Rhodobacter sphaeroides Uses a Reductive Route via Propionyl Coenzyme A To Assimilate 3-Hydroxypropionate

Abstract

3-Hydroxypropionate is a product or intermediate of the carbon metabolism of organisms from all three domains of life. However, little is known about how carbon derived from 3-hydroxypropionate is assimilated by organisms that can utilize this C₃ compound as a carbon source. This work uses the model bacterium Rhodobacter sphaeroides to begin to elucidate how 3-hydroxypropionate can be incorporated into cell constituents. To this end, a quantitative assay for 3-hydroxypropionate was developed by using recombinant propionyl coenzyme A (propionyl-CoA) synthase from Chloroflexus aurantiacus. Using this assay, we demonstrate that R. sphaeroides can utilize 3-hydroxypropionate as the sole carbon source and energy source. We establish that acetyl-CoA is not the exclusive entry point for 3-hydroxypropionate into the central carbon metabolism and that the reductive conversion of 3-hydroxypropionate to propionyl-CoA is a necessary route for the assimilation of this molecule by R. sphaeroides. Our conclusion is based on the following findings: (i) crotonyl-CoA carboxylase/reductase, a key enzyme of the ethylmalonyl-CoA pathway for acetyl-CoA assimilation, was not essential for growth with 3-hydroxypropionate, as demonstrated by mutant analyses and enzyme activity measurements; (ii) the reductive conversion of 3-hydroxypropionate or acrylate to propionyl-CoA was detected in cell extracts of R. sphaeroides grown with 3-hydroxypropionate, and both activities were upregulated compared to the activities of succinate-grown cells; and (iii) the inactivation of acuI, encoding a candidate acrylyl-CoA reductase, resulted in a 3-hydroxypropionate-negative growth phenotype.

INTRODUCTION

The C₃ compound 3-hydroxypropionate (CH₂OH-CH₂-COO⁻) is increasingly being recognized as an important intermediate or end product of carbon metabolism in a variety of organisms. So far, there are at least five known metabolic processes involving 3-hydroxypropionate. One process is propionyl coenzyme A (propionyl-CoA) metabolism in plants. Propionyl-CoA is derived from the breakdown of chlorophyll, odd-chain fatty acids, or amino acids like isoleucine and is oxidized to 3-hydroxypropionate and probably further oxidized to acetyl-CoA (18, 30, 36). Some animals and algae may also metabolize propionate via a similar route (11, 20, 28). Another process involves autotrophic CO₂ fixation pathways. In bacteria and archaea, the reductive conversion of acetyl-CoA and CO₂ to propionyl-CoA via 3-hydroxypropionate is part of two CO₂ fixation pathways; however, different enzymes are used in either pathway to catalyze the common steps in the conversion of acetyl-CoA and CO₂ to propionyl-CoA (8, 9, 21, 39). For example, the reductive conversion of 3-hydroxypropionate to propionyl-CoA is catalyzed by a fusion protein, named propionyl-CoA synthase, in Chloroflexus aurantiacus (3-hydroxypropionate bi-cycle), whereas Metallosphaera sedula (hydroxypropionate/4-hydroxybutyrate cycle) requires three separate enzymes to catalyze the same reaction sequence (1, 42). A third process is bacterial dimethylsulfonopropionate (DMSP) metabolism. 3-Hydroxypropionate is also an intermediate in the metabolism of the secondary metabolite DMSP by microorganisms (5, 44). DMSP is synthesized by marine algae and some land plants, and there are currently three different mechanisms known for the initial step of DMSP degradation: demethylation to methylmercaptopropionate (23, 24), cleavage by a DMSP lyase into dimethylsulfide and acrylate (5, 12, 45), and the cleavage of DMSP into 3-hydroxypropionate and dimethylsulfide by an unusual CoA-transferase (43, 44). The acrylate or 3-hydroxypropionate generated from the cleavage of DMSP may be further metabolized to acetyl-CoA and CO₂; in the case of acrylate, this proceeds via 3-hydroxypropionate (5, 6, 44). However, some bacteria use DMSP solely as a sulfur source and may therefore release 3-hydroxypropionate or acrylate as an end product (19). Another metabolic processes involving 3-hydroxypropionate is uracil degradation. 3-Hydroxypropionate has also been identified as the end product of two different pathways for uracil degradation in bacteria like Escherichia coli as well as in the yeast Saccharomyces kluyveri (4, 29, 32). Yet another metabolic process involving 3-hydroxypropionate is the anaerobic metabolism of glycerol. There have also been reports of 3-hydroxypropionate formation by the fermentation of glycerol by lactic acid bacteria (31, 41) and the anaerobic oxidation of glycerol by a sulfate-reducing bacterium (33). In summary, 3-hydroxypropionate is likely to play an important role in the overall carbon cycle as an end product or intermediate in the carbon metabolism of a variety of compounds.

Rhodobacter sphaeroides is an abundant purple nonsulfur bacterium that utilizes a variety of carbon substrates and likely encounters 3-hydroxypropionate in its environment, as this compound is released from other organisms. In addition, R. sphaeroides strain 2.4.1 was recently shown to liberate dimethylsulfide from DMSP and to contain a DddL-type DMSP lyase; the cleavage of DMSP would lead to the formation of acrylate, the dehydration product of 3-hydroxypropionate (14). Here we show that R. sphaeroides 2.4.1 is able to utilize 3-hydroxypropionate as a sole carbon source. For R. sphaeroides, it is possible to separate carbon assimilation from energy metabolism during photoheterotrophic growth, thereby allowing the monitoring of the sole conversion of 3-hydroxypropionate into cell carbon. If 3-hydroxypropionate were exclusively oxidized to acetyl-CoA and CO₂, a functional ethylmalonyl-CoA pathway for acetyl-CoA assimilation would be essential to convert acetyl-CoA to other precursor metabolites needed for cell carbon biosynthesis. However, a mutant with a deletion of ccr, encoding crotonyl-CoA carboxylase/reductase, a key enzyme of the ethylmalonyl-CoA pathway, had no effect on 3-hydroxypropionate-dependent growth. Instead, acuI, encoding a candidate acrylyl-CoA reductase, was shown to be essential for growth with 3-hydroxypropionate and for the reductive conversion of 3-hydroxypropionate to propionyl-CoA. Propionyl-CoA is then further converted to succinyl-CoA to enter the central carbon metabolism.

MATERIALS AND METHODS

Materials.

3-Hydroxypropionate was purchased from Tokyo Chemical Industry (TCI) America (Portland, OR) and, according to the manufacturer, also contained 3,3′-oxydipropionic acid. Crotonyl-CoA and propionyl-CoA were synthesized from their anhydrides (38).

Bacterial strains and growth conditions.

R. sphaeroides strain 2.4.1 (DSMZ 158) was grown at pH 6.7 and at 30°C aerobically in the dark or in 2-liter bottles anaerobically in the light (3,000 lx) in defined medium supplemented with 10 mM the appropriate carbon source as described previously (3). Growth was monitored by determining the optical density at 578 nm (OD₅₇₈), and cells were harvested in the mid-exponential phase at an OD₅₇₈ of 0.5 to 0.8. For growth studies, R. sphaeroides mutant and the wild-type cells were pregrown anaerobically in 10 ml minimal medium containing 10 mM sodium succinate, and 0.1 ml was transferred into stoppered screw-cap (Hungate) tubes with 10 ml minimal medium and the appropriate carbon source. Cells of the mutant strain were grown in the presence of 20 μg ml⁻¹ kanamycin and 25 μg ml⁻¹ spectinomycin, if appropriate. Escherichia coli strains DH5α and SM10 were grown in Luria-Bertani (LB) broth. For conjugation, R. sphaeroides cells were grown aerobically on LB medium in the dark.

Quantitative assay for 3-hydroxypropionate.

3-Hydroxypropionate concentrations were determined by an endpoint assay using recombinant propionyl-CoA synthase from C. aurantiacus produced in E. coli. The assay mixture (0.5 ml) contained 100 mM Tris-HCl (pH 8.0), 5 mM MgCl₂, 10 mM KCl, 2 mM dithiothreitol (DTT), 0.5 mM CoA, 3 mM ATP, 0.3 mM NADPH, and 5 to 10 μg partially purified recombinant propionyl-CoA synthase. The reaction was started with 20 μl of sample for which the 3-hydroxypropionate concentration was to be determined. The change in the absorbance at 365 nm was recorded, and the 3-hydroxypropionate concentration was determined by using an extinction coefficient of 3,400 M⁻¹ cm⁻¹ for NADPH and a stoichiometry for 3-hydroxypropinate used to NADPH oxidized of 1:1. The concentration of the 3-hydroxypropionate solution purchased from TCI was 2.7 M (3.6 M was determined by the titration of the solution by the manufacturer), meaning that it contained 10% 3,3′-oxydipropionic acid, as this was the only other compound present according to the manufacturer.

Determination of dry weight.

Cells from a 10- to 30-ml culture obtained at different time points during growth were collected by filtration using a 0.2-μm filter, completely dried, and weighed. On average, at an OD₅₇₈ of 1, the dry weight was 0.41 ± 0.01 g liter⁻¹.

Enzyme assays.

For measurements of the activities in R. sphaeroides cell extracts, 400 to 600 mg frozen cells was resuspended in 0.6 ml 25 mM Tris-HCl (pH 8.0) buffer containing 5 mM MgCl₂ and 0.1 mg ml⁻¹ DNase I. After the addition of 1 g glass beads (diameter, 0.1 to 0.25 mm), the cell solution was treated in a mixer mill (type MM2; Retsch, Haare, Germany) for 9 min at 30 Hz. Cell debris and glass beads were removed by centrifugation at 14,000 × g for 10 min at 4°C. The protein content of the cell extract was 2 to 15 mg ml⁻¹. Protein concentrations were determined according to the method of Bradford (10), using bovine serum albumin as a standard. The crotonyl-CoA-dependent oxidation of NADPH was monitored spectrophotometrically at 365 nm (ε_NADPH = 3,400 M⁻¹ cm⁻¹), using a cuvette with a path length of 0.1 cm. The reaction mixture (0.2 ml) contained 100 mM Tris-HCl buffer (pH 7.9), 4 mM NADPH, 2 mM crotonyl-CoA, and 0.04 to 0.8 mg cell extract. The reaction was started by the addition of 33 mM NaHCO₃ to the mixture. The ATP-, CoA-, and 3-hydroxypropionate- or acrylate-dependent oxidation of NADPH was monitored spectrophotometrically at 365 nm by using a cuvette with a path length of 1 cm. Dithiothreitol (2 mM) was added to cell extracts immediately after cell lysis. The reaction mixture (0.5 ml) contained 100 mM Tris-HCl buffer (pH 8.0), 5 mM MgCl₂, 10 mM KCl, 0.5 mM CoA, 3 mM ATP, 0.3 mM NADPH, 2 mM DTT, and 0.15 to 1.0 mg cell extract. The reaction was started with 20 mM 3-hydroxypropionate or acrylate.

HPLC analysis.

An optimized reaction mixture (1.24 ml) was used to monitor the product formation of the reductive conversion of acrylate. The assay mixture contained 100 mM Tris-HCl (pH 8.0), 2 mM DTT, 5 mM MgCl₂, 10 mM KCl, 3.6 mM ATP, 1.2 mM CoA, 1 mM NADPH, and 0.5 to 1 mg ml⁻¹ cell extract from 3-hydroxypropionate-grown cells. The enzymatic reaction was started by the addition of 26 mM acrylate to the mixture and stopped after 5 min by transferring 200 μl of the reaction mixture into 4 μl 25% HCl. After 5 min, NADPH (final concentration of 1.7 mM) was added to the mixture, and the reaction was stopped at different time points as described above. The samples were centrifuged at 10,000 × g for 5 min to remove the precipitated protein. CoA-thioesters were analyzed by reverse-phase high-performance liquid chromatography (HPLC) using an RP-C₁₈ column (CC 250/4 Nucleodur C₁₈ Gravity, 5 μm; Macherey-Nagel, Düren, Germany). A 30-min gradient from 2% (vol/vol) to 20% (vol/vol) acetonitrile in 50 mM potassium phosphate buffer (pH 6.7) at a flow rate of 0.6 ml min⁻¹ was used. All CoA-esters were detected at 260 nm. The retention times for standards were 10 min for free CoA, 11.5 min for 3-hydroxypropionyl–CoA, 13 min for acetyl-CoA, 17 min for propionyl-CoA, 18 and 20 min for acrylyl-CoA, and 20 min for crotonyl-CoA. 3-Hydroxypropionyl–CoA and acrylyl-CoA were synthesized from the free acids by using recombinant 3-hydroxypropionyl–CoA synthetase (2).

Mutant construction, ΔccrMC4 strain isolation, and complementation.

Chromosomal DNA from R. sphaeroides was isolated by using standard techniques. Two fragments containing 500- to 600-nucleotide (nt) flanking regions on either side of the kanamycin insertion site were amplified and cloned into conjugation vector pJQ200mp18 (34). For the upstream region, primers 5′-TCT CGA CGA ATT CGA TCT GCA TGG CCT GAT C-3′ and 5′-ATG GAG GTA CCG GCA TAG GTG CGG AAC AGC C-3′ were used, and for the downstream region, primers 5′-CAA GGA AGC TTC TCT AGA CGA GGG CAA GGA TAT TG-3′ and 5′-GGG CAA GCT TGA GCG TCT AGA CGA GCC CGG-3′ and nested PCR primers 5′-CAT CAG AAG CGT CTG CAG GGC AGC CAC TTC G-3′ and 5′-ATC TCG AGC ATA TGT CTA CAG AGC CCG GCC GCC AGC-3′ were used (introduced restriction sites are underlined). A kanamycin resistance cassette was amplified by using forward primer 5′-CCA ATC TGC AGT TAG AAA AAC TCA TCG AGC AT-3′, reverse primer 5′-CAT GAA TTC GAA AGC CAC GTT GTG TCT C-3′, and pUC4KSAC (Promega) as a template. The three fragments were cloned in tandem into pUC19 (for the downstream fragment, an internal PstI site was used), resulting in plasmid pMC06. An 891-nt fragment within the ccr gene was therefore replaced by the kanamycin resistance cassette. Plasmid pMC06 was digested with XbaI, and the fragment containing the interrupted ccr gene was ligated into pJQ200mp18, resulting in plasmid pMC21R. This plasmid was transferred into R. sphaeroides cells by conjugation, as described previously (17), with E. coli SM10 carrying pMC21R. The deletion/insertion mutation of the ccr gene was verified by PCR analyses (forward primer 5′-GCC AGC ACC GAG GAC TTG CCT TC-3′ and reverse primer 5′-GGT TGG TGC GAT AGA GCG CAT TCG AG-3′) and sequencing.

For the complementation construct, the ccr coding region was amplified from chromosomal DNA from R. sphaeroides by using primers 5′-GGT ACA GTA AGC TTA CCA TGG CCC TCG ACG TGC AG-3′ and 5′-ATG ATG GTA CCA CGG CGG CTG AGA CTT G-3′. A 367-nt upstream region of ccr was amplified by using primers 5′-TGG GTC ACA AGC TTG GGA CTC CAC TCG ACG AGG-3′ and 5′-TGC ACT AAG CTT ACC ATG GTT GCC TCC TGT TGG GCC GAA AG-3′, and both fragments were cloned in tandem into pBBRsm2MCS5, resulting in plasmid pMC19_2. Plasmid pBBRsm2MCS5 was constructed by replacing the gentamicin resistance cassette of pBBR1MCS5 (7, 25, 26) with a streptomycin-spectinomycin resistance cassette from vector pCDFDuet1 (Novagen), using primers 5′-GGT GTG CGT CCA TGG GCA AAT ATT ATA CGC AAG G-3′ and 5′-GGC TTC CCG GAG ATC TAG TTG TTC GGT AAA TTG-3′ and pBBR1MCS5 as a template and using primers 5′-AGT CTC ATC CAT GGA GCG TAG CGA CCG AGT GAG-3′ and 5′-ATG TGG CGA GAT CTC TTG AAC GGA TTG TTA GAC-3′ and pCDFDuet1 as a template. Plasmid pMC19_2 or pBBRsm2MCS5 was transferred into the ΔccrMC4 strain by conjugation with E. coli SM10.

Mutant construction, ΔacuI::kan LB2 strain isolation, and complementation.

Two fragments containing 700- to 800-nt flanking regions on either side of the kanamycin insertion site were amplified and cloned. For the upstream region, primers 5′-GAA GAA TTC TAG AGG CAG GCG CTT CAC AAC CGC AC-3′ and 5′-CTC CGG TAC CGA AAC GGA CTG CGT ATC-3′ were used, and for the downstream region, primers 5′-AGT AGG ATC CGG CAA AGC TCG AGG AGA TG-3′ and 5′-TAA CCT GCA GTC TAG ACC AGC AGG CAG GGC GAC AG-3′ were used. A kanamycin resistance cassette was amplified by using primers 5′-CCG AGG ATC CTA GAA AAA CTC ATC GAG CAT C-3′ and 5′-CAT CGG TAC CGA AAG CCA CGT TGT GTC TC-3′ and pUC4KSAC as a template. The three fragments were cloned in tandem into pJQ200mp18, resulting in pLB11. Plasmid pLB11 was transferred into R. sphaeroides by conjugation, and a double crossover was selected for and confirmed by the sequencing of the appropriate chromosomal region using primers 5′-CTT CCG CGG CGA ACT GAG TTT CAT TCC-3′ and 5′-CCC GAC CTC GCG CTG CTT CTT TAT TAT C-3′.

For the complementation construct, the acuI coding region was amplified from chromosomal DNA from R. sphaeroides by using primers 5′-GGA GAA GCA TAT GAG AGC CGT TCT GAT AG-3′ and 5′-CTA CGC GTT CGG ATC CGG AGT GCA TAC-3′ and cloned behind a 327-nt tetA promoter region that was amplified from plasmid pRL27 (27) by using primers 5′-AGA AGT TAT CAT ATG TGG CCT CCG-3′ and 5′-CCG CCG GAA TTC TCT AGA ATG-3′ and plasmid pBBRsm2MCS5, resulting in plasmid pMA5-1. Plasmid pMA5-1 or pBBRsm2MCS5 was transferred into the ΔacuI::kan LB2 strain by conjugation.

RESULTS

Assay to specifically and quantitatively detect 3-hydroxypropionate.

3-Hydroxypropionate cannot be quantified by commonly used methods such as HPLC, as no pure 3-hydroxypropionate solution that could be used as a standard is available. Therefore, propionyl-CoA synthase from C. aurantiacus was used to establish an enzymatic endpoint assay to quantify 3-hydroxypropionate. In order to obtain a large amount of this enzyme, the pcs gene (GenBank accession number AF445079) from C. aurantiacus, encoding propionyl-CoA synthase, was heterologously expressed in E. coli, and the recombinant enzyme was partially purified by heat precipitation (data not shown). The 3-hydroxypropionate-, ATP-, CoA-, and K⁺-dependent oxidation of NADPH catalyzed by propionyl-CoA synthase is essentially irreversible, because the acrylyl-CoA/propionyl-CoA couple has a much higher standard redox potential (+69 mV) than the NADP⁺/NADPH couple (37). In addition, due to its high substrate specificity and low K_m value for 3-hydroxypropionate (15 μM) (1), the enzyme is ideally suited for a specific and quantitative assay for 3-hydroxypropionate. The spectrophotometric endpoint assay measures the amount of NADPH oxidized after the addition of 3-hydroxypropionate. Based on a 1:1 ratio of the NADPH oxidized to the 3-hydroxypropionate reduced, the concentration of 3-hydroxypropionate was calculated; the lower detection limit was a final concentration of 2 μM 3-hydroxypropionate in the assay mixture (data not shown).

Growth of R. sphaeroides with 3-hydroxypropionate.

The photoheterotrophic (anaerobic in the light) or chemoheterotrophic (aerobic in the dark) growth of R. sphaeroides with 3-hydroxypropionate was monitored concomitantly with a decrease in the 3-hydroxypropionate concentration in the culture supernatant (Fig. 1). The results showed that R. sphaeroides was able to grow with 3-hydroxypropionate as the sole carbon source under phototrophic conditions, with a doubling time of 5 h. The biomass yield (g dry cell mass per mol substrate) was 60 g mol⁻¹ for anaerobic phototrophic growth with 3-hydroxypropionate. For photoheterotrophic growth with acetate, a carbon substrate in which the carbon has the same oxidation state as that of 3-hydroxypropionate but contains two instead of three carbons, the biomass yield was 40 g mol⁻¹. During aerobic chemotrophic growth in the dark with 3-hydroxypropionate as the carbon and energy source, the growth yield was 34 g mol⁻¹, indicating that about half of the substrate was completely oxidized to CO₂ to supply electrons for aerobic respiration.

Fig 1.

Growth of wild-type Rhodobacter sphaeroides 2.4.1 in defined medium containing 3-hydroxypropionate as a sole carbon source under photoheterotrophic (anaerobic/light) (A) and chemoheterotrophic (aerobic/dark) (B) growth conditions. Experiments were carried out with triplicate cultures. Growth was monitored by measuring the OD₅₇₈ (solid line). The extracellular 3-hydroxypropionate concentration (dotted line) was determined by an enzymatic endpoint assay using recombinant propionyl-CoA synthase from C. aurantiacus. Standard deviations are indicated by the error bars.

Crotonyl-CoA carboxylase/reductase is not involved in 3-hydroxypropionate metabolism.

In the following section, the focus is on photoheterotrophic growth, as the carbon substrate supplied was used exclusively for assimilation. An oxidative route for 3-hydroxypropionate utilization that leads to the formation of acetyl-CoA and CO₂ from 3-hydroxypropionate was proposed previously (18, 36). If R. sphaeroides also uses this proposed pathway, acetyl-CoA would be the entry point into the central carbon metabolism. However, for the conversion of acetyl-CoA into cell mass, a specialized anaplerotic pathway is required to convert acetyl-CoA to other precursor metabolites such as oxaloacetate, pyruvate, and α-ketoglutarate, from which cell carbon biosynthesis starts. R. sphaeroides uses the ethylmalonyl-CoA pathway for the assimilation of acetyl-CoA by converting three molecules of acetyl-CoA and two molecules of inorganic carbon to succinyl-CoA and malate (16). Malate and succinyl-CoA are further converted to the other precursor metabolites by known reactions of the central carbon metabolism. Because the committed step of the ethylmalonyl-CoA pathway is catalyzed by crotonyl-CoA carboxylase/reductase (Ccr), we examined the requirement for ccr during the growth of R. sphaeroides with 3-hydroxypropionate. For this reason, the ccr gene was inactivated, and the growth phenotype with 3-hydroxypropionate was examined; acetate and succinate were used as control carbon sources. As expected, the growth of the mutant strain with an insertional deletion in the ccr gene (ΔccrMC4) was unaffected with succinate (Fig. 2A), but the mutant was unable to use acetate as a sole carbon source (Fig. 2B). The introduction of the ccr gene together with 367 nucleotides of its upstream region on a plasmid [ΔccrMC4(pMC19_2)] restored growth with acetate, whereas the empty vector [ΔccrMC4(pBBRsm2MCS5)] did not (Fig. 2B). With 3-hydroxypropionate, the ΔccrMC4 mutant grew normally compared to the wild-type strain (Fig. 2C). Furthermore, Ccr activity was measured in cell extracts of R. sphaeroides. As expected, the activities were upregulated during photoheterotrophic growth on acetate (210 ± 30 nmol min⁻¹ mg⁻¹) compared to the activities during growth on succinate (<5 nmol min⁻¹ mg⁻¹). Notably, Ccr activity was undetectable in cell extracts of wild-type R. sphaeroides grown with 3-hydroxypropinate (<5 nmol min⁻¹ mg⁻¹). Together, these results clearly indicate that Ccr is not required during growth on 3-hydroxypropionate and suggest that 3-hydroxypropionate is not exclusively oxidized to acetyl-CoA and CO₂.

Fig 2.

Photoheterotrophic (anaerobic/light) growth of wild-type Rhodobacter sphaeroides 2.4.1 (○), the Δccr MC4 mutant (▵), the ΔccrMC4(pMC19_2) ccr-complemented mutant (×), and the ΔccrMC4(pBBRsm2MCS5) mutant (□) in defined medium containing succinate (A), acetate (B), or 3-hydroxypropionate (C) as a sole carbon source (10 mM final concentration). pBBRsm2MCS5 was the vector used for the construction of ccr complementation plasmid pMC19_2.

Reductive conversion of 3-hydroxypropionate to propionyl-CoA.

Extracts of R. sphaeroides grown photoheterotrophically with 3-hydroxypropionate catalyzed the MgATP-, CoA-, and 3-hydroxypropionate-dependent oxidation of NADPH with a specific activity of 24 ± 8 nmol min⁻¹ mg⁻¹. When the assay was started with acrylate instead of 3-hydroxypropionate, the specific activity was 120 ± 0 nmol min⁻¹ mg⁻¹. Both activities rapidly decreased when no DTT was added to the cell extract. However, if the cell extract was kept anaerobically or DTT was added, the activity still decreased over time but more slowly, with 50% activity remaining after 4 h. About 90% of the activity was recovered in the supernatant after centrifugation at 130,000 × g, indicating that the enzyme system is not membrane bound. The reduction of 3-hydroxypropionate or acrylate was not detectable in cell extracts of succinate-grown or propionate/CO₂-grown cells (both at <5 nmol min⁻¹ mg⁻¹), indicating that the activity is regulated in response to the growth substrate. The products of the reductive conversion of acrylate with cell extracts of 3-hydroxypropionate-grown R. sphaeroides were analyzed by HPLC. Acrylate was used as the assay substrate, because it can be obtained in a pure form. Propionyl-CoA was identified as the product of the reaction after coelution with a standard from an HPLC column (Fig. 3). When NADPH was initially omitted from the assay mixture, the main product was 3-hydroxypropionyl-CoA, which is expected, as the ratio of 3-hydroxypropionyl-CoA to acrylyl-CoA at equilibrium is greater than 50:1 (W. Buckel, personal communication).

Fig 3.

HPLC analysis of CoA-thioesters formed during the reductive conversion of acrylate by cell extracts of R. sphaeroides grown photoheterotrophically with 3-hydroxypropionate as the carbon source. Shown are data for the reaction mix before the addition of acrylate and NADPH (A), 5 min after the addition of acrylate but not NADPH (B), 2 min after the addition of NADPH to the reaction mixture (C), and after 5 min (D) and 10 min (E) of additional incubation in the presence of NADPH and acrylate. The identity of compound X is unknown but is also formed when acrylyl-CoA is synthesized using acyl-CoA synthetase from S. tokodaii and may represent an adduct between acrylyl-CoA and free CoA.

AcuI is involved in reductive conversion of 3-hydroxypropionate to propionyl-CoA.

A propionyl-CoA synthase similar to the one from C. aurantiacus that catalyzes the activation of 3-hydroxypropionate to its CoA-ester, the dehydration of 3-hydroxypropionyl-CoA to acrylyl-CoA, and the reduction of acrylyl-CoA to propionyl-CoA (1) is not encoded by the genome of R. sphaeroides. However, the predicted gene product of acuI (RSP_1434) is a member of the medium-chain dehydrogenase/reductase superfamily, which also contains acrylyl-CoA reductase from Sulfolobus tokodaii (42) and the acrylyl-CoA reductase domain of propionyl-CoA synthase from C. aurantiacus (1) despite all three proteins having limited sequence similarities (Fig. 4). For R. sphaeroides, acuI was recently shown to be cotranscribed with the DMSP lyase gene (dddL) and the gene whose product is involved in acrylate metabolism, but its function was left uncharacterized (40). Furthermore, a protein with 54% amino acid sequence identity to the R. sphaeroides RSP_1434 gene product is encoded by a gene (SPO1914) located next to dmdA, the gene for DMSP demethylase from Ruegeria pomeroyi DSS-3 (2) (Fig. 4). We reasoned that different mechanisms for the initial DMSP metabolism may lead to the same intermediate, acrylate or 3-hydroxypropionate, which is then further metabolized by a common pathway. In order to study the role of AcuI, 815 nucleotides of the coding region of acuI were replaced by a kanamycin resistance cassette on chromosome 1 of R. sphaeroides. The ΔacuI::kan LB2 mutant grew normally with succinate but was unable to grow with 3-hydroxypropionate as the carbon source (Fig. 5). The introduction of the acuI gene in trans on a plasmid under the control of the tetA promoter of E. coli restored growth (Fig. 5B). The MgATP-, CoA-, and 3-hydroxypropionate- or acrylate-dependent oxidation of NADPH was detectable in wild-type R. sphaeroides cells grown in the presence of both 3-hydroxypropinate and succinate (12 ± 0 nmol⁻¹ min⁻¹ mg⁻¹ for 3-hydroxypropionate reduction and 57 ± 6 nmol⁻¹ min⁻¹ mg⁻¹ for acrylate reduction); these activities were not detected in cell extracts of the ΔacuI::kan LB2 mutant grown under the same conditions (<5 nmol min⁻¹ mg⁻¹ for both 3-hydroxypropionate and acrylate reductions). These results clearly show that AcuI is involved in the reductive conversion of 3-hydroxypropionate to propionyl-CoA and that this route is essential for 3-hydroxypropionate assimilation by R. sphaeroides.

Fig 4.

Alignment of AcuI-like protein sequences and acuI-like genes. (A) Amino acid sequence alignment of AcuI-like proteins. The proposed acrylyl-CoA reductase from R. sphaeroides strain 2.4.1, AcuI (GenBank accession number ABA77575), was aligned with acrylyl-CoA reductase from Sulfolobus tokodaii strain 7 (accession number ACJ71675) and partially aligned with the acrylyl-CoA reductase domain of propionyl-CoA synthase from Chloroflexus aurantiacus strain J-10-f (accession number AAL47820) and 1914, a protein with unknown function from Ruegeria pomeroyi strain DSS-3 (accession number AAV95191). Red boxes indicate residues that are identical between two proteins, blue boxes indicate residues that are identical between three proteins, and green boxes indicate residues that are the same between all four proteins (and therefore apply only to the region where Pcs is coaligned) and contain an NAD(P)H binding domain, as indicated. (B) Genomic context of genes encoding AcuI-like proteins. acuR encodes a transcriptional regulator (40), acuI encodes the proposed acrylyl-CoA reductase (this work), and dddL encodes DMSP lyase (14). acr encodes an acrylyl-CoA reductase (42); the functions of the gene products from genes next to acr are unknown. pcs encodes propionyl-CoA synthase (1); the three domains refer to the acyl-CoA synthetase domain (pink), the acrylyl-CoA hydratase domain (blue), and the acrylyl-CoA reductase domain (brown; the region that was not aligned is shown in a fainter color). dmdA encodes DMSP demethylase (23), and SPO_1914 encodes a protein of unknown function.

Fig 5.

Photoheterotrophic (anaerobic/light) growth of wild-type (○), ΔacuI::kan mutant (▵), ΔacuI::kan(pMA5_1) acuI-complemented mutant (▴), and ΔacuI::kan(pBBRsm2MCS5) mutant (♢) strains in minimal medium containing succinate (10 mM) (A) or 3-hydroxypropionate (10 mM) (B) as a sole carbon source. pBBRsm2MCS5 is the vector used for the construction of acuI complementation plasmid pMA5-1.

DISCUSSION

R. sphaeroides can use 3-hydroxypropionate as a carbon and energy source. A theoretical growth yield was calculated based on the ash-free biomass elemental molar composition of another photosynthetically growing purple nonsulfur bacterium, Rhodopseudomonas palustris (13). Assuming the same composition for R. sphaeroides and a stoichiometric assimilation of 3-hydroxypropionate, the theoretical growth yield (dry cell mass per mol substrate) for phototrophic growth with 3-hydroxypropionate is 60 g mol⁻¹, as determined by using the following equation: 1.5 C₃H₆O₃ + 0.72 NH₃→4 CH_1.8N_0.18O_0.38 + 0.5 CO₂ + 1.98 H₂O.

The growth yield determined under photoheterotrophic conditions was 60 g cells (dry weight) per mol of 3-hydroxypropionate. For the C₂ compound acetate, the biomass yield was two-thirds of the biomass yield from the C₃ compound 3-hydroxypropionate, as expected. This also means that the 3,3′-oxydipropionic acid present in the purchased 3-hydroxypropionate solution was not assimilated by R. sphaeroides.

Figure 6 summarizes our current proposed scheme for 3-hydroxypropionate metabolism by R. sphaeroides. The assimilation of 3-hydroxypropionate by R. sphaeroides requires the reductive conversion of 3-hydroxypropionate to propionyl-CoA. A gene (acuI [RSP_1434]) was shown to be essential for growth with 3-hydroxypropionate for R. sphaeroides. Based on (i) its amino acid sequence, which places it in the large family of medium-chain dehydrogenases/reductases, and (ii) its involvement in the reductive conversion of acrylate to propionyl-CoA, AcuI is likely to be an acrylyl-CoA reductase. The heterologous expression of acuI in E. coli confirmed that AcuI catalyzes the acrylyl-CoA-dependent oxidation of NADPH; however, the instability of the enzyme so far hampers detailed characterizations (M. Asao and B. E. Alber, unpublished results). The instability of AcuI is reflected in the loss of activity over time for the overall conversion of 3-hydroxypropionate or acrylate to propionyl-CoA in cell extracts of 3-hydroxypropionate-grown R. sphaeroides cells.

Fig 6.

Proposed scheme for 3-hydroxypropionate assimilation by R. sphaeroides. The reductive conversion of 3-hydroxypropionate to propionyl-CoA involving AcuI is essential for the growth of R. sphaeroides with 3-hydroxypropionate as the sole carbon source. The ethylmalonyl-CoA pathway involving Ccr is not required for 3-hydroxypropionate assimilation. The dotted line represents an oxidative path that cannot be ruled out based on the results presented here and might be used for the biosynthesis of cell constituents derived directly from acetyl-CoA. However, in principle, acetyl-CoA may also be formed from succinyl-CoA oxidation via pyruvate.

The presence of a route that assimilates 3-hydroxypropionate via propionyl-CoA is consistent with the fact that crotonyl-CoA carboxylase/reductase was not required during growth with 3-hydroxypropionate. Crotonyl-CoA carboxylase/reductase is the key enzyme of the ethylmalonyl-CoA pathway, a reaction sequence that is essential for growth with carbon compounds, such as acetate, that enter central carbon metabolism exclusively at the level of acetyl-CoA (Fig. 6). However, the entry point of carbon derived through the reduction of 3-hydroxypropionate to propionyl-CoA is succinyl-CoA, as propionyl-CoA carboxylase and methylmalonyl-CoA mutase are also essential for growth with 3-hydroxypropionate for R. sphaeroides (data not shown). An oxidative path from 3-hydroxypropionate to acetyl-CoA and CO₂ may still be important solely for the formation of building blocks derived directly from acetyl-CoA, such as fatty acids. Other precursor metabolites can be provided by succinyl-CoA, which results from the reduction of 3-hydroxypropionate and the subsequent carboxylation of propionyl-CoA. In addition, the complete oxidation of 3-hydroxypropionate to CO₂ during aerobic growth is likely to proceed via acetyl-CoA in order to supply electrons for respiration.

There are now two paths suggested for the conversion of 3-hydroxypropionate to precursor metabolites by different organisms. The first is 3-hydroxypropionate oxidation to acetyl-CoA and CO₂ (18, 36). Experimental evidence for the proposal that 3-hydroxypropionate enters the central carbon metabolism exclusively at the level of acetyl-CoA has been provided for the aerobic obligatory heterotrophic alga Prototheca zopfi. The glyoxylate bypass essential for acetyl-CoA assimilation is required to convert 3-hydroxypropionate-derived acetyl-CoA to other precursor metabolites (11). The second suggested path, the reductive conversion of 3-hydroxypropionate to succinyl-CoA via propionyl-CoA, as described here for R. sphaeroides, may also be operating in other organisms. A protein with 60% sequence identity to the herein-described AcuI from R. sphaeroides is encoded by a gene that is part of a cluster for DMSP metabolism in Alcaligenes faecalis (15). Furthermore, there is an acuI-like gene (SPO_1914) located next to dmdA, encoding DMSP demethylase in R. pomeroyi DSS-3. Hence, we reasoned that AcuI is involved in the metabolism of DMSP. While the catabolism of DMSP involving DMSP demethylase by R. pomeroyi was recently shown to proceed by a route independent of 3-hydroxypropionate as an intermediate (35), the AcuI-like protein may still be required for the assimilation of carbon derived from DMSP by a route through acrylyl-CoA and propionyl-CoA. The absence of an acuI-like gene in other DMSP-utilizing organisms, on the other hand, does not necessarily rule out the assimilation of carbon by a reductive route. In principle, enzymes with very limited sequence identity to AcuI or members of different enzyme classes might catalyze the reduction of acrylyl-CoA to propionyl-CoA (2, 22, 42).

The primarily marine substrate DMSP is by far not the only source of 3-hydroxypropionate in nature, but instead, 3-hydroxypropionate is the product or intermediate of the breakdown of many substrates in a variety of environments (4, 18, 29, 30, 31, 33, 41). It will therefore not be surprising to also find many nonmarine organisms that will be able to derive carbon from 3-hydroxypropionate. For example, very recently, it was suggested that C. aurantiacus coassimilates 3-hydroxypropionate during mixotrophic growth using propionyl-CoA synthase (46).