Two Polyhydroxyalkanoate Synthases from Distinct Classes from the Aromatic Degrader Cupriavidus pinatubonensis JMP134 Exhibit the Same Substrate Preference

Cupriavidus pinatubonensis JMP134 utilizes a variety of aromatic substrates as sole carbon sources, including meta-nitrophenol (MNP). Two polyhydroxyalkanoate (PHA) synthase genes, phaC1 and phaC2, were annotated and categorized as class I and class II PHA synthase genes, respectively. In this study, both His-tagged purified PhaC1 and PhaC2 were shown to exhibit typical class I PHA synthase substrate specificity to make short-chain-length (SCL) PHA from 3-hydroxybutyryl-CoA and failed to make medium-chain-length (MCL) PHA from 3-hydroxyoctanoyl-CoA. The phaC1 or phaC2 deletion strain could also produce SCL PHA when grown in fructose or octanoate, but the double mutant of phaC1 and phaC2 lost this ability. The PhaC2 also exhibited substrate preference towards SCL substrates when expressed in Pseudomonas aeruginosa PAO1 phaC mutant strain. On the other hand, the transcriptional level of phaC1 was 70-fold higher than that of phaC2 in MNP-grown cells, but 240-fold lower in octanoate-grown cells. Further study demonstrated that only phaC1 was involved in PHA synthesis in MNP-grown cells. These findings suggested that phaC1 and phaC2 genes were differentially regulated under different growth conditions in this strain. Within the phaC2-containing gene cluster, a single copy of PHA synthase gene was present clustering with genes encoding enzymes in the biosynthesis of PHA precursors. This is markedly different from the genetic organization of all other previously reported class II PHA synthase gene clusters and this cluster likely comes from a distinct evolutionary path.


Introduction
Polyhydroxyalkanoate (PHA) is a class of biopolymer that is produced by a variety of microorganisms as carbon and energy storage components. It is usually synthesized when the carbon source is available in excess, and growth is limited by a lack of other essential nutrients [1,2].
Based on the chain-length of the monomers for PHA synthesis, PHAs are typically categorized into two major types: short-chain-length (SCL) polymers (C3~C5 monomers) and mediumchain-length (MCL) polymers (C6 and longer) [3,4]. The monomer types of PHA depend on the carbon sources, their metabolic routes and PHA synthases classes [1]. During PHA biosynthesis, PHA synthases are the key enzymes catalyzing the polymerization of hydroxyacyl-coenzyme A (CoA) thioesters concomitantly with the release of free CoA [3,4]. Based on their substrate specificities and subunit compositions, PHA synthases have been categorized into four classes [3]. Class I and class II PHA synthases consist of a single type of subunit, while class III and IV synthases have two types of subunits. Class I, III and IV PHA synthases prefer substrates that are SCL hydroxyalkanoic acid CoA thioesters (C3~C5). In contrast, class II PHA synthases prefer MCL substrates (!C6). Of the bacteria capable of producing PHA, some have more than one PHA synthase gene. These are usually homologs and belong to the same class of PHA synthases as reviewed by Bernd H. Rehm. [3]. Nevertheless, there are a few reported cases of bacterial strains carrying multiple functional PHA synthase homologs that come from different PHA synthase classes.
Cupriavidus pinatubonensis JMP134 (DSM4058, formerly known as Cupriavidus necator JMP134) is a versatile aromatic compounds degrader isolated from polluted environments. It has been deduced to contain two putative PHA synthase genes: phaC1 (Reut_A1347) and phaC2 (Reut_A2138) [5]. It has been predicted that phaC1 encodes a class I PHA synthase, while phaC2 encodes a class II PHA synthase [5]. However, the function of these two putative PHA synthase genes remains to be elucidated. On the other hand, strain JMP134 is capable of utilizing over 60 aromatic compounds as a sole carbon and energy source for growth, including the environmental pollutant meta-nitrophenol (MNP) [6][7][8]. Microbial MNP degradation has been given a great deal of attention in C. pinatubonensis JMP134. Its reductive catabolic pathway has been clearly illustrated at molecular and biochemical levels by Schenzle et al. [6] and our lab [9,10]. However, the relationship between the aromatic degradation and PHA production has not yet been investigated in this strain, or any other aromatic utilizers. Research experience with the MNP degradation by strain JMP134 has naturally led us to explore the possibility of its PHA production when utilizing MNP as the carbon source.
In this study, we report that the phaC1 and phaC2 genes code for two functional PHA synthases in C. pinatubonensis JMP134 using genetic and biochemical approaches. Despite being classified into two classes based on their protein sequences, both these two PHA synthases exhibited activities towards SCL but not MCL substrate in vitro. Further, PHA can be produced when MNP was the sole carbon source in vivo, but only in the presence of phaC1. This not only provides us with a better understanding of the genetic and biochemical diversity of PHA synthesis in bacteria, but also reveals the link between PHA production and MNP degradation.

Bacterial strains and culture conditions
Bacterial strains and plasmids used in this study are listed in Table 1. Unless noted otherwise, all Cupriavidus pinatubonensis and Pseudomonas aeruginosa strains were grown at 30°C and shaken at 200 rpm in minimal salt medium (MSM) described previously [11] supplemented with 0.5 mM meta-nitrophenol or 2% (w/v) fructose or octanoate as the sole carbon source. All E. coli strains were grown in Lysogeny Broth (LB) medium at 37°C and shaken at 200 rpm. Antibiotics were added to the media at the following concentrations (μgÁmL -1 ): chloramphenicol (Cm), 34; spectinomycin (Spec), 50; and tetracycline (Tc), 20.

Generation of phaC mutants and their complemented strains
All phaC mutants were generated by conjugation using a modified mobility plasmid for the transfer between E. coli WM3064 and C. pinatubonensis or Pseudomonas aeruginosa PAO1 as described previously [19]. The mobility plasmids pEX18Tc-phaC1cm, pEX18Tc-phaC2spec and pEx18Tc-phaC Pa cm were constructed by fusing PCR products of chloramphenicol resistance gene cat from pEX18Tc-cmgfp [16] or spectinomycin resistance gene aadA from pML300 [17] into the upstream and downstream fragments of targeting genes. These plasmids were transformed into E. coli WM3064 (2,6-diaminopimelic acid auxotroph) respectively before transferred into C. pinatubonensis or Pseudomonas aeruginosa PAO1 as described previously [19]. The double-crossover recombinants were screened on LB or MSM agar plates containing 0.5 mM MNP and 34 μg/mL chloramphenicol or 100 μg/mL spectinomycin to get desired mutants. The double mutant strain (JMP134ΔphaC1ΔphaC2) was constructed by deleting phaC2 in the C. pinatubonensis phaC1 mutant strain (JMP134ΔphaC1). The mutant strains were further confirmed by sequencing the PCR product generated from the modified region. All primers are listed in Table 2. Complementation experiments were performed by first cloning the complete coding sequences of phaC1 and phaC2 (amplified by PCR with primers phaC1For/phaC1Rev and phaC2For/phaC2Rev, respectively) into the broad host range vector pRK415 [18] via EcoR1 and Kpn1 restriction sites. The resulting plasmids were designated pRK415-phaC1 and pRK415-phaC2, respectively, and were transformed into E. coli WM3064 and conjugated with phaC1 and phaC2 mutant strains respectively to obtain the complemented strains JMP134ΔphaC1[pRK415-phaC1] and JMP134ΔphaC2[pRK415-phaC2]. For complementation of the double mutant strain JMP134ΔphaC1ΔphaC2, both phaC1 and phaC2 genes were amplified by PCR and cloned into pRK415 via EcoR1 and Kpn1 sites. The resulting plasmid was transformed into the double mutant strain. The phaC2 gene along with phaA and phaB from C. pinatubonensis JMP134 were amplified and cloned into pRK415, and the resulting plasmid was transformed into P. aeruginosa PAO1 phaC mutant strain.

Bioinformatics and in silico analyses of PHA synthases
Clustal Omega was used to perform multiple-sequence alignment of the reported PHA synthases of known or unconfirmed functions as well as PhaC1 and PhaC2 from C. pinatubonensis JMP134 [20]. The phylogenetic analysis was performed using the maximum likelihood method with bootstrap analysis (1000 iterations) as described by Quelas, J. I. et al [21]. The tree was constructed using MEGA (version 5.1) software [22]. The occurrence of conserved residues in the catalytic triad and the lipase box was checked manually.

Real-time quantitative PCR
The bacterial strains were grown in minimal medium with 0.5 mM MNP or 0.2% octanoate as the sole carbon source to the early stationary phase (approximately 12 hours after inoculation). Total RNA was isolated with an RNA prep pure bacterial kit (Tiangen Biotech, Beijing, China) and reversely transcribed into cDNA using a PrimeScript RT Reagent Kit with gDNA Eraser (Perfect Real Time) (TaKaRa, Dalian, China). Real-time quantitative PCR (RT-qPCR) was performed on a CFXConnect™ Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) using iQ SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) with primers specific to individual phaC genes or 16S rRNA gene listed in Table 2. The amount of target mRNA was normalized to that of 16S rRNA gene with the 2 -ΔΔC T method as described previously [19].

PHA monomer determination and quantification
The intracellular PHA contents and PHA composition were determined by gas chromatography-mass spectrometry (GC-MS) as described previously [23,24]. Briefly, bacterial cells cultured for 48 hours (late stationary phase) were harvested by centrifugation at 15,000 × g for 15 min, washed with phosphate buffered saline (PBS) (pH 7.4, 20 mM) and lyophilized overnight. Lyophilized bacterial cells (15 mg) were subjected to methanolysis in the presence of 3% H 2 SO 4 (v/v) at 95°C for 4 hours, washed with ddH 2 O and extracted with chloroform. The resulting methyl esters of the constituent 3-hydroxyalkanoic acids were diluted and analyzed by GC-MS according to the literature [23]. The GC analysis was performed by injecting 50 μl of each sample into a TRACE GC Ultra gas chromatography (Thermo Fisher scientific Inc., MA, USA) using a capillary column TRACE TR-1701 (0.25 mm × 30 meters, Thermo Fisher scientific Inc., MA, USA). The mass spectrometry was performed as described previously [25]. Table 2. Oligonucleotides used in this study. Oligonucleotides Polyhydroxybutyrate (PHB) (Sigma-Aldrich Co., St. Louis, MO, USA) or 3-hydroxyoctanoate (3-HO) (Sigma-Aldrich Co., St. Louis, MO, USA) was used as a standard, and benzoic acid was used as an internal standard for PHA quantification. Under these conditions, authentic PHB, 3HO, and benzoic acid had retention times of 6.02 min, 11.76 min, and 9.18 min, respectively. The methyl esters can be used to indicate methanolysized PHA and were identified with a NIST98 MS data library based on the comparison of both the GC retention time and the mass spectra with those of the authentic compounds.

Protein expression and purification
The PHA synthase genes phaC1 and phaC2 of strain JMP134 and phaC1 (PA5056) from P. aeruginosa PAO1 were amplified using primers ( Table 2) and then cloned into pET28a via NcoI and HindIII restriction sites to obtain pET-phaC1, pET-phaC2 and pET-phaC1 Pa , respectively. E. coli BL21(DE3) strains carrying the resulting expression constructs were grown at 37°C to an OD 600 of 0.4~0.6 and then induced with 0.1 mM isopropyl-D-thiogalactopyranoside (IPTG) at 30°C for 5 hours or at 16°C for overnight. Bacterial cells were collected by centrifugation at 15,000 × g for 10 min and lysed by sonication. Cell crude extracts were centrifuged at 12,000 × g for 30 min to remove the cell debris, and the supernatants were applied to a Ni 2+ -NTA (nickel-nitrilotriacetic acid) agarose chromatography column (Novagen, Madison, WI) to purify the C-terminal His-tagged PhaC1-H 6 , PhaC2-H 6 and PhaC1 Pa -H 6 . The non-specific binding proteins were washed by phosphate buffered saline (PBS) containing 80 mM imidazole and the His-tagged proteins were eluted with PBS containing 200 mM imidazole as described [19]. Proteins were analyzed by SDS-PAGE and stored at 4°C.

PHA synthase activity determination
The standard assay (final volume of 1 mL) used 0.

Measurement of bacterial growth on MNP
The growth of strain JMP134 and its derivatives in 0.5 mM MNP was measured at OD 600 . The growth curves to the time were fitted by the modified Gompertz equation with OriginPro 8.0 software as described previously [27,28]. Their maximum specific growth rates (μ max ) were calculated.

Results
In silico analysis revealed that phaC1 and phaC2 encode two distinct classes of PHA synthases The C. pinatubonensis JMP134 genome contains two genes-phaC1 and phaC2-predicted to encode two PHA synthases according to the sequence analysis. As shown in Fig 1A, these two genes are in two distinct PHA synthesis clusters at different loci in the genome. The phaC1 gene is located in front of the putative genes encoding acetyl-CoA acetyltransferase (phaA) and acetoacetyl-CoA reductase (phaB). These are highly similar to the genes involved in the synthesis of polyhydroxybutyrate (PHB) in Ralstonia eutropha H16. The phaC2 gene is located in a five-gene cluster that contains genes encoding the putative acyl-CoA dehydrogenase (fadE), MaoC-like dehydratase (phaJ), PHA granule-associated protein (phaP) and PHA depolymerase (phaZ)-all of these are predicted to be involved in PHA metabolism. PhaC1 and PhaC2 contain 591 and 579 amino acids respectively, with 39% sequence identity. Their major differences come from the N-terminus of the proteins. The C-terminus has high sequence similarity including the catalytic triad Cys296, Asp451 and His479, as well as the conserved lipase box sequence GXCXG (Fig 1B), which are common features in PHA synthases. A phylogenetic tree was constructed based on a multiple-sequence alignment of PHA synthase protein sequences with known or unconfirmed functions using a maximum likelihood method with 1000 bootstraps (Fig 1C). PhaC1, closely related to the well-studied PhaC1 of Ralstonia eutropha H16, was classified as a class I PHA synthase. The PhaC2 was grouped with class II PHA synthases based on the phylogenetic tree. Interestingly, PhaC2 showed a moderate sequence identity (approximately 50%) to the class II PHA synthases of known functions from Pseudomonas spp. strains. But its closer homologs were several putative class II PHA synthases genes of unconfirmed functions. Of the clusters containing functionally identified class II PHA synthase genes so far, two different class II PHA synthase genes were present and separated by the phaZ gene encoding an intracellular PHA depolymerase [3,29,30]. In contrast, the phaC2 gene cluster in strain JMP134 has a distinct gene organization without the additional synthase gene (Fig 1A).

Both PhaC1 and PhaC2 were functional for PHA synthesis in vivo
To investigate if the two phaC genes are involved in PHA accumulation, the phaC1, phaC2 or both genes were deleted from the strain JMP134 genome by allelic replacement to generate JMP134ΔphaC1, JMP134ΔphaC2 and JMP134ΔphaC1ΔphaC2 mutant strains, respectively. PHA production from these strains when grown in 0.2% (w/v) fructose or octanoate was monitored with GC-MS. Here, the only hydroxyalkanoic methyl ester detected in samples of strain JMP134 grown in MNP, fructose or octanoate was 3-hydroxybutyric methyl ester, indicating that strain JMP134 can only produce PHB. In samples of ΔphaC1 and ΔphaC2 strains when grown in fructose or octanoate, 3-hydroxybutyric methyl ester was also detected, suggesting that the ΔphaC1 and ΔphaC2 mutant strains could still produce PHB as the wild type strain (Table 3). In contrast, no hydroxyalkanoic methyl ester was detected in the double mutant strain JMP134ΔphaC1ΔphaC2, suggesting that there are no additional functional PHA  Table 3). The 3-hydroxybutyric methyl ester was detected in the complemented strain of the JMP134ΔphaC1ΔphaC2 double mutant with either phaC1 or phaC2 gene. This suggests that both PhaC1 and PhaC2 are functional PHA synthases. Furthermore, the quantification analysis showed that no significant difference in PHB content of strain JMP134 grown in 0.2% (w/v) fructose or octanoate (Table 3). In addition, both JMP134ΔphaC1 and JMP134ΔphaC2 strains accumulated similar amounts of PHB as JMP134 when grown in these carbon sources (Table 3). This suggests that different carbon sources did not affect the amount of PHB produced in these strains regardless of the presence of functional PhaC1 or PhaC2 or both.
Interestingly, with 0.5 mM MNP (0.0068% w/v) as the sole carbon source, PHA was detected in both JMP134 and JMP134ΔphaC2 samples. There was very little PHA accumulated in the mutant strains JMP134ΔphaC1 and JMP134ΔphaC1ΔphaC2 when grown in MNP. This indicated that PhaC1 but not PhaC2 is involved in the synthesis of PHB when grown in 0.5 mM MNP. Both JMP134 and JMP134ΔphaC2 strains accumulated similar PHB contents when grown with MNP, comparing to that with fructose or octanoate (Table 3).

Transcriptional levels of phaC1 and phaC2
To investigate the transcription of phaC1 and phaC2 when strain JMP134 was grown on different carbon sources, their transcriptional levels were assessed by qRT-PCR. With 0.5 mM MNP ends at 2361552 on the Crick strand in the genome. Open reading frames were annotated with a guide by BLAST results. PhaA: β-ketothiolase; PhaB: acetoacetyl-CoA reductase; PhaZ: PHA depolymerase; PhaP: phasin; PhaJ: R-specific enoyl-CoA hydratase; FadE: acyl-CoA dehydrogenase. (B) Amino acid sequence alignment of PhaC1 and PhaC2 from C. pinatubonensis JMP134, a typical class I PHA synthase PhaC1 Re from R. eutropha H16 and a typical class II PHA synthase PhaC1 Pa from P. aeruginosa PAO1. The boxed region indicates the signature motif of lipase sequence. The arrows indicate conserved residues involved in catalysis (catalytic triad). The triangle indicates a critical residue affecting substrate specificity. (C) Phylogenetic tree constructed using a multiple sequence alignment of several PHA synthases. Classes of different PHA synthases are indicated on the right side of the tree. The triangles indicate PHA synthases of unconfirmed function. The asterisks indicate PHA synthases from strain JMP134. All protein accession numbers are given in brackets.
doi:10.1371/journal.pone.0142332.g001 as the sole carbon source, the transcriptional level of phaC1 was approximately 70-fold (70 ±9.17) higher than that of phaC2. In contrast, when grown in 0.2% octanoate, the transcriptional level of phaC2 was approximately 240-fold (244±45.21) higher than that of phaC1. Interestingly, phaC2 was co-transcribed with its neighbor genes that are predicted to be involved in fatty acid metabolism (Fig 2). The data indicated that phaC1 and phaC2 in strain JMP134 were differentially transcribed when induced with different substrates or metabolites.

PhaC1 and PhaC2 activities in vitro
To investigate whether PhaC1 and PhaC2 exhibit activities towards SCL hydroxyacyl-CoA substrate 3-hydorxybutyryl-CoA (3HBCoA) and MCL hydroxyacyl-CoA substrate 3-hydroxyoctanoyl-CoA (3HOCoA) in vitro, these two enzymes were recombinantly produced and purified, with a similar molecular mass of approximately 65 kDa (Fig 3). Both PhaC1 and PhaC2 showed activity towards 3HBCoA but not 3HOCoA. In contrast, the typical class II PHA synthase PhaC1 Pa from P. aeruginosa as a control displayed enzyme activity towards 3HOCoA but not 3HBCoA (Table 4). This suggested that both PhaC1 and PhaC2 had the same substrate specificity, only catalyzing the polymerization of SCL substrate 3-hydroxyacyl-CoA rather than MCL substrate 3-hydroxyacyl-CoA. The kinetics parameters of PhaC1 and PhaC2 towards 3HBCoA were listed in Table 4. The K m value of PhaC1 for 3HBCoA was lower than that of PhaC2, suggesting that PhaC1 has a higher affinity to 3HBCoA than PhaC2.

PhaC2 also heterologously functions as a class I PHA synthase
PhaC2 along with PhaAB from strain JMP134 were heterologously expressed in P. aeruginosa PAO1 phaC mutant strain (PAO1ΔphaC1 Pa -Z Pa -C2 Pa ). The PHA content and composition in strain PAO1 and its derived strains were analyzed after cultivation in 0.2% (w/v) octanoate. The results indicated that strain PAO1 produced 22.8 ± 2.2% 3-hydroxyoctanoate (3HO)- based PHA of cell dry weight but its phaC mutant did not produce detectable PHA (Table 5). Interestingly, its mutant strain expressing PhaC2, PhaA, and PhaB from strain JMP134 accumulated 42.2 ± 12.5% 3-hydroxybutyrate (3HB)-based PHA. This suggested that PhaC2 from C. pinatubonensis still preferred SCL substrates in a Pseudomonas spp. as in its native background.

Bacterial growth of strain JMP134 and its phaC mutant strains
The optical density at 600 nm of strains JMP134, JMP134ΔphaC1, JMP134ΔphaC2, and JMP134ΔphaC1ΔphaC2 cultured in MSM with 0.5 mM MNP were assessed every 3 hours to investigate if the formation of intracellular PHA would affect bacterial growth. All strains could grow in media containing MNP as the sole carbon source (Fig 4). The maximum growth rate of strain JMP134 (0.78 h -1 ) in MNP was higher than those of JMP134ΔphaC1 (0.22 h -1 ), JMP134ΔphaC2 (0.29 h -1 ), and JMP134ΔphaC1ΔphaC2 (0.18 h -1 ) strains indicating a correlation between the PHA accumulation (stated in Table 3) and the growth.

Discussion
Of the PHA synthase genes with known functions in PHA-producing bacteria, phaC paralogs tend to come from a same PHA synthase class if two or more such gene copies are present in  the genome [3,31]. In this study, phaC1 and phaC2 in the versatile aromatic degrader C. pinatubonensis JMP134 encode functional PHA synthases with same substrate preference but belong to PHA synthase class I and class II, respectively. Subsequent biochemical and genetic analyses demonstrated that, although their products have the same substrate preference, these genes coding for the two PHA synthase were transcribed differentially when different carbon sources were utilized by this strain. It is generally accepted that the class I and class II PHA synthases prefer SCL and MCL substrates, respectively [3]. In this study, biochemical and genetic evidence has demonstrated that, although being categorized in class II, the PhaC2 displayed a distinct substrate preference  [34,35]. Substituting of leucine at amino acid residue 484 (most conserved residue at the corresponding position in class II PHA synthases) to valine in class II PHA synthase from Pseudomonas putida Gpo1 altered its substrate specificity from SCL only to both SCL and MCL [36] (valine and isoleucine are the most conserved residues at the corresponding position in class I and III PHA synthases). Interestingly, an isoleucine at residue 484 was also found in PhaC2. This may be responsible for its unique substrate specificity in this case. This substrate preference of PhaC2 from strain JMP134 is different from those of class II PHA synthase (PhaC1 and PhaC2 from Pseudomonas sp. 61-3 or PhaC2 from Pseudomonas stutzeri 1317), indicating a requirement for new parameters in PHA synthase classification and prediction.
Despite having the same substrate preference to synthesize PHB in vitro, the genes encoding PhaC1 and PhaC2 are regulated differently in strain JMP134 as illustrated by qRT-PCR analysis of their transcriptional levels. The higher transcriptional level of phaC1 in MNP-grown cells indicates that intermediates of MNP catabolism may be involved in the triggering of phaC1 transcription. This correlates with the sole involvement of PhaC1 in PHB synthesis of MNPgrown cells and suggests its role in PHA synthesis when utilizing an unfavorable carbon source (MNP). In contrast, the higher transcriptional level of phaC2 was shown in cells grown in octanoate-a favorable carbon source for this bacterium. Generally, the regulation of PHA synthesis genes is affected by the growth stage and nitrogen limitation [3]. The observation of differential transcription of phaC paralogs when utilizing different carbon sources is of interest but has not yet been reported. These obvious variations in transcription on phaC1 and phaC2 may well reflect the differences in their transcriptional regulations towards unfavorable carbon sources or readily usable carbon sources.
In the PHA synthesis involving class II PHA synthase, the encoding gene clusters exhibit a common feature that two different phaC genes in the same class are separated by phaZ encoding for an intracellular depolymerase. Genes encoding enzymes related to the biosynthesis of PHA precursors are normally scattered in the genomes of PHA producers [3]. In contrast, the cluster in strain JMP134 has a single copy of the class II PHA synthase-encoding gene. This gene is clustered and co-transcribed with genes encoding enzymes predicted to be involved in the synthesis of MCL 3-hydroxyalkanoate. Considering the PhaC2 substrate preferences, we speculated that this gene cluster was evolved from a cluster encoding a set of enzymes responsible for synthesizing MCL-PHA. This unique phaC2-containing genetic and transcriptional organization may well present a different evolutionary path among PHA producers with class II PHA synthases.