelope biogenesis, general function prediction only, secondary metabolic-related genes, function unknown, and defense mechanisms ( Figure 2B). Our previous studies have revealed that the most DEGs associated with ergosterol biosynthesis inhibitors are transport-and metabolism-related genes [23]. Consistent with these results, in the present study, the most common DEGs were also transport-and metabolism-related genes, which accounted for 38.1% and 59.4% of the DEGs for CK vs. OE and CK vs. RNAi, respectively. These transport-and metabolism-related genes were divided into six classes ( Figure 2). Among these six classes, the most common for both CK vs. OE and CK vs. RNAi are carbohydrate transporters ( Figure 2C,D). only, secondary metabolic-related genes, function unknown, and defense mechanisms ( Figure 2B).

118
Our previous studies have revealed that the most DEGs associated with ergosterol biosynthesis 119 inhibitors are transport-and metabolism-related genes [23]. Consistent with these results, in the 120 present study, the most common DEGs were also transport-and metabolism-related genes, which 121 accounted for 38.1% and 59.4% of the DEGs for CK vs. OE and CK vs. RNAi, respectively. These 122 transport-and metabolism-related genes were divided into six classes ( Figure 2). Among these six 123 classes, the most common for both CK vs. OE and CK vs. RNAi are carbohydrate transporters ( Figure   124 2C,D).

134
DEGs were associated with cell membrane proteins or transporters, accounting for 41%; the second 135 were associated with energy production and conversion, and transcription. There were only five

Comparison of Gene Expression Profiles in OE and RNAi Strains
Results show that there are 22 common DEGs (Table 3). Except for EIT82403 and EIT77689, the most common DEGs showed the same trends in CK vs. OE and CK vs. RNAi. Most of the common DEGs were associated with cell membrane proteins or transporters, accounting for 41%; the second were associated with energy production and conversion, and transcription. There were only five commonly shared DEGs among the three groups: EIT76067, EIT82403, EIT76073, EIT77689, and EIT79995.1, which are shown in bold in Table 3. The details regarding expression differences and function of these genes are shown in Table 3.

Expression of Cytochrome P450 Family Genes in the Over-Expression Strain
By analyzing the different DEGs in OE and RNAi strains, it was discovered that major differences could be classified into secondary metabolic-related genes, which can be seen in Figure 2B. There were 22 and three DEGs associated with secondary metabolism in CK vs. OE and CK vs. RNAi, respectively. A detailed analysis of these genes revealed that 12 DEGs (55%) belonged to the cytochrome P450 family in CK vs. OE; however, only one DEG belonged to the cytochrome P450 family for CK vs. RNAi. Fourteen DEGs belonged to the cytochrome P450 family for OE vs. RNAi. The name and expression level of these genes are shown in Table 4.

Expression Pattern of Genes Involved Lipid Biosynthesis and Metabolism
A heatmap of the expression patterns of these lipid-related genes revealed that almost all of the genes displayed altered expression patterns, either via induction or suppression, in the OE and RNAi strains ( Figure 3A). Then, we analyzed the expression patterns of ergosterol biosynthesis genes and found that most of these genes also displayed altered expression patterns ( Figure 3B). The DEGs involved in lipid biosynthesis and metabolism were also analyzed. As shown in Table 3, there were nine DEGs found in CK vs. OE, including five ergosterol biosynthesis genes; only three DEGs were found in CK vs. RNAi. Among all the DEGs involved in lipid biosynthesis and metabolism, only two genes (EIT79966 and EIT77179) were down-regulated; the others were up-regulated. The detailed functions of these genes are shown in Table 5. Besides, qRT-PCR was performed to confirm the reliability and availability of the DEGs involved lipid and ergosterol biosynthesis and transportation. The results of qRT-PCR correlated with that of transcriptome analysis ( Figure 4).  Figure 3A). Then, we analyzed the expression patterns of ergosterol biosynthesis genes and 155 found that most of these genes also displayed altered expression patterns ( Figure 3B). The DEGs 156 involved in lipid biosynthesis and metabolism were also analyzed. As shown in Table 3 Table 5. Besides, qRT-PCR was performed to confirm the 161 reliability and availability of the DEGs involved lipid and ergosterol biosynthesis and transportation.

162
The results of qRT-PCR correlated with that of transcriptome analysis ( Figure 4).

Fatty Acid Composition in Over Expression and RNAi A. oryzae Strains
As the initial substrates for ergosterol and fatty acid biosynthesis pathways are acetyl-CoA. Our previous study showed that ergosterol contents were less in OE and RNAi strains compared to the control strain [22]. Thus, we analyzed whether the fatty acid compositions were also affected in OE and RNAi strains. Results showed that in OE strains, the compositions of fatty acid were changed, for example C14:0, C16:1, C18:0, C18:1n9c, and C24:0 increased a little, but the contents of total fatty acid, unsaturated fatty acid (UFA) and saturated fatty acid (FA) showed no significant change compared with CK strain (Table 6). However, in the RNAi strain, both the fatty acid compositions and contents showed differences compared with CK strain. In RNAi strains, the total fatty acid content was 12.1% greater than that of the control strain, which is approximately equal to the differences in UFA (12.0%) and FA (12.4%) content. Therefore, the total fatty acid content in the RNAi strain increased, while no difference in total acid content was found between the overexpression strain and the control strain.

Discussion
Cellular lipid metabolism involves a complex network of pathways, and is precisely regulated and coordinated to maintain lipid homeostasis [3,17]. Our previous study showed that both the overexpression and RNAi of AoErg19, one of the key genes in MVA pathway, were associated with lower ergosterol content in A. oryzae. In this study, we highlighted the transcription profile changes that occur in AoErg19 overexpression and RNAi strains. The fatty acid contents of these strains were also analyzed. This study lays a foundation for understanding the gene expression regulation along the MVA, ergosterol biosynthesis, and fatty acid biosynthesis pathways.

Effects of AoErg19 on Gene Transcription in A. oryzae
We identified 256 and 74 DEGs for CK vs. OE and CK vs. RNAi, respectively. The number of DEGs appears to be smaller compared with other transcriptome analyses, indicating that the effect on the gene transcription profiles by modification of AoErg19 is specific. Thus, these DEGs are likely to be genes that are regulated directly by AoErg19. The number of DEGs found in CK vs. OE was far greater than the number in CK vs. RNAi, which can be explained by the differences in phenotype, mRNA levels, and ergosterol content of the OE strain, which were more pronounced than those of RNAi strains [22]. Former transcriptome analysis revealed that the most common DEGs associated with the inhibition of ergosterol biosynthesis were transport-and metabolism-related genes [23,24]. Consistent with these results, the most common DEGs for CK vs. OE and CK vs. RNAi were also transportand metabolism-related genes, which may be caused by the changes in ergosterol content, as this is a component of cell membranes, and most transporters are located within the cell membrane. We also compared transport-and metabolism-related genes with those identified from ergosterol biosynthesis inhibitor treatment and found that there were no common genes. This indicates that different treatment methods or transgenic strains resulted in different changes in cell membrane components, resulting in differences in transporter gene expression. Our former studies revealed that the OE and RNAi strains had defects in growth and sporulation [22]. However, little DEGs involved in growth and sporulation were identified. This may because of that the RNA-seq analysis was performed in 48-h-old strains, the growth differences between transgenic strains and CK is not obvious and sporulation usually happens after 48-h cultivation.

Comparison of Gene Expression Profiles in OE and RNAi Strains
A comparison of DEGs for CK vs. OE and CK vs. RNAi revealed that there were 22 genes affected by both overexpression and RNAi of AoErg19, and only five DEGs were shared among CK vs. OE, CK vs. RNAi, and OE vs. RNAi. Most of the DEGs found were associated with cell membrane proteins or transporters. The second most common were associated with energy production and conversion, and transcription related genes. The changes in energy production and conversion-related genes can be explained by how altered ergosterol content impairs mitochondrial structure [25][26][27]. In investigating the differences of DEGs between CK vs. OE and CK vs. RNAi, a large number of cytochrome P450 family genes were found. For example, there were 12 DEGs associated with the cytochrome P450 family in CK vs. OE, but only 1 was found for CK vs. RNAi. This can be explained by how in the OE strain, the intermediates IPP, dimethylallyl-PP, geranyl diphosphate, or farnesyl-PP could have accumulated, which can be toxic to cells. Therefore, cytochrome P450 family genes may have been regulated to metabolize these intermediates.

Expression Pattern of Lipid Biosynthesis and Metabolism Genes, and Fatty Acid Content in OE and RNAi Strains
Although expression patterns of almost all genes related to lipid biosynthesis and metabolism displayed altered expression patterns ( Figure 3A), only nine DEGs for CK vs. OE (including five ergosterol biosynthesis genes) and three DEGs for CK vs. RNAi, were identified. Fatty acid composition analysis showed that although compositions of RNAi and OE fatty acids were different to those of the control, the contents of total fatty acid, UFA, and FA were only increased in RNAi strains. A possible reason is that reduction of AoErg19 mRNAi levels may have led to a decrease in the acetyl-CoA available for the biosynthesis of ergosterol. Therefore, more acetyl-CoA may have been used for fatty acid biosynthesis. Few DEGs involved in lipid and transportation were identified in CK vs. RNAi, possibly because of acetyl-CoA in fatty acid biosynthesis boosting the overall fatty acid biosynthesis pathway with minimal involvement of specific genes; altered expression levels may have been too insignificant to be detected. This is consistent with the result that the total fatty acid difference was approximately equal to that of UFA (12.0%) and FA in RNAi strains. A model for genetic modification of AoErg19 mRNAi levels affecting gene transcription profiles and fatty acid composition in A. Oryzae is shown in Figure 5.   Oryzae. In the AoErg19 over expression strain, more acetyl-CoA was consumed by the MVA pathway, and IPP, dimethylallyl-PP, geranyl diphosphate, or farnesyl-PP accumulated as a result. Cytochrome P450 family genes were activated to metabolize these intermediates.
In the RNAi strain, more acetyl-CoA was consumed by fatty biosynthesis. Both OE and RNAi strains showed differences in ergosterol contents, which can affect the cell membrane structure and mitochondrial structure, further affecting transport and cell membrane related gene expression and energy production and conversion related gene expression. Meanwhile, changes in fatty acid in cell resulted in the altered expression of lipid biosynthesis and transportation related genes. The bold arrowhead indicates that more substrate enters into the pathway.

Preparation of cDNA Libraries and RNA Sequencing
Approximately 0.5 g mycelia after 48 h cultivation were used for RNA sequencing. Total RNA isolation was performed using a fungal RNA kit (Omega Bio-tek, Norcross, GA, USA) according to the manufacturer's instructions, with the addition of RNase-free DNase I treatment (Omega). RNA concentration and integrity were analyzed using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA), and using a Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA, USA). Equal quantities of RNA of each pool from three individual cultures were used for cDNA library construction to ensure reliability and reproducibility. The total RNA was then enriched using oligo (dT) magnetic beads. After enrichment, mRNA was digested into short pieces in fragmentation buffer at 94 • C for 5 min. Then, mRNA fragments were used as templates to synthesize first-strand cDNA using random hexamer-primers. Second-strand cDNA was synthesized using DNA Polymerase I and RNase H. Following end-repair and adaptor ligation, the products were amplified by PCR and purified with a QIAquick PCR purification kit (Qiagen, Beijing, China) to create a cDNA library. Finally, the constructed libraries were sequenced by Biomarker Technologies Co. using an Illumina HiSeq 2500 platform (Illumina, San Diego, CA, USA) to generate 125 bp paired-end reads. Two biological replicates for RNA seq were used for each strain. The sequencing data were deposited in the NCBI/SRA database (Bioproject: PRJNA558619; BioSample: SAMN12496167).

Mapping Reads to the A. oryzae Reference Genome and Normalized Gene Expressions
Low-quality reads, adaptor sequences and reads with N's (uncertain bases) were removed to obtain high quality clean-read datasets, which were aligned to the A. oryzae genome (https: //www.ncbi.nlm.nih.gov/genome/526?genome_assembly_id=29881) using Tophat (v2.0.7) software. The Q20, Q30, GC-content and sequence duplication rate of the clean data were calculated. Transcript abundances were normalized by the reads per kilobase per million mapped reads (RPKM) metric to identify differentially expressed genes (DEGs). The false discovery rate (FDR) within 0.05 and log 2 (fold change) over 1 were set as the threshold for identification DEGs between samples [29][30][31].

Quantitative Real-time Reverse Transcription-PCR
Total RNA was isolated as described above. The cDNAs were synthesized from 1 µg of total RNA using the Prime Script™ RT Reagent Kit (Perfect Real Time; Shiga, Takara). All of the quantitative real-time reverse transcription-PCR (qRT-PCR) measurements were performed using a CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) with TB-Green Premix Ex Taq (Takara, Shiga, Japan), according to the manufacturer's instructions. The housekeeping gene GAPDH was used as a normalization control. The relative expression was calculated by using the formula 2 -∆∆Ct . All the experiments were performed for each biological replicate. The primer sequences for qRT-PCR are provided in Table S5.

Measurement of Fatty Acid Contents
Total lipids extraction was performed according to a previously described method [32,33]. Briefly, 48-h-old A. oryzae mycelia were collected and vacuum freeze-dried to a constant weight. Then, the mycelia were pulverized, weighed, and placed in a 50-mL Falcon tube for lipid extraction. Subsequently, 2.3 mL of dimethyl sulfoxide, p-hydroxymercuribenzoic acid (1 mmol/L final concentration), and 2,6-di-tert-butyl-p-cresol (7 mmol/L final concentration) were added. The suspension was maintained in a water bath at 100 • C for 1 h and then cooled at room temperature. Then, 13.3 mL of 1:1 chloroform-methanol was added and the mixture was gently shaken for 3 min then maintained at 4 • C overnight. To this, 6.7 mL of chloroform was added and 2.5 mL water was introduced and the tube was gently shaken for 3 min; after centrifugation at 6000 rpm for 10 min, the aqueous phase was eliminated and the organic phase washed 3× with 1.6 mL of 2 mol/L KCl·H2O. Finally, the lipid extract was then dried in a rotavapor and dissolved in 5 mL of chloroform. The lipid extracts were incubated in anhydrous methanol with 2% H 2 SO 4 at 70 • C for 2 h to obtain fatty acid methyl esters (FAMEs). The FAME components were extracted by N-hexane, methyl salicylate as internal standard. A mixed fatty acid methyl ester standard (Nu-Chek Prep Inc. Elysian, MN, USA) was used to draw standard curve. Samples were separated and analyzed using a GC-MSD (Agilent 7890A/5975C Gas-Mass Spectromete) equipped with an Agilent DB-WAX capillary polar column (30 m × 0.25 mm ID × 0.25 µm). MSD ChemStation software was used to extract the peak area and retention time. Fatty acid contents were calculated using the standard curve and are expressed as mg/g dry weight cells.

Data Analysis
All of the data in this study are presented as mean ± SE of three replicates. Significant differences were determined through a one-tailed Student's t-test or one-way analysis of variance (ANOVA) followed by Tukey's honestly significance difference (HSD) test for mean comparison.

Conclusions
To conclude, this study have investagited the gene transcription profile in AoErg19 OE and RNAi strains of A. oryzae. The comparison of the DEGs in CK vs OE and CK vs RNAi revealed that the most common DEGs were transport-and metabolism-related genes and the greatest difference existed in the number of genes belonging to the cytochrome P450 family. Moreover, the total fatty acid content in the RNAi strain was 12.1% greater than the control strain, but no difference in total acid content was found between the overexpression strain and the control strain.