* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download New insight into the photoheterotrophic growth of the
Peptide synthesis wikipedia , lookup
Lactate dehydrogenase wikipedia , lookup
Light-dependent reactions wikipedia , lookup
Protein–protein interaction wikipedia , lookup
Magnesium transporter wikipedia , lookup
Mitogen-activated protein kinase wikipedia , lookup
Western blot wikipedia , lookup
NADH:ubiquinone oxidoreductase (H+-translocating) wikipedia , lookup
Electron transport chain wikipedia , lookup
Two-hybrid screening wikipedia , lookup
Metalloprotein wikipedia , lookup
Evolution of metal ions in biological systems wikipedia , lookup
Ribosomally synthesized and post-translationally modified peptides wikipedia , lookup
Fatty acid metabolism wikipedia , lookup
Glyceroneogenesis wikipedia , lookup
Photosynthesis wikipedia , lookup
Fatty acid synthesis wikipedia , lookup
Biochemistry wikipedia , lookup
Proteolysis wikipedia , lookup
Oxidative phosphorylation wikipedia , lookup
Paracrine signalling wikipedia , lookup
Biochemical cascade wikipedia , lookup
Biosynthesis wikipedia , lookup
Microbial metabolism wikipedia , lookup
Microbiology (2015), 161, 1061–1072 DOI 10.1099/mic.0.000067 New insight into the photoheterotrophic growth of the isocytrate lyase-lacking purple bacterium Rhodospirillum rubrum on acetate B. Leroy,1 Q. De Meur,1 C. Moulin,1 G. Wegria2 and R. Wattiez1 Correspondence 1 B. Leroy [email protected] 2 Received 24 September 2014 Accepted 27 February 2015 Laboratory of Proteomics and Microbiology, Research Institute for Biosciences, University of Mons, Place du Parc 20, 7000 Mons, Belgium Biotech Materia Nova, Parc Initialis, Avenue Copernic 1, 7000 Mons, Belgium Purple non-sulfur bacteria are well known for their metabolic versatility. One of these bacteria, Rhodospirillum rubrum S1H, has been selected by the European Space Agency to ensure the photoheterotrophic assimilation of volatile fatty acids in its regenerative life support system, MELiSSA. Here, we combined proteomic analysis with bacterial growth analysis and enzymatic activity assays in order to better understand acetate photoassimilation. In this isocitrate lyaselacking organism, the assimilation of two-carbon compounds cannot occur through the glyoxylate shunt, and the citramalate cycle has been proposed to fill this role, while, in Rhodobacter sphaeroides, the ethylmalonyl-CoA pathway is used for acetate assimilation. Using proteomic analysis, we were able to identify and quantify more than 1700 unique proteins, representing almost one-half of the theoretical proteome of the strain. Our data reveal that a pyruvate : ferredoxin oxidoreductase (NifJ) could be used for the direct assimilation of acetyl-CoA through pyruvate, potentially representing a new redox-balancing reaction. We additionally propose that the ethylmalonyl-CoA pathway could also be involved in acetate assimilation by the examined strain, since specific enzymes of this pathway were all upregulated and activity of crotonyl-CoA reductase/carboxylase was increased in acetate conditions. Surprisingly, we also observed marked upregulation of glutaryl-CoA dehydrogenase, which could be a component of a new pathway for acetate photoassimilation. Finally, our data suggest that citramalate could be an intermediate of the branched-chain amino acid biosynthesis pathway, which is activated during acetate assimilation, rather than a metabolite of the so-called citramalate cycle. INTRODUCTION Rhodospirillum rubrum belongs to the purple non-sulfur bacteria, a highly metabolically versatile group of bacteria. The S1H strain of this bacterium has been selected by the European Space Agency as a major component of its regenerative life support system MELiSSA (Micro-Ecological Life Support System Alternative) (Hendrickx et al., 2006). In this artificial ecosystem, bioreactors inhabited by different organisms will recycle crew-produced waste to generate oxygen, water and food. The first compartment, which treats crude waste through thermophilic fermentation, will produce large amounts of volatile fatty acids, the most abundant Abbreviations: EMC, ethylmalonyl-CoA; Icl2, isocytrate lyase-lacking; ILV, isoleucine, leucine and valine; IlvC, ketol-acid reductoisomerase; IlvD, dihydroxy-acid dehydratase; MELiSSA, Micro-Ecological Life Support System Alternative; NifJ, pyruvate : ferredoxin oxidoreductase; PHB, poly(3-hydroxybutyrate); XIC, extracted ion chromatogram. Two supplementary tables are available with the online Supplementary Material. 000067 G 2015 The Authors of which will be acetate. Rs. rubrum, which is included in the second compartment of the MELiSSA loop, will be used to photoassimilate these volatile fatty acids. The assimilation of carbon sources that enter central carbon metabolism through acetyl-CoA requires an alternative cycle to replenish the TCA cycle intermediates used for the synthesis of all cell components (Dunn et al., 2009). During photoheterotrophic metabolism, the TCA cycle is primarily employed for intermediate metabolite production, rather than acetyl-CoA oxidation and energy production. The first identified shunt allowing two-carbon compound assimilation was the glyoxylate pathway (Kornberg & Krebs, 1957), through which two acetyl-CoA molecules are converted into malate to replenish the intermediates consumed. The key enzyme involved in the glyoxylate shunt is isocitrate lyase, which converts isocitrate into succinate and glyoxylate. Glyoxylate is then combined with acetyl-CoA by malate synthase to produce malate. Through the consumption of one citrate molecule and one acetyl-CoA molecule, this cycle results in the net production of two molecules of malate, Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 Printed in Great Britain 1061 B. Leroy and others which will be converted into oxaloacetate and 2-oxoglutarate, which are at the root of several biosynthetic pathways. A number of bacteria, including Rs. rubrum, lack a functional isocitrate lyase and are thus designated Icl2 (Kornberg & Lascelles, 1960). In these organisms, an alternative to the glyoxylate shunt is required to produce oxaloacetate and 2oxoglutarate to feed the reactions of cell component biosynthesis (Anthony, 2011). The methylaspartate cycle was recently demonstrated to occur in haloarchaea (Khomyakova et al., 2011). Moreover, a second pathway allowing acetylCoA assimilation, the ethylmalonyl-CoA (EMC) pathway, has been identified in Rhodobacter sphaeroides by Alber and colleagues (Erb et al., 2009a). All the enzymes belonging to this pathway in Rb. sphaeroides have been identified, purified and extensively characterized (Alber et al., 2006; Erb et al., 2007, 2008, 2009b; Meister et al., 2005). Schneider et al. (2012) recently showed that Methylobacterium extorquens also employs the EMC pathway. On the other hand, Rs. rubrum has been proposed to assimilate acetate through the citramalate cycle (Ivanovsky et al., 1997), based on the early finding that acetate assimilation is accompanied by the production of a compound identified as citramalate (Hoare, 1963) and on the presence of enzymatic activity allowing the production of glyoxylate and propionyl-CoA from C5 compounds (Osumi & Katsuki, 1977). Berg and Ivanovsky’s group also reported the detection of enzymatic activity related to this pathway in Rs. rubrum cell extracts (Berg & Ivanovskiı̆, 2009; Berg et al., 2002; Filatova et al., 2005), but the enzyme responsible for citramalate synthesis or conversion to mesaconate has not been identified/purified to date, nor has a putative gene encoding this enzyme been detected in the genome. Recently, the same group observed the expression of crotonyl-CoA carboxylase/reductase, the key enzyme of the EMC pathway, in Rs. rubrum grown photoheterotrophically on acetate, but the authors concluded that EMC involvement was not significant in this strain (Berg & Ivanovskiı̆, 2009). In both the citramalate and EMC pathways, glyoxylate and propionyl-CoA are produced from mesaconyl-CoA, with 3methylmalyl-CoA generated as an intermediate, but the two pathways differ in the first step, leading to production of mesaconyl-CoA as well as propionyl-CoA assimilation. Here, we used growth analysis, proteomics and enzymatic activity tests to analyse acetate photoassimilation in Rs. rubrum strain S1H. More than 1700 proteins, covering a significant part of the expressed proteome, were identified and quantified. Except for those shared with poly(3-hydroxybutyrate) (PHB) formation, all the proteins belonging to the EMC pathway exhibited a higher abundance under the conditions tested, suggesting that the EMC pathway could effectively be operational in our strain. In addition, activity of crotonyl-CoA carboxylase/reductase was also higher under acetate conditions. An enzyme that could be responsible for detected citramalate synthesis has also been observed in higher abundance in acetate-growing culture. Citramalate synthesis is here proposed to be a consequence of operation of part of the branched-chain amino acid biosynthetic pathway. In addition, several other proteins showed higher abundances 1062 and could be linked to the cellular redox balance of the cells and also to a potential assimilation pathway through glutarylCoA dehydrogenase that has not so far been identified. METHODS Bacterial strain, culture medium and growth conditions. Rs. rubrum strain S1H (ATCC 25903), hereafter named S1H, was used to study acetate assimilation under anaerobic and phototrophic conditions, in accordance with the requirements of the MELiSSA project. The S1H strain differs from the S1 type strain only in its lower sensitivity to growth inhibition by L-threonine. The bacterium was cultivated on the basal salt medium of Segers & Verstraete (1983) as described by Suhaimi et al. (1987), with acetate (62 mM) or succinate (124 mM) and ammonium chloride (35 mM) as the sources of C and N, respectively, and biotin as the only vitamin provided. Carbonate was included in the medium at a concentration of 3 mM under both conditions. When acetate and propionate were used as a mixed carbon source, both components were supplied at an equivalent net carbon concentration (acetate 31 mM, propionate 20.6 mM). The cultures were grown in 50 ml serum bottles containing 40 ml medium. Pure nitrogen was used to remove oxygen from the upper gas phase and the flasks were then hermetically sealed. Unless otherwise stated, an illumination intensity of 50 mmol m22 s21, as determined with an underwater quantum meter (LI-193; LI-COR), was supplied with a halogen light (10 W, 100 lm; 2650 K; Sencys). The cultures were incubated at 30 uC with rotary shaking at 150 r.p.m. Growth was monitored by measuring OD680. Proteomic analysis. Bacteria were acclimated and subcultivated in acetate or succinate starting at OD680 0.4–0.5. On reaching OD680 0.9– 1.2, bacteria were harvested via centrifugation at 16 000 g at 4 uC, and proteins were extracted using guanidinium chloride (6 M) and ultrasonication (3610 s, amplitude 40 %, IKA U50 sonicator). The protein concentration was determined using the Bradford method (Bradford, 1976), with bovine gamma-globulin as a standard. Aliquots of protein (50 mg) were reduced, alkylated and precipitated with acetone. The protein pellets obtained were solubilized using 50 mM ammonium bicarbonate containing 1 mg trypsin, and incubated overnight at 37 uC. Digestion was stopped with 0.1 % formic acid (v/ v, final concentration). Protein identification and quantification were performed following a label-free strategy on a UHPLC-HRMS platform (Eksigent 2D UltraAB SCIEX Triple TOF 5600). Peptides (2 mg) were separated in a 25 cm C18 column (Acclaim PepMap100, 3 mm; Dionex) using a linear acetonitrile (ACN) gradient [5–35 % (v/v), in 120 min] in water containing 0.1 % (v/ v) formic acid, at a flow rate of 300 nl min21. To achieve the greatest possible retention time stability, which is required for label-free quantification, the column was equilibrated with 10 vols 5 % ACN prior to each injection. Mass spectra were acquired over the range 400– 1500m/z in high-resolution mode (resolution .35 000), with a 500 ms accumulation time. The instrument was operated in data-dependent acquisition mode, and MS/MS spectra were acquired over the range 100–1800m/z. The precursor selection parameters were as follows: intensity threshold, 200 c.p.s.; maximum precursors per cycle, 50; accumulation time, 50 ms; exclusion time after one spectrum, 15 s. These parameters lead to a duty cycle of 3 s per cycle to ensure that highquality extracted ion chromatograms (XICs) are obtained for peptide quantification. ProteinPilot software (v4.1) was used to perform database searches against the UniProt database, restricted to Rhodospirillum entries (ATCC 11170 plus F11 strains). The search parameters included differential amino acid mass shifts for carbamidomethyl cysteine, oxidized methionine, all biological modifications, amino acid substitutions and missed trypsin cleavage sites. Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 Microbiology 161 Acetate photoassimilation in Rhodospirillum rubrum For peptide quantification, the ‘quant’ application of PeakView was employed to calculate XICs for the five top peptides from all proteins identified at a false discovery rate below 1 %. Only unmodified and unshared peptides were subjected to quantification. Peptides were also excluded if the confidence in their identification was below 0.99, as determined in ProteinPilot. A retention time window of 2 min and a mass tolerance of 0.015m/z were applied. The calculated XICs were exported into MarkerView, which was used to normalize the chromatograms based on the summed area of the entire run. Only proteins exhibiting a fold change greater/less than 1.5/0.66 with a Pvalue lower than 0.05 across the three biological replicates and quantified with at least two peptides were employed for metabolic characterization. with 1 % (w/v) OsO4/cacodylate buffer at room temperature, and washed again. They were subsequently dehydrated in a graded ethanol series up to 100 %, and finally embedded in Spurr resin (TAAB Laboratories Equipment). Ultrathin slices (70–90 mm) obtained with a Leica Ultracut UCT ultramicrotome were stained with uranyl acetate and lead citrate. Prepared samples were observed with a Zeiss LEO 906E transmission electron microscope operated at 60 kV, and transmission electron microscopy images were acquired with analySIS software (Soft Imaging System). Monitoring of volatile fatty acid consumption by HPLC analysis. Culture supernatants were obtained from culture samples after centrifugation at 16 000 g for 10 min at 4 uC. Aliquots (50 ml) of Rs. rubrum growth with acetate as unique carbon source – light stress sensitivity culture supernatants were analysed in isocratic mode using a Shodex Sugar SH1011 column (30068 mm) with aqueous H2SO4 (5 mM) as the mobile phase. Acetate and propionate were assayed by integrating their specific peaks (retention times 11 and 13 min, respectively) and comparison with a reference curve constructed for acetate and propionate standards. Detection was performed using a refractometer. Enzymatic activity tests. Cells were harvested by centrifugation at 16 000 g for 10 min at 4 uC in the middle of the exponential growth phase (OD680 ~1). Cells were mechanically lysed in Tris/HCl buffer (50 mM pH 7.9) using bead-beating [bead size ,106 mm (Sigma Aldrich); 1 : 3 beads : sample ratio; 3000 r.p.m., 5 min at 4 uC). Cellfree extract was obtained by centrifugation at 16 000 g for 10 min at 4 uC. The protein concentration was determined using the Bradford method (Bradford, 1976), with bovine gamma-globulin as standard. Acetolactate synthase activity was tested as described previously (Muhitch, 1988), using leucine and valine (10 mM each) as inhibitors of acetolactate synthase to obtain specific activity as recommended. The specific acetolactate synthase activity is defined as the part of the measured activity that can be inhibited by the presence of leucine and valine. The final activity observed in the presence of leucine and valine (the non-specific activity is attributed to pyruvate decarboxylase) was subtracted from the activity measured in their absence. Acetolactate produced after 1 h was assayed at a single end point by conversion to acetoin, which was detected by the reaction described by Westerfield (1945). The amount of acetoin produced was determined through the use of a standard curve. Values were normalized using the amount of protein determined by Bradford assay. Crotonyl-CoA carboxylase/ reductase activity was measured as described previously (Erb et al., 2007) by following crotonyl-CoA-dependent oxidation of NADPH spectrophotometrically at 380 nm (e51210 l mol21 cm21) for 3 min. The test started when protein extract (20–40 mg) was added to the reaction mixture. Glutaryl-CoA dehydrogenase activity was measured as described previously (Blázquez et al., 2008) by following glutarylCoA-dependent reduction of ferricenium hexafluorophosphate spectrophotometrically at 280 nm (e54300 l mol21 cm21) for 50 min, the test starting when protein extract (20–40 mg) was added to the reaction mixture. Values obtained for each enzymatic activity were normalized using the amount of protein determined by Bradford assay. All tests were also carried out in the absence of protein extract to ensure the protein dependence of the test (data not shown). Transmission electron microscopy. Cells were harvested by centrifugation at 16 000 g for 10 min at 4 uC in the middle of the exponential growth phase (OD680 ~1). Cells were washed with a cacodylate buffer at pH 7.8 [0.1 M cacodylate, 8 mM NaCl, 0.05 % (w/v) ruthenium red] and immediately fixed for 2 h at 4 uC with 3 % (v/v) glutaraldehyde in cacodylate buffer. After two additional 10 min washing steps with cacodylate buffer, samples were post-fixed for 2 h http://mic.sgmjournals.org RESULTS AND DISCUSSION To better characterize the growth behaviour of our strain when using acetate as the carbon source and to define culture conditions allowing almost parallel growth on acetate and succinate, which is a requirement for unbiased proteomic analysis, we first established growth curves for the acetate and succinate cultures under various conditions. In contrast with succinate, we observed a particularly long lag phase (more than 150 h) for the acetate cultures under conditions involving a small amount of succinateobtained inoculum (starting OD680 ~0.1) and high light intensity (150 mmol m22 s21) (Fig. 1a). Because this difference would bias the proteomic analysis, we decided to optimize culture conditions to decrease the lag phase for acetate photoassimilation. Even though such a long lag phase is unlikely to be attributable to metabolic adaptation, we tested the effect of using acetate-grown inocula without any decrease in the lag phase (data not shown). No improvement was observed when the carbonate concentration was doubled (data not shown), suggesting that the long lag phase observed did not result from the already described dependence of acetate assimilation on the presence of CO2 (Hoare, 1962; Porter & Merrett, 1972). In contrast, we were able to significantly decrease the length of the lag phase by decreasing the light intensity from 150 to 50 mmol m22 s21 (Fig. 1b). A second major improvement was achieved by increasing the size of the inoculum so that the culture started with OD680 0.5–0.6 (Fig. 1c), in place of 0.2 initially. Based on these results, it appears likely that the long lag phase could be due to a higher sensitivity to light stress when S1H is cultivated with acetate as unique carbon source. Indeed, dilution of the culture at the time of inoculation significantly increases the light intensity experienced at the cellular level. In contrast, increasing the OD at inoculation and lowering the light intensity reduce the level of ‘inoculation stress’. To confirm the higher sensitivity to light stress of S1H growing on acetate vs succinate as the carbon source, cultures reaching OD680 1.1 (the middle of the exponential growth phase) were subjected to a twofold increase in the light intensity (from 50 to 100 mmol m22 s21). As shown in Fig. 2, the increase in the light intensity had almost no effect on the growth rate of the bacteria on succinate but completely abolished growth on acetate. This experiment demonstrates that Rs. rubrum Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 1063 B. Leroy and others (a) 2.5 1.5 1.0 Time (h) 2.0 Log OD680 0 200 –0.5 300 400 600 OD680 0.5 1.5 Succinate 1.0 Acetate –1.0 0.5 –1.5 Succinate –2.0 Acetate –50 –2.5 0 50 Time, t (h) 100 150 –3.0 (b) 1.0 Time (h) 0.5 Log OD680 0 50 –0.5 100 150 250 200 –1.0 –1.5 succinate –2.0 acetate –2.5 –3.0 –3.5 (c) 1.0 0.8 Log OD680 0.6 –0.4 Time (h) 0.2 0 –0.2 20 40 60 80 100 120 140 –0.4 succinate –0.6 acetate –0.8 explain the difference in sensitivity observed between these two culture conditions. A rapid increase in the light intensity would trigger an increase in the photosynthetically produced proton gradient, which would, in turn, increase the rate of NAD+ photoreduction to NADH through reverse NADH dehydrogenase activity (Golomysova et al., 2010; Herter et al., 1998; McEwan, 1994). Acetate is a more reduced carbon source than succinate (McKinlay & Harwood, 2010), and the NAD+/NADH ratio could be more difficult to maintain if the intensity of the light received is suddenly increased. In support of this hypothesis, the elemental composition of acetate-grown Rs. rubrum exposed to different lighting regimes clearly indicates that a more reduced biomass is produced when the bacterium is cultivated at a biomass concentration of 1 g l21 (high light) than at 4 g l21 (low light) (Favier-Teodorescu et al., 2003). Proteomic analysis of acetate photoassimilation by Rs. rubrum S1H Fig. 1. Optimization of culture conditions for Rs. rubrum when using acetate as the sole carbon source. Growth was followed by measuring OD680 (expressed as log OD680) for cultures submitted to high light intensity (150 mmol m”2 s” 1) and inoculated at OD680 0.1 (a), submitted to low light intensity (50 mmol m”2 s”1) and inoculated at OD680 0.1 (b), and submitted to low light intensity (50 mmol m”2 s”1) and inoculated at OD680 0.5–0.6 (c). All inocula were taken from precultures already acclimated to the corresponding carbon source. Means±SD; n53 for all conditions. S1H shows greater sensitivity to light stress when growing on acetate as the sole carbon source. We assume that a redox imbalance is responsible for this light stress and can 1064 Fig. 2. Comparison of the effect of light stress on Rs. rubrum S1H grown on acetate or succinate as the sole carbon source. Acetate and succinate cultures submitted to low light intensity (50 mmol m”2 s”1) and inoculated at OD680 0.5–0.6 were subjected when they reached the middle of the exponential growth phase (OD680 ~1.1) to a twofold increase in light intensity (t50). Growth was followed by measuring OD680. All inocula were taken from precultures already acclimated to the corresponding carbon source. Means±SD; n53 for all conditions. For proteomic analysis, cultures were grown under a low light intensity (50 mmol m22 s21), starting at OD680 between 0.4 and 0.5, to reduce the delay in the growth of acetate cultures as much as possible. To avoid bias due to the different light intensities effectively perceived by the cells, biomass samples were collected at a constant cell density (OD680 1.1–1.4) from the cultures grown on acetate and succinate. Three biological replicates, originating from different clonal populations, were analysed under acetate and succinate conditions. Altogether, more than 1700 unique proteins, representing almost 50 % of the theoretical proteome (3461 and 3878 entries for ATCC 11170 and strain F11, respectively, Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 Microbiology 161 Acetate photoassimilation in Rhodospirillum rubrum in UniProt), were identified and quantified. The complete dataset can be found in Table S1, available in the online Supplementary Material. Among the proteins for which a significant (P,0.05) fold change (.1.5 or ,0.66) was detected between the acetate and succinate conditions, 352 and 225 proteins exhibited lower and higher abundances, respectively, under acetate conditions. For further interpretation, only proteins quantified with at least two peptides were taken into account. Ethylmalonyl-CoA pathway involvement in acetate assimilation by Rs. rubrum S1H Proteomic data revealed that all the enzymes proposed to belong to the EMC pathway (Table 1) were effectively upregulated (Fig. 3b), suggesting that this pathway could be involved in acetate assimilation in our strain. These enzymes were often badly annotated in the database but comparison to sequences of EMC enzymes studied in Rb. sphaeroides (Alber, 2011) allowed us to correct their annotation and clearly link them to this pathway. (Table S2 presents the best homologue candidates of all the enzymes of the EMC pathway of Rb. sphaeroides in the genome of Rs. rubrum.) The data obtained through differential proteomic analyses are therefore largely in favour of the significant use of the EMC pathway for the assimilation of acetate in our strain. As the upregulation of an enzyme does not evidence its higher activity, crotonyl-CoA carboxylase/reductase (Ccr) activity has been measured in cell-free extracts of Rs. rubrum S1H grown with acetate or succinate as unique carbon source. As shown in Fig. 4b, Ccr activity was significantly higher (8.8 times higher; P,0.01) in acetate than in succinate conditions. This result is in contradiction with that obtained previously by Berg & Ivanovskiı̆ (2009), who observed no increase in Ccr activity in acetate conditions compared with malate conditions. This difference between the two studies could arise from the different experimental set-up (such as light intensity, which was not sufficiently detailed to be compared with our study) and also in Ccr activity test procedures. Notably, Ivanovsky’s group used incorporation of radiolabelled bicarbonate while we used cofactor oxidation, and their extracts were produced using a French press while we used bead-beating. Also, the exact control used in their study is unclear. Citramalate involvement in acetate assimilation by Rs. rubrum S1H In the Icl2 purple bacterium Rs. rubrum, Ivanovsky and coworkers proposed that the citramalate cycle takes place, in which acetate is assimilated via the condensation of acetylCoA with pyruvate through the action of citramalate synthase (Ivanovsky et al., 1997). This pathway was proposed based on the observation that citramalate is the earliest molecule to be labelled when 13C-acetate is photoassimilated (Porter & Merrett, 1972). Here, we observed slight upregulation of a http://mic.sgmjournals.org protein that is currently annotated as 2-isopropylmalate synthase (LeuA) (Rru_A0695; fold change, 1.6; P-value, 0.01; identified peptides, 12), but for which the PRIAM method [for automated enzyme detection (Claudel-Renard et al., 2003)] predicts citramalate synthase activity with a higher probability (E-values, 6e254 for citramalate synthase versus 1e241 for isopropylmalate synthase). Both 2-isopropylmalate synthase and citramalate synthase are members of the LeuA dimer superfamily (Frantom, 2012). The best-characterized citramalate synthase among proteobacteria is that of Geobacter sulfurreducens (Risso et al., 2008a), and this protein shares 51 % identity with Rru_A0695. These data favour the annotation of this protein as a true citramalate synthase. Nevertheless, upregulation of another group of enzymes leads us to propose a new hypothesis regarding involvement of citramalate in acetate assimilation by Rs. rubrum S1H. Indeed, among the identified upregulated proteins, we detected a cluster of three genes encoding ketolacid reductoisomerase (IlvC) (Rru_A0469; fold change, 4.1; P-value, 0.0005; identified peptides, 54) and the acetolactate synthase small (Rru_A0468; fold change, 3.1; P-value, 0.003; identified peptides, 22) and large subunits (Rru_A0467; fold change, 3.1; P-value, 0.003; identified peptides, 33). These proteins are involved in the first step of branched-chain amino acid (isoleucine, leucine and valine, ILV) synthesis (Fig. 3a). Dihydroxy-acid dehydratase (IlvD), which is also involved in this pathway, was also more abundant under acetate conditions (Rru_A1786; fold change, 1.6; P-value, 0.04; identified peptides, 32). Citramalate is an intermediate in the atypical isoleucine synthesis pathway, which is independent of threonine supply (Wu et al., 2010). In this alternative pathway, citramalate is converted to 2-oxobutanoate by the action of 3-isopropylmalate/(R)-2-methylmalate dehydratase large and small subunits (LeuC and LeuD, Rru_A1189 and Rru_A1190) and 3-isopropylmalate dehydrogenase (LeuB, Rru_A1191). These three proteins were identified in this study but were not significantly differentially regulated under acetate condition (Fig. 3a). Acetolactate synthase can equivalently transfer an acetaldehyde from pyruvate to 2-oxobutanoate, forming 2-ethyl-2-hydroxy-3oxobutanoate, a reaction involved in the ILV biosynthesis pathway. As the upregulation of an enzyme does not evidence its higher activity, we measured the acetolactate synthase activity in cell-free extracts of Rs. rubrum. As shown in Fig. 4a, acetolactate synthase activity was slightly but significantly higher in acetate conditions than in the succinate control (P,0.05), suggesting this enzyme could be involved in acetate photoassimilation. However, we observed a decrease in abundance of some branched-chain amino acid transaminases (Rru_A2223; fold change 0.4, P-value 0.01, identified peptides 16; Rru_A1040, fold change 0.03, P-value 0.001, identified peptides 16), suggesting that the outcome of acetate assimilation through reactions belonging to the ILV pathway is not the production of branched-chain amino acids. 3-Methyl-2-oxopentanoate and/or 2-oxoisovalerate could, for example, be directed to another metabolic pathway, such as the valine degradation Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 1065 B. Leroy and others Table 1. Differentially expressed proteins UniProt accession number Locus tag P-value Fold change (acetate/succinate)* No. identified peptidesD Description tr|Q2RYD8 tr|Q2RYD7 tr|Q2RXX3 tr|Q2RXR7 tr|Q2RXR6 tr|Q2RXR4 tr|Q2RV44 tr|Q2RV43 tr|Q2RU23 tr|Q2RT18 tr|Q2RRG6 tr|Q2RQI1 tr|Q2RQ36 tr|Q2RPT8 tr|Q2RPT7 tr|Q2RPT6 tr|Q2RWQ6 sp|P80448 tr|Q2RX73 tr|Q2RX72 tr|G2T8X1 tr|Q2RWJ6 tr|Q2RVK4 tr|G2TF50 tr|G2TF51 tr|G2TFQ4 tr|Q2RTR3 tr|G2T9T2 tr|Q2RT67 tr|Q2RT66 tr|Q2RT64 tr|Q2RS72 sp|Q53046 tr|G2T7S3 tr|G2T7S5 tr|Q2RU35 tr|G2T7T0 tr|Q2RU29 tr|G2TF86 tr|Q2RRP3 sp|Q2RRP2 tr|Q2RRP1 tr|Q2RRP0 tr|Q2RW98 tr|Q2RW96 tr|Q6US84 tr|Q6US85 tr|Q6US86 tr|Q6US87 tr|Q2RSU0 tr|Q2RST9 Rru_A0052 Rru_A0053 Rru_A0217 Rru_A0273 Rru_A0274 Rru_A0276 Rru_A1200 Rru_A1201 Rru_A1572 rru_A1927 Rru_A2479 Rru_A2817 Rru_A2964 Rru_A3062 Rru_A3063 Rru_A3064 Rru_A0635 Rru_A0077 Rru_A0467 Rru_A0468 Rru_A0469 Rru_A0695 Rru_A1040 Rru_A1189 Rru_A1190 Rru_A1191 Rru_A1682 Rru_A1786 Rru_A1878 Rru_A1879 Rru_A1881 Rru_A2223 Rru_A2398 Rru_A1556 Rru_A1558 Rru_A1560 Rru_A1563 Rru_A1566 Rru_A2400 Rru_A2402 Rru_A2403 Rru_A2404 Rru_A2405 Rru_A0793 Rru_A0795 Rru_A2264 Rru_A2265 Rru_A2266 Rru_A2267 Rru_A2005 Rru_A2006 0.000003 0.000045 0.002630 0.070980 0.001060 0.022770 0.001320 0.000340 0.008110 0.004190 0.005550 0.000870 0.489180 0.000002 0.002520 0.006950 0.000630 0.002830 0.003230 0.003480 0.000500 0.012260 0.001140 0.080750 0.341380 0.017120 0.02998 0.041030 0.000130 0.000059 0.000310 0.012260 0.010130 0.000280 0.033260 0.032290 0.014250 0.000009 0.037140 0.017310 0.001050 0.002590 0.001440 0.011820 0.001720 0.000320 0.002080 0.000720 0.000720 0.000007 0.000900 4.75 5.14 2.08 0.86 0.53 1.60 2.31 4.46 1.58 1.83 1.98 0.66 0.81 8.80 3.20 2.79 0.21 0.26 3.12 3.07 4.07 1.59 0.03 1.38 1.15 1.36 1.70 1.62 0.16 0.21 0.40 0.44 1.83 2.43 1.54 1.82 1.91 80.97 0.52 2.12 4.82 2.05 2.63 0.06 0.02 6.23 4.15 5.35 6.14 176.27 8.05 54 40 39 51 119 10 25 20 9 34 28 10 15 7 27 35 5 8 33 22 54 12 16 22 19 31 27 32 13 20 20 16 80 12 8 42 5 3 54 8 22 24 39 3 12 6 28 25 32 26 26 Biotin carboxylase Carboxyltransferase L-Malyl-CoA/b-methylmalyl-CoA lyase 3-Oxoacyl-(acyl-carrier-protein) reductase Acetyl-CoA acetyltransferase Polyhydroxyalkanoate synthesis repressor Malyl-CoA thioesterase Mesaconyl-CoA hydratase Ethylmalonyl-CoA/methylmalonyl-CoA epimerase Acetyl-CoA hydrolase Methylmalonyl-CoA mutase Phasin MaoC-like dehydratase Ethylmalonyl-CoA mutased Crotonyl-CoA carboxylase/reductase Methylsuccinyl-CoA dehydrogenase Na+/solute symporter Ferredoxin-2 Acetolactate synthase, large subunit Acetolactate synthase, small subunit Ketol-acid reductoisomerase 2-Isopropylmalate synthase Leucine dehydrogenase 3-Isopropylmalate dehydratase, large subunit 3-Isopropylmalate dehydratase,small subunit 3-Isopropylmalate dehydrogenase Branched-chain amino acid aminotransferase Dihydroxy-acid dehydratase Dihydrolipoamide dehydrogenase Dihydrolipoamide acetyltransferase Pyruvate dehydrogenase (lipoamide) Branched-chain amino acid aminotransferase Pyruvate : ferredoxin oxidoreductase NADH-quinone oxidoreductase subunit B NADH dehydrogenase subunit D NADH-quinone oxidoreductase, subunit F NADH-quinone oxidoreductase subunit I NADH-plastoquinone oxidoreductase, chain 5 Ribulose bisphosphate carboxylase Ribulose-phosphate 3-epimerase Fructose-1,6-bisphosphatase class 1 Phosphoribulokinase Transketolase Nitrogenase Nitrogenase iron protein 1 FixX§ FixC§ FixB§ FixA§ Glutaryl-CoA dehydrogenase L-Carnitine dehydratase/bile acid-inducible protein F The complete dataset can be found in the supplementary data. *The fold change is the ratio of the abundance of a protein under acetate conditions to the abundance under succinate conditions. D‘No. identified peptides’ represents the number of identified peptides with a confidence higher than 95 %. dDue to interference in MS signal, fold change for this protein was manually corrected and results from quantification of four peptides rather than five. §Fix, Electron transfer flavoprotein. 1066 Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 Microbiology 161 Acetate photoassimilation in Rhodospirillum rubrum (b) O OH O HO CH3 D-Citramalate Rru_A1190 #19; x1.15 #22; x1.4 CH3 D-Erythro-3- methylmalate Rru_A1191 S CoA O H3C CoA S Crotonyl-CoA Rru_A2398 O H 3C H3C 2× Acetyl-CoA #15; x1 CO2 NADH + H+ CO2 #31; x1.4 #34; x1.8 O Rru_A2964 S CoA Acetyl-CoA NAD+ Rru_A1927 CoA S OH Acetate (R)-3-Hydroxybutyryl-CoA O H3C O H3C #34; x1.8 #12; x1.6 O OH Rru_A1927 OH NADP+ #51; x0.85 Acetate O NADPH + H+ Rru_A0273 OH H3C Rru_A0695 OH O H3C CoA S Acetoacetyl-CoA O Rru_A1189 HO O H3C #22; x0.8 O HO Rru_A1380 (a) CO2 NADPH + H+ NADP+ Rru_A3063 #80; x1.8 #27; x3.2 Rru_A2005 CO2 #26; x176 OH O O 2-Oxobutanoate OH H3C O CO2 O O OH H3C (2S)-Ethylmalonyl-CoA #33; x3/#22; x3.1 #9; x1.6 CO2 O OH O HO O OH OH CH3 NADPH + H+ FAD Rru_A3062 Acetolactate FADH2 #7; x8.8 O HO NADPH + H+ Rru_A0469 #54; x4 NADP+ (2R)-Ethylmalonyl-CoA OH HO CH3 (S)-2-Aceto-2hydroxybutanoate CoA S O H3C CH3 CoA S Glutaryl-CoA Rru_A1572 Rru_A0467/Rru_A0468 H 3C O HO CH3 Pyruvate O CoA S O Pyruvate O O HO S CoA CH3 O NADP+ Methylsuccinyl-CoA HO CH3 O HO CH3 O H 3C Rru_A3064 #35; x2.8 OH H3C OH O OH OH (R)-2,3-Dihydroxy-3methylpentanoate HO (R)-2,3-Dihydroxy-3methylbutanoate CoA Mesaconyl-CoA Rru_A1786 H Rru_A1201 #32; x1.6 #20; x4.4 CH3 O OH CH3 O H3C S CH3 O O CH3 β-Methylmalyl-CoA 2-Oxoisovalerate (S)-3-Methyl-2oxopentanoate S CoA OH O H O HO H3C OH O O Rru_A0217 O #39; x2.1 Rru_A1040 Rru_A1040 #16; x0.03 #16; x0.03 CH3 O H3C O OH Isoleucine CH3 HO CH3 NH2 CH3 NH2 Valine Glyoxylate H3C O CH3 HO NH2 OH O O Leucine S CoA Propionyl-CoA HCO3– ATP ADP + Pi Rru_A0052/0053 #54; x4.7/#40; x5 O Rru_A0217 #39; x2.1 OH O HO S CoA HO CH3 (R)-Methylmalonyl-CoA Rru_A1572 O S CoA O Malyl-CoA Rru_A1200 #25; x2.3 #9; x1.6 O Fold change <0.2 <0.5 <0.66 <1.5 <2 <5 >5 OH O S CoA HO O HO CH3 (S)-Methylmalonyl-CoA OH O Malate Rru_A2479 #28; x2 O HO S CoA O Succinyl-CoA http://mic.sgmjournals.org Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 1067 B. Leroy and others Fig. 3. Schematic representation of the metabolic pathways highlighted through differential proteomic analysis. (a) Metabolic reactions linked to branched-chain amino acid (ILV) biosynthesis. (b) Ethylmalonyl-CoA pathway. Water molecules were omitted to preserve clarity of the schematic representation. x: fold change observed in differential proteomic analysis; #, number of peptides detected for the protein. Colour indicates the fold change: red represents downregulation under the acetate conditions, green represents upregulation under the acetate conditions, and light grey represents not significantly affected by carbon source. Arrows in blue represent enzymatic reactions that have been monitored by an enzymatic activity test. pathway (Sykes et al., 1987), which is conducive to propionylCoA production. However, despite intensive researches for an enzyme similar to the already described keto-acid dehydrogenase in the genome of Rs. rubrum, we could not find any candidate fulfilling this function. Further investigation, notably including metabolite analyses, will thus be needed in order to clarify the involvement of reactions belonging to the ILV synthesis pathway in acetate photoassimilation in Rs. rubrum. (a) 35 Specific activity (nmol min–1 mg–1) 30 25 20 15 10 5 (b) 60 Specific activity (nmol min–1 mg–1) 0 50 Acetate Succinate Acetate Succinate Acetate Succinate 40 30 20 10 (c) 25 Specific activity (nmol min–1 mg–1) 0 20 15 10 5 0 –5 Fig. 4. Enzymatic activity test. Cell-free extracts of Rs. rubrum S1H grown with acetate or succinate as unique carbon source were tested for acetolactate synthase activity (n56) (a), crotonylCoA reductase/carboxylase activity (n55) (b), and glutaryl-CoA dehydrogenase activity (n54) (c). Means±SE. 1068 Acetolactate synthase combines two molecules of pyruvate to produce one molecule of acetolactate and releases one molecule of CO2. Substantial activation of this pathway would result in enhanced consumption of pyruvate, which would need to be replenished continuously. We detected the presence of a pyruvate : ferredoxin oxidoreductase, also known as NifJ (Rru_A2398; fold change, 1.8; P-value, 0.01; identified peptides, 80), which is commonly reported to catalyse oxidative decarboxylation of pyruvate to generate acetyl-CoA. It has been shown that this enzyme is able to catalyse the reverse reaction as well, thus producing pyruvate via the reductive carboxylation of acetyl-CoA (Furdui & Ragsdale, 2000; St Maurice et al., 2007). At physiological concentrations, Furdui & Ragsdale (2000) measured a significant carboxylating activity and showed that the rate of the reaction depends only on the concentration of acetyl-CoA and pyruvate. As the intracellular concentration of acetyl-CoA is likely to be high in cells grown on acetate and as pyruvate level should be kept low through consumption by acetolactate synthase, the net flux through this enzyme is expected to favour pyruvate synthesis when the bacteria are grown on acetate. Noticeably, the high level of acetyl-CoA can be maintained in the cell through the upregulation of acetyl-CoA transferase (Rru_A1927; fold change, 1.8; P-value, 0.004; identified peptides, 34), which could be responsible for acetate activation, as mentioned by Alber et al. (2006). The preference for reductive carboxylation activity is further supported by the observation of major downregulation of all subunits of the pyruvate dehydrogenase (lipoamide) complex (Rru_A1878, Rru_A1879, Rru_A1881). Indeed, pyruvate dehydrogenase (lipoamide), dihydrolipoamide acetyltransferase and dihydrolipoamide dehydrogenase were all clearly downregulated under acetate conditions, suggesting that oxidative decarboxylation of pyruvate activity is lowered in cells grown on acetate. Importantly, reductive carboxylation activity relies on the presence of an electron supplier, which has been proposed to be ferredoxin (Furdui & Ragsdale, 2000). In our dataset, we identified 4Fe–4S ferredoxin as differentially regulated under acetate conditions, but its abundance was greatly decreased (Rru_A0077; fold change, 0.25; P-value, 0.003; identified peptides, 8). Nevertheless, a second ferredoxin encoded in the fix gene cluster was clearly upregulated under acetate conditions (Rru_A2264; fold change, 6.2; P-value, 0.0003; identified peptides, 6) and could therefore serve as the electron supplier for reductive carboxylation of acetate. Altogether, these data are in favour of a major involvement of carboxylation of acetyl-CoA to pyruvate, which in turn feeds the branched-chain amino acid synthesis pathway. This interpretation of our result is also in agreement with the data of Porter & Merrett (1972), who Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 Microbiology 161 Acetate photoassimilation in Rhodospirillum rubrum Differential proteomic analysis also revealed the intense upregulation of the clustered glutaryl-CoA dehydrogenase (Rru_A2005; fold change, 176; P-value, 76106; identified peptides, 26) and carnitine hydratase (best annotated as a CoA transferase) (Rru_A2006; fold change, 8; P-value, 96104; identified peptides, 26). Glutaryl-CoA dehydrogenase is generally observed to be involved in the anaerobic degradation of benzoate in bacteria (Härtel et al., 1993; Wischgoll et al., 2009), in which it catalyses the decarboxylation of glutaryl-CoA to produce crotonyl-CoA. Under our acetate conditions, this compound could be easily assimilated through the EMC pathway (Fig. 3b), but the origin of glutaryl-CoA and the role of the linked CoA transferase remain enigmatic. Glutaryl-CoA dehydrogenase has been shown to use electron transfer flavoproteins as electron acceptors (Ramsay et al., 1987). Under acetate conditions, we observed a large increase of the abundance of the four proteins encoded by the fix gene cluster (FixABCX; Rru_A2267, Rru_A2266, Rru_A2265 and Rru_A2264, respectively). FixA and FixB are electron transfer flavoproteins that are usually responsible for electron transfer to nitrogenase in diazotrophically growing Rs. rubrum (Edgren & Nordlund, 2004). However, nitrogenase was clearly downregulated in response to the use of acetate as the carbon source for growth (Rru_A0793, fold change 0.06, P-value 0.01, identified peptides 3; and nitrogenase iron protein Rru_A0795, fold change 0.02, P-value 0.0017, identified peptides 12). We thus analysed the glutaryl-CoA dehydrogenase activity in cell-free extracts of Rs. rubrum S1H grown in acetate and succinate conditions. As shown in Fig. 4(c), glutaryl-CoA dehydrogenase activity was not detected in the succinate conditions but was readily detected in the acetate conditions [17 nmol min21 (mg protein)21, P,0.01]. The detection of glutarylCoA dehydrogenase activity only in acetate-grown cells is in total agreement with the very high fold change observed in proteomic data which reveal absence/presence of the enzyme. We also tested (data not shown) glutaryl-CoA dehydrogenase activity in cultures assimilating butyrate and propionate as sole carbon source in the same conditions but neither showed glutaryl-CoA dehydrogenase activity. Non-decarboxylating glutaryl-CoA dehydrogenase exists, which releases the http://mic.sgmjournals.org Acetate transport in Rs. rubrum S1H Among the observed downregulated proteins, we identified a Na+/solute symporter (Rru_A0635; fold change, 0.2; Pvalue, 0.0006; identified peptides, 5) that is highly similar (70 % ID) to the acetate permease (ActP) of Escherichia coli. Downregulation of the acetate transporter in response to acetate in the medium could appear to be counterproductive, but Gimenez et al. (2003) clearly demonstrated that, in E. coli, ActP was only needed for growth at low acetate concentration. In addition, Risso et al. (2008b) have shown that acetate transporter genes in Geobacter species were only expressed at low concentration of acetate. Moreover, we observed that propionate, which has been shown to inhibit acetate transport through ActP in E. coli (Gimenez et al., 2003), had no effect on acetate assimilation in Rs. rubrum when the two carbon sources were present simultaneously and at the same concentration in the medium (Fig. 5). Altogether, these observations lead us to 35 2.5 30 2.0 25 20 1.5 15 1.0 10 0.5 5 OD680 Glutaryl-CoA dehydrogenase involvement in acetate assimilation by Rs. rubrum S1H intermediate glutaconyl-CoA in place of crotonyl-CoA. While the released compound was not identified in our enzymatic test, we consider Rru_A2005 most probably belongs to the decarboxylating group of glutaryl-CoA dehydrogenases, based on conservation of key amino acids highlighted by the Boll group (Schaarschmidt et al., 2011; Wischgoll et al., 2009), namely E91, S99 and Y373. Our data thus suggest that glutaryl-CoA dehydrogenase could play a previously unsuspected role in acetate photoassimilation. It is clear that only a detailed flux analysis could definitely identify the relative importance of the multiple pathways proposed here. Nevertheless, comparison of specific activities obtained for crotonyl-CoA carboxylase/reductase [47 nmol min21 (mg protein)21] and glutaryl-CoA dehydrogenase [17 nmol min21 (mg protein)21] suggests that the latter potentially contributes significantly to acetate assimilation. VF A in medium (mM) reported that citramalate was one of the earliest and most abundant radiolabelled compounds in short-term exposure experiments with NaH14CO3 and acetate-assimilating Rs. rubrum. Altogether, these data lead us to propose that the extensively reported contribution of citramalate synthase in acetate assimilation by Rs. rubrum is linked to the use of part of the ILV biosynthesis pathway (Fig. 3a) rather than to the so-called citramalate cycle, and that the ILV biosynthesis pathway plays a significant but not yet fully understood role in acetate assimilation. Further experiments such as flux analyses will be needed to definitely demonstrate and quantify the involvement of this proposed pathway in acetate photoassimilation in Rs. rubrum. Acetate Propionate OD680 0 0 0 20 40 Time (h) 60 80 Fig. 5. Photoassimilation of acetate and propionate when supplied together at an equivalent net carbon concentration. Rs. rubrum S1H was cultivated with a 1 : 1 mixture of acetate and propionate as carbon source. Abundances of acetate and propionate in the culture supernatant were monitored by HPLC during growth. Means±SD; n53. VFA, volatile fatty acid. Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 1069 B. Leroy and others (a) (b) 2 mm 5 mm 2 mm 5 mm Fig. 6. PHB production in Rs. rubrum S1H cultivated in acetate or succinate conditions. Transmission electron microscopy photomicrographs were obtained from Rs. rubrum S1H grown in the presence of succinate (a) or acetate (b). Arrows indicate PHB granules. suggest that ActP is not involved in acetate transport in our strain under our conditions. Respiratory chain modification during acetate assimilation by Rs. rubrum S1H Five of the ten members of the NADH-quinone oxidoreductase complex identified in the present study were detected at higher abundances under acetate conditions. The activation of respiratory chain electron transport makes sense only if a final diffusible/excretable electron acceptor is present, such as O2 or nitrate. In this study, cells were grown under anaerobic conditions with ammonium as the nitrogen source, and neither O2 nor nitrate was present as a final electron acceptor. The absence of O2 was confirmed based on pigment expression, which is immediately repressed at a very low O2 partial pressure (Sistrom, 1965). Upregulation of members of NADH-quinone oxidoreductase in acetate conditions will need further investigation. Redox balance during acetate assimilation by Rs. rubrum S1H Photoheterotrophic growth of purple bacteria has been extensively reported to depend on the Calvin–Benson– Bassham (CBB) cycle playing the role of an ‘electron sink’ (Falcone & Tabita, 1991; McKinlay & Harwood, 2010; Wang et al., 2010). In our dataset, we also observed a decreased abundance of RuBisCO when acetate is used as sole carbon source (Rru_A2400; fold change, 0.52; P-value, 0.04; identified peptides, 54) as compared with succinate-containing cultures. This finding seems to be in agreement with the conclusion of Laguna et al. (2011) that, in Rb. sphaeroides, CO2 fixation through the EMC pathway is sufficient to equilibrate redox balance even in the absence of a functional CBB cycle. Nevertheless, we observed that other redoxbalancing mechanisms could be playing a role during acetate assimilation. Notably, we observed that, as already mentioned by Kim et al. (2012) for Rb. sphaeroides, our strain produced larger amounts of PHB during acetate assimilation (Fig. 6). Hauf et al. (2013) have shown that the reduction state 1070 of the cofactor pool can influence PHB production. PHB production observed here could be the result of an increased level of NAD(P)H and indirectly be a way to deal with NAD+ photoreduction (Herter et al., 1998), notably occurring at the beginning of the growth curve. Further experiments are needed to confirm this hypothesis, notably by following PHB accumulation along the complete growth curve in acetate conditions. NifJ catalyses the reductive carboxylation of acetyl-CoA, which could also play the role of redox-balancing reaction. Interestingly, in a kinetic experiment in which we compared the proteome of a culture in lag phase with the proteome of the same culture just after the initiation of growth on acetate as a carbon source (data not shown), NifJ was the most highly overexpressed protein when growth was initiated (fold change, 11.2; P-value, 0.003; identified peptides, 86). This result suggests that NifJ plays a major role at the beginning of the growth curve, corresponding to the period during which light stress due to culture dilution likely triggered the previously mentioned redox imbalance (Favier-Teodorescu et al., 2003). CO2 fixation through acetyl-CoA carboxylation could therefore represent a new redox-balancing reaction developed by bacteria to combine acetate assimilation and maintenance of the redox balance. CONCLUSIONS In conclusion, our proteomic data suggest that Rs. rubrum S1H assimilates acetate through multiple pathways. Combination of proteomic analysis and enzymatic activity testing reveal that the EMC pathway could be operational in Rs. rubrum when growing photoheterotrophically on acetate. Indeed, all the enzymes specific to the EMC pathway were detected in higher abundances and crotonylCoA carboxylase/reductase activity was increased under acetate culture conditions. Our data also lead us to propose a new outcome for citramalate. Indeed, citramalate could be an intermediate of an alternate route of acetate assimilation through pyruvate and part of the ILV biosynthesis pathway. NifJ, which has been shown to produce pyruvate through acetyl-CoA reductive carboxylation, was upregulated in Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 Microbiology 161 Acetate photoassimilation in Rhodospirillum rubrum acetate conditions. This could represent a new redoxbalancing reaction, and the pyruvate generated could be directed to acetolactate synthase and the IlvC and IlvD enzymes, which were also detected in higher abundances in acetate conditions. The outcome of the activation of part of the ILV biosynthesis pathway during acetate assimilation is still not understood and requires further investigation. Finally, glutaryl-CoA dehydrogenase could play a previously unsuspected but important role in photoheterotroph acetate assimilation since this protein was the mostly upregulated in this study. Erb, T. J., Fuchs, G. & Alber, B. E. (2009a). (2S)-Methylsuccinyl-CoA dehydrogenase closes the ethylmalonyl-CoA pathway for acetyl-CoA assimilation. Mol Microbiol 73, 992–1008. Erb, T. J., Brecht, V., Fuchs, G., Müller, M. & Alber, B. E. (2009b). Carboxylation mechanism and stereochemistry of crotonyl-CoA carboxylase/reductase, a carboxylating enoyl-thioester reductase. Proc Natl Acad Sci U S A 106, 8871–8876. Falcone, D. L. & Tabita, F. R. (1991). Expression of endogenous and foreign ribulose 1,5-bisphosphate carboxylase-oxygenase (RubisCO) genes in a RubisCO deletion mutant of Rhodobacter sphaeroides. J Bacteriol 173, 2099–2108. Favier-Teodorescu, L., Cornet, J. F. & Dussap, C. G. (2003). Modelling continuous culture of Rhodospirillum rubrum in photobioreactor under light limited conditions. Biotechnol Lett 25, 359–364. ACKNOWLEDGEMENTS This research was supported by the European Space Agency and BELSPO (GSTP ‘Melgen-3’). This work was also partially financed by the FNRS under grant ‘Grand Equipement’ no. 2877824. Filatova, L. V., Berg, I. A., Krasil’nikova, E. N., Tsygankov, A. A., Laurinavichene, T. V. & Ivanovskiı̆, R. N. (2005). [The mechanism of acetate assimilation in purple nonsulfur bacteria lacking the glyoxylate pathway: acetate assimilation in Rhodobacter sphaeroides cells]. Mikrobiologiia 74, 313–318 (in Russian). REFERENCES Frantom, P. A. (2012). Structural and functional characterization of a-isopropylmalate synthase and citramalate synthase, members of the Alber, B. E. (2011). Biotechnological potential of the ethylmalonyl- Furdui, C. & Ragsdale, S. W. (2000). The role of pyruvate ferredoxin CoA pathway. Appl Microbiol Biotechnol 89, 17–25. oxidoreductase in pyruvate synthesis during autotrophic growth by the Wood-Ljungdahl pathway. J Biol Chem 275, 28494–28499. LeuA dimer superfamily. Arch Biochem Biophys 519, 202–209. Alber, B. E., Spanheimer, R., Ebenau-Jehle, C. & Fuchs, G. (2006). Study of an alternate glyoxylate cycle for acetate assimilation by Rhodobacter sphaeroides. Mol Microbiol 61, 297–309. Anthony, C. (2011). How half a century of research was required to understand bacterial growth on C1 and C2 compounds; the story of the serine cycle and the ethylmalonyl-CoA pathway. Sci Prog 94, 109–137. Berg, I. A. & Ivanovskiı̆, R. N. (2009). [Enzymes of the citramalate cycle in Rhodospirillum rubrum]. Mikrobiologiia 78, 22–31 (in Russian). Berg, I. A., Filatova, L. V. & Ivanovsky, R. N. (2002). Inhibition of acetate and propionate assimilation by itaconate via propionylCoA carboxylase in isocitrate lyase-negative purple bacterium Rhodospirillum rubrum. FEMS Microbiol Lett 216, 49–54. Blázquez, B., Carmona, M., Garcı́a, J. L. & Dı́az, E. (2008). Gimenez, R., Nuñez, M. F., Badia, J., Aguilar, J. & Baldoma, L. (2003). The gene yjcG, cotranscribed with the gene acs, encodes an acetate permease in Escherichia coli. J Bacteriol 185, 6448–6455. Golomysova, A., Gomelsky, M. & Ivanov, P. S. (2010). Flux balance analysis of photoheterotrophic growth of purple nonsulfur bacteria relevant to biohydrogen production. Int J Hydrogen Energy 35, 12751– 12760. Härtel, U., Eckel, E., Koch, J., Fuchs, G., Linder, D. & Buckel, W. (1993). Purification of glutaryl-CoA dehydrogenase from Pseudo- monas sp., an enzyme involved in the anaerobic degradation of benzoate. Arch Microbiol 159, 174–181. Hauf, W., Schlebusch, M., Hüge, J., Kopka, J., Hagemann, M. & Forchhammer, K. (2013). Metabolic changes in Synechocystis Identification and analysis of a glutaryl-CoA dehydrogenase-encoding gene and its cognate transcriptional regulator from Azoarcus sp. CIB. Environ Microbiol 10, 474–482. PCC6803 upon nitrogen-starvation: excess NADPH sustains polyhydroxybutyrate accumulation. Metabolites 3, 101–118. Bradford, M. M. (1976). A rapid and sensitive method for the Hendrickx, L., De Wever, H., Hermans, V., Mastroleo, F., Morin, N., Wilmotte, A., Janssen, P. & Mergeay, M. (2006). Microbial ecology of quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Anal Biochem 72, 248–254. Claudel-Renard, C., Chevalet, C., Faraut, T. & Kahn, D. (2003). Enzyme-specific profiles for genome annotation: PRIAM. Nucleic Acids Res 31, 6633–6639. Dunn, M. F., Ramı́rez-Trujillo, J. A. & Hernández-Lucas, I. (2009). Major roles of isocitrate lyase and malate synthase in bacterial and fungal pathogenesis. Microbiology 155, 3166–3175. Edgren, T. & Nordlund, S. (2004). The fixABCX genes in Rhodospirillum rubrum encode a putative membrane complex participating in electron transfer to nitrogenase. J Bacteriol 186, 2052–2060. Erb, T. J., Berg, I. A., Brecht, V., Müller, M., Fuchs, G. & Alber, B. E. (2007). Synthesis of C5-dicarboxylic acids from C2-units involving the closed artificial ecosystem MELiSSA (Micro-Ecological Life Support System Alternative): reinventing and compartmentalizing the Earth’s food and oxygen regeneration system for long-haul space exploration missions. Res Microbiol 157, 77–86. Herter, S. M., Kortlüke, C. M. & Drews, G. (1998). Complex I of Rhodobacter capsulatus and its role in reverted electron transport. Arch Microbiol 169, 98–105. Hoare, D. S. (1962). The photometabolism of [1-14C]acetate and [2-14C]acetate by washed-cell suspensions of Rhodospirillum rubrum. Biochim Biophys Acta 59, 723–725. Hoare, D. S. (1963). The photo-assimilation of acetate by Rhodo- spirillum rubrum. Biochem J 87, 284–301. Ivanovsky, R. N., Krasilnikova, E. N. & Berg, I. A. (1997). A proposed crotonyl-CoA carboxylase/reductase: the ethylmalonyl-CoA pathway. Proc Natl Acad Sci U S A 104, 10631–10636. citramalate cycle for acetate assimilation in the purple non-sulfur bacterium Rhodospirillum rubrum. FEMS Microbiol Lett 153, 399–404. Erb, T. J., Rétey, J., Fuchs, G. & Alber, B. E. (2008). Ethylmalonyl- Khomyakova, M., Bükmez, O., Thomas, L. K., Erb, T. J. & Berg, I. A. (2011). A methylaspartate cycle in haloarchaea. Science 331, 334–337. CoA mutase from Rhodobacter sphaeroides defines a new subclade of coenzyme B12-dependent acyl-CoA mutases. J Biol Chem 283, 32283– 32293. http://mic.sgmjournals.org Kim, M. S., Kim, D. H., Cha, J. & Lee, J. K. (2012). Effect of carbon and nitrogen sources on photo-fermentative H2 production associated with Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 1071 B. Leroy and others nitrogenase, uptake hydrogenase activity, and PHB accumulation in Rhodobacter sphaeroides KD131. Bioresour Technol 116, 179–183. A dehydrogenase by site-directed mutagenesis. FEBS Lett 585, 1317– 1321. Kornberg, H. L. & Krebs, H. A. (1957). Synthesis of cell constituents from Schneider, K., Peyraud, R., Kiefer, P., Christen, P., Delmotte, N., Massou, S., Portais, J. C. & Vorholt, J. A. (2012). The ethylmalonyl- C2-units by a modified tricarboxylic acid cycle. Nature 179, 988–991. Kornberg, H. L. & Lascelles, J. (1960). The formation of isocitratase by the Athiorhodaceae. J Gen Microbiol 23, 511–517. Laguna, R., Tabita, F. R. & Alber, B. E. (2011). Acetate-dependent CoA pathway is used in place of the glyoxylate cycle by Methylobacterium extorquens AM1 during growth on acetate. J Biol Chem 287, 757–766. photoheterotrophic growth and the differential requirement for the Calvin–Benson–Bassham reductive pentose phosphate cycle in Rhodobacter sphaeroides and Rhodopseudomonas palustris. Arch Microbiol 193, 151–154. Segers, L. & Verstraete, W. (1983). Conversion of organic acids to H2 McEwan, A. G. (1994). Photosynthetic electron transport and of bacteriochlorophyll in Rhodospirillum molischianum. J Bacteriol 89, 403–408. anaerobic metabolism in purple non-sulfur phototrophic bacteria. Antonie van Leeuwenhoek 66, 151–164. McKinlay, J. B. & Harwood, C. S. (2010). Carbon dioxide fixation as a central redox cofactor recycling mechanism in bacteria. Proc Natl Acad Sci U S A 107, 11669–11675. Meister, M., Saum, S., Alber, B. E. & Fuchs, G. (2005). L-Malylcoenzyme A/b-methylmalyl-coenzyme A lyase is involved in acetate assimilation of the isocitrate lyase-negative bacterium Rhodobacter capsulatus. J Bacteriol 187, 1415–1425. by Rhodospirillaceae grown with glutamate or dinitrogen as nitrogen source. Biotechnol Bioeng 25, 2843–2853. Sistrom, W. R. (1965). Effect of oxygen on growth and the synthesis St Maurice, M., Cremades, N., Croxen, M. A., Sisson, G., Sancho, J. & Hoffman, P. S. (2007). Flavodoxin : quinone reductase (FqrB): a redox partner of pyruvate : ferredoxin oxidoreductase that reversibly couples pyruvate oxidation to NADPH production in Helicobacter pylori and Campylobacter jejuni. J Bacteriol 189, 4764–4773. Suhaimi, M., Liessens, J. & Verstraete, W. (1987). NH+ 4 -N assimilation by Rhodobacter capsulatus ATCC 23782 grown axenically and non-axenically in N and C rich media. J Appl Bacteriol 62, 53–64. Muhitch, M. J. (1988). Acetolactate synthase activity in developing Sykes, P. J., Burns, G., Menard, J., Hatter, K. & Sokatch, J. R. (1987). maize (Zea mays L.) kernels. Plant Physiol 86, 23–27. Osumi, T. & Katsuki, H. (1977). A novel pathway for L-citramalate Molecular cloning of genes encoding branched-chain keto acid dehydrogenase of Pseudomonas putida. J Bacteriol 169, 1619–1625. synthesis in Rhodospirillum rubrum. J Biochem 81, 771–778. Wang, D., Zhang, Y., Welch, E., Li, J. & Roberts, G. P. (2010). Porter, J. & Merrett, M. J. (1972). Influence of light intensity on reductive pentose phosphate cycle activity during photoheterotrophic growth of Rhodospirillum rubrum. Plant Physiol 50, 252–255. Elimination of Rubisco alters the regulation of nitrogenase activity and increases hydrogen production in Rhodospirillum rubrum. Int J Hydrogen Energy 35, 7377–7385. Ramsay, R. R., Steenkamp, D. J. & Husain, M. (1987). Reactions of Westerfield, W. W. (1945). A colorimetric determination of blood electron-transfer flavoprotein and electron-transfer flavoprotein : ubiquinone oxidoreductase. Biochem J 241, 883–892. Risso, C., Van Dien, S. J., Orloff, A., Lovley, D. R. & Coppi, M. V. (2008a). Elucidation of an alternate isoleucine biosynthesis pathway acetoin. J Biol Chem 161, 495–502. Wischgoll, S., Taubert, M., Peters, F., Jehmlich, N., von Bergen, M. & Boll, M. (2009). Decarboxylating and nondecarboxylating glutaryl- in Geobacter sulfurreducens. J Bacteriol 190, 2266–2274. coenzyme A dehydrogenases in the aromatic metabolism of obligately anaerobic bacteria. J Bacteriol 191, 4401–4409. Risso, C., Methé, B. A., Elifantz, H., Holmes, D. E. & Lovley, D. R. (2008b). Highly conserved genes in Geobacter species with expression Wu, B., Zhang, B., Feng, X., Rubens, J. R., Huang, R., Hicks, L. M., Pakrasi, H. B. & Tang, Y. J. (2010). Alternative isoleucine synthesis patterns indicative of acetate limitation. Microbiology 154, 2589–2599. pathway in cyanobacterial species. Microbiology 156, 596–602. Schaarschmidt, J., Wischgoll, S., Hofmann, H. J. & Boll, M. (2011). Con- version of a decarboxylating to a non-decarboxylating glutaryl-coenzyme 1072 Edited by: P. O’Toole Downloaded from www.microbiologyresearch.org by IP: 88.99.165.207 On: Wed, 14 Jun 2017 15:02:18 Microbiology 161