Genetic Plasticity and Ethylmalonyl Coenzyme A Pathway during Acetate Assimilation in Rhodospirillum rubrum S1H under Photoheterotrophic Conditions

ABSTRACT

Purple nonsulfur bacteria represent a promising resource for biotechnology because of their great metabolic versatility. Rhodospirillum rubrum has been widely studied regarding its metabolism of volatile fatty acid, mainly acetate. As the glyoxylate shunt is unavailable in Rs. rubrum, the citramalate cycle pathway and the ethylmalonyl-coenzyme A (CoA) pathway are proposed as alternative anaplerotic pathways for acetate assimilation. However, despite years of debate, neither has been confirmed to be essential. Here, using functional genomics, we demonstrate that the ethylmalonyl-CoA pathway is required for acetate photoassimilation. Moreover, an unexpected reversible long-term adaptation is observed, leading to a drastic decrease in the lag phase characterizing the growth of Rs. rubrum in the presence of acetate. Using proteomic and genomic analyses, we present evidence that the adaptation phenomenon is associated with reversible amplification and overexpression of a 60-kb genome fragment containing key enzymes of the ethylmalonyl-CoA pathway. Our observations suggest that a genome duplication and amplification phenomenon is not only involved in adaptation to acute stress but can also be important for basic carbon metabolism and the redox balance.

IMPORTANCE Purple nonsulfur bacteria represent a major group of anoxygenic photosynthetic bacteria that emerged as a promising resource for biotechnology because of their great metabolic versatility and ability to grow under various conditions. Rhodospirillum rubrum S1H has notably been selected by the European Space Agency to colonize its life support system, called MELiSSA, due to its capacity to perform photoheterotrophic assimilation of volatile fatty acids (VFAs), mainly acetate. VFAs are valuable carbon sources for many applications, combining bioremediation of contaminated environments with the generation of added-value products. Acetate is one of the major volatile fatty acids generated as a by-product of fermentation processes. In Rs. rubrum, purple nonsulfur bacteria, the assimilation of acetate is still under debate since two different pathways have been proposed. Here, we clearly demonstrate that the ethylmalonyl-CoA pathway is the major anaplerotic pathway for acetate assimilation in this strain. Interestingly, we further observed that gene duplication and amplification, which represent a well-known phenomenon in antibiotic resistance, also play a regulatory function in carbon metabolism and redox homeostasis.

INTRODUCTION

Rhodospirillum rubrum is an alphaproteobacterium that is known for its great metabolic versatility. In particular, its ability to display photoheterotrophic metabolism led the European Space Agency (ESA) to select the S1H strain as a microbe implemented in the MELISSA loop, its microecological regenerative life support system, to recycle volatile fatty acids (VFAs) (1, 2). Purple nonsulfur bacteria are well-studied for their ability to grow under photoheterotrophic conditions using energy from light and various VFAs as carbon and electron sources. VFAs are common fermentation by-products, and among them, acetate is usually one of the most abundant VFAs produced during the acidogenic and acetogenic phases of anaerobic digestion (3, 4). Rs. rubrum S1H belongs to the same group of isocitrate lyase-negative (ICL⁻) organisms as Rhodobacter sphaeroides and Methylobacterium extorquens (5, 6). The absence of this key enzyme implies that C2 compounds, such as acetate, cannot be metabolized through the well-known glyoxylate shunt. In this context, the ethylmalonyl-coenzyme A (CoA) (EMC) pathway was demonstrated to be an efficient anaplerotic pathway for circumventing the absence of the glyoxylate cycle for acetate metabolism in ICL⁻ organisms (7–11). In the EMC pathway, two molecules of acetyl-CoA are condensed into acetoacetyl-CoA for further conversion into crotonyl-CoA. A carboxylation step leads to the formation of ethylmalonyl-CoA, which is then converted into methylsuccinyl-CoA, mesaconyl-CoA, and, finally, 3-methylmalyl–CoA. The latter is converted into propionyl-CoA and glyoxylate. Propionyl-CoA is transformed into succinyl-CoA, which is finally converted into a first molecule of malate, while the condensation of glyoxylate with a third molecule of acetyl-CoA leads to the formation of a second malate molecule.

Nevertheless, the nature of acetate photoassimilation in Rhodospirillum rubrum is still under debate, with the potential involvement of at least two pathways. The citramalate cycle was proposed by Ivanovsky et al. as the condensation of acetyl-CoA and pyruvate with the formation of citramalate, which is finally converted into glyoxylate and propionyl-CoA (12). In contrast, Leroy and collaborators suggested that acetate assimilation in Rs. rubrum S1H involves, at least in part, the ethylmalonyl-CoA pathway (13). A proteomic analysis showed higher abundances of all the enzymes specifically implicated in this pathway. The activity of crotonyl-CoA carboxylase/reductase, a key enzyme in the EMC pathway, was also increased when cells were using acetate as the sole carbon source. In addition, citramalate could be an intermediate metabolite of an alternative pathway for acetyl-CoA in the branched-chain amino acid biosynthetic pathway. In this work, a combination of functional genomics and targeted mutagenesis allowed us to demonstrate the essentiality of the EMC pathway in acetate assimilation. Furthermore, we also highlighted an unexpected long-term adaptation of Rs. rubrum strain S1H to cultivation under acetate conditions characterized by a drastic decrease in the long lag phase of growth usually associated with this culture condition. Analysis of the adapted strains revealed the upregulation of some key enzymes in the EMC pathway associated with a reversible gene duplication and amplification (GDA) phenomenon. Although GDA has often been described in antibiotic resistance or other environmental stresses, the involvement of GDA in carbon metabolism in a bacterium is documented only poorly today.

RESULTS

Higher bicarbonate concentrations reduce the lag phase of growth characterizing acetate assimilation.

The photoheterotrophic growth of Rs. rubrum with acetate as a sole carbon source is characterized by a long lag phase typically lasting more than 5 days under our conditions. Recently, this lag phase has been hypothesized to be linked to a probable acetate-induced intracellular redox imbalance (13). The acetate assimilation has been described to rely on the presence of CO₂ (14, 15), the fixation of which has been described as a central redox cofactor recycling mechanism in bacteria (16). To better understand the growth lag phase as well as the involvement of carbon dioxide fixation in this growth phase, the effect of the bicarbonate concentration in the medium on bacterial growth was tested. The initial bicarbonate concentration (i.e., 3 mM NaHCO₃) in the medium was first doubled without a significant growth phenotype modification (data not shown). Only a higher bicarbonate concentration significantly shortened the lag phase of Rs. rubrum grown with acetate as a sole carbon source, to such an extent that supplementation of the culture medium with 50 mM NaHCO₃ led to complete abolition of the growth lag phase (Fig. 1A).

FIG 1.

Effect of bicarbonate supplementation on the growth of wild type Rhodospirillum rubrum S1H and essentiality of the ethylmalonyl-CoA pathway. (A) Growth monitoring of Rs. rubrum S1H measuring OD₆₈₀ under light anaerobic (filled markers) and dark aerobic (open markers) conditions (cond.) using succinate (blue closed circles/blue open circles; n = 5) or acetate supplemented with 50 mM (red filled squares; n = 5) or 3 mM (green filled triangles/green open triangles; n = 5) bicarbonate as the carbon source(s). Error bars indicate standard deviations. (B) Fitness of the gene set of the ethylmalonyl-CoA pathway measured for the mutant library grown with succinate (blue; n = 3) or acetate (red; n = 3) as the carbon source(s) supplemented with 50 mM bicarbonate. The fitness of acetyl-CoA hydrolase (Rru_A1927) is also presented. Error bars indicate standard deviations. The significance of the difference of fitness between succinate and acetate condition is represented as single asterisks (*; P value < 0.05) or double asterisks (**; P value < 0.01). d, days.

The ethylmalonyl-CoA pathway is essential for acetate photoassimilation in Rs. rubrum.

To determine whether the EMC pathway is necessary for photoheterotrophic assimilation of acetate under our experimental conditions, genome-wide mutant fitness assays were performed using a transposon mutant library with a random bar code transposon-site sequencing (RB-TnSeq) approach. RB-TnSeq is a variation of the transposon sequencing method in which the mutant library is characterized only once to link random DNA bar codes incorporated in transposons to their insertion location. The subsequent fitness assays are characterized by DNA barcode sequencing (BarSeq), allowing high-throughput fitness profiling of a mutant library (17).

The Rs. rubrum mutant library was produced under the least selective conditions for our strain: in rich medium under aerobic conditions without light. This condition was quite different from the experimental photoheterotrophic conditions used in this work (i.e., minimal medium with a single carbon source, anaerobic conditions with light). To discriminate genes specifically involved in acetate photoassimilation from genes involved in the adaptation to minimal medium and photoheterotrophic conditions, we compared the fitness of the library grown in the presence of acetate to the fitness profile obtained using succinate as a control carbon source. The gene fitness values were determined as the average of the strain fitness values measured for the strains carrying insertions in this specific gene (average number of different strains per gene, 27). The strain fitness values were defined as the log₂ of the ratio between the strain abundance after 5 generations and its abundance at the beginning of the experiment under the relevant conditions (17). The experiment was conducted as three independent replicates consisting of independent amplifications of different stocks of the library. The minimum number of read counts per strain was set at 10 to ensure the acquisition of high-confidence data. Only genes for which the fitness value was lower than −0.5 under the acetate condition and significantly different from the fitness value under the succinate condition (P value < 0.05; unpaired t test) were considered to be involved in acetate assimilation. Of the 2,138 genes for which fitness data were obtained, only 28 exhibited fitness that was specifically associated with adaptation to acetate conditions. The complete data set is available as Table S1 in the supplemental material.

Among those 28 acetate-specific genes, the EMC pathway appeared to be of foremost importance since all the genes for which data were obtained presented negative fitness under the acetate condition (Table 1 and Fig. 1B). Supporting the robustness of these results, an average of 20 different mutant strains were found for each gene in the EMC pathway in our mutant library. Individual mutant strains are presented in Table 1. Only 2 genes from the EMC pathway were missing from our mutant fitness assay: the gene encoding ethylmalonyl-CoA/methylmalonyl-CoA epimerase (Rru_A1572) was absent from the data set because of its exclusion due to short size criteria (17), and the gene encoding malyl-CoA thioesterase (Rru_A1200) was not included in the data set because its fitness could be accurately measured in only 2 of the 3 replicates. Nevertheless, the fitness data measured for this gene in these 2 replicates were also negative and specific for the acetate condition. These results are in agreement with data previously obtained in a proteomic analysis (13) and clearly demonstrate that the EMC pathway is required for optimal growth under photoheterotrophic conditions with acetate as a sole carbon source. Recently, a larger gene fitness data set, including 25 bacterial strains and hundreds of sets of experimental conditions, was described (18). Among the hundreds of conditions recorded for the 25 organisms already tested, the genes encoding EMC pathway components were noticed as essential only when acetate was used as a carbon source for Phaeobacter inhibens strain BS107, a heterotroph marine ICL⁻ member of the Rhodobacteraceae cultivated under dark aerobic conditions. These data reinforce the hypothesis that the EMC pathway is involved in acetate assimilation by other ICL⁻ organisms.

TABLE 1.

Gene fitness values and differential protein expression in Rhodospirillum rubrum S1H

Locus	Description	Phaeobacter inhibens BS107						Phaeobacter inhibens BS107
		Protein fold changec	P valuec	Succinate gene fitnessd	Acetate gene fitnessd	P valued	No. of strainse	Acetate gene fitnessf
Ethylmalonyl-CoA pathway
Rru_A1927	Acetyl-CoA hydrolase	1.83	4.2e−3	0.0	2.1	8.7e−5	19	n.a.g
Rru_A2964	MaoC-like dehydratase	0.81	4.9e−1	0.0	−0.8	1.3e−2	7	−1.4
Rru_A3063	Crotonyl-CoA carboxylase/reductase	3.20	2.5e−3	−0.2	−1	1.4e−3	14	−2.7
Rru_A1572a	Ethylmalonyl-CoA/methylmalonyl-CoA epimerase	1.58	8.1e−3	n.a.	n.a.	n.a.	5	n.a.
Rru_A3062	Ethylmalonyl-CoA mutase	8.80	7.9e−5	0.0	−1.0	1.9e−2	35	−2.8
Rru_A3064	Methylsuccinyl-CoA dehydrogenase	2.79	7.0e−3	0.0	−1.2	1.5e−2	26	−3.0
Rru_A1201	Mesaconyl-CoA hydratase	4.46	3.4e−4	0.1	−1.3	6.3e−3	17	−2.7
Rru_A0217	l-Malyl-CoA/b-methylmalyl-CoA lyase	2.08	2.6e−3	0.0	−1.5	1.9e−2	17	−2.5
Rru_A1200b	Malyl-CoA thioesterase	2.31	1.3e−3	n.a.	n.a.	n.a.	6	−1.9
Rru_A0052	Biotin carboxylase	4.75	3.1e−6	0.0	−2.7	2.9e−4	34	n.a.
Rru_A0053	Carboxyltransferase	5.14	4.5e−5	−0.1	−2.6	2.3e−3	21	n.a.
Rru_A2479	Methylmalonyl-CoA mutase	1.98	5.6e−3	−0.1	−3.4	2.4e−3	29	n.a.
Branched-chain amino acid biosynthesis pathway
Rru_A0467	Acetolactate synthase, large subunit	3.12	3.2e−3	−2	−4.0	3.4e−4	33	−2.1
Rru_A0468	Acetolactate synthase, small subunit	3.07	3.5e−3	n.a.	n.a.	n.a.	5	−0.3
Rru_A0469	Ketol-acid reductoisomerase	4.07	5.0e−4	−2.2	−3.5	2.8e−2	20	n.a.
Rru_A1786	Dihydroxy-acid dehydratase	1.62	4.1e−2	−2.4	−3.5	7.0e−2	28	−3.3
Rru_A1040	Leucine dehydrogenase	0.03	1.1e−3	0.0	0.0	9.0e−1	18	1.0
Citramalate cycle
Rru_A06 95	2-Isopropylmalate synthase	1.59	1.2e−2	−0.1	−0.2	4.6e−1	23	1.1

^aCentral sequence too short to be included in the data set based on acceptance criteria used in a previous study (17).

^bNot included in the data set because fitness could be accurately measured in only 2 of 3 of the replicates.

^cProtein fold change is defined as the ratio of the abundance of a protein under acetate conditions to the abundance under succinate conditions. The statistical significance of the results of comparisons of the acetate and succinate conditions was determined with an unpaired t test. Data are from Leroy et al. (13).

^dThe strain fitness is defined as the log₂ value of the ratio between the strain abundance reached after 5 generations and the abundance at time zero (T₀) under the relevant condition. The gene fitness values were calculated by averaging the strain fitness values for each gene. The presented gene fitness values are the average values resulting from three independent fitness assays for each condition. The statistical significance of the results of comparisons of the acetate and succinate conditions was determined with an unpaired t test. Italicized loci are plotted in Fig. 1B. The complete data set is available in the supplemental material.

^eNo. of strains, the number of strains with one independent central transposon insertion per gene in the mutant library.

^fThe gene fitness values for Phaeobacter inhibens BS107 were obtained from the 3H35 experiment set available on the Fitness Browser webtool (http://fit.genomics.lbl.gov) and were determined under dark aerobic conditions with acetate as the carbon source.

^gn.a., not assayed.

In addition to the EMC pathway, the potential involvement of acetolactate synthase and part of the branched-chain amino acid (isoleucine, leucine, and valine [ILV]) biosynthesis pathway in acetate photoheterotrophic assimilation has been previously reported (13). Interestingly, the mutant fitness assay data corroborated these observations since acetolactate synthase (large subunit), the ketol-acid reductoisomerase, and the dihydroxy-acid dehydratase were more essential for acetate assimilation than for succinate assimilation, even if the level of statistical significance of the data obtained for the latter did not reach that specified by the significance criteria applied here. The actual role of the ILV biosynthesis pathway in acetate assimilation remains unclear.

The mutant fitness assay also revealed an unexpected result for acetyl-CoA hydrolase/transferase (Rru_A1927), the mutation of which appeared to be beneficial under acetate conditions. Previous proteomic data have shown that this protein is specifically upregulated in the presence of acetate, leading to consideration of the possibility that its transferase activity is mainly involved in acetate activation (13). The surprising positive fitness results from the strains carrying the mutated Rru_A1927 led us to consider that this enzyme could be acting in the direction of succinyl-CoA production, transferring CoA from acetyl-CoA to succinate. This reaction would be competing with acetate activation, which would explain the beneficial mutation of this enzyme observed in the mutant fitness assay. The ambiguous activity of Rru_A1927 and its involvement in acetate activation are therefore more complex than expected.

To validate these data and the involvement of the EMC pathway in acetate assimilation in Rs. rubrum, we constructed a knockout mutant of the ccr gene (Rru_A3063) encoding crotonyl-CoA carboxylase/reductase (Fig. 2A and B). This enzyme is known to connect the C4 branch and C5 branch of the pathway via NADPH-dependent reductive carboxylation of crotonyl-CoA to ethylmalonyl-CoA (9) and has been targeted because of its involvement (unique, to our knowledge) in the EMC pathway. As expected, the Δccr::Km^r (kanamycin resistance) mutant did not present any significant growth phenotype modifications while growing (both aerobically and anaerobically) in succinate-containing medium (Fig. 2C). In contrast, no growth was detected for the Δccr::Km^r mutant when acetate was used as the sole carbon source, regardless of the bicarbonate level (3 or 50 mM) provided in the medium. Moreover, the growth of the Δccr::Km^r strain was also impaired under aerobic conditions. Regardless of which culture conditions were used, knockout of the ccr gene led to a total disruption of acetate assimilation, thus demonstrating that the EMC pathway is absolutely required for acetate assimilation by Rs. rubrum.

FIG 2.

Targeted mutagenesis of the crotonyl-CoA carboxylase/reductase gene. (A) Scaled diagram of the ccr gene region in both the wild-type strain and the Δccr::Km^r strain. “H.R.1” and “H.R.2” represent the homologous regions used for the final double homologous recombination between the genomic DNA and cloning vector, and “Km^R Cass.” represents the kanamycin antibiotic resistance cassette. (B) PCR amplification targeting the ccr gene and the Km^r cassette are indicated as “Fg^t. 1” and “Fg^t. 2,” respectively, in the Δccr::Km^r strain. Primers, represented as colored arrows, are listed in Table 3. Ctrl. Pos., positive control; Ctrl. Neg., negative control. (C) Monitoring of the growth of Rhodospirillum rubrum S1H strain Δccr::Km^r by measuring OD₆₈₀ under light anaerobic (filled markers) and dark aerobic (open markers) conditions using succinate (blue closed circles/blue open circles; n = 5) or acetate supplemented with 50 mM (red filled squares; n = 5) or 3 mM (green filled triangles/green open triangles; n = 5) bicarbonate as the carbon source(s). Error bars indicate standard deviations.

Repeated acetate exposure induces an acclimatization involving a reversible variation in the ccr gene copy number.

During the process of site-directed mutagenesis, we very often obtained strains that were able to grow on selective medium (kanamycin resistance) but that also remained positive for PCR amplification of the deleted ccr gene (“Fg^t 1”; see Fig. S1A and B in File S1 in the supplemental material). Successive spreading of the strains on the plate was realized to ensure purity of the clonal population. Those strains were also positive for PCR amplification of the kanamycin resistance cassette (“Fg^t 2”; Fig. S1A and B in File S1). To test if those strains arose from the insertion of the kanamycin resistance cassette into alternative regions of the genome, we amplified the fragment overlapping the Rru_A3062 locus upstream of the H.R.1 region and the kanamycin resistance cassette (“Fg^t 3”; Fig. S1A and B in File S1) or the ccr gene itself (“Fg^t 4”; Fig. S1A and B in File S1). Positive PCR amplification of both Fg^t 3 and Fg^t 4 ensured that both the ccr gene and the kanamycin resistance cassette were localized at the expected position in the genome, suggesting that this region existed in this strain in multiple copies before insertion of the kanamycin resistance cassette. To test this hypothesis, we determined the relative ccr copy numbers in the original wild-type strain used for targeted mutagenesis in the ccr-negative Δccr::Km^r strain (the true Δccr::Km^r mutant) and in the ccr-positive Δccr::Km^r strain (used here as a calibrator) through quantitative PCR (qPCR) (validation of the procedure is available as text in the supplemental material). This experiment revealed that the wild-type strain used for site-directed mutagenesis effectively contained twice as many ccr copies as the ccr-positive Δccr::Km^r strain (see Fig. S1C in File S1).

In accordance with Ohno's model of molecular evolution (19), gene duplication is an important evolutionary tool in all three kingdoms of life to produce new genes. The gene copy number variation has also been reported in the last few decades to be a regulatory response-like adaptation to environmental stresses, particularly in antibiotic resistance in bacteria (20). Given the key role played by the crotonyl-CoA carboxylase/reductase in acetate assimilation and the copy number variation observed in our wild-type strain, we decided to further explore the acclimatization to acetate through a long-term acetate repeated batch cultivation experiment. After approximately 30 generations (3 generations per batch), a shortening of the lag phase, evaluated by monitoring the increase in optical density at 680 nm (OD₆₈₀) during the first 48 h of the batch cultivation, was observed for the culture maintained under acetate conditions compared with the data obtained at the beginning of the fed batch experiment. After 40 generations, a statistically significant difference between the growth rates on succinate and acetate was no longer observed (Fig. 3A). This long-term acclimatization to acetate clearly did not result from a rapid specific metabolic response to a modification of the environmental conditions but was highly similar to the extensively described antibiotic resistance acquisition previously reported to occur through gene duplication and amplification (GDA) (20).

FIG 3.

Effect of acetate acclimatization on the lag phase and gene copy number. (A) Cultures were maintained in fed batch cultures with relevant carbon sources for several generations (based on the cell doubling time), and the lag phase was measured by monitoring OD₆₈₀ during the first 48 h. The experiment was performed using both acetate (black) and succinate (white) as carbon sources. Error bars indicate standard deviations. Significant differences between samples are represented as single asterisks (*; P value < 0.05) or double asterisks (**; P value < 0.01). (B) The relative ccr copy number was measured by qPCR in monoclonal cultures of acetate-acclimatized cells (white) or nonacclimatized cells (black). A single dagger (†) and a double dagger (††) denote the samples used as calibrators for monoclonal cultures, with 1 to 5 and 6 to 10 relative copy number computations, respectively. Error bars indicate standard deviations. Values represent averages of results from six technical replicates, except for culture 6, with only four replicates. (C) The reversibility of the acetate acclimatization was monitored by cultivating acetate-acclimatized (AA; n = 3) Rs. rubrum in acetate (black, selective conditions favoring gene duplication) and succinate (white, nonselective conditions) media after 9, 18, and 27 generations. The ccr gene copy number was determined by qPCR using nonacclimated strains that were maintained under nonselective (succinate) conditions as a calibrator for copy number normalization. Error bars indicate standard deviations. Values represent averages of results from six technical replicates, except for sample AA1 of generation 27 in succinate, with only five replicates. (D) Growth of acetate-acclimatized (green open circles/green closed circles; n = 3), nonacclimatized (black open circles/black closed circles; n = 3), and deacclimatized (blue open circles/blue closed circles; n = 3) cells under succinate (open circles) and acetate (filled circles) conditions was monitored by measuring OD₆₈₀. Error bars indicate standard deviations. Significant differences between samples are represented as single asterisks (*; P value < 0.05) or double asterisks (**; P value < 0.01).

To evaluate if the acetate acclimatization resulted from a modification of the copy number of genes encoding enzymes of the EMC pathway, we compared the ccr gene copy number in clonal populations obtained from acetate-nonacclimatized (AnA) and acetate-acclimatized (AA) strains. Interestingly, all 10 independent clones obtained from the AA strains contained approximately 3 times more ccr gene copies (Fig. 3B) than the AnA control strains, suggesting that the observed acetate acclimatization process results from a GDA phenomenon. The occurrence of a GDA phenomenon as an adaptation of the central carbon metabolism to a specific carbon source has been poorly documented to date.

Gene duplication and amplification are often linked to the reduced fitness of the strains seen under nonselective conditions due to the gene dosage (with a fitness cost depending more on the genetic content than on the size of the duplicated region [21]) or to protein overproduction (with probable additional deleterious effects on protein folding and chaperone function [20, 22]). Nevertheless, under selective conditions, GDA offers advantages such that the gene copy number under conditions of extreme GDA can reach several hundred (23, 24). Here, we observed a surprisingly stable number of ccr copies, with a maximum of 3 to 4 copies. This limitation most likely reflects a balance between the advantages and disadvantages caused by the ccr GDA, allowing rapid growth for a limited fitness cost.

The homologous recombination occurring between identical amplified genes constitutes the inherent instability of gene duplication and amplification (21). When acclimatized cells were again grown under succinate conditions, the disruption of the selective pressure favoring GDA and induced by acetate led to the counterselection of the subpopulation containing amplified regions of the genome due to the fitness cost of a high gene copy number. We observed, in less than 30 generations, a decrease in the ccr gene copy number to the initial copy number (Fig. 3C). Therefore, the so called “deacclimatized cells” recovered an acetate-nonacclimatized phenotype characterized by a long lag phase (Fig. 3D).

A 62-kb fragment is amplified under acetate acclimatization conditions.

In response to the GDA phenomenon, the size of duplicated and amplified regions is known to range from kilobases to megabases. To determine whether the region amplified under acetate conditions was limited to the ccr gene or could extend to a wider region of the genome, we first used a proteomic approach to search for a global fold change shift in enzymes encoded by genes located in the neighborhood of ccr. Proteins from five biological replicates of AA cells cultivated in acetate or in succinate were identified and quantified using a SWATH acquisition strategy. The data (Table S2) were compared to the data previously obtained from nonacclimatized strains (13). To evaluate the impact of GDA on protein abundance at the genome scale but also with the aim of reducing the impact of punctual fold change modifications, we plotted the median of a sliding window of 30 quantified proteins over the whole quantified proteome (Fig. S2 in File S1). Since we were also trying to evaluate the spread of the amplified region, we also plotted the median of a sliding window of 70 genes covering an average amount of 30 quantified proteins over the whole Rs. rubrum genome (Fig. 4A). In both cases, we found a cluster of enzymes that were specifically upregulated in acetate-acclimatized cells, ranging from Rru_A3004 to Rru_A3115, corresponding to a 132-kb fragment centered on ccr.

FIG 4.

Impact of acetate acclimatization on the Rhodospirillum rubrum S1H proteome and genome. (A) The results of the proteomic analysis performed on acetate-acclimatized cells (red) were compared with previously obtained results determined with nonacclimatized cells (black) by computing the fold change shift in nearby enzymes using the median of a sliding window over the complete ordered genome (70 genes and an average of 30 quantified proteins per window). (B) Whole-genome sequencing of acetate-acclimatized and nonacclimatized Rs. rubrum genomes using Illumina short read methods. The average read depth coverage was measured for 1,000 (blue) and 10,000 (black) nucleotide windows and normalized to the expected read count for each sequence (average read count across the whole genome, excluding the 60-kb amplified region).

To confirm this result, the genomes of an acclimatized strain and a nonacclimatized strain were sequenced. The average read depth coverage was measured for windows of 1,000 and 10,000 nucleotides across the whole genome and clearly showed an amplified fragment between positions 3496500 and 3558300 of the reference genome (PubMed identifier [ID]: NC_007643.1) (Fig. 4B). The region ranging from Rru_A3037 to Rru_A3092 was present in 3 copies in the genome, confirming the qPCR copy number evaluation. Surprisingly, this fragment was smaller than expected based on the proteomic observation.

To better understand the GDA phenomenon observed here, we searched for main mechanisms that have already been shown to be involved in duplication and amplification (25). In the IS Finder (26) database, among the 7 full and 22 partial insertion sequences (IS) reported for Rs. rubrum ATCC 11170, none were detected in the region of the genome affected by duplication and amplification as described here. We reached the same conclusions concerning rRNA operons, short tandem repeats, and interspersed repeats. Further investigations should focus on explaining the development of the duplication and amplification.

Acetate acclimatization provides improved fitness during the early growth phase.

Since the amplified region was not restricted to the ccr gene and since the nature of the influence of the duplication on the locus expression was still not completely clear, we further explored genes included in the region centered on the amplified genome fragment for which a global fold change shift had been observed in the proteomic data (Table S2 and Fig. 4A). Considering, therefore, a larger region of the genome ranging from locus Rru_A3000 to Rru_A3120, we inspected the data from the mutant fitness assay to determine the gene included in the amplified gene cluster that could be responsible the observed acetate acclimatization. Reliable fitness data were found for 91 of the 121 genes (Table S3). Only 6 of these 91 genes presented a negative fitness value lower than −0.5 under the acetate condition. Excluding genes for which the fitness value was also below −0.5 under the succinate condition to avoid taking into account genes related to adaptation to photoheterotrophic conditions, only genes belonging to the EMC pathway (Rru_A3062 to Rru_A3064) were specifically involved in adaptation to acetate conditions.

Interestingly, those three genes (Rru_A3062 to Rru_A3064) also presented the highest fold change values in our proteomic analysis and thus represented good candidates to explain the improved fitness in our acclimatized strain. The crotonyl-CoA reductase/carboxylase (Rru_A3063) abundance was nearly three times higher in acclimatized strains (fold change, 10.3; P value, 3.6e−4) than in the nonacclimatized strains (fold change, 3.2; P value, 2.5e−3). Interestingly, only the lag phase was shortened in AA cells, and no significant differences were observed regarding the growth rate or the total biomass. This result suggests that the only advantage provided by the amplification phenomenon concerns growth conditions involving high light stress. A higher abundance of CCR would result in enhanced reductive carboxylation of crotonyl-CoA with a higher recycling rate of reduced cofactors produced during the lag phase, finally allowing the cell to begin to grow at an earlier time point. The two other EMC pathway-related enzymes (Rru_A3062 and Rru_A3064) were also quantified with higher abundance in the acclimatized strain (fold change values, 7. 9 and 6.4, respectively). Nevertheless, the benefits of their amplification seemed less obvious and were probably a consequence of the clustering of these 3 genes.

The 3-hydroxybutyryl–CoA dehydrogenase (Rru_A3079) (S-specific enzyme) is a fourth protein that is likely to be related to acetate metabolism in acclimatized strains. This enzyme was clearly overexpressed (fold change, 2.5) and could complement the acetoacetyl-CoA reductase (Rru_A0273) that was retained as an acetate-specific candidate to reduce 3-acetoacetyl-CoA into (R)-3-hydroxybutyryl–CoA. Moreover, the reaction catalyzed by this enzyme consumed reduced cofactors and could thus help to maintain the cellular redox balance, possibly contributing to the fitness improvement of the acetate-acclimated strains.

Interestingly, enzymes catalyzing the subsequent conversion of mesaconyl-CoA in the ethylmalonyl-CoA pathway were no longer upregulated in acetate-grown acclimatized strains, in contrast to enzymes that converted acetyl-CoA into crotonyl-CoA, the abundance of which remained unchanged (Table 2). This result may indicate that at least one of the enzymes that were differentially expressed in acetate-acclimatized strains functions as a rate-limiting enzyme in the EMC pathway. The overabundance of these enzymes in acclimatized strains could thus have reduced the need for downstream enzyme overexpression.

TABLE 2.

Differential protein expression in acetate-acclimatized strain of Rhodospirillum rubrum S1H^a

Uniprot accession no.	Locus tag	P valueb	Protein fold changeb (acetate/succinate)	No. of identified peptidesc	Description
Q2RQ36	Rru_A2964	3.3e−2	0.83	5	MaoC-like dehydratase
Q2RPT7	Rru_A3063	3.6e−4	10.29	6	Crotonyl-CoA carboxylase/reductase
Q2RU23	Rru_A1572	1.8e−1	1.19	3	Ethylmalonyl-CoA/methylmalonyl-CoA epimerase
Q2RPT8	Rru_A3062	1.4e−3	7.89	2	Ethylmalonyl-CoA mutase
Q2RPT6	Rru_A3064	3.4e−4	6.37	6	Methylsuccinyl-CoA dehydrogenase
Q2RV43	Rru_A1201	2.1e−4	1.34	5	Mesaconyl-CoA hydratase
Q2RXX3	Rru_A0217	3.3e−4	0.68	6	l-Malyl-CoA/b-methylmalyl-CoA lyase
Q2RV44	Rru_A1200	5.1e−4	1.28	6	Malyl-CoA thioesterase
Q2RYD8	Rru_A0052	1.2e−6	1.83	6	Biotin carboxylase
Q2RYD7	Rru_A0053	4.4e−6	1.83	6	Carboxyltransferase
Q2RRG6	Rru_A2479	3.4e−1	0.85	5	Methylmalonyl-CoA mutase

^aThe complete data set is available in File S1 in the supplemental material.

^bThe protein fold change is defined as the ratio of the abundance of a protein under acetate condition to the abundance under succinate condition. The statistical significance of results of comparisons of the acetate and succinate conditions was determined with an unpaired t test.

^cNo. of identified peptides, number of identified peptides used for quantification with a confidence level higher than 99%; maximum number = 6.

The involvement of these genes in the acclimatization process requires further validation using expression vectors; however, such analyses are beyond the scope of this paper.

DISCUSSION

The present study clearly demonstrated that the ethylmalonyl-CoA pathway is the main driver of acetate metabolism in Rhodospirillum rubrum. This conclusion was drawn from a RB-TnSeq-based mutant fitness assay performed in biological triplicate and from a site-directed mutagenesis experiment targeting the key enzyme of the EMC pathway, the crotonyl-CoA reductase/carboxylase associated with the proteomic analysis. While the proteomics data showed that only the EMC pathway was operational, here, the mutant fitness profiling of acetate assimilation in Rs. rubrum under photoheterotrophic conditions highlighted a strong dependency on EMC pathway enzymes.

Although acetate assimilation has been reported to be dependent on the presence of bicarbonate in the culture medium (14), small amounts are sufficient to allow complete assimilation of the VFA. This phenomenon is in contrast to results obtained with propionate and butyrate, for which a net CO₂ uptake is required (27, 28). Net CO₂ release is even observed during photoheterotrophic growth on acetate (28, 29). Nevertheless, we observed that a larger amount of bicarbonate in the medium translates into a strong decrease of the still poorly understood lag phase that is usually associated with acetate assimilation under photoheterotrophic conditions in Rs. rubrum. The redox balance is an essential parameter of photoheterotrophic growth on reduced substrates such as acetate, particularly during the early growth phase. A higher sensitivity to light stress with acetate as the carbon source which was assumed to result from unbalanced redox homeostasis has been previously reported (13). Bicarbonate fixation through the EMC pathway has been reported as a redox balancing mechanism (29, 30). Therefore, the use of a higher bicarbonate concentration could be an efficient way to stimulate the reductive carboxylation of crotonyl-CoA, allowing in turn a higher level of consumption of reducing equivalents and a better balance of the redox potential, finally enabling earlier cell growth. Interestingly, our mutant with respect to crotonyl-CoA reductase/carboxylase was unable to grow with acetate under dark aerobic conditions. Additionally, a recent study also reported the use of the EMC pathway under these conditions (31), in which the use of oxygen as an electron acceptor abolishes the coupling of carbon metabolism and redox homeostasis. In contrast with Rs. rubrum, Rhodobacter sphaeroides does not require the presence of bicarbonate in the medium in order to grow photoheterotrophically with acetate as the sole carbon source. Assuming that both strains use the EMC pathway for acetate assimilation, this difference between Rh. sphaeroides and Rs. rubrum is puzzling. The dependency of acetate assimilation on the presence of bicarbonate in the medium has been known for a very long time (14, 15); it could be of interest to demonstrate that it is effectively the case under our specific conditions. In addition, it could be interesting to test this dependency with our acetate-acclimatized strain.

The key role played by the EMC pathway during early growth was also emphasized by acetate acclimatization. Surprisingly, repeated exposure to acetate led to the duplication of a 62-kb region of the genome that phenotypically translated into a drastic decrease in the lag phase. The duplication and amplification phenomenon in prokaryotes is known to confer the phenotypic advantages needed to adapt to extreme conditions or episodic unfavorable situations. It has been extensively described as a means to confer or increase resistance to environmental stresses such as antibiotics (20) or heavy metals (32). The gene amplification could also be involved in adaptation to environments with limited nutrient concentration, usually by increasing limiting substrate uptake or processing (33–35). Here, we report the amplification of central carbon metabolism actors in response to a redox imbalance. Indeed, amplification is known to confer selectable advantages through increased overexpression of genes of interest. Here, the acetate acclimatization-related amplicon included the crotonyl-CoA reductase/carboxylase, which is involved in redox homeostasis as discussed above. A higher abundance of this enzyme would allow the cell to manage redox imbalances even with small amounts of bicarbonate in the medium, probably through improved mobilization of the endogenously released carbon dioxide. This observation has to be considered in relation to the low affinity of this enzyme for CO₂. As mentioned by Erb et al. (36), with a K_m of 14 mM for bicarbonate, the activity of this enzyme could be low under low-bicarbonate conditions such as the concentration used in this study (3 mM bicarbonate). Considering this low affinity, the increase in the abundance of the enzyme could adequately increase the metabolic flux through the EMC pathway to compensate for the low bicarbonate level.

The 3-hydroxybutyryl–CoA dehydrogenase (Rru_A3079) is also present in the amplicon. Since this enzyme was downregulated in our nonacclimatized strains cultivated with acetate compared to succinate, the conversion of acetoacetyl-CoA into 3-hydroxybutyryl–CoA was assigned to the acetoacetyl-CoA reductase (Rru_A0273) (13). The situation is reversed through acetate acclimatization: Rru_A3079 was clearly overexpressed, while Rru_A0273 was slightly downregulated. Interestingly, these enzymes are specific for different isomers of 3-hydroxybutyryl–CoA (Fig. 5). The formation of (S)-3-hydroxybutyryl–CoA was catalyzed by Rru_A3079, while Rru_A0273 was specific to the (R)-isomer. In contrast, the first steps of the EMC pathway are shared with the poly(3-hydroxybutyrate) biosynthesis pathway, and the reduced equivalent-consuming conversion of acetoacetyl-CoA into 3-hydroxybutyryl–CoA has been previously proposed to have a redox balancing effect (13, 31). However, only the (R)-hydroxybutyryl–CoA monomers can be polymerized (37, 38). Thus, the use of Rru_A3079 in addition to Rru_A0273 might allow acetate-acclimatized strains to benefit from a higher recycling rate of reduced cofactors while limiting the carbon flux to poly-(3)-hydroxybutyrate (PHB) storage and focusing on biosynthesis precursors (Fig. 5).

FIG 5.

Impact of acetate acclimatization on shared steps of the EMC and PHB biosynthesis pathways. The figure presents a schematic view of the first steps of the ethylmalonyl-CoA (EMC) pathway shared with the poly-(3)-hydroxybutyrate biosynthesis pathway. CoASH, (S)-3-hydroxybutyryl–CoA; ECMpw, ECM pathway.

Another interesting result regarding the amplification phenomenon was the significant increase in the abundance of nearly all proteins encoded by the amplified region. The literature usually focuses on the gene of interest in the amplified region, but to our knowledge, only a few references show an upregulation of gene expression affecting the entire amplicon in a prokaryote (39) or even in the context of aneuploidy in eukaryotes (40). Moreover, the upregulation observed in our study also encompassed amplicon satellite genes. The global upregulation of the amplicon remains misunderstood and will require further detailed analyses. Nevertheless, this upregulation of a large number of proteins is very often embedded with a function that is not related to stress, which could explain the rapid decrease in gene copy number after removal of the selective pressure, as often reported in the literature (21, 41–44).

Altogether, these results clearly demonstrate the requirement of the EMC pathway for acetate assimilation in Rs. rubrum. In addition, the previously reported involvement of this pathway in dissipating the excess reducing power and maintaining the redox homeostasis allowing photoheterotrophic growth on acetate (29, 30) is strongly supported by the present observations. Finally, this work showed that GDA is a more widespread phenomenon than expected, also occurring in central carbon metabolism regulation.

MATERIALS AND METHODS

Bacterial strain, medium composition, and cultivation conditions.

The bacterial strain used in this study was Rhodospirillum rubrum S1H (ATCC 25903). The growth medium used for photoheterotrophic culture conditions contained MOPS (morpholinepropanesulfonic acid; 100 mM), KH₂PO₄ (3.5 mM), K₂HPO₄ (3 mM), FeSO₄ · 7H₂O (0.07 mM), Na₂SO₄ (4 mM), EDTA (0.07 mM), MnCl₂ · 4H₂O (0.05 mM), MgSO₄ · 7H₂O (0.08 mM), CaCl₂ · H₂O (0.34 mM), NiSO₄ · 6H₂O (1.9 mM), ZnSO₄ · 7H₂O (0.35 mM), CuSO₄ · 5H₂O (0.02 mM), H₃BO₃ (1.6 mM), and Na₂MoO₄ · 2H₂O (0.2 mM). This basal medium was supplemented with acetic acid (62 mM) as the carbon source, NH₄Cl (35 mM) as the nitrogen source, and biotin (0.06 mM). Following the establishment of the culture conditions, the medium was supplemented with 3 or 50 mM NaHCO₃. The carbon source used in the control condition was succinate (provided as succinic acid) (31 mM) at the same net carbon concentration. The pH was adjusted to 6.9.

The growth was monitored by optical density measurements at 680 nm. Under anaerobic phototrophic conditions, Rs. rubrum was grown in 50-ml penicillin flasks with a working volume of 40 ml. The flasks were hermetically sealed, and pure nitrogen was used to purge oxygen from the upper gas phase. The cultures were inoculated at a starting OD₆₈₀ of between 0.5 and 0.7 and incubated using 50 μmol · m⁻² · s⁻¹ of light intensity supplied by halogen lamps (Sencys) (10 W; 100 lumens; 2,650 K) at 30°C with rotary shaking at 150 rpm. Under dark aerobic conditions, Rs. rubrum cells were grown in 50-ml penicillin flasks sealed with aluminum foil with a working volume of 15 ml. The cultures were inoculated starting at an OD₆₈₀ of between 0.1 and 0.2 and incubated at 30°C with rotary shaking at 150 rpm. Three clonal biological replicates were used for each culture condition.

Mutant library and mutant fitness assay.

The mutant library was produced following the protocol described by Wetmore et al. (17). Briefly, the mutant library was produced by conjugation of Rhodospirillum rubrum S1H with the APA752 library (diaminopimelic acid [DAP] auxotrophic Escherichia coli donor strain WM3064 transformed with the randomly bar-coded mariner transposon delivery vector pKMW3). The strains were combined at a 1:1 ratio and incubated for 6 h at 30°C on 0.45-μm-pore-size nitrocellulose filters overlaying SMN (supplemented malate-ammonium) (45) agar plates containing DAP. The conjugated cells were resuspended and plated on SMN agar plates containing 50 μg/ml kanamycin and incubated at 30°C for 2 days. The resulting kanamycin-resistant mutant library was harvested, resuspended in SMN with 50 μg/ml kanamycin, incubated at 30°C, and, when an OD₆₈₀ of 1.0 was reached, divided into aliquots and placed into glycerol −80°C freezer stocks.

The genomic DNA of the mutant library was extracted with a QIAamp DNA minikit (Qiagen), and the transposon insertion sites were determined and linked to random bar codes by TnSeq following the protocol described by Wetmore et al. (17). Considering only strains of the mutant library with insertions lying within the central 10% to 90% of a gene, 86,027 insertions were found in total. These insertions covered 3,167 of the 3,838 protein-coding genes. A mean of 27.0 different mutant strains (median, 19) was found per protein.

The mutant fitness assays were performed in triplicate starting with three different glycerol stocks of the mutant library. Each aliquot of the mutant library was independently amplified in SMN with 50 μg/ml kanamycin under the standard photoheterotrophic conditions described above. The amplified library was rinsed and resuspended in defined medium to reach a starting OD₆₈₀ of 0.1. The biomass was harvested at time zero (T₀) and after 5 generations (1.5 ml of culture per sample), and the genomic DNA was extracted with a QIAamp DNA minikit (Qiagen). The concentration and purity of the extracted DNA were assessed with a BioSpec-Nano spectrophotometer (Shimadzu Biotech), and if required, the samples were concentrated and purified with a Vivacon 500 kit (Sartorius) before storage at −80°C.

We performed DNA barcode sequencing (BarSeq) as described by Wetmore et al. (17) to quantify the bar codes and, consequently, the abundance of each mutant under each experimental condition. The strain fitness was determined as the normalized log₂ ratio of bar code counts between samples after 5 generations and the “time-zero” reference samples. Gene fitness was calculated as the weighted average of the individual strain fitness values for a given gene, as described by Wetmore et al. (17).

Construction of a ccr knockout strain of Rhodospirillum rubrum S1H.

The ccr knockout strain (Δccr::Km^r) of Rhodospirillum rubrum was obtained by double homologous recombination of the genomic DNA with a cloning vector containing a kanamycin resistance cassette flanked with homologous regions located upstream and downstream of the targeted gene. To construct the cloning vector, we amplified four different fragments: fragments 1 and 2, representing homologous regions 1 (HR1) and 2 (HR2) from Rs. rubrum genomic DNA; fragment 3, representing the kanamycin resistance cassette (Km^r) from the pACYC177 plasmid; and fragment 4, representing the suicide cloning vector from the pKNOCK-Gm plasmid (46). The Rs. rubrum genomic DNA was extracted with a QIAamp DNA minikit (Qiagen). The pACYC177 and pKNOCK-Gm plasmids were isolated from E. coli MG1655 and K-12 MA8 strains, respectively, using a Wizard Plus SV Minipreps DNA purification system (Promega). All the purification steps were conducted according to the manufacturer's instructions. The primers used for the amplifications are listed in Table 3. Amplified DNA insertions were assembled using a GeneArt seamless cloning system (Invitrogen) following the manufacturer's recommendations. The resultant construct was used to transform the E. coli BW29427 strain, a DAP auxotroph donor strain (K. Datsenko and B. L. Wanner, unpublished data), by heat shock. The transformed colonies were selected on SMN agar plates supplemented with kanamycin (50 μg · ml⁻¹) and DAP (0.3 mM). The cloning vector was finally transferred into Rs. rubrum by conjugation performed with the transformed E. coli BW29427 strain at a 1:1 ratio of the two strains. The Δccr::Km^r strains were isolated on SMN agar plates supplemented with kanamycin (50 μg · ml⁻¹) and screened by PCR for the kanamycin resistance cassette.

TABLE 3.

Oligonucleotides used for targeted mutagenesis and qPCR amplification^a

Source	Nameb	Sequence, 5′–3′c
E. coli K-12 MC1061
pACY177	Km_Fw	TCTCTGATGTTACATTGCAC
	Km_Rv	CAGCGTAATGCTCTGC
	LKm_Rv	GATTGTCGCACCTGATTGCC
E. coli K-12 MA8
pKNOCK-Gm	pKNOCK_Fw	CATAAGCCTGTTCGGT
	pKNOCK_Rv	CGGGGGATCCACTAG
Rhodospirillum rubrum S1H
Genomic DNA	ccrHR1_Fw	AGAACTAGTGGATCCGGCGTTCTTTGACAAGACCA
	ccrHR1_Rv	AATGTAACATCAGAGAGTTGAAACCTCTAGGCCGCT
	ccrHR2_Fw	GCAGAGCATTACGCTGAAGAGGTGCGCAAGTTCG
	ccrHR2_Rv	ACCGAACAGGCTTATGTCGCTCAGGTGCCATTC
	LKm_Fw	TATCGGCGCCTTCGGTGTAT
	Ccr_Fw	GTCAACTACAACGGGATCTGGG
	Ccr_Rv	GAAATCGCGCTTGGTCTCATC
	Glts_Fw	GATCGGCTATTGCAACTGCG
	Glts_Rv	CAGATCCTCCATCGACTGGC

^aAll the oligonucleotides were designed using Primer3Plus v2.4.0 and checked for specificity using Snapgene Viewer software (GSL Biotech) and for the formation of dimers and secondary structures using the Multiple Primer Analyzer webtool (ThermoFisher Scientific) and were obtained from Eurogentec S.A.

^bFw, forward primer; Rv, reverse primer.

^cNucleotide sequences based on the sequences of the pACY117 plasmid (GenBank accession no. X06402.1), the pKNOCK-Gm plasmid, and Rhodospirillum rubrum strain ATCC 11170 (NCBI reference sequence; NC_007643.1). The underlined sequences are end-terminal homology sequences required for GeneArt Seamless cloning.

Genomic DNA extraction and relative gene copy number measurements.

To avoid any gene copy number bias due to growth phase variations under the different culture conditions, cells were harvested during the early plateau phase when the OD₆₈₀ reached 4.5 to 5.0. For each condition, genomic RNA-free DNA was isolated from 1 ml of bacterial culture using a QIAamp DNA minikit (Qiagen) according to the manufacturer's instructions, including RNase A digestion. The eluted DNA samples were rinsed and concentrated using Vivacon 500 (Sartorius) centrifugal units. The concentration and quality of the isolated DNA samples based on UV spectra were measured on a BioSpec-Nano (Shimadzu) microvolume spectrophotometer. The samples were adjusted to a concentration of 1 ng/μl.

The relative gene copy number measurements were based on the initial abundance of a target gene compared with a reference gene determined by qPCR. The target gene was the ccr (Rru_A3063) gene encoding crotonyl-CoA reductase. The reference gene was the glts (Rru_A0019) gene encoding the glutamate synthase (NADPH) large subunit. The glts gene was selected as a housekeeping gene that presented a unique copy number in the genome (in comparison to the commonly used 16S genes) located at a site distant (±850,000 bp) from the region suspected to have been duplicated. The sequences of the primers are listed in Table 3. The qPCR was performed using a StepOnePlus real-time PCR system (Applied Biosystems). Each reaction mixture contained 10 μl of 2× SYBR green PCR master mix (Applied Biosystems), 0.2 μl of 100 μM forward and reverse primer, 5 μl of template DNA, and 4.6 μl of PCR-grade water to reach a total volume of 20 μl/well. Four serial dilutions of each DNA sample were used to achieve 2.5, 1.25, 0.625, and 0.312 ng of DNA per reaction volume. The reaction mixtures were prepared on ice and distributed on a MicroAmp Fast optical 96-well reaction plate (Applied Biosystems) sealed with MicroAmp optical adhesive film (Applied Biosystems). The thermocycling parameters consisted of an initial denaturation step at 95°C for 10 min, followed by 40 amplification cycles performed at 94°C for 30 s, 67°C for 30 s, and 72°C for 30 s. Fluorescence measurements were collected after the elongation phase. A melting curve analysis was performed after the qPCR to check the amplification specificity. All the amplifications were carried out in duplicate.

The raw data were analyzed with StepOne software v2.3 (Applied Biosystems). The analytical sensitivity and robustness of the qPCR assays were determined based on the amplification efficiency, which was calculated following serial dilutions (47), and on linear regression data determined by the log(fluorescence)-per-cycle-number method (48). The relative ccr copy number was determined as a ratio (fold change) of the normalized ccr and glts initial abundances. For each PCR, the threshold cycle (C_T) values were converted into initial DNA abundance values (2^−CT), adjusted to their respective dilution factors, and averaged. For each qPCR assay, a calibrator was arbitrarily selected as a DNA known sample or considered to contain a unique copy of the target gene (i.e., for acetate-nonacclimatized strains of Rs. rubrum). Finally, the relative copy number of the target gene was obtained for each DNA sample by calculating the calibrator-normalized ratio between the initial abundance of the target gene and the average initial abundance of the reference gene as follows:

$$\text{RCN}(\text{target})=\frac{A_i(\text{target})\times \overline{A_l^{Cal}}(\text{reference})}{\overline{A_l}(\text{reference})\times \overline{A_l^{Cal}}(\text{target})}$$ (1)

where RCN is the relative copy number, A_i is the initial abundance obtained from the C_T values, A_l is the average initial abundance, and $\overline{A_l^{Cal}}$ is the average initial abundance of the calibrator.

Proteomic analysis using mass spectrometry (MS).

Quantitative proteomic analysis was performed using protein extracts of the Rs. rubrum S1H acetate-acclimatized strain cultivated in medium containing acetate or succinate. Cells from 5 biological replicates were harvested by centrifugation (16,000 × g, 10 min, 4°C) when the OD₆₈₀ reached 0.9 to 1.2. Proteins were extracted by sonication (3 × 10 s, amplitude of 40%, IKA U50 sonicator) in 6 M guanidinium chloride solution. Extracted proteins were reduced, alkylated, and precipitated with acetone. The proteolytic peptides were obtained by overnight enzymatic digestion using trypsin at a ratio of 1:50 (wt/wt).

MS experiments were conducted following a label-free strategy on an ultra-high-performance liquid chromatography–high-resolution tandem mass spectroscopy (UHPLC-HRMS/MS) platform (Eksigent 2D Ultra and AB Sciex TripleTOF 5600 system) in the SWATH mode of acquisition (data-independent acquisition [DIA]). Peptides (2 μg) were separated on a C₁₈ column (Acclaim PepMap100; Dionex) (3-μm pore size, 150 μm by 25 cm) with a linear acetonitrile gradient (5% to 35% [vol/vol] of acetonitrile [300 nl · min⁻¹], 120 min) in water containing 0.1% (vol/vol) formic acid. Each MS survey scan (400 to 1,500 m/z, 50 ms accumulation time) was followed by 50 SWATH acquisition overlapping windows covering the precursor m/z range. For each window, ions were fragmented using rolling collision energy, and fragment ion spectra were accumulated for 95 ms in high-sensitivity mode.

SWATH spectra were identified by comparison to a reference spectral library obtained using data-dependent acquisition (DDA) experiments performed on proteins extracted from R. rubrum S1H grown on different carbon sources (acetate, succinate, and butyrate). Sample preparation and separation procedures used to construct the library were identical to those previously mentioned. DDA spectra were acquired with the following parameters: MS scan (400 to 1,500 m/z, 500 ms accumulation time) in high (>35,000)-resolution mode followed by 50 MS/MS scans (100 to 1,800 m/z, 50 ms accumulation time, intensity threshold at 200 cps). The DDA mass spectrometry data were processed with AB Sciex ProteinPilot 4.5 software. Identification of the spectra was performed by searching against the Rhodospirillum rubrum ATCC 11170 UniProt entries (release date, 22 May 2017) with the relevant parameters, including carbamidomethyl cysteine, oxidized methionine, all biological modifications, amino acid substitutions, and missed cleavage site. Proteins identified at a false-discovery rate of below 1% were used as the SWATH reference spectral library.

SWATH wiff files were processed using AB Sciex PeakView 2.1 software and the SWATH Acquisition MicroApp. Up to 6 peptides were selected with 6 transitions per peptide with at least 99% confidence. The XIC extraction window was set to 25 min and the XIC width to 75 ppm. The XIC peak area was extracted and exported to AB Sciex MarkerView 1.2 software for normalization and statistical analysis.

Sequencing and read depth coverage.

Whole-genome sequencing was performed by MicrobesNG (http://www.microbesng.uk). The genomic DNA of acetate-acclimatized and nonacclimatized strains was extracted from steady-state cultures to avoid replication bias of the read count and was prepared following the requirements of MicrobesNG. The resulting 2 × 250 paired-end reads were aligned on the reference genome (PubMed ID: NC_007643.1) using Geneious (v10.1.3, Biomatters Ltd.). The average number of reads was computed for nonoverlapping windows of 1,000 or 10,000 nucleotides. The final read depth coverage was normalized using the expected read count (i.e., the average read count across the whole genome with the amplified region excluded).