The phototrophic purple non‐sulfur bacteria Rhodomicrobium spp. are novel chassis for bioplastic production

Abstract

Petroleum‐based plastics levy significant environmental and economic costs that can be alleviated with sustainably sourced, biodegradable, and bio‐based polymers such as polyhydroxyalkanoates (PHAs). However, industrial‐scale production of PHAs faces barriers stemming from insufficient product yields and high costs. To address these challenges, we must look beyond the current suite of microbes for PHA production and investigate non‐model organisms with versatile metabolisms. In that vein, we assessed PHA production by the photosynthetic purple non‐sulfur bacteria (PNSB) Rhodomicrobium vannielii and Rhodomicrobium udaipurense. We show that both species accumulate PHA across photo‐heterotrophic, photo‐hydrogenotrophic, photo‐ferrotrophic, and photo‐electrotrophic growth conditions, with either ammonium chloride (NH₄Cl) or dinitrogen gas (N₂) as nitrogen sources. Our data indicate that nitrogen source plays a significant role in dictating PHA synthesis, with N₂ fixation promoting PHA production during photoheterotrophy and photoelectrotrophy but inhibiting production during photohydrogenotrophy and photoferrotrophy. We observed the highest PHA titres (up to 44.08 mg/L, or 43.61% cell dry weight) when cells were grown photoheterotrophically on sodium butyrate with N₂, while production was at its lowest during photoelectrotrophy (as low as 0.04 mg/L, or 0.16% cell dry weight). We also find that photohydrogenotrophically grown cells supplemented with NH₄Cl exhibit the highest electron yields – up to 58.89% – while photoheterotrophy demonstrated the lowest (0.27%–1.39%). Finally, we highlight superior electron conversion and PHA production compared to a related PNSB, Rhodopseudomonas palustris TIE‐1. This study illustrates the value of studying non‐model organisms like Rhodomicrobium for sustainable PHA production and indicates future directions for exploring PNSB metabolisms.

The purple non‐sulfur bacteria Rhodomicrobium vannielii and Rhodomicrobium udaipurense produce polyhydroxyalkanoate under various phototrophic growth conditions using N₂ gas or ammonium chloride as nitrogen sources. This study adds to a growing body of literature exploring metabolically versatile microbes as biocatalysts for heterotrophic and autotrophic bioproduction.

INTRODUCTION

Plastics are foundational to modern life. They are inexpensive, lightweight, durable, and versatile materials with uses in packaging, transportation, agriculture, construction, healthcare, and energy (Geyer et al., 2017; Rosenboom et al., 2022). Unsurprisingly, we are producing – and discarding – more plastic than ever: from 2000 to 2019, plastic production and waste doubled, reaching 460 million and 353 million tonnes, respectively (OECD, 2022). Of all the plastic waste produced in 2019, only 9% was recycled, while the remaining 91% was either landfilled, incinerated, or otherwise mismanaged/uncollected (e.g., diverted to uncontrolled dumpsites) (OECD, 2022). Plastic production currently accounts for 3.4% of global greenhouse gas emissions, and by 2050 it will be responsible for almost 13% of the global carbon budget – equal to the combined emissions of 615 coal‐fired power stations (HBS, 2019; OECD, 2022). A long‐term solution to these problems lies in circularizing the plastics economy (Rosenboom et al., 2022). To do so, we must transition from single‐use, non‐biodegradable, petrochemical‐based plastics to more sustainable alternatives.

Polyhydroxyalkanoates (PHAs) are polyesters produced by diverse microorganisms as intracellular carbon and energy reserves in response to nutrient and redox imbalance, light stress, and other environmental stimuli (Batista et al., 2018; Bayon‐Vicente et al., 2020; Diankristanti et al., 2024). PHAs comprise over 150 monomers with varying physical and structural properties, though they are generally thermoplastic, biocompatible, biodegradable, non‐toxic, inert, and hydrophobic (Choi et al., 2020; Diankristanti et al., 2024; McAdam et al., 2020). These properties make PHA a suitable bioplastic candidate with potential applications in packaging, agriculture, biomedicine, and pharmaceuticals (Vaishnav & Choudhary, 2021). While a small number of companies are in the commercial PHA space, industrial‐scale production suffers from low yields, reliance on energy‐ or cost‐intensive inputs, and narrow metabolisms (McAdam et al., 2020). Yet addressing these barriers to market entry alone is not enough; solutions must also account for broader climate concerns, e.g., greenhouse gas emissions and competition with agriculture for organic feedstock to fuel heterotrophic bioproduction.

With all of these concerns in mind, autotrophic microbes that fix CO₂ using light (photoautotrophs) or chemical compounds (chemolithotrophs) show great promise as chassis for bioproduction (see Srisawat et al., 2022 for a review). These include purple non‐sulfur bacteria (PNSB), a group of phototrophic bacteria that synthesize various bioproducts including PHAs (Higuchi‐Takeuchi & Numata, 2019; Monroy & Buitrón, 2020). PNSB are particularly noteworthy for their metabolic flexibility, which affords opportunities to investigate diverse biosynthetic pathways while deepening our understanding of PNSB metabolisms. Yet PNSB remain understudied as biological chassis for PHA production, with only a few genera – namely Rhodobacter, Rhodospirillum, Afifella, Rhodovulum, and Rhodopseudomonas – among those reported in the literature (Monroy & Buitrón, 2020; Shaikh et al., 2023). Absent from the bioproduction literature is Rhodomicrobium, a genus of microaerobic to anaerobic PNSB found in freshwater soils and sediments, especially as constituents of plant microbiomes (Town et al., 2023; Wei et al., 2020; Zuo et al., 2022). The Rhodomicrobium genus includes three primary members: R. vannielii, R. udaipurense, and R. lacus, all of which exhibit a polymorphic lifecycle with filamentous non‐motile cells, motile swarmer cells, and exospores (Duchow & Douglas, 1949; G et al., 2020; Ramana et al., 2013; Richter et al., 2023; Whittenbury & Dow, 1977). R. vannielii and R. udaipurense, the most well‐characterized species, fix dinitrogen (N₂) gas and perform chemoheterotrophic, photoheterotrophic, and photoautotrophic metabolisms, including photosynthetic iron oxidation (i.e., photoferrotrophy) (Duchow & Douglas, 1949; Ehrenreich & Widdel, 1994; Whittenbury & Dow, 1977). Recently, R. vannielii and R. udaipurense were shown to conduct extracellular electron uptake (EEU), wherein outer membrane protein complexes facilitate electron uptake from solid phase conductive substances (e.g., poised electrodes or iron minerals) (Gupta et al., 2019). EEU capabilities open up the possibility of bioproduction via microbial electrosynthesis (MES), wherein whole‐cell biocatalysts use electrons supplied by a cathode to reduce CO₂ to complex carbon compounds (Nevin et al., 2010). This metabolic versatility makes Rhodomicrobium spp. appealing for further study, as it affords the opportunity to fill gaps in our knowledge about PNSB metabolisms in the context of bioproduction.

The primary objective of this study was to quantify PHA accumulation by R. vannielii and R. udaipurense across four phototrophic growth conditions with either N₂ gas or NH₄Cl as a nitrogen source. We also characterized the relationship between carbon yield, electron yield, and PHA production. Our data show that Rhodomicrobium produces the most PHA during N₂‐fixing growth on sodium butyrate, while photoautotrophic growth leads to comparable or superior carbon and electron yields despite lower intracellular PHA content. We also compared Rhodomicrobium to a better‐studied PNSB with similar metabolic capabilities, Rhodopseudomonas (Rh.) palustris TIE‐1. We find that Rhodomicrobium spp. are comparable or superior to TIE‐1 across the metrics we assessed – particularly electron yields and PHA content per cell. These data highlight the importance of studying non‐model organisms for bioproduction and suggest future avenues for improving PHA production by PNSB.

EXPERIMENTAL PROCEDURES

Bacterial strains, media, and non‐electrochemical growth conditions

Rhodomicrobium vannielii ATCC 17100 was purchased from DSMZ (Leibniz Institute, Braunschweig, Germany). Rhodomicrobium udaipurense JA643 was acquired from the University of Hyderabad (Hyderabad, India). Both strains were incubated anaerobically in bicarbonate‐buffered freshwater (FW) medium prepared under the flow of N₂‐CO₂ (80%/20%) gas (Ehrenreich & Widdel, 1994). Either 5.61 mM NH₄Cl (non‐N₂‐fixing) or N₂ gas at ~103 kPa (N₂‐fixing) was provided as a nitrogen source. For all non‐electrochemical growth conditions, the headspace of sterile and sealed Balch tubes was vacuumed and flushed with the appropriate gas three times before being pressurized as described below. Photoheterotrophic and photohydrogenotrophic cultures were grown in airtight 27 mL Balch tubes with 15 mL of cell culture and 12 mL of headspace. Photoferrotrophic cultures were grown in airtight 150 mL serum bottles with 80 mL of cell culture and 70 mL of headspace. All cell cultures were incubated at 30°C without shaking at a 30‐cm distance from a 60‐W incandescent light bulb providing a total irradiance of approximately 40 W/m². Time‐course cell growth was monitored using a Spectronic 200 spectrophotometer (Thermo Fisher Scientific, USA). All sample manipulations were performed inside an anaerobic chamber with H₂‐N₂‐CO₂ (5%/75%/20%) atmosphere.

For photoheterotrophic growth, 15 mL of FW medium supplemented with anoxic sodium butyrate to a final concentration of 10 mM was dispensed in Balch tubes sealed with sterile butyl stoppers and an aluminium crimp. Cultures were pre‐grown in FW medium with sodium butyrate and N₂ or NH₄Cl to an optical density (OD₆₆₀) of ~1.0, pelleted, washed with FW medium (without NH₄Cl), and inoculated to a final OD₆₆₀ of ~0.01. Then, the headspaces of all Balch tubes were flushed with N₂‐CO₂ (80%/20%) gas and vacuumed three times, with a final vacuum step before pressurization with N₂‐CO₂ gas to 103 kPa. For photohydrogenotrophic growth with H₂ gas as an electron donor, 15 mL cell cultures in FW medium were prepared in Balch tubes and pressurized following the same procedure, except cells were pre‐grown photohydrogenotrophically with N₂ or NH₄Cl. For N₂‐fixing photohydrogenotrophy, the headspace was flushed with N₂‐CO₂ gas and vacuumed three times, with a final vacuum step before pressurization to 48 kPa with N₂‐CO₂ and then topped off to 103 kPa with H₂‐CO₂ (80%/20%). Additionally, N₂‐fixing cultures were inoculated to an initial OD₆₆₀ of 0.1 to promote cell growth, as cells inoculated at a lower initial OD₆₆₀ failed to grow. For photoferrotrophic growth with Fe(II) as an electron donor, 500 mL of FW medium was prepared in a sterile, capped, and crimped serum bottle that was flushed with N₂‐CO₂ gas and vacuumed three times, with a final vacuum step before pressurization with N₂‐CO₂ to 103 kPa. The media was then supplemented with anoxic sterile stocks of nitrilotriacetic acid (NTA) to a final concentration of 10 mM and iron chloride (FeCl₂) to a final concentration of 5 mM. Then, 80 mL of the FW/FeCl₂/NTA medium was aliquoted to sterile, capped, and crimped serum bottles pre‐flushed with N₂‐CO₂ (80%/20%) gas. Cultures were then inoculated with cells pre‐grown photohydrogenotrophically to a starting OD₆₆₀ of 0.1 and pressurized to 103 kPa as described above.

Photoelectrotrophic growth

Photoelectrotrophic growth was performed using single‐chamber three‐electrode configured seal‐type bioelectrochemical cell (BECs) (C001 Seal Electrolytic Cell, Xi'an Yima Opto‐electrical Technology Com., Ltd, China) as described by Rengasamy et al. (2018), with slight modifications. The working electrode was fabricated from a 1 cm² section of carbon felt (Fuel Cell Earth, Woburn, Massachusetts, USA) poised at +100 mV vs. standard hydrogen electrode (SHE) using a multichannel potentiostat (Gamry Instruments, Warmister, PA) that was operated continuously for 120 hrs. (N₂‐fixing photoelectrotrophy) or 140 hrs. (non‐N₂‐fixing photoelectrotrophy). To prepare cell cultures, 75 mL of FW media was dispensed into sterile BECs which were sealed and made anaerobic by purging and pressurizing to 48 kPa with N₂‐CO₂ (80%/20%). BECs were connected to the potentiostat and monitored overnight for evidence of residual oxygen presence or contamination prior to inoculation. 5 mL of cells pre‐grown photohydrogenotrophically in FW media to an OD₆₆₀ of 1.0 were used to inoculate BECs to an OD₆₆₀ of 0.2, leaving a headspace volume of 20 mL. The resulting cultures were purged and pressurized to 48 kPa with N₂‐CO₂ before reattaching to the potentiostat. BECs were operated continuously at 26°C at a 30‐cm distance from a 60‐W incandescent light bulb providing a total irradiance of approximately 40 W/m². Cell growth was monitored visually and final OD₆₆₀ readings made up of combined planktonic and biofilm cells attached to the cathode were taken using Spectronic 200 (Thermo Fisher Scientific, USA).

Cell enumeration

To enumerate cells, R. vannielii and R. udaipurense cultures grown in FW medium supplemented with NH₄Cl and 10 mM sodium butyrate were harvested at mid‐exponential phase (OD₆₆₀ ~ 0.5) and diluted in FW media to generate a range of optical densities between the early‐ to late‐exponential phases of the growth curve. The number of cells in each of these cell suspensions was quantified using a Petroff‐Hauser chamber (Hausser Scientific, Horsham, PA, USA) according to the manufacturer's protocol. Cells were visualized and counted via phase contrast microscopy using 10X and 40X objectives (Olympus BX43 manual system microscope, Olympus Life Science). From this, an average cell number/mL of cell culture was determined, which was plotted against the corresponding OD₆₆₀ of the cell cultures to obtain a standard curve. The standard curve was used to approximate the number of cells/mL at a given OD₆₆₀, which allowed us to normalize PHA production to cell number, biomass, and biovolume.

Total protein quantification for biomass estimates

Normalizing PHA production to biomass (typically dry biomass) is valuable for assessing productivity. However, at the volumes and cell densities reported here, we were unable to attain reliable direct measures of biomass. Therefore, we opted to indirectly estimate biomass using two orthogonal approaches: a microscopy‐based method and a total protein‐based method. The Methods S1 describe the microscopy‐based approach, which strongly agrees with the total protein‐based estimates. Here, we describe our total protein‐based method, which is reported throughout the main body of the text.

At time of PHA sampling, 2–5 mL of cell culture was removed and centrifuged at 10,000 × g for 5 min. After removing the supernatant, cell pellets were stored at −80°C until protein extraction. To quantify total protein, cell pellets were thawed and processed using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, USA) following manufacturer's protocol. In brief, three technical replicates for each sample were mixed with the supplied reagents in a 96‐well plate and incubated in the dark at 37°C for 30 min. Then, the absorbance at 562 nm for each sample was measured using a BioTek Synergy HTX Multimode plate reader (Agilent, USA). Absorbance values were converted to total protein based on a standard curve generated from a bovine serum albumin standard. The total protein for each group was calculated from the average total protein of three (photohydrogenotrophy, photoferrotrophy, and photoheterotrophy) or two (photoelectrotrophy) biological replicates. From here, total protein values were converted to cell dry weight (reported as % cdw_prot) based on an assumed protein/dry weight ratio of 52.4%. (Stouthamer, 1973). PHA production was normalized to this value to give PHA % cdw_prot.

Analytical procedures

PHA measurement

10 mL (photoheterotrophy, photohydrogenotrophy with H₂) or 40–50 mL (photoferrotrophy, photoelectrotrophy) of cell cultures at mid‐ to late‐exponential phase (Figure S3) were pelleted at 8000 × g for 10 min and stored at −80°C until PHA extraction was performed. Extraction and quantification of PHA was accomplished via its conversion to crotonic acid, which here serves as a proxy for PHA based on a standard curve using a commercially available polyhydroxybutyrate standard. In brief, cell pellets were thawed and resuspended in 425 μL of methanol, 500 μL of HPLC grade chloroform, and 75 μL of 98% sulfuric acid, followed by digestion via 1‐h incubation at 95°C. Following digestion, samples were separated into organic and aqueous phases via centrifugation at 500 × g for 10 min, and the organic phase was dried overnight via Savant SC210A Speedvac concentrator (Thermo Fisher Scientific, USA). Dried samples were resuspended in 1:1 acetonitrile: water and filtered for subsequent LC–MS analysis. LC–MS analysis protocol was previously described (Ranaivoarisoa et al., 2019). PHA was detected as crotonic acid with mass‐to‐charge ratio (m/z) = 87 and normalized to bacterial cell number and total biomass (see “Cell enumeration” and “Total protein quantification for biomass estimates”). Additional details on PHA extraction, PHA carbon yield, and PHA electron yield calculations are described in Methods S1.

H₂ , N₂ , and CO₂ measurement

Time‐course H₂, N₂, and CO₂ were analysed via gas chromatography (Shimadzu BID 2010‐plus, equipped with RT – Silica BOND PLOT Column: 30 m × 0.32 mm; Restek, USA) with helium as a carrier gas following the manufacturer's protocol as described previously (Ranaivoarisoa et al., 2019). Standard curves were generated using known quantities of H₂, N₂, and CO₂ gas, wherein the area under the curve for each gas was plotted against moles of gas, the latter of which was calculated using the ideal gas law. Moles of gas in the media and headspace of cell cultures were calculated based on the linear equation from the standard curve.

Sodium butyrate measurement

Time‐course consumption of sodium butyrate was quantified via ion chromatography as described previously (Ranaivoarisoa et al., 2019). In brief, ion chromatography was performed using a Metrohm 881 Compact Pro with a Metrosep organic acid column (250 mm) at 27°C with 0.5 mM H₂SO₄/15% acetone as the eluent using a 10 mM LiCl suppression solution. The butyrate peak was detected at 21.1 min on the conductivity detector, which was confirmed with a standard butyrate solution at known concentrations. The standard butyrate solution was used to determine the concentration of butyrate in our samples.

Fe(II) measurement

Time‐course Fe(II) concentration was measured using the ferrozine assay as described previously (Bose & Newman, 2011).

Bioelectrochemical analyses

To measure microbial electron uptake during photoelectrotrophy, chronoamperometry was conducted for 120 h. (N₂‐fixing photoelectrotrophy) or 140 h. (non‐N₂‐fixing photoelectrotrophy). The working electrode was poised at +100 mV vs. SHE. Current response was assessed using Gamry Echem Analyst software (Gamry Instruments, Warmister, PA). We determined current response across two parameters: first, we measured current density vs. time, which shows the real‐time current response during bacteria‐electrode electron transfer. Second, we integrated the current density over time to determine the total current uptake in Coulombs (ampere second). Coulombs was then mathematically converted to total electrons transferred to perform electron conversion calculations (see Methods S1).

Identifying PHA cycle genes in Rhodomicrobium spp.

Rhodomicrobium vannielii ATCC17100 (JAEMUJ010000000) and R. udaipurense JA643 (JAEMUK010000000) genomes were downloaded from the National Center for Biotechnology Information database. We searched the downloaded genomes for putative PHA cycle genes with RAST's (https://rast.nmpdr.org/) BLASTp function using the amino acid sequences of known TIE‐1 homologues as queries. The genes coding for each protein and the corresponding locus tags were also identified.

RESULTS

R. vannielii and R. udaipurense genomes encode putative PHA synthesis genes

Rhodomicrobium vannielii and R. udaipurense, like Rh. palustris TIE‐1, are freshwater phototrophic PNSB in the order Hyphomicrobiales. Thus, we reasoned that the amino acid sequences of the proteins in TIE‐1's PHA biosynthesis pathway would be reasonable queries for homologues in the Rhodomicrobium spp. genomes (Ranaivoarisoa et al., 2019). After mining the Rhodomicrobium spp. genomes for homologues of TIE‐1 PHA biosynthesis proteins, we constructed a putative pathway based on the canonical PHA biosynthesis pathway (Figure 1A). In this hypothesized pathway, β‐ketothiolase (PhaA) condenses two acetyl‐CoAs into acetoacetyl‐CoA. Acetoacetyl‐CoA is then reduced to 3‐hydroxybutyryl‐CoA via acetoacetyl‐CoA reductase (PhaB). The PHA polymerase (PhaC) and depolymerase (PhaZ) are likely responsible for PHA polymerization and depolymerization, respectively (Quelas et al., 2016). PhaR plays a role in regulating PHA biosynthesis, though it has been characterized as both a repressor and activator of PHA synthesis in different species, with recent evidence in TIE‐1 suggesting that its role may vary based on the growth conditions (Maehara et al., 2001; Quelas et al., 2016; York et al., 2002; Ranaivoarisoa et al., 2024). BLASTp analysis of PhaR shows over 50% amino acid identity to putative and known PHA repressors, including PhaR in B. diazoeficiens (53% amino acid identity), suggesting that the PhaR homologues we identified may also repress PHA polymerase in Rhodomicrobium. PhaP likely dictates the number and size of PHA granules, based on BLASTp analysis showing a high degree (>50%) of amino acid identity to putative and known phasin proteins. In total, we identified five acetyl‐CoA acetyltransferases (phaA), one acetoacetyl‐CoA reductase (phaB), two PHA polymerases (phaC), three phasins (phaP), one PHA depolymerase (phaZ), and one PHA polymerase repressor (phaR) in each genome, though additional work is needed to confirm each protein's role in the biosynthetic pathway (Figure 1A). Both strains show similar genetic architecture, with phaA and phaB forming a putative operon nearby phaR, which is encoded on the opposite strand (Figure 1B).

FIGURE 1.

The putative PHA synthesis cycle and genes in R. vannielii and R. udaipurense. (A) The putative PHA synthesis cycle of R. vannielii (Rv) and R. udaipurense (Ru) based on the hypothesized PHA cycle of Rhodopseudomonas palustris TIE‐1. Acetyl‐CoA is produced from organic or inorganic carbon sources followed by condensation of two acetyl‐CoAs into acetoacetyl‐CoA by PhaA. NAD‐dependent acetoacetyl‐CoA reductase (PhaB) reduces acetoacetyl‐CoA, producing 3‐Hydroxybutyryl‐CoA. 3‐Hydroxybutyryl‐CoA is polymerized into PHA granules by PHA polymerase (PhaC). When cells call upon PHA granules for carbon and/or energy reserves, PHA depolymerase (PhaZ) degrades the granules back into 3‐Hydroxybutyrate and then to acetyl‐CoA via multiple enzymatic reactions (depicted as double white lines). PhaR likely represses the expression of PhaC, while phasins (PhaP) dictate size and number of PHA granules. Locus tags for each strains' homologues are listed under their respective genes. (B) The putative genes involved in the PHA synthesis cycle of R. vannielii (Rv) and R. udaipurense (Ru) based on the highest degree of similarity to TIE‐1 homologues. Locus tags corresponding to the genes with the highest similarity to those in TIE‐1 are listed below each gene for Rv and Ru. PHA: Polyhydroxyalkanoate. CoA‐SH: Coenzyme A. NADPH: Nicotinamide adenine dinucleotide phosphate. Adapted from Ranaivoarisoa et al. (2019).

Photoheterotrophy led to the fastest generation times and highest PHA production for both R. vannielii and R. udaipurense

We assessed PHA production across several metrics, all of which can be found in Table S1. Here, we highlight % cell dry weight based on total protein (% cdw_prot). We first investigated PHA production during photoheterotrophic growth on butyrate, which has been used as a substrate for PHA synthesis in diverse PNSB (Brandl et al., 1989; Carlozzi et al., 2019; Ghimire et al., 2016; Guerra‐Blanco et al., 2018; Ranaivoarisoa et al., 2019). Between both Rhodomicrobium species, photoheterotrophy resulted in the fastest generation times (15.2 (±2.3) – 25.5 (±2.3) hours, Table 1). PHA production ranged from 5.94% to 43.58% cdw_prot across all photoheterotrophic groups (Figure 2).

TABLE 1.

Growth metrics for Rhodomicrobium species under different carbon, electron, and nitrogen regimes.

Strain	Electron donor/substrate	Nitrogen source	Lag time (h)	Generation time (h)	p‐value
R. vannielii	H2 / CO2	N2	20	53.5 (2.5)	2.34E‐04
		NH4Cl	12	26.9 (0.47)
	Butyrate	N2	20.2	25.5 (2.3)	1.07E‐02
		NH4Cl	12.2	15.2 (2.3)
	FeCl2 / CO2	N2	40	342.1 (20.5)	5.22E‐02
		NH4Cl	34.5	464.4 (270.9)
	Carbon Felt / CO2	N2	70.1	140.3 (24.8)	2.50E‐01
		NH4Cl	41.8	83.6 (7.7)
R. udaipurense	H2 / CO2	N2	20	61.4 (6.3)	1.69E‐03
		NH4Cl	12	27.4 (1.4)
	Butyrate / CO2	N2	20.2	17.5 (0.3)	2.55E‐01
		NH4Cl	12.2	15.9 (1.2)
	FeCl2 / CO2	N2	34.5	401.0 (42.3)	3.15E‐04
		NH4Cl	22	120.3 (23.2)
	Carbon Felt / CO2	N2	52.1	117.1 (17.2)	3.15E‐04
		NH4Cl	58.6	104.3 (5.8)

Note: See Experimental Procedures for the description of generation time calculations. Lag time is defined as the period preceding exponential phase or half of the time required to reach the first doubling based on growth curves. For growth on carbon felt cathode, lag time is half of the generation time because cells immediately enter exponential phase and only initial and final optical densities were measured. N₂: nitrogen‐fixing growth with dinitrogen gas as sole nitrogen source. NH₄Cl: non‐nitrogen fixing growth with ammonium chloride as sole nitrogen source. p‐values are calculated between N₂ and NH₄Cl generation times for each respective condition. Significant p‐values are bolded. () = standard deviation values from N ≥ 3 except for R. vannielii grown on Carbon Felt / CO₂ + N₂ and growth on H₂CO₂ + NH₄Cl for both species (N = 2).

FIGURE 2.

PHA production by Rhodomicrobium spp. during photoautotrophic and photoheterotrophic growth. Intracellular PHA production as a percentage of cell dry weight based on total protein (cdw_prot) during photoautotrophic growth with the following electron and carbon sources: Hydrogen (H₂) gas and carbon dioxide (CO₂) (photohydrogenotrophy), ferrous chloride (FeCl₂) and CO₂ (photoferrotrophy), and sodium butyrate (photoheterotrophy). Cdw_prot was calculated as described in the Materials and Methods. N₂: Nitrogen‐fixing growth with dinitrogen gas as sole nitrogen source. NH₄Cl: Non‐nitrogen fixing growth with ammonium chloride as sole nitrogen source. p‐values are calculated between N₂ and NH₄Cl growth conditions. Statistically significant differences are indicated with horizontal bars and p‐values.

Next, we assessed photoautotrophic PHA production with two different electron donors: hydrogen gas (H₂, i.e., photohydrogenotrophy) and ferrous iron (FeCl₂, i.e., photoferrotrophy). Photohydrogenotrophy showed faster generation times (26.9 (±0.47) – 61.4 (±6.3) hours) compared to photoferrotrophy (120.3 (± 23.2) – 464.4 (± 270.9) hours). Photoferrotrophy led to slightly higher PHA production in both species (4.64%–47.03% cdw_prot) than photohydrogenotrophy (1.10%–6.19% cdw_prot), with overlap between non‐N₂‐fixing photohydrogenotrophy and N₂‐fixing photoferrotrophy (Figure 2). Altogether, these data confirm that Rhodomicrobium spp. produce PHA via photoheterotrophic growth on butyrate and photoautotrophic growth on CO₂ with two different electron donors, H₂ and Fe²⁺.

N₂ fixation promotes PHA accumulation during photoheterotrophy, but inhibits it during photoautotrophy

We also assessed the impact of nitrogen fixation on PHA production. Non‐N₂₋fixing growth was achieved by providing NH₄Cl in the growth medium. When NH₄Cl was not supplied, N₂ gas in the headspace served as the sole nitrogen source via N₂ fixation. Photoheterotrophic PHA production from butyrate was 5‐ to 7‐fold greater during N₂‐fixation (29.57%–43.58% cdw_prot) than with NH₄Cl (5.94%–5.95% cdw_prot) in both species (Figure 2). On the other hand, PHA production during photohydrogenotrophy and photoferrotrophy was greatest when media was supplemented with NH₄Cl (Table S1; Figure 2). Photoferrotrophic growth of R. vannielii led to 47.03% cdw_prot PHA under non‐N₂‐fixing conditions, compared to only 4.64% cdw_prot PHA during N₂‐fixation (p = 1.10E‐02) (Figure 2). Likewise, photoferrotrophic PHA production by R. udaipurense was higher with NH₄Cl supplementation (24.13% vs. 4.86% cdw_prot, p = 3.34E‐05) (Figure 2). The same trends were true of photohydrogenotrophy, with NH₄Cl stimulating greater PHA production than N₂. Altogether, these data show that the nitrogen source impacts PHA accumulation and that this effect is related to the carbon source, at least with respect to the conditions described here. Specifically, cells grown on butyrate produce more PHA when N₂ gas is the sole nitrogen source compared to NH₄Cl. Conversely, N₂ fixation has an inhibitory effect on PHA production when cells are using CO₂ as a carbon source.

Rhodomicrobium spp. produce comparatively low amounts of PHA via microbial electrosynthesis

Microbial electrosynthesis (MES) is a method of generating bioproducts using bacterial biocatalysts that accept electrons from poised cathodes. Here, we constructed an anoxic MES system with a 1 cm² carbon felt cathode poised at +100 mV continuously for 120 h (N₂‐fixing) or 144 h (non‐N₂‐fixing). During N₂‐fixing growth, R. vannielii achieved a maximum average current density of −3.97 μA/cm² (Figure 3B, Table S3) and a total charge transfer of −0.83 Coulombs (Figure S1; Table S3). When NH₄Cl was supplied, R. vannielii reached a slightly higher maximum current density of −5.77 μA/cm² with a corresponding charge transfer of −2.01 Coulombs (Figure 3B; Table S3). R. udaipurense reached a maximum average current density of −2.20 μA/cm² (Figure 3B; Table S3) with a corresponding charge transfer of −0.31 Coulombs (Figure S1; Table S3) during N₂‐fixing photoelectrotrophy. Under NH₄Cl supplementation, R. udaipurense reached a maximum current density of −1.86 μA/cm² and a corresponding charge transfer of −1.01 Coulombs (Figure 3B; Figure S1; Table S3). Throughout the growth period, we observed small increases in the optical density of the combined biofilm and planktonic fractions. R. vannielii showed a ΔOD₆₆₀ of 0.09 during N₂‐fixing growth and 0.31 when supplemented with NH₄Cl. R. udaipurense showed a ΔOD₆₆₀ of 0.18 during N₂‐fixing growth and 0.21 when supplemented with NH₄Cl (Table S16). These data agree with a previous report of phototrophic EEU by R. vannielii and R. udaipurense (Gupta et al., 2019) and indicate that phototrophic EEU stimulates the growth and division of Rhodomicrobium cells, albeit much less so than the other conditions we investigated.

FIGURE 3.

PHA production by Rhodomicrobium spp. during photoelectrotrophy. (A) Intracellular PHA production as a percentage of cell dry weight based on total protein (cdw_prot) during photoelectrotrophic growth using a carbon felt cathode poised at +100 mV versus standard hydrogen electrode. Cdw_prot was calculated as described in the Materials and Methods. p‐values are calculated between N₂ and NH₄Cl growth conditions. Statistically significant differences are indicated with horizontal bars and p‐values. (B) Current uptake by R. vannielii and R. udaipurense in microbial electrosynthesis reactors. N = 3 for biological MES reactors. N = 2 for abiotic controls. N₂: Nitrogen‐fixing growth with dinitrogen gas as sole nitrogen source. NH₄Cl: Non‐nitrogen fixing growth with ammonium chloride as sole nitrogen source.

In accordance with low biomass accumulation, we measured low intracellular PHA, along with a statistically significant effect of nitrogen source (Figure 3). PHA production by R. vannielii during N₂‐fixing photoelectrotrophic growth was about 6.6‐fold greater than growth with NH₄Cl (2.02% vs. 0.33% cdw_prot, p = 8.35E‐04) (Figure 3A). Likewise, R. udaipurense showed significantly more PHA accumulation during N₂‐fixing growth (0.78% vs. 0.24% cdw_prot, p = 1.47E‐02) (Figure 3A). From these data, we can conclude that (1) N₂‐fixing photoelectrotrophy stimulates greater PHA accumulation than growth with NH₄Cl and (2) photoelectrotrophy results in far less PHA accumulation than the other conditions. It is important to note that due to biomass limitations, measures of PHA, OD₆₆₀, and total protein from photoelectrotrophic cultures are representative of the combined planktonic and biofilm fractions. Therefore, while these data support the claim that Rhodomicrobium spp. conduct electron uptake from a poised cathode to support growth and PHA accumulation, the relative contribution of the biofilm compared to the planktonic cells remains unclear.

Carbon and electron conversion efficiency

We next sought to understand the efficiency with which cells convert carbon and electrons into intracellular PHA (see Methods S1 for detailed calculations). Carbon yields to PHA were highest during N₂ fixing growth on butyrate (5.85%–8.24%) and photoferrotrophic growth on NH₄Cl (5.12%–7.18%) (Table S2). The lowest carbon yields occurred during photoelectrotrophic growth, ranging from 0.04% to 0.58% (Table S2). In accordance with our observations surrounding PHA content and nitrogen source, photohydrogenotrophy and photoferrotrophy showed the highest carbon yields to PHA with NH₄Cl as a nitrogen source, while carbon yields during photoheterotrophy and photoelectrotrophy were greatest during N₂‐fixation (Table S2). We performed one‐way ANOVA tests to determine if either strain shows a statistically significant difference in carbon yields among all growth conditions. We found statistically significant differences in the carbon yield values for R. vannielii (F(7, 15) = 50.45, p = 2.67E‐09) and R. udaipurense (F(7, 15) = 4.78, p = 0.005) (Tables S4 and S5). Games‐Howell post hoc tests for multiple comparisons revealed significantly higher carbon yields during N₂‐fixing photoheterotrophy in R. vannielii compared to most other growth conditions (Table S6). For R. udaipurense, we observed significantly greater carbon yields under photoferrotrophy compared to photoelectrotrophy and N₂‐fixing photohydrogenotrophy (Table S7). Finally, all conditions showed carbon yields <10%, suggesting that PHA is a relatively small carbon sink regardless of the carbon, electron, or nitrogen source (Table S2).

We also calculated electron yields based on the ratio of electrons predicted to go toward PHA production/total consumed electrons (see Methods S1 for detailed calculations). Both strains showed the highest electron yields during photoautotrophic growth, with photohydrogenotrophy + NH₄Cl peaking at 35.34% and 46.64% for R. vannielii and R. udaipurense, respectively (Table S2). Additionally, we calculated electron yields >20% during N₂‐fixing photoelectrotrophy and non‐N₂‐fixing photoferrotrophy, suggesting that these conditions are also effective in directing electrons toward PHA (Table S2). Conversely, photoheterotrophy resulted in the lowest electron yields, ranging from 0.27% to 1.39% (Table S2). One‐way ANOVA confirmed significant differences among the respective electron yields of R. vannielii (F(7, 15) = 18.31, p = 2.74E‐06) and R. udaipurense (F(7, 14) = 69.52, p = 8.51E‐10) (Tables S8 and S9). However, Games‐Howell post hoc tests revealed few statistically significant differences among the mean electron yields for each species (Tables S10 and S11). While we cannot determine any statistically significant differences between the electron yields of specific growth conditions, a qualitative comparison suggests that a greater fraction of electrons is diverted toward PHA during photosynthetic CO₂ fixation compared to photoheterotrophy (Figure S2).

Taken together, the carbon and electron yield data imply a complex dynamic between carbon and electron conversion in the context of PHA accumulation, with photoheterotrophy favouring PHA carbon yield and photoautotrophy favouring PHA electron yield. To better understand how carbon and electron yields relate to PHA accumulation, we conducted linear regression analyses. Carbon yield showed a strong correlation to PHA % cdw_prot (R ² = 0.956, p = 6.84E‐11), whereas electron yield did not correlate with PHA % cdw_prot (R ² = 3.88E‐04, p = 0.942) (Tables S12 and S13). To visualize these relationships, we plotted the carbon and electron yield values against the corresponding PHA % cdw_prot values on x‐y axes. We then separated the plot into four quadrants by bisecting each axis at one‐half of the largest value in the x and y directions. This resulted in four groups for categorizing growth conditions based on their respective carbon/electron yields and PHA production: (1) low carbon/electron yield and low PHA, (2) high carbon/electron yield and low PHA, (3) low carbon/electron yield and high PHA, and (4) high carbon/electron yield and high PHA (Figure 4). Figure 4A illustrates that N₂‐fixing photoheterotrophy and photoferrotrophy + NH₄Cl exhibit high carbon yields with moderate to high PHA production (Group 4). Figure 4B shows that photoferrotrophy and photohydrogenotrophy supplemented with NH₄Cl along with N₂‐fixing photoelectrotrophy tend to fall within Groups 2 and 4 due to their high electron yields and low to high PHA production. Cells grown on butyrate show drastically lower electron yields with moderate to high PHA yields and fall into Groups 1 and 3 (Figure 4B). Altogether, this framework shows that carbon and electron yields are useful parameters by which to classify PHA production across different microbes and growth conditions, even in cases where the parameters do not clearly correlate with PHA production.

FIGURE 4.

Carbon and electron yields compared to PHA production. (A) PHA carbon yield (%) and (B) PHA electron yield (%) vs. PHA production as a percentage of cell dry weight (cdw_prot). Samples have been grouped by bisecting the horizontal and vertical axes at one‐half the maximum x and y values, respectively. Trendline and corresponding R ² value reflects all samples. Cdw_prot was calculated as described in the Materials and Methods. Blue, Rv: R. vannielii. Orange, Ru: R. udaipurense. B: Butyrate. H₂: Hydrogen electron donor. Fe: FeCl₂. E: Poised electrode. Triangles, N₂: Nitrogen gas. Circles, NH₄: Ammonium chloride.

Rhodomicrobium spp. exhibit superior electron yields and PHA production compared to Rh. Palustris TIE‐1

Previously, our lab explored PHA production by the photosynthetic PNSB Rhodopseudomonas palustris (Rh. palustris) TIE‐1 grown in the same conditions and assessed using similar experimental and analytical methods (Ranaivoarisoa et al., 2019). This methodological consistency allows us to compare PHA accumulation in two PNSB with similar metabolic capabilities but different genomic and morphological contexts. Superimposing the normalized PHA production data from Ranaivoarisoa et al. (2019) onto our data in Figure 4 reveals that both Rhodomicrobium species contained up to 5 to 6 orders of magnitude more PHA per cell than TIE‐1 (Figure 5). Although carbon yields tend to be similar between all three species (Figure 5A), Rhodomicrobium exhibits comparable or superior electron yields to TIE‐1 (Figure 5B). In fact, half of our Rhodomicrobium cultures showed average electron yields greater than 10%, with some as high as 35%–46% (Table S2). Conversely, the majority of TIE‐1's electron yields across the same growth conditions were below 1% (Ranaivoarisoa et al., 2019). Although we did not perform statistical tests to compare the electron yields from the present study with those reported by Ranaivoarisoa et al., Rhodomicrobium spp. appear to show higher electron yields than TIE‐1 across almost all comparable growth conditions (Figure S2). Overall, our data indicates that, in certain contexts, Rhodomicrobium converts electrons into PHA more effectively than TIE‐1 while also managing much more PHA production per cell. However, we cannot conclusively determine a causal link between these phenomena.

FIGURE 5.

Rhodomicrobium spp. and TIE‐1 carbon and electron yields compared to normalized PHA production. (A) PHA carbon yield (%) and (B) PHA electron yield (%) vs. PHA production per litre normalized to calculated cell numbers. Samples have been grouped by bisecting the horizontal and vertical axes at one‐half the maximum x and y values, respectively. Black: R. palustris TIE‐1 (Ranaivoarisoa et al., 2019). Blue: R. vannielii. Orange: R. udaipurense. Triangles: Nitrogen gas. Circles: Ammonium chloride.

DISCUSSION

Conventional plastics impose significant environmental and economic costs throughout their lifetime. Biopolymers like PHA that are derived from renewable materials are promising alternatives. However, microbially synthesized PHA faces barriers to commercialization including low‐product yields, substantial capital and operational costs, and narrow metabolisms for bioproduction. Non‐conventional microbes may provide insight into these challenges. In that vein, we investigated two novel biological chassis in the genus Rhodomicrobium that produce PHA using sustainable carbon sources, minimal nutrients, and energy from light. To the best of our knowledge, this is the first quantitative analysis of PHA production by Rhodomicrobium spp.

Photoheterotrophic PHA production

The first substrate we assessed was sodium butyrate. Butyrate and similar organic compounds are known to be ideal substrates for PHA production in phototrophs (Alsiyabi et al., 2021; Carlozzi et al., 2019; Luongo et al., 2017; Monroy & Buitrón, 2020; Ranaivoarisoa et al., 2019). However, it is important to consider the economic and environmental costs associated with heterotrophic bioproduction. Butyrate is a costlier substrate than others (e.g., acetate, methanol/CO₂, and H₂/CO₂), although the total feedstock costs of bioproduction can change based on nitrogen source and the market price of electron donors (Alloul et al., 2023). There are also environmental concerns such as land and water use, greenhouse gas emissions, and soil and water health (Koch et al., 2023). These costs can be mitigated with sustainable practices aimed at circularizing the bioproduction of PHAs. For example, butyric acid can be sourced via microbial fermentation using feedstocks like corn husk (Xiao et al., 2018), straw (Baroi et al., 2015), corn fibre (Zhu et al., 2002), oilseed rape straw (Huang et al., 2016), and sugarcane bagasse (Wei et al., 2013), as well as waste streams and anaerobic digestors (Atasoy et al., 2018). Moreover, certain microbes can convert CO₂ to butyrate, though this technology is in its infancy (Batlle‐Vilanova et al., 2017; Benito‐Vaquerizo et al., 2022; Romans‐Casas et al., 2024; Ueki et al., 2014). In any case, there are more sustainable methods for obtaining organic feedstocks for PNSB than chemical synthesis, which can drastically impact the economics of heterotrophic bioproduction (Alloul et al., 2023). This is especially true considering that raw materials account for over half of production costs for biopolymers and 70%–80% of raw materials costs are attributed to the carbon source alone (Sirohi et al., 2021). Therefore, circularizing the bioproduction pipeline through sustainable sourcing of organic feedstocks is key to making heterotrophic bioproduction economically competitive.

Setting aside the environmental and economic considerations, butyrate and similar organic compounds are known to be ideal substrates for PHA production in phototrophs (Alsiyabi et al., 2021; Carlozzi et al., 2019; Luongo et al., 2017; Monroy & Buitrón, 2020; Ranaivoarisoa et al., 2019). Our data agree with this body of literature. We observed some of the highest PHA yields when cells were grown on butyrate, especially with N₂ as a nitrogen source (Figure 2). As described by Alsiyabi et al. (2021), this is likely because butyrate is a favourable organic substrate given the constraints of the PHA biosynthesis pathway. When PHA biosynthesis commences, the thiolase enzyme (PhaA) converts acetyl‐CoA to acetoacetyl‐CoA (Figure 1). This is a thermodynamically unfavourable reaction (ΔG′ ⁰ = 26 kJ/mol) that requires a large amount of acetyl‐CoA to proceed (>12 acetyl‐CoA per 1 CoA) (Alsiyabi et al., 2021). However, butyrate breakdown produces acetoacetyl‐CoA, thus bypassing phaA and negating the thermodynamic burden. Future studies of Rhodomicrobium and other PNSB with homologous biosynthetic pathways should take advantage of this by exploring other substrates that are predicted to bypass the PhaA step, e.g., p‐coumarate (Alsiyabi et al., 2021). Additionally, genetic tools can be used to evaluate this claim by assessing PHA production in phaA knockout mutants grown on substrates expected to bypass the thiolase enzyme. Following the PhaA reaction, PhaB catalyses acetoacetyl‐CoA reduction to 3‐hydroxybutyryl‐CoA (Figure 1). This is considered the rate‐limiting step in PHA biosynthesis and requires a high NAD(P)H/NAD(P) ratio (Alsiyabi et al., 2021; Tyo et al., 2010). Butyrate, as a highly reduced substrate, likely results in a highly reduced intracellular environment, thus providing the reducing equivalents needed to drive the PhaB reaction (Alsiyabi et al., 2021; Guzman et al., 2019). Again, future studies can use genetic tools along with assays tracking the intracellular redox state to validate this claim in Rhodomicrobium and other PNSB. Assuming that these dynamics are at play in Rhodomicrobium, our findings support the hypothesis that butyrate and other carbon‐ and electron‐rich substrates are well‐suited to overcome the thermodynamic limitations inherent to PHA biosynthesis. Future studies should explore PHA synthesis using substrates with similar characteristics acquired from existing waste streams (e.g., lignocellulose breakdown products). Not only would this be useful for a sustainability but it would also shed light on how PNSB metabolize different organic substrates.

Photoautotrophic PHA production

We also investigated photoautotrophic metabolisms relying on CO₂ and one of three electron donors: H₂, Fe²⁺, or a poised carbon felt cathode. Compared to photoheterotrophy, photoautotrophy generally resulted in less PHA production, apart from photoferrotrophy + NH₄Cl (Figure 2). This is in line with the idea that more reduced substrates like butyrate lead to more PHA production than oxidized substrates like CO₂ (Alsiyabi et al., 2021). Despite this, we found similar carbon yields and superior electron yields across most photoautotrophic grown conditions compared to photoheterotrophy. Considering this along with the environmental upsides of CO₂ as a carbon source (e.g., mitigating greenhouse gas emissions and non‐competition with agricultural feedstocks), photoautotrophy's strength may lie in its efficiency and sustainability. But to make photoautotrophic PHA production feasible, biomass and PHA yields must improve. One approach is to exploit genetic engineering. A recent study found that photohydrogenotrophic (+NH₄Cl) PHA production by Rh. palustris TIE‐1 increased two‐told by overexpressing the RuBisCO genes responsible for CO₂ fixation, deleting the two nifA genes that activate the nitrogen fixating machinery, or inhibiting glycogen production by deleting the glycogen synthase (gly) (Ranaivoarisoa et al., 2024). This suggests that overexpressing the enzymes responsible for photosynthetic CO₂ fixation and/or removing competing electron sinks like nitrogen fixation or glycogen synthesis could also promote greater PHA yields in Rhodomicrobium, though this effect may be context‐dependent (see below). Other studies suggest alternative genetic engineering strategies for enhancing PHA yields in purple bacteria. For example, heterologously overexpressing the membrane‐bound transhydrogenase PntAB from Escherichia coli MG1655 and phaB1 gene from Ralstonia eutropha H16 allowed Rhodospirillum rubrum S1 to synthesize 5.1% cdw poly(3‐hydroxybutyrate‐co‐3‐hydroxyvalerate) from CO₂ and syngas (Heinrich et al., 2015). The authors noted a 13‐fold increase of 3‐hydroxyvalerate content compared to wild‐type, corresponding to a decrease in the brittleness of the biopolymer. This suggests that heterologously expressing genes important for PHA synthesis from distantly related microbes can favourably alter the amount and composition of intracellular autotrophically synthesized biopolymers. Crucially, it underscores the need to explore the genetics and biochemistry of PHA production in non‐model organisms like Rhodomicrobium spp. not only to design more efficient PHA producers but to better understand the underlying physiology.

PHA production via microbial electrosynthesis (MES)

MES is a unique instance of photoautotrophic growth with its own considerations. We observed the lowest PHA production during photoelectrotrophy (Figure 3), which may be explained in part by different sub‐populations of cells within electrochemical reactors. Unlike the other growth conditions that consist solely of suspended cells, the single‐chamber electrochemical reactors are home to both planktonic cells and a biofilm attached to a poised cathode. These populations likely experience significant differences in electron uptake, carbon acquisition, biomass accumulation, and other factors relevant for PHA production. For example, simultaneous direct and indirect electron transfer processes are likely at play, with the former dominating biofilms and the latter dominating planktonic cells (Bai et al., 2020; Conners et al., 2022; Ha et al., 2017; Karthikeyan et al., 2019; Nevin et al., 2010; Villano et al., 2010). If this leads to differences in the intracellular redox state, biofilms and planktonic cells may have different intracellular PHA content. Additionally, biofilms and planktonic cells experience different microenvironments (Beyenal & Babauta, 2012; Bisht & Wakeman, 2019; Imdahl et al., 2020; Mann & Wozniak, 2012; Moormeier et al., 2013; Obana et al., 2020). Localized nutrient limitation within the biofilm could promote faster PHA accumulation relative to the planktonic fraction, where nutrients are homogenously distributed throughout the medium. Finally, biofilms interact dynamically with planktonic cells, which likely complicates the relative contribution of each population to total biomass, final PHA titers, and the carbon/nitrogen/electron acquisition dynamics (Berlanga & Guerrero, 2016; Song et al., 2022). In any case, any potential differences between the two populations would be lost when they are combined for bulk measurements, as was the case in our study due to biomass limitations. Future work with large‐scale reactors can achieve the necessary biomass to reliably measure PHA accumulation in the biofilm and planktonic populations separately. Ultimately, these gaps in knowledge highlight the need to better understand how biofilms and planktonic cells behave during MES, particularly as it pertains to PHA synthesis.

The role of nitrogen and redox homeostasis in governing PHA content

In accordance with previous studies, we observed that nitrogen source (NH₄Cl vs. N₂) is an important determinant of PHA production (Ciesielski et al., 2010; Johnson et al., 2010; Montiel‐Jarillo et al., 2017; Ranaivoarisoa et al., 2019; Schlegel et al., 1961; Zhou et al., 2022). Specifically, we found that N₂ fixation stimulated greater PHA production during photoheterotrophic and photoelectrotrophic growth, while cells grown with NH₄Cl accumulated more PHA during photohydrogenotrophy and photoferrotrophy. To understand the potential mechanisms underlying these observations, it is important to consider how PNSB maintain redox homeostasis. In general, PNSB use biosynthesis, CO₂ fixation, PHA accumulation, H₂ production (via N₂ fixation), and organic acid secretion as electron sinks to maintain redox balance. Aside from biosynthesis, cells preferentially allocate reducing power toward the Calvin cycle and PHA production, which are more thermodynamically favourable than N₂ fixation (Alloul et al., 2023; McKinlay & Harwood, 2011). Although we cannot conclusively determine the redox hierarchies at play within each growth condition, our data support the idea that different carbon, electron, and nitrogen sources impose different redox stress, which ultimately effects PHA content.

Photoheterotrophic growth on an electron‐rich substrate like butyrate results in a highly reduced intracellular environment (Alsiyabi et al., 2021; Guzman et al., 2019). Thus, cells need to maintain redox homeostasis by diverting reducing power toward the electron sinks mentioned above (Alloul et al., 2023). Supplementing growth media with NH₄Cl decreases nitrogenase activity, thereby removing N₂ fixation as an outlet for electrons and potentially increasing electron flux through the other pathways. In this case, we would predict increased biomass, CO₂ fixation, and PHA synthesis in cultures with NH₄Cl compared to N₂. In accordance with this, photoheterotrophic cultures supplemented with NH₄Cl showed slightly more net CO₂ fixation normalized to biomass (gas chromatography data not shown) and faster generation times (Table 1). However, we saw lower PHA electron yields (Figure S2) and lower intracellular PHA content (Figure 2). Together, this suggests that eliminating N₂ fixation does not necessarily redirect reducing power toward PHA synthesis. In fact, it has been suggested that PNSB do not use PHA synthesis as an electron sink when grown on butyrate because there is no cofactor re‐oxidation (Alloul et al., 2023; Cabecas Segura et al., 2021). Therefore, removing electron sinks may not be an effective strategy for improving PHA production in PNSB grown on butyrate alone. This is corroborated by a recent genetic study in Rh. palustris TIE‐1 wherein deleting the nifA genes responsible for activating the nitrogen fixating machinery significantly decreased photoheterotrophic PHA synthesis from butyrate with NH₄Cl (Ranaivoarisoa et al., 2024). Alternatively, it may be useful to view N₂ as imposing a nutrient limitation, thereby stimulating PHA synthesis (Johnson et al., 2010; Ranaivoarisoa et al., 2019; Zhou et al., 2022). Additionally, nitrogenase activity may provide a much‐needed outlet for reducing power when cells are catabolizing butyrate, which likely benefits overall cellular health and facilitates carbon assimilation to biomass and PHA. Still, it remains unclear whether and to what extent the various redox balancing pathways are at play in Rhodomicrobium. Time‐course quantification of total organic and inorganic carbon along with intracellular redox state would paint a more complete picture of overall carbon and electron flux. Relatedly, time‐course differential gene expression analyses could shed light on the activation of carbon and electron‐sinking pathways. While the small‐scale experiments here clearly show that N₂ fixation stimulates PHA accumulation during growth on butyrate (Figure 2) and that this condition leads to some of the highest carbon yields we observed (Figure 4), large‐scale experiments that generate sufficient biomass for more extensive sampling are needed to answer questions about the underlying mechanisms.

Contrary to photoheterotrophy, PHA synthesis was higher during photohydrogenotrophy and photoferrotrophy with NH₄Cl (Figure 2). With CO₂ as a carbon source, we must recontextualize N₂ fixation and PHA biosynthesis with respect to redox balance. As noted above, CO₂ fixation and N₂ fixation are among the primary electron sinks employed by PNSB (Alloul et al., 2023). Furthermore, NH₄Cl assimilation is far less energetically intensive than N₂ fixation (Alloul et al., 2023). Therefore, supplementing media with NH₄Cl likely has the dual effects of reducing the energetic burden of nitrogen assimilation and nullifying N₂ as a competing electron sink, thereby freeing reducing equivalents and ATP for CO₂ fixation and PHA synthesis. On the other hand, N₂‐fixing growth with CO₂ likely results in electrons and energy being spread thin between N₂ fixation and the Calvin cycle, with little electron or carbon mobilization toward PHA. This suggests that PHA synthesis using CO₂ as a carbon source could benefit from the removal of competitive electron sinks. In accordance with this, Ranaivoarisoa et al. (2024) found that deleting the nitrogenase activating machinery (nifA) or the glycogen synthase (gly) from Rh. palustris TIE‐1 improved photohydrogenotrophic PHA production with NH₄Cl. On the other hand, the same study found that these mutants resulted in less (gly) or similar (nifA) PHA during photoferrotrophy with NH₄Cl. Therefore, the relationship between various electron sinks may change based on the electron donor in question. It is also important to consider how PHA synthesis overlaps with various regulatory processes, which can be elucidated via metabolic modelling and experimental validation (Alsiyabi et al., 2021; Hädicke et al., 2011).

This reasoning may imply that photoelectrotrophic growth – wherein CO₂ is the sole carbon source – would follow a similar PHA accumulation pattern to other CO₂‐fixing conditions. However, photoelectrotrophically grown cells showed higher PHA synthesis when fixing N₂ (Figure 3). In this case, the constant supply of reducing power from a poised cathode likely imposes substantial redox stress compared to the other phototrophic growth conditions we examined (Guzman et al., 2019). In the absence of N₂ fixation, redox stress may be too severe, thus harming overall cellular health and hindering the cell's ability to assimilate carbon to biomass or PHA. This relationship between N₂ fixation and PHA accumulation during photoelectrotrophy reflects previous observations in Rh. palustris TIE‐1, suggesting conservation across related electrotrophic PNSB (Ranaivoarisoa et al., 2019; Ranaivoarisoa et al., 2024). Overall, our data support the idea that the highly reduced growth conditions imposed by electrotrophy may be harmful to cells' ability to assimilate carbon to biomass and PHA and that N₂ fixation has a moderately beneficial effect on PHA content.

The mechanisms governing redox homeostasis and PHA production in Rhodomicrobium are speculative and require further study. As alluded to above, it is especially important to understand cell physiology over time. We sampled all conditions in mid‐ to late‐exponential phase, when PHA content is predicted to peak just before cells begin consuming it during the stationary phase (Segura et al., 2022). This also ensured sufficient biomass for sampling and allowed us to quantify PHA across comparable growth phases. However, PHA production may start, peak, and decrease at distinct stages of growth based on (1) the sources of carbon and nitrogen, (2) the carbon/nitrogen (C/N) ratio, and (3) redox stress. For example, previous studies indicate that C/N ratios impact PHA accumulation in pure and mixed microbial cultures (Bhatia et al., 2018; Cui et al., 2017; Johnson et al., 2010; Montiel‐Jarillo et al., 2017; Yang et al., 2010; Zhou et al., 2022). Thus, changing C/N ratios over the course of growth could dynamically impact PHA abundance, which our one‐time sampling strategy does not capture. Given these complexities, future studies should track total carbon, nitrogen, electron donors, and PHA intermittently throughout growth. Similarly, it is necessary to measure redox state within the cell, e.g., via time‐course NAD(P)H/NAD(P) quantification. Acquiring these data will require large‐volume cell cultures to meet biomass requirements for sampling, particularly in photoautotrophic conditions that generate low biomass (e.g., N₂‐fixing photoferrotrophy). This will also make direct biomass measurements more feasible, which will improve the accuracy of PHA normalization and allow us to calculate carbon and electron flow to biomass with more certainty. At present, our approach is well suited to determine relative amounts of PHA between large numbers of samples; but given the proxy measures and assumptions being used, it is less reliable to make broader assessments (e.g., about overall carbon and electron flux). In the meantime, small‐scale experiments like those described here can provide insights into a wide range of growth conditions and strains, which informs how best to dedicate resources toward large‐scale experiments.

Comparing Rhodomicrobium spp. and Rh. palustris TIE‐1

Finally, we compared Rhodomicrobium spp. to a related PNSB, Rh. palustris TIE‐1. The former is a severely understudied PNSB, while the latter is commonly studied as a model for PNSB metabolism with established genetic and biochemical tools. Our comparison shows that TIE‐1 underperforms along multiple PHA bioproduction metrics relative to Rhodomicrobium spp. Therefore, we argue that non‐model organisms like Rhodomicrobium warrant further study to expand our understanding of PNSB metabolism, particularly as pertains to bioproduction. The most salient difference between TIE‐1 and Rhodomicrobium is their morphology; TIE‐1 are rod‐shaped cells, while Rhodomicrobium's polymorphic cell cycle includes ovoid cells interconnected by stalks and hyphae, motile swarmer cells, and exospores (Richter et al., 2023; Whittenbury & Dow, 1977). Rehm and Qi (2001) indicated that Caulobacter crescentus, a distantly related Alphaproteobacteria with similar prosthecate morphology, accumulates PHA granules primarily in stalked cells in the presence of excess carbon. But it remains unclear whether different Rhodomicrobium cell types accumulate PHA at different rates. Questions such as this could be addressed by developing high‐resolution imaging protocols, culturing techniques, and reliable genetic tools in Rhodomicrobium, the latter of which have recently been explored in the context of cell differentiation and magnetosome biosynthesis (Dziuba et al., 2024; Richter et al., 2023). Additionally, the molecular mechanisms governing Rhodomicrobium's metabolic flexibility are poorly understood. For example, although R. vannielii and R. udaipurense encode homologues of the PioABC protein complex responsible for extracellular electron uptake (EEU) from solid‐phase conductive substances, PioABC‐mediated EEU has only been explored in Rh. palustris TIE‐1 (Bose et al., 2014; Gupta et al., 2019; Guzman et al., 2019). The availability of genomes for the strains used in this study will facilitate inquiries into the molecular underpinnings of these metabolisms, which can lead to strain design for better bioproduction (Conners et al., 2021).

CONCLUSIONS

Here, we present the first quantitative analysis of PHA production by Rhodomicrobium vannielii and Rhodomicrobium udaipurense under various carbon, nitrogen, and electron regimes. We report the highest PHA titers during photoheterotrophic growth on sodium butyrate, despite resulting in the lowest electron yields. On the other hand, we find comparable carbon conversion and superior electron conversion during growth on CO₂. We also show that Rhodomicrobium spp. are highly efficient at converting electrons to PHA compared to a related phototroph, Rh. palustris TIE‐1, and that Rhodomicrobium cells appear to accumulate a large amount of PHA per cell. The trends we describe are likely the result of redox homeostasis mechanisms, which require additional study to unravel. This study paves the way for future investigations of Rhodomicrobium as a potential biological chassis for PHA production and as a model for unique bacterial metabolisms.

AUTHOR CONTRIBUTIONS

Eric M. Conners: Conceptualization; data curation; formal analysis; funding acquisition; investigation; methodology; validation; visualization; writing – original draft; writing – review and editing. Karthikeyan Rengasamy: Data curation; formal analysis; investigation; methodology; software; validation; visualization; writing – original draft; writing – review and editing. Tahina Ranaivoarisoa: Data curation; methodology; validation; visualization; writing – review and editing. Arpita Bose: Conceptualization; funding acquisition; resources; software; validation; writing – original draft; writing – review and editing.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supporting information

Figure S1.

Figure S2.

Figure S3.

Figure S4.

Figure S5.

Data S1.

Table S1.

Table S2.

Table S3.

Table S4.

Table S5.

Table S6.

Table S7.

Table S8.

Table S9.

Table S10.

Table S11.

Table S12.

Table S13.

Table S14.

Table S15.

Table S16.

Conners, E.M. , Rengasamy, K. , Ranaivoarisoa, T. & Bose, A. (2024) The phototrophic purple non‐sulfur bacteria Rhodomicrobium spp. are novel chassis for bioplastic production. Microbial Biotechnology, 17, e14552. Available from: 10.1111/1751-7915.14552

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.