Efficient Production of Active Polyhydroxyalkanoate Synthase in Escherichia coli by Coexpression of Molecular Chaperones

Abstract

The type I polyhydroxyalkanoate synthase from Cupriavidus necator was heterologously expressed in Escherichia coli with simultaneous overexpression of chaperone proteins. Compared to expression of synthase alone (14.55 mg liter⁻¹), coexpression with chaperones resulted in the production of larger total quantities of enzyme, including a larger proportion in the soluble fraction. The largest increase was seen when the GroEL/GroES system was coexpressed, resulting in approximately 6-fold-greater enzyme yields (82.37 mg liter⁻¹) than in the absence of coexpressed chaperones. The specific activity of the purified enzyme was unaffected by coexpression with chaperones. Therefore, the increase in yield was attributed to an enhanced soluble fraction of synthase. Chaperones were also coexpressed with a polyhydroxyalkanoate production operon, resulting in the production of polymers with generally reduced molecular weights. This suggests a potential use for chaperones to control the physical properties of the polymer.

INTRODUCTION

Polyhydroxyalkanoates (PHAs) are one of the leading candidates to replace petrochemical plastics in many everyday applications (1). The key enzyme in production of PHAs is polyhydroxyalkanoate synthase (PhaC), as its substrate specificity determines which monomers can be polymerized.

PHA synthases are divided into four groups depending on the number of subunits constituting an active enzyme and on their specificity for monomers of different chain lengths (2). The PHA synthase from Cupriavidus necator has been adopted as the archetype for group I synthases and consequently is very well studied. C. necator was previously called Ralstonia eutropha, and the synthase protein is still widely known as PhaC_Re, a convention that is followed here.

As a type I synthase, PhaC_Re preferentially catalyzes the polymerization of short-chain (R)-hydroxyalkanoic acids (4 to 6 carbon atoms), particularly (R)-3-hydroxybutyrate-coenzyme A (3HB-CoA) to poly(hydroxybutyrate) (PHB) (3). 3HB-CoA is produced from acetyl-CoA by the sequential action of two enzymes: β-ketothiolase (PhaA) and acetoacetyl-CoA reductase (PhaB). The genes encoding these three enzymes constitute an operon in the genome of C. necator and are sufficient to allow production of PHB up to more than 90% of dry cell weight (DCW) when heterologously expressed in Escherichia coli (2).

Unfortunately, no crystal structure exists for any PHA synthase, because the enzyme tends to form inclusion bodies when overexpressed in a bacterial host or to aggregate during purification. Therefore, the catalytic mechanism is not fully understood, and studies of the enzyme must be carried out with the small quantities that can be produced.

There is interest in developing in vitro PHA production processes (4). This would facilitate the development of continuous production systems, eliminate the costly and environmentally damaging process of PHA recovery from within bacterial cells, allow the polymerization of monomeric units that are not easily produced by bacteria, and extend the range of physical properties of the final product. For in vitro production to be feasible on an industrial scale would require production of far larger quantities of PhaC than is currently possible. Therefore, it is important to improve upon the existing methods for PhaC expression both to assist scientific understanding and for more efficient industrial production.

Inclusion body formation is a well-known phenomenon in heterologous protein production, particularly in E. coli (5). Many methods to combat the problem have been suggested, including reducing the concentration of protein by modulating inducer quantities, expression at reduced temperatures (≤30°C), use of specialized “folding” strains such as E. coli Origami, and in vitro refolding of proteins from isolated inclusion bodies (6, 7). However, these techniques increase the time required to obtain active protein and suffer from drawbacks, including reduced protein yield and high cost.

Despite the difficulties involved with PhaC production, little work on improving the efficiency of the process has been reported. Reduced temperatures (typically 30°C) are often used to prevent inclusion body formation, and addition of a mild nonionic detergent such as 6-O-(N-heptylcarbamoyl)-methyl-alpha-d-glucopyranoside (Hecameg) is essential to prevent agglomeration of the enzyme during and after purification (3). Some attempts have also been made to resolubilize inclusion bodies formed by the type II synthases from Pseudomonas oleovorans using S-Sepharose (8) and from Pseudomonas putida by denaturing with 6 M guanidine hydrochloride, followed by refolding (9).

A widely used technique to aid production of biologically active heterologous proteins is simultaneously to overexpress one or more chaperone proteins, which constitute a diverse family, many of which belong to the group of heat shock proteins and are upregulated when the cell is under stress (7, 10, 11). This has been used with success for many different proteins, particularly eukaryotic or membrane-bound proteins (reviewed in reference 12), although the appropriate expression conditions must usually be determined experimentally on a case-by-case basis.

We report here the results of a study on chaperone-assisted PhaC_Re expression and PHA production. Three chaperone systems were chosen for coexpression with PhaC_Re: GroEL/GroES (plasmid pGro7), trigger factor (Tf; plasmid pTf16) and DnaK/DnaJ/GrpE (plasmid pKJE7). Additionally, Tf and GroEL/GroES were expressed together (plasmid pG-Tf2), as were GroEL/GroES and DnaK/DnaJ/GrpE (plasmid pG-KJE8). The Tf and DnaK/DnaJ/GrpE systems both bind to hydrophobic residues on nascent proteins, preventing aggregation until other folding steps have completed (13, 14). Poorly folded proteins can be targeted to the GroE complex, composed of a barrel with a hydrophobic core of GroEL and a lid of GroES. Conformational changes in GroEL, driven by ATP, force proteins within the barrel into more compact, properly folded conformations (7).

MATERIALS AND METHODS

Strains, plasmids, and genetic manipulation.

The strains and plasmids used in this study are described in Table 1.

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant features	Source or reference
Strains
E. coli BL21(DE3)	E. coli B F− ompT rB− mB− (λDE3)	Novagen
E. coli W3110hnsΔ93	E. coli K12 F− λ− rph-1 hnsΔ93 INV(rrnD, rrnE)	Chen et al.a
Cupriavidus necator H16	Wild type (ATCC 17699)	ATCC
Plasmids
pET15phaCRe	Expression vector providing N-terminal His6 tag containing phaCRe; T7 promoter; Apr	15
pASG1phaCRe	Expression vector containing phaCRe with C-terminal His6 tag; tet promoter; Apr	This study
pGro7	Expression vector for GroEL/GroES; araB promoter; Cmr	Clontech
pTf16	Expression vector for Tf; araB promoter; Cmr	Clontech
pG-Tf2	Expression vector for GroEL/GroES/Tf; pzt1 promoter; Cmr	Clontech
pKJE7	Expression vector for DnaK/DnaJ/GrpE; araB promoter; Cmr	Clontech
pG-KJE8	Expression vector for GroEL/GroES with pzt1 promoter; DnaK, DnaJ, GrpE with araB promoter; Cmr	Clontech
pTrcphaCAB	Expression vector containing C. necator phaCAB operon; trc promoter; Apr	16

^aC.-C. Chen, R. Walia, K. J. Mukherjee, and D. K. Summers, submitted for publication.

To construct plasmid pASG1phaC_Re, the StarGate cloning system (IBA, Germany) was utilized according to the manufacturer's instructions. The phaC_Re gene was amplified by PCR with Phusion DNA polymerase (Fisher Scientific) using 5′-phosphorylated primers to allow blunt-ended ligation into the linearized and dephosphorylated vector pENTRY10. pET15phaC_Re was used as the template. The forward primer sequence was AATGGCGACCGGCAAAGGC. The reverse primer sequence was TCCCGTGGTGGTGGTGGTGGTGTGCCTTGGCTTTGACGTATCG. Underlined bases represent partial recombination sites that are not expressed in the protein. Bases in bold type encode the His₆ tag. Following verification of gene entry by digestion with XbaI and HindIII, the His₆-tagged phaC_Re gene was transferred to plasmid pASG-IBAwt1 by recombination to create pASG1phaC_Re. The integrity of the new plasmid was verified by Sanger sequencing.

Growth conditions.

For PhaC_Re expression, strains were grown in 1-liter Erlenmeyer flasks with 200 ml of Luria-Bertani (LB) medium supplemented with ampicillin (100 μg ml⁻¹) and chloramphenicol (30 μg ml⁻¹) as necessary for plasmid selection. Overnight cultures in LB medium (2 ml) were used for inoculation, and the cultures were then incubated at 37°C while shaking at 250 rpm until the optical density at 600 nm (OD₆₀₀) was approximately 0.6. At this point, PhaC_Re production was induced as described below, and the incubation temperature was reduced to 30°C for a further 6 h. The cells were harvested by centrifugation at 4,000 × g for 15 min, washed once with deionized water, and stored at −80°C with 100,000 U lysozyme and 10 μl protease inhibitor for His-tagged proteins (both from Sigma-Aldrich).

The N-terminally tagged PhaC_Re was expressed from the T7 promoter by addition of isopropyl β-d-1-thiogalactopyranoside (IPTG, 1 mM). To accumulate chaperone proteins before beginning PhaC_Re expression, each culture was supplemented with arabinose (0.5 mg ml⁻¹) and/or anhydrotetracycline (0.2 μg ml⁻¹) depending on the promoter(s) used. The C-terminally tagged PhaC_Re was expressed from the tet promoter, which is also induced by anhydrotetracycline. Therefore, in this case only those chaperones controlled by the araB promoter were induced from inoculation, while PhaC_Re and the remaining chaperones were induced by addition of anhydrotetracycline (0.2 μg ml⁻¹) at an OD₆₀₀ of 0.6.

For PHB production, 200-ml cultures were grown in LB medium supplemented with glucose (20 g liter⁻¹). Ampicillin (100 μg ml⁻¹) and chloramphenicol (30 μg ml⁻¹) were added as required, and both chaperone expression and PHB production were induced at the time of inoculation by addition of IPTG (1 mM), arabinose (0.5 mg ml⁻¹), and, if necessary, anhydrotetracycline (0.2 μg ml⁻¹). Cultures were incubated at either 30 or 37°C with shaking at 250 rpm for 72 h. Cells were harvested by centrifugation at 4,000 × g for 15 min, washed once with deionized water, transferred to preweighed polypropylene tubes, freeze-dried, and then weighed to determine DCW.

Protein purification and analysis. (i) Cobalt affinity purification.

Cells were suspended in lysis buffer (2.65 mM NaH₂PO₄, 47.35 mM Na₂HPO₄, 300 mM NaCl, 10 mM imidazole, 5% [vol/vol] glycerol and 0.05% [wt/vol] Hecameg) (pH 8.0) and ruptured by 5 cycles of sonication using a Sonopuls HD2200 sonicator (Bandelin) and MS73 tip, set to a 30-s duration, 10% duty cycle, and 10% power, with 5 min on ice between cycles. Cell debris was removed by centrifugation at 4°C followed by filtration of the supernatant with a cellulose acetate syringe filter with a 0.44-μm pore size. Talon His tag purification resin (Clontech) was used for cobalt affinity protein purification according to the manufacturer's instructions. Lysis buffer was used to equilibrate and wash the column, and purified PhaC_Re was eluted using the same buffer containing 100 mM imidazole. Elution fractions containing PhaC_Re were combined and concentrated, and the buffer was replaced with storage buffer (2.65 mM NaH₂PO₄, 47.35 mM Na₂HPO₄, 5% [vol/vol] glycerol and 0.05% [wt/vol] Hecameg) (pH 8.0) using centrifugal filters with a molecular mass cutoff of 10,000 Da.

(ii) PhaC_Re activity assay.

The final protein concentration of each sample was determined by measuring the absorbance at 280 nm (A₂₈₀) with a Nanodrop spectrophotometer (Fisher Scientific). The molar extinction coefficient was taken to be 162,000 M⁻¹ cm⁻¹ (17). Enzyme activity was determined by measuring the decrease in A₂₃₆ resulting from the cleavage of the thioester bond with an extinction coefficient of 4,500 M⁻¹ cm⁻¹ (18). The reaction mixture consisted of 50 mM phosphate buffer, 200 μM (dl)-3-hydroxybutyryl-CoA (Sigma-Aldrich), 4 μg PhaC_Re, and double-distilled water (ddH₂O) with a total volume of 400 μl. One enzyme unit was defined as the amount of enzyme that catalyzed the release of 1 μmol CoA per min.

(iii) Polyacrylamide gel electrophoresis and Western blotting.

The purity of PhaC_Re was checked by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). A 20-μl portion of each sample of purified PhaC_Re was mixed with 10 μl of 3× SDS-PAGE buffer and incubated at 95°C for 3 min before being loaded onto a 10% polyacrylamide gel in Tris-glycine buffer. The loading volume was adjusted so that 1.22 μg protein was loaded per sample. Gels were stained with PageBlue protein-staining solution (Fermentas).

For Western blotting, 10 ml samples of cell culture were taken prior to harvesting. The cells were recovered by centrifugation and washed once in deionized H₂O. The cell pellet was resuspended in 500 μl BugBuster cell lysis buffer (Novagen) with 250 μl of Benzonase nuclease (Sigma-Aldrich) and shaken at room temperature for 30 min. The lysis mixture was centrifuged at 17,000 × g for 30 min at 4°C. The supernatant constituted the soluble protein fraction, and the pellet (the insoluble fraction) was dissolved in 500 μl of 1% (wt/vol) SDS. The proteins were separated by SDS-PAGE, and transferred to a polyvinylidene difluoride (PVDF) membrane using a Trans-Blot semidry blotting system (Bio-Rad). His₆-tagged PhaC_Re was specifically stained using the HisProbe HRP conjugate kit and metal-enhanced DAB substrate (Fisher Scientific). Densitometry analysis was performed using the gel analysis function of ImageJ (National Institutes of Health).

(iv) Analysis of PHB production.

Approximately 20 mg of dried cells were used for methanolysis to convert PHB into its methyl ester by mixing with 2 ml each of chloroform and methanol-sulfuric acid (85:15 vol/vol) and heating at 100°C for 140 min (19). After addition of 1 ml ddH₂O to induce phase separation, the filtered lower chloroform layer was subjected to gas chromatography (GC) analysis (Shimadzu GC 2014), with methyl benzoate (0.05%) as an internal standard, in order to calculate the amount of accumulated PHB.

PHB was extracted from dried cells by stirring with 50 ml chloroform per 100 mg PHB for 72 h. After filtering to remove cell debris, the polymer was precipitated by slowly dropping the chloroform solution into methanol and recovered by a further filtration step. The purified polymer was then redissolved in chloroform at a concentration of 1 mg ml⁻¹ for gel permeation chromatography (GPC) to determine the molecular weight. The GPC analysis was conducted at 40°C using a Shimadzu 10A GPC system with Shodex K-806 M and K-802 columns and a 10A refractive index detector. Chloroform was used as the eluent at a flow rate of 0.8 ml min⁻¹. Polystyrene standards (M_p [peak molecular mass] = 1.3 × 10³ to 7.5 × 10⁶) with low polydispersity were used to produce a calibration curve.

RESULTS

Production and purification of PhaC_Re.

To test the productivity and activity of PhaC_Re, with or without coexpression of chaperone proteins, N-terminally His₆-tagged PhaC_Re was purified from 200-ml shake flask cultures of E. coli BL21(DE3) containing pET15phaC_Re, by cobalt affinity chromatography. It has been reported that modifications to the C terminus of PhaC_Re can be introduced without affecting enzyme activity only if the hydrophobic environment surrounding the C terminus is maintained (20). To test whether the coexpression of chaperones could prevent the reduction in PhaC_Re activity caused by a C-terminal His₆ tag, the experiment was repeated using pASG1phaC_Re to express a C-terminally tagged protein.

Growth of each culture was quantified by measuring the average final OD₆₀₀ (Table 2). If necessary, cultures were diluted 10-fold with LB medium to bring the optical density within the linear range of the spectrophotometer. Growth was not affected by expression of GroEL/GroES (pGro7) or Tf (pTf16), but was substantially reduced by expression of GroEL/GroES/Tf (pG-Tf2), DnaK/DnaJ/GrpE (pKJE7) and GroEL/GroES/DnaK/DnaJ/GrpE (pG-KJE8). Chaperone proteins were efficiently produced at high concentration, as determined by SDS-PAGE analysis (data not shown), and so were assumed to be saturating.

Table 2.

Effect of chaperone co-expression on yields of soluble PhaC_Re^a

Plasmid content	Chaperone(s) expressed	Final OD600b	Protein recovered (mg liter−1)	Specific productivity (mg liter−1 OD unit−1)
pET15phaCRec
Alone		4.77 ± 1.7	14.55 ± 6.2	3.55 ± 2.7
+pGro7	GroEL/GroES	4.89 ± 1.4	44.37 ± 9.4	9.99 ± 5.2
+pTf16	Tf	4.73 ± 1.2	36.14 ± 9.5	8.20 ± 4.0
+pG-Tf2	GroEL/GroES/Tf	0.74 ± 0.3	13.44 ± 5.5	23.13 ± 20.1
+pKJE7	DnaK/DnaJ/GrpE	1.65 ± 0.5	10.10 ± 4.0	7.10 ± 4.7
+pG-KJE8	GroEL/GroES/DnaK/DnaJ/GrpE	2.00 ± 0.9	15.84 ± 5.9	8.71 ± 4.6
pASG1phaCRed
Alone		5.17 ± 1.4	1.77 ± 1.1	0.32 ± 0.1
+pGro7	GroEL/GroES	5.20 ± 0.2	13.40 ± 2.3	2.57 ± 0.4
+pTf16	Tf	4.93 ± 0.3	17.67 ± 3.5	3.58 ± 0.6
+pG-Tf2	GroEL/GroES/Tf	0.48 ± 0.1	8.00 ± 4.1	16.70 ± 8.7
+pKJE7	DnaK/DnaJ/GrpE	1.38 ± 0.8	1.68 ± 0.9	1.22 ± 0.1
+pG-KJE8	GroEL/GroES/DnaK/DnaJ/GrpE	3.49 ± 1.1	10.35 ± 6.7	2.78 ± 0.9

^aAll values are averages for cultures grown in triplicate ± standard deviations.

^bOptical density was measured at 600 nm immediately before cells were harvested.

^cN-terminally His₆ tagged.

^dC-terminally His₆ tagged.

Table 2 also shows the average amount of soluble PhaC_Re recovered per liter of bacterial culture. The N-terminally His₆-tagged version of PhaC_Re (pET15phaC_Re) was recovered in much larger quantities than the C-terminally His₆-tagged version (pASG1phaC_Re). Within the set of strains expressing the N-terminally His₆-tagged version, coexpression with GroEL/GroES and with Tf resulted in approximately 3-fold increases in soluble protein, with 44.37 mg liter⁻¹ and 36.14 mg liter⁻¹, respectively, being recovered, compared to 14.55 mg liter⁻¹ for PhaC_Re alone. The other chaperone systems resulted in either a small increase (GroEL/GroES/DnaK/DnaJ/GrpE) or a small decrease (GroEL/GroES/Tf and DnaK/DnaJ/GrpE) compared to the control.

A similar pattern was seen for the set of C-terminally tagged PhaC_Re strains. GroEL/GroES (13.40 mg liter⁻¹) and Tf (17.67 mg liter⁻¹) resulted in 8- to 10-fold increases compared to the control (1.77 mg liter⁻¹). GroEL/GroES/Tf and GroEL/GroES/DnaK/DnaJ/GrpE coexpression resulted in 5- to 6-fold increases, to 8.00 mg liter⁻¹ and 10.35 mg liter⁻¹, respectively.

To allow for the differences in final culture density, the specific productivity of PhaC_Re (mg protein per liter per OD₆₀₀ unit) was also calculated (Table 2). This confirms that the production of PhaC_Re is far more efficient with an N-terminal His₆ tag (3.55 mg liter⁻¹ OD unit⁻¹) than a C-terminal His₆ tag (0.32 mg liter⁻¹ OD unit⁻¹). Interestingly, despite causing a serious reduction in growth, coexpression with GroEL/GroES/Tf resulted in by far the largest specific productivity for both N-terminally (23.13 mg liter⁻¹ OD unit⁻¹) and C-terminally (16.70 mg liter⁻¹ OD unit⁻¹) tagged synthase.

A one-step purification protocol was used to reduce sample losses. Therefore, some coeluting proteins remained in the final PhaC_Re solutions. To assess the degree of contamination from coeluting proteins, samples (approximately 1.2 μg) of PhaC_Re solution from either pET15phaC_Re or pASG1phaC_Re with or without the coexpression of chaperones were electrophoresed through a 10% polyacrylamide gel (Fig. 1). Individual protein bands were excised and identified by trypsin digestion followed by mass spectrometry to produce a protein fingerprint.

Fig 1.

SDS-PAGE gels stained with Coomassie blue, showing PhaC_Re purified by cobalt affinity chromatography after expression in BL21(DE3)/pET15phaC_Re (a) or BL21(DE3)/pASG1phaC_Re (b) with and without chaperone protein coexpression. Bands were identified by digestion with trypsin followed by mass spectrometry fingerprinting. M, prestained molecular mass markers; lane 1, PhaC_Re only; lanes 2 to 6, coexpression from pGro7, pTf16, pG-Tf2, pKJE7, and pG-KJE8, respectively. PhaC_Re purities shown in the figure were calculated from densitometry analysis.

The major band in each sample was identified as PhaC_Re, which comprised 76 to 90% of the protein in each sample as determined by densitometry using ImageJ software (Fig. 1). Additional bands were observed, particularly in the samples with coexpression of DnaK/DnaJ/GrpE. The most abundant of these was an ∼56-kDa digestion product of PhaC_Re (10 to 24% of total protein), which most likely resulted from cleavage of 90 to 100 amino acids from the end of the protein. Low concentrations of the DnaK and DnaJ chaperones were also detected in the samples where they were overexpressed. Two additional proteins, which were identified as a putative ligase and SlyD (a peptidyl-prolyl cis-trans isomerase similar to trigger factor), were detected in samples of PhaC_Re expressed from pASG1phaC_Re but not in samples from pET15phaC_Re. None of the copurified proteins, including the PhaC_Re degradation product, are expected to display PHA polymerase activity.

Using estimated absorption coefficients generated by the online ProtParam program (web.expasy.org/protparam), we determined the percentage contribution to the A₂₈₀ of PhaC_Re in each sample (data not shown). Due to the relatively high molecular weight and extinction coefficient of PhaC_Re, the coeluted protein contributed disproportionately less to the estimate of protein concentration than PhaC_Re. Therefore, while the average purity of PhaC_Re (in the case of the C-terminally tagged protein, which contained more coeluted proteins) was 82%, it contributed toward an average of 89% of the absorbance. This indicates that the average error attributed to impurities following a one-step purification procedure is approximately 11%.

Effect of chaperones on PhaC_Re solubility.

To find out whether the increased yield of soluble PhaC_Re when coexpressed with chaperones was due to a net increase in PhaC_Re synthesis or to an increase in the proportion of soluble protein, soluble and insoluble protein fractions from cells containing pET15phaC_Re alone or with each of the chaperone systems were compared by SDS-PAGE. Because DnaK (70 kDa), GroEL (60 kDa), and trigger factor (56 kDa) all have sizes similar to that of PhaC_Re, they obscured the PhaC_Re band on an SDS-PAGE gel. Therefore, the PhaC_Re was selectively visualized by Western blotting (Fig. 2). The loading volume in each lane was normalized according to OD₆₀₀.

Fig 2.

Distribution of PhaC_Re between the insoluble (I) and soluble (S) protein fractions. PhaC_Re was expressed from BL21(DE3) containing pET15phaC_Re on its own or coexpressed with chaperone proteins. The relative densities were calculated using ImageJ software. Density relative to PhaC_Re is the ratio of the density of each band to the soluble fraction of PhaC_Re alone (bold). Relative densities are the ratios of soluble to insoluble fractions within each sample. pGro7, GroEL/GroES; pTf16, Tf; pG-Tf2, GroEL/GroES/Tf; pKJE7, DnaK/DnaJ/GrpE; pG-KJE8, GroEL/GroES/DnaK/DnaJ/GrpE.

With the exception of protein from cells expressing GroEL/GroES/Tf, which grew very poorly, making recovery of protein difficult, the intensities of the bands for the insoluble fractions were similar (ranging from 1.3 to 2.8 times the intensity of the soluble control). However, the intensity of the bands for the soluble fractions increased when chaperone proteins were coexpressed. The two best-performing systems from the purification experiment, GroEL/GroES and Tf, had 1.6- and 2.0-fold more soluble PhaC_Re than the control. However, the relative proportions of PhaC_Re in these two strains were 0.57 and 0.80, respectively. This is larger than that of PhaC_Re expressed without chaperones (0.4) but smaller than those for the other three strains. This suggests that the other chaperone combinations were more efficient at solubilizing the PhaC_Re that they produced but either produced less PhaC_Re per cell or resulted in less cell growth.

These values are smaller than the relative increases in soluble protein recovered during protein purification, possibly because antibody binding was inhibited by the large quantities of chaperone proteins present in the Western blots. Therefore, these results should be considered only as a semiquantitative indication of the relative amounts of PhaC_Re in each fraction. As such, they suggest that chaperones successfully increased both the total amount of protein produced and its solubility, although substantial amounts of protein remained in the insoluble fraction.

Optimization of inducer concentrations.

The most successful coexpression combination in the initial study was N-terminally tagged PhaC_Re with GroEL/GroES, which resulted in a 3-fold increase in soluble protein recovered (Table 2). To investigate whether further increases could be achieved, cultures of BL21(DE3) containing pET15phaC_Re and pGro7 were grown as before with 0.1, 0.5, or 1.0 mM IPTG and 0, 0.25, 0.5, 1.0, 2.5, or 5.0 mg ml⁻¹ arabinose. The results are displayed in Table 3.

Table 3.

Optimization of PhaC_Re production by varying inducer concentration in BL21(DE3) containing pET15phaC_Re and pGro7

Arabinose concn (mg/ml)	Protein recovery (mg liter−1) at IPTG concn (mM) of:			Specific productivity (mg liter−1 OD unit−1) at IPTG concn (mM) of:
	0.1	0.5	1.0	0.1	0.5	1.0
0	12.71	10.43	13.89	2.93	2.35	2.67
0.25	51.44	49.48	49.90	10.03	10.48	8.09
0.50	66.53	51.12	52.04	10.35	10.08	7.95
1.00	81.65	50.33	59.89	14.49	10.16	9.90
2.50	78.92	58.94	66.90	14.47	12.38	11.54
5.00	82.37	59.88	64.60	15.43	10.97	11.71

Lower concentrations of IPTG resulted in larger quantities of PhaC_Re, suggesting that gentle induction is necessary to allow time for the protein to fold (Table 3). Increasing the concentration of arabinose also increased PhaC_Re production. Consequently, both the largest amount of PhaC_Re (82.37 mg liter⁻¹) and the highest specific productivity (15.43 mg liter⁻¹ OD unit⁻¹) were achieved with 0.1 mM IPTG and 5.0 mg ml⁻¹ arabinose. This corresponds to a 5.7-fold increase compared to production from pET15phaC only. Gentle expression of PhaC_Re combined with coexpression of large amounts of GroEL/GroES is therefore the best strategy to increase soluble PhaC_Re production.

Influence of chaperone coexpression on PhaC activity.

To calculate the specific activity of our PhaC_Re preparations, the total yield of PhaC_Re protein from each 200-ml culture was calculated from the final volume and concentration of purified protein solution after desalting. Activity assays were performed in triplicate for each sample by measuring the decrease in absorbance at 236 nm, corresponding to the 3HB-CoA thioester bond that is broken during polymerization. Analyses for statistical significance were performed using two-tailed t tests assuming unequal sample variance.

The average specific activity of N-terminally His₆-tagged PhaC_Re produced without chaperone protein coexpression was 12.1 U mg⁻¹ (Fig. 3). The cultures in which protein recovery was increased showed no significant change in specific activity (P > 0.05). However, the specific activity was significantly reduced (P < 0.01) by coexpression of GroEL/GroES/Tf (9.0 U mg⁻¹) and DnaK/DnaJ/GrpE (8.8 U mg⁻¹). When the specific activity was multiplied by the protein productivity to give the average total yield of enzyme (U liter⁻¹), GroEL/GroES and Tf again resulted in significant (P < 0.01) increases (586.3 and 439.7 U liter⁻¹, respectively) compared to PhaC_Re alone (168.7 U liter⁻¹). Only DnaK/DnaJ/GrpE significantly (P < 0.01) decreased the yield (82.4 U liter⁻¹). Thus, the most influential effect of the GroEL/GroES and Tf systems (plasmids pGro7 and pTf16) was to increase the amount of soluble protein produced, rather than increasing its specific activity.

Fig 3.

Average specific activities (hatched bars) and total yield per liter (crosshatched bars) for PhaC_Re produced from pET15phaC_Re (N-terminal His₆ tag) (a) and pASG1phaC_Re (C-terminal His₆ tag) (b) with and without coexpression of chaperone proteins. Error bars represent the standard deviations of results from cultures grown in triplicate, with three replicates of each assay. Values that are significantly different (P < 0.01) from that for PhaC_Re only are indicated by asterisks.

The specific activity of C-terminally tagged PhaC_Re (expressed from pASG1phaC_Re) was significantly lower (P < 0.01) than that of the N-terminally tagged protein (Fig. 3b), so chaperones were not able to rectify the misfolding of PhaC_Re caused by His₆ tag addition to the C terminus. Compared to an average specific activity of 1.7 U mg⁻¹ for protein produced in cells containing pASG1phaC_Re alone, chaperone coexpression either showed no change (P > 0.05) or resulted in a decrease (P < 0.01). Coexpression with GroEL/GroES or Tf both resulted in average specific activities of 1.7 U mg⁻¹. The other three chaperone combinations all resulted in decreased specific activities, with the smallest being 0.9 U mg⁻¹ for DnaK/DnaJ/GrpE. As seen for the N-terminally His₆-tagged PhaC_Re, the differences between the yields of purified protein were more important than the variation in specific activity. The greatest yield was 30.2 U liter⁻¹ for Tf coexpression, which represents a 10-fold increase over the yield of 3.0 U liter⁻¹ when PhaC_Re was expressed alone. Coexpression with every chaperone combination except DnaK/DnaJ/GrpE resulted in a significant increase (P < 0.01) in protein yield compared to PhaC_Re alone.

Effect of chaperones on PHB production.

In E. coli, the molecular weight of PHB varies in inverse proportion to the amount of PhaC_Re produced (21). As an initial investigation into the effects of improved PhaC_Re production (due to chaperone coexpression) on PHB biosynthesis, E. coli W3110hnsΔ93-1 carrying plasmid pTrcphaCAB alone and equivalent strains carrying each of the five chaperone expression plasmids were grown for 72 h in LB medium supplemented with glucose (20 g liter⁻¹) at either 30 or 37°C. The long culture time was chosen to allow sufficient time for PHB accumulation, even if chaperone coexpression negatively influenced cell growth, as for PhaC_Re production.

In every case, growth (determined by measuring DCW) was better at 30°C than at 37°C (Table 4). As seen in the PhaC_Re production experiments, chaperone protein coexpression sometimes inhibited bacterial growth, and this was particularly evident when cells contained the pGro7 or pG-Tf2 chaperone expression plasmids.

Table 4.

Production of PHB during overexpression of molecular chaperone proteins

Temp	Protein(s) expressed	DCW (g liter−1)	PHB content (% DCW)	Mn (106)	Mw (106)	Mw/Mn
30°C	PhaCAB
	Alone	6.37	65	1.37	3.09	2.3
	+GroEL/GroES	2.12	27	1.29	2.78	2.2
	+Tf	6.85	63	1.07	2.18	2.0
	+GroEL/GroES/Tf	2.24	3	NDa	ND	ND
	+DnaK/DnaJ/GrpE	7.47	49	1.42	3.00	2.1
	+GroEL/GroES/DnaK/DnaJ/GrpE	7.76	59	0.86	1.93	2.2
37°C	PhaCAB
	Alone	2.46	71	2.34	4.46	1.9
	+GroEL/GroES	0.93	13	1.46	2.75	1.9
	+Tf	2.06	70	2.07	4.71	2.3
	+GroEL/GroES/Tf	1.08	1	ND	ND	ND
	+DnaK/DnaJ/GrpE	2.20	44	2.42	4.66	1.9
	+GroEL/GroES/DnaK/DnaJ/GrpE	2.53	28	1.60	3.35	2.1

^aND, not determined.

There were large variations in the amount of PHB produced by each culture, both between temperatures and between chaperone expression systems. In the absence of chaperone coexpression, PHB accumulated to 65% DCW at 30°C and to 71% DCW at 37°C. In almost every case, coexpression of chaperone proteins substantially reduced the yield of PHB. The biggest reductions were for GroEL/GroES and GroEL/GroES/Tf, with the smallest yields being for GroEL/GroES/Tf coexpression (3% and 1% at 30 and 37°C, respectively).

The number-averaged molecular weights (M_n) were generally decreased by chaperone coexpression (Table 4). At 30°C, Tf and GroEL/GroES/DnaK/DnaJ/GrpE reduced the M_n of PHB by 21.8% and 37.2%, respectively. At 37°C the largest reductions were with GroEL/GroES and GroEL/GroES/DnaK/DnaJ/GrpE (37.6% and 31.6%). Additionally, M_n for each sample was larger at 37°C than at 30°C. Polydispersities (defined as the ratio of M_w to M_n) were not significantly affected by chaperone protein coexpression. Due to poor growth and low PHB yield, molecular weight data were not gathered for GroEL/GroES/Tf.

DISCUSSION

Chaperone protein coexpression was found to assist in the production of soluble PhaC_Re in BL21(DE3). The most significant improvements were seen with the production of N-terminally tagged PhaC_Re and chaperone expression was not able to restore reduced activity caused by a C-terminal His₆ tag. Coexpression of the GroEL/GroES operon (pGro7) was the most successful strategy, increasing soluble PhaC_Re recovery 5.7-fold. Enzyme activity was largely unaffected by chaperone protein coexpression, suggesting that soluble PhaC_Re is correctly folded in E. coli. The increase in yield was due to substantial increases in the quantity of soluble, active protein that could be recovered by cobalt affinity chromatography. Therefore, the chaperone proteins assisted the folding of polypeptide chains that would otherwise have accumulated in inclusion bodies or be targeted for protein degradation.

As shown in Fig. 2, considerable quantities of insoluble PhaC_Re remained in each strain, although the fraction of soluble protein was increased by coexpressing chaperones. This suggests that chaperone proteins are able to allow greater quantities of PhaC_Re to be produced in total, either by increasing production or by reducing the amount of protein that is targeted for protease degradation.

Optimization of the expression strategy resulted in even higher yields of soluble PhaC_Re (Table 3). The most successful strategy was to use gentle induction of PhaC_Re expression (0.1 mM IPTG) but high-level expression of the GroEL/GroES system (5 mg ml⁻¹ arabinose). This ensures a plentiful supply of chaperones to prevent aggregation and assist folding of each PhaC_Re polypeptide as soon as it is produced. Further improvements may be possible by combining chaperone coexpression with other strategies, such as expression at reduced temperature, use of longer culture times, or expression in fermentor cultures.

Reduced bacterial growth, a known side effect of chaperone protein coexpression (7), contributed to the difference in total protein yields between the strains used in this study. Another potential cause of these variations is plasmid instability. We attempted to quantify the magnitude of this effect by plating samples of each culture on antibiotic-free and antibiotic-containing LB agar plates, prior to induction (OD₆₀₀ = 0.6) and after the 6-h production period, and comparing the number of colonies (data not shown). There was no evidence for plasmid instability prior to induction. However, fewer colonies grew with or without antibiotics following protein production, indicating a loss of colony-forming ability during expression, as suggested by others (22). This is in keeping with the observation that the high specific productivity of strains containing pG-Tf2 appeared to be deleterious to growth and suggests that most cells retain the plasmids during protein accumulation. Modulation of the level of chaperone expression could potentially reduce the negative effects on bacterial growth while maintaining the beneficial effects on protein folding, although this effect was not seen in our optimization experiment.

There were differences between the specific activities of PhaC_Re expressed with different chaperone systems. It is likely that much of the difference in specific activities is due to variations in the relative amounts of contaminating proteins in each sample. In particular, coexpression from pKJE7 had a larger number of coeluting proteins than other samples and also had a low specific activity (Fig. 1 and 3). The contaminating proteins were identified as a mixture of chaperones (DnaK/J), SlyD, and degradation products of PhaC_Re. Both DnaK and DnaJ, as well as SlyD, are common contaminants of His₆ tag-purified proteins due to naturally occurring oligo(His) sequences in their primary structure (23). Therefore, it is not surprising to find traces of them, particularly in a strain overexpressing DnaK and DnaJ proteins.

The specific activity of PhaC_Re has previously been reported to be 40 U mg⁻¹, which is substantially higher than the activities reported here (17). The protein preparation in the previous report involved an extra purification step using size exclusion chromatography to remove soluble aggregations of PhaC_Re. Although such a procedure would be expected to increase the specific activities in this case (and to reduce variance), absolute specific activity had far less influence on relative yields between cultures than the increase in total soluble protein. Therefore, the conclusions following a one-step purification procedure remain valid.

The C-terminally tagged protein produced from pASG1phaC_Re had much lower specific activities than N-terminally tagged protein from pET15phaC_Re and was also produced in lower quantities. This suggests that chaperone proteins were not able to rectify the misfolding of PhaC_Re caused by the addition of a His₆ tag to the C terminus. It is not clear whether the lower quantity of protein was due to differences in the expression signals and/or copy number between the plasmids, increased accumulation in inclusion bodies, or recognition of the badly folded protein and targeting for degradation. However, since chaperone proteins were still able to increase the amount of soluble PhaC_Re, it is evident that they play a part in the folding mechanism and are at least partially successful in preventing the aggregation of badly folded protein.

Chaperone protein coexpression generally resulted in a decrease in both the number-averaged and weight-averaged molecular weights of PHB produced in E. coli, particularly for the GroEL/GroES, Tf, and GroEL/GroES/DnaK/DnaJ/GrpE systems. This can be attributed at least in part to the production of a larger number of active PhaC_Re molecules. These compete for the available monomeric units and catalyze the production of a larger number of shorter chains than when only the PHB operon is expressed. Consistent with this, pKJE7 caused a small increase in M_n and was also responsible for a reduction in PhaC_Re productivity.

Several studies have shown that the addition of exogenous hydroxylated molecules can usefully alter the molecular weights of PHAs generated in vivo. This is thought to occur by transmembrane migration of the agents to assist in chain termination of the polymer (24, 25). Our results point to the use of chaperone-mediated alteration of the monomer-enzyme balance as an alternative route to altering the molecular weight distribution of PHAs. Modulation of PhaC activity has previously been used to vary the molecular weight and yield of PHA produced in E. coli (26) and P. putida (27). Our results agree with the previous finding that PHA molecular weights are inversely related to PhaC activity. While product control by an exogenous agent can be toxicity limited, the expression of the endogenous chaperone could be more flexibly tuned to control the molecular weight of the resulting polymer in order to suit a range of applications.

PHB production relies on an interconnected web of related biochemical pathways, including cellular respiration, which provides the acetyl-CoA that is converted into the monomeric unit for PHB. Since chaperone proteins are actively involved in the folding of the majority of bacterial proteins, it is very likely that overexpression of any of the systems described in this study will simultaneously also affect the levels of the PhaA and PhaB proteins as well as many more diverse native proteins. This hinders a simple interpretation of the results but suggests that the large variations in bacterial growth and PHB accumulation could be due to imbalances in the core biochemical pathways of the cells. Further studies are required to fully understand the mechanism of PHB molecular weight reduction during chaperone coexpression and how this could be translated into a commercially relevant production process.

Chaperone protein coexpression is an effective method for increasing the yield of PhaC_Re. By using a widely available set of expression plasmids, this study has demonstrated increases in total enzyme yield of up to 6 times without the need for costly and time-consuming processes such as expression at reduced temperature or in vitro protein refolding. This will facilitate the laboratory-scale study of the enzyme for the purposes of elucidating its structure and catalytic mechanism as well as developing in vitro PHB production methods. The method should also be easily transferrable to the study of other PHA synthases.