Comparative proteomic analysis of mitochondria isolated from Euglena gracilis under aerobic and hypoxic conditions

The unicellular microalga Euglena gracilis produces wax esters for ATP acquisition under low-oxygen conditions. The regulatory mechanism of wax ester production is not yet understood. Indeed, our previous transcriptomic analysis showed that transcript levels of genes involved in the wax ester synthesis hardly changed under hypoxic conditions, suggesting contribution of post-transcriptional regulation. In this study, we conducted a proteome analysis of E. gracilis mitochondria, as this organelle employs the fatty-acid synthesis pathway under hypoxic conditions. Mitochondria were isolated from E. gracilis SM-ZK strain treated with both aerobic and hypoxic conditions and used for shotgun proteomic analysis. Three independent proteomic analyses succeeded in identifying a total of 714 non-redundant proteins. Of these, 229 were detected in common to all experiments, and 116 were significantly recognized as differentially expressed proteins. GO enrichment analysis suggested dynamic changes in mitochondrial metabolic pathways and redox reactions under aerobic and hypoxic conditions. Protein levels of bifunctional enzymes isocitrate lyase and malate synthase in glyoxylate cycle were 1.35-fold higher under hypoxic conditions. Abundances of the propionyl-CoA synthetic enzymes, succinyl-CoA synthetase and propionyl-CoA carboxylase, were also 1.35- and 1.47-fold higher, respectively, under hypoxic conditions. Protein levels of pyruvate:NADP+ oxidoreductase, a key enzyme for anaerobic synthesis of acetyl-CoA, which serves as a C2 donor for fatty acids, showed a 1.68-fold increase under hypoxic conditions, whereas those of pyruvate dehydrogenase subunits showed a 0.77–0.81-fold decrease. Protein levels of the fatty-acid synthesis enzymes, 3-ketoacyl-CoA thiolase isoforms (KAT1 and KAT2), 3-hydroxyacyl-CoA dehydrogenases, and acyl-CoA dehydrogenase were up-regulated by 1.20- to 1.42-fold in response to hypoxic treatment. Overall, our proteomic analysis revealed that wax ester synthesis-related enzymes are up-regulated at the protein level post-transcriptionally to promote wax ester production in E. gracilis under low-oxygen conditions.

Introduction Euglena gracilis is a motile unicellular flagellate capable of photosynthesis and has been proposed as a biocatalyst for the production of bioenergy and various valuable compounds [1]. E. gracilis cells have the ability to synthesize a large amount of the storage polysaccharide β-1,3-glucan, known as "paramylon", under aerobic conditions. These cells accumulate up to 50% (on a dry weight basis) of paramylon [2]. However, upon cell culture under low-oxygen conditions, the paramylon is degraded and converted into wax esters consisting of saturated fatty acids and alcohols with C14 chains as the major species. This phenomenon has been designated as wax-ester fermentation due to the generation of ATP without net ATP loss during the production of wax esters [3,4]. Myristic acid (C14:0) has more potential as an alternative bio-jet fuel than medium-length fatty acids produced by other algae, such as palmitic acid (C16:0) and stearic acid (C18:0), because myristic acid has a low freezing point and an adequate Cetane number (66.2) [5].
In E. gracilis mitochondrial wax-ester synthesis, fatty acids are synthesized by a reversal of β-oxidation process [6]. Some of the metabolic enzymes involved in this process have been biochemically analyzed. Acetyl-CoA stemming from pyruvate by a peculiar oxygen-sensitive enzyme, pyruvate:NADP + oxidoreductase (PNO), serves as a primer and a C2 donor [7,8]. Synthesis of odd-numbered fatty acids starts from propionyl-CoA, which is synthesized via the methylmalonyl-CoA pathway. Fatty acid synthesis then proceeds via a condensing step catalyzed by 3-ketoacyl-CoA thiolase (KAT) [9]. Acyl-CoA was believed to be generated from enoyl-CoA by the catalysis of trans-2-enoyl CoA reductase (TER) [10] and then exported from mitochondria; however, the substantial contribution of TER enzyme to fatty acid synthesis remains unclear [11]. Acyl-CoA and the fatty alcohol subsequently generated by a fatty acyl-CoA reductase are then esterified by wax synthase/diacylglycerol acyltransferase to produce wax esters in the microsomes [12,13].
In addition to wax esters, E. gracilis also produces a large amount of succinate, a useful material for the production of bioplastics, under anaerobic conditions. A previous study suggested that E. gracilis produces succinate via the reductive TCA cycle in mitochondria [14]. Thus, the metabolic flux shift in mitochondria in response to low-oxygen conditions is an important process to produce valuable compounds in E. gracilis, such as wax esters and succinate.
Aiming to elucidate the regulatory mechanism of hypoxic wax ester synthesis, we previously performed a comparative transcriptomic analysis of E. gracilis and found that most genes differentially expressed under low-oxygen conditions are related to photosynthesis, nucleotide metabolism, oxidative phosphorylation, and fatty acid metabolism, but not to wax ester synthesis [15]. Thus, the metabolic regulation of wax ester production in response to low-oxygen conditions is likely to occur at the post-transcriptional level. Herein, we focused on analyzing the mitochondrial proteome and performed a shotgun proteomic analysis of mitochondria isolated from E. gracilis cells under aerobic and hypoxic conditions to understand the metabolic regulatory mechanism during wax ester synthesis in E. gracilis.

Strain and culture
E. gracilis strain SM-ZK (non-photosynthetic mutant) was grown in Koren-Hutner (KH) medium [16] under continuous light conditions (50 μmol m −2 s −1 ) at 26˚C with rotary shaking (120 rpm). For aerobic treatment, cells cultured for 6 d were aerated with aseptic air for an additional 24 h. On the other hand, for hypoxic treatment, 6 d-cultures were completely sealed and left to stand still for 24 h without shaking after the replacement of air with nitrogen gas. Notably, cell numbers did not change during additional 24 h aerobic or hypoxic treatments, because the 6 d-cultures have reached to the stationary phase.

Mitochondria isolation from E. gracilis
E. gracilis cells were collected and resuspended in lysis buffer (25 mM HEPES-KOH, pH 7.5, 0.25 M sucrose and 0.5 mM EDTA) and partially disrupted by gas nebulizer-mediated collision using BioNeb disruption system (Glas-Col, Terre Haute, IN). After centrifugation at 1,000 × g for 5 min at 4˚C, the supernatants were centrifuged at 12,000 × g for 10 min at 4˚C. Precipitations were resuspended in the lysis buffer and mitochondrial fractions were purified by stepwise Percoll density (10, 30, and 50%) centrifugation at 9,000 × g for 40 min at 4˚C. The purified mitochondrial fractions were recovered at the interface of 30% and 50% Percoll. After centrifugation at 13,000 × g for 10 min at 4˚C, supernatants were removed to collect the mitochondria pellets ready for use. The extent of non-mitochondrial contaminants such as plastids were estimated very low (1% or less) by calculating with normalized abundances of some representative proteins.

Peptide preparation
E. gracilis mitochondrial fractions were resuspended in extraction buffer (50 mM Tris-HCl, pH 7.5, 10% glycerol, 1 mM 6-aminocapric acid, 1.3 mM benzamidine, 2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 10 mg/mL polyvinylpyrrolidone) and disrupted by sonication. The mitochondria homogenates were centrifuged at 13,000 × g for 15 min at 4˚C, and the supernatant was collected. Protein contents were determined by the Bradford method [17]. Protein extracts containing 150 μg of proteins were purified by a methanol/chloroform precipitation method. Protein extracts were dissolved in 8 M urea, diluted to 800 μL with 50 mM NH 4 HCO 3 , and, then, reduced with 5 mM dithiothreitol for 30 min at 60˚C. These steps were followed by alkylation with 0.74 mM iodoacetamide at room temperature for 30 min. Finally, the reduced and alkylated protein solutions were incubated with 5 μg of trypsin overnight at 37˚C. The reaction was stopped with 0.57% trifluoroacetic acid. The resulting peptide solutions were desalted and concentrated using MonoSpin C18 (GL Sciences, Tokyo, Japan). Peptide solutions were then subjected to fractionation in Agilent 3100 OFFGEL Fractionator (Agilent Technologies, Santa Clara, CA), desalted, and concentrated using GL-Tip SDB (GL Sciences). The final treated samples were dissolved in a solution containing 0.1% formic acid and 5% acetonitrile and used for mass spectrometry.

Data processing, searching, and analysis
The progenesis QI software (Waters Ltd./Nonlinear Dynamics, Newcastle upon Tyne, UK) was used to analyze the raw data files imported from the Synapt G2 mass spectrometer. Each run in the experiment was shown as an ion intensity map, which was representative of the sample MS signal in terms of m/z and retention time. To combine and compare results from different runs, the Progenesis QI aligned the data to compensate for the between-run variation during chromatography. Processed data were analyzed using trypsin as the cleavage protease, one missed cleavage was allowed; fixed modification was set to carbamidomethylation of cysteines, and variable modification was set to oxidation of methionine. Minimum identification criteria included 1 fragment ions per peptide, 2 fragment ions per protein, and a minimum of 1 peptide per protein. Peptide identification was performed using the identity algorithm implemented in the Progenesis QI by searching against E. gracilis protein database constructed by translating cDNA sequences having high BUSCO completeness (91.09%) [15]. False discovery rates (FDR) threshold for peptide and protein identification were set at 4%. To quantify the relative abundance of identified proteins, each run was automatically normalized using an ensemble data set constructed from the aligned runs as recommended by the Progenesis QI software manual. Protein abundance is then calculated from the sum of all unique normalized peptide ion abundances for a specific protein on each run. Only diagnostic peptides were used for quantification at the protein level, and data on abundance changes between aerobic and hypoxic treatments are provided after using ANOVA statistical significance (p<0.05). The scoring algorithm implemented in Progenesis software was used to determine confidence values. To show changes in the abundance of each protein under aerobic and hypoxic conditions, we compared maximum fold change values in peptide counts. Proteins with significantly different abundances (ANOVA: p<0.05) between aerobic and hypoxic conditions were defined as differentially expressed proteins (DEPs). Gene ontologies (GOs) of identified proteins for the categories of biological process, molecular function and cellular compartment were identified as described previously [15].

Protein identification and pathway analysis
In this study, we used SM-ZK strain, an E. gracilis non-photosynthetic mutant, which has been used for many studies so far to understand physiological and biochemical functions on wax ester fermentation including mitochondrial metabolic enzymes [e.g., 3,9]. Previously, we have shown that the E. gracilis cells accumulate significant amount of wax esters after 24 h low-oxygen treatment [11,13], indicating that the metabolic enzymes related to wax-ester production can be expected to fully express in the cells by the treatment. Thus, heterotrophically grown E. gracilis SM-ZK cells were treated with aerobic or hypoxic conditions for 24 h, and intact mitochondria were isolated from each treatment. After digestion of soluble proteins extracted from mitochondria, nanoLC/MS/MS analysis, protein identification, and quantification were performed. Three independent experiments identified total 714 proteins without duplication. Among them, 229 proteins were detected through all three independent experiments and 116 proteins were identified as differentially expressed proteins (DEPs) with significant differences (ANOVA: p<0.05) (S1 Table). Of the 116 DEPs identified, 60 were upregulated under hypoxic conditions. GO enrichment analysis was performed to understand the functional categories of DEPs. The GO terms mitochondrion (GO:0005739), mitochondrial matrix (GO:0005759), and mitochondrial inner membrane (GO:0005743) as cellular component, oxidation-reduction process (GO:0055114) as biological process, and oxidoreductase activity (GO:0016491) as molecular function, showed high DEP rates (Fig 1). These results confirmed that we have successfully obtained proteins localized in mitochondria, and that dynamic expression changes in mitochondria metabolic pathways and redox reactions take place at the protein levels in response to oxygen availability. We then focused on enzymes involved in TCA cycle, glyoxylate cycle, and mitochondrial fatty acid synthesis, as these are particularly responsible for wax ester synthesis, among mitochondrial metabolic pathways.

TCA cycle
The relative protein levels (hypoxic/aerobic) of enzymes involved in the TCA cycle, glyoxylate cycle, and mitochondrial fatty acid synthesis are summarized in Table 1 and their pathways are illustrated in Fig 2. The pyruvate dehydrogenase (PDH) complex plays a critical role in aerobic cell growth as a supplier of acetyl-CoA to the TCA cycle [8]. However, under low-oxygen conditions, acetyl-CoA is generated with a different enzyme PNO [4]. This was supported by the fact that the relative protein levels of PDH E1 α and β subunits decreased under hypoxic conditions (0.81-and 0.77-fold, respectively), whereas those of PNO increased, as described below (Table 1). A previous proteomic approach using two-dimensional gel electrophoresis and MS analysis showed a similar trend [18]. Incidentally, the downregulation of PDH under anoxia has been reported at transcript and protein levels in rice plants [19]. Further observations in other species, especially in protozoa and algae, are needed to evaluate whether it is a common phenomenon or not.
Enzyme 2-oxoglutarate decarboxylase (OGDC), which is a unique enzyme involved in the E. gracilis-specific pathway of the TCA cycle, catalyzes the conversion of 2-oxoglutarate to  succinate semialdehyde and is essential for aerobic growth [8]. Succinate semialdehyde is then converted into succinate by succinate semialdehyde dehydrogenase (SSDH). E. gracilis lacks 2-oxoglutarate dehydrogenase but shows an alternative route composed of OGDC and SSDH in the TCA cycle [8,20]. Succinate is converted to propionyl-CoA by a sequential catalysis of succinyl-CoA synthetase (SCS), methylmalonyl-CoA mutase, and propionyl-CoA carboxylase (PCC) to provide C3 donor for odd-chain fatty acid synthesis. The relative protein levels of the SCS β subunit and the PCC α and β subunits increased by 1.35-to 1.47-fold under hypoxic conditions ( Table 1), suggesting that upregulation of these enzymes contributed to the promotion of odd-chain fatty acid synthesis. The NADP + -dependent malic enzyme (malate dehydrogenase, MDH) plays an important role in electron transport from the cytosol to the mitochondria across the mitochondrial membrane [21]. The relative protein expression level of NADP + -dependent malic enzyme (comp36596) increased 1.71-fold under hypoxic conditions (Table 1). This agreed with previous report of increased NADP + -dependent malic enzyme activity in E. gracilis under anaerobiosis [22]. Therefore, the upregulation of the NADP + -dependent malic enzyme may contribute to the electron supply required for wax ester synthesis.

Glyoxylate cycle
The glyoxylate cycle is important for gluconeogenesis and the conversion of wax esters to paramylon in E. gracilis when the cells are transferred from hypoxic to aerobic conditions [23]. The glyoxylate cycle enzyme (GCE), which is a unique bifunctional enzyme having isocitrate lyase and malate synthase activities, is a key enzyme in the glyoxylate cycle of E. gracilis [24]. Interestingly, the relative protein level of GCE was 1.35-fold higher under hypoxic conditions, despite its role in wax-ester degradation (Table 1). Because the GCE reaction produces succinate from isocitrate without carbon loss and, then, provides propionyl-CoA, it might be possible that this enzyme is involved in the synthesis of odd-chain fatty acids in response to lowoxygen conditions. This could be supported by the fact that E. gracilis accumulated a large amount of succinate under low-oxygen and dark conditions [14,25].

Mitochondrial fatty acid synthesis
PNO catalyzes the oxidative decarboxylation of pyruvate to acetyl-CoA in hypoxic mitochondria. A recent study showed that suppression of PNO expression markedly inhibits wax-ester production under anaerobic conditions in E. gracilis SM-ZK, suggesting that PNO plays a critical role in hypoxic fatty acid synthesis in mitochondria [8]. As shown in Table 1, proteomic analysis showed that the relative protein level of PNO increased 1.68-fold under hypoxic conditions, thus underlining the importance of this enzyme in mitochondrial fatty acid synthesis. In contrast, our previous transcriptomic analysis using wild-type Z strain showed that the PNO transcript level decreased 0.60-fold upon hypoxic treatment for 24h [15]. This discrepancy suggests that PNO expression is regulated post-transcriptionally. For the regulation of gene expression in E. gracilis in response to various environmental conditions, many studies have confirmed that post-transcriptional control is more important than transcriptional control [18,[26][27][28][29][30][31][32][33][34][35][36]. Thus, a similar mechanism might apply for PNO regulation under low-oxygen conditions. KAT is a condensing enzyme of mitochondrial fatty-acid synthesis, and E. gracilis contains at least six predicted KAT isoforms (KAT1-6). It has been considered that KAT1, KAT2, and KAT3 are major 3-ketoacyl-CoA thiolases involved in mitochondrial fatty-acid synthesis in E. gracilis, and Nakazawa et al. (2015) [9] proposed that KAT1 and KAT2 catalyze medium-and long-chain acyl-CoA synthesis, while KAT3 catalyzes short-and medium-chain acyl-CoA synthesis. Proteomic analysis indicated that the relative protein levels of KAT1, KAT2, and KAT3 increased 1.35-, 1.23-, 1.19-fold under hypoxic conditions. In contrast, the relative protein levels of KAT6 decreased 0.78-fold under hypoxic conditions. As the knockdown of KAT6 was found to not affect wax ester production under anaerobic conditions [9], this enzyme may not be involved in the process of wax ester fermentation. Neither KAT4 nor KAT5 isoforms were identified in mitochondrial proteomic analysis, strongly supporting that their localization is not in mitochondria.
Previously, TER was believed to catalyze the final step of fatty acid synthesis from enoyl-CoA to acyl-CoA and enable malonyl-CoA independent fatty acid synthesis in E. gracilis mitochondria [6,10]. However, our recent study showed that TER1 is dispensable for wax ester production, although it is still necessary for chloroplast development [11]. Moreover, it has been suggested that acyl-CoA dehydrogenase isoform 1 (ACD1) catalyzes the reduction of trans-2-enoyl-CoA in E. gracilis cells, although knockdown of ACD1 does not completely inhibit wax ester production [37]. Proteomic analysis showed that the relative protein levels of TER1 and ACD1 increased 1.80-and 1.42-fold under hypoxic conditions, respectively (Table 1). However, the normalized abundance of TER1 was very low and represented only 1.9% that of ACD1 under hypoxic conditions (Table 1). Thus, the protein expression level of TER1 is likely to be very low even under low-oxygen conditions, although it has been proved that an enzyme activity against a short chain CoA (crotonyl-CoA) derived from TER1 is dominant based on results of our previous TER1 gene knockdown experiment [11].

Proteolysis
The proteins comp29057, 35721, 28384, 36301, and 28283, which are annotated as cysteine proteases, including cathepsins, were identified twice in three-independent analyses, and their relative protein levels decreased 0.31-0.38-fold under hypoxic conditions ( Table 2). One of these, comp29057 protein was actually expected to be localized in mitochondria based on Tar-getP prediction ( Table 2). Cathepsins are ubiquitous proteases essential for the survival of Trypanosoma brucei, a protozoan parasite phylogenetically close to E. gracilis [38]. It has been reported that the transcript levels of proteases, including those of cysteine protease, increase during anoxia acclimation in Chlamydomonas reinhardtii, suggesting a correlation between protease expression and low-oxygen treatment [39]. Although the physiological role of cathepsins in hypoxic metabolism is unknown, it is possible that these enzymes regulate the protein levels of the fatty acid synthesis-related enzymes in response to low-oxygen conditions.

Conclusions
Intact mitochondria were isolated from E. gracilis under aerobic and hypoxic conditions and used for shotgun proteomic analysis to understand the regulatory mechanism of wax ester synthesis in this organism. Our study demonstrated that the relative protein levels of enzymes involved in the TCA cycle (MDH), the glyoxylate cycle (GCE), and mitochondrial fatty-acid synthesis (SCS, PCC, PNO, KAT1, KAT2, HADHs, and ACD1) all increased (1.20-to 1.71-fold) under hypoxic conditions, suggesting that hypoxic mitochondrial metabolism for wax ester synthesis is activated by the upregulation of enzymes involved at the protein level, but not at the transcript level [15]. Our mitochondrial proteome analysis provides new insight into the regulatory mechanism underlying the metabolic shift under hypoxic conditions. It might be possible to speculate that an unknown mechanism which facilitates translocation of precursor enzymes into mitochondria in response to low-oxygen condition is functioning. Or, a small RNA molecule might regulate translational levels of some particular proteins as recently reported in Trypanosoma brucei [40]. It should be noted, however, that the rate of increase in mitochondrial proteins was less than double, which is probably not enough to explain the drastic metabolic shift. Thus, the process of wax ester fermentation is likely to be regulated at the post-translational level as well as the translational level. In fact, a recent reverse genetic analysis of starch degradation-related kinase genes (STD1 and STD2) suggested the possibility of the protein phosphorylation-mediated regulation of wax ester production in E. gracilis [41]. A functional proteomic approach, such as a phospho-proteome analysis, will help us to expand our understanding of the regulatory mechanism of wax ester production in E. gracilis in future studies.
Supporting information S1 Table. List of differentially expressed proteins in mitochondria isolated from E. gracilis cells treated with aerobic and hypoxic conditions. (XLSX)