Metabolic responses of Aspergillus terreus under low dissolved oxygen and pH levels

Abstract

The metabolic responses of Aspergillus terreus NRRL1960 to stress conditions (low dissolved oxygen and pH with limited nitrogen and phosphate) in the two-phase fermentation were investigated in this study. The fermentation kinetics suggested that itaconate production was suppressed under low dissolved oxygen (DO) concentrations. A slight change in pH caused a significant change in itaconate production. The transcriptomic data revealed that under low DO concentration, the glycolytic pathway was uncoupled from the oxidative phosphorylation, resulting in the activation of substrate-level phosphorylation as an alternative route for ATP regeneration. The downregulation of pdh genes, the genes encoding ATP synthase and succinate dehydrogenase, confirmed the observation of the uncoupling of the oxidative TCA cycle from glycolysis. It was found that the upregulation of pyc resulted in a large pool of oxaloacetate in the cytosol. This induced the conversion of oxaloacetate to malate. The upregulation of the gene encoding fumarate hydratase with the subsequent formation of fumarate was found to be responsible for the regeneration of NADPH and ATP under the condition of a low dissolved oxygen level. The large pool of oxaloacetate drove itaconic acid production also via the oxidative TCA cycle. Nevertheless, the downregulation of ATP synthase genes resulted in the deficiency of the proton-pumping H⁺ ATPase and the subsequent stress due to the failure to maintain the physiological pH. This resulted in itaconate production at a low titer. The fermentation kinetics and the transcriptomic data provided in this study can be used for further process optimization and control to improve itaconate production performance.

Introduction

Aspergillus sp. has gained much commercial interest as the filamentous fungi in this genus have been reported as the industrial producers of enzymes and metabolites. Besides Aspergillus niger and Aspergillus oryzae, Aspergillus terreus has been used extensively in the commercial production of enzymes (cellulase and xylanase), organic acid (itaconic acid), pharmaceutical product (lovastatin), and several secondary metabolites (geodin, cyclosporin A, questrin, citrinin, and aspulvinone) (Schimmel 1998; Hajjaj et al. 2001; Nazir et al. 2010; Klement and Buchs 2013; Kocabas et al. 2014; Boruta and Bizukojc 2016). Due to its ability to produce several metabolites, many researchers have attempted to identify and to reconstruct the genomic sequences of A. terreus for better understanding and for controlling its metabolic network (Birren et al. 2005; Liu et al. 2013). The reconstructed genome-scale metabolic model based on genome annotation and literature mining accurately predicted the physiological responses to the environmental conditions (Liu et al. 2013). The bottleneck reactions in the metabolic pathway could be also elucidated. In bioprocess development, the model that explained the function of the metabolic network is necessary for fermentation process optimization as it reflects the titer, yield, and production rate of the desirable product. A suitable manipulation of A. terreus metabolism to reduce the formation of byproducts (e.g., geodin in lovastatin production and cell biomass in itaconate production) is required. Itaconate production by A. terreus has long been studied; however, the synthetic pathway has not been clearly elucidated (Liu et al. 2013). The rate-limiting reaction in itaconate production is still unclear and the metabolic evolution proposed to improve the production performance seems insufficient. To overcome the problems addressed previously, a comprehensive understanding of the physiology of A. terreus is important.

In this study, the metabolic responses of A. terreus and its end product formation were investigated during the cultivation under low dissolved oxygen concentrations and pH. The fermentation kinetics and the transcriptomic data analysis were employed to explain the physiological responses of A. terreus to the changes in the environmental conditions, especially under stress conditions (low dissolved oxygen concentration and pH). The results obtained in this study could be deposited in the literature mining as the database for manipulating A. terreus cultivation to achieve a high production performance of the desirable end products.

Materials and methods

Microorganisms, inoculum preparation, and medium compositions

A. terreus NRRL1960 was kindly provided by the Agricultural Research Service Culture Collection, US Department of Agriculture, Peoria, IL, USA. For the inoculum preparation, the culture was maintained on Czapek Dox agar and was incubated at 30 °C for 7 days. The spores were harvested with sterile deionized water. The spore concentration was adjusted to 10⁶ spores/mL by dilution with sterile deionized water. A total of 10 mL of the spore suspension was used to inoculate the bioreactor.

The medium composition of the one-phase fermentation consisted of (per liter) 180 g glucose, 0.1 g KH₂PO₄, 3 g NH₄NO₃, 1 g MgSO₄·7H₂O, 5 g CaCl₂·2H₂O, 1.67 mg FeCl₃·6H₂O, 8 mg ZnSO₄·7H₂O, and 15 mg CuSO₄·5H₂O. The pH of the medium was adjusted to 3.1 with 0.5 M H₂SO₄ (Krull et al. 2017).

In the two-phase fermentation, medium A containing (per liter) 30 g glucose and 5 g yeast extract was used initially for inducing spore germination and initial cell growth. The pH of medium A was adjusted to 3.0. Later, medium A was replaced by medium B to enhance metabolite production with the limited cell growth (Pimtong et al. 2017). The compositions of medium B were (per liter) 100 g glucose, 2.36 g (NH₄)₂SO₄, 0.11 g KH₂PO₄, 2.08 g MgSO₄·7H₂O, 0.13 g CaCl₂·2H₂O, 0.074 g NaCl, 0.2 mg CuSO₄·5H₂O, 5.5 mg FeSO₄·7H₂O, 0.7 mg MnCl₂·4H₂O, and 1.3 g ZnSO₄·7H₂O. The pH of medium B was adjusted to 2.0.

Cultivation of A. terreus in a 7-L stirred fermentor

The cultivation was performed in the 7-L stirred fermentor (Sartorius Biostat® B) with a filled medium volume of 3 L. Before the cultivation, the fermentor was autoclaved at 121 °C and 15 psig, for 30 min. After sterilization, the fermentor was cooled down before starting the control system. A dissolved oxygen (DO) sensor (Mettler Toledo InPro®6820) was calibrated with sterile pure nitrogen and air. A dissolved CO₂ sensor (Mettler Toledo InPro®50,009(i)) was calibrated with CO₂ gas before the cultivation. The fermentation was initiated by inoculating the fermentor with the spore suspension.

For the typical one-phase fermentation cultivation, after the inoculation, the fermentor was controlled at 30 °C, 350 rpm, and 20% DO, with an initial pH of 3.1. The cultivation was carried out for 132 h (Krull et al. 2017).

For the two-phase fermentation process, the cultivation started with the inoculation of the fermentor that was filled with medium A. The fermentor was operated at 30 °C, 100 rpm, and 0.5 vvm air, for 48 h. During the growth phase, spores were germinated, and the hyphal growth was initiated until a sufficiently high cell concentration was obtained. At the end of the growth phase, medium A was aseptically discharged using the peristaltic pump and the fermentor was filled with the sterile medium B. The culture was incubated at the same temperature but with varied pH and DO levels to induce the production of fermentation metabolites. During the production phase, the pH was automatically controlled by the addition of 5 M KOH. The DO profile and the CO₂ concentration were monitored during the cultivation in the production phase.

Broth samples were collected from the fermentor every 12 h throughout the cultivation for analyses of the remaining glucose and ammonium and for the presence of fermentation metabolites. The cell biomass was collected for dry weight (DW) and transcriptomic analyses. All the experiments were conducted in duplicate for each condition. All the datasets presented for this study are the average values obtained from the duplication.

Complementary DNA library preparation and transcriptomic sequencing and analysis

To understand better the metabolic changes of A. terreus NRRL1960 upon the manipulation of the process conditions during the two-phase fermentation, the fermentation sample at the end of the first phase and those samples during the second phase were collected for RNA extraction, subsequent transcriptome sequencing, and differential expression analysis.

Barcoded RNA libraries were prepared using Lexogen’s Quant-Seq 3′ mRNA seq kit from Ion Torrent (Lexogen, Vienna, Austria). This approach generated libraries of the sequences close to the 3′ end of the polyadenylated RNAs, and only one fragment was produced per transcript. The RNA input was quantified by spectrophotometry and by fluorometry. Approximately 300–500 ng of the RNA input was required for library generation. The modified Quant-Seq protocol was performed as suggested for this low input, partially degraded RNA. The External RNA Controls Consortium (ERCC) obtained from Life Technologies (Carlsbad, CA, USA) was spiked into each of the prepared library reactions at the manufacturer’s recommended concentration. The library quantification and quality control were performed using the high sensitivity DNA kit and the Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA).

The DNA templates for sequencing were prepared using the 200 bp v3 OT2 kit and the Ion One Touch 2 platform (Life Technologies). Sequencing was performed on the Ion Proton, with signal processing and base calling performed by the Ion Torrent Suite v5.0.4 (Life Technologies). The raw sequences from each sample were uploaded to Partex Flow. The adapter sequences were trimmed. The bases were then trimmed according to the visual representation and the quality score to a size below 170 bp. The resulting reads were then aligned to A. terreus NIH2624 genome references downloaded from http://fungi.ensembl.org/info/website/ftp/index.html using Star 2.4.1d (UniProtKB) (Protein Knowledgebase (UniProt KB) 2017). The transcript quantification was performed using the Partex E/M method available in Partek Flow.

The trimming of the adapters including TGGAATTCTCGGGTG, CACCCGAGAATTCCA, AATCTCGTATGCCGTCTTCTGCTTGC, AGATCGGAAGAGCTCGTATGCCGTCTTCTGCTTG, AGATC GGAAGAGCGTCGTCTAGGGAAAGAGTGT, and AGATCGGAAGAGCGGTTCAGCAGGAATGCCGAG was performed. The sequence alignment with the A. terreus NIH2624 (assembly ASM14961v1) genome was performed using STAR v2.4.2a. Samtools v0.1.19-96b5f2294a was used with the default setting to generate the bam index files. The quantification of gene expression was performed by Htseq v0.6.0 with–f and–s options indicating bam inputs and un-stranded reads, respectively. Cuffdiff v2.1.1 was used to estimate the transcript abundance with the --no-diff and default options to generate the differential analysis files for each of the described comparisons.

Sample analyses

The cell growth and product formation during the fermentation were determined. The product yield was calculated from the ratio of the product formed to the amount of glucose consumed. The volumetric productivity was defined as the total amount of the product formed per unit volume per unit time.

The cell biomass concentration was determined from the cell dry weight. The collected fermentation sample was filtered through Whatman filter paper no.4 to harvest the cell biomass. The cell biomass was rinsed thoroughly with deionized water and dried at 80 °C until a constant dry weight was obtained. The cell concentration (in g/L) was calculated.

High-performance liquid chromatography (HPLC) was used to analyze the remaining glucose and the presence of fermentation metabolites. The fermentation broth was centrifuged at 10000 g for 7 min. The supernatant was collected for the HPLC sample preparation. The sample was diluted and was filtered through a hydrophilic PTFE membrane. The particle-free sample (15 μL) was injected automatically (Shimadzu DGU-20A 5R, Shimadzu SIL-20A HT) into an organic acid column (Biorad, Aminex HPX-87H ion exclusion organic acid column; 300 mm × 7.8 mm) maintained at 40 °C in a column oven (Shimadzu, CTO-20A). A solution of 0.005 M H₂SO₄ was used as the mobile phase at the flowrate of 0.6 mL/min (Shimadzu, LC-20AD). The refractive index detector (Shimadzu RID-20A) was used to detect the remaining glucose while the organic compounds, e.g., glucose, itaconic acid, citric acid, cis-aconitic acid, succinic acid, oxaloacetic acid, L-malic acid, pyruvic acid, fumaric acid, and ethanol, present in the sample were detected by the UV detector at 210 nm (Shimadzu SPD-20A). The standards containing 0–2 g/L of each compound mentioned previously were injected as the references to determine the concentration of each compound in the sample. The chromatogram peak area was selected for comparison basis.

The residual ammoniacal nitrogen was determined by the indophenol blue method (Tzollas et al. 2010). Before the assay, the following reagents were prepared accordingly: (1) phenol-alcohol solution was prepared by dissolving 10 g phenol (reagent grade) into 100 mL ethanol (95% v/v); (2) 0.5% sodium nitroprusside was prepared by dissolving 1 g sodium nitroprusside into 200 mL water; (3) alkaline solution was prepared by dissolving 100 g trisodium citrate and 5 g sodium hydroxide into 500 mL water; (4) sodium hypochlorite solution was prepared from commercial hypochlorite (Chlorox); and (5) oxidizing solution was prepared on the same day before use from a mixture of 100 mL sodium citrate solution and 25 mL hypochlorite solution. The reaction assay was started by adding a 500 μL sample into the mixture of phenol solution (20 μL), sodium nitroprusside solution (50 μL), and oxidizing solution (125 μL). The reaction was developed for 1 h in darkness. The optical purity of the mixture was measured at the wavelength of 640 nm (Thermo Scientific Multiskan GO), and the ammonia concentration in the sample was calculated from the standard curve.

The residual phosphate was determined by the malachite green method (Baykov et al. 1988). The dye solution was prepared before the assay by adding 60 mL concentrated sulfuric acid into 300 mL water. The solution was allowed to cool down to room temperature before 0.44-g malachite green was supplemented. The orange solution was then prepared on the same day before use by mixing 2.5 mL ammonium molybdate (7.5%) and 0.2 mL Tween 20 (11%) into 10 mL dye solution. For the reaction assay, a 100 μL sample was mixed with a 400 μL orange solution. The absorbance at 630 nm (Thermo Scientific Multiskan GO) was read within 10 min after mixing. The phosphate concentration of the sample was calculated from the standard curve.

Results and discussion

Metabolite production profiles by A. terreus NRRL1960 from two cultivation approaches

The fermentation kinetics during the 1-phase fermentation cultivation of A. terreus NRRL1960 were observed (Fig. 1). Growth started immediately after inoculation; however, the production of the metabolites was observed after 48 h. This resulted in the low overall volumetric productivity of the metabolites. From the fermentation profiles, it was found that glucose was mainly consumed by A. terreus NRRL1960 for cell biomass production (45.1 g/L). Fumaric acid was the major metabolite found in the 1-phase fermentation at the concentration of 41.6 g/L. Succinic acid was found as the second major end product at a concentration lower than 10 g/L. Other intermediates in the TCA cycle, e.g., citric acid and cis-aconitic acid were present in trace amounts while no malic acid and itaconic acid were observed. The evidence of fumarate formation from the 1-phase fermentation by A. terreus has been occasionally reported in the previous literature; nevertheless, the finding in this study was confirmed by the HPLC chromatograms of the fermentation samples compared with those of the standards (data not shown) (Jimenez-Quero et al. 2016).

Fig. 1

Metabolite production profiles in 1-phase fermentation by A. terreus cultivated at 30 °C, 350 rpm, 20% DO, and initial pH 3.1

The two-phase fermentation was introduced in the cultivation of A. terreus to investigate the production of fermentation metabolites under different environmental conditions. In the two-phase fermentation, spore germination and fungal growth were initiated during the first phase; whereas in the second phase, the growth was limited, allowing the fungi to convert the substrates to other fermentation metabolites (Thitiprasert et al. 2016; Pimtong et al. 2017). By this approach, it can be seen clearly that cell biomass concentration remained relatively constant during the second phase (Fig. 2). Itaconic acid was found to be produced as the major product during the second phase. Fumaric acid and CO₂ production were observed with the trace amount of other metabolites in the TCA cycle (data not shown); however, the concentration was relatively low in comparison to those observed in the typical two-phase cultivation. From the findings in the two-phase fermentation, further experimental observation was conducted under varied environmental conditions for a better understanding of the metabolic response of A. terreus NRRL1960 to the process conditions, especially under stress conditions.

Fig. 2

Metabolite production profiles of A. terreus during the production phase in two-phase fermentation. The fermentor was controlled at 30 °C, 100 rpm, 10% DO, and pH 2.0. The C/N weight ratio of the production medium was 100/2.36

Dissolved oxygen and pH regulated metabolite production in the 2-phase fermentation

Table 1 represents the fermentation kinetics data of A. terreus NRRL1960 during the second phase in the two-phase fermentation cultivated at low DO levels (10, 15, and 20%). Itaconic acid and fumaric acid were observed in the fermentation broth during the second phase in all of the conditions studied. From the kinetics profiles (data not shown), ethanol was observed later in some operating conditions at the end of the fermentation (Table 1). The formation of ethanol revealed the evidence of oxygen limitation during the cultivation. The final cell concentration and yield fluctuated because of morphological changes occurring during the production phase, which caused difficulties in sampling and quantitative measurement. The results show that increasing the DO level and pH rather lowered the production of the end metabolites (itaconic acid and fumaric acid). Nevertheless, the metabolite production was still limited as observed from the low yield and productivity when compared with the fermentation study operated at the high DO level (Kuenz et al. 2012; Krull et al. 2017).

Table 1 Effect of dissolved oxygen and pH on the metabolic responses of A. terreus NRRL1960 cultivated in medium B at 30 °C and 100 rpm during the second phase in the two-phase fermentation

The phenomenon of low cell biomass production during the second phase in the two-phase fermentation can be explained by the activation of the reductive pyruvate carboxylation together with the oxidation of pyruvate via TCA cycle while the oxidative phosphorylation was uncoupled from the glycolysis. It was reported that the TCA cycle was active during fumaric acid production in case the reductive pyruvate carboxylation was the sole pathway. As a result, ATP generation for cell maintenance and metabolites transportation was limited (Kenealy et al. 1986). With the CO₂ fixation under the aerobic condition, pyruvate carboxylase catalyzed the conversion of pyruvate to oxaloacetic acid and subsequently to fumaric acid. Consequently, citric acid flux in the TCA cycle was somewhat low during the first phase when growth was presumably promoted. However, the glucose metabolism and CO₂ fixation could continue and could lead to the accumulation of citric acid flux in the TCA cycle later when nitrogen became limited which gradually decreased the cell biomass production (Romano et al. 1967).

In the typical 1-phase fermentation cultivation, end product formation followed the growth-associated product-formation kinetics model. Kuenz et al. reported that itaconic acid was produced simultaneously with the growth during the cultivation (Kuenz et al. 2012). It should be noted that itaconic acid fermentation by A. terreus is an incomplete oxidation process involving decarboxylation of cis-aconitic acid by cis-aconitate decarboxylase, which interrupts the TCA cycle (Gyamerah 1995a). On the other hand, in the two-phase cultivation, it was found that the metabolite production (both itaconic acid and fumaric acid) followed non-growth-associated product-formation kinetics. This phenomenon could be explained by oxidative phosphorylation being uncoupled from glycolysis under the low DO level with limited ammonium concentration (Fig. 3) (Gyamerah 1995b; Riscaldati et al. 2000; Karaffa et al. 2015). Consequently, the glycolytic flux was enhanced rapidly for ATP regeneration via substrate-level phosphorylation under this condition (Klement and Buchs 2013). With the limited nitrogen and oxygen, pyruvate flux was shifted toward the itaconic acid production route instead of completing the oxidative TCA cycle for biosynthesis.

Fig. 3

The fermentation kinetics of A. terreus under the limited nitrogen and phosphate during the production phase. The fermentor was controlled at 30 °C, 100 rpm, 20% DO, and pH 2.0. The C/N weight ratio of the production medium was 100/2.36

The evidence of both nitrogen and phosphate limitation confirmed the non-growth-associated product-formation kinetics of end metabolite production during the second phase in the two-phase cultivation (Riscaldati et al. 2000; Papagianni et al. 2005; Boer et al. 2010). Itaconic acid and fumaric acid were found from the fermentation after 96 h when the concentration of ammonium and phosphate became low at 0.29 and 13.0 mg/L, respectively (Fig. 3). This finding was consistent with our previous study indicating that itaconic acid was first observed as the major end product in the fermentation broth when the growth was limited (Songserm et al. 2015). Klement and Buchs also reported that the overproduction of itaconic acid by A. terreus required nutrient limitation to uncouple glycolysis from oxidative phosphorylation (Klement and Buchs 2013). Riscaldati et al. reported that itaconic acid production was observed in the fermentation broth when the phosphate concentration in the medium dropped to approximately 10 mg/L, with the continuously decreasing ammonium concentration from 20 mg/L to less than 1 mg/L, while the cell concentration was slowly increasing from 3 to 10–11 g/L (Riscaldati et al. 2000). Hevekerl et al. claimed that, under the optimized operating conditions, itaconic acid was produced by the 1-day cultivation when the phosphate concentration was decreased (Hevekerl et al. 2014a).

It was reported that in the absence of oxygen, itaconic acid production was immediately stopped in A. terreus NRRL1960 (Gyamerah 1995b). The production mechanism involving the protein synthesis was restored slowly only under the aerobic condition (Gyamerah 1995b). To be more specific, itaconic acid production stopped as ATP generation was inhibited (Kuenz et al. 2012; Klement and Buchs 2013). ATP was claimed to be responsible for maintaining a proper physiological pH inside the cells (i.e., near neutral pH), counteracting the acid produced in the fermentation process and the low external pH (Riscaldati et al. 2000; Hevekerl et al. 2014b; Krull et al. 2017). When the extracellular pH was decreased, the permeability of the cell membrane was decreased. The transport of acid metabolites and proton across the membrane only occurred when the sufficient ATP was present (Corte-Real and Leao 1990). In the yeast cultivation, it was stated that the diffusion rate of undissociated acids across the membrane was increased with the decreasing extracellular pH. The same phenomenon was found during the second phase in the two-phase cultivation of A. terreus in this study. The different extracellular concentration of itaconic acid was observed from the cultivation with the small change in pH. This can be explained by the different acid dissociation degree at the different cultivation pH which resulted in the different concentration ratio of the undissociated acid and its salt. A detailed explanation on the synthesis of end metabolites in A. terreus and the transport across the membrane can be found below.

Itaconic acid was appeared as the major end metabolite produced in the second phase of the two-phase fermentation. It was produced in an undissociated form (H₂IA) in the cytosol. The degree of dissociation shifted from H₂IA to the single-dissociated form (HIA⁻) and the double-dissociated form (IA²⁻) based on the intracellular pH, causing the release of protons and the acidification of the cytosol, which led to stress or growth inhibition of the cells. The dissociated forms (HIA⁻) and (IA²⁻) were not able to pass across the cytoplasmic membrane, while the acid form (H₂IA) could diffuse through the membrane freely via the major facilitator superfamily (Mfs) protein transporters, using the energy from electrochemical gradients across the membrane. The transport of itaconic acid across the cytoplasmic membrane also involved the proton-pumping H⁺ ATPase, which required ATP to drive the transport of protons across the membrane (Krull et al. 2017).

In case of fumaric acid production at low extracellular pH, it was reported that the undissociated form could passively diffuse back through the cytoplasmic membrane decreasing the intracellular pH. Nonetheless, the transport mechanism of fumaric acid in fungi has not been clearly understood. It is believed that increasing the activity of the dicarboxylic acid transporters could lower the intracellular fumarate concentration in the cytosol and, therefore, could provide the positive effect on the fermentation process performance (Roa Engel et al. 2008, 2011; Xu et al. 2012).

From the previous explanation, it can be summarized that a low pH is necessary in the cultivation of A. terreus NRRL1960 for the end metabolite production in order to provide the sufficient amount of proton in exchange with the acid metabolite transport. In addition, ethanol formation and low extracellular concentration of itaconate and fumarate confirmed the evidence of oxygen limitation and insufficient ATP regeneration in the two-phase fermentation system. For further improvement in production performance, the fermentation under high DO level is necessary.

Gene expression explained metabolic responses of A. terreus

To better understand the fermentation kinetics and the metabolic response of A. terreus NRRL1960 and, thus, to improve the fermentation performance for desirable end product further, we observed gene expression during the second phase in the two-phase fermentation where the end metabolite production was enhanced. From the selection of one fermentation condition as the representative of the fermentation operation under stress condition (low DO and pH) to conduct transcriptomic analysis, we attempted to correlate the fermentation kinetics results with the regulation of the genes of interest under the specified process conditions. Figure 4 shows the central metabolic pathway of A. terreus. Glucose was converted to the key intermediate pyruvate. The key responsible enzyme in this route was phosphofructokinase. Pyruvate was either converted to acetyl CoA by the pyruvate dehydrogenase complex or was carboxylated to oxaloacetate by pyruvate carboxylase. The formation of itaconic acid involved the shuttle of intermediate metabolites between the cytosol and the mitochondria and the formation utilized the different enzymes present in both of the cell compartments. Cis-aconitate was transported into the cytosol via the mitochondrial TCA transporter. Cis-aconitate was then converted to itaconic acid by cis-aconitate decarboxylase. Later, undissociated form of itaconic acid was secreted from the cytosol across the cytoplasmic membrane in exchange for protons via the Mfs transporter (Krull et al. 2017). Fumaric acid synthesis pathway started with the carboxylation of pyruvate by pyruvate carboxylase to oxaloacetate with the presence of ATP and CO₂. Later, oxaloacetate was converted to malic acid by malate dehydrogenase and then to fumaric acid by fumarase or fumarate hydratase (Xu et al. 2012). The transport mechanism of fumaric acid in filamentous fungi has not been studied extensively; however, the transport mechanism in yeast has been studied thoroughly. This was presumably applied for the transport of fumarate in the filamentous fungi (Roa Engel et al. 2008).

Fig. 4

Metabolic pathway of A. terreus and the key genes responsible in glycolysis and TCA cycle. The key genes in the green boxes were overexpressed during the production phase, while the genes in the red boxes were downregulated. The genes in the gray boxes were found to be expressed at certain levels

Figure 5 describes the transcriptional levels of genes encoding key enzymes involved in glycolysis, TCA cycle, oxidative phosphorylation, and biosynthesis of acid metabolites (itaconic acid and fumaric acid) during the second phase in the two-phase fermentation. The fermentation process was controlled at 30 °C, 100 rpm, 10% DO, and pH 2.0 (see Supplementation Material Online, Table S1). The expression of phosphofructokinase (pfk) genes of glycolysis was changed slightly during the production phase (Fig. 5a). The evidence of the downregulation of pyruvate dehydrogenase (pdhB, pdhC, pdhA, and pdhX) genes with the upregulation of the pyruvate carboxylase (pyc) gene from the heat maps indicated that more pyruvate was carboxylated to oxaloacetate as the fermentation proceeded (Fig. 5a). It was suggested that the limited nitrogen and phosphate and the low DO level were responsible for this metabolic shift (Wynn et al. 2001). The downregulation of the citrate synthase (cs) gene was the consequence of the downregulated pdh genes, while the expression of the gene encoding aconitase fluctuated (Fig. 5b) (Patel et al. 2014). The upregulation of pyc at 144 h resulted in a large pool of oxaloacetate in the cytosol. This induced the conversion of oxaloacetate to malate, to be transported into the mitochondria, while cis-aconitate was transported across the mitochondrial TCA transporter. As a result, the major glucose consumption went to itaconic acid synthesis. This phenomenon was confirmed by the upregulation of the genes encoding mitochondrial TCA transporter (mttA), cis-aconitate decarboxylase (cadA), and Mfs transporter (mfsA) responsible for secreting itaconic acid across the cytoplasmic membrane to the fermentation broth (Fig. 5c) (Hossain et al. 2016). It was also found that the gene encoding fumarate hydratase was slightly upregulated (Fig. 5d). The time course data explaining the upregulation of the genes in itaconic acid cluster was consistent with the fermentation kinetics data describing the production of itaconic acid after 96 h cultivation.

Fig. 5

Heat maps summarizing the differential expression of the key genes responsible in glycolysis and TCA cycle during the production phase of A. terreus. The number in the box indicates the differential gene expression level compared with that at the starting time that the gene was expressed (48 or 72 h). The green box represents the upregulation of the gene. The red box represents the downregulation of the gene. The black box indicates a slight change in gene expression level (log2(fold_change) less than 2). ND represents the fragments per kilobase per million reads (FPKM) at that time was zero (no expression). a Glycolytic cluster, b citrate isomer, c itaconic acid cluster, d glyoxylate/dicarboxylate metabolism, and e ATP synthase cluster

The low production rate of itaconic acid could be explained by the downregulation of four ATP synthase genes (for the ATP synthase subunits d, 4, 5, and 9) (Fig. 5e). As a result, the proton-pumping H⁺ ATPase which requires ATP to drive the transport of proton across the membrane in exchange for intracellular H₂IA was shunted (Krull et al. 2017). Together with the upregulated genes of the itaconic acid cluster, this would result in the accumulation of H₂IA and eventually intracellular metabolic stress. In addition to the downregulation of ATP synthase genes, the downregulation of the succinate dehydrogenase gene confirmed that glycolysis was uncoupled from oxidative phosphorylation under the condition of limited nitrogen, phosphate, and dissolved oxygen (Fig. 5d). Succinate dehydrogenase was the only enzyme that participated in both the TCA cycle and the electron transport chain; therefore, the downregulation of this gene resulted in an incomplete TCA cycle (Hartman et al. 2014; Jimenez-Quero et al. 2016). This finding confirmed that glycolysis and oxidative phosphorylation were uncoupled, resulting in limitation of ATP availability and reduction of power regeneration (in the form of NADH and FADH) (Jimenez-Quero et al. 2016). Substrate-level phosphorylation was initiated to allow ATP regeneration, where fumaric acid was the electron acceptor, confirmed by the slight upregulation of the fumarate hydratase gene (Fig. 5d).

Conclusion

The central metabolic pathway of A. terreus involves both the biosynthesis and the transport of metabolites across cell compartments. In this study, two-phase fermentation was conducted to study the metabolic responses of A. terreus during cultivation under low DO and pH levels. The transcriptomic data analysis was employed to correlate the gene expression level with the metabolite production kinetics. Under the stress conditions with low DO level and limited nitrogen and phosphate, A. terreus produced itaconic acid as the major end product. Fumaric acid was found to the byproduct under these conditions. The production of both itaconic acid and fumaric acid indicated that oxaloacetate simultaneously entered both the reductive carboxylation and the oxidative TCA pathways. The transcriptomic data suggested that a high flux of itaconic acid occurred due to the upregulation of pyc. As a result, pyruvate was subsequently converted to oxaloacetate and malate was generated. Malate entered the mitochondria, while cis-aconitate was transported to the cytosol at the mttA transporter resulting in the incomplete oxidative TCA cycle and the uncoupled glycolysis and oxidative phosphorylation. It was observed that the gene encoding fumarate hydratase in the reductive TCA cycle was slightly upregulated. Although the genes in the itaconic acid cluster including mfsA were upregulated, the downregulation of ATP synthase caused a strong effect on itaconic acid production as it controlled the transport of protons in exchange for itaconic acid across the cytoplasmic membrane. From the fermentation kinetics and the transcriptomic analysis obtained in this study, it can be suggested that low pH is necessary in the cultivation of A. terreus NRRL1960 for the end metabolite production in order to provide the sufficient amount of proton in exchange with the acid metabolite transport. In addition, ethanol formation and low extracellular concentration of itaconate and fumarate confirmed the evidence of oxygen limitation and insufficient ATP regeneration in the two-phase fermentation system. For further improvement in production performance, the fermentation under high DO level is necessary.