SEARCH

SEARCH BY CITATION

Keywords:

  • brown algae (Phaeophyceae);
  • carbon-concentrating mechanisms;
  • Ectocarpus siliculosus;
  • metabolite profiling;
  • nyctemeral cycle;
  • photorespiration;
  • transcriptomic profiling;
  • γ-aminobutyric acid (GABA)

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Knowledge about primary metabolic processes is essential for the understanding of the physiology and ecology of seaweeds. The Ectocarpus siliculosus genome now facilitates integrative studies of the molecular basis of primary metabolism in this brown alga.
  • Metabolite profiling was performed across two light–dark cycles and under different CO2 and O2 concentrations, together with genome and targeted gene expression analysis.
  • Except for mannitol, E. siliculosus cells contain low levels of polyols, organic acids and carbohydrates. Amino acid profiles were similar to those of C3-type plants, including glycine/serine accumulation under photorespiration-enhancing conditions. γ-Aminobutyric acid was only detected in traces.
  • Changes in the concentrations of glycine and serine, genome annotation and targeted expression analysis together suggest the presence of a classical photorespiratory glycolate pathway in E. siliculosus rather than a malate synthase pathway as in diatoms. Several metabolic and transcriptional features do not clearly fit with the hypothesis of an alanine/aspartate-based inducible C4-like metabolism in E. siliculosus. We propose a model in which the accumulation of alanine could be used to store organic carbon and nitrogen during the light period. We finally discuss a possible link between low γ-aminobutyric acid contents and the absence of glutamate decarboxylase genes in the Ectocarpus genome.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Brown algae (Phaeophyceae), the dominant photosynthetic organisms in the intertidal zone, are members of the Chromoalveolate kingdom. As such, they are phylogenetically distant from red and green algae, as well as land plants (Baldauf, 2008). Basic knowledge about the primary metabolic processes in marine Stramenopiles (which include diatoms and brown algae) is essential for the understanding of the physiology and ecology of these organisms. Recently, insights into the molecular bases of primary metabolism have been gained in temperate diatoms through the genome annotation of the centric Thalassiosira pseudonana (Armbrust et al., 2004) and the pennate Phaeodactylum tricornutum (Bowler et al., 2008). The access to these genomic resources has paved the way for more targeted analysis, as illustrated by the comparative whole genome analysis which enabled the construction of a model for the carbohydrate metabolism in P. tricornutum (Kroth et al., 2008), and the study of the regulation of key enzymes of primary carbon and nitrogen metabolism in T. pseudonana (Granum et al., 2009). Nevertheless, this work was based mainly on genome and targeted transcriptomic analyses, and was not combined with metabolite profiling data.

In brown seaweeds, information on the molecular bases of primary metabolism is rather scattered. A few targeted investigations have been conducted in different species, with emphasis on carbon metabolism, such as the biosynthesis of the carbon storage compound and compatible osmolyte mannitol (Reed et al., 1985), and processes involved in light-dependent or light-independent carbon fixation (see, for example, Kremer & Küppers, 1977; Axelsson et al., 1989a,b; Johnston, 1991; Moulin et al., 1999; Busch & Schmid, 2001; Hillrichs & Schmid, 2001; Schmid & Hillrichs, 2001). Most of this work was based on the measurement of 14C incorporation into primary assimilates, and was completed by determination of the activity of enzymes related to carbon-concentrating metabolism and the photorespiratory pathway. Several studies have also been conducted on the regulation of nitrogen assimilation (Davison & Reed, 1985; Rosell & Srivastava, 1985; Young et al., 2007).

Amino acid metabolism in brown algae has been described in only a few reports, mainly using field samples (Smith & Young, 1955; Nasr et al., 1967; Nagahisa et al., 1995). Akagawa et al. (1972) observed that alanine was one of the first amino acids to be synthesized from radiolabeled carbon in several brown algae. Rosell & Srivastava (1985) showed that, in Macrocystis integrifolia and Nereocystis luetkeana, the predominant amino acids were alanine, glutamate, aspartate and glutamine, together accounting for c. 90% of the total free amino acids. Unfortunately, these authors did not examine changes in amino acid content during the course of the diurnal cycle or their regulation under different physiological conditions.

The small and filamentous brown alga Ectocarpus siliculosus constitutes an excellent model to investigate brown algal primary metabolism. First, it can be cultivated easily under controlled conditions in the laboratory, therefore alleviating possible influences of seasonal and environmental variations. Moreover, E. siliculosus has been selected as a genetic model (Peters et al., 2004; Charrier et al., 2008), and both its complete genome sequence (Cock et al., 2010) and genome-scale microarrays (Dittami et al., 2009) are available. The access to genomic data allowed us to combine manual genome annotation with targeted gene expression analysis and metabolite profiling of amino acids, carbohydrates and sugar alcohols performed during the diurnal cycle and under different physiological conditions. This approach enabled us to extend beyond the bioinformatic and phylogenetic analysis of the proteins involved in central and storage carbohydrate metabolism conducted for this alga by Michel et al. (2010), and to shed light on several interesting molecular features of E. siliculosus.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Experimental set-up

For the diurnal cycle experiment, three replicate cultures of Ectocarpus siliculosus (Dillwyn) Lyngbye (Ectocarpales, Phaeophyceae) unialgal strain 32 (accession CCAP 1310/4, origin San Juan de Marcona, Peru) were cultivated in 10 l plastic flasks in a culture room at 14°C using artificial seawater (ASW) (450 mM Na+, 532 mM Cl, 10 mM K+, 6 mM Ca2+, 46 mM Mg2+, 16 mM SO42−) enriched with Provasoli nutrients (Starr & Zeikus, 1993). Photosynthetically active radiation (PAR) was provided by Philips daylight fluorescence tubes at a photon flux density of 40 μmol m−2 s−1 for 10 h d−1. Cultures were aerated with filtered (0.22 μm) compressed air to avoid CO2 depletion. The diurnal cycle experiment was started c. 15 min after the beginning of the light phase. Every 3 h, c. 200 ml of each of the replicate cultures were concentrated on 40 μm nylon mesh filters (Cell Strainer; Franklin Lakes, BD, NJ, USA), quickly dried with a paper towel, divided into two subsamples and frozen in liquid nitrogen. This whole process took < 1 min. The samples were used to extract total RNA and, after freeze–drying, used for gas chromatography (GC) and ultra performance liquid chromatography (UPLC) analysis.

Carbon starvation and enrichment experiments were performed under the same conditions, with a few exceptions: natural seawater (NSW) adjusted with HCl/NaOH to pH 8.3–8.5 was used instead of ASW, PAR (30 μmol m−2 s−1) was provided 12 h a day and cultures were kept in 250 ml glass flasks. In these experiments, four different CO2 and O2 conditions were tested: a control condition, aerated with compressed air; a carbon depletion condition, in which the culture medium was acidified to pH 2.0 with HCl, stirred for 4 h before restoring to pH 8.4, and then aerated with a mixture of pure N2 and O2 (80% and 20%, respectively); a carbon saturation condition, in which 10 g l−1 NaHCO3 was added to seawater and the pH was again adjusted to 8.3–8.5 (here the air used for bubbling was also enriched with CO2 by aeration through a saturated solution of NaHCO3); and, finally, an O2-enriched condition, in which the control medium was aerated with 50% O2, 50% air starting 12 h before the experiment. This experiment was started at the beginning of the light phase by the addition of c. 2 g FW of E. siliculosus to the prepared culture media. Three biological replicates were made on three subsequent days. Each culture was harvested as described above after 6 and 12 h.

In order to be able to compare E. siliculosus with terrestrial model plants, mature, nonsenescing leaves (leaves 5–8) from 3-wk-old Arabidopsis thaliana plantlets, cultivated as described previously by Lugan et al. (2009), were harvested and immediately frozen in liquid nitrogen. Three independent samples were collected 6 h after the beginning of the light phase and treated exactly as the E. siliculosus samples.

Metabolite profiling

For metabolite profiling, 10 mg of the freeze-dried sample were ground using a ball mill. A methanol–chloroform–water-based extraction was performed according to the following procedure: ground samples were suspended in 400 μl of methanol containing two internal standards: 200 μM 3-aminobutyric acid (BABA) (for amino acid quantification) and 400 μM ribitol (GC analysis). Suspensions were agitated for 15 min at room temperature. Then, 200 μl of chloroform were added, followed by a 5 min agitation step. Finally, 400 μl of water were added, and samples were vortexed vigorously and centrifuged at 13 000 g for 5 min to induce phase separation. The upper phase, which contained nonstructural carbohydrates, polyols, organic acids and amino acids, was transferred to a clean microtube and used for subsequent analysis.

For amino acid profiling, 50 μl of each methanol–water extract were dried under vacuum. Dry residues were resuspended in 50 μl of ultrapure water and 10 μl were used for the derivatization employing the AccQ-Tag Ultra derivatization kit (Waters, Milford, MA, USA). Derivatized amino acids were analyzed using an Acquity UPLC-DAD system (Waters) according to Jubault et al. (2008). BABA was used as internal standard. As this protocol provides only poor separation of glutamate and citrulline, an alternative gradient was also run to verify that no citrulline was present in our samples: initial, 99.9% A; 2 min, 99.9% A, curve 6; 6 min, 98% A, curve 7; 10 min, 96% A, curve 7; 12 min, 80% A, curve 6; 15 min, 40.4% A, curve 6; 16 min, 40.4% A, curve 6; 17 min, 99.9% A, curve 6; and 18 min, 99.9% A, curve 6.

Nonstructural carbohydrates, polyols and organic acids were analyzed by GC-flame ionization detection (FID) according to Adams et al. (1999) and Lugan et al. (2009). Fifty microliters of polar extract were dried under vacuum. The dry residue was dissolved in 50 μl of 20 mg ml−1 methoxyaminehydrochloride in pyridine at 30°C for 90 min under orbital shaking. Then, 50 μl of N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) were added and samples were incubated at 37°C for 30 min, and then at room temperature overnight before injection. One microliter of the mixture was injected into a Trace 2000 GC-FID (Thermo-Fisher Scientific, Waltham, CA, USA) equipped with an AS2000 autosampler (Thermo-Fisher Scientific), a split/splitless injector (split mode set to 1 : 25) at 230°C, a J&W DB5 30 m × 0.32 mm × 0.25 mm column and a flame ionization detector at 250°C. The temperature gradient was as follows: 5 min at 70°C, 5°C min−1 until 220°C, 2°C min−1 until 260°C, 20°C min−1 until 300°C and, finally, 5 min at 300°C. Ribitol was used as internal standard.

1H-NMR fingerprinting was carried out as described by Gagneul et al. (2007) with some modifications. Briefly, 100 mg of freeze-dried material were first delipidated by the addition of 1.5 ml of deuterated chloroform, vigorous shaking and 4 h of incubation at room temperature. After centrifugation (5 min at 15 000 g), chloroform was discarded and the pellet was air-dried before the addition of 1 ml of D2O 99.9% containing 0.5 mM tert-butanol as an internal standard. After 4 h of incubation at room temperature and further centrifugation (5 min at 15 000 g), the supernatant was used as the crude preparation for 1H-NMR recording. This was performed on a Bruker NMR spectrometer (Bruker BioSpin, Rheinstetten, Germany) operating at a 1H frequency of 300 MHz. The processing of the spectra was carried out using MestReNova software (Mestrelab Research, Santiago de Compostela, Spain), and the main peaks were identified according to chemical shifts and spin–spin coupling by comparison with published spectra.

Genome annotation and gene expression analysis

All genes were identified by sequence similarity in version 2 of the E. siliculosus genome (Cock et al., 2010). The subcellular protein localization of genes was predicted using HECTAR, a prediction method developed for heterokonts (Gschlössl et al., 2008), and PTS1 predictor, a method used to predict the C-terminal peroxisomal targeting signal 1 (PTS1; Neuberger et al., 2003). Real-time quantitative polymerase chain reaction (qPCR) analyses were performed for the time points 0, 6, 12, 18 and 24 h of the diurnal cycle experiment and for the 6 h samples of the carbon starvation and enrichment experiments (three replicates each). Primer design, RNA extraction, reverse transcription and qPCR analyses were performed as described previously (Le Bail et al., 2008).

Expression profiles were examined for the following genes (Supporting Information Table S1): two aminotransferases (ALT1 and GGAT, where GGAT has a peroxisomal target peptide PTS1); three genes involved in the photorespiratory pathway [one phosphoglycolate phosphatase (PGP), one glycolate oxidase (GOX1) and one protein of the glycine cleavage system (GCS-L)]; two pyruvate dehydrogenases (acetyl transferase E1 subunit, PDH1 and PDH2); three genes potentially involved in known C4 metabolic pathways [one phosphoenolpyruvate carboxykinase (PEPCK), one phosphoenolpyruvate carboxylase (PEPc) and one cytosolic NADP-malic enzyme (ME1)]; and three carbonic anhydrases (CA2, CA3, CA4). Genes coding for an elongation factor 1 alpha (EF1alpha) and a ubiquitin-conjugating enzyme (UBCE) were used as reference genes for normalization (Le Bail et al., 2008).

Statistical analysis

Statistical analysis was carried out using R 2.9.0 [http://www.r-project.org] together with R Commander (Fox, 2005). All experiments were carried out in three replicates and all data are presented as the mean and standard error. For the diurnal cycle experiment, both metabolites and log2-transformed gene expression data were analyzed by means of a one-way analysis of variance (ANOVA) with time as predictor. Hierarchical clustering of autoscaled (mean-centering and variance scaling) diurnal metabolite profiles was performed using TIGR MeV 4.4.1 (Saeed et al., 2003) and the average linkage function. The resulting clusters were refined manually. In the carbon starvation and enrichment experiments, metabolite data were analyzed using a two-way ANOVA with time and treatment as predictors, and log2-transformed gene expression data were tested using a Student’s t-test. In this experiment, every treatment was compared separately with the control and a Bonferroni correction was applied.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Metabolic profile of E. siliculosus at a glance

An overview of the primary metabolites detected during the light period is presented, as these species have been described only partly so far.

GC analysis showed a surprising paucity of compounds in the chromatograms. As expected, mannitol was the most abundant solute in E. siliculosus, and citrate was the major organic acid found (Table 1). Furthermore, small amounts of glucose and succinate were detected, but many primary metabolites classically found in vascular plants, such as sucrose, fructose, malate, fumarate, trehalose and oligosaccharides of the raffinose family, were not detectable in E. siliculosus, even after concentration of the samples (Table 1). To complete the GC analysis, we confirmed the absence of the internal standard (ribitol) in nonspiked samples (data not shown).

Table 1.   Organic polar solute profiling in Ectocarpus siliculosus (6 h time point) in comparison with 3-wk-old rosette leaves of Arabidopsis thaliana
 EctocarpusArabidopsis
  1. All values are given in μmol g−1 DW and represent the means of three replicates ± standard error. GABA, γ-aminobutyric acid; nd, not detected; ?, not examined.

Total amino acids88.182.4
 Aspartate14.5 ± 2.010.9 ± 0.9
 Asparagine0.7 ± 0.16.1 ± 0.2
 Methionine0.3 ± 0.10.2 ± 0.0
 Isoleucine0.2 ± 0.10.2 ± 0.0
 Lysine0.3 ± 0.10.3 ± 0.0
 Threonine0.7 ± 0.14.5 ± 0.3
 Glutamate22.9 ± 4.623.5 ± 1.1
 Glutamine10.8 ± 2.116.0 ± 1.4
 Arginine0.1 ± 0.00.2 ± 0.0
 GABA< 0.051.3 ± 0.4
 Proline1.3 ± 0.23.2 ± 0.5
 Ornithine0.1 ± 0.00.2 ± 0.1
 α-Alanine29.9 ± 5.15.8 ± 0.3
 Valine0.7 ± 0.40.9 ± 0.0
 Leucine0.3 ± 0.10.3 ± 0.0
 Tyrosine< 0.050.2 ± 0.0
 Phenylalanine0.2 ± 0.10.3 ± 0.1
 Tryptophan0.0 ± 0.00.1 ± 0.0
 Serine3.1 ± 1.06.8 ± 0.6
 Glycine1.7 ± 0.71.4 ± 0.2
 Glycine : serine0.5 ± 0.00.2
Organic acids18.4178.2
 Fumaratend86.8 ± 31.9
 Malatend34.7 ± 5.4
 Citrate16.7 ± 2.156.7 ± 7.1
 Succinate1.7 ± 0.2?
Sugars and polyols334.750.6
 Fructose< 0.0511.4 ± 0.7
 Glucose3.3 ± 0.721.7 ± 0.7
 Myo-inositol< 0.055.5 ± 0.7
 Sucrose< 0.0512.0 ± 0.1
 Mannitol331.4 ± 68.7nd
Total organic solutes441.1311.3

Glutamate and alanine were the major free amino acids in E. siliculosus samples, with mean contents of 22.9 ± 4.6 and 29.9 ± 5.1 μmol g−1 DW, respectively (Table 1). Aspartate and glutamine were the next most abundant, with mean concentrations of 13.9 ± 2.2 and 10.7 ± 1.7 μmol g−1 DW, respectively. Except for glycine and serine, which were found in similar concentrations in E. siliculosus as in terrestrial plants, other proteinogenic amino acids occurred at comparatively lower levels in their free form. Nonproteinogenic amino acids, such as γ-aminobutyric acid (GABA) and ornithine, were present only in traces (< 0.05 μmol g−1 DW; Table 1). Furthermore, 34 peaks potentially corresponding to as yet unidentified biogenic amines were detected through the AccQ-Tag derivation in Ectocarpus samples (data not shown).

A 1H-NMR fingerprint of an aqueous extract validated that our chromatography-based analytical procedure allowed the detection of the main occurring polar metabolites (Fig. S1), because most could be characterized through their chemical shifts. Characteristic chemical shifts at 4.30 and 5.40 – corresponding to protons of malate and sucrose, respectively – were not detected, which is consistent with the GC analysis.

Nyctemeral fluctuations of metabolites

The mannitol content was subject to strong variations during the course of the light–dark cycle, ranging from 200 to 350 μmol g−1 DW (3.6–6.4% DW), with a clear increase during light periods, and a subsequent decrease in the dark (Fig. 1a). The overall intracellular amino acid concentration remained stable during the photoperiodic cycle (mean of 86.1 ± 10.8 μmol g−1 DW; Fig. 1a), but this obscures the large fluctuations in the contents of major amino acids during the photoperiod (Fig. 1b). The light phase was characterized by a decrease in glutamate and an increase in alanine content. In the first part of the dark period, a strong decrease in alanine and a parallel accumulation of glutamate were observed, followed by a slight decrease in glutamate and an increase in alanine in the second half of the dark phase. Detailed nyctemeral profiles of all the examined metabolites are presented in Table S2.

image

Figure 1.  Changes in mannitol (closed circle) and total free amino acid (grey square) concentrations (a) as well as the relative concentrations of the predominant amino acids (Asp, aspartate; Glu, glutamate; Gln, glutamine; Ala, α-alanine) (b) during the course of two light–dark cycles. Grey areas indicate the dark phases. The graphs show the means of three replicates ± standard errors.

Download figure to PowerPoint

Clustering of autoscaled (mean-centering and variance scaling) data allowed the comparison of the fluctuations of compounds with highly contrasting basal levels. This analysis revealed three major clusters of metabolites, the first aggregating around alanine, the second around mannitol and the third around glutamate (Fig. 2). In addition, there were several amino acids with diurnal changes that were delayed with respect to these three main clusters. The first cluster around alanine (CL 1, Fig. 2) also comprised glutamine, threonine, methionine, glycine, serine and citrate. These metabolites were present at higher concentrations during the light than during the dark period. Mannitol, valine, lysine, leucine, arginine, asparagine and aspartate (CL 2) followed a similar pattern, but with a delay. These compounds were found at highest concentrations at the end of the light period and the beginning of the dark period. The third main cluster (CL 3) contained glutamate and isoleucine, both of which were predominant during the dark phase. For phenylalanine, tryptophan, succinate, ornithine, glucose and tyrosine, no significant changes in concentration could be detected over the course of the nyctemeral cycle.

image

Figure 2.  Clustering of metabolites according to their diurnal oscillations. Only metabolites exhibiting significant changes (P < 0.05) are shown. All values represent the means of three replicates. Data were centered and normalized (mean = 0, SD = 1) to facilitate the comparison of metabolites with different basal levels and different degrees of variation (original means and SDs are given in parentheses to the right of the compound name). The statistical significance of the metabolite changes is indicated for each metabolite as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001. The 00:00–09:00 h time points were taken during the light period, and the 12:00–21:00 h time points during the dark period.

Download figure to PowerPoint

Changes in amino acid content under altered experimental conditions

Alanine, glycine and serine, all part of cluster CL 1 (Fig. 2), were among the most variable amino acids in our dataset (Fig. 3). These observations, together with the current uncertainty regarding the occurrence of an organic carbon concentration mechanism in E. siliculosus, prompted us to design additional experiments to test two hypotheses: (1) glycine and serine accumulation at the end of the day period potentially reflects an enhancement of the photorespiratory process caused by inorganic carbon depletion; (2) the accumulation of alanine during the light period may be associated with carbon assimilation similar to plant C4 metabolism, which is regulated by the availability of inorganic carbon.

image

Figure 3.  Changes in the concentrations of glycine (closed circle) and serine (grey square) (a) as well as in the ratio of glycine to serine (mol mol−1) (b) during the course of two light–dark cycles. Grey areas indicate the dark phases. The graph shows the means of three replicates ± standard errors.

Download figure to PowerPoint

In order to potentially affect photorespiration activity, the availability of CO2 and O2 was modified, and amino acid contents were determined after 6 and 12 h of treatment. No variation in the amount of total free amino acids was observed in this experiment. O2 enrichment, a treatment that should enhance photorespiration, triggered significant accumulation of glycine (P = 0.022), serine (P = 0.023) and an increase in the glycine to serine ratio (P = 0.027; Fig. 4). In contrast, CO2 enrichment, which was expected to lower photorespiration, clearly led to a decrease in all three biochemical markers of this process (serine, P < 0.001; glycine, P = 0.004; glycine : serine, P = 0.006). Under CO2 starvation, a slight increase in all three markers was observed, although it was not statistically significant (P > 0.130). Finally, we monitored a decrease in the NH4+ concentration in the CO2-enriched condition (Fig. 4; P < 0.001). Considering that NH4+ may be viewed as a coproduct of photorespiration, these results support our first hypothesis that glycine and serine accumulation reflects a plant-like photorespiratory process.

image

Figure 4.  Changes in the relative metabolite concentrations in response to different growth conditions. All values are displayed as ratios of treatment (closed circle, high O2; closed square, low CO2; closed triangle, high CO2) and control (mean of three replicates ± standard errors).

Download figure to PowerPoint

Following our second hypothesis, a condition that increases photorespiration could result in an increase in alanine and/or aspartate. However, we observed a decrease in alanine in response to O2 enrichment (16%, P = 0.008), and a decrease in aspartate (45%, P = 0.003) in response to CO2 starvation (see Table S3 for a complete list of metabolite concentrations in this experiment). Therefore, these observations do not provide support for the hypothesis that a putative alanine/aspartate organic carbon-concentrating mechanism (CCM) would be regulated by CO2 availability in Ectocarpus.

Analysis of selected genes related to carbon metabolism

Genes related to photorespiration  We manually annotated genes potentially related to photorespiration (Table S4, Fig. 5). Several genes involved in the glycolate photorespiratory pathway were identified in the genome, including a sequence encoding a putative glycerate kinase. However, we did not find any genes coding for a serine-glyoxylate aminotransferase, a glycerate dehydrogenase or a hydroxypyruvate reductase. For most of the genes identified, subcellular targeting predictions are consistent with the classical photorespiratory paradigm in terrestrial plants, except for the candidate glycerate kinase. A gene coding for a malate synthase, an enzyme potentially involved in a diatom-like alternative glyoxylate cycle-based photorespiratory pathway, is also present, but we did not detect a peroxisomal target peptide.

image

Figure 5.  Putative glycolate-based photorespiratory pathway in Ectocarpus siliculosus (after Kroth et al., 2008). Enzymatic reactions are symbolized by arrows; the catalyzing enzyme is shown in italics. Grey arrows and crossed out abbreviated enzyme names indicate that no corresponding gene was identified in the E. siliculosus genome; a question mark after the abbreviation indicates unexpected subcellular targeting. GCS, glycine cleavage system; GGAT, glutamate/alanine:glyoxylate aminotransferase; GK, glycerate kinase; GOX, glycolate oxidase; HPR, hydropyruvate reductase; MLS, malate synthase; 3-PGA, 3-phospho-d-glycerate; PGP, 2-phosphoglycolate-phosphatase; RUBISCO, ribulose-1,5-bisphosphate carboxylase oxygenase; SGAT, serine glyoxylate aminotransferase; SHMT, serine hydroxymethyltransferase.

Download figure to PowerPoint

Expression profiles were investigated under the previously described experimental conditions for four genes directly involved in this photorespiratory pathway: phosphoglycolate phosphatase (PGP), glycolate oxidase (GOX1), peroxisomal glutamate/alanine:glyoxylate aminotransferase (GGAT) and glycine cleavage system subunit L (GCS-L) (Figs 6, S2). A second glyoxylate oxidase present in the genome (GOX2) was not examined in this study because of problems designing specific primers. However, this gene model is supported by only one expressed sequence tag (EST) vs 25 for GOX1. GGAT was up-regulated during CO2 starvation (3.4-fold, P = 0.009) and in the O2-enriched condition (2.7-fold, P = 0.055), and down-regulated in the CO2-enriched condition (2.5-fold, P = 0.035; Fig. 6a), corroborating the potential role of this predicted peroxisomal aminotransferase in the glycolate cycle-based photorespiration. The other examined genes exhibited only minor (< 1.5-fold) or nonsignificant changes in expression (Fig. S2), whatever the conditions tested.

image

Figure 6.  Gene expression profiles of selected genes during the light–dark cycle and in different growth conditions (means of three replicates ± standard error). The profiles of each gene during the diurnal cycle show the copy number, and changes in expression (log2 ratios with the control condition) in response to the different growth conditions are shown in the adjacent graph to the right. In the latter graph, positive values indicate that a gene is up-regulated, and negative values indicate that it is down-regulated; P values are indicated as follows: ‘P < 0.1; *, P < 0.05; **, P < 0.01; ***, P < 0.001. The shaded area indicates the dark phase (from 09:45 to 23:45 h). See Table 1 for a definition of the examined genes.

Download figure to PowerPoint

Genes related to carbon concentration mechanisms  Considering that carbon concentration mechanisms are of major importance in improving photosynthetic efficiency and lowering photorespiration in E. siliculosus, we examined the regulation exerted by CO2 availability on the genes possibly involved in inorganic CCMs. Among the carbonic anhydrases (CA1CA5) with EST support in Ectocarpus (Table S4), the predicted cytosolic CA4 was the most highly expressed carbonic anhydrase-related gene, and was also the only one to be differentially expressed in our experiments. This gene was up-regulated during the dark phase (P = 0.024) and strongly down-regulated in CO2-enriched conditions (3.7-fold, P < 0.001, Fig. 6b). Consistent with the presence of carbonic anhydrases, we found two genes coding for putative bicarbonate transporters containing a eukaryotic bicarbonate transporter domain (IPR003020/IPR011531) according to InterProScan. Both proteins had closest homologs in P. tricornutum (51 and 52% identity) and shared 35% identity with Drosophila melanogaster anion transporters.

Additional genes potentially involved in organic CCMs were also studied, including PEPc, PEPCK and malic enzymes (Table S4). Among these genes, ME1 exhibited a high basal level of transcription. Only PEPCK (predicted to be mitochondrial) showed significant variations (≥ 2-fold, P ≤ 0.05; Figs 6c, S2) during the day/night cycle, but not in response to modified CO2/O2 conditions. These observations did not provide an explanation for the observed fluctuations of aspartate and alanine.

Genes involved in pyruvate/alanine metabolism  The alanine aminotransferase ALT1, predicted to be targeted to the cytosol, is likely to catalyze the transfer of an amino group from glutamate to pyruvate, thereby producing alanine and α-ketoglutarate. It was therefore hypothesized to be involved in the synchronous opposite fluctuations observed between alanine and glutamate during the day/night cycle. This gene was constitutively expressed (Fig. S2), which neither supports nor excludes the control of alanine and glutamate fluctuations by ALT activity. We then monitored the changes in expression of two genes coding for pyruvate dehydrogenases, enzymes that compete with ALT1 for the use of pyruvate as a substrate, to furnish acetyl-CoA for the tricarboxylic acid (TCA) cycle. The TCA cycle represents a metabolic crossroads between carbon and nitrogen metabolism, providing carbon skeletons as precursors for alanine, aspartate and glutamate biosynthesis. PDH1 and PDH2 exhibited similar expression profiles, although changes were significant only in PDH1, the expression of which decreased during the light phase and increased during the dark phase – a feature that could suggest an increase in TCA cycle activity during the dark phase. Both PDH1 and PDH2 exhibited only minor changes in response to altered CO2/O2 conditions (≤ 1.5-fold; Figs 6d, S2).

Genes involved in the GABA shunt  Finally, the detection of traces of GABA in the free amino acid pool of E. siliculosus led us to consider its metabolism in this species and whether the GABA shunt, as an anaplerotic pathway, operates in brown algae. We searched for genes classically involved in the GABA shunt, and found that E. siliculosus, as well as P. tricornutum and T. pseudonana, lacked genes coding for a glutamate decarboxylase and a GABA-transaminase, key enzymes of GABA biosynthesis and GABA catabolism, respectively. Interestingly, genes coding for these enzymes were identified in the four Phytophthora genomes (data not shown). Thus, the low level of GABA detected in our samples could have been synthesized from putrescine by the activity of diamine oxidases and an aminobutyraldehyde dehydrogenase (see Table S4 for candidate genes with these functions).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

General considerations from metabolic profiles and day/night fluctuations

This study presents the most extensive profiling approach of primary metabolites carried out in brown algae so far, and reveals the prevalence of only a small number of organic acids, polyols and soluble nonstructural carbohydrates in E. siliculosus samples.

The predominant GC-detected soluble compounds were mannitol and citrate, representing 90% and 6% of the total polyols and organic acids measured, respectively. These results are in accordance with early biochemical reports on brown algae (Jones, 1956; Yamaguchi et al., 1966). The paucity of malate, confirmed by both chromatographic analysis of concentrated extracts (data not shown) and NMR analysis, is also significant, not only because it is a strong argument against the presence of crassulacean acid metabolism (CAM) in E. siliculosus, but also because malate is a central compound widely involved in several processes, including intracellular shuttles for reducing power in plants.

The most abundant amino acids in our biological samples were alanine, glutamate, aspartate and glutamine, together accounting for c. 90% of the total free amino acids. These results are in accordance with previous data obtained from the brown algae Macrocystis integrifolia and Nereocystis luetkeana (Rosell & Srivastava, 1985). This amino acid pattern is quite similar to profiles observed in higher plants, as illustrated in Table 1 for A. thaliana. However, shifted changes of mannitol, alanine and glutamate during the day/night cycle are of special interest for at least three reasons. Firstly, glutamate has been described as a metabolite with one of the lowest day/night variations in Arabidopsis (Gibon et al., 2006), as well as in other terrestrial plants (Forde & Lea, 2007). Therefore, it appears that mechanisms suspected to underpin glutamate homeostasis in plants are different (or absent) in E. siliculosus. Secondly, the opposite profiles of alanine and glutamate are likely to be directly coupled to the photoperiod, and thus to the photoassimilation of carbon and/or nitrogen. Thirdly, most of the fluctuations affecting minor metabolites can be linked to variations in these three major metabolites, suggesting their central role in E. siliculosus physiology.

Gene analysis and metabolite profiling support a glycolate-based photorespiration pathway in E. siliculosus

Photorespiration is a process that occurs in plants, and is generally intensified under high light, high temperature or CO2 deficiency. Under these stress conditions, an increased proportion of ribulose-1,5-bisphosphate (RuBP) may be oxygenated rather than carboxylated. Although, in terrestrial plants, RuBP is regenerated by the photorespiratory glycolate pathway (C2 cycle; Foyer et al., 2009), the diatom genomes of P. tricornutum and T. pseudonana do not code for a glycerate kinase, an essential enzyme involved in this process. These organisms are therefore thought to primarily use a peroxisomal malate synthase-based photorespiratory glyoxylate pathway or to excrete glycolate (Winkler & Stabeneau, 1995; Kroth et al., 2008). In the green lineage, the accumulation of glycine and serine, and an increase in the ratio of glycine to serine, are considered to be reliable markers of photorespiration (Foyer et al., 2003; Igarashi et al., 2006). In contrast, in diatoms, activation of the photorespiratory glyoxylate pathway would not affect serine/glycine metabolism, as demonstrated by Burris (1977).

In Ectocarpus, the observed changes in the glycine to serine ratio were statistically significant, despite high variability between replicates, which could have been caused, for example, by the killing procedure. This suggests that classical glycolate cycle-based photorespiration could be induced during the light period and under high-O2 and low-CO2 conditions, as well as a reduction of photorespiration in response to CO2 enrichment. Similar observations have been reported in different Fucales and Laminariales using 14C-labeling techniques (Kremer, 1980), but changes in the pools of glycine and serine after incubation under conditions altering photorespiration have never been described in Ectocarpus. Classical glycolate-based photorespiration would also lead to the production of NH4+ through activity of the glycine cleavage system, and could therefore (partly) explain the strong decrease in the content of this cation under high-CO2 conditions. However, we did not observe any clear diurnal NH4+ fluctuations, and the availability of carbon skeletons for NH4+ assimilation should also be taken into account.

Additional support for a potential glycolate-based photorespiratory pathway can be derived from genome annotation (Fig. 5) and gene expression profiling. Among the genes tested, changes in the transcript levels for a glutamate:glyoxylate aminotransferase (GGAT), predicted to be targeted to the peroxisome, were highly correlated with the expected level of photorespiration in CO2/O2-altered growth conditions. Moreover, the identification of a putative glycerate kinase gene in E. siliculosus is of great interest, as this gene was not identified in the genome of either P. tricornutum or T. pseudonana. Two genes of the glycolate-based photorespiratory cycle, a serine glyoxylate aminotransferase and a hydroxypyruvate reductase, were not identified in the Ectocarpus genome; however, the latter activity was demonstrated experimentally in the brown alga Egregia menziesii (Gross, 1990), and all enzymatic activities necessary to complete the glycolate cycle were detected previously in the brown alga Spatoglossum pacificum (Iwamoto & Ikawa, 1997). Thus, one possible explanation could be that these genes were not indentified in E. siliculosus merely as a result of the lack of well-characterized heterokont sequences. Together, our findings provide molecular and biochemical support for the presence of a photorespiratory glycolate pathway, as suggested by previous biochemical studies on brown algae (Burris, 1977; Gross, 1990).

Potential role of CA4 in CCMs

A major role of intracellular carbonic anhydrases in inorganic carbon concentration in E. siliculosus has been demonstrated previously using pharmacological approaches (Schmid & Dring, 1996; Schmid, 1998). The strong regulation of CA4, the carbonic anhydrase with the highest EST support, in response to changes in carbon availability suggests that this enzyme could be a key component of an inorganic CCM operating in E. siliculosus. In agreement with this, we report the presence of two putative bicarbonate transporters that could be of interest in future studies on carbon assimilation. In contrast, despite our efforts, no evidence could be found linking diurnal changes in aspartate and alanine concentrations to a C4-like mechanism in E. siliculosus.

Alanine as a potential storage compound for pyruvate-derived carbon skeletons

The observed antagonism between alanine and glutamate (Fig. 1b) remains one of the most striking phenomena in our dataset. Alanine can be synthesized by the reversible transfer of an amino group from glutamate to pyruvate, also producing one molecule of α-ketoglutarate. The latter can then be recycled for further glutamate/glutamine synthesis via the glutamine synthetase/glutamine:2-oxoglutarate aminotransferase (GS/GOGAT) system (Fig. 7). Although the mitochondrial alanine transaminase thought to catalyze the synthesis of alanine (ALT1) was not regulated at the transcriptomic level under our experimental conditions, a direct conversion of glutamate to alanine would provide a plausible explanation for the observed inverse correlation between the concentrations of these two compounds. One hypothesis could be that, under laboratory conditions, when nitrate is not limiting, this accumulation constitutes a way of temporarily storing carbon skeletons produced from photosynthesis, a process that might be complementary to the role of mannitol as storage compound. At the onset of the dark phase, the supply of pyruvate from photosynthesis will cease. During the early dark phase, respiration is not likely to decrease, as supported by the increase in PDH1 expression (Fig. 6d). As the alanine transaminase works in both directions, this would shift the alanine/glutamate balance towards glutamate, and therefore liberate pyruvate, explaining both the rapid decrease in alanine concentration at the transition from the light to the dark phase and the parallel increase in glutamate. Finally, after a latency period of a few hours, the rapid changes observed at the beginning of the dark phase come to a halt. This could be caused by an activation of another pyruvate source, for example glycolysis. Such fluxes between pyruvate and alanine have been reported previously in terrestrial plants, for instance in maize phloem (Valle & Heldt, 1991), and would, moreover, be consistent with the work reported by Akagawa et al. (1972), who illustrated that alanine was one of the first amino acids to be synthesized from radiolabeled carbon in several brown algae. Although this hypothesis provides a possible explanation for the observed relationship between alanine and glutamate, further experimental validation, for example using pulse-chase experiments, will be required to explore links between this alanine/glutamate antagonism and the above-mentioned questions concerning aspartate and β-carboxylation in Ectocarpus.

image

Figure 7.  Schematic representation of reactions possibly underlying the changes in alanine and glutamate concentrations. Arrows pointing up indicate an increase or activation, and arrows pointing down indicate a decrease. Boldface arrows indicate a strong flow of metabolites in the direction of the arrow. ALT, alanine aminotransferase; GS/GOGAT, glutamine synthetase/glutamine:2-oxoglutarate aminotransferase; PDH, pyruvate dehydrogenase; PEP, phosphoenolpyruvate; PS, photosynthesis; TCA, tricarboxylic acid cycle; α-KG, α-ketoglutarate.

Download figure to PowerPoint

Absence of the GABA shunt in E. siliculosus

Compared with previous studies in brown algae, we used more sensitive methods, facilitating the separation and detection of several compounds present only at very low concentrations in our samples. This allowed the detection of traces of GABA, which was not detected in a previous study of the brown alga Himanthalia elongata (Jones, 1956). The quasi-absence of this compound is an unusual feature, as this molecule is generally considered to be ubiquitous (reviewed in Clark et al., 2009). The GABA shunt is believed to be the main pathway of GABA synthesis in both plants and animals (reviewed in Petroff, 2002). In plants, this pathway is also thought to furnish intermediates for the TCA cycle (Allan & Shelp, 2006). The quasi-absence of GABA in E. siliculosus in our culture conditions could be related to the apparent absence of a gene encoding a glutamate decarboxylase in its genome (Table S4).

Conclusion

By combining gene analysis and extensive metabolite profiling, this reports highlights several new features related to primary metabolism in the model brown alga E. siliculosus. Firstly, we give a global overview of primary metabolite concentrations and fluctuations during the nyctemeral cycle. These results show that, with the exception of mannitol, E. siliculosus tissue samples contain low levels of polyols, organic acids and sugars. Moreover, a surprising inverse relationship between glutamate and alanine, corresponding to light–dark transitions and to the regulation of downstream amino acids, was observed. Secondly, several observations, such as the absence of malate, the increase in aspartate during the light phase and gene expression patterns, do not match those classically found in C4 or CAM plants, whereas the high degree of regulation of the carbonic anhydrase CA4 in response to changes in CO2 could indicate the implication of this gene in inorganic CCM. Thirdly, changes in the concentrations of glycine and serine, in combination with genome annotation and targeted gene expression profiling, suggest the presence of a classical photorespiratory glycolate pathway in E. siliculosus, rather than a malate synthase pathway as described in diatoms. Finally, E. siliculosus samples contained very low levels of GABA, a quality most probably related to the apparent absence of a glutamate decarboxylase. The hypotheses raised in this report pave the way for further investigations, which should include experimental determination of the subcellular localization for a number of enzymes, as well as pulse-chase experiments, in order to further elucidate the pathways involved in the primary metabolism in Ectocarpus.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We would like to thank Constance de Villardi for practical help in setting up the carbon starvation and enrichment experiments, Jean-Paul Guégan (ENSCR, Rennes, France) for technical assistance with the NMR experiments, and François Larher and Hugues Renault for helpful discussions. Rennes Métropole is acknowledged for its financial support for the acquisition of the UPLC equipment. S.D. received funding from the European community’s Sixth Framework Programme (ESTeam contract no MESTCT 2005-020737).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
NPH_3400_sm_fS1.pdf64KSupporting info item
NPH_3400_sm_fS2.pdf403KSupporting info item
NPH_3400_sm_tS1.xls23KSupporting info item
NPH_3400_sm_tS2.xls44KSupporting info item
NPH_3400_sm_tS3.xls84KSupporting info item
NPH_3400_sm_tS4.xls29KSupporting info item