Isolated Poly(3-Hydroxybutyrate) (PHB) Granules Are Complex Bacterial Organelles Catalyzing Formation of PHB from Acetyl Coenzyme A (CoA) and Degradation of PHB to Acetyl-CoA

Abstract

Poly(3-hydroxybutyrate) (PHB) granules isolated in native form (nPHB granules) from Ralstonia eutropha catalyzed formation of PHB from ¹⁴C-labeled acetyl coenzyme A (CoA) in the presence of NADPH and concomitantly released CoA, revealing that PHB biosynthetic proteins (acetoacetyl-CoA thiolase, acetoacetyl-CoA reductase, and PHB synthase) are present and active in isolated nPHB granules in vitro. nPHB granules also catalyzed thiolytic cleavage of PHB in the presence of added CoA, resulting in synthesis of 3-hydroxybutyryl-CoA (3HB-CoA) from PHB. Synthesis of 3HB-CoA was also shown by incubation of artificial (protein-free) PHB with CoA and PhaZa1, confirming that PhaZa1 is a PHB depolymerase catalyzing the thiolysis reaction. Acetyl-CoA was the major product detectable after incubation of nPHB granules in the presence of NAD⁺, indicating that downstream mobilizing enzyme activities were also present and active in isolated nPHB granules. We propose that intracellular concentrations of key metabolites (CoA, acetyl-CoA, 3HB-CoA, NAD⁺/NADH) determine whether a cell accumulates or degrades PHB. Since the degradation product of PHB is 3HB-CoA, the cells do not waste energy by synthesis and degradation of PHB. Thus, our results explain the frequent finding of simultaneous synthesis and breakdown of PHB.

Poly(3-hydroxybutyrate) (PHB) is the most prominent member of the bacterial polyhydroxyalkanoates (PHA). PHA are osmotic neutral reservoirs of carbon and energy and are synthesized when more carbon sources are available than can be consumed by the bacteria. PHA are reutilized (mobilized) during times of starvation and greatly enhance the survival of bacteria in the absence of a suitable exogenous carbon source.

PHB is synthesized by condensation of two molecules of acetyl coenzyme A (CoA) to acetoacetyl-CoA (with a thiolase encoded by phaA), subsequent reduction to 3-hydroxybutyryl-CoA (with a reductase encoded by phaB), and polymerization to PHB (with a synthase encoded by phaC). The polymer can be hydrolyzed to 3-hydroxybutyrate by PHB depolymerases (PhaZs). Biosynthesis and biodegradation of PHA have been investigated by many research groups for about three decades, and a series of books and reviews have been published (11, 13, 30, 34, 39-42). PHB exists in two different forms. In vivo PHB granules consist of an amorphous polymer and are covered by a dense layer consisting of mainly proteins (phasins, PHB synthase, PHB depolymerases, and other proteins) (14, 22, 31, 38). Such granules are called native PHB (nPHB) granules and can be isolated in the native form by glycerol density gradient centrifugation (12, 23, 45). Isolated PHB granules that have been treated with solvents or with compounds that remove the surface layer rapidly crystallize and are referred to as denatured PHB granules (23, 24). (For more details on the impact of the biophysical state of the polymer granules on their susceptibility to enzymatic hydrolysis see references 5, 8, and 12.)

Despite the progress made in understanding the function of individual proteins involved in PHA metabolism, the molecular tools and mechanisms with which a cell decides whether it should synthesize or degrade (mobilize) PHB are not known. Several reports have indicated that PHB synthesis and PHB degradation can happen simultaneously in Ralstonia eutropha, the model organism for PHB metabolism (3, 43). The finding that there is constitutive expression of PHB synthase and PHB depolymerases in R. eutropha is in agreement with these findings (20). However, synthesis of PHB by condensation of 3-hydroxybutyryl-CoA (3HB-CoA) monomer units to PHB and free CoA by PHB synthase and simultaneous hydrolysis of PHB to 3-hydroxybutyrate (3HB) by intracellular PHB depolymerase make no physiological sense as this would be a futile cycle and waste energy in the form of hydrolyzed thioester bonds. A convincing explanation for this apparent contradiction cannot be given. The literature contains many reports of the existence of putative intracellular PHA depolymerases in R. eutropha and other bacteria (2, 12, 34). Meanwhile, as many as seven PHB depolymerases and two 3HB oligomer hydrolase genes are thought to be involved in PHB metabolism in R. eutropha (1, 9, 16, 17, 32, 33, 47). The experimental data showing that the gene products are physiologically important intracellular PHB depolymerases are, however, poor. Only for PhaZa1 is involvement in PHB mobilization supported by independent contributions (9, 32, 47). Nevertheless, we were not able to show significant in vitro PHB depolymerase activity with nPHB granules as the substrate and added PhaZa1 (unpublished data). Either PHB depolymerase has no depolymerase activity in vitro under the conditions used or nPHB granules are so densely covered with proteins that excess depolymerase protein cannot bind to the polymer core. We therefore decided to examine nPHB granules as a whole system. For this, we first confirmed that PhaZa1 is a PHB granule-bound protein in vivo by performing a fusion analysis with green fluorescent protein. Isolated nPHB granules from R. eutropha were then examined for various metabolic functions in PHB synthesis and PHB mobilization.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. R. eutropha was grown in nutrient broth (NB) or mineral salts medium with sodium gluconate (or fructose) as described previously at 30°C (1, 5). Escherichia coli strains were maintained in Luria-Bertani broth (37°C). For induction of genes under control of rhamnose or the lac promoter, 0.2% (wt/vol) rhamnose or 0.1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) was added. Sodium acetate (0.4%, wt/vol) was added for synthesis of PHB. Plasmids harboring various combinations of PHB biosynthetic genes (Table 1) were constructed by standard methods. Fusion of phaZa1 with egfp was performed like apdA-egfp fusion (10).

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant characteristics	Source or reference
Ralstonia eutropha strains
H16	Wild type, accumulation of PHB	DSM428
H16-SK1544	phaZa1 deletion strain	9
H16 Re1052	ΔphaP1	46
Escherichia coli strains
BL21(DE1)	Expression strain	Novagen
S17-1	Mobilizing strain	36
JM109	Cloning strain
Plasmids
pBBRMCS2	Broad-host-range plasmid, Kmr	18
pBBRMCS2::phaZa1-egfp	Expression of PhaZa1-Egfp fusion	This study
pHWG640	Expression plasmid with rhamnose promoter
pBHR68	Source of phaCAB operon	37
pHWG640::phaCAB	Accumulation of PHB	This study
pHWG640::phaP1-phaCAB	Accumulation of PHB with PhaP1	This study
pET171H	pET23 harboring phaZa1 under T7 promoter	32

Isolation of PHB granules.

nPHB granules were isolated from French press-lysed cells (in the presence of 1 mM dithiothreitol [DTT]) by two glycerol gradient centrifugations as described previously (5, 12, 23). Cholate-coated artificial PHB (aPHB) granules were prepared as described elsewhere (1).

Thiolysis of PHB.

A reaction mixture (0.5 ml) containing 250 μl nPHB granules (100 to 200 mg/ml depending on the batch of nPHB granules), 50 mM potassium phosphate buffer (pH 7.0), 1 mM CoA, and 1 mM DTT was incubated at 30°C. At selected time points samples (0.1 ml) were taken, acidified with 10 μl of 1 N HCl, and centrifuged (20,800 × g, 5 min). The supernatant was analyzed by high-performance liquid chromatography (HPLC). In some experiments nPHB granules were replaced by cholate-coated aPHB granules (final concentration, 2.5 mg/ml). Experiments with a cofactor regeneration system were performed in the presence of 0.1 M glucose, 1 mM NAD⁺, and 0.5 U glucose dehydrogenase (220 U/mg) or in the presence of 0.1 M glucose-6-phosphate, 1 mM NADP, and glucose-6-phosphate dehydrogenase (2.5 U).

Synthesis of PHB from acetyl-CoA.

A reaction mixture (0.1 ml) containing 10 μl nPHB granules, 50 mM potassium phosphate buffer (pH 7.0), 1 mM DTT, 1 mM NADPH, 0.1 M glucose-6-phosphate, 2.5 U glucose-6-phosphate dehydrogenase, 1 or 2 mM acetyl-CoA, and 0.17 mM [1-¹⁴C]acetyl-CoA (2.18 GBq/mmol; GE Healthcare) was incubated at 30°C for the times indicated below. Samples (50 μl) were acidified with 2.5 μl trichloroacetic acid, and 1 ml of water was added. The precipitate formed after centrifugation (20,800 × g, 5 min) was dissolved in trichloromethane (50°C) and washed three times with 1 ml water. The solvent was evaporated, and the solids were dissolved in 0.1 ml trichloromethane. After addition of 10 ml of scintillation fluid, the radioactivity was determined with a scintillation counter. The radioactivity in the supernatant of the stopped reaction mixture was also determined. A heat-treated control (20 min, 100°C) of nPHB granules served as a control.

HPLC conditions and other methods.

One to 5 μl of the supernatant of a reaction mixture was loaded onto a reverse-phase C₁₈ HPLC column (5 μm; 4.6 by 150 mm; Eclipse XDB-C18; Agilent). Samples were eluted at a flow rate of 0.8 ml/min. The buffers were 0.05 M potassium phosphate buffer (pH 4.7) (solution A) and pure methanol (solution B). The elution conditions were as follows: a gradient from 5% solution B to 25% solution B within 25 min, a gradient to 50% solution B within 8 min (33 min), and a gradient to pure solution B within 1 min (34 min). After this, the column was run isocratically with 5% solution B for 5 min (39 min). Solutions of CoA, acetyl-CoA, and 3-HB-CoA (Sigma) served as standards. pH stat experiments were performed with isolated nPHB, and released products were determined as described previously (5). All other experiments were performed by using standard procedures.

RESULTS

PhaZa1 is a PHB granule-bound protein in vivo.

The putative PHB depolymerase gene, phaZa1, was fused to the green fluorescent protein gene, egfp, cloned in broad-host-range vector pBBRMCS2, and transferred to R. eutropha. Transconjugant cells that were grown in NB for a few hours or in mineral salts medium with gluconate contained several PHB granules, as revealed by detection of refractive globular structures by phase-contrast microscopy. The same PHB granules showed green fluorescence (Fig. 1), indicating that the fusion protein was expressed and colocalized with PHB granules. When the cells were stained with Nile red, the granules showed the typical fluorescence of enhanced green fluorescent protein (Egfp) and Nile red (not shown). After 30 h of incubation NB cultures had lost all PHB granules, and green fluorescence could not be detected anymore. We concluded that PhaZa1-Egfp was expressed during phases of PHB accumulation and was associated with PHB granules.

FIG. 1.

Fluorescence microscopy analysis of R. eutropha H16 harboring pBBRMCS2::phaZa1-egfp during exponential growth. Bacteria were visualized with a Zeiss Axioplan fluorescence microscope using an F41-54 Cy2 filter. Pictures were taken under phase-contrast and fluorescence conditions with a digital camera (Coolsnap) and were processed with the Metaview/Metamorph software (Visitron Systems).

Catalytic properties of isolated nPHB granules.

Isolated nPHB granules were examined for PHB depolymerase activity by pH stat analysis as described previously (5). Wild-type granules isolated from gluconate-grown cells had a specific acid release rate of 4 nmol min⁻¹ mg nPHB⁻¹. When the same experiment was performed with nPHB granules isolated from the R. eutropha ΔphaZa1 mutant, a reduced value, 3 nmol min⁻¹ mg nPHB⁻¹, was obtained. However, the acid-releasing activity of nPHB granules varied by a factor of 3 for different batches of isolated nPHB granules, and the significance of the reduced activity of nPHB granules isolated from the ΔphaZa1 mutant is difficult to assess. Product analysis by enzymatic determination of 3HB using 3HB dehydrogenase and by HPLC analysis after derivatization with bromophenacylbromide showed that monomeric 3HB was the only detectable product of in vitro hydrolysis of isolated nPHB granules from the wild type and from the ΔphaZa1 mutant (data not shown). No indication of release of dimers or higher oligomers was obtained. To investigate the catalytic capabilities of nPHB-bound enzymes, nPHB granules isolated from R. eutropha were incubated in the presence of different substrates and potential cofactors. Samples were taken at intervals as indicated below and screened for soluble products by HPLC analysis.

Thiolysis of nPHB.

Isolated nPHB granules were incubated in the presence of 1 mM CoA. HPLC analysis of the soluble products showed that the area of the CoA peak decreased from 0.91 to 0.48 mM and that two new peaks appeared (Fig. 2 and Table 2, experiment 1). The two new peaks were identified as 3HB-CoA and acetyl-CoA by comparison with standard compounds. In addition, the masses (m/z values) of the ions at both peaks were determined by HPLC-electrospray ionization (ESI)-mass spectrometry (MS) and corresponded to the expected masses of 3HB-CoA (m/z 854) and acetyl-CoA (m/z 810). Detection of 3HB-CoA and acetyl-CoA strictly depended on the presence of CoA. Interestingly, the size of the acetyl-CoA peak increased with prolonged incubation time of the granules up to 0.21 mM, while the size of the 3HB-CoA peak slightly decreased from 0.29 to 0.24 mM (Fig. 3). This result suggests that nPHB granules are able to cleave PHB via thiolysis and that intermediately formed 3HB-CoA can be converted to acetyl-CoA.

FIG. 2.

HPLC analysis of products formed during thiolysis. nPHB granules isolated from R. eutropha wild-type strain H16 (wt) and from the ΔphaZa1 mutant were incubated in the presence of CoA as described in Materials and Methods. The reaction was stopped by acidification at the times indicated, and 5 μl of the supernatant obtained after centrifugation was analyzed by HPLC. Peaks identified as CoA, acetyl-CoA, and 3HB-CoA by comparison with standards and by HPLC-ESI-MS are indicated.

TABLE 2.

Assay of in vitro thiolysis^a

Expt	Conditions	Concn of CoA (mM) at:				Concn of acetyl-CoA (mM) at:				Concn of 3HB-CoA (mM) at:
		0 h	0.5 h	1 h	2 h	0 h	0.5 h	1 h	2 h	0 h	0.5 h	1 h	2 h
1	H16 nPHB, CoA	0.91	0.60	0.51	0.48	<0.03	0.06	0.12	0.21	<0.03	0.29	0.27	0.24
2	H16 nPHB, CoA, NADP	1.20	0.89	0.79	0.48	<0.03	0.10	0.21	0.36	<0.03	0.22	0.26	0.22
3	H16 nPHB, CoA, NAD	0.94	0.70	NDb	ND	<0.03	0.33	ND	ND	<0.03	<0.03	ND	ND
4a	H16 nPHB, CoA	0.96	0.63	ND	ND	<0.03	0.07	ND	ND	<0.03	0.27	ND	ND
4b	H16 nPHB, CoA + NADH system	0.94	0.53	ND	ND	<0.03	0.03	ND	ND	<0.03	0.21	ND	ND
4c	H16 nPHB, CoA + NADPH system	1.01	0.89	ND	ND	<0.03	<0.03	ND	ND	<0.03	0.15	ND	ND
4d	H16 nPHB, CoA + NADH system + NADPH system	0.96	0.86	ND	ND	<0.03	<0.03	ND	ND	<0.03	0.10	ND	ND
5	E. coli (phaCAB) nPHB, CoA	1.01	1.05	1.01	1.03	<0.03	<0.03	0.03	<0.03	<0.03	<0.03	<0.03	<0.03
6	E. coli (phaCAB + phaP1), nPHB, CoA	0.98	1.06	1.06	1.08	<0.03	<0.03	<0.03	<0.03	<0.03	<0.03	<0.03	<0.03
7	E. coli (phaCAB + phaZa1), nPHB, CoA	1.03	0.98	0.98	0.99	<0.03	0.03	0.03	<0.03	<0.03	<0.03	<0.03	<0.03
8	E. coli (phaCAB + phaP1 +phaZa1), nPHB, CoA	1.05	0.74	0.60	0.51	<0.03	<0.03	<0.03	<0.03	<0.03	0.06	0.14	0.21
9	H16 ΔphaP1 nPHB, CoA	1.08	0.79	0.67	0.67	<0.03	0.19	0.31	0.43	<0.03	0.13	0.08	0.03
10	H16 ΔphaZa1 nPHB, CoA	0.96	0.88	0.77	0.67	<0.03	0.05	0.13	0.24	<0.03	0.10	0.12	0.13
11	aPHB, CoA + PhaZa1	0.98	ND	0.79	0.70	<0.03	ND	<0.03	<0.03	<0.03	ND	0.15	0.21

^aThiolysis experiments were performed as described in Materials and Methods. The values were calculated from HPLC peak areas.

^bND, not determined.

FIG. 3.

Time course of consumption of CoA and formation of acetyl-CoA and 3HB-CoA during thiolysis of wild-type (wt) and ΔphaZa1 mutant nPHB granules as revealed by determination of peak areas after HPLC. The peak area for free CoA at time zero corresponds to 1 mM. The experimental conditions are described in the legend to Fig. 2.

Conversion of 3HB-CoA to acetyl-CoA requires the presence of cofactors, such as NAD⁺ or NADP⁺, which presumably are present at only trace levels in isolated nPHB granules. We therefore repeated the thiolysis experiments in the presence of added 1 mM NAD⁺ or 1 mM NADP⁺ (Table 2, experiments 2 and 3). Peaks corresponding to 0.36 mM acetyl-CoA and 0.22 mM 3HB-CoA were detected in the presence of 1 mM NADP⁺. Apparently, NADP⁺ has no significant effect on product formation. Interestingly, acetyl-CoA (0.33 mM), but no detectable 3HB-CoA, was found in the presence of 1 mM NAD⁺. We assumed that intermediately formed 3HB-CoA is rapidly converted to acetyl-CoA in NAD⁺-dependent reactions so that the concentration of 3HB-CoA is below the detection limit.

Next, we tested the influence of the presence of reduced cofactors [NAD(P)H] on thiolysis and subsequent reactions. To keep the concentration of the reduced cofactors constant and high, we added an NAD(P)H regeneration system to nPHB granules as described in Materials and Methods. Without the regeneration system we detected 0.27 mM 3HB-CoA and traces of acetyl-CoA (0.07 mM) when nPHB granules were incubated with CoA (Table 2, experiment 4a). However, in the presence of (i) the NADH regeneration system (experiment 4b), (ii) the NADPH regeneration system (experiment 4c), or (iii) both regeneration systems (experiment 4d) only 3HB-CoA (0.1 to 0.21 mM) was produced; no acetyl-CoA or only traces (0.03 mM) of acetyl-CoA could be detected. Apparently, high NADH/NAD⁺ or NADPH/NADP⁺ ratios prevented oxidation of the 3HB-CoA formed.

The experiments described above clearly show that isolated nPHB granules can cleave PHB to 3HB-CoA via thiolysis. To identify the enzyme responsible for the thiolysis reaction, we performed experiments with nPHB granules isolated from recombinant E. coli harboring different combinations of pha genes (Table 1). Isolated nPHB granules of strains harboring phaCAB of R. eutropha as the only PHB-related genes did not produce levels of 3HB-CoA in the presence of added CoA that were significantly above the detection limit (Table 2, experiment 5). Thiolysis did also not occur when phaP1 (encoding phasin [29]) or phaZa1 (encoding PHB depolymerase [9, 32, 47]) of R. eutropha was present in E. coli (experiments 6 and 7). Interestingly, when both the phaP1 and phaZa1 genes were present in a phaCAB background, nPHB granules of this strain produced 3HB-CoA (0.21 mM) in the presence of CoA (experiment 8). Significant amounts of acetyl-CoA were not detected when nPHB granules of recombinant E. coli were used, indicating that downstream mobilizing enzymes were absent. We concluded that E. coli nPHB granules are able to perform the thiolysis reaction in the presence of PhaP1 and PhaZa1. However, when the experiment was performed with nPHB granules that had been isolated from an R. eutropha phaP1 deletion mutant, 3HB-CoA (0.13 mM) was detected (experiment 9). Presumably, other phasin proteins of R. eutropha (19, 29) can replace the function of PhaP1. Thiolysis with nPHB granules from an R. eutropha ΔphaZa1 strain (experiment 10) resulted in production of slightly reduced but still significant amounts of 3HB-CoA, indicating that other PHB depolymerases are present in R. eutropha H16 (Fig. 2 and 3). To confirm that PhaZa1 and not PHB synthase is responsible for the thiolysis reaction, aPHB granules that contained no proteins were prepared and were incubated in the presence of CoA and PhaZa1 (experiment 11). Up to 0.21 mM 3HB-CoA was detected with these granules, confirming that PhaZa1 is able to catalyze thiolysis. When PhaZa1 was heated to 90°C for 10 min before the reaction was performed, no 3HB-CoA was formed (not shown).

PHB synthesis by nPHB granules from acetyl-CoA.

The experiment described above clearly showed that isolated nPHB granules are able to perform thiolysis of PHB and can catalyze oxidative cleavage of 3HB-CoA to acetyl-CoA. We were interested in whether nPHB granules could also catalyze the reverse reaction, synthesis of PHB from acetyl-CoA. Synthesis of PHB from 3HB-CoA catalyzed by PHB synthase has been repeatedly shown by other workers (4, 6, 35) and therefore was not examined by us. When isolated nPHB granules were incubated in the presence of acetyl-CoA and NAD(P)H, formation of free CoA and a decrease in the acetyl-CoA peak were determined by HPLC analysis (Table 3, experiments 1 and 2). This result indicated either that acetyl-CoA was cleaved to acetate and free CoA or that acetyl-CoA was converted to PHB via 3HB-CoA. The same results were obtained when nPHB granules of the R. eutropha H16 ΔphaZa1 mutant were used, indicating that PhaZa1 is not important for the reaction observed (Table 3, experiments 3 and 4). When the experiment was performed in the presence of an NADPH regeneration system, 3HB-CoA (0.17 mM) was detected in addition to free CoA and a reduced amount of acetyl-CoA (Table 3, experiment 5). The identity of 3HB-CoA was confirmed by HPLC-ESI-MS and determination of the corresponding m/z value, m/z 854. We also detected ions with the m/z value characteristic of acetoacetyl-CoA (m/z 852). These findings indicated that PHB could be synthesized from acetyl-CoA by catalysis with enzymes present in isolated nPHB granules. To exclude the possibility that acetyl-CoA is cleaved to free acetate and CoA, with the latter converted to 3HB-CoA via thiolysis, we repeated the experiment with ¹⁴C-labeled acetyl-CoA. After incubation of nPHB in the presence of [¹⁴C]acetyl-CoA and NADPH, the suspension was centrifuged and the radioactivity of the PHB-containing pellet was determined (Table 4). Significant label (33 to 40 μM) was present in the PHB fraction (experiments 1 and 3). Controls in which nPHB granules had been heat treated prior to incubation showed no label in the polymer fraction (Table 4, experiments 2 and 4). Addition of ¹⁴C-labeled acetyl-CoA to nPHB without incubation (time zero) did not result in detection of radioactivity in the PHB fraction, indicating that nonspecific binding of acetyl-CoA to nPHB granules was not significant. When NADPH was omitted, no radioactivity was detected in the polymer fraction, indicating that the reaction is NADPH dependent (Table 4, experiment 5). We concluded that isolated nPHB granules have the ability to catalyze all reactions leading from acetyl-CoA to PHB.

TABLE 3.

Assay of in vitro synthesis of PHB^a

Expt	Conditions	Concn of CoA (mM) at:		Concn of acetyl-CoA (mM) at:		Concn of 3HB-CoA (mM) at:
		0 h	0.5 h	0 h	0.5 h	0 h	0.5 h
1	H16 nPHB + acetyl-CoA + NADH	<0.03	0.31	0.69	0.24	<0.03	<0.03
2	H16 nPHB + acetyl-CoA + NADPH	<0.03	0.24	0.72	0.38	<0.03	<0.03
3	H16 ΔphaZa1 nPHB + acetyl-CoA + NADH	<0.03	0.46	0.69	0.22	<0.03	<0.03
4	H16 ΔphaZa1 nPHB + acetyl-CoA + NADPH	<0.03	0.38	0.93	0.29	<0.03	<0.03
5	H16 nPHB + acetyl-CoA + NADPH with regeneration	0.09	0.70	2.23	1.46	<0.03	0.17

^aIsolated nPHB granules were incubated at 30°C as described in Materials and Methods. At zero time and 30 min samples were taken, and the products formed were investigated by HPLC. Note that in experiment 5 the concentration of acetyl-CoA was doubled to 2 mM. The values were calculated from HPLC peak areas.

TABLE 4.

Synthesis of PHB from ¹⁴C-labeled acetyl-CoA

Expt	Conditions	Label in supernatant at:				Label in precipitate (PHB fraction) at:
		0 h		0.5 h		0 h		0.5 h
		Radioactivity (dpm)	Concn (μM)	Radioactivity (dpm)	Concn (μM)	Radioactivity (dpm)	Concn (μM)	Radioactivity (dpm)	Concn (μM)
1	H16 nPHB + [14C]acetyl-CoA + complete assay mixture	1,170,000	1,100	955,000	920	82	0.08	34,100	33
2	H16 nPHB (heat treated) + [14C]acetyl-CoA	1,170,000	1,100	979,000	950	70	0.07	176	0.17
3	Same as expt 1, complete assay mixture	1,280,000	1,200	1,180,000	1,100	109	0.11	41,100	40
4	Same as expt 2, complete assay mixture, nPHB heat treated	1,310,000	1,300	1,290,000	1,200	89	0.09	211	0.20
5	Same as expt 1, assay mixture without NADPH	1,310,000	1,300	1,330,000	1,300	72	0.07	214	0.21

^anPHB granules isolated from R. eutropha H16 were incubated in the presence of [¹⁴C]acetyl-CoA (total concentration, 2.17 mM; 37 kBq), 1 mM DTT,1 mM NADPH, and the cofactor regeneration system at 30°C for 30 min as described in Materials and Methods. The radioactivity of the soluble and insoluble fractions was determined. Experiments with heat-inactivated nPHB and without NADPH served as controls.

DISCUSSION

Biochemical investigation of intracellular mobilization of PHB was hindered in the past by an inability to detect the PHB depolymerase activity of isolated putative intracellular PHB depolymerases with the natural substrate (nPHB granules). A second reason probably was that researchers interested in intracellular mobilization of PHB were influenced by detailed knowledge concerning extracellular PHB depolymerases (PhaZs) (11, 13). Extracellular PhaZs generally have 3HB (or oligomers of 3HB) as end products, and presumably researchers, including us, expected that intracellular PHB depolymerases also release free acids as products. The finding that isolated nPHB granules have thiolytic activity and release 3HB-CoA suggests that researchers could have looked for the “wrong” activity in the past. Only recently, the presence of thiolytic activity was proposed for PHB synthase (44). However, the observed activity was rather low (about 0.01 mM 3HB-CoA). PhaZa1-catalyzed thiolysis of aPHB granules in the presence of free CoA produced ca. 0.2 mM 3HB-CoA (Table 2, experiment 11), and this confirmed that PhaZa1 is at least one of the important intracellular PhaZs and that 3HB-CoA is a primary mobilization product. We cannot yet judge the importance of the other putative intracellular PhaZs of R. eutropha, but we now have a suitable tool to test individual PHB depolymerases for activity. Investigation of the amino acid sequence of PhaZa1 and site-directed mutagenesis experiments showed that Cys183 is the active site in PhaZa1 (15). The presence of a thiol group is in agreement with the presence of thiolytic activity of PhaZa1. The thiol group is sensitive to oxidation, and therefore addition of reduced compounds (DTT) was necessary in our experiments to keep the nPHB granules in a reduced state. Since PhaZa1 is expressed intracellularly under reduced conditions, oxidation of the thiol group in vivo is unlikely. The finding that isolated nPHB granules release free 3HB at a rate of about 4 nmol min⁻¹ mg⁻¹ PHB in vitro may be explained by the fact that hydrolysis of an ester bond in an aqueous environment is highly exergonic and may occur as an in vitro reaction of PHB depolymerase and PHB synthase.

Another unexpected result of this study was the finding that isolated nPHB granules are able to catalyze the conversion of acetyl-CoA to PHB and vice versa. Synthesis of PHB from acetyl-CoA was unequivocally shown by the incorporation of ¹⁴C label from [¹⁴C]acetyl-CoA (Tables 3 and 4), and the mobilization reactions were demonstrated by formation of acetyl-CoA from nPHB in the presence of CoA (Table 2). Under appropriate conditions the same granules catalyzed formation of PHB from acetyl-CoA and degradation of PHB to acetyl-CoA. This means that all activities (thiolase, reductase, synthase, depolymerase) are present in the isolated nPHB granule fraction in vitro. We do not know whether the phaCAB and phaZa1 gene products alone are responsible for the formation of PHB and acetyl-CoA. The R. eutropha genome contains several dozen potential thiolase and reductase isologs (26). At least one ketothiolase has been found to be bound to isolated nPHB granules (29). Data for in vivo localization of enzymes on the surface of nPHB granules exist only for the PHB synthase PhaC (7, 25), for the PHB depolymerase PhaZa1 (Fig. 1), and for phasins and the phasin-related protein ApdA in Rhodospirillum rubrum (10). nPHB granules were isolated by layering a soluble crude extract onto the surface of a glycerol gradient. The PHB granules migrated into the glycerol fraction, while the soluble proteins remained in the surface layer of the gradient. However, we cannot exclude the possibility that some enzymes, such as a reductase and a thioloase, artificially bound to nPHB granules during preparation of the granules and that the finding that there were thiolase and reductase activities in the nPHB granule fraction was an experimental artifact. Lysozme, which is often added during cell lysis processes, has been found to be attached to isolated PHB granules (21). Therefore, it is necessary to investigate in vivo localization of thiolase and reductase enzymes. Unfortunately, the R. eutropha genome contains many thiolase and reductase isologs (26).

If nPHB granules harbor all proteins necessary for the formation and mobilization of the polymer, nPHB is more than just storage material. PHB granules could have at least three functions, (i) synthesis of PHA, (ii) storage of PHA, and (iii) mobilization of PHA. Consequently, PHB granules contain many proteins, including at least four phasin isoenzymes, PHB synthase, an acetoacetyl-CoA thiolase(s), an acetoacetyl-CoA reductase(s), a PHB depolymerase(s), and a regulatory protein, PhaR, which are known to be at least present on PHB granules (26-29, 48). We therefore consider PHA granules subcellular organelles rather than simple storage tanks.

A striking consequence of our findings is that the ability of the cells to simultaneously synthesize and degrade PHB (3, 43) does not lead to a futile cycle. If the primary product of PHB mobilization is 3HB-CoA instead of the free acid, there is no loss of energy, and the results of Doi et al. (3) and Taidi et al. (43) showing simultaneous synthesis and degradation of PHB can now be fully understood. Our experiments performed with different precursor molecules and cofactors suggest that the balance between net biosynthesis and net mobilization is controlled by the concentration of key metabolites. In the future, it will be necessary to perform in vitro experiments with nPHB granules and different combinations and different concentrations of putative key metabolites and to determine the formation of products. The in vitro values obtained should be compared with the concentrations of key metabolites determined with living cells cultured under PHB synthesis and PHB mobilization conditions.