Chlamydomonas carries out fatty acid β-oxidation in ancestral peroxisomes using a bona fide acyl-CoA oxidase Fantao Kong, Yuanxue Liang, Bertrand Légeret, Audrey Beyly-Adriano, Stéphanie Blangy, Richard Haslam, Johnathan Napier, Fred Beisson, G. Peltier, Yonghua Li-Beisson To cite this version: Fantao Kong, Yuanxue Liang, Bertrand Légeret, Audrey Beyly-Adriano, Stéphanie Blangy, et al.. Chlamydomonas carries out fatty acid β-oxidation in ancestral peroxisomes using a bona fide acyl- CoA oxidase. Plant Journal, 2017, 90 (2), pp.358-371. ￿10.1111/tpj.13498￿. ￿hal-01692495￿ HAL Id: hal-01692495 https://hal-amu.archives-ouvertes.fr/hal-01692495 Submitted on 25 Jan 2018 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Chlamydomonas carries out fatty acid β-oxidation in ancestral peroxisomes using a bona fide acyl-CoA oxidase Fantao Kong1, Yuanxue Lian1, Bertrand Légeret1, Audrey Beyly-Adriano1, Stéphanie Blangy1, Richard P. Haslam2, Johnathan A. Napier2, Fred Beisson1, Gilles Peltier1, Yonghua Li-Beisson1 1 Commissariat a l’Energie Atomique et aux Energies Alternatives, CNRS, Aix Marseille Universite, UMR7265, Institut de Biosciences et Biotechnologies Aix Marseille, 13108 Cadarache, France, 2 Department of Biological Chemistry and Crop Protection, Rothamsted Research, Harpenden, UK Summary Peroxisomes are thought to have played a key role in the evolution of metabolic networks of photosynthetic organisms by connecting oxidative and biosynthetic routes operating in different compartments. While the various oxidative pathways operating in the peroxisomes of higher plants are fairly well characterized, the reactions present in the primitive peroxisomes (microbodies) of algae are poorly understood. Screening of a Chlamydomonas insertional mutant library identified a strain strongly impaired in oil remobilization and defective in Cre05.g232002(CrACX2), a gene encoding a member of the acyl-CoA oxidase/dehydrogenase superfamily. The purified recombinant CrACX2 expressed in Escherichia coli catalyzed the oxidation of fatty acyl-CoAs into trans-2-enoyl-CoA and produced H O . This result demonstrated that CrACX2 is a genuine acyl-CoA oxidase, which 2 2 is responsible for the first step of the peroxisomal fatty acid (FA) β-oxidation spiral. A fluorescent protein-tagging study pointed to a peroxisomal location of CrACX2. The importance of peroxisomal FA β-oxidation in algal physiology was shown by the impact of the mutation on FA turnover during day/night cycles. Moreover, under nitrogen depletion the mutant accumulated 20% more oil than the wild type, illustrating the potential of β-oxidation mutants for algal biotechnology. This study provides experimental evidence that a plant-type FA β-oxidation involving H O -producing acyl-CoA oxidation activity has already evolved in 2 2 the microbodies of the unicellular green alga Chlamydomonas reinhardtii. Introduction The β-oxidation of fatty acids (FAs) plays a pivotal role in eukaryotic cells. This catabolic pathway generates acetyl-CoAs via breakdown of FAs acquired from the environment or released upon hydrolysis of membrane structural lipids and storage triacylglycerols (TAGs) by lipolytic enzymes (lipases), and could also be de novo synthesized FAs when the rate of synthesis bypass downstream metabolic needs (Marchesini and Poirier, 2003; Poirier et al., 2006; Graham, 2008). Lipid degradation therefore provides cells with carbon skeletons and energy to drive anabolic processes whilst also ensuring membrane function and cell fitness through the elimination of oxidized, toxic or unusual FAs produced either after being exposed to adverse growth conditions or via transgenic means (Marchesini and Poirier, 2003; Moire et al., 2004; Poirier et al., 2006; Napier, 2007). Lipid catabolic processes and the enzymes involved have been studied intensively in mammalian cells (Eaton, 2002), in germinating oilseeds (Graham, 2008) and in fungal species that can utilize FAs as a carbon source (Daum et al., 2007), and have also recently been explored in plant leaves (Kunz et al., 2009; Troncoso-Ponce et al., 2013; Fan et al., 2014). In mammalian cells β- oxidation of FAs occurs in both mitochondria and peroxisomes (Eaton, 2002), but in plant/fungal cells it occurs exclusively in the peroxisome (Purdue and Lazarow, 2001; Poirier et al., 2006; Graham, 2008). Peroxisomes (also called microbodies) are small single membrane-bound entities, and were originally defined as organelles that carry out oxidative reactions leading to the production of hydrogen peroxide (H O ); thus the occurrence of peroxisomes allows the separation of 2 2 otherwise dangerous oxidative reactions from the remaining cellular metabolism. Since the discovery of peroxisomes in the 1950s (Rhodin, 1954), the metabolic processes occurring within them have been well studied in a few model organisms including human and Saccharomyces cerevisiae (Purdue and Lazarow, 2001), Pichia pastoris(Gasser et al., 2013), Yarrowia lipolytica (Ledesma-Amaro and Nicaud, 2016) and the higher plant Arabidopsis thaliana (Poirier et al., 2006; Graham, 2008; Hu et al., 2012). Plant peroxisomes are the best characterized in the green lineage, and are known to perform a plethora of functions including lipid metabolism, detoxification, nitrogen metabolism, photorespiration and synthesis of some hormones (Kaur et al., 2009; Hu et al., 2012). Most of the metabolic processes in the peroxisome are a part of large metabolic networks spanning several other subcellular organelles, notably plastids and mitochondria. Indeed, physical associations of peroxisomes with mitochondria or plastids have been observed in 7-day-old Arabidopsis seedlings under transmission electron microscopy (Kaur et al., 2009). Essential roles of peroxisomes in coordinating plant metabolism can also be seen through their dynamic nature (increases in both size and number), and that plants without peroxisomes are not viable (Kaur et al., 2009; Hu et al., 2012). Moreover, the peroxisome has been shown to be useful target for re-programming plant metabolism to produce bioplastics and to increase plant biomass productivity (Poirier, 2002; Moire et al., 2004; Poirier et al., 2006; Kessel- Vigelius et al., 2013). Despite the essential functions attributed to peroxisomes in higher plants, and the growing interest in microalgae for green biotechnology, little is known about the metabolic repertoire of algal peroxisomes or about degradation of FAs in microalgae. Earlier literature suggests that, depending on algal species, FA degradation can occur in mitochondria, in peroxisomes or in both organelles (Stabenau et al., 1984, 1989; Winkler et al., 1988). Differences in the compartmentalization of the enzymes of the FA oxidative pathway have been suggested as a consequence of different phylogenetic development. Of particular note, it is reported that a few algal species in the genera Mougeotia, Pyramimonas and Eremosphaeraharbor a peroxisomal acyl-CoA oxidizing (ACX) enzyme which uses O as an electron acceptor and 2 produces H O rather than H O (Winkler et al., 1988; Stabenau et al., 1989). This feature is 2 2 2 often observed in peroxisomes that lack catalase (Stabenau et al., 1989). Thus certain algal species may harbor ACXs with activities different from the H O -producing ones present in 2 2 higher plants. In this study, we used the green unicellular microalga Chlamydomonas reinhardtii as a model to uncover factors involved in lipid hydrolysis. To this end, we employed a forward genetic approach to screen for mutants compromised in oil remobilization. We report here the detailed genetic, biochemical and cell biological characterization of a mutant defective in a member of the acyl-CoA oxidase/dehydrogenase superfamily. We show that this protein encodes an enzyme with a genuine acyl-CoA oxidase activity (producing H O ) that is 2 2 required for breakdown of FAs during lipid remobilization. We provide experimental evidence for the physiological roles of FA β-oxidation in algal diurnal growth. Finally we demonstrate that shutting down core enzymes of the FA β-oxidation spiral increases oil content in green microalgae. Results Isolation of a Chlamydomonas mutant compromised in oil degradation To understand lipid turnover processes in green microalgae, we screened an insertional mutant library for strains perturbed in their capacity to remobilize oil reserves. The screening procedure has previously been described in detail (Cagnon et al., 2013). Briefly, cells were cultivated in acetate-containing medium (TAP) and then boosted to accumulate oils via removal of nitrogen (N; TAP-N72 h). In the wild-type (WT) strains oil is usually used rapidly to power regrowth upon resupply of N (Siaut et al., 2011). Oil content was determined 24 h after addition of N back to the culture (MM24h). Several mutants were found to be impaired in their capacity to remobilize oil upon N resupply, including Nb7D4 (Figure 1a,b). Oil quantification based on densitometry showed that in this mutant 50–70% of oils accumulated at the height of the oil accumulation phase were retained in the mutant 24 h after N resupply when kept in the dark, contrasting with WT cells where 20–30% of oil is retained (Figure 1c). The retention of oils in the lipid droplet (LD) was observed using confocal microscopy (Figure 1d). Flow cytometry, chemical lipid analysis and microscopy therefore pointed to a severe defect in oil degradation in the Nb7D4 mutant. Figure 1. The mutant Nb7D4 (cracx2) is compromised in oil turnover. (a) Oil content screening by flow cytometry based on Nile red fluorescence. (b) Chemical quantification of oil content by thin layer chromatography (TLC). (c) Defect in triacylglycerol (TAG) remobilization. Cells were starved nitrogen for 3 days, and samples were taken for analysis during the recovery phase (+N24 h). Each lane represents one biological replicate from that genotype. Total lipids were extracted from a fixed number of cells, then deposited onto a TLC plate, and TAGs were revealed after staining with a CuSO containing solution (see Experimental procedures for details). This is a 4 representative of at least four or six biological replicates done at different times. Error bars represent standard deviation. (d) Retention of lipid droplet in the mutant cells as revealed by staining with Nile red. Lipid droplets are colored green and chlorophyll autofluorescence is in red. Bars = 10 µm. AU, artificial unit. Nb7D4 is disrupted in Cre05.g232002 encoding an acyl-CoA oxidase (ACX) The insertion site of the paromomycin resistance gene (AphVIII) in Nb7D4 was identified by restriction enzyme site-directed amplification (RESDA)-PCR (Gonzalez- Ballester et al., 2011). Using a combination of specific and degenerate primers, a fragment of 1000 bp was amplified, including 426 bp of the AphVIII gene and 574 bp of flanking sequences. The fragment was sequenced and then Blasted against the genome of C. reinhardtii v5.5 (Phytozome); AphVIII was found inserted in the third intron of the locus Cre05.g232002 (Figure 2a). A RT-PCR analysis using gene-specific primers demonstrated that the insertion resulted in null expression of Cre05.g232002 in the mutant background (Figure 2b). Cre05.g232002 encodes a protein of 76 kDa and is annotated as acyl-CoA oxidase in v5.5 of the C. reinhardtii genome (Merchant et al., 2007). In many organisms, acyl-CoA oxidase (EC1.1.1.3) catalyzes the first committed step in FA β- oxidation, thus exerting major control on this pathway (Klein et al., 2002; Haddouche et al., 2010; Theodoulou and Eastmond, 2012). BlastP analysis of this protein on The Arabidopsis Information Resource (TAIR) website revealed that its amino acid sequence is mostly similar to Arabidopsis AtACX2 (At5g65110) (54% identity, and 68% similarity). Based on this sequence homology, we named the protein encoded by the locus Cre05.g232002 as CrACX2, and the mutant Nb7D4 as cracx2-1. Figure 2. Nb7D4 is defective in an acyl-CoA oxidase. (a) The insertion site of the cassette AphVIII in the mutant Nb7D4 (cracx2-1), and in the second line cracx2-2 (background strain CC-4533), respectively. (b) RT-PCR valid zero expression of CrACX2 in the cracx2-1 and cracx2-2 mutants, respectively. (c) Genetic complementation of cracx2-1. C1: a representative of the complemented strain. (d) Immunoblot detection of the presence of CrACX2 protein in the C1 line. (e) Defect in oil remobilization in the mutant cracx2-2 line. (c,e) Cells were first starved of N for 3 days, then transferred to an N-containing medium to initiate oil remobilization. Cells were harvested for lipid analysis 24 h after the onset of oil degradation. The triacylglycerol (TAG) quantification data by TLC are representative of three biological replicates done at different times. Data are means of two biological replicates with two technical replicates each, and error bars indicate standard deviations. (b) CBLP (Cre06.g278222) codes for a receptor of activated protein kinase C; (d) V5 is the recombinant protein fused at its N-terminus to the epitope V5. Genetic complementation and isolation of a second allele of the mutant cracx2-2 To confirm that the cracx2-1 mutant phenotype was a result of the disruption of the gene Cre05.g232002, complementation of the mutant with a cDNA of the WT locus Cre05.g232002 was conducted. To this, we first cloned the full-length transcript (2025 bp; corresponding to Cre05.g232002.t2.1) into the vector pChlamy4 in frame to the 3′ end of the epitope V5. Despite several trials, we could not clone the cDNA corresponding to Cre05.g232002.t1.1. The promoters and gene structure information are shown in Figure S1(a) in the Supporting Information. After screening about 100 independent zeocin- resistant clones, one clone (cracx2-1;V5-CrACX2) recovered almost its full capacity to remobilize oil (Figure 2c), and the presence of the expressed protein is validated by immunoblot against the anti-V5 antibodies (Figure 2d). Results for some representative clones possessing varying degrees of complementation are shown in Figure S2. Due to the notorious low expression of transgenes in the nuclear genome of C. reinhardtii, only a few complemented lines were obtained here. A second allele (LMJ.SG0182.014586) was identified in the mutant library made by the Jonikas group (Li et al., 2016). It harbors an insertion in the first intron (Figure 2a). The null expression of CrACX2 in this mutant was confirmed by RT-PCR (Figure 2b). Analyses of oil content during the N recovery phase showed impairment in oil remobilization, i.e. the same defect as observed for the cracx2-1 mutant (Figure 2e). We thus named this line cracx2-2. Dynamic changes in TAG content in both mutant alleles (cracx2-1, cracx2-2) and their corresponding WT strains are shown in Figure S3. Taken together, these data firmly establish that the impairment in oil remobilization in the mutants is due to disruption in the normal expression of CrACX2. cracx2 is defective in β-oxidation of FAs To determine if the failure of cracx2 to utilize TAG is caused by impaired β-oxidation, we tested the growth of Chlamydomonas on minimal medium (MM) containing oleic acid as the sole carbon source in the dark. Indeed, Chlamydomonas cells are able to utilize oleic acid supplied in the medium to drive TAG synthesis in the presence of light and acetate (Fan et al., 2011). The utilization of oleic acid as a source of carbon requires a functional β- oxidation cycle in which oleic acid is reduced to acetyl-CoA then to sugars through the glyoxylate and gluconeogenesis pathways (Graham, 2008). We reasoned that mutants defective in the β-oxidation of FAs should display reduced growth when cultivated in the presence of oleic acid as the only carbon source. To test this, cells were grown under strict photoautotrophic conditions, transferred to oleic acid-supplemented MM medium and then cell density was monitored. We first evaluated the optimal oleic acid concentration for such a test in the WT. We observed that when the oleic acid concentration exceeds 0.8 mM cells started to bleach and eventually died (Figure S4), most likely due to the detergent property of FAs. An oleic acid concentration of 0.5 mM is optimal when added to a cell culture of 2 × 106cells ml−1. We observed that 24 h after addition of 0.5 mM oleic acid, WT cells grew at a rate twice that of the mutant cracx2-1 (Figure 3). Growth was arrested eventually, probably due to the exhaustion of oleic acid in the medium, because when additional oleic acid (0.5 mM) was then added to the same culture, regrowth was observed with the WT strain but not the cracx2 mutant (Figure 3). We also observed a slower growth of the mutant compared with the WT in the control experiment (dotted lines in Figure 3), probably due to the fact that under strict carbon starvation, β-oxidation of FAs released from membrane lipids could provide another source of the carbon skeletons required for maintenance of growth. This test allows us to attribute the defect in oil utilization in the mutant to a block in FA β-oxidation. These data show that functional FA β-oxidation is therefore essential in the redistribution of carbon skeletons occurring under strict carbon starvation in C. reinhardtii. Figure 3. Oleic acid feeding test in Chlamydomonas reinhardtii. Cells were grown to mid log phase, then diluted to around 2 million cells ml−1. Either ethanol (as a control) or 0.5 mM oleic acid was added to each culture, and cells were then kept in the dark. The growth kinetics were followed up for a few days after the addition of oleic acid. Potential growth is seen as a result of oleic acid utilization. This is a representative figure for three independent biological replicates. Data are means of three technical replicates, and error bars indicate standard deviations. Turnover of FAs during diurnal growth is compromised in cracx2 mutant cells It has been observed in the marine unicellular stramenopile Nannochloropsis oceanica that total FA content varies during a day/night cycle, i.e. FAs accumulate during the day and degrade at night (Poliner et al., 2015). This phenomenon was also observed here in C. reinhardtii when it was cultivated photoautotrophically under a day/night cycle (12 h/12 h) (Figure 4a). At the end of night period, the mutant retained >80% of total FAs accumulated at the end of day, in contrast to 60% in the WT (Figure 4b). A role of CrACX2 in lipid turnover during the day/night cycle is consistent with high transcription of CrACX2 during the night (Figure 4c, adapted from Zones et al., 2015). Therefore, this study provides experimental evidence that functional FA β-oxidation is involved in lipid homeostasis during nutrient stress but also plays a role in lipid turnover following natural diurnal cycles. Figure 4. The cracx2 mutant is impaired in fatty acid (FA) turnover during day/night cycles. (a) Fluctuation of FA content during the day and night cycle in the wild type and the mutant. (b) Percentage of FA retained during the night. (c) Expression profile of CrACX2 within a day/night (12 h/12 h) cycle (data are based on Zones et al., 2015). (a,b) Data are means of three biological replicates and with two technical replicate each. Error bars represent standard deviations. (c) RPKM stands for ‘reads per kilobase per million mapped reads’. CrACX2 localizes to peroxisomes No subcellular localization could be assigned for the CrACX2 using the PredAlgo prediction tool. This is not surprising, because in the design of the PredAlgo program, peroxisomes/microbodies are not included due to lack of known peroxisome protein sequences in algae (Tardif et al., 2012). In order to determine in which compartment (peroxisomes or mitochondria) FA β-oxidation occurs in Chlamydomonas, we determined the subcellular localization of CrACX2, the first enzyme of the pathway. Sequence examination of the C-terminus or N-terminus of CrACX2 did not reveal obvious sequence similarity to either the peroxisome-targeting sequence (PTS1) [(S/C/A)(K/R/H)(L/M)] or the PTS2 [(R/K)(L/V/I)X (H/Q)(L/A)] consensus sequence (Klein et al., 2002; Hu et al., 2012). This is 5 different from the homolog of the two Arabidopsis proteins – AtACX2 contains a PTS2 sequence, whereas AtACX1 contains a typical PTS1 signal (Eastmond et al., 2000) – but similar to ScPOX1 from S. cerevisiae, where no apparent PTS sequence is present but the protein is known to be imported into the peroxisomes through a novel non-PTS1 pathway (Klein et al., 2002). The knockout mutant of ScPOX1 cannot grow when oleic acid is present as the sole carbon source, thus demonstrating its essential role in the oxidative degradation of FAs in yeast peroxisomes (Dmochowska et al., 1990). To avoid a potential effect of protein mis-targeting or the occurrence of ‘untypical’ targeting sequences at its N- or C- terminus, we made two constructs, one with mCherry protein fused at the N-terminus and the other with mCherry fused at the C-terminus of CrACX2 (Figure S1b,c). CrMDH2 contains a typical PTS2 signal at its N-terminus and has previously been localized to peroxisomes in C. reinhardtii(Hayashi et al., 2015); it was used here as a positive marker for peroxisomes. Either of the two constructs was co-transformed, independently, into the WT strain with the PTS2 sequence from CrMDH2 fused to GFP (Figure S1d). Protein fluorescence analyses using confocal microscopy demonstrated that the co-transformation of mCherry-CrACX2 and PTS2(MDH2)-GFP co-localize to peroxisomes in C. reinhardtii (Figure 5). Despite several attempts, no signals could be detected when mCherry was fused at the C-terminus of CrACX2 (CrACX2-mCherry). One of the reasons for this could be that the fusion of mCherry to the C-terminus of the CrACX2 could potentially interfere with correct protein targeting due to the likely presence of PTS1(-like) internal signal sequences close to the C-terminus (Kaur et al., 2009). Figure 5. CrACX2 localizes to peroxisomes in Chlamydomonas reinhardtii. Peroxisomal localization of mCherry-CrACX2 by confocal microscopy in a representative of transgenic lines expressing mCherry-tagged CrACX2. This line was created by co- transformation with a GFP-tagged PTS2 signal from the protein CrMDH2. Pseudo-colors are used. Chlorophyll autofluorescence is shown as magenta. GFP, green fluorescent protein; PTS, peroxisomal targeting signal; TRITC, tetramethylrhodamine isothiocyanate. CrACX2 catalyzes the conversion of acyl-CoA to trans-2-enoyl-CoA and produces H O 2 2 Previous studies reported the occurrence of acyl-CoA oxidases in certain algal peroxisomes which do not produce H O but instead transfer the energy into water (Stabenau et al., 1989). 2 2 The absence of catalase in the peroxisomes of Chlamydomonas raised the question of whether H O -producing activities were present in its peroxisomes. To understand the molecular 2 2 mechanism of this oxidative step in Chlamydomonas, we characterized the catalytic activity of CrACX2. A codon-optimized version of CrACX2 was cloned into an E. coli expression vector (Figure S5). In parallel, we also expressed AtACX2 in E. coli (Figure 6a). The AtACX2 protein is known to produce H O while oxidizing long chain fatty acyl-CoAs 2 2 (Hooks et al., 1999; Eastmond et al., 2000). Purified recombinant CrACX2 protein catalyzed the conversion of acyl-CoAs to their respective trans-enoyl-CoA products, and produced H O . CrACX2 is more active toward long chain acyl-CoAs (C18:1-, C18:0-, C20:0-, C16:0- 2 2 CoAs) than to medium chain acyl-CoA (C12:0-CoA) (Figure 6b). AtACX2 showed higher activity toward C18:1-CoA followed by C12:0-CoA, and had lower activity with C16:0-, C18:0- and C20:0-CoA (Figure 6b). The preference for mono-unsaturated CoAs over saturated CoA is consistent with a previous finding (Hooks et al., 1999), whereas in our assay the Arabidopsis protein can also utilize C12:0-CoA.