Trapping of Intermediates with Substrate Analog HBOCoA in the Polymerizations Catalyzed by Class III Polyhydroxybutyrate (PHB) Synthase from Allochromatium Vinosum

Abstract

Polyhydroxybutyrate (PHB) synthases (PhaCs) catalyze the formation of biodegradable PHB polymers that are considered as an ideal alternative to petroleum-based plastics. To provide strong evidence for the preferred mechanistic model involving covalent and noncovalent intermediates, a substrate analog HBOCoA was synthesized chemoenzymatically. Substitution of sulfur in the native substrate HBCoA with an oxygen in HBOCoA enabled detection of (HB)_nOCoA (n = 2–6) intermediates when the polymerization was catalyzed by wild-type (wt-)PhaEC_Av at 5.84 hr⁻¹. This extremely slow rate is due to thermodynamically unfavorable steps that involve formation of enzyme-bound PHB species (thioesters) from corresponding CoA oxoesters. Synthesized standards (HB)_nOCoA (n = 2–3) were found to undergo both reacylation and hydrolysis catalyzed by the synthase. Distribution of the hydrolysis products highlights the importance of the penultimate ester group as previously suggested. Importantly, the reaction between primed synthase [³H]-sT-PhaEC_Av and HBOCoA yielded [³H]-sTet-O-CoA at a rate constant faster than 17.4 s⁻¹, which represents the first example that a substrate analog undergoes PHB chain elongation at a rate close to that of the native substrate (65.0 s⁻¹). Therefore, for the first time with a wt-synthase, strong evidence was obtained to support our favored PHB chain elongation model.

Polyhydroxyalkanoates (PHAs) are energy storage ploymers that are synthesized by a variety of bacteria under nutrient-limited conditions except for carbon.^{1, 2} Accumulation of PHA can reach up to 90% of dry cell weight.³ When the conditions become hospitable, these polymers are degraded to release energy for other biological processes.¹ PHAs have received much attention because they are considered as an ideal alternative to petroleum-based plastics that are nonbiodegradable.^{4, 5} Huge amounts of plastic waste are generated each year and a significant portion ends up in oceans, which will eventually destroy the marine ecosystem and in turn our planet if nothing is done to prevent it.

In contrast, PHAs are biodegradable and can be produced from renewable sources.^{1, 2} Approximately 150 different monomers are found to be incorporated into PHAs that can be developed into various applications such as packaging, drug delivery, and tissue engineering.^{2, 6} However, progress in PHA commercialization has been limited due to high cost of production.^7-9 Thus, study of PHA biosynthesis will facilitate us to engineer relevant enzymes so that PHAs can be produced as economically as the petroleum-based plastics.

As depicted in Scheme 1, PHAs are generated from 3-hydroxyalkanoate (HA) coenzyme A (CoA) catalyzed by PHA synthases, PhaCs.^{10, 11} Depending on the size of the R group, the resulting polymers can have properties ranging from thermoplastics to elastomers. When R is a methyl group, the polymer is named polyhydroxybutyrates (PHBs). It is well known that PhaCs are crucial enzyme in PHA biosynthesis as they determine polymer productivity as well as the properties such as composition, length, and polydispersity.^{6, 12, 13}

Scheme 1.

PHA synthesis by PhaCs

Bacterial PhaCs are divided into four classes according to their subunit composition and substrate specificity. ^10-12 The synthase from Allochromatium Vinosum, PhaEC_Av represents a prototypical class III PHB synthase. It consists of two subunits PhaC and PhaE, each having a molecular weight (MW) at 39 and 40K Da, respectively.¹⁴ No homology has been found for PhaE in any other proteins. However, its presence is required in order for PhaC to catalyze polymerization reactions.¹⁵

Two mechanistic models have been proposed for PHB chain elongation as depicted in Scheme 2.^{10, 11} Model A is based on the analogy to fatty acid biosynthesis,¹⁶ in which there is one PHB chain per dimer of PhaC and the growing chain alternates between the two monomers as each additional 3-hydroxybutyrate (HB) unit is added. According to this model, the active site has to be close to the surface in order to facilitate chain transfer. However, sequence alignment involving all four classes of PhaCs suggests that they have strong homology with the α/β-hydrolase superfamily and particularly lipases.^{17, 18} Available crystal structures of lipases have indicated that the active site is deeply buried.¹⁹ Therefore, model B based on the polyketide synthases is proposed.²⁰ For both models, the triad consisting of Cys, His, and Asp is catalyzing the polymerization reactions, in which Asp activates the 3-OH of a second HBCoA for ester bond formation. However, the difference between the two models is the way the second HBCoA is loaded to the synthases. While model A involves a covalent catalysis with the second HBCoA, model B requires noncovalent catalysis with the second HBCoA as well as the generation of a noncovalent chain-extension intermediate. Therefore, detection of the noncovalent intermediates (boxed in Scheme 2B) will provide an effective way to distinguish between these two models.

Scheme 2.

Mechanistic models of PHB chain elongation

Differentiation of chain elongation modes has significant implications in polymer length control. Such efforts have been reported for revealing the MW control of galactans.²¹ Processive and distributive mechanisms are commonly referred in template-independent polymerizations.²¹ Due to the processive character in model A, the resulting polymers are predicted to have low polydispersity index (PDI). In contrast, if model B is operative, the formed PHB-CoA after each catalytic addition may be released into the solution upon CoA removal. Thus, the polymers display a Poisson distribution with high PDI. In addition, the MW of PHB produced in model A is expected to be higher than that in model B. By comparing the prediction and experimental results, it can be deduced whether PHB formation requires other partners for polymer length control in addition to PhaCs.

Our previous study with a class III active site mutant has demonstrated that model B is preferred for PHB chain elongation.²² However, a distinction between these models is not yet possible with the wild-type (wt) synthase. Previous attempts to trap the intermediates with wt-PhaEC_Av were unsuccessful even with a rapid flow quench (RFQ) apparatus.²² Thus, we had to design a substrate analog that could slow down or arrest the chain elongation to facilitate the detection. This strategy led to the preparation of a methylene analog, HBCH₂CoA, in which the sulfur atom in HBCoA was replaced with a methylene group.²³ However, contrary to our expectations that the methylene analog would elongate one HB unit after incorporation, it instead participates in chain termination.²³

To further pursue our efforts in testing the preferred model with the wt-enzyme, here we report chemoenzymatic synthesis of an analog HBOCoA 1, in which the S in HBCoA is substituted with an O, and its applications as a mechanistic probe to trap the intermediates during PHB polymerization reactions.

RESULTS AND DISCUSSION

Design of HBOCoA as a Mechanistic Probe

Based on our previous successful studies with C149S-PhaEC_Av,²² as well as our hypothesis that the oxygen is a much closer analog to sulfur than the methylene group (CH₂), it is anticipated that the oxo analog, HBOCoA 1 may serve as an informative mechanistic probe to study PHB chain elongation. As described in Scheme 3A and if the preferred model B (Scheme 2) is operative, the rate of HBOCoA polymerization is predicted to be very slow because it involves enzymatic (re)acylation steps that are thermodynamically unfavorable (from an oxoester to thioester). The reduced rate will greatly enhance the possibility to detect noncovalent intermediates such as dimer (HB)₂OCoA 2a. Moreover, when HBOCoA reacts with a radiolabeled primed synthase [³H]-sT-PhaEC_Av as depicted in Scheme 3B, it is expected that the [³H]-sT chain will be extended by one HB unit at a rate similar to the native substrate to give a radiolabeled tetramer [³H]-sTet-O-CoA 3b. Additional elongation will be eliminated on the short-time scale due to subsequent reacylation step that is thermodynamically unfavorable.

Scheme 3.

Strategies of using HBOCoA as a probe to test the preferred mechanistic model B

Chemoenzymatic Synthesis of HBOCoA and Related Analogs

The approach to prepare HBOCoA 1 and its related analogs 2–4 is described in Scheme 4. Starting with p-methoxyphenyl (PMP) protected pantothenic acid 5,²⁴ condensation with 2-aminoethanol in the presence of ethyl chloroformate gave the amide 6 in a 76% yield.²⁵ The key step is formation of ester bond between the terminal hydroxyl group in 6 and carboxylic group in 7. Depending on the structure of acid 7, this step was accomplished in the presence of different coupling reagents. When the benzyl (Bn) protected monomeric HB acid 7a was employed, N, N′-dicyclohexylcarbodiimide (DCC) and catalytic 4-dimethylaminopyridine (DMAP) were used to promote the ester formation 8a in a 85% yield. However, combination of DCC and DMAP resulted in low yields for other acids (7b-e). Thus, they were first converted to acyl chloride in the presence of oxalyl chloride and catalytic amount of N, N′-dimethylformamide (DMF) followed by nucleophilic acyl substitution with 6 to afford esters 8b-e in good yields. Synthesis of 7a–d has been reported elsewhere.¹⁷ The protecting groups in 8 and 6 were removed by hydrogenation in excellent yields to give enzymatic precursors 9 and 10, respectively.

Scheme 4.

Chemoenzymatic synthesis of HBOCoA and related analogs

Introduction of adenosine 3′,5′-triphosphate was accomplished with three enzymes involved in CoA biosynthesis: a pantothenate kinase from Staphylococcus aureus (SaPanK),²⁶ a phosphopantetheine adenylyltransferase (EcCoaD),²⁷ and a dephospho-CoA kinase (EcCoaE) from Escherichia coli.²⁸ Various CoA analogs have been prepared by this method and employed as inhibitors or mechanistic probes to investigate CoA-dependent enzymes.^{23, 25, 29-33} The final compounds were isolated by reverse-phase HPLC in reasonable yields and characterized by ¹H, ¹³C, and ³¹P NMR and HRMS.

It should be noted that since the pantothenamide 10 contains two terminal hydroxyl groups, concerns of regioselectivity were raised for the enzymatic conversion. Although CoAOH 4 was previously synthesized from a pre-CoA thioester,³⁴ only a low resolution MS was reported for its characterization. To make sure phosphate installed at the correct hydroxyl group by SaPanK, both the amides 6 and 10 were subject to phosphorylation. It was found that no reaction occurred with 6 while 10 was converted to phosphopantothenamide according to HPLC profile (data not shown). Thus, demonstrating that the phosphorylation occurred at the desired hydroxyl group. The identity of 4 was fully characterized and is reported in Supporting Information.

Detection of (HB)_n CoA (n=2–6) Intermediates in the Reaction between HBOCoA and wt-PhaEC_Av

To look for (HB)_nOCoA intermediates, wt-PhaEC_Av and HBOCoA 1 were incubated at 30 °C and the reaction was monitored by HPLC at 260 nm (Figure 1A). A decrease in the amount of HBOCoA (retention time t_R = 17.7 min) was observed. Additionally, three new peaks at 9.80, 29.0, and 37.7 min appeared. UV-Vis scan of these peaks revealed that they all had a maximal absorbance at 260 nm.

Figure 1.

(A) Formation of (HB)_nCoA intermediates during incubation of 100 μM wt-PhaEC_Av with 20.0 mM HBOCoA at 30 °C for 1 (black), 2 (red), 3 (green), 4 (blue), 5 (orange), and 6 hrs (pink). HPLC condition is described in EXPERIMENTAL SECTION; (B) MALDI-TOF MS of species eluted at 9.80 (I), 29.0 (II), 37.7 (III), and 40.0–60.0 min (IV); (C) Assignment of peaks shown in B; (D) Conversion of the terminal HB unit into a crotonate in the presence of acid.

The new species were isolated, concentrated, and characterized by low-resolution MALDITOF MS in a negative reflective mode. The results are shown in Figure 1B I–III. The species eluted at 9.80, 29.0, and 37.7 min (Figure 1A) were identified as CoAOH 4, (HB)₂OCoA 2a, and (HB)₃OCoA 2b, respectively. Their MS are identical to the synthesized standards (data not shown). The assignment of the mass peaks is summarized in Figure 1C. Further confirmation arises from their respective co-elution with the synthesized standards.

Since PHB is highly hydrophobic, it is predicted that CoA analogs attached with a large number of HB units will be eluted at high percentages of methanol. Examination of HPLC profiles at 260 nm did not find apparent peaks between 40.0 and 60.0 min. Nevertheless, the eluents at 40.0–60.0 min were collected, concentrated, and analyzed by MALDI-TOF MS. The result is shown in Figure 1B IV with the peak assignment summarized in Figure 1C. CoA analogs containing up to hexameric HB units were identified. Additionally, peaks corresponding to H₂O elimination were observed in MS, suggesting that the terminal HB unit had been converted into a crotonate. Because MALDI-TOF MS uses soft ionization, elimination was not expected. Thus, formation of crotonate was attributed to the presence of acidic matrix during sample preparation as shown in Figure 1D. The terminal hydroxyl group is protonated by matrix α-cyano-4-hydroxycinnamic acid (CHCA). A subsequent E2 elimination will generate the analog with an end crotonate unit.

Determination of Kinetic Parameters of HBOCoA and Rates of the Formation of Intermediates (HB)_nOCoA (n = 2–3)

Since the analog HBOCoA was recognized as a substrate, its kinetic parameters were determined with wt-PhaEC_Av. The analog underwent monophasic release of CoAOH with a K_M and k_cat at 2.38 mM and 5.84 hr⁻¹, respectively (Figure 2A). This translates into catalytic efficiency (k_cat/K_M ) of 0.68 M⁻¹s⁻¹, which is 7.35×10⁵-fold slower than for the native substrate HBCoA (k_cat/K_M = 5.00×10⁵ M⁻¹s⁻¹). Furthermore, the turnover number of HBOCoA is about 4.00×10⁵-fold slower than that of HBCoA. The dramatically lowered catalytic efficacy and extremely slow rate are consistent with the prediction that the reaction between the Cys and (HB)_nOCoA (an oxoester) to produce an enzyme-bound PHB (a thioester) is thermodynamically unfavorable (Scheme 3A).

Figure 2.

(A) Kinetic analysis of HBOCoA with 100 μM wt-PhaEC_Av. The data were fitted to Michaelis–Menten equation to give a K_M = 2.38 ± 0.23 mM and V_max = 9.73 ± 0.28 μM•min⁻¹. The reaction was performed in duplicate. (B) Rates of formation of (HB)₂OCoA (solid line) and (HB)₃OCoA (dash line) in the presence of 100 μM wt-PhaEC_Av and 20 mM HBOCoA. Linear regression of the data gave the slopes of 3.47 × 10⁻¹ and 3.51 × 10⁻² mM•hr-1 to (HB)₂OCoA and (HB)₃OCoA, respectively. The reaction was performed in duplicate.

Kinetic competence implies that the rates of formation and breakdown of the intermediates must be equal to or faster than the overall rate of the reaction (k_cat). To examine that, the rates of formation of (HB)₂OCoA and (HB)₃OCoA were measured under the condition that the synthase was saturated with the substrate (8K_M). As shown in Figure 2B, (HB)₂OCoA and (HB)₃OCoA were formed at a rate constant of 3.47 and 0.35 hr⁻¹, respectively. Thus, the overall rate of intermediate formation is 7.99 hr⁻¹ (3.47×2+0.35×3=7.99 hr⁻¹) plus the rate of formation of PHB attached to the enzyme. The latter is unknown due to lack of the radioactivity. Nevertheless, the rate of formation of intermediates (>7.99 hr⁻¹) is indeed faster than the overall rate of the reaction (k_cat=5.84 hr⁻¹), which suggests that these intermediates could be kinetically competent.

Products from Incubation of (HB)_nOCoA (n = 2–3) with wt-PhaEC_Av

As described above, kinetic competence of an intermediate also requires that the rate of its utilization is greater than or equals to the rate of substrate consumption. Therefore, efforts were made to measure the reacylation rates of (HB)_nOCoA (n=2–3) with wt-PhaEC_Av. It has been reported that (HB)₂CoA and (HB)₃CoA can function as primers to PHB synthases by covalently attaching (HB)_n (n=2–3) chain to the active site Cys with the concurrent release of CoA.³⁵ However, when the oxo analogs (HB)_nOCoA (n=2–3) were incubated with wt- PhaEC_Av followed by HPLC analysis, new species were found in addition to the expected release of CoAOH (t_R = 9.80 min). As shown in Figure 3A, incubation with dimer (HB)₂OCoA (solid line, t_R = 29.0 min) resulted in an additional species that corresponded to HBOCoA (t_R = 17.7 min). Similarly, when trimer (HB)3OCoA (dash line, t_R = 37.7 min) was employed as the substrate, both HBOCoA (t_R = 17.7 min) and (HB)₂OCoA (t_R = 29.0 min) were identified along with CoAOH (t_R = 17.7 min). Characterization of these species was confirmed with MALDI-TOF MS and co-elution with authentic standards. Their rates of formation are summarized in Figure 3B.

Figure 3.

(A) HPLC profiles of incubating 50.0 μM wt-PhaEC_Av with 5.00 mM (HB)₂OCoA (solid line) or (HB)₃OCoA (dash line) for 30 min. The t_R of CoAOH, HBOCoA, (HB)₂OCoA, and (HB)₃OCoA are 9.80, 17.7, 29.0, and 37.7 min, respectively; (B) Estimated rates of product formation in hr⁻¹; (C) Proposed pathways for product formation from incubation of (HB)_nOCoA (n = 2–3) with wt-PhaEC_Av.

Formation of these products is postulated in Figure 3C. Acylation of the synthase by (HB)_nOCoA (n=2–3) will release the expected CoAOH (red arrows). Their rates were estimated at ~2-fold faster than the overall rate of the reaction (5.84 hr⁻¹), which suggests that (HB)_nOCoA (n=2–3) could serve as kinetically competent interme diates during polymerization reactions. Additionally, the rate of acylation of (HB)₃OCoA is slightly faster than that of (HB)₂OCoA, which is consistent with previous observations that rate of reacylation increases as the chain elongates.^{22, 36}

Products other than CoAOH generated during incubation are attributed to enzyme-catalyzed hydrolysis. It is known that PHB synthases are hydrolytic.¹⁷ As shown in Figure 3A, the major hydrolysis product is HBOCoA though the starting materials were different. This means the water preferentially attacks the pneultimate HB unit (black solid arrows in Figure 3C) independent of primer length, which coincidentally agrees with the hydrolysis studies performed recently with [³H]-sT-PhaEC_Av (structure shown in Scheme 3B).^{23, 37} It was discovered that the pneultimate HB unit of the sT chain attached to a synthase is preferentially attacked by a nucleophile.^{23, 37} However, the rate of hydrolysis of sT chain is ~2-5-fold slower than that of (HB)_nOCoA (n=2–3), suggesting that the synthase-bound PHB chain is oriented for elongation, the native function of the enzyme.

Rapid Formation of Elongated Products with Radiolabeled Primed Synthase [³H]-sT-PhaEC_Av

It is known that the rate of PHB chain elongation is much faster than that of initiation.³⁸ This poses a challenge to detect noncovalently bound intermediates due to heterogeneous formation of multiple species in undetectable amounts. In order to overcome this problem, artificial primers such as saturated trimeric-CoA (sTCoA) were developed to uniformly load the synthases so that the elongation process could be examined.¹⁵

Such a strategy has recently been employed with the analog HBCH₂CoA to investigate PHB chain elongation. However, the attempt was unsuccessful, possibly due to misalignment of the substrate binding.²³ In the present study, intermediates (HB)_nOCoA (n=2-6) were trapped during polymerization, suggesting that the binding mode of HBOCoA is similar to the native substrate and might overcome the misalignment associated with HBCH₂CoA in the reaction with a primed synthase. As described in Scheme 3B and since k_cat of the native substrate is 65.0 s⁻¹,³⁹ it is predicted that the proposed reaction will be rapid for the first turnover as it does not involve formation of enzyme-bound PHB that is thermadynamically unfavorable.

Thus, the wt-PhaEC_Av was acylated with radiolabeled [³H]-sTCoA and subsequently purified by Sephadex G-50 size-exclusion chromatography (SEC) to remove excess [³H]-sTCoA. The primed synthase was immediately reacted with HBOCoA. The reaction was terminated at selected times using RFQ. After the protein precipitate was separated from the mixture, the supernatant was neutralized, concentrated, and analyzed by HPLC and scintillation counting. The profiles of elution at 260 nm and radioactivity are shown in Figure 4A.

Figure 4.

(A) Typical HPLC (solid line) and radioactivity (bar) profiles of the reaction between 4.20 μM [³H]-sT-PhaEC_Av and 5.00 mM HBOCoA stopped by RQF. The dash line represents HPLC profile of synthesized standards, sT-O-CoA (44.1 min) and sTet-O-CoA (47.4 min); (B) MALDI-TOF MS of species eluted at 47.0-49.0 min. (C) Rate of formation of [³H]-sTet-O-CoA in the reaction between 4.20 μM [³H]-sT-PhaEC_Av and 5.00 mM HBOCoA. The data were fitted to y=a(1-e-^bx), where a = burst amplitude and b = k_obs. The fit gives an amplitude of 139.7±2.2 cpm (counts per minute) and a k_obs of 17.4±0.8 s⁻¹.

Indeed, the results were consistent with our prediction. The eluents at 47.0–49.0 min contained the radioactivity (bar in Figure 4A) and had maximal absorbance at 260 nm (solid line in Figure 4A). Mass analysis by MALDI-TOF under a negative reflective mode showed a value of 1078.818 (Figure 4B), which corresponds to the [M–H]⁻ ions (calcd. 1078.283) of predicted elongation product [³H]-sTet-O-CoA 3b (Scheme 3B). This species also eluted at the same t_R (47.5 min) as the synthesized authentic compound (dash line in Figure 4A).

Moreover, it has to be pointed out that sT-O-CoA 3a (Scheme 4) was also prepared as a standard, which has a t_R at 44.1 min (dash line in Figure 4A). Recent study with HBCH₂CoA has revealed that the oxoester group that is next to thioester bond can be attacked by a nucleophile.²³ To determine whether compound 3a was produced, eluents at 43.5-45.5 min were collected, concentrated, and analyzed by MALDI-TOF MS and scintillation counting. The results suggested that sT-O-CoA was not generated during the reaction.

The rate of formation of [³H]-sTet-O-CoA was measured for the reaction between 32.0 μM primed synthase and 5.00 mM HBOCoA. It was observed that only 0.13 equiv of the synthase (~4.20 μM) was attached with a radiolabelled [³H]-sT chain ([³H]-sT-PhaEC_Av) after SEC purification. This agrees with previous established study that 0.15 equiv per PhaEC_AV was radiolabelled.¹⁵ The reaction was stopped by perchloric acid from 10 to 300 ms. The total recovery of radioactivity was more than 85% after HPLC separation. A typical radioactivity profile is shown as the bars in Figure 4A. The data were fitted to a single exponential rise to the maximum and gave a k_obs of 17.4 s⁻¹ (Figure 4C). Due to limited amount of [³H]-sTCoA and HBOCoA, the experiment was only performed once. Although the obtained 17.4 s⁻¹ is less than the k_cat of native substrate (65.0 s⁻¹⁾, it could be due to the fact that the enzyme was not saturated with the substrate analog. Thus, for the first time, we have designed a mechanistic probe that can be used with a primed wt-synthase in a kinetically competent fashion to interrogate PHB chain elongation.

CONCLUSION

This study demonstrates a new way to study the challenging PHB polymerization processes using the substrate analog HBOCoA. Incubation of the analog with wt-PhaEC_Av resulted in detection of multiple intermediates (HB)_nOCoA (n=2-6) that have been proven chemically and kinetically competent. This represents the first example of employing a wt-synthase to provide direct evidence to support the preferred model involving covalent and noncovalent intermediates. The rate of polymerization with HBOCoA is extremely slow (5.84 hr⁻¹) due to a step that is thermodynamically unfavorable (from oxoester to thioester).

Moreover, reaction between wt-PhaEC_Av and (HB)_nOCoA (n=2-3) yielded hydrolysis products in addition to the expected release of CoAOH from reacylation. The distribution suggests that not only are the hydrolysis products independent of the primers being used, but also the penultimate HB unit (from the side of CoA moiety) is susceptible to synthase-catalyzed hydrolysis. The latter observation coincidentally agrees with previous studies done by us and others showing that the penultimate HB of a sT chain attached to synthases is preferentially attacked by a nucleophile. ^{23, 37}

Reactions between the primed synthase [³H]-sT-PhaEC_Av and HBOCoA formed the elongated product [³H]-sTet-O-CoA at a rate of at least 17.4 s⁻¹. This represents the first example that a wt-synthase catalyzes the polymerization with an analog at the rate close to that of the native substrate. To the best of our knowledge, it is by far the most decisive evidence to support our favored model shown in Scheme 2.

Modes of PHB polymerizations will facilitate the determination of other factors in addition to PhaCs that are required for controlling polymer length and polydispersity. The distributive character associated with our preferred chain elongation model suggests that PHB should have a high PDI. Although no report was found that addressed the polymer properties obtained from native A. vinosum, studies with other PHB-producing organisms suggest that the polydispersity of PHB is relatively low.^40-44 Thus, our results presented here provide additional evidence to support the hypothesis that PHB polymer length and polydispersity are controlled by multiple factors in addition to PhaCs.^45-47 Such factors could include phasin proteins PhaP,⁴⁸ recently identified membrane-associated protein PhaM,⁴⁹ and others yet to be identified. Understanding their relationships to PhaCs is essential in order to engineer polymer properties as well as to lower the cost of production.

METHODS

Chemoenzymatic Synthesis of HBOCoA 1 and Related Analogs

The chemoenzymatic approaches to synthesize HBOCoA 1 and related analogs are shown in Schemes 3. The experimental details are summarized in Supporting Information.

General HPLC Conditions

The HPLC was performed on a Waters Breeze 2 system equipped with a 1525 pump and 2998 PDA detector. Except for compound purification, all HPLC were performed on a Luna C18-2 column (5 μm, 4.6×250 mm) that was eluted with 40.0 mM NH₄OAc (pH 4.70) and MeOH at 1.0 mL/min using a linear gradient (1) 5-30% MeOH in 0-30 min; and (2) 30-90% MeOH in 30-60 min.

Reaction of HBOCoA with wt-PhaEC_Av and Analysis of Small Molecules by HPLC and MALDI–TOF

In a final volume of 100 μL, 100 μM wt-PhaEC_Av in 200 mM KPi (pH 7.80) was incubated with 20.0 mM HBOCoA at 30 °C. Aliquots of 13.0 μL were withdrawn at selected time intervals (1.0 to 6.0 hr) and the enzyme was deactivated by heat treatment at 95 °C for 5 min. At the end of the time course, the reaction mixture was centrifuged and the supernatant was removed and placed on ice. The precipitated protein was washed with ddH₂O (100 μL ×2). The pooled washes were combined with the supernatant and then filtered using a Whatman microsyringe filter. The filtrate was subject to HPLC. Each peak was lyophilized to dryness, re-dissolved in minimal ddH₂O, and desalted by a Zip Tip. The concentrated peak was analyzed by MALDI–TOF MS in a negative mode using α-cyano-4-hydroxycinnamic acid (CHCA) as the matrix.

Determination of Kinetic Parameters of HBOCoA and Rate of the Formation of Intermediate (HB)_nOCoA (n=2-3)

Reactions and HPLC analysis were conducted in the same way as described above except that adenosine 5′-triphosphate (ATP) was added as an internal standard. The HBOCoA concentration varied from 0.50 to 20.0 mM and each concentration was performed in duplicate. The kinetic parameters were determined by fitting the data to Michaelis-Menten equation 1 using SigmaPlot. Rate of the formation of intermediate (HB)_nOCoA (n=2-3) was determined when the enzyme was saturated with the substrate ([HBOCoA] = 20.0 mM).

$$v=\frac{v_{\max }\left[\mathrm{S}\right]}{{\mathrm{K}}_{\mathrm{M}}+\left[\mathrm{S}\right]}$$ (1)

Incubation of (HB)_nOCoA (n=2–3) with wt-PhaEC_Av

In a final volume of 150 μL, 50.0 μM wt-PhaEC_Av in 200 mM KPi (pH 7.80) was incubated with 5.00 mM (HB)_nOCoA (n=2-3) at 30 °C. Aliquots of 13.0 μL were removed at different time points (10 min to 3.0 hrs) and the enzyme was deactivated by heat treatment at 95 °C for 5 min. Each sample was added ATP as an internal standard and then processed and analyzed by HPLC and MALDI-TOF MS as described above. The rate of formation of each peak was determined by initial slope of the reaction.

Reaction of HBOCoA with Primed Synthase [³H]-sT-PhaEC_Av and Analysis of Small Molecules by HPLC and MALDI–TOF MS

In a final volume of 50 μL, 1.30 mM wt-PhaEC_Av and 13.0 mM [³H]-sT-CoA (832 cpm/nmol) in 200 mM KPi (pH 7.80) were incubated at 25 °C for 3 min before loading the mixture onto a Sephadex G-50 column (5 × 80 mm) pre-equilibrated with 200 mM KPi (pH 7.80) at 4 °C. The acylated protein [³H]-sT-PhaEC_Av was isolated in 90% yield and then immediately reacted with 5.00 mM HBOCoA. The reaction was performed with KinTeK RQF-3 at room temperature. The final concentration of synthase was 32.0 μM, of which 0.13 equiv was attached with a [³H]-sT chain (4.20 μM). At defined time points (10-300 ms), the reaction was quenched with 2% HClO₄. Each sample was then centrifuged and the supernatant was removed and placed on ice. The precipitated protein was washed with ddH₂O (75 μL ×3), redissolved in 100 μL 20% SDS, and analyzed by scintillation counting. The pooled washes were combined with supernatant. The pH of the solution was adjusted to ~6.0 using pre-determined amount of chilled 2M NaOH. The solution was then centrifuged at 4 °C to remove any insoluble particles. Aliquots of 410 μL were injected onto HPLC column. Fractions of 1.00 mL were collected and each fraction was analyzed by scintillation counting. The total recovery yields of radioactivity (including supernatants and protein precipitates) ranged from 85% to 90%.

Eluents at 43.5-45.5 min and 47.0-49.0 min were combined. They were processed and analyzed by MALDI-TOF MS in the same way as described above.

The observed rate of formation of [³H]-sTet-O-CoA was determined by fitting the data to exponential rise to maximum y=a(1–e^–bx) using SigmaPlot.