The structural insight of class III of polyhydroxyalkanoate synthase from Bacillus sp. PSA10 as revealed by in silico analysis

PhaC synthase is an enzyme responsible for PHA polymerization. In this work, the catalytic mechanism class III of PhaC synthase from Bacillus sp. PSA10 (BacPhaCSynt) was reported through in silico modelling approach based on the primary sequence of the PhaC synthase. The open reading frame BacPhaCSynt has been successfully isolated, cloned and overexpressed the recombinant protein in Escherichia coli BL21(DE3). To know the global architecture and catalytic mechanism, the structural prediction of BacPhaCSynt has been carried out by using MODELLER. The recombinant BacPhaCSynt exhibited monomeric molecular weight (MW) of 43.6 kDa, when it was analyzed on 12% SDS‐PAGE gel. Based on the structural prediction, BacPhaCSynt exhibited global architecture of α/β hydrolase fold, with the root mean square deviation (r.m.s.d) value of 0.94Å. The catalytic residues composition of BacPhaCSynt consists of C151, D307, and H336, but the H336 and D307 residues of the model have been distorted 62.8o and 175.2o from the corresponding residues of the template. Since the D307 is quite a distance from the H336, it might act as a general base for the activation of ‐OH group of the substrate. The results strongly suggested that the mode of action of BacPhaCSynt obeyed the covalent catalysis mechanism.


Introduction
Nowadays, plastic is indispensable material that can be used in immense fields of application, such as packaging materials, parts in the electronics and automotive indus try, and biomedical devices, etc. Despite the high impact of the plastics in our daily life, plastics also can cause se vere the environmental problems. The growing accumu lation of waste consisting of nondegradable carbon back bone plastic polymer, elevated CO 2 and increasing toxic level in the atmosphere caused by the plastic incineration, are the most visible example of the negative impact of the plastic usage (Braunegg et al. 2004; Iwata 2015. There fore, an alternative way to replace petroleumbased plastic is necessary. A biobased material with plasticlike prop erties is regarded as the one alternative to overcome such a petroleumbased plastics usage problem. Biobased mate rial with plasticlike properties is more ecofriendly to the environment because of its biodegradability. Biobased material is nonfossilbased; therefore it does not face the ongoing depletion of fossil resources.
Polyhydroxyalkanoates (PHA) is the one biobased polymer that can be used as an alternative material to replace the petroleumbased plastics. Polyhydroxyalka noates (PHA) are natural polyester comprising various hy droxylalkanoates (HAs) that have been considered as a feasible substitute to conventional petroleumbased plas tics and elastomers (Sudesh et al. 2000; Zhang et al. 2009; Chen and Patel 2012; Chek et al. 2017. Various mi croorganisms have synthesized PHA as an energy and re dox storage material typically under nutrientlimited con ditions in the presence of an excess of carbon source (Hu et al. 2005). The PHA from Bacillus sp., Cupriavidus necator (formerly Ralstonia eutropha) and Pseudomonas sp. have been studied profoundly. However, the lack of understanding on PHA biosynthesis and the weak prop erties of the polymer, bacterial PHA has not been able to completely replace the petroleumbased polymer (Witten born et al. 2016; Sagong et al. 2018. Therefore, a com plete understanding of the PHA biosynthesis is required for the sake of its application to replace the petroleum based polymer. PHA is synthesized by the reaction three enzymes, which consist of phaC synthase (PhaC)(EC.2.3.1.), ace tocaetyl CoA 3ketotiolase (PhaA)(EC. 2.3.1.9), and ace toacetyl CoA reductase (PhaB)(EC.1.1.1.36) (Sagong et al. 2018). Among of them, PhaC synthase is the most important enzyme, since it is responsible for PHA polymerization (Chek et al. 2019). Based on their ki netic properties and catalytic mechanism, PhaC syn thases are classified into four classes. The grouping of the PhaC synthase into each class is dependent on its structure of PhaC alone or in association with other subunit and substrate specificity: Class I, III, and IV produce shortchain length polymer depending on 3 hydroxypropionate (3HP), 3hydroxy or 4hydroxy bu tyrate (3HB, 4HB), 3hydroxyvalerate (3HV), and 3 hydroxyhexanoate (3HH) precursors, while class II pro duce medium chain length of polymer depending on the 3 hydroxyhexanoate (3HH), 3hydroxyhepatanoate (3HHp), 3hydroxyocatanoate (3HO), 3hydroxy decanoate (3HD), 3hydroxyundecanoate (3HUD), 3hydroxydodecanoate (3HDD) (C6 to C12), and availability of the corresponding CoA thioester substrates, originating from three different metabolic pathways (Mezolla et al. 2018).
Class I and II are formed by a single protein, with a molecular weight of 60 kDa, but enzymatically active in homodimeric form. Each monomer consists of two do mains, N domain which is located at the end of the N terminus and catalytic domain. The N domain functions for proteinprotein interaction during dimerization pro cesses. Class I and II also contain the lid or cap which has an important role for open and closing the substrate entrance gate (Chek et al. 2017).
In contrast to class I and II, class III and IV are com posed of the two different protein subunits. The class III and IV of PhaC synthases requires the PhaE and PhaR sub unit, respectively. In the case of class III of PhaC synthase, PhaE requires for PHA polymerization. Class IV of PhaC synthase is prevalent for the Bacillus group and as mention before that the protein is composed of the PhaC synthase and PhaR as an accessory subunit. The PhaC synthase sub unit, with a molecular weight of 40 kDa, of class IV, con tains PhaC box sequence ([GS]XCX[GA]G) that lo cated at the catalytic core. PhaR of the subunit composi tion of class IV of PhaC synthase has a molecular weight of 20 kDa which is comparable to the archaealtype PhaEs. It seems to be that the PhaR also functions for PHA polymer ization since the PhaC synthase alone shows enzymatic activity. PhaR has an additional function for alcoholysis mechanism (Tsuge et al. 2015; Kihara et al. 2017. Alco holysis mechanism is important for the regulation of the PHA molecular weight and also for modifying of PHA terminus. The amino acids sequence identity between the class III and IV of PhaC synthase subunit is 34 percent, which is considered as high enough homology, by the fact that among the members of class III of PhaC synthase have an identity of 38 percent (Tsuge et al. 2015).
Regarding the PhaC synthase catalytic mechanism, a various mechanism has been proposed. One mecha nism is referred to as a nonprocessive pingpong model. The pingpong model requires two cysteine residues for monomer chain transfer during the polymer elongation. In order to perform such a mechanism, the PhaC synthase must be in dimeric form. The second mechanism is the processive model. The processive model does not neces sarily require two cysteine residues. Therefore, the pro cessive model can be taken place in a single catalytic site (Mezolla et al. 2018). The third mechanism that has been proposed based on the C. necator PhaC synthase crys tal structure implies that once newly 3HBCoA enters the catalytic core, it would be nucleophile attacked by cys teine residue to produce 3HBCys. Following the binding of incoming substrate (3HB)2CoA, the hydroxyl group of 3HBCys is deprotonated by the histidine residue, facili tated through basicity modulation by the aspartate residue (Wittenborn et al. 2016). All the proposed catalytic mech anisms are based on the crystal structure of class I of PhaC synthases that have been recently determined. For ex ample, the PhaC synthase from C. necator, a class I of PhaC synthase, has been determined from two indepen dent groups (Wittenborn et al. 2016; Kim et al. 2017. Another class I of PHA synthase came from Chromobac terium USM2 has also been determined (Chek et al. 2017). It remains elusive though, how the PhaC synthases per form their catalytic mechanism, as well as the class III of PhaC synthase. To the best of our knowledge, there is no crystal structure either from class II, III and IV have been reported.
Recently we have isolated the PhaC synthase encod ing gene from Bacillus sp. PSA10. BlastX result indi cated that the PhaC synthase from Bacillus sp. PAS10 (BacPhaCSynt) belongs to the class III PhaC synthase. Al though class I and class III produce a similar type of PHA polymer, class III requires accessory subunit (PhaE) for its catalysis (Liebergesell andSteinbuchel 1992; Müh et al. 1999). The results showed that the global architecture of BacPhaCSynt was similar to the class I PhaC synthase, as well as the catalytic mechanism. Interestingly, the analysis also observed that Bacillus sp. PSA10 has PhaR instead of PhaE, in which both proteins have lower homology. Taken together of the finding, it might suggest that class III and IV of PhaC synthase are similar enzymes, but they have different accessories subunit.

Bacterial cell and Polyhydroxyalkanoate (PHA) accumulation test
Bacillus sp. PSA10 (previously isolated from the waste of sago starch processing) was used in this experiment (Yanti et al. 2009). To confirm whether the bacterial isolate able to produce PHA, Bacillus sp. PSA10 was examined for its ability to produce polyhydroxyalkanoate (PHA) polymer on Ramsay minimal media containing 2% of soluble starch as a carbon source. The bacterial cell was cultivated at room temperature for 24 h (Berger et al. 1989; Ramsay et al. 1990). The PHA accumulation was examined after 24 h of cultivation by a light microscope and transmission electron microscope (TEM) (Mesquita et al. 2015).

Bacterial cell and genome preparation
For the genomic preparation, Bacillus sp. PSA10 was cul tivated in LuriaBertani (LB) media for 48 h at room tem perature. After that, cell was harvested by centrifugation at 8000 x g for 5 min. The cell pellet was collected and further subjected for genomic isolation. Bacterial genome isolation was carried out using the Wizard Genomic DNA Extraction Kit (Promega). All the genomic DNA isola tion followed the manufacture's recommendations. The obtaining genomic DNA was then stored in 20 o C for fur ther analysis.

PhaC synthase open reading frame isolation and cloning
To amplify the class III of PhaC synthase and PhaR open reading frame (orf) from Bacillus sp. PSA10, the following pairs of primer was designed base on the PhaC synthase orf sequences from Bacillus group (Accession no. KU233683.1, AB525784 for phaC and phaR, respectively). The sequence of the primes were phbC_f: 5ATGGCAATTCCTTACGTGCAAGAG 3, phbC_r: 5TTATTTAGAGCGTTTTTCTAGCC3, phaR_f: 5ATGGAACAGCAAAAAGTATTTGATCC3 and phaR_r : 5TTACTTGCGAGCTGGCTGCTC3 for the forward and reverse primers, respectively. For the cloning purpose of PhaC synthase orf, the same sequence of the primers was used, except the addi tion of recognition site for NcoI and EcoI. Therefore, the sequence of the primers were phbC_f_NcoI: 5 ATATATCCATGGCAATTCCTTACGTGCAAGAG3, phbC_r_EcoRI: 5ATATATATGAATTCTTATTTAGA GCGTTTTTCTAGCC3. Underlined letters indicate the recognition sites for NcoI and EcoRI, respectively. Poly merase chain reaction (PCR) was performed with T100 ThermoCycler (BioRad) using KOD Polymerase (Toy obo) according to the procedures recommended by the supplier. All DNA oligomers for PCR were synthesized by Macrogen. The DNA sequence was carried out by 1st BASE to further confirm whether the amplified product was the open reading frame of PhbC synthase. To clone the orf of PhaC synthase, the orf was reamplified by PCR using a pair of primers carrying the recognition site for NcoI and EcoRI. The PCR product was then purified by GeneHplow (Geneaid) and then subjected for DNA di gestion. Fast Digest NcoI and EcoRI (Thermo Scientific) restriction enzymes were employed for this purpose. For the expression vector, pET26a (Novagen) was used in this experiment. pET26a vector was linearized by double digestion with the same restriction enzymes. Ligation of the DNA fragment into the expression vector was carried out using Fast DNA ligation kit. The ligation product was named as pETBacPhaCSynt hereafter, and used to transform E. coli BL21(DE3) (Novagen). The growing colonies were then subjected for colony PCR using the T7 promoter and T7 terminator primers (Macrogen) to select the positive clones.

Overexpression of recombinant BacPhaCSynt
To examine the expression of the recombinant PhaC syn thase, the E. coli BL21(DE3) carrying recombinant pET BacPhaCSynt, was grown in LB media supplemented with 50 mg/mL of kanamycin, at 37 ºC for 48 h. On the follow ing day, the culture was transferred into fresh LB media containing 50 mg/mL of kanamycin and the growth was continued for 2 h until the optical density (OD600) of the culture reached 0.5. The culture was then induced by 1 mM of isoprophylthioβDgalactopyranoside (IPTG) for an additional 3 h. After 3 h the 5 mL cell culture was then harvested by centrifugation at 8000 x g for 5 min. The cell pellet was collected and subjected for SDSPAGE analysis (Laemmli 1970).

Phylogenetic tree construction
To analyze the evolutionary history of PhaC synthase, the phylogenetic tree of all class PhaC synthase was con structed. A total of 46 amino acid sequences of PhaC syn thase were used for the phylogenetic construction. Among 46 amino acid sequences, one was the amino acid se quence of BacPhaCSynt. The phylogenetic tree was con structed by using the MEGA7 suite. The evolutionary his tory of the tree was inferred by using the NeighborJoining method (Saitou and Nei 1987) and the evolutionary dis tance was computed using the Poisson correction method (Zuckerkandl and Pauling 1965). To test the integrity of the tree, 500 replicates of the bootstrapping was employed.

Modeling class III BacPhaCSynt
Class III of BacPhaCSynt was modeled by using the MODELLER v.9.20 suite (Webb and Sali 2016). I TASSER suite was used to find the most appropriate tem plate of the model (https://zhanglab.ccmb.med.umich.edu/ FIGURE 2 Evolutionary relationships of PhaC synthases from bacteria and archaeal.The evolutionary history was inferred using the Neighbor-Joining method (Saitou and Nei 1987) The optimal tree with the sum of branch length = 9.19428895 is shown. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree.
The evolutionary distances were computed using the Poisson correction method (Zuckerkandl and Pauling 1965)  ITASSER/) . I TASSER analysis produces several appropriate templates to model the BacPhaCSynt. The best template for model ing of BacPhaCSynt was class I of PhaC synthase struc ture (PDB accession no. 5t6o) from C. necator with the identity value of 27 percent. The five models of BacPhaC Synt generated by MODELLER was then selected for the best model based on the lowest value of molpdf and DOPE (Webb and Sali 2016). The selected model was then fur ther refined by running on the ModRefiner (https://zhangl ab.ccmb.med.umich.edu/ModRefiner/)  to produce the protein structure model with the best physical quality of the local structure. The model structure was refined iteratively and stopped when the best structure model was obtained. Every re fined product was then evaluated to their stereochemical quality of the model by analyzing residue by residue and overall structural geometry by using SAVES v.5.0 suite program (https://servicesn.mbi.ucla.edu/SAVES/) (Mor ris et al. 1992; Hooft et al. 1996. The best quality of the model would be expected to have over 90 percent of residues in the most favored regions. PyMol Version 2.1.1 was used to visualize the model (PyMol), measurement of the residueresidue distance, residuesubstrate distance, and torsion angle. PyMol also used for manipulation of the substrate analog docking. To be noted that the energy binding value of the docking result was not calculated. The CAVER 3.0.3 plugin suite was employed to predict the possible substrate tunnel of the model. Standard param eters of the tunnel prediction were as follows: minimum probe radius was set at 0.9Å; shell depth and shell radii were 4 and 3Å, respectively; clustering threshold was set to 3.5Å; starting point used was all the structure fold and the starting point for optimization was set for maximum distance was set to 3Å and the desired radius was set to 5Å (Chovancová et al. 2012).

Cloning and expression of class III of BacPhaCsynt open reading frame
Previously it has been reported that Bacillus sp. PSA10 could accumulate the polyhydroxyalkanoate (PHA) poly mer (Yanti et al. 2009). To further confirm the previous result, the ability of the Bacillus sp. PSA10 to accumulate PHA was tested. The bacterial cell was grown in Ramsay minimal medium containing 2% of soluble starch. The PHA accumulation was examined microscopically (light microscope and TEM) after 24 h cultivation. The result confirmed that Bacillus sp. PSA10 showed an ability to accumulate PHA with the soluble starch as a carbon source ( Fig. 1a and b). Previously, it has been reported that the PHA type accumulated by Bacillus sp. PSA10 was poly hydroxybutyrate (PHB) (Yanti et al. 2009 (Figure 2). As shown in Figure 2, (Figure 3), the positive cloned was further examined for BacPhaCSynt recombinant pro tein expression. Since the expression is under control of lac promoter, the isoprophyl thioβDgalactopyranoside (IPTG) was used to induce the expression of the recombi nant class III of BacPhaCSynt. The recombinant protein was induced after 2 h of incubation with IPTG and pro longed the incubation time to another 1 hour did not sig nificantly increase the expression level (Figure 4). Disrup tion of cell suspension in 10 mM TrisHCl bufffer pH 8.0 by sonication indicated that the cell suspension turbidity did not turn into translucent, indicating that the recombi nant BacPhaCSynt formed inclusion body (Pradani, per sonal communication). Previously, it was also reported that overexpression of PhaC synthase from Chromatium vinosum under the control of T7lac promoter produced only 25% of solubility (Müh et al. 1999). The insolubility of the recombinant protein indicates that the recombinant protein might not properly fold. Several proteins require a chaperone to help their proper folding. It is already known that the class III of PhaC synthase requires another subunit (PhaE or PhaR) for its activity. Our recombinant protein may require the presence of PhaE or PhaR to present in a soluble form. Therefore, it is noteworthy to construct the coexpression system to simultaneously overexpress the PhaC synthase and PhaE or PhaR subunit.

The global fold class III of BacPhaCSynt
To understand the catalytic mechanism, in silico modelling of class III of BacPhaCSynt has been carried out. Accord ing to the best of our knowledge, the only class I of PhaC synthase structure is available, there is no report regarding the structure of other classes of PhaC synthase. Class I, III, and IV produce similar polyhydroxyalkanoate polymer depending on the type of precursor, however, the catalytic mechanism of class III remains elusive. Unlike the class I PhaC synthase, class III requires another subunit that is encoded by a different gene (phaE) (Liebergesell andSteinbuchel 1992; Müh et al. 1999). To model the class III of BacPhaCSynt, the orf of class III of PhaC synthase Bacillus sp. PSA10 was translated into the amino acid se quence. Based on the sequence alignment, BacPhaCSynt showed the sequence identity of 18% to 27% to the class I of PhaC synthase. All the catalytic residues of BacPhaC Synt are located on the fully conserved region. Secondary prediction indicates that BacPhaCSynt contains 14 of α helix and 8 of βsheet ( Figure 5).
To predict the 3D structure of BacPhaCSynt, the amino acid sequence was then used to find the appropriate tem plate to model the structure. By running the ITASSER suite, it was found that the most appropriate template to model the class III of BacPhaCSynt was the structure class I of PhaC synthase (PhaCCn) from C. necator (PDB ac cession no. 5t6oA) with the identity of 27% which is in agreement with the amino acid sequence alignment. The ITASSER result was then used to model the BacPhaCSynt by using the MODELLER. From the modelling results, 5 models of BacPhaCSynt were constructed and among them, model 1 (phbC.B99990001.pdb) (Table 1S, Supple mentary material) was the best, with the molpdf and DOPE scores were 2146.21 and 39167.52, respectively. The re sulting model was then refined by using the crystal struc ture of class I of C. necator (PDB accession no. 5hz2). Af ter several times refinement of the model, the most appro

FIGURE 7
The substrate tunnel entrance of BacPhaCSynt shows the substrate entrance mechanism. The green spares represent the tunnel diameter and the red arrow indicates the entrance site. The substrate analog β-octyl glucoside (β-OG)(grey) is also shown. Black dashed line indicates the tunnel path. The orange circle indicates the catalytic core location. Inset indicates the distance of the β-OG to the C151 residue. The distance was 4.8 Å.
priate model was obtained with over 90% of the amino acid residues in the most favoured regions of the Ramachan dran plot (Fig. 2S, Supplementary material). The resulting model has a similar fold to the template with the root mean square deviation (r.m.s.d) value of 0.94Å. The model of the BacPhaCSynt has α/β fold which is the common super secondary structure of the α/β hydrolase family ( Figure 6). The structural model of BacPhaCSynt also contains CAP and CAT domains. The CAP domain functions for closing and opening the substrate channel. Open and close mech anism of the CAP domain important for the regulation of the accessing of the catalytic center. When the CAP do main flips away, it will open the entrance of the catalytic center that makes product release and substrate entrance from or to the catalytic center (Mezolla et al. 2018; Chek et al. 2017. In this work, the prediction had been focused on the single subunit of the protein, since the single subunit was sufficient to explain the global fold of the protein. Al though, the biological activity of class I of PhaC synthase is in homodimeric form, the symmetrical dimeric architec ture in which the active site of each monomer is separated from one another by 33Å across an extensive dimer in terface, suggesting that the catalytic mechanism occurs in single subunit (Wittenborn et al. 2016).

Substrate entrance channel of BacPhaCSynt
By using the model of BacPhaCSynt, the substrate channel pattern of the model structure was predicted by employ ing the CAVER ver. 3.0.3 suite. The substrate channel was located in proximity to the CAP domain (Figure 7). When the substrate analog βoctyl glucoside (βOG) was fitted onto the predicted channel, it clearly showed that the substrate analog was fit to the path of the channel (as in dicated the black dashed line, Figure 7) The end of the substrate analog was also directed into the catalytic core. The end of the substrate analog seems to be approaching to the cysteine residue (C151). The distance of the end of the substrate analog to the C151 was 4.8Å. Such dis tance, however, is quite far for the donoracceptor proton process. Therefore, the βOG is only to show the path of the substrate entry, not for showing the nucleophile attack by the C151 residue (see possible catalytic mechanism). Substrate analog of βOG is the inhibitor that is commonly used for cocrystallization of lipase to make the lid of li pase performs in the open conformation (Roussel et al. 2002).

FIGURE 9
The possible interaction between substrate analog and catalytic residues of BacPhaCSynt. Substrate analog was C11Y4 phosphonate, an inhibitor of the dog gastric lipase (DGL). The distance of C151 residue is only 2.5Å from the ester bond of the substrate. The distance of C151 and H336 residue is 3.4Å (yellow dashed line). The distance of Oδ of D307 to either Nδ1 or Nε2 of H 336 is 6.2Å and 6.4Å (yellow dashed line), respectively. Black curve dashed arrows indicate the electron transfer.

Possible catalytic mechanism of BacPhaCSynt
By the model, the orientation of the catalytic residues of class III BacPhaCSynt could be determined. The cat alytic residues of the class III BacPhaCSynt were cysteine (C151), aspartate (D307) and histidine (H336), which were similar to the catalytic triad of class I of PhaCCn synthase. The catalytic residues of model class III of BacPhaCSynt have a similar orientation to that of class I of PhaCCn synthase (Figure 8). However, the residue D307 and H336 showed little distortion from the corresponding residues of the template. The D307 distorts 175.2º from the configuration of D480, while H336 distorts 62.8º from the configuration of H508. Despite the distortion orienta tion of the D307 and H336, the presentation of all these three residues suggested that BacPhaCSynt had a similar catalytic mechanism to the class I of PhaC synthase. When the substrate analog of C11Y4 phosphonate was docked on to the catalytic core of the BacPhaCSynt, the ester bond of the substrate is close enough to the C151 residue, which is only 2.5Å (Figure 9). Within this distance, it makes possi ble for the C151 residue to do nucleophile attacks the ester bond of the substrate analog. The atomic distance within 2.22.5Å is categorized as a strong electrostatic interac tion (Jeffrey 1997). The Nε1 of H336, which is only 3.4Å from the C151, makes H336 possible to performs hydro gen atom abstraction of sulfhydryl (SH) group of C151 residue, although such distance is categorized as the weak electrostatic. From figure 9, it also clearly showed that the position of D307 quite far from the H336. The distance between the Oδ2 of D307 to the Nδ1 and Nε2 of H336 was 6.2 and 6.4Å, respectively, which is quite far for the deprotonation of H336 by the D307 residue. Therefore, the D307 is not part of the catalytic triad but functions as a general base catalyst in the activation of the 3OH group of HBCoA (Tian et al. 2005).

Conclusions
In this work the open reading frame of BacPhaCSynt, a class III of PhaC synthase from Bacillus sp. PSA10 was successfully isolated and overexpressed the recombinant protein in the E. coli system. In silico analysis clearly showed that the global fold of the BacPhaCSynt was sim ilar to the class I of PhaC synthase with the r.m.s.d value of 0.97Å. The model structure of BacPhaCSynt also sug gested that class III of PhaC synthase has a similar catalytic mechanism to that of class I of PhaC synthase, specifically it obeys the covalent catalysis mechanism. Therefore, our results give a hints the catalytic mechanism of class III PhaC synthase. Figure 1S. Alignment of PhaR from Bacillus sp. PAS10 with its homolog; Tabel 1S. Five model generated by MODELLER based on the template class I polyhydroxy alkanoate synthase C. necator (5t6o); Figure 2S. Ra machandran plot of the refined model BacPhaCSynt.