Comparative Genomics of Mortierellaceae Provides Insights into Lipid Metabolism: Two Novel Types of Fatty Acid Synthase

Fungal species in the family Mortierellaceae are important for their remarkable capability to synthesize large amounts of polyunsaturated fatty acids, especially arachidonic acid (ARA). Although many genomes have been published, the quality of these data is not satisfactory, resulting in an incomplete understanding of the lipid pathway in Mortierellaceae. We provide herein two novel and high-quality genomes with 55.32% of syntenic gene pairs for Mortierella alpina CGMCC 20262 and M. schmuckeri CGMCC 20261, spanning 28 scaffolds of 40.22 Mb and 25 scaffolds of 49.24 Mb, respectively. The relative smaller genome for the former is due to fewer protein-coding gene models (11,761 vs. 13,051). The former yields 45.57% of ARA in total fatty acids, while the latter 6.95%. The accumulation of ARA is speculated to be associated with delta-5 desaturase (Delta5) and elongation of very long chain fatty acids protein 3 (ELOVL3). A further genomic comparison of 19 strains in 10 species in three genera in the Mortierellaceae reveals three types of fatty acid synthase (FAS), two of which are new to science. The most common type I exists in 16 strains of eight species of three genera, and was discovered previously and consists of a single unit with eight active sites. The newly revealed type II exists only in M. antarctica KOD 1030 where the unit is separated into two subunits α and β comprised of three and five active sites, respectively. Another newly revealed type III exists in M. alpina AD071 and Dissophora globulifera REB-010B, similar to type II but different in having one more acyl carrier protein domain in the α subunit. This study provides novel insights into the enzymes related to the lipid metabolism, especially the ARA-related Delta5, ELOVL3, and FAS, laying a foundation for genetic engineering of Mortierellaceae to modulate yield in polyunsaturated fatty acids.

Compared with other microorganisms such as microalgae and algae, oleaginous fungi are distinct in possessing a shorter life cycle, assimilating a variety of carbon sources, and Bellefonte, PA, USA), and nonanoic acid (C9:0) was selected as internal standard according to Zhao et al. [1].

Genome Sequencing and Assembly
Mortierella alpina CGMCC 20262 and M. schmuckeri CGMCC 20261 were incubated with PDA at 20 • C for five days. Total cell DNAs were extracted from mycelia using a kit (O-GPLF-400, GeneOnBio Corporation, Changchun, China) according to the manufacturer's protocol, and then detected by DNA/Protein Analyzer and 1% agarose gel electrophoresis. High-quality DNAs were sequenced at Beijing Novogene Bioinformatics Technology Co., Ltd. (Beijing, China) by using a PacBio Sequel and Illumina NovaSeq 6000 platform with 20 kb and 350 bp library, respectively. Low-quality reads (less than 500 bp) were removed by quality control from the raw data produced with the PacBio Sequel platform, and then the controlled high-quality reads were de novo assembled using SMRT Link v5.1.0 [34]. The assembled genomes were assessed by Quast v5.0.2 [35] and BUSCO v5.2.2 [36].

Phylogenomic and Phylogenetic Analyses
A total of 40 strains of early diverging fungi, including 38 strains downloaded from online databases and two strains sequenced herein, were used for phylogenomic analyses (Table 1). A total of 192 clusters of orthologous proteins were identified with HMMER v3.3.1 [49] and Trimal v1.4.4 [50], following the methods described by Spatafora et al. [51] and James et al. [52], and then their amino acid sequences were aligned with MAFFT v7 [53]. Phylogenomic analyses were carried out with a Maximum Likelihood (ML) algorithm using RaxML v8.1.12 [54]. Maximum Likelihood analyses adopted the PROTGAMMALGX substitution model with 100 bootstrap replications. Phylogenetic analyses on amino acids of delta-5 desaturase and fatty acid synthase were performed using IQ-TREE v1.0 [55] with an ML algorithm, WAG substitution model, and 1000 bootstrap replications. Note: The star mark "*" represents members in the Mortierellaceae. The two genomes newly generated in this study are in bold.

Comparative Genomic Analyses
The genome sequences of Mortierella alpina CGMCC 20262 and M. schmuckeri CGMCC 20261 were aligned using MCScanX for all protein-coding gene models, and then genomic collinearity was analyzed using a dual synteny plotter package [61]. For comparison of genes in lipid metabolism, especially those encoding fatty acid synthase and delta-5 desaturase, all 19 genomes of Mortierellaceae were reannotated using the KEGG database [41].

Synteny between Mortierella alpina and M. schmuckeri
A total of 13,726 gene models located in all scaffolds are collinear between Mortierella alpina CGMCC 20262 and M. schmuckeri CGMCC 20261, accounting for 55.32% of all the 24,812 gene models in both strains. Figure 3 shows the synteny of the ten largest scaffolds where 16,927 gene models are predicted and 11,777 (69.58%) gene models are syntenic.

Synteny between Mortierella alpina and M. schmuckeri
A total of 13,726 gene models located in all scaffolds are collinear between Mortierella alpina CGMCC 20262 and M. schmuckeri CGMCC 20261, accounting for 55.32% of all the 24,812 gene models in both strains. Figure 3 shows the synteny of the ten largest scaffolds where 16,927 gene models are predicted and 11,777 (69.58%) gene models are syntenic.

Lipid Metabolism in Mortierellaceae
To explore the lipid metabolism in Mortierellaceae, genomes of 19 strains in 10 species are de novo annotated or reannotated. We focus on the 13 genes associated with NADPH, precursors, desaturation, and elongation ( Figure 4 and Supplementary Figure S2). The gene Delta5 (encoding delta-5 desaturase) is not found in M. gamsii NVP60, and neither is Delta6 (encoding delta-6 desaturase) in M. alpina LL118. With these two exceptions, each gene possesses between one and six copies. The genes HK (encoding hexokinase, a key enzyme in EMP or the glycolytic pathway) and ME (encoding malate dehydrogenase) are remarkable as both have between three and six copies (Supplementary Figure S2 and Supplementary Table S1).
Three types of fatty acid synthase are identified in the family Mortierellaceae (

Lipid Metabolism in Mortierellaceae
To explore the lipid metabolism in Mortierellaceae, genomes of 19 strains in 10 species are de novo annotated or reannotated. We focus on the 13 genes associated with NADPH, precursors, desaturation, and elongation ( Figure 4 and Supplementary Figure S2). The gene Delta5 (encoding delta-5 desaturase) is not found in M. gamsii NVP60, and neither is Delta6 (encoding delta-6 desaturase) in M. alpina LL118. With these two exceptions, each gene possesses between one and six copies. The genes HK (encoding hexokinase, a key enzyme in EMP or the glycolytic pathway) and ME (encoding malate dehydrogenase) are remarkable as both have between three and six copies (Supplementary Figure S2        Four desaturases, i.e., Delta5, Delta6, Delta9, and Delta12, are found in the 19 strains of Mortierellaceae (Supplementary Figure S2 and Supplementary Table S1). Among these, the Delta5 catalyzes the synthesis of arachidonic acid (ARA, C20:4; Figure 4), which has diverged into three clades in Mortierellaceae (Figure 6 and Supplementary File S2).
The elongation of very long chain fatty acids protein (ELOVL) family consists of four members in Mortierellaceae, i.e., ELOVL2, ELOVL3, ELOVL4, and ELOVL6 (Supplementary Figure S1 and Supplementary File S3). ELOVL2 is shared by all strains of Mortierellaceae, while ELOVL3, ELOVL4, and ELOVL6 are species-specific. In detail, ELOVL3 is identified in M. alpina and M. amoeboidea, and ELOVL4 and ELOVL6 are annotated in M. verticillata and D. globulifera.
The draft genome of M. alpina was published in 2011, and consequently lipid synthesis pathway was predicted [12]. Since then, more genomes of members in the family Mortierellaceae have shed light on their lipid synthesis mechanism. For example, Mortierella sp. BCC40632 has an n-6 series fatty acid synthesis pathway due to a lack of delta-15 or omega-3 desaturase [25][26][27]30,31]. In this paper, the genomes of two more strains, M. alpina CGMCC 20262 and M. schmuckeri CGMCC 20261, are sequenced to provide more insights into lipid metabolism (Table 2).
A large number of genes hitherto have been found to affect lipid synthesis in Mortierella alpina, such as acetyl-CoA carboxylase, AMP deaminase, malate dehydrogenase, delta-5 desaturase, delta-6 desaturase, and elongase 2 [12,22,23,[66][67][68][69]. In this study, we compare 13 genes of 19 strains in Mortierellaceae (Supplementary Figure S2), finding differences in genes encoding hexokinase, delta-5 desaturase, and elongation of very long chain fatty acids protein. Compared with M. schmuckeri CGMCC 20261, the M. alpina CGMCC 20262 possesses one more copy of the HK gene, more copies of Delta5 and ELOVL genes but fewer copies of the Delta9 gene (Supplementary Figure S2 and Supplementary  Table S1). Delta-5 desaturase and elongation of very long chain fatty acids protein play an essential role in the synthesis of polyunsaturated fatty acids (PUFAs, Figure 4), which might be one of the reasons for M. alpina CGMCC 20262 yielding more PUFAs than M. schmuckeri CGMCC 20261.
PUFA biosynthesis requires an activity of elongation of very long chain fatty acids proteins, as well as fatty acid desaturases such as delta-5, delta-6, delta-12, and omega-3 desaturase [12,16]. For example, the M. alpina 1S-4 mutant in delta-5 desaturase resulted in a huge accumulation of dihomo-γ-linolenic acid (DGLA), up to 43.3% of total fatty acids [70]. In the present study, the delta-5 desaturase of M. alpina is significantly different from
The draft genome of M. alpina was published in 2011, and consequently lipid synthesis pathway was predicted [12]. Since then, more genomes of members in the family Mortierellaceae have shed light on their lipid synthesis mechanism. For example, Mortierella sp. BCC40632 has an n-6 series fatty acid synthesis pathway due to a lack of delta-15 or omega-3 desaturase [25][26][27]30,31]. In this paper, the genomes of two more strains, M. alpina CGMCC 20262 and M. schmuckeri CGMCC 20261, are sequenced to provide more insights into lipid metabolism (Table 2).
A large number of genes hitherto have been found to affect lipid synthesis in Mortierella alpina, such as acetyl-CoA carboxylase, AMP deaminase, malate dehydrogenase, delta-5 desaturase, delta-6 desaturase, and elongase 2 [12,22,23,[66][67][68][69]. In this study, we compare 13 genes of 19 strains in Mortierellaceae (Supplementary Figure S2), finding differences in genes encoding hexokinase, delta-5 desaturase, and elongation of very long chain fatty acids protein. Compared with M. schmuckeri CGMCC 20261, the M. alpina CGMCC 20262 possesses one more copy of the HK gene, more copies of Delta5 and ELOVL genes but fewer copies of the Delta9 gene (Supplementary Figure S2 and Supplementary Table S1). Delta-5 desaturase and elongation of very long chain fatty acids protein play an essential role in the synthesis of polyunsaturated fatty acids (PUFAs, Figure 4), which might be one of the reasons for M. alpina CGMCC 20262 yielding more PUFAs than M. schmuckeri CGMCC 20261.
PUFA biosynthesis requires an activity of elongation of very long chain fatty acids proteins, as well as fatty acid desaturases such as delta-5, delta-6, delta-12, and omega-3 desaturase [12,16]. For example, the M. alpina 1S-4 mutant in delta-5 desaturase resulted in a huge accumulation of dihomo-γ-linolenic acid (DGLA), up to 43.3% of total fatty acids [70]. In the present study, the delta-5 desaturase of M. alpina is significantly different from other species in the family Mortierellaceae based on the phylogenetic analyses ( Figure 6 and Supplementary File S2), which is probably one of the reasons for synthesizing high content of ARA [12,71,72]. Elongation of very long chain fatty acids proteins (ELOVLs) are involved in the long-chain polyunsaturated fatty acid synthetic pathway, and four gene types, ELOVL2, ELOVL3, ELOVL4, and ELOVL6, are identified herein in 19 strains of Mortierellaceae (Supplementary Figure S1 and Supplementary File S3).
The fatty acid synthase plays an important role in lipid synthesis, catalyzing the synthesis of saturated fatty acids from acetyl-CoA and malonyl-CoA. A variety of fatty acid synthases were found in fungi, such as Aspergillus oryzae, Mortierella alpina, Saccharomyces cerevisiae, and Yarrowia lipolytica [6,12,73,74]. In this study, three types of fatty acid synthases are found in Mortierellaceae (Supplementary Figure S2 and Supplementary File S1). Type I is the most common and similar to that reported in Mortierella alpina ATCC 32222, a single subunit consisting of eight catalytic domains/active sites [17]. Types II and III are new to science. Type II presents in M. antarctica KOD 1030 only, and type III in Dissophora globulifera REB-010B and Mortierella alpina AD071 (Supplementary Figure S2 and Supplementary File S1).
Overall, the rapid increase in genomic data is providing more materials for tackling fatty acid biosynthesis, discovering more potential genes, and finalizing the understanding of the genetic foundation of lipid metabolism.

Conclusions
In this paper, two new genomes, Mortierella alpina CGMCC 20262 and M. schmuckeri CGMCC 20261, were sequenced using the PacBio Sequel and Illumina NovaSeq 6000 platform. To explore the lipid metabolism in Mortierellaceae, a total of 19 genomes were reannotated. The results suggest that delta-5 desaturase and elongation of very long chain fatty acids protein 3 probably promoted the accumulation of polyunsaturated fatty acids, especially arachidonic acid. Besides, three types of fatty acid synthase, including two novel ones, were identified. Consequently, with the increase in public genomic data, a comprehensive analysis on lipid metabolism will discover more genes or protein-encoding models, providing more information for subsequent genetic engineering of oleaginous fungi.   (Table 2).