Role of Genetic Redundancy in Polyhydroxyalkanoate (PHA) Polymerases in PHA Biosynthesis in Rhodospirillum rubrum

Abstract

This study investigated the apparent genetic redundancy in the biosynthesis of polyhydroxyalkanoates (PHAs) in the Rhodospirillum rubrum genome revealed by the occurrence of three homologous PHA polymerase genes (phaC1, phaC2, and phaC3). In vitro biochemical assays established that each gene product encodes PHA polymerase. A series of single, double, and triple phaC deletion mutants were characterized with respect to PHA production and growth capabilities on acetate or hexanoate as the sole carbon source. These analyses establish that phaC2 contributes the major capacity to produce PHA, even though the PhaC2 protein is not the most efficient PHA polymerase biocatalyst. In contrast, phaC3 is an insignificant contributor to PHA productivity, and phaC1, the PHA polymerase situated in the PHA biosynthetic operon, plays a minor role in this capability, even though both of these genes encode PHA polymerases that are more efficient enzymes. These observations are consistent with the finding that PhaC1 and PhaC3 occur at undetectable levels, at least 10-fold lower than that of PhaC2. The monomers in the PHA polymer produced by these strains establish that PhaC2 is responsible for the incorporation of the C₅ and C₆ monomers. The in vitro characterizations indicate that heteromeric PHA polymerases composed of mixtures of different PhaC paralogs are more efficient catalysts, suggesting that these proteins form complexes. Finally, the physiological role of PHA accumulation in enhancing the fitness of R. rubrum was indicated by the relationship between PHA content and growth capabilities of the genetically manipulated strains that express different levels of the PHA polymer.

INTRODUCTION

Polyhydroxyalkanoates (PHAs) are polyester polymers that many microbes use as a means of storing carbon and energy. These molecules are deposited within cells mainly in the form of protein- and lipid-bound granular inclusions (15, 22, 36, 38). PHA biosynthesis and accumulation is enhanced when microbes are under conditions of excess carbon but are limited for other nutrient(s), such as phosphorus, nitrogen, or sulfur (15, 22, 36, 38). Additional functionalities that have been ascribed to PHAs include the transport of calcium phosphate and DNA across membranes. This is thought to be achieved by the formation of calcium polyphosphate membrane-spanning channels that can enhance transformation efficiency and regulate intracellular calcium concentrations. Such changes in intracellular calcium concentrations also influence calcium signaling (29–31) and resistance to stresses (e.g., radiation, desiccation, and osmotic pressure) (40).

PHAs have technological applications as biodegradable plastics (46). Because these applications are dependent on the chemical and physical properties of the PHA polymer, which are themselves dependent on the monomeric constituents and their relative order in the polymer, intensive studies have been conducted to understand the mechanisms of PHA biosynthesis (15, 22, 38). These studies have established that the type of polymer that is produced depends on the diversity of substrates that are available to the PHA polymerase (product of the phaC gene) that assembles the final polymer (2, 23). The substrate for all known PhaC enzymes is (R)-configured 3-hydroxyacyl-coenzyme A (CoA) of different acyl chain lengths (from 3 to 14 carbons) (26).

Most bacteria produce PHAs that are composed of monomers that are either of short chain length (C₃ to C₅) or medium chain length (C₆ to C₁₄) (1, 16, 37, 38). A few organisms produce PHAs with a broader range of monomer chain lengths (C₄ to C₁₂) (4, 10, 21, 23). The purple and phototrophic bacterium Rhodospirillum rubrum is known to produce PHA composed of both short- and medium-chain-length monomers, and it can produce up to 50% (dry weight) PHA (4, 19, 45). This metabolically versatile bacterium can grow under aerobic or anaerobic conditions or in the presence or absence of light, and in the latter condition it can use a variety of different carbon substrates (39, 43). Previous studies and genome sequencing have revealed that R. rubrum can express three PHA polymerases (5, 12). Because of its metabolic flexibility, R. rubrum offers the potential for converting many different carbon sources to PHA, and therefore there are considerable industrial interests in exploring this flexibility (6, 35). In this study, we characterized the significance and roles of each phaC paralog in PHA production and growth of R. rubrum.

MATERIALS AND METHODS

Chemicals and enzymes.

All chemicals were obtained from Sigma-Aldrich Corporation (St. Louis, MO) and Fisher Scientific Inc. (Pittsburgh, PA). All DNA-manipulating enzymes were obtained from Invitrogen Corporation (Carlsbad, CA).

Bacterial strains and plasmids.

Rhodospirillum rubrum ATCC 11170 and Escherichia coli DH5α were used in this study. The primers used in this study are listed in Table S1 in the supplemental material.

Bacterial growth conditions.

E. coli was grown at 37°C in LB medium. R. rubrum was grown in supplemented malate-ammonium medium (SMN medium) (13). When needed, gentamicin (25 μg/ml) or kanamycin (25 μg/ml) was added to medium to maintain selection for plasmids. Isopropyl-β-d-thiogalactopyranoside (IPTG) and 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) were used at concentrations of 20 and 40 mg/ml, respectively.

PHA production was assessed in R. rubrum cultured in RRNCO medium (but omitting ammonium chloride, hydrogen sulfide, carbon monoxide, and carbon dioxide) (14). In these experiments, a 0.2-ml aliquot of a normalized SMN culture (5 optical density [OD] units) was collected by centrifugation at 13,000 × g for 2 min; the cells were washed once with RRNCO medium and resuspended in 20 ml RRNCO medium containing either 10 mM acetate or 5 mM hexanoate as the carbon source. Cultures were shaken at 150 rpm at 25°C under 5,000-lx light intensity. For anaerobic growth, R. rubrum cultures were grown with an argon headspace in 18- by 150-mm anaerobic tubes (Bellco Biotechnology, Vineland, NJ). For biochemical analysis, aliquots of 4 ml were withdrawn from the cultures at 0, 72, 96, 120, and 192 h postinoculation. Cell density was determined by monitoring the A₆₈₀ using a Spectronic 20D+ spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA). The doubling time (T_d) was determined from the initial four time points of cultures.

GC-MS analysis of PHA.

Cells were collected from liquid cultures by centrifugation at 6,000 × g for 10 min, and following washing with 10 mM Tris-HCl buffer (pH 7.5), the cell pellets were lyophilized and stored at −70°C until analysis. PHA content and composition were determined using 4- to 20-mg dry cell samples as described by Brandl et al. (3). Adipic acid was added as an internal standard. Derivatized samples (methylated esters) were concentrated under a stream of nitrogen gas when necessary. A 1-μl portion of the methyl esters was assayed by splitless injection into a gas chromatograph-mass spectrometer (GC-MS; Agilent 6890 GC) interfaced to an electron impact ionization mass spectrometer detector (Model 5973; Agilent Technologies, Santa Clara, CA). The GC was equipped with a DB-WAX column (30 m by 0.25 mm; inner diameter, 0.5 μm), separation was with helium as the carrier gas (1.2 ml/min), and the temperature gradient was programmed from 80 to 200°C at 10°C/min. In the MS, the operating parameters were set to 70 electron voltage (eV) for the ionization voltage, and the interface temperature was 280°C. The temperatures of the injector and detector were 250 and 240°C, respectively. The GC-MS data files were deconvoluted with the NIST AMDIS software (http://chemdata.nist.gov/mass-spc/amdis/). The retention times of the different 3-hydroxycarboxylic acid methyl esters (C₄, C₅, and C₆) were 6.75, 7.66, and 8.70 min, respectively. The PHA content was calculated as percent cell dry weight, using 3-hydroxybutyric acid (10 to 1,000 μg/ml) to construct a standard curve, normalized for losses based on the recovery of the methyl ester of adipic acid.

DNA isolation and manipulation.

Genomic DNA was isolated from R. rubrum as described by Kerby et al. (13). Plasmids were isolated from E. coli cells grown in LB medium by using a QIAprep Spin miniprep kit (Qiagen Inc., Valencia, CA). Agarose gel electrophoresis and transformation of E. coli were carried out as described by Sambrook and Russell (33). PCR products were cloned into pCR2.1-TOPO using a TOPO TA cloning kit (Invitrogen Corporation, Carlsbad, CA).

Construction of gene deletion plasmids.

The 1,018-bp DNA fragment upstream of the phaC1 open reading frame (ORF) was amplified by PCR with the primers mC1USf-SacI and mC1USr-XbaI, and this amplification product contained SacI and XbaI sites at the 5′ and 3′ ends. The PCR product was purified and ligated into the TA cloning vector pPCR2.1-TOPO. The PCR fragment was removed from the resulting vector by digestion with SacI and XbaI, gel purified, and cloned into pJQ200SK at the SacI and XbaI sites, forming the vector pJQ200SKmC1U. The 1,023-bp DNA fragment downstream of the phaC1 ORF was cloned into the vector pJQ200SKmC1U via the same strategy, forming the ΔphaC1 gene deletion vector pJQ200SKmC1UD. The same strategy was used to construct vectors pJQ200SKmC2UD and pJQ200SKmC3UD, which were used to create gene deletion alleles for phaC2 and phaC3 genes, respectively. The primers used for amplifying R. rubrum genomic DNA fragments for the latter vectors are listed in Table S1 in the supplemental material.

Enzymological and protein methods.

Recombinant His-tagged fusion proteins were purified via nickel-affinity chromatography (BD Biosciences, San Jose, CA) from extracts of E. coli BL21-AI or Arctic Express strains harboring the appropriate pDEST17-derivatized expression vector. To generate antibodies, the purified recombinant proteins were used to immunize mice maintained at the Iowa State University Hybridoma Facility (http://www.biotech.iastate.edu/facilities/hybridoma/). PHA biosynthetic proteins were immunologically detected after protein extracts were separated by SDS-PAGE, and immunoblot analysis were performed as described previously (17); there was no detectable cross-reactivity among the three PhaC antibodies that were prepared. The PHA polymerase activity was assayed spectrophotometrically at 37°C in a 500-μl reaction mixture that contained 50 mM Tris-HCl (pH 7.5), 0.5 mM hydroxybutyryl-CoA substrate, 50 μl of cell lysate containing 5 μg of recombinant PhaC, and 1 mM DTNB. Polymerase activity was monitored by the rate of increase in absorbance at 412 nm due to the reaction of the CoA thiol with DTNB (24). The concentration of CoA was determined using an extinction coefficient (412 nm) of 13,700 M⁻¹ cm⁻¹ (8, 32). One unit of polymerase activity is defined as 1 μmol CoA released/min/μg polymerase.

Conjugation.

The three phaC gene deletion plasmids were mobilized from E. coli strain 17-1 into R. rubrum by conjugation (18). R. rubrum conjugants were selected on gentamicin-containing MN medium (MN medium is SMN medium in which yeast extract and casein enzyme hydrolysate are omitted). Single-colony R. rubrum conjugants were isolated by repeated transfer to gentamicin-containing MN medium.

Construction of R. rubrum deletion mutant strains.

Seven deletion strains of R. rubrum were generated by homologous recombination via a two-step procedure using the suicide vector pJQ200SK (25). These consisted of the ΔphaC1, ΔphaC2, and ΔphaC3 single-site deletion mutant strains, the ΔphaC1 ΔphaC2, ΔphaC1 ΔphaC3, and ΔphaC2 ΔphaC3 double mutant strains, and the ΔphaC1 ΔphaC2 ΔphaC3 triple mutant strain. The molecular confirmation of each deletion allele was achieved by PCR amplifying each allele with flanking primers (see Table S1 in the supplemental material) and determining the nucleotide sequence of each amplification product. Primers imC1f and imC1r were used to confirm the ΔphaC1 allele, primers imC2f and imC2r were used to confirm the ΔphaC2 allele, and primers imC3f and imC3r were used to confirm the ΔphaC3 allele.

RESULTS AND DISCUSSION

Three phaC paralogs in the R. rubrum genome.

Previous studies identified and characterized two PHA polymerases (PhaCs) (5, 12) in R. rubrum. Using these two sequences as the query, homology searches of the R. rubrum genome of strain ATCC 11170 revealed three genes encoding PHA polymerases, Rru_A0275, Rru_A2413, and Rru_A1816, which we designated phaC1, phaC2, and phaC3, respectively. One of them (phaC1) is located adjoining the phaA and phaB homologous genes in the PHA biosynthetic operon. The three phaC genes encode proteins of 414, 598, and 600 amino acids, respectively, and they all belong to the class I PHA polymerase family (26 and http://genome.ornl.gov/microbial/rrub/.). Figure 1 shows the sequence comparison of the amino acid sequences of PhaC1, PhaC2, and PhaC3 proteins. This analysis indicates that PhaC2 and PhaC3 share the highest sequence conservation (50.2% identity), and PhaC1 is equally distinct from PhaC2 and PhaC3 (sharing 14.3 and 18.4% sequence identity, respectively).

Fig 1.

Amino acid sequence alignment of PhaC3, PhaC2, and PhaC1 from R. rubrum constructed with ClustalW (41). Black-shaded residues are identical, and gray-shaded residues are similar.

Recombinantly expressing each phaC gene in E. coli and assaying the resulting cell lysate for PHA polymerase activity provided experimental evidence that these three homologs encode PHA polymerase enzymes (Table 1). All three individual proteins are capable of utilizing 3-hydroxybutyl-CoA to produce PHA, demonstrating that each phaC gene encodes a functional PHA polymerase. Of the three gene products, PhaC1 is the most efficient catalyst and PhaC2 has the lowest specific activity, about 10% of that of PhaC1.

Table 1.

PHA polymerase activity in vitro^a

Enzyme(s)	Expected activity	Activity (U)b
PhaC1	NA	17.9 ± 2.4
PhaC2	NA	2.1 ± 0.4
PhaC3	NA	6.6 ± 0.6
PhaC1 + PhaC2	10.0	48.6 ± 1.8*
PhaC1 + PhaC3	12.3	22.7 ± 1.8*
PhaC2 + PhaC3	4.4	3.0 ± 1.3

^aNA, not applicable. An asterisk indicates that the activity of the heteromeric enzymes is significantly higher than the expected activity (P < 0.05).

^bData represent averages from triplicate determinations ± standard errors.

The occurrence of multiple PHA polymerases in a single genome has been noted before in a number of Pseudomonas and Ralstonia species (11, 23, 42). A BLASTP analysis (conducted in October 2011) of sequenced microbial genomes identified 315 strains that contain PhaC homologs, and 156 of these strains contain two or more paralogs. The distribution of these multiple phaC loci within individual genomes has some commonalities. For example, as with R. rubrum, often one phaC locus is situated in a PHA biosynthetic operon that also contains the phaB and phaA genes. The additional phaC loci are sometimes situated in operons that also contain PHA-regulatory functions, such as phaD and phaZ (23). However, there are also examples of phaC loci that reside in independent, non-operon loci, as is the situation for the phaC2 and phaC3 loci of R. rubrum. For example, the Pseudogulbenkiania sp. NH8B genome contains three phaC genes, one (NH8B_1486) of which is situated in an operon that also contains the phaA gene (NH8B_1485), but the other two paralogs (NH8B_2463 and NH8B_1804) are at independent non-operon loci.

To investigate the metabolic functions of the three phaC genes in R. rubrum, single-locus deletion mutants (ΔphaC1, ΔphaC2, and ΔphaC3 strains), double-locus phaC deletion mutants (ΔphaC1 ΔphaC2, ΔphaC1 ΔphaC3, and ΔphaC2 ΔphaC3 strains), and the triple phaC deletion mutant (ΔphaC1 ΔphaC2 ΔphaC3 strain) were generated as described in Materials and Methods. These mutant strains were characterized in terms of growth, PHA yields, and monomer composition of the polymer. These characterizations were conducted by growing these strains with either acetate or hexanoate as the sole carbon source.

Effect of single-locus phaC deletions on PHA production and growth.

With the exception of the ΔphaC2 strain, all single mutant strains grew with kinetics similar to those of the wild-type (WT) strain with acetate (Fig. 2B) or hexanoate (Fig. 2D) as the carbon source; the ΔphaC2 strain grew significantly slower (doubling time of 14.2 h, compared to 10.3 h for the WT strain) and to a lower cell density when grown on acetate. The most significant difference among these strains in terms of PHA production was the observation that the ΔphaC2 strain lost the ability to accumulate PHA, to about 10% of WT levels with either carbon source (<3% cell dry weight [CDW]). In contrast, the ΔphaC1 and ΔphaC3 mutant strains showed a PHA accumulation pattern that was similar to that of the WT with either acetate (Fig. 2A) or hexanoate (Fig. 2C) as the carbon source, reaching values of about 25 and 15% of dry weight, respectively. An interesting observation was that with either carbon source the ΔphaC1 strain accumulated larger amounts of PHA than the WT, an increase of between 10 and 30% depending on the carbon source and phase of growth. These in vivo data indicate that PhaC2 is the major polymerase enzyme contributing to the synthesis of PHA. Furthermore, PhaC1 appears to have an inhibitory effect on PHA biosynthesis.

Fig 2.

PHA content (A and C) and growth curves (B and D, with doubling times indicated) of R. rubrum strains (WT and ΔphaC1, ΔphaC2, and ΔphaC3 mutants) grown in acetate (A and B) or hexanoate (C and D) as the carbon source. Data represent averages from triplicate biological samples, and error bars indicate the standard errors.

Effect of single-locus phaC deletions on PHA monomer composition.

In parallel to the studies described above, we determined the effect of the individual phaC knockouts on PHA monomer composition (Table 2). These analyses indicate that in any individual mutant strain, PHA monomer composition was unaffected by the phase of growth of the culture; the data presented in Table 2 where gathered from cells collected at the 72-h time point. In all strains grown on acetate, the vast majority (>99%) of the PHA was composed of the 3-hydroxybutyrate monomer (3HB), but in hexanoate-grown cells, about 6% of the monomer was either 3-hydroxyvalerate (3HV) or 3-hydroxyhexanoate (3HHx). The most significant monomer compositional change was observed with the strain lacking the PhaC2 polymerase (ΔphaC2), which resulted in the loss of the longer chain monomers (i.e., 3HV and 3HHx). This was particularly discernible when these strains were grown on hexanoate as the carbon source. A somewhat more subtle effect was the increased proportion of the longer chain monomers that occurred in the ΔphaC1 and ΔphaC3 strains; although relatively small amounts were incorporated in the polymer, statistical Student t tests indicate that the 3HV and 3HHx monomer contents in these two strains are significantly increased (P = 0.008). These data indicate that PhaC2 is the unique polymerase responsible for integrating 3HV and 3HHx monomers into the PHA polymer, and this is a property that is absent from PhaC1 and PhaC3 polymerases.

Table 2.

PHA composition of R. rubrum strains^a

Strain and carbon source	PHA composition (mol %) of:
	3HB (C4)	3HV (C5)	3HHx (C6)
Acetate
WT	99.6 ± 0.2	0.4 ± 0.2	<0.1*
ΔphaC1	99.5 ± 0.2	0.5 ± 0.2	<0.1*
ΔphaC2	100	<0.1*	<0.1*
ΔphaC3	99.4 ± 0.3	0.6 ± 0.3	<0.1*
ΔphaC1 ΔphaC2	NA	NA	NA
ΔphaC1 ΔphaC3	99.3 ± 0.3	0.7 ± 0.3	<0.1*
ΔphaC2 ΔphaC3	100	ND	ND
ΔphaC1 ΔphaC2 ΔphaC3	NA	NA	NA
Hexanoate
WT	95.1 ± 0.4	0.8 ± 0.1	4.1 ± 0.3
ΔphaC1	93.3 ± 0.5	0.9 ± 0.1	5.8 ± 0.4
ΔphaC2	100	<0.1*	<0.1*
ΔphaC3	93.2 ± 0.4	1.0 ± 0.1	5.8 ± 0.3
ΔphaC1 ΔphaC2	NA	NA	NA
ΔphaC1 ΔphaC3	93.6 ± 0.4	0.8 ± 0.1	5.6 ± 0.3
ΔphaC2 ΔphaC3	100	ND	ND
ΔphaC1 ΔphaC2 ΔphaC3	NA	NA	NA

^a3HB, 3-hydroxybutyrate; 3HV, 3-hydroxyvalerate; 3HHx, 3-hydroxyhexanoate; NA, not applicable; ND, not detected; *, less than the detection limit (0.5 μg/ml for each monomer). Samples were collected at the 72-h time point. Data represent averages ± standard errors from triplicate biological samples.

These compositional changes in the PHA polymer may be due to the substrate specificities of the three polymerases or to the availability of the appropriate monomer precursors. The 3HV and 3HHx monomers may be derived from intermediates of fatty acid β-oxidation or, in the case of 3HV, may be generated by the condensation of a propionate and acetate moiety (34). Regardless of these two mechanisms, our data indicate that PhaC2 has a unique capability to access these intermediates and incorporate them into the PHA polymer, whereas the other two polymerases exclude the incorporation of the larger precursors. An analogous situation has been analyzed in a Pseudomonas species, in which the phaC gene that resides in the operon that also contains phaA and phaB is responsible for poly(3-hydroxybutyrate) (PHB) biosynthesis (i.e., incorporating 3HB monomers). In this case, two independent phaC paralogs are responsible for incorporating the C₆ to C₁₂ monomers, and these two paralogs are class II polymerases (23).

Effect of double and triple phaC deletions on PHA production.

To further dissect the significance of the gene redundancy in PHA polymerase genes, we generated all possible double mutant combinations and the triple mutant, which lacked all three phaC genes. As would be expected, the ΔphaC1 ΔphaC2 ΔphaC3 triple mutant strain failed to accumulate any PHA when either carbon source was tested. This finding indicates that these are the only polymerases that assemble the PHA polymer. The inability to accumulate PHA influences the initial rate of growth of the culture with either acetate or hexanoate as the carbon source (Fig. 3A and B), and this is particularly more evident with acetate as the carbon source.

Fig 3.

Growth curves (A and B, with doubling times indicated) of R. rubrum strains (WT and ΔphaC1 ΔphaC2 ΔphaC3 mutant) grown in acetate (A) or hexanoate (B) as the carbon source. Data represent averages from triplicate biological samples, and error bars indicate the standard errors.

The double mutant strains provide a means of assessing the ability of individual PHA polymerases to support PHA production. Thus, these experiments are a corollary to the single-gene deletion mutants and provide an independent means of evaluating the function of each phaC paralog in PHA production. The data gathered from these double mutants grown on either acetate or hexanoate indicate that phaC2 is the most effective gene in determining PHA levels, followed by phaC1 and phaC3; indeed, the latter strain, expressing only phaC3, does not accumulate any detectable PHA (Fig. 4A and C). These conclusions are consistent with those reached from the characterization of the single-gene deletion mutants.

Fig 4.

PHA content (A and C) and growth curves (B and D, with doubling times indicated) of R. rubrum strains (WT and ΔphaC1 ΔphaC2, ΔphaC1 ΔphaC3, and ΔphaC2 ΔphaC3 mutants) grown in acetate (A and B) or hexanoate (C and D) as the carbon source. Data represent averages from triplicate biological samples, and error bars indicate the standard errors.

The effects of these genetic manipulations influence the growth of the cultures, in that strains that fail to accumulate significant levels of PHA (<1% dry weight) grow slower than the wild type (Fig. 4B and D). In contrast, the double mutant strain that expresses only the PhaC2 polymerase and accumulates elevated levels of PHA grows at the same rate as the wild-type strain.

Of the double and triple mutants, only the ΔphaC1 ΔphaC3 double mutant accumulated sufficient PHA for a reliable determination of the monomer composition of the polymer. In this strain, which expressed only the PhaC2 polymerase, the proportion of the 3HV and 3HHx monomers increased compared to the WT (P = 0.01) when it was grown on hexanoate (Table 2), consistent with previous findings that phaC2 is primarily responsible for the incorporation of 3HV and 3HHx monomers into the polymer.

Characterization of the expression of the PHA biosynthetic proteins.

To ensure that the genetic deletion of phaC genes does not elicit an alteration in other PHA biosynthetic genes, particularly the phaC homologs, which would confound the interpretation of these experiments, we compared the expression of the PHA biosynthetic genes between the WT and mutant strains. These analyses were conducted by using antibodies raised against each PHA biosynthetic protein, and we evaluated whether the accumulation of each protein was altered by the genetic manipulations. To ensure that the Western blot analysis quantitatively detected each PHA biosynthetic protein, each antibody was used to detect different amounts of the respective recombinant protein in the range of 1 to 100 ng of protein (Fig. 5A). These analyses indicate that in our immunological assays the detection limits for PhaC1, PhaC2, PhaC3, PhaA, and PhaB are approximately 10, 1, 10, 1, and 1 ng, respectively (Fig. 5A).

Fig 5.

Immunological characterization of PHA biosynthetic proteins. (A) Immunological detection limit of each PHA biosynthetic protein. The indicated amounts of each purified recombinant protein were subjected to SDS-PAGE and Western blot analysis. (B) Western blot analysis of PhaA, PhaB, and PhaC2 proteins in extracts of R. rubrum strains, grown in acetate as a carbon source (after 72 h of cultivation) or hexanoate as a carbon source (after 192 h of cultivation). Not shown in B are immunoassays for PhaC1 and PhaC3, which were undetectable or below the detection limit of the assay. For PhaA, PhaB, and PhaC2, each lane contained 5 μg of total protein, and for PhaC1 and PhaC3, each lane contained 50 μg of total protein.

Using this immunological assay, we quantified the expression of each PHA biosynthetic protein in the WT strain. These analyses indicate that of the three PhaC paralogs, the most abundantly expressed is PhaC2 (2 to 4 ng/μg total protein), and the other two paralogs were below the detection limit of the assay (<0.2 ng/μg of total protein). This biochemical finding that the PhaC2 paralog is the most highly expressed polymerase protein, being at least 10-fold more abundant than the other two paralogs, is consistent with the genetic-based conclusions that PhaC2 plays a major role in producing the PHA polymer and the other paralogs play minor roles. Of the three proteins that are encoded by the PHA operon, the most abundantly expressed is PhaA (5 to 10 ng/μg total protein), followed by PhaB (2 to 4 ng/μg total protein), and PhaC1 was below the detection limit (<0.2 ng/μg total protein).

In response to the genetic manipulation of each phaC paralog, we analyzed the expression level of those proteins that were detectable by this immunological assay (i.e., PhaA, PhaB, and PhaC2) (Fig. 5B). These analyses indicate that of the three proteins that were detectable, only the expression of PhaA and PhaC2 was altered in response to the genetic manipulations; PhaB accumulation was unaltered. PhaA levels increased in all three strains that carried a ΔphaC2 deletion, but this occurred only in strains grown on hexanoate as a carbon source. This is consistent with the involvement of PhaC2 as the only polymerase that has the capacity to utilize 3HHx monomer generated from the β-oxidation of hexanoate; hence, PhaA expression is enhanced in the ΔphaC2 deletions to accommodate carbon flux through the PHA operon pathway (i.e., needing PhaA functionality).

In vitro enzymological assay of PHA polymerase activities.

The observation that eliminating phaC1 or phaC1-phaC3 combinations results in increased accumulation of PHA suggests that these two genes interact to affect PHA biosynthetic capability. Evidence for potential interactions between the PhaC proteins was obtained by mixing E. coli cell lysates that contain each of the PhaC paralogs. For example, mixing lysates that contained 0.5 μg of PhaC1 and PhaC2, we expected a yield of about 10 U of PHA polymerase activity; however, we observed a specific activity of 49 U/μg (Table 1). This synergistically enhanced PHA polymerase activity is an indication that these two paralogs are interacting. This level of synergistic activation of PHA polymerase activity was also observed when these two paralogs were mixed at a ratio of 1:10 (PhaC1-PhaC2) (data not shown), which is more reflective of the in vivo accumulation of these two polymerases (Fig. 5). Similar but less synergistic enhancement in PHA polymerase-specific activity was found when PhaC1 and PhaC3 enzymes were mixed, but no such enhancement was obtained when PhaC2 and PhaC3 were mixed (Table 1). These results indicate that PhaC1-containing heteromeric polymerase mixtures are more active than the homomeric enzymes, indicating that these paralogs probably form complexes.

The finding that interactions among PhaC proteins enhance PHA polymerase activity makes it difficult to directly extrapolate from the in vitro results to explain the in vivo genetic deletion results. Most of our studies have considered the effect of modifying phaC paralogs on PHA production, but PHA accumulation is a more integrated system. For example, previous studies have shown that PHA biosynthesis and degradation occurs simultaneously, and PHA granule-associated proteins need to coordinate for optimal PHA accumulation (7, 27, 28, 44). Thus, in our ΔphaC mutants we need to consider not only the effect on PHA polymerization but also that the global PHA biosynthetic, catabolism, and storage machinery may modulate in response to the genetic modifications to accommodate the discrete ΔphaC mutant changes. However, our studies provide insights into the fact that PHA polymerase enzyme activity can be modulated at the biochemical level by the ability to produce heteromeric polymerase enzymes, and this attribute may have biotechnological applications to enhance PHA production.

The interrelationships among PHA monomer composition, PHA levels, and growth.

In the characterizations of the mutant R. rubrum strains that express qualitative and quantitative changes in PHA, we noted correlations between these traits and the growth characteristics of the strains. Specifically, we observed a correlation between the occurrence of 3HHx monomer-containing polymer and the maintenance of PHA content into the late stationary growth phase of the culture. For example, the phaC2-containing strains, which accumulate significant levels of 3HHx monomers in the PHA polymer, maintain higher levels of PHA in the stationary phase of culture growth when grown in hexanoate as a carbon source. However, this maintenance of PHA into stationary phase does not occur in acetate-grown cultures, and the PHA that is produced does not contain 3HHx monomer. The likely explanation for this phenomenon is that the PHA depolymerase (encoded by phaZ1 [Rru_A1585]) of R. rubrum has poor activity against PHA polymers that contain medium-chain-length monomers and greater specificity for degrading polymers that are composed of short-chain-length monomers (3HB and 3HV) (9).

Another observation that may be indicative of the physiological function of PHA in the growth characteristics of R. rubrum is the correlation between PHA content and the initial growth rate and final density of the cultures. These correlations are illustrated in Fig. 6 and indicate positive correlations between PHA content and growth fitness of the cultures, achieving higher cell density (Fig. 6A) and a higher growth rate (Fig. 6B). These correlations are also dependent on the carbon source, there being a stronger growth dependency on PHA content for the cultures grown on acetate than on hexanoate.

Fig 6.

Physiological consequence of genetic changes in PHA accumulation. (A) Correlation between final culture cell density and PHA content of R. rubrum strains. The OD of strains grown on acetate as the carbon source was obtained after 72 h of cultivation (△), and the OD of strains grown on hexanoate as the carbon source was obtained after 240 h of cultivation (□). (B) Correlation between culture doubling time and maximum PHA content of R. rubrum strains when grown on acetate (△) or hexanoate (□) as a carbon source. Data represent averages from triplicate biological samples, and error bars indicate the standard errors.

Although the underlying mechanisms for these correlations are unclear, previous studies have shown that accumulation of PHA has positive contributions to growth by enhancing resistance to stresses (e.g., radiation, desiccation, and osmotic pressure) (40). PHA serves as an electron and energy sink by acting as a store of highly reduced carbon and therefore provides a mechanism for cellular homeostasis of energy through carbon storage. Our data indicate that the capacity of the culture to accumulate PHA is correlated with the growth capabilities of the culture. A future direction of this research could be to investigate the role of the major regulator (guanosine-tetraphosphate) in coordinating growth and PHA production, as occurs in Pseudomonas oleovorans (20).