The E. coli strain DH5α was used as the host in this study for the cloning and plasmid construction. E. coli strains were routinely cultured at 37°C in Luria-Bertani (LB) media (1% tryptone, 0.5% yeast extract, and 1% sodium chloride contained) or on LB agar plates supplemented with 100 μg/ml of ampicillin. Y. lipolytica Po1g KU70Δ was used as the base strain in this study, which has been generated from the parental strain Po1g (a commonly used host strain for protein expression). This strain was used as it is known that the rate of precise homologous recombination (HR) increased substantially for deletion of the KU70 gene in Po1g (Yu et al., 2016). Routine cultivation of Y. lipolytica strains was carried out at 30°C in YPD medium (1% yeast extract, 2% peptone and 2% dextrose contained) while the yeast synthetic complete medium (YNB) (0.67% yeast nitrogen base without amino acids, 2% glucose, 1.5% bacto agar) lacking the appropriate nutrients was used for the screening of transformants. The fermentation experiment used YPO medium containing WCO (1% yeast extract, 2% peptone, 1.18% WCO and 0.2% tween-80 contained), and the initial pH of cultivation media was 5.73. Among them, the amount of WCO added is calculated based on the same C atoms as glucose in YPD medium. The strains and plasmids used in this study are listed in Supplementary Table S1. The PCR primers used in this study were synthesized by Genewiz (Jiangsu, China) and are listed in Supplementary Table S2.

The Y. lipolytica expression vector pYLEX1 used in this study possesses the strong promoter hp4d, and its detailed information was provided in Li et al. (2021). Using primers CAD1-F/R and CAD2-F/R that were synthesized according to the existing sequence (GenBank ID: AB326105.1) in NCBI GenBank, two fragments of the CAD gene without introns were amplified from the A. terreus HAT418 genome and cloned into pYLEX1 to yield pYLEX1-CAD through adapted homologous recombination. The construction process of plasmid pYLEX1-CAD is depicted in Supplementary Figure S1. The sequences of the oligonucleotides used to amplify all the genes are listed in Supplementary Table S2 in the Additional file. Subsequently, the expression cassettes of other gene candidates were cloned into pYLEX1-CAD individually (Supplementary Figure S2). All recombinant plasmids were constructed using the One Step Cloning Kit from Vazyme Biotech Co., Ltd. (Nanjing, China). Transformants were plated on LB-ampicillin agar plates and incubated overnight at 37°C. Single colonies were inoculated into LB-ampicillin and cultured overnight at 37°C with shaking at 225 rpm. Plasmids were isolated, and the genes were verified by DNA sequencing.

Following that, all plasmids were linearized using the Spe Ⅰ enzyme and then transformed into the Y. lipolytica Po1g KU70Δ competent cells using lithium acetate/single-stranded vector DNA/polyethylene glycol method. The linearized plasmids introduced were integrated at the pBR322 locus of the strain Po1g KU70Δ. After 2 to 3 days of culture, the positive Y. lipolytica transformants were selected on YNB-LEU plates and subsequently confirmed by genomic DNA PCR analysis (Yu et al., 2016). Accordingly, in this study, the engineered Y. lipolytica Po1g KU70Δ strain was used as the host for all genetic modifications with gene knockouts and chromosomal expression constructs introduced via engineered pYLEX1 plasmids.

Seed inoculum of Y. lipolytica were first cultured in a 20 ml tube with 5 ml YPD medium and incubated for 24 h in a shaking incubator set at 30°C and 220 rpm. Next, a 250 ml flask was filled with 50 ml YPO medium and inoculated at the seeding density of OD₆₀₀ 0.1. The inoculated finished shake flasks were grown in a shaking incubator set at 30°C and 220 rpm. Fermented yeast cultures were collected on the fourth day and analyzed by GC-MS to determine and identify the IA content.

The ICLΔ strain was generated by knocking out the ORF region gene of ICL via the homologous recombination (HR) mechanism, which replaced ICL with the hygromycin B resistance marker gene (HPH) amplified from pSH69-Hph using the primer pairs ICL-Hph-F/R. To this end, two targeting arms (upstream and downstream flanking sequences of ICL), each approximately 1,000 bp in length, were amplified using PCR from the genomic DNA of Po1g-2G and ligated to the 5′ and 3′ ends of the HPH gene, respectively. After transformation of the ICL disruption cassette into Y. lipolytica cells, a gene replacement event occurs via double-crossover homologous recombination within the two flanking homology arms at the targeted locus. Transformants were grown in the YPDH solid medium (30°C, under dark conditions) supplemented with hygromycin and chosen randomly. The correct ICLΔ strain was confirmed by PCR with ICL-Hph-knock-F and ICL-Hph-knock-R primers. The construction of the CATΔ strain was carried out using a similar procedure.

To test the peroxisomal targeting ability of enhanced peroxisome targeting signal ePTS1, yeast cells expressing hrGFPO-ePTS1 were cultured in 50 ml YPD medium for 24 h. For simultaneous visualization of hrGFPO and Nile red, precultures incubated in 50 ml YPD were stained by adding Nile red solution (1 mg/ml) in acetone to the cell suspension (0.1 v/v) and incubated for 60 min in the dark at room temperature. The stained cells were washed with normal saline and resuspended in potassium phosphate buffer (pH 7.4) before being transferred onto glass slides to visualize hrGFPO at 488 nm and Nile red at 561 nm with an Olympus FV1000 confocal laser scanning microscope.

2 ml of the fermented supernatant was added to 1.5 ml of 10% HCl-CH₃OH solution, which was esterified at 62°C for 3 h. Then, 2 ml of n-hexane was added and the resultant mixture was violently shaken for 1 min to dissolve the dimethyl itaconate. After centrifugation (6,000 rpm, 5 min), the upper organic phase was transferred into another clean bottle for detection.

0.6 μl of the upper organic phase from the above Section was analyzed by GC-MS using an Agilent 7890A GC with a 5975C MSD equipped with an HP-5MS column (30 m × 0.25 mm × 0.25 μm, Agilent, Santa Clara, CA, United States). The GC oven temperature was initially held at 60°C for 2 min, and then ramped up to 250°C at a rate of 10°C/min and held for 9 min. The split ratio was 10:1. Helium was used as the carrier gas, with an inlet pressure of 13.8 psi. The injector was maintained at 250°C and the ion source temperature was set to 220°C. The final data analysis was performed using the Enhanced Data Analysis software (Agilent, Santa Clara, CA, United States) to obtain the standard curve of dimethyl itaconate, and the area obtained after the sample is analyzed and detected by the instrument is brought into the formula of the standard curve to obtain the output of dimethyl itaconate. The titer of IA is obtained by converting with the esterification rate obtained in the above section.

Differences in titers between the control strain and other strains were evaluated using SPSS 22.0 software for Windows (SPSS, Chicago, IL, United States). One-way ANOVA analyses were carried out with a confidence interval of 95% and statistical significance between the groups and the relevant control was considered if p-value < 0.05.

Bioreactor fermentation was batched processed using an optimal medium formulation containing 59 g/L WCO, 16 g/L yeast extract, 8 g/L peptone and 10 g/L tween-80. The strain was first seeded in 50 ml YPD medium in 250 ml shake flasks, cultured at 30°C and 220 rpm for 16 h. Following that, the bioreactor containing 3 L of YPO medium were inoculated with the seed cultures at an OD₆₀₀ of 1.

Fermentation without any pH control was carried out in a 5 L stirred fermenter (Shanghai Baoxing Bioengineering Equipment Co., Ltd., Shanghai, China) at 30°C and 1 vvm. The bioreactor pressure was maintained at 0.06 MPa. The impeller stirring speed was 400 rpm.

In A. terreus, IA is generated from the decarboxylation of the TCA intermediate cis-aconitic acid by the CAD enzyme (Bonnarme et al., 1995; Tevz et al., 2010). To test if A. terrus’s CAD gene can be expressed successfully in Y. lipolytica without codon optimization, we first cloned the associated gene from A. terreus HAT418 strain into Y. lipolytica strain Po1g KU70Δ, with the gene’s intron spliced out. In the gene sequencing analysis that followed, we discovered that the actual PCR-amplified gene sequence was different from the genome sequence shown in NCBI database. Our sequence data for A. terreus HAT418 CAD gene was submitted to GenBank under the accession number MT862134.1. Overexpression of the CAD gene in Y. lipolytica Po1g KU70Δ resulted in the creation of strain Po1g-CAD. We subjected both the engineered strain with cytosolic CAD and control strain without CAD to shake flask fermentation and assayed for IA continuously over a period of 6 days. We confirmed that IA was produced only in the engineered Y. lipolytica but not in its wild type. IA levels were first detected in the supernatant on day 2 and they increased gradually with time until a maximum yield of 33.12 mg/L was obtained on day 4 (Figure 2). This is contrasted with the Po1g KU70Δ original strain where no IA production detected, thereby confirming that the heterologous expression of the A. terreus CAD is necessary for IA production in Y. lipolytica.

β-oxidation of long chain fatty acids in eukaryotes are known to occur mainly in the peroxisomes (Wache et al., 2001; Hanko et al., 2018). In Y. lipolytica, this process produces acetyl-CoA which then enters the glyoxylate cycle for synthesis of the IA precursor, cis-aconitic acid (Dominguez et al., 2010; Koivistoinen et al., 2013; Xu et al., 2017). Several studies have shown that subcellular localization of specific enzymes or metabolic pathways not only increase product conversion efficiency, but is also able to suppress the undesirable effects of competitive metabolic inhibition (Zhu et al., 2018; Yang et al., 2019; Zhu et al., 2021). As such, this approach of subcellular compartmentalization is adopted in our study and complemented with the use of WCO as the substrate to enable sustainable, efficient and low-cost production of IA. To this end, IA production from the glyoxylate cycle in Y. lipolytica was ensured by targeting the involved heterologous enzymes to the peroxisomal matrix through the addition of enhanced peroxisomal targeting signal (ePTS1) after its gene sequence. The ePTS1 applied in this instance has been shown to be localized in S. cerevisiae (DeLoache et al., 2016).

Two separate dyes, Nile red and green fluorescence, were employed for staining of the yeast cells to validate the peroxisomal targeting ability of ePTS1. In an earlier study, it was shown that hrGFPO, encoding the green fluorescence protein, was most strongly expressed in Po1g KU70Δ (Zhao et al., 2021a). The plasmid with sequence ePTS1 added after the hrGFPO protein sequence was retransformed into yeast, resulting in strain Po1g-hrGFPO-ePTS1 (Figure 3A). Nile red fluorescence, on the other hand, was used to stain the peroxisomes of the yeast cells. To determine if ePTS1 could be successfully localized to peroxisomes in Y. lipolytica, Laser Scanning Confocal Microscopy (LSCM) was performed to observe the location of the two different fluorescence in yeast cells. As shown in Figure 3B, a green fluorescent protein with localization signal ePTS1, which exhibits green light under microscope irradiation, was expressed in the engineered yeast. Yeast cells after Nile red staining also show localized red fluorescence under the microscope. Combining these two images, we observed that the green and red shades overlap almost completely and produce a bright yellow light. Therefore, it can be confirmed that ePTS1 plays a role in determining the location of the peroxisome could be used as a peroxisomal targeting sequence for Y. lipolytica.

Subsequently, the plasmid pYLEX1-CAD-ePTS1 constructed through the ligation of ePTS1 downstream of the CAD gene was integrated into the Y. lipolytica Po1g KU70Δ chromosomes of the strain. The resulting engineered strain was cultured in the YPO medium and the 6 days course of IA production titers and biomass were shown in Figure 4. The titers of IA increased continuously from the beginning of cultivation up to day 4 with the highest titer having reached 1.58 g/L. Following this, the titers of IA gradually stabilized, likely owing to WCO depletion. Notably, we also compared the use of WCO and glucose in this subcellular compartmentalized approach to generate IA under the same conditions. The use of WCO had resulted in almost 100-folds increase in IA titer as compared to glucose (13.68 mg/L of IA) as the carbon source, hence implying that WCO was superior to glucose for IA production in these conditions. We also observed that the overproduction of IA has a positive effect on the cell growth. Together, our results demonstrate that the expression and localization of CAD in the peroxisomes of Y. lipolytica can lead to substantial increase in IA production.

To further enhance IA production in Y. lipolytica, we attempted to study the pathway genes involved in the conversion of oils to fatty acids and the utilization of fatty acids to raise the flux of precursor acetyl-CoA. The β-oxidation of fatty acids is a four-reaction cycle comprising of oxidation, hydration, dehydrogenation, and thiolysis, which results in one molecule of acetyl-CoA released in the peroxisome (Braga and Belo, 2016). In Y. lipolytica, the first step of fatty acid β-oxidation can be catalyzed by six different acyl-CoA oxidases (POX1-6) (Beopoulos et al., 2008). The second and third steps of β-oxidation are catalyzed by a multifunctional enzyme (MFE1) (Black et al., 2000; Dulermo et al., 2013), and the final step is catalyzed by peroxisomal thiolase (POT1) (Wang et al., 2020). As such, the genes involved in the β-oxidation pathway were overexpressed in an attempt to increase the flux towards IA. Ten genes, consisting of LIP2 (encoding lipases, Zhang et al., 2021b), POX1-6 (Ledesma-Amaro and Nicaud, 2016), MFE1 (Haddouche et al., 2010), POT1 (Smith et al., 2000), and PEX10 (encoding a proteins required for peroxisome assembly, Zhang et al., 2021b), were overexpressed individually and investigated for their effects on IA overproduction to determine the genes critical for IA biosynthesis in the acetyl-CoA production pathway. To this end, ten strains were constructed on the basis of the strain expressing CAD-ePTS1 gene, including the ten endogenous genes in the acetyl-CoA production pathway of Y. lipolytica; all genes were integrated into the chromosomes of Y. lipolytica Po1g KU70Δ. These ten engineered strains were then cultured in YPO medium for 6 days in shake flasks. The IA titers of the strains showed that the overexpression of the individual corresponding genes could improve IA production compared to the control strain expressing only the respective CAD-ePTS1 gene. Among them, the POT1-overexpressed strain (hereafter named Po1g-2G), achieved the highest titers of 2.42 g/L for IA after 4 days of cultivation (Figure 5). The results indicated that overexpression of this key enzymes can effectively promote the fatty acid degradation process and release the most acetyl-CoA molecules for IA biosynthesis. This observation is consistent with several other studies in which POT1 has already been demonstrated to be the key rate-limiting enzyme in the β-oxidation pathway (Ma et al., 2020; Zhang et al., 2021b). Therefore, the engineered strain Po1g-2G was used for subsequent engineering efforts to boost IA production.

The yield of IA can be further improved by reducing the loss of the precursor acetyl-CoA and preventing the synthesis of cis-aconitic acid from the glyoxylate cycle into downstream products such as succinic acid. The carnitine acetyltransferases (CAT) is responsible for transporting acetyl-CoA between different organelles, which can reversibly link the acetyl units to the carrier molecule carnitine (Strijbis et al., 2008; Strijbis et al., 2010). Meanwhile, the isocitrate lyase (ICL) manages the conversion of isocitrate into succinic acid and glyoxylic acid (Koivistoinen et al., 2013). To verify if either of these enzymes assume a major role in IA production, the corresponding genes singly were deleted from Po1g-2G, resulting in the creation of strains Po1g-2G-CATΔ and Po1g-2G-ICLΔ. After cultivating the resulting strains in shake flasks in YPO medium, it was found that higher IA production reaching 3.33 g/L was observed in ICL knockout strain as compared to CAT knockout strain with 2.8 g/L titers. This suggests that blocking the downstream pathway improves IA production while blocking the efflux effect of the acetyl coenzyme in the peroxisome is not as advantageous. Therefore, the Po1g-2G-ICLΔ strain was selected as the final optimized strain.

One of the most crucial issues in platform chemicals production is in achieving a high product titers consistently (Gao et al., 2016; Li et al., 2020). To investigate the performance of IA-producing Y. lipolytica at conditions that are more relevant for large-scale application, a 5 L bioreactor was employed. Unlike the procedure conducted in the shaking flask fermentation method, here, the composition of the growth medium was altered and the approach of adding sufficient WCO substrate at once was adopted to avoid the problems caused by fed batch fermentation. In the phase of active cell growth between 24 and 96 h, the Po1g-2G-ICLΔ strain intensively produced IA. During this period, the average specific rate of IA synthesis was 0.8 g/L/h, and the maximum specific rate of 2.3 g/L/h was observed between the 76–96-h intervals (Figure 6). Hence, the maximum titer of IA was 54.55 g/L after the 96-h reaction in the fermenter. At the time of writing, this is the highest IA production achieved by a yeast host reported worldwide. As such, Y. lipolytica would be a promising industrial host for IA production from renewable feedstock. Our study also demonstrated that the circular bioeconomy concept can be an effective model for scale-up production of valuable biochemical, in particular with the valorization of WCO as raw material.