Overexpression of Citrate Synthase Increases Isocitric Acid Biosynthesis in the Yeast Yarrowia lipolytica

Yarrowia lipolytica is a non-conventional yeast producing valuable compounds, such as citric acids, from renewable raw materials. This study investigated the impact of citrate synthase overexpression on the biosynthesis of citric and isocitric acid in Y. lipolytica. Two transformants of Y. lipolytica A101.1.31 strain (efficient citric acid producer), overexpressing CIT1 or CIT2 gene (encoding proteins with citrate synthase activity), were constructed. The results revealed that overexpression of either of these genes enhances citrate synthase activity. Additionally, the cit1 knockout strain was unable to use propionate as the sole carbon source, which proves that CIT1 gene encodes a dual activity protein–citrate and 2-methylcitrate synthase. In the overexpressing mutants, a significant increase in isocitric acid biosynthesis was observed. Both CIT1 and CIT2 overexpressing strains produced citric and isocitric acid from vegetable oil in a ratio close to 1 (CA/ICA ratio for wild-type strain was 4.12).


Introduction
Bulk chemicals such as organic acids can be readily produced with the use of microorganisms. These compounds are often intermediates of primary metabolism, as in the case of citric acid (CA), the Krebs cycle intermediate. Due to its natural origin, low toxicity and chemical properties, CA has found many applications in food and pharmaceutical industries as an acidifying and flavoring agent, antioxidant, preservative and stabilizer. This compound is also used as a cleaning agent in household chemicals as a replacement for environmentally hazardous, phosphate-containing cleaners. The global CA market was worth 2.5 billion USD in 2016 [1]. The demand for CA is constantly growing, especially for the food and beverages industry (57.9% of the CA market in 2016), where a CA growth rate of 5% from 2017 to 2025 is expected. This high demand can be met through its production in microbial-based processes. The well-established technology harnessing the fungus Aspergillus niger to produce CA from glucose or molasses provides high product yields [2]; however, it causes some environmental issues [3,4].
The yeast Yarrowia lipolytica, known for its ability to accumulate large amounts of lipids [5], is also an excellent CA producer [6][7][8][9][10]. This feature, combined with its ability to use a wide spectrum of substrates, such as n-alkanes, oils, hydrolyzed molasses and glycerol [11], makes CA production by Y. lipolytica an interesting alternative to processes with A. niger. Moreover, this yeast is non-pathogenic, and several processes, including yeast biomass and CA production, were classified as GRAS (Generally genes in Y. lipolytica were published. In this study, two genes (CIT1 and CIT2) encoding proteins with citrate synthase activity were identified and two Y. lipolytica transformants overexpressing these genes, able to produce both CA and ICA from renewable raw materials, were characterized. Sustainability 2020, 12, x FOR PEER REVIEW 3 of 15 transformants overexpressing these genes, able to produce both CA and ICA from renewable raw materials, were characterized.

Strains
The Y. lipolytica strains used in this study are listed in Table 1. All the targeted transformants were derived from the Y. lipolytica A101.1.31 strain, which is a UV mutant obtained from the wild-type Y. lipolytica A101 strain [36].

Growth Media and Conditions
E. coli strains were cultivated in 37 • C in LB medium consisting of 5 g dm −3 tryptone (Merck, Darmstadt, Germany), 10 g dm −3 yeast extract (Merck) and 10 g dm −3 NaCl. For plates, 2% agar was added. Solid LB medium with 0.1% kanamycin (Thermo Fisher Scientific, Waltham, MA, USA) was used for clone selection.

General Genetic Techniques, Plasmid Construction and Transformation
Standard molecular genetic techniques were adopted from Sambrook and Russell [38]. Restriction enzymes were obtained from New England Biolabs (Ipswich, MA, USA). Genomic DNA from yeast was extracted and purified as described in [39]. Nucleotide sequences of YALI0E00638g (CIT1) and YALI0E02684g (CIT2) genes were retrieved from GRYC database [40] Genomic DNA from Y. lipolytica A101 wild-type strain was used as a template for PCR using primers listed in Table S1. The BclI or BglII and AvrII restriction sites were used for forward and reverse primers, respectively (underlined sequences). PCR was carried out using a Biometra T1 Personal thermal cycler (Biometra, Göttingen, Germany) using Phusion High-Fidelity DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA). The PCR products were verified by gel electrophoresis and the DNA fragments were recovered from the agarose gels using Gel-Out Kit (A&A Biotechnology, Gdynia, Poland). Amplified genes were cloned into JMP62URA plasmid digested with appropriate restriction enzymes and transformed into E. coli DH5α according to [38]. The obtained plasmids for gene overexpression contained a strong constitutive pTEF promoter, URA3 gene for selection of Y. lipolytica transformants and KanR marker for E. coli selection. The deletion cassettes were prepared according to [41]. In brief, the promoter (P) and terminator (T) regions (~1kb) were amplified from Y. lipolytica A101 genomic DNA with the gene-specific P1/P2 and T1/T2 primers (Table S1). Additionally, P2 and T1 primers contained an extension bearing I-sceI restriction site. For the CIT1 gene, primer pairs CIT1-P1/CIT-P2 and CIT1-T1/CIT2-T2 and for CIT2-CIT2-P1/CIT2-P2 and CIT2-T1/CIT2-T2 were used. The obtained P and T regions were subsequently purified and used in a second round of PCR. The fused P-T fragments were ligated into pCR™-Blunt II-TOPO ® vector (Invitrogen, Carlsbad, CA, USA). Lastly, the URA3 marker was inserted at the I-sceI site yielding the constructs containing either CIT1-PUT or CIT2-PUT disruption cassette. Both the overexpression (random chromosomal integration) and deletion vectors (homologous recombination) were digested by NotI restriction enzyme to release the desired cassette. All Y. lipolytica transformations were performed using the lithium acetate procedure [42]. Positive Y. lipolytica transformants were verified by PCR and gel electrophoresis after genomic DNA extraction.

Bioreactor Studies
To test the ability of Y. lipolytica transformants to produce CA and ICA, as well as to determine the citrate synthase activity, bioreactor cultures were performed. Strains were maintained in 5-dm 3 stirred-tank BIOSTAT B-PLUS bioreactors (Sartorius, Frankfurt, Germany) under the following conditions: working volume-2 dm 3 , temperature-28 • C, stirring rate-2.3 Hz and aeration rate-1 dm 3 dm −3 min −1 . The CA production media consisted of 150 g glycerol, 3.0 g NH 4 Cl, 0.2 g KH 2 PO 4 , 1.0 g MgSO 4 × 7H 2 O and 1.0 g yeast extract in 1 dm 3 of tap water. pH was automatically maintained at 6.0 using 5 mol dm −3 NaOH solution. Medium for CA production using vegetable oil contained 200 g rapeseed oil (50 g at the beginning of the culture, 50 g after 24 h and 100 g after 48 h), 6 g NH 4 SO 4 , 2.0 g KH 2 PO 4 , 0.2g Na 2 HPO 4 , 1.4 g MgSO 4 × 7H 2 O, 0.8 g Ca(NO 3 ) 2 × 4H 2 O, 0.5 g NaCl and 0.5 g yeast extract in 1 dm 3 of tap water. pH was automatically maintained at 6.0 using 5 mol dm −3 NaOH solution. The inoculum (48 h) for all bioreactor cultures constituted 5% of the working volume and the initial optical density (OD 600 ) was adjusted to 0.5. Bioreactor cultures were performed in triplicate.

Analytical Methods
The analyses of substrate concentration, biomass and sum of citric acids during the bioreactor cultures were performed by taking two 10-cm 3 samples from each culture 10 min after inoculation (time = 0); subsequent sampling was conducted at 12th, 18th, 25th, 44th, 60th and 72nd hours of the culture. Each sample was centrifuged for 5 min at 2700 RCF; the supernatants and cell pellets were collected and subjected to further analyses.
To determine biomass concentration, cell pellets were washed twice with distilled water and filtered under vacuum using membranes with a pore size of 0.22 µm. Biomass was determined gravimetrically after cell drying at 105 • C. Biomass was expressed as grams of cell dry weight per 1 dm 3 (CDW g dm −3 ).
The concentrations of glycerol and citric acids (sum) in the culture supernatants were measured using HPLC (Dionex-Thermo Fisher Scientific, UK) with a carbohydrate H+ column (Thermo Scientific, Waltham, MA, USA) coupled to UV (λ = 210 nm) and refractive index (RI) detectors (Shodex, Ogimachi, Japan). The eluate used was 25 mM trifluoroacetic acid (TFA). The conditions of the analyses were as follows: temperature-65 • C and flow rate-0.6 cm 3 min −1 .
Citrate synthase activity assay using Y. lipolytica whole cell extracts (10 min of sonication using Sonics VCX500 sonicator, amplitude 100%, 0.5 s intervals) was performed with Citrate Synthase Assay Kit (Sigma-Aldrich, MAK193) using Evolution 300 UV-VIS spectrophotometer (Thermo-Scientific), according to the manufacturer's protocol. Total protein concentration was measured using Lowry's method [43]. All the assays were performed in triplicate.

Calculation
The mass yield coefficient of product formation (Y), expressed as grams of product per gram of substrate consumed, was calculated from the following equation: The specific productivity (q, grams of product per gram of biomass per hour) was calculated using The volumetric productivity of citric acids (Q, grams per liter per hour) was calculated from where P is the total amount of CA or ICA in the culture medium at the end of cultivation [g], S is the total amount of substrate consumed [g], V is the initial volume of culture [dm −3 ], X is the concentration of biomass in a bioreactor [g] and t represents the time of cultivation [h].

Genes Encoding Citrate Synthases in Y. lipolytica
To identify citrate synthase genes in the yeast Y. lipolytica, amino acid sequences of Saccharomyces cerevisiae citrate synthases (ScCit1-YNR001C, ScCit2-YCR005C and ScCit3-YPR001W) retrieved from SGD were compared to the available protein sequences of Y. lipolytica CLIB122 using Protein BLAST [48]. It emerged that two sequences, annotated as YALI0E00638p (YlCit1, 54% amino acid sequence identity to ScCit1, 52% to ScCit2 and 50% to ScCit3) and YALI0E02684p (YlCit2, 76% amino acid sequence identity to ScCit1, 68% to ScCit2 and 46% to ScCit3) were found. The alignment of the amino acid sequences ( Figure 2) showed substantial amounts of conserved motifs among the compared proteins. Interestingly, the YlCit1 protein contains ScCit3-like domains, important for conversion of propionyl-CoA into 2-methylcitrate. This observation suggests that YlCit1 may show 2-methylcitrate synthase activity.

CIT1 and CIT2 Overexpression Enhances Citrate Synthase Activity
Citrate synthase activity was analyzed using whole cell extracts of the corresponding transformants grown in glycerol-based medium. The results presented in Figure 3 show that an additional copy of either the CIT1 or CIT2 gene in Y. lipolytica enhanced citrate synthase activity 3.93fold (CIT1) and 6.34-fold (CIT2) at the 24th hour of culture compared to the control strain. The CA synthase activity in the control strain reached only 0.98 U/mg of protein, a value similar to the results demonstrated previously for Y. lipolytica VKM Y-2373 [30]. Notably, citrate synthase activity for both strains decreased over time ( Figure 3).

CIT1 and CIT2 Overexpression Enhances Citrate Synthase Activity
Citrate synthase activity was analyzed using whole cell extracts of the corresponding transformants grown in glycerol-based medium. The results presented in Figure 3 show that an additional copy of either the CIT1 or CIT2 gene in Y. lipolytica enhanced citrate synthase activity 3.93-fold (CIT1) and 6.34-fold (CIT2) at the 24th hour of culture compared to the control strain. The CA synthase activity in the control strain reached only 0.98 U/mg of protein, a value similar to the results demonstrated previously for Y. lipolytica VKM Y-2373 [30]. Notably, citrate synthase activity for both strains decreased over time (Figure 3).

CIT1 is Indispensable for Propionate Utilization in Y. lipolytica
In S. cerevisiae, the CIT3 gene encodes the dual activity citrate and methylcitrate synthase [49]. To evaluate the methylcitrate synthase activity of proteins encoded by CIT1 and CIT2 genes in Y. lipolytica, deletion mutants of A101.1.31 with either Δcit1 or Δcit2 were constructed. Growth comparison in a media with 1% glucose and 1% propionate was evaluated ( Figure 4). All tested strains exhibited usual growth on glucose; however, the Δcit1 transformant was unable to grow in the presence of propionate. This result clearly indicates that the CIT1 gene is essential for propionate metabolism ( Figure 1) and its product possesses the methylcitrate synthase activity which confirmed the prediction described above (Section 3.1).

Overexpression of CIT1 or CIT2 Increases ICA Biosynthesis
Y. lipolytica A101.1.31, an acetate negative mutant and CA overproducer [7,50], and its derivatives with overexpression of either CIT1 or CIT2 genes were cultivated in a glycerol-based medium promoting CA overproduction. Final biomass concentrations obtained by the transformants (18.35 g dm −3 and 18.80 g dm −3 for CIT1 and CIT2, respectively) were slightly higher (8% and 10% higher for CIT1 and CIT2, respectively) compared to the parental strain ( Figure 5, Table 2).

CIT1 is Indispensable for Propionate Utilization in Y. lipolytica
In S. cerevisiae, the CIT3 gene encodes the dual activity citrate and methylcitrate synthase [49]. To evaluate the methylcitrate synthase activity of proteins encoded by CIT1 and CIT2 genes in Y. lipolytica, deletion mutants of A101.1.31 with either ∆cit1 or ∆cit2 were constructed. Growth comparison in a media with 1% glucose and 1% propionate was evaluated ( Figure 4). All tested strains exhibited usual growth on glucose; however, the ∆cit1 transformant was unable to grow in the presence of propionate. This result clearly indicates that the CIT1 gene is essential for propionate metabolism ( Figure 1) and its product possesses the methylcitrate synthase activity which confirmed the prediction described above (Section 3.1).

CIT1 is Indispensable for Propionate Utilization in Y. lipolytica
In S. cerevisiae, the CIT3 gene encodes the dual activity citrate and methylcitrate synthase [49]. To evaluate the methylcitrate synthase activity of proteins encoded by CIT1 and CIT2 genes in Y. lipolytica, deletion mutants of A101.1.31 with either Δcit1 or Δcit2 were constructed. Growth comparison in a media with 1% glucose and 1% propionate was evaluated ( Figure 4). All tested strains exhibited usual growth on glucose; however, the Δcit1 transformant was unable to grow in the presence of propionate. This result clearly indicates that the CIT1 gene is essential for propionate metabolism ( Figure 1) and its product possesses the methylcitrate synthase activity which confirmed the prediction described above (Section 3.1).

Overexpression of CIT1 or CIT2 Increases ICA Biosynthesis
Y. lipolytica A101.1.31, an acetate negative mutant and CA overproducer [7,50], and its derivatives with overexpression of either CIT1 or CIT2 genes were cultivated in a glycerol-based medium promoting CA overproduction. Final biomass concentrations obtained by the transformants (18.35 g dm −3 and 18.80 g dm −3 for CIT1 and CIT2, respectively) were slightly higher (8% and 10% higher for CIT1 and CIT2, respectively) compared to the parental strain ( Figure 5, Table 2).

Overexpression of CIT1 or CIT2 Increases ICA Biosynthesis
Y. lipolytica A101.1.31, an acetate negative mutant and CA overproducer [7,50], and its derivatives with overexpression of either CIT1 or CIT2 genes were cultivated in a glycerol-based medium promoting CA overproduction. Final biomass concentrations obtained by the transformants (18.35 g dm −3 and 18.80 g dm −3 for CIT1 and CIT2, respectively) were slightly higher (8% and 10% higher for CIT1 and CIT2, respectively) compared to the parental strain ( Figure 5, Table 2). The total amount of citric acids at the end of the culture reached 10% and 23% higher values for CIT1 and CIT2 overexpressing transformants, respectively, compared to Y. lipolytica A101.1.31 (75.41 g dm −3 ). However, only when a closer look at the concentrations of CA and ICA separately was taken into consideration was a better picture of the metabolism revealed ( Figure 6, Table 2). CIT1 overexpression resulted in decreased CA biosynthesis and pushed the carbon flux towards ICA production. This was expressed by an almost 9.5-fold increase in ICA concentration compared to the parental strain. A similar, albeit not identical, situation was observed for the CIT2 transformant. In this case, the ICA amount at the end of the culture increased 6.8-fold, with a concomitant minor increase in CA concentration in comparison to A101.1.31. Therefore, the resulting CA to ICA ratios for both Y. lipolytica transformants were significantly lower and decreased from 29.8 for the parental A101.1.31 strain to 2.6 and 4.6 for CIT1 and CIT2 overexpressing strains, respectively. A high CA/ICA ratio for A101.1.31 cultivated on crude glycerol was previously described [14]. Additionally, Levinson and colleagues [51] reported CA/ICA ratios for wild strains of Y. lipolytica grown on glycerol ranging from 1.7 to 11.3.  The total amount of citric acids at the end of the culture reached 10% and 23% higher values for CIT1 and CIT2 overexpressing transformants, respectively, compared to Y. lipolytica A101.1.31 (75.41 g dm −3 ). However, only when a closer look at the concentrations of CA and ICA separately was taken into consideration was a better picture of the metabolism revealed ( Figure 6, Table 2). CIT1 overexpression resulted in decreased CA biosynthesis and pushed the carbon flux towards ICA production. This was expressed by an almost 9.5-fold increase in ICA concentration compared to the parental strain. A similar, albeit not identical, situation was observed for the CIT2 transformant. In this case, the ICA amount at the end of the culture increased 6.8-fold, with a concomitant minor increase in CA concentration in comparison to A101.1.31. Therefore, the resulting CA to ICA ratios for both Y. lipolytica transformants were significantly lower and decreased from 29.8 for the parental A101.1.31 strain to 2.6 and 4.6 for CIT1 and CIT2 overexpressing strains, respectively. A high CA/ICA ratio for A101.1.31 cultivated on crude glycerol was previously described [14]. Additionally, Levinson and colleagues [51] reported CA/ICA ratios for wild strains of Y. lipolytica grown on glycerol ranging from 1.7 to 11.3. The total amount of citric acids at the end of the culture reached 10% and 23% higher values for CIT1 and CIT2 overexpressing transformants, respectively, compared to Y. lipolytica A101.1.31 (75.41 g dm −3 ). However, only when a closer look at the concentrations of CA and ICA separately was taken into consideration was a better picture of the metabolism revealed ( Figure 6, Table 2). CIT1 overexpression resulted in decreased CA biosynthesis and pushed the carbon flux towards ICA production. This was expressed by an almost 9.5-fold increase in ICA concentration compared to the parental strain. A similar, albeit not identical, situation was observed for the CIT2 transformant. In this case, the ICA amount at the end of the culture increased 6.8-fold, with a concomitant minor increase in CA concentration in comparison to A101.1.31. Therefore, the resulting CA to ICA ratios for both Y. lipolytica transformants were significantly lower and decreased from 29.8 for the parental A101.1.31 strain to 2.6 and 4.6 for CIT1 and CIT2 overexpressing strains, respectively. A high CA/ICA ratio for A101.1.31 cultivated on crude glycerol was previously described [14]. Additionally, Levinson and colleagues [51] reported CA/ICA ratios for wild strains of Y. lipolytica grown on glycerol ranging from 1.7 to 11.3.   Table 2. Parameters of bioreactor cultures of Y. lipolytica A101.1.31, CIT1 and CIT2 at 72nd hour of culture in a glycerol-based medium. Culture conditions were set up as follows: glycerol 150 g dm −3 , 28 • C, pH 6.0, aeration rate 1 dm 3 dm −3 min −1 and stirring rate 2.3 Hz. All values are expressed as means derived from triplicate experiments. The SD did not exceed 5%. Abbreviations: X-biomass, CA-sum of citric acids, CA-citric acid, ICA-isocitric acid. The volumetric and specific productivity reflect the results described above ( Table 2). The highest volumetric productivity of CA was achieved by the CIT2 strain, while the lowest was achieved by the CIT1 overexpressing transformant. Inversely, this strain (CIT1) was the fastest ICA producer (0.16 g dm −3 h −1 ). In terms of specific CA productivity, the highest result was achieved by the parental A101.1.31 strain.
The best CA-producing strain from one gram of glycerol was the Y. lipolytica CIT2 transformant; however, the parental A101.1.31 strain was characterized by similar performance. The CIT1 transformant was the least efficient CA producer, converting only 42% of available glycerol into CA. In contrary, the CIT1 strain was the most robust ICA producer (15% conversion), while A101.1.31 converted only 2% of glycerol into ICA. These data are in agreement with previous observations [7]. The parental strain, Y. lipolytica A101.1.31, did not consume all the available glycerol, leaving around 16 g dm −3 of the substrate in the medium at the end of the process ( Figure 5). In contrast, the CIT1 overexpressing strain showed increased substrate consumption, which was reflected by the total utilization of glycerol during 72 h of the process. Similarly, overexpression of CIT2 resulted in a robust glycerol-utilizing phenotype. In the process using this strain, the substrate was depleted already in the 60th hour of the culture.
Taken together, overexpression of both genes (CIT1 and CIT2) does not increase CA but ICA biosynthesis for strains cultured in a medium containing glycerol as the sole carbon source.

ICA Production from Hydrophobic Substrate
Efficient biosynthesis of ICA from vegetable oil by Y. lipolytica was previously demonstrated [21]. Therefore, in the present study, the ICA-overproducing phenotype of Y. lipolytica transformants was analyzed also in a medium based on rapeseed oil optimized for ICA biosynthesis [21]. The A101.1.31 investigated during the first part of this study could not be used as a control strain, due to its inability to utilize even carbon fatty acids ( Figure S1); this is why the parental strain of A101.1.31-Y. lipolytica A101 was used instead. The final biomass concentration reached by the transformants (26.9 g dm −3 for CIT1 strain and 30.9 g dm −3 for CIT2 strain) varied significantly compared to the value measured for the control strain (55 g dm −3 , Table 3).
All analyzed strains secreted higher amounts of citric acids at the end of the cultures (Table 3) compared to the processes in glycerol-based media ( Table 2). In contrast to the glycerol-based cultures, the highest sum of citric acids was obtained by the CIT1 transformant (133.2 g dm −3 ), while the CIT2 strain reached 109.9 g dm −3 , which represented the lowest value observed for cultures on rapeseed oil (Table 3). However, as mentioned above, it is difficult to assess the true effect of the analyzed genetic modifications without a closer look to the CA and ICA concentration. The titers of both citric acids at the end of the transformants' cultures were very similar, which resulted in a CA/ICA ratio close to 1 (1.04 for CIT1 and 0.95 for CIT2), whereas the final CA/ICA ratio for the control strain reached 4.12. A higher amount of ICA produced on hydrophobic substrates by Y. lipolytica A101 was also reported by Wojtatowicz and colleagues [50]. According to [21], similar results to these presented in the current study were reported-CA:ICA ratios equal to 1.06:1 and 1.14:1 for Y. lipolytica 710 and 68, respectively. However, the best described strain, Y. lipolytica 704-UV4-A/NG50, reached a CA:ICA ratio of 0.37:1 [21]. In the pilot-plant scale culture on refined sunflower oil, as demonstrated in [20], the wild-type Y. lipolytica EH59 produced 93 g dm −3 of ICA, with an ICA:CA ratio of 1.14:1. Table 3. Parameters of bioreactor cultures of Y. lipolytica A101, CIT1 and CIT2 in a rapeseed oil-based medium. Culture conditions were set up as follows: rapeseed oil 200 g dm −3 (50 g at the beginning of the culture, 50 g after 24 h and 100 g after 48 h), 28 • C, pH 6.0, aeration rate 1 dm 3 dm −3 min −1 and stirring rate 2.3 Hz. All values are expressed as means derived from triplicate experiments. The SD did not exceed 5%. Abbreviations: X-biomass, CA-sum of citric acids, CA-citric acid, ICA-isocitric acid.  In contrary to glycerol cultures, however, according to the concentration of CA and ICA described above, the control A101 strain as well as the CIT1 transformant reached high volumetric productivity of citric acids during the processes with rapeseed oil ( Table 3). The CIT2 transformant reached only 0.58 g dm −3 h −1 . Both strains A101 and CIT1 presented also similar conversion of oil into citric acids. On the other hand, the highest specific production rate of citric acids was obtained by CIT1 and it was twice as high as the value reached by the control strain. This is due to the two times lower biomass concentration noted for the CIT1 transformant. All the analyzed parameters considering CA separately from ICA were higher for the A101 strain ( Table 3). The CIT1 and CIT2 transformants reached similar values for all parameters regarding both CA and ICA.
It seemed plausible that overexpression of citrate synthase would cause an increased concentration of CA in the mitochondria, which would therefore be available in abundance for ICA-generating aconitase, leading in turn to high ICA biosynthesis. This compound cannot be further metabolized due to inactivation of ICDH by nitrogen-limiting conditions [52]. It is also important to note that overexpression of citA citrate synthase in the CA-producing fungus A. niger did not improve CA biosynthesis [53].
Both CA and ICA are valuable natural compounds. Y. lipolytica strains constructed in this study were able to produce these two acids in similar quantities from renewable substrates, e.g., glycerol and rapeseed oil. In combination with the abilities of Y. lipolytica to use waste glycerol from the biodiesel industry [33,54] and waste cooking oil [55,56] to produce various chemicals, it would be of great interest to culture the transformants using these substrates in order to obtain ICA and/or CA, which can be readily separated by esterification [20]. The amounts of ICA obtained in this study are very promising compared to the literature data [6,30,37]. Higher concentrations of ICA from glycerol were obtained by increasing the glycerol uptake through overexpression of the GUT genes [9] or from ethanol by supplementing the culture medium with zinc and itaconic acid [13]. Moreover, different cultivation modes (repeated batch fermentation) positively affect ICA biosynthesis [23]. None of this research, however, directly modified the Krebs cycle to prove the possibilities to improve ICA biosynthesis. Further enhancement of ICA biosynthesis by this peculiar yeast, except medium supplementation and/or the "push" strategy, represented by increased glycerol uptake, could be presumably reached by overexpression of ACO1 gene encoding aconitase; however, this hypothesis remains to be verified.

Conclusions
The yeast Y. lipolytica is a highly appreciated industrial microorganism due to its wide range of applications in biotechnology. To our knowledge, this is the first study on the identification of genes encoding the citrate synthase in the yeast Y. lipolytica. We aimed to improve ICA biosynthesis by overexpressing either CIT1 or CIT2 genes. The activity assays showed that both genes encode proteins with citrate synthase activities; however, YlCit2 proved to be the main citrate synthase, whereas the protein encoded by the CIT1 gene possessed dual citrate and 2-methylcitrate synthase activity. The constructed Y. lipolytica transformants overexpressing either CIT1 or CIT2 genes produced noticeably more ICA in both glycerol-and vegetable oil-based media compared to the control strain. These results lay the foundation for future advancements in ICA biosynthesis by the means of genetic engineering and process optimization.