R. rubrum characterization under different levels of CO in light and dark conditions

CO, which is one of the main components of syngas, has been extensively studied in R. rubrum (Karmann et al. 2019; Kerby et al. 1995; Rodríguez et al. 2021; Shelver et al. 1995). However, the physiological effects of different doses of CO have not been assessed in a systematic study, specially at high levels of this gas. To obtain this information, cells of R. rubrum Rr02_01 were cultivated in sealed bottles with different concentrations of CO in the headspace. This strain can reach higher levels of PMP than the wild-type strain S1, both under light and in darkness (Fig. S1).

Growth kinetics

Through the range of initial ppCO explored in photoheterotrophic conditions, two different behaviors in the growth kinetics can be noticed (Fig. 2). The simplest one could be associated with the canonical bacterial growth curve, with an initial lag phase, followed by exponential growth and concluded with a stationary phase. Cultures exposed to light with an initial ppCO of 0.0 bar and 1.0 to 2.5 bar fit in this type of curve. However, an extended growth limit was observed for the range of ppCO from 0.2 to 0.8 bar. In these cases, a brief re-adaptation phase followed the exponential growth, after which cells re-started biomass formation at a slower rate, eventually reaching the stationary phase. This curve matches with a typical diauxic growth. With simpler kinetics, cultures in darkness behaved similarly despite varying initial ppCO concentrations. As acetate cannot sustain growth per se in darkness (Kerby et al. 1995), bottles without CO were not included in the experimental design. Without light, just one growth phase was observed, after which OD600 decreased to eventually remain steady in the stationary phase (Fig. 2C). This decrease after the maximum of turbidity could reflect lysis or, more likely, changes in their morphology/compositions triggered by the exhaustion of a limiting nutrient.

Fig. 2

Growth curves under different levels of CO with and without light. Phototrophic cultures grown with an initial ppCO ranging from (A) 0.0 to 0.8 bar and from (B) 1.0 to 2.5 bar. The condition of ppCO = 0.0 bar is displayed in both graphs to facilitate the comparison. (C) Cultures grown in darkness with an initial ppCO ranging from 0.2 to 1.0 bar. Curves represent the mean value of OD660 (n = 3). Standard deviation was less than 10% in all cases. (D) A schematic representation of the growth dynamics observed in this experiment is depicted: (a) Initial lag phase; (b) exponential growth (μmax); (c) turbidity remains steady (light) or c*) rapidly decreases (darkness). (d) Some cultures (light, ppCO 0.2–0.8 bar) exhibit a second growth phase with a μ < μmax after phase C. (e) Turbidity decreases gradually (stationary phase). In the case of phototrophic cultures with ppCO = 0.0 bar or ≥ 1.0 bar, there is no d transition between c and e. The timepoints/lapses where the different parameters were measured are indicated with grey arrows. LagP, lag phase; Acet, acetate

Lag phase and growth rate

The most conspicuous effect of CO on cell growth was related to the lag phase. In phototrophic cultures, while the lag phase remained steady for doses up to 1.5 bar, this variable increased for higher initial ppCO values, ultimately reaching 1.4 ± 0.3 days at ppCO of 2.5 bar (Fig. 3C, Fig. S2). In the case of the maximal growth rate (μmax), the optimum was obtained at ppCO 0.6 bar (1.8 ± 0.2 day−1), and the lowest value was for the highest level of initial ppCO (0.8 ± 0.2 day−1). In darkness, this effect was more pronounced (Fig. 3D). The lag phases in these conditions ranged from 1.4 ± 1.2 days (ppCO 0.2 bar) to 19.5 ± 6.1 days (ppCO 1.0 bar). The μmax decreased in darkness in a much smoother fashion than the increase in the lag phase: from 1.1 ± 0.2 day−1 to 0.7 ± 0.2 day−1 (ppCO of 0.2 and 1.0 bar respectively). These facts reflect that the hardest challenge for cells at high doses of this gas is to adapt to CO, which once accomplished, leads to a similar evolution of the culture.

Fig. 3

General profiling of R. rubrum grown under different levels of CO with and without light. Using RRNCO medium, strain Rr02_01 was cultured under increasing levels of initial ppCO, from 0.0 to 2.5 bar under light (A, C, E, G) and 0.2 to 1.0 bar in darkness (B, D, F, H). Cultures were monitored and assessed for several relevant parameters, such as the production of biomass and PMP (A and B); the growth rate and lag phase (C and D); the presence of CO2, H2, and CO in the headspace; and the consumption of acetate (E and F) and the accumulation of PHB (G and H). For more details on the sampling times, see Fig. 2C. Values represent the mean, and error bars represent the standard deviation of at least three independent biological replicates

Acetate consumption and transformation of CO into CO2/H2

The acetate (10 mM) of the cultures grown photoheterotrophically was completely depleted by the end of the first exponential growth phase (Fig. 3E). Gas consumption/production profiles varied depending on the initial ppCO. Without CO, R. rubrum did not produce significant amounts of any of the studied gases. In cultures fed with CO, different amounts of H2 and CO2 were produced in a CO-dependent manner: the higher the initial ppCO was, the more H2 and CO2 were accumulated in the headspace of the bottles. At the end of this experiment, cultures with an initial ppCO from 0.2 to 0.8 had no H2 or CO remnant. On the contrary, for higher initial ppCO levels, various amounts of H2, CO2, and CO were detected. This indicates that the consumption of H2 and CO2 occurred only in cases where CO was completely depleted from the headspace (ppCO 0.2 to 0.8 bar). H2 consumption to drive photosynthesis as a source of reductive power will be elaborated in the “Discussion” section.

In darkness, the final biomass and acetate consumption were also directly dependent on the initial ppCO (Fig. 3F). In this case, unlike photoheterotrophic cultures, acetate was not depleted for any initial ppCO, meaning it was in excess in respect to CO, the limiting growth factor, whereas CO was totally consumed by the end of the experiment.

PHB accumulation and pigment production

At the end of the exponential growth in light or during the stationary phase both in light and dark cultures, the intracellular storage of PHB was almost completely depleted (data not shown). To compare the different conditions, its concentration was measured at the end of the exponential phase, when the OD660 was maximal. Under light, the amount of PHB was higher at ppCO 0.6 (8.3 ± 3.4 % CDW, 34.9 ± 7.9 μg·mL−1) compared to the other conditions where CO was present (Fig. 3G). In the absence of CO, the accumulation of PHB was similar, reaching 9.5 ± 2.7 % CDW or 46±18 μg·mL−1. In darkness, the upward trend on biomass along the increasing initial ppCO was followed by an upward volumetric accumulation of PHB, keeping the % PHB unvaried along the range studied (Fig. 3H). While the latter reached around 10% in the different conditions, the volumetric production of the polymer reached as much as 35 ± 13 μg·mL−1 with ppCO of 1.0 bar.

Observing the values of PMP represented in Fig. 3A, at the end of the first growth phase, the highest PMP was observed at ppCO of 0.6 bar (2.1 ± 0.3 Abs880/660) and the minimum at ppCO of 2.5 bar (1.6 ± 0.1 Abs880/660). This value was similar to the PMP produced in the absence of CO (1.67 ± 0.04 Abs880/660). Cultures grown in darkness showed a PMP similar for all the initial ppCO studied (~1.7 Abs880/660) (Fig. 3B).

Kinetic analysis of main culture parameters at selected CO levels

In the previous section, two types of growth kinetics were observed for R. rubrum cultured under light depending on the initial ppCO. On the other hand, in darkness, no such differences were observed: all the cultures exhibited equivalent properties regarding biomass, pigment and PHB production, and the carbon substrate (acetate and CO) consumption. Consequently, we decided to zoom in on the physiological aspects of three typical growth kinetics for phototrophic cultures (ppCO of 0.0 bar, 0.6 bar, and 2.0 bar) and one grown in darkness (ppCO=0.6 bar) (Fig. 4).

Fig. 4

Culture evolution in selected conditions. Curves of acetate and CO consumption, OD660, and H2/CO2 production for each of the selected conditions. Photoheterotrophic cultures were conducted with an initial ppCO of (A) 0.0 bar, (B) 0.6 bar, or (C) 2.0 bar. No gas was detected in the condition with no CO (A). (D) Cultures in darkness were evaluated with ppCO of 0.6 bar. Curves represent the mean of at least three independent biological replicates, and the shadow denotes the standard deviation. For PHB values. The standard deviation was lower than 15% of the mean

As expected, acetate and CO consumption were associated with light and dark growth. For a ppCO of 0.6 bar in light, CO and acetate were depleted approximately at the same time (Fig. 4B). On the other hand, CO could not be completely consumed at the expense of acetate in cultures with an initial ppCO of 2.0 bar (Fig. 4C). Consequently, a fraction of this gas remained in the bottle until the end of the experiment. Therefore, CO depletion seems to trigger the second growth phase. When cultures run out of acetate in light, PHB undergoes a rapid mobilization. In darkness, acetate is consumed all along cell growth, when biomass and PHB are produced (Fig. 4D). When CO is over, acetate consumption ceases immediately, and its concentration increases in the supernatant, meaning this acid is secreted probably at the expense of PHB depolymerization.

To verify if the general behavior of strain Rr02_01 was similar for its parental strain S1, the growth curves of both were tested in these conditions, giving very similar results in terms of biomass formation (OD660), with a higher level of PMP for the strain Rr02_01 in all the conditions, as expected (Fig. S3). Under light without CO, differences were minimal. Interestingly, in cultures grown with light and ppCO 0.6 bar, strain Rr02_01 grew faster during the second growth phase than the parental strain (0.066 ± 0.009 day−1 vs 0.045 ± 0.004 day−1, respectively) and reached a higher OD (4.94 ± 0.44 vs. 3.97 ± 0.21, respectively) (Table S3). At this point, the differences with respect to the PMP were also maximal, with a PMP value 35% higher than the parental strain. Moreover, the stationary phase of S1 presented a much faster decline in the OD as can be seen in Fig. S3. These showed that while both strains behave similarly for variables such as OD and growth rate during growth in CO, after the depletion of this gas, strain Rr02_01 seems to grow more efficiently using CO2 and H2.

Expediting adaptation to CO in darkness

Great metabolic versatility is one of the main characteristics of PNSB, and this fact is manifested in the high tolerance we observe in R. rubrum under CO. However, this gas has a significant impact on the physiology of R. rubrum during dark growth, particularly during the lag phase. To test whether we could further enhance the metabolic tolerance of R. rubrum in these conditions, we intended to reduce its long lag phase by randomly modifying its genetics. As a proof of concept, we decided to conduct an experiment based on adaptive laboratory evolution (ALE) principles. Given that a traditional ALE requires several generations to find positive mutations and R. rubrum has a high duplication time in darkness, a random mutagenesis factor was incorporated into the conventional ALE experiment to accelerate the emergence of mutant clones under high CO levels. We called this approach a UV-accelerated (UVa) ALE or simply UVa-ALE.

A scheme of the UVa-ALE is shown in Fig. 5 (the details can be found in the “Materials and methods” section). After exposing bacteria to UV, they were allowed to recuperate aerobically in a rich medium prior to being incubated anaerobically with N2. This step is supposed to minimize stress before the first exposure to CO. The selective pressure was exerted in two levels (ppCO of 0.8 and 2.0 bar) to avoid an excessive stringent condition that could imply a high mortality of cells. It is worth mentioning that we had never seen the growth of R. rubrum in 2.0 bar of CO in our experiments when light was not supplied. The isolation of clones was done in an atmosphere of CO (0.8 bar) to maintain the selective pressure, using bottles with a solid medium (Fig. S4). In this step, 18 clones were screened in liquid RRNCO medium at an initial ppCO of 1.0 bar, from which 3 did not grow and only 7 showed a significantly shorter lag phase compared to the parental strain Rr02_01 (Fig. 6). When grown in a rich medium for routine handling, clone 1.4-2B showed more robust growth compared to clones isolated from the same bottle (series 1.4). Clone 1.7-3A was particularly interesting because it formed pigmented colonies on the agar plate, in contrast to the Rr02_01 strain (data not shown). For these reasons, clones 1.4-2B and 1.7-3A were selected for further analysis.

Fig. 5

Scheme of the UVa-ALE. A culture of R. rubrum Rr02_01 grown in rich medium was exposed to UV for 20 min, after which, they were allowed to recuperate. The idea to use UV radiation was to augment the mutation rate, given that the replication of R. rubrum in CO is slow under dark conditions, making it necessary to increase the genetic variability and thus reduce the duration of the experiment. The selective pressure consisted on a first step of ppCO of 0.8, followed by other of ppCO 2.0 bar (5 passages). The isolation of clones was done in a controlled atmosphere (ppCO 0.8 bar) using solid medium

Fig. 6

Screening of clones with shorter lag phase. 18 clones isolated after the UVa-ALE (blue) were tested and compared to the parental strain Rr02_01 (yellow) in RRNCO medium (ppCO 1.0 bar). The three clones which did not grow (ø) were discarded. Clones with a significant difference with respect to parental were colored in green (ANOVA, Dunnett’s post hoc test). The values represent the mean, and the error bars represent the standard deviation of three experimental units

Increased CO adaptability impacted PHA or pigment production

The mutations in both 1.4-2B and 1.7-3A strains might have pleiotropic effects beyond the increased adaptability to CO. Four main products were assessed to test the feasibility of using these strains for biotechnological processes: PHA, H2, pigments (PMP), and biomass (OD660) (Fig. 7).

Fig. 7

Characterization of strains 1.4-2B and 1.7-3A. (A) The growth curves are represented at the corresponding average lag phase of each strain. (B) The values of lag phase are represented with boxplots and a normal curve is fitted to the experimental data (dots). The consumption of (C) acetate and the production of (D) PMP, (E) H2, and (F) PHB was measured. The cultures were done in RRNCO medium with ppCO of 1.0 bar. Values represent the mean, and the error bars the standard deviation of at least three independent biological replicates. Asterisks denote significance with p value < 0.05 (ANOVA, Tukey’s test)

When grown in RRNCO medium with CO (1.0 bar) in the absence of light, no variations in hydrogen production among the three strains were observed. The levels of PHB were similar among the three strains: Rr02_01 (11 ± 2 %), 1.7-3A (10 ± 2 %), and 1.4-2B (8 ± 2 %). However, this strain showed a significant increase in the level of PMP (+19%). The maximal OD660 reached by strain 1.7-3A was 26% lower than that of the parental strain (0.97 vs 1.30). Concomitantly, the amount of consumed acetate was proportionally lower in strain 1.7-3A, as it consumed one-fourth less of the acetate in comparison to the parental strain (4.1 ± 0.5 mM vs. 5.2 ± 0.8 mM, respectively).

The genetic variations behind increased CO adaptability in mutants 1.7-3A and 1.4-2B

In order to identify the genetic differences between clones 1.7-3A and 1.4-2B with the parental strain Rr02_01, both mutant genomes were sequenced. A total of 5 and 7 mutations were found in the genome of strains 1.4-2B and 1.7-3A, respectively (Table 1).

Table 1 Genetic variations found in strains 1.4-2B and 1.7-3A compared to the parental strain Rr02_01

In strain 1.4-2B, five mutations were found, impacting on a DegT/DnrJ/EryC1/StrS aminotransferase (Rru_A0933), a short-chain dehydrogenase/reductase (Rru_A1264), a NADH dehydrogenase (ubiquinone) (Rru_A1424), and an intergenic region (position 3,782,100). Locus Rru_A1424 (cooU) is part of the cooMKLXUH operon, which codes for CO-induced hydrogenase-related proteins under the control of the transcription factor CooA (Fox et al. 1996). This mutation does not annul the protein but substitutes an alanine (a hydrophobic amino acid) at position 155 for a threonine (a polar amino acid), with phenotypic consequences that are difficult to predict. Additionally, a mutation affected the gene coding for a member of the subfamily of multidrug and toxic compound extrusion (MATE)-like proteins (Rru_A0697). These proteins are involved in basic mechanisms for homeostasis maintenance by facilitating the excretion of metabolic waste products and xenobiotics (Moriyama et al. 2008).

In the case of strain 1.7-3A, seven differences in its genomic sequence with respect to the parental strain were found. This included a signal transduction histidine kinase (Rru_A0741), a DeoR family transcriptional regulator (Rru_A1265), a periplasmic sensor diguanylate cyclase/phosphodiesterase (Rru_A2602), a prolyl-tRNA synthetase (Rru_A1576), and an intergenic region (position 3,820,243). More interestingly, in this strain, a reversion of a single nucleotide deletion was produced in locus Rru_A0625. This locus codifies a particular version of the anti-repressor PpaA, which in PNSB modifies PpsR repressor activity. PpaA in R. rubrum has been recently proved to affect various biological processes related to adaptation to microarobiosis/anaerobiosis (Godoy et al. 2023b). Notably, the same mutation at the Rru_A0697 locus present in stain 1.4-2B was also identified in 1.7-3A. It is worth highlighting that both strains come from completely independent clones.

In general terms, mutations present in 1.4-2B seem to be more related to genes coding for metabolic enzymes, some of them related to redox reactions (dehydrogenases/reductases), whereas mutations present in 1.7-3A are more related to regulatory proteins (anti-repressor, histidine kinase, diguanylate cyclase, etc), showing that the two genetic profiles selected after CO selective pressure are clearly differentiated.