Homologous Overexpression of Acyl-CoA Thioesterase 8 Enhanced Free Fatty Acid Accumulation in Oleaginous Fungus Mucor circinelloides WJ11

: Thioesterases play an essential role in the metabolism of fatty acids since they are considered one of the key enzymes to change the total amount and composition of fatty acid in an organism’s cells. Acyl-coenzyme A thioesterase 8 (ACOT8) exhibits substrate speciﬁcity mainly for short-to long-chain acyl-CoA. To identify and characterize the ACOT8 enzyme’s superfamily in Mucor circinelloides , three genes were characterized and homologously expressed in M. circinelloides WJ11 which has been used as a model organism to investigate the mechanism of lipid accumulation. Multiple sequence alignment showed that McACOT8s had signiﬁcant conserved motifs in the ACOT8 family. Experimental data indicated that the biomass of the three recombinant strains (McACOT8a, McACOT8b and McACOT8c) was slightly lower than that of the control strain Mc2075, but the total fatty acid (TFA) content was signiﬁcantly increased by 30.3, 21.5 and 23.9%, respectively, and the free fatty acid production increased from 12.9% (control strain) to 19.2, 25.2 and 26.2% (recombinant types), respectively, with a maximum increase of 103.1% in McACOT8c. The three strains did not show any signiﬁcant improvement in fatty acid composition compared to the control strain. However, in terms of fatty acid composition of free fatty acid, McACOT8b and McACOT8c contained γ -linolenic acid (C18:3), which was not detected in the control strain or in McACOT8a, indicating that ACOT8b and ACOT8c had substrate speciﬁcity for C18:3. These results displayed that ACOT8 can increase TFA accumulation and that it may be an important target of genetic manipulation for microbial oil production.


Introduction
Thioesterases are a large superfamily of enzymes catalyzing the cleavage of thioester bonds and releasing free fatty acids, both medium-and long-chain fatty acids. These thioester bonds are present in a variety of activated fatty acid acyl-coenzyme A (acyl-CoA) substrates, acyl-acyl carrier proteins (acyl-ACPs), acyl-glutathione and other cellular molecules [1]. At present, the importance of the enzymatic activity of acyl-CoA/ACP thioesterases can be identified on the basis of the acyl-CoA/ACP esters and free fatty acids involved, in addition to the genetic manipulation of thioesterases to increase lipid content and alter fatty acid composition, which is a common technique for enhancing microbial lipid production [2].
Acyl-ACP thioesterases are an important enzyme in fatty acid synthesis, and the fatty acids synthesized in the plastid are hydrolyzed into free acyl-CoA/ACP under its lipid metabolism plays a role in stopping fatty acid chain elongation has not been reported. Therefore, we identified and studied the role of ACOT8 in the intracellular lipid metabolism of M. circinelloides WJ11, and three genes, ACOT8a, ACOT8b and ACOT8c, were homologously overexpressed in M. circinelloides to investigate the role of the genes in lipid accumulation in this microbial fungus.

Microorganisms, Cultivation and Transformation Conditions
M. circinelloides WJ11 (CCTCC no. M2014424) was used as the source of three genes (ACOT8a, ACOT8b and ACOT8c). M. circinelloides M65 (a urea-assisted strain derived from WJ11), which has similar oil-producing properties to the wild-type strain, was used as a background strain for all transformation experiments to overexpress these three genes. The cultures were cultured at 28 • C in complete yeast extract-peptone glycerol (YPG), minimal microbial microdroplet culture (MMC) and yeast nitrogen base media (YNB), which were supplemented with uracil at 600 mg/L when needed [18]. E. coli Top10 was grown in lysogeny broth (LB) at 37 • C with shaking at 220 rpm and used to maintain and propagate the recombinant plasmids in all cloning experiments [19].

Construction of ACOT8 Overexpressing Strains of M. circinelloides
Total RNA of M. circinelloides WJ11 was isolated using Trizol reagent (Takara, Dalian, China) according to standard protocols and reverse transcribed into cDNA. The cDNA was used as a template to obtain gene fragments of McACOT8a, McACOT8b and McACOT8c via PCR using the corresponding primers, as shown in Table S1. These fragments were cloned into the expression vector pMAT2075, which was constructed and preserved by our laboratory using the restriction endonucleases XhoI and NheI to produce a recombinant plasmid pMAT-ACOT8a, pMAT-ACOT8b and pMAT-ACOT8c. The plasmids expressing McACOT8a, McACOT8b and McACOT8c were transfected into M. circinelloides using the electroporation-mediated procedure described by Torres-Martinez et al. [20]; the CarRP gene associated with pigment synthesis was replaced and the transformants were color screened until all colonies turned white, and a mutant strain carrying the empty vector pMAT2075 was used as a negative control.

Fermentation of M. circinelloides Transformants
Approximately 10 6 spores of ACOT8 overexpression and the control strains were initially inoculated into 100 mL of Kendrick and Ratledge (K&R) medium [21] in 500 mL baffled flasks and cultured in a rotating shaker at 130 rpm, 28 • C for 24 h, and then 200 mL medium was placed in a 1.0 L baffled triangular flask and placed in a constant temperature shaker with 150 rpm and 28 • C to select the appropriate transformants. Finally, fermentation was carried out in a 3.0 L fermenter with 1.5 L medium maintained at 28 • C with 700 rpm agitation and 1.0 vvm aeration. Automatic addition of 2 M sodium hydroxide was used to maintain the pH at 6.0. Fermentation broth (1.0 L) consisted of 80.0 g glucose, 1

Analysis of Cell Dry Weight (CDW) and Culture Supernatant
Mycelia of overexpressing strains were cultured in a fermenter for 4 days, filtered through a Buchner funnel under reduced pressure and then washed with distilled water. The collected mycelium was placed in a centrifuge tube after weighing for freeze-drying, and the CDW of mycelium in the centrifuge tube was determined using the differential weight method. The residual glucose concentration in the medium was determined using a glucose oxidase kit (Shanghai Rongsheng Biotech Co., Ltd., Shanghai, China) and the residual ammonium concentration was determined using the indophenol test [22]. The ratio of CDW to time yielded the growth rate. The ratio of the mass of generated cells to the mass of consumed substrate was substrate cell yield (Yx/s). The specific growth rate (µ) was calculated using the following formula; N 2 and N 1 are the CDW at t 2 and t 1 , respectively. µ = (lnN 2 − lnN 1 )/(t 2 − t 1 ) (/h) (1)

Total Fatty Acid (TFA) Content and Lipid Composition Analysis
After freeze-drying treatment, cells of mycelium were cleaved using hydrochloric acid (4 mol/L) with the addition of pentadecanoic acid (C15:0) as the internal standard; chloroform/methanol (2:1, v/v) was added to extract TFA [23], and then fatty acid methyl ester (FAME) was prepared by adding 10% (w/w) methanol hydrochloride at 60 • C for 3 h.
Lipid composition of overexpressed and control strains was determined using thin layer chromatography (TLC): chloroform/methanol/water (2:1:1, v/v/v) was added for extraction. Lipids in the organic phase were dissolved with ethyl ether, then dropwise added to the silica gel plate and placed in a chromatographic chamber with the developing agent (n-hexane/ethyl ether/acetic acid, 50:50:1). The silica gel plate was placed in an iodine chamber for color development. The gray scanning software ImageJ was used to process the developing strip and obtain the gray peak area. The stained bands were scraped off and methylated with 10% (w/w) methanol hydrochloride at 60 • C for 3 h. FAME were extracted with n-hexane, determined using gas chromatography with an Agilent 123-3232 DB-Waxetr column (30 m × 320 µm × 0.25 µm) column, and then the procedure was as follows according to the previous study [24]: 120 • C for 3 min, ramp to 200 • C for 5 • C/min, ramp to 220 • C for 4 • C/min, hold for 2 min. The peaks were identified using standard sample of FAME (FAME mixed label from Sigma-Aldrich, St. Louis, MO, USA).

Gene Expression and RT-qPCR Analysis
Genetically modified strains of M. circinelloides were cultured in fermenters and mycelium was collected at 6, 12, 24 and 96 h. After grinding the mycelium under liquid nitrogen, total RNA of M. circinelloides was extracted using Trizol and reversed to cDNA using a qPCR SYBR Green Master Mix kit (Accurate Bio-Medical, Changsha, China). Relative expression of mRNA was analyzed with real-time quantitative polymerase chain reaction (RT-qPCR) using the 2 −∆∆C method [25], with mRNA expression of each overexpression strain at 6 h being one. Three independent biological replicates were analyzed.

Statistical Analysis
All data are presented as mean ± standard deviation (SD). Statistical analyses were performed using SPSS version 20.0 (SPSS Inc., Chicago, IL, USA). Differences were considered significant when p < 0.05 and very significant when p < 0.01.  (Figure 1), indicating that the ACOT8s identified in this species are homologous to genes of the already-identified family, suggesting that ACOT8a-c belongs to the thioesterase 4 family and is able to hydrolyze the acyl-CoA bond. tional center for biotechnology information (NCBI) showed that McACOT8a, McACOT8b and McACOT8c possessed the same superfamily domain. Multiple sequence alignment of four other identified ACOT8 genes and three ACOT8 from M. circinelloides WJ11 was performed (Figure 1), indicating that the ACOT8s identified in this species are homologous to genes of the already-identified family, suggesting that ACOT8a-c belongs to the thioesterase 4 family and is able to hydrolyze the acyl-CoA bond.

Construction of Plasmids and Screening of Transformants
Transformation of M. circinelloides WJ11 was carried out using homologous recombination to verify whether the target genes were associated with fatty acid accumulation. Transformants with fragmented empty vectors were used as a negative control to obtain control strain Mc2075, and transformants with fragments of the uracil synthesis gene (pyrF), promoter (Pzrt1) and ACOT8a-c were used as positive transformants to obtain different overexpression strains McACOT8a, McACOT8b and McACOT8c (Figure 2A). Genomic DNA from the transformants was extracted and used as a template for PCR validation using the primer CarRP-up-F/CarRP-down-R (sequence in Table S1). The band sizes of the transformants McACOT8a, McACOT8b and McACOT8c after amplification were 6372 bp, 6213 bp and 6447 bp, and Mc2075 was approximately 5361 bp ( Figure 2B). As shown in Figure 2C,D, by measuring the growth of three groups of transformants (two per group) with biomass and fatty acid content, it was evident that homologous overexpression of ACOT8 in M. circinelloides WJ11 had no significant effect on the biomass in the fungus, but had a significant promotional effect on the TFA content. The TFA content of the engineered strains exceeded the control strain Mc2075 (20.8%), and finally, three overexpression strains McACOT8a-1 (25.4%), McACOT8b-2 (26.9%) and McACOT8c-2 (29.3%) were selected as late fermentation strains based on the results of the TFA content.

Construction of Plasmids and Screening of Transformants
Transformation of M. circinelloides WJ11 was carried out using homologous recombination to verify whether the target genes were associated with fatty acid accumulation. Transformants with fragmented empty vectors were used as a negative control to obtain control strain Mc2075, and transformants with fragments of the uracil synthesis gene (pyrF), promoter (Pzrt1) and ACOT8a-c were used as positive transformants to obtain different overexpression strains McACOT8a, McACOT8b and McACOT8c ( Figure 2A). Genomic DNA from the transformants was extracted and used as a template for PCR validation using the primer CarRP-up-F/CarRP-down-R (sequence in Table S1). The band sizes of the transformants McACOT8a, McACOT8b and McACOT8c after amplification were 6372 bp, 6213 bp and 6447 bp, and Mc2075 was approximately 5361 bp ( Figure 2B). As shown in Figure 2C,D, by measuring the growth of three groups of transformants (two per group) with biomass and fatty acid content, it was evident that homologous overexpression of ACOT8 in M. circinelloides WJ11 had no significant effect on the biomass in the fungus, but had a significant promotional effect on the TFA content. The TFA content of the engineered strains exceeded the control strain Mc2075 (20.8%), and finally, three overexpression strains McACOT8a-1 (25.4%), McACOT8b-2 (26.9%) and McACOT8c-2 (29.3%) were selected as late fermentation strains based on the results of the TFA content.

Effect of ACOT8's Overexpression on Cell Growth in the Fermenter
The effects of McACOT8a-1, McACOT8b-2 and McACOT8c-2 overexpression on growth were analyzed by measuring CDW, residual glucose and ammonium in the culture of each expression strain during the whole fermentation process ( Figure 3). Generally, all overexpression strains showed a similar and typical growth profile to the control. The biomass of McACOT8a, McACOT8b and McACOT8c (18.0, 17.6 and 17.5 g/L) was significantly lower than that of the control (18.7 g/L) after 4 days of growth. Cell growth and lipid accumulation require the supply of carbon and nitrogen sources. In the first 12 h, cells grow, at which point nitrogen sources are completely consumed. In 12 h to 24 h, cells are in a state of adjustment, so there is a short plateau. By calculating the growth rate, specific growth rate and Yx/s data of fungi cell biomass in Figure S1, the maximum rate of the four strains was in 12 h, and the growth rate decreased significantly thereafter. After 40 h growth, the consumption rate of glucose as a carbon source slowed down and the growth rate of oleaginous fungus cells decreased. Under this culture condition, the TFA accumulation of the strain was saturated, so the TFA accumulation rate of the strain slowed down after 40 h. As shown in Figure S1D, the accumulation of TFA yield in the cells was significantly increased, and TFA yield in McACOT8a-1 cells reached 5.9 g/L at 96 h, while that in the control strain was only 4.7 g/L. The glucose consumption trend of these strains was almost similar. Glucose was consumed rapidly before 48 h, and then the cells grew slowly after 48 h and the glucose consumption rate slowed down, generating a

Effect of ACOT8's Overexpression on Cell Growth in the Fermenter
The effects of McACOT8a-1, McACOT8b-2 and McACOT8c-2 overexpression on growth were analyzed by measuring CDW, residual glucose and ammonium in the culture of each expression strain during the whole fermentation process (Figure 3). Generally, all overexpression strains showed a similar and typical growth profile to the control. The biomass of McACOT8a, McACOT8b and McACOT8c (18.0, 17.6 and 17.5 g/L) was significantly lower than that of the control (18.7 g/L) after 4 days of growth. Cell growth and lipid accumulation require the supply of carbon and nitrogen sources. In the first 12 h, cells grow, at which point nitrogen sources are completely consumed. In 12 h to 24 h, cells are in a state of adjustment, so there is a short plateau. By calculating the growth rate, specific growth rate and Yx/s data of fungi cell biomass in Figure S1, the maximum rate of the four strains was in 12 h, and the growth rate decreased significantly thereafter. After 40 h growth, the consumption rate of glucose as a carbon source slowed down and the growth rate of oleaginous fungus cells decreased. Under this culture condition, the TFA accumulation of the strain was saturated, so the TFA accumulation rate of the strain slowed down after 40 h. As shown in Figure S1D, the accumulation of TFA yield in the cells was significantly increased, and TFA yield in McACOT8a-1 cells reached 5.9 g/L at 96 h, while that in the control strain was only 4.7 g/L. The glucose consumption trend of these strains was almost similar. Glucose was consumed rapidly before 48 h, and then the cells grew slowly after 48 h and the glucose consumption rate slowed down, generating a slower TFA accumulation rate in the strains. Finally, glucose remained above 30 g/L at the end of fermentation ( Figure 3C). The medium provided sufficient carbon sources for the growth of M. circinelloides, but it was limited by the space of the fermentation tank and the oxygen concentration. M. circinelloides did not fully utilize these carbon sources. The organic nitrogen content in the medium was complicated and difficult to be extracted and detected. The consumption of the nitrogen source can be explored by measuring the change in the content of ammonium in the medium. There was no difference in ammonium consumption between the four strains. At 6 h, the consumption of nitrogen source was fast ( Figure 3D) and completely exhausted at 12 h. slower TFA accumulation rate in the strains. Finally, glucose remained above 30 g/L at the end of fermentation ( Figure 3C). The medium provided sufficient carbon sources for the growth of M. circinelloides, but it was limited by the space of the fermentation tank and the oxygen concentration. M. circinelloides did not fully utilize these carbon sources. The organic nitrogen content in the medium was complicated and difficult to be extracted and detected. The consumption of the nitrogen source can be explored by measuring the change in the content of ammonium in the medium. There was no difference in ammonium consumption between the four strains. At 6 h, the consumption of nitrogen source was fast ( Figure 3D) and completely exhausted at 12 h.

Effect of ACOT8's Overexpression on TFA Accumulation and Composition in M. circinelloides WJ11
All of the strains consumed nitrogen rapidly in the early stages of growth and division, and TFA began to accumulate in the cells when the nitrogen source was depleted. McACOT8c-2, respectively). Throughout the fermentation process, the TFA mass of the McACOT8a-1 overexpression strain (32.7%) was higher than that of the other two engineered strains and increased by 30.3% compared with the control (25.1%). This result confirms that the enzymes of ACOT8 have a significant effect on TFA accumulation in M. circinelloides WJ11. In terms of TFA composition, as shown in Table 1   Duncan's multiple range test was conducted and data with different letters indicate statistically significant differences among groups at p < 0.05.
The distance traveled on the thin silica gel plate in TLC varied with the mobility (Rf) in the unfolding agent, with triglycerides (TAG), free fatty acids, 1,3-diglycerides (1,3-DAG), 1,2-diglycerides (1,2-DAG) and monoglycerides (MAG) increasing in polarity and decreasing in distance from the origin, in that order ( Figure 4A). After coloration development on TLC silica gel plates, it was found that the lipid product of M. circinelloides WJ11 had the highest proportion of TAG in the total lipid content, with its band appearing at the top of the thin silica gel plate; followed by free fatty acids, 1,3-DAG and 1,2-DAG near each other; and the lowermost band near the spot sample was glycerol monoester, which was less abundant. The area of the gray scale peaks obtained after processing the developed bands with the gray scale scanning software ImageJ was plotted to scale ( Figure 4B   Duncan's multiple range test was conducted and data with different letters indicate statistically significant differences among groups at p < 0.05. The distance traveled on the thin silica gel plate in TLC varied with the mobility (Rf) in the unfolding agent, with triglycerides (TAG), free fatty acids, 1,3-diglycerides (1,3-DAG), 1,2-diglycerides (1,2-DAG) and monoglycerides (MAG) increasing in polarity and decreasing in distance from the origin, in that order ( Figure 4A). After coloration development on TLC silica gel plates, it was found that the lipid product of M. circinelloides WJ11 had the highest proportion of TAG in the total lipid content, with its band appearing at the top of the thin silica gel plate; followed by free fatty acids, 1,3-DAG and 1,2-DAG near each other; and the lowermost band near the spot sample was glycerol monoester, which was less abundant. The area of the gray scale peaks obtained after processing the developed bands with the gray scale scanning software ImageJ was plotted to scale    Duncan's multiple range test was conducted and data with different letters indicate statistically significant differences among groups at p < 0.05.

Expression Levels of Different ACOT8 Genes in the Overexpressing Strains of M. circinelloides WJ11
The mRNA levels of three recombinant strains were analyzed with RT-qPCR at 6, 12, 24 and 96 h of growth in fermenters ( Figure 5). Taking

Expression Levels of Different ACOT8 Genes in the Overexpressing Strains of M. circinelloides WJ11
The mRNA levels of three recombinant strains were analyzed with RT-qPCR at 6, 12, 24 and 96 h of growth in fermenters ( Figure 5). Taking

Discussion
Thioesterases play a key role in lipid accumulation and metabolism, they are widely found in prokaryotes and eukaryotes and are an important processor of various products. The largest proportion of thioesterase is still acyl-CoA/ACP thioesterase, and a variety of thioesterases perform multiple biologically important functions by enabling the cleavage of thioester bonds in a wide range of substrates. Therefore, lipoyl thioesterases genes are often genetically modified to enhance fatty acid production in oil-producing microorganisms [26]. Acyl-CoA thioesterases are critical and essential in the lipid synthesis pathway of microorganisms. As shown in Table 3, it was found that homologous overexpression of acyl-CoA thioesterase (tesA) and two genes encoding acyl-CoA synthetase (fadD5 and fadD15) in Corynebacterium glutamicum create a cyclic route between acyl-CoAs and fatty acids; and the resulting fadDs-disrupted and tesA-amplified strains increased their production by 72% and produced fatty acids consisting primarily of oleic acid, palmitic acid

Discussion
Thioesterases play a key role in lipid accumulation and metabolism, they are widely found in prokaryotes and eukaryotes and are an important processor of various products. The largest proportion of thioesterase is still acyl-CoA/ACP thioesterase, and a variety of thioesterases perform multiple biologically important functions by enabling the cleavage of thioester bonds in a wide range of substrates. Therefore, lipoyl thioesterases genes are often genetically modified to enhance fatty acid production in oil-producing microorganisms [26]. Acyl-CoA thioesterases are critical and essential in the lipid synthesis pathway of microorganisms. As shown in Table 3, it was found that homologous overexpression of acyl-CoA thioesterase (tesA) and two genes encoding acyl-CoA synthetase (fadD5 and fadD15) in Corynebacterium glutamicum create a cyclic route between acyl-CoAs and fatty acids; and the resulting fadDs-disrupted and tesA-amplified strains increased their production by 72% and produced fatty acids consisting primarily of oleic acid, palmitic acid and stearic acid [27]. Overexpression of thioesterase ACOT8I in the oleaginous fungus Mortierella alpina enhanced the transformation from acyl-CoA to free fatty acids, increased the production of free fatty acids and adjusted the fatty acid type of lipid species [2]. Further, in the engineered E. coli strain containing an endogenous thioesterase, at least 50% of the fatty acids produced can be present in free-acid form [28]. The acyl-CoA thioesterase of M. musculus was expressed in S. cerevisiae, and the unsaturated fatty acid content in the final engineered strain was higher [15]. Clearly, overexpression of acyl-CoA thioesterases can increase fatty acid production in oil-producing microorganisms.
Thioesterase not only plays a role in the accumulation of microbial lipid content, but also has specificity to fatty acids of different substrates. According to previous studies, ACOT8 in its active state is substrate-specific for all acyl-CoA substrates, from two to twenty carbon atoms, saturated and unsaturated fatty acids, and the experiment proved it. In the specific hydrolysis substrate of acyl-CoA, there is a significant difference with acyl-ACP. The eukaryotic homolog of ACH2 identified in A. thaliana is peroxisome acyl-CoA thioesterase, which is up-regulated during the period of increased fatty acid oxidation, and studies have shown that ACH2-6His hydrolyzes medium-to long-chain fatty acyl-CoA, but has the highest activity for long-chain unsaturated acyl-CoA [14]. Four ACOTs were identified in A. thaliana, of which two are likely peroxisomes (AtACH1 and AtACH2, with C-terminal PTS1 signals of eAKL and eSKL, respectively) and two others, AtACH4 and AtACH5, were located in the endoplasmic reticulum [29]. Of these, ACH2 was characterized in detail and shown to hydrolyze long-chain saturated and unsaturated acyl-CoA [14]. That ACOT8 is located in the peroxisome and has broad activity against long-chain acyl-CoAs, with the greatest activity levels for CoA esters of primary bile acids choloyl-CoA and chenodeoxycholoyl-CoA, suggests that ACOT8 is a regulator of peroxisomal lipid metabolism [9]. These data suggest that McACOT8 may be involved in lipid biosynthesis and play an important role in free fatty acid accumulation and fatty acid alteration in M. circinelloides. Table 3. Different sources of Acyl-ACP/CoA thioesterases were overexpressed in different microbial species on lipid or TFA accumulation and fatty acid composition.
to the application of its complex genetic engineering tools and the high production of polyunsaturated fatty acids from its genome sequence [12]. Although the role of key enzymes in the lipid synthesis pathway has been studied by many researchers at home and abroad, the study of endogenous thioesterases has not yet been reported. The present experiment was designed to investigate the functions of three acyl-CoA thioesterases in M. circinelloides WJ11, deepening the exploration of the lipid synthesis pathway in M. circinelloides WJ11, enriching the study of key enzymes in lipid synthesis and the gene pool of the thioesterase family, providing a basis for future research on the specific production of certain fatty acid products, offering further possibilities for the construction of M. circinelloides cell factories and thus making the conversion of M. circinelloides to biodiesel more efficient. This will provide a further possibility for the construction of a M. circinelloides cell factory and, thus, make the conversion of M. circinelloides to biodiesel more possible and practical.
Author Contributions: F.X. and M.G. contributed equally to this work and are considered the first authors. F.X. and M.G. designed the experiment, performed the experimental work and wrote the original draft. R.W. and Y.C. were involved in the fermentation testing. Z.X., W.D. and H.Y. performed the data curation. Y.Z. carried out the result interpretation. Y.S., C.R. and H.Z. conceived the study, contributed to designing the research protocol and finalized the manuscript. All authors have read and agreed to the published version of the manuscript.