Organophosphate hydrolase interacts with ferric-enterobactin and promotes iron uptake in association with TonB dependent transport system.

Our previous studies have shown the existence of organophosphate hydrolase (OPH) as a part of the inner membrane associated TonB complex (ExbB/ExbD and TonB) of Sphingobium fuliginis. We now show its involvement in iron uptake by establishing direct interactions with ferric-enterobactin. The interactions between OPH and ferric-enterobactin were not affected even when the active site architecture is altered by substituting active site aspartate with either alanine or asparagine. Protein docking studies further substantiated these findings and predicted the existence of ferric-enterobactin binding site that is different from the catalytic site of OPH. A lysine residue (82 K) found at the predicted ferric-enterobactin binding site facilitated interactions between OPH and ferric-enterobactin. Substitution of lysine with alanine did not affect triesterase activity, but it abrogated OPH ability to interact with both ferric-enterobactin and ExbD, strengthening further the fact that the catalytic site is not the site for binding of these ligands. In the absence of interactions between OPHK82A and ExbD, OPHK82A failed to target membrane in E. coli cells. The Sphingobium fuliginis TonB dependent transport (SfTonBDT) system was reconstituted in E. coli GS027 cells generated by deleting the exbD and tonB genes. The E. coli GS030 cells having SfTonBDT system with OPH showed increased iron uptake. Such an increase was not seen in E. coli GS029, cells having SfTonBDT system generated either by omitting OPH or by including its variants, OPHD301A, OPHD301N suggesting a role for OPH in enhanced iron uptake.


INTRODUCTION
Phosphotriesterases (PTEs), also known as organophosphate hydrolases (OPH), are present in a number of soil bacteria. These binuclear metallo-enzymes hydrolyze P-O, P-S and P-C bonds found in a variety of organophosphate insecticides and nerve agents (1). The inner membrane associated OPH contains a 23 amino acid long signal peptide with unique characteristic features. It contains a Twin Arginine Transport (TAT) motif (MQTRRVVLK) at the N-terminus and a lipobox motif (LAGC) at the signal peptidase cleavage site (2). OPH is targeted to the inner membrane in a prefolded conformation and remains anchored to it through a diacyl glycerol moiety linked to an invariant cysteine residue present at the junction of signal peptidase cleavage site (2,3). There are no transmembrane domains in OPH and the entire protein exists in the periplasmic space (3). Recent studies have shown the existence of OPH as part of a multiprotein complex (3,4). The TonB dependent transport system components, TonB, ExbB/ExbD were co-eluted when the membrane associated OPH was affinity purified from either S. fuliginis or S. wildii (2,3). The OPH specifically interacts with the inner membrane associated proton motif force component ExbD and energy transducer TonB to form a four component-TonB complex (4). OPH is targeted to the inner membrane only when co-expressed with ExbB/ExbD. In their absence it remains in the cytoplasm suggesting that formation of a multi-protein complex in the cytoplasm is essential for OPH before it targets the membrane. TonB dependent transport system plays a key role in the transport of nutrients across the energy deprived outer membrane (5). It contains an outer membrane transporter also known as TonB dependent transporter (TBDT) and an inner membrane associated TonB complex comprising of an energy generating proton motive force (PMF) components ExbB/ExbD and energy transducer TonB. Upon binding to the substrate, the TBDT Downloaded from https://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200299/888537/bcj-2020-0299.pdf by guest on 29 July 2020 Biochemical Journal. This is an Accepted Manuscript. You are encouraged to use the Version of Record that, when published, will replace this version. The most up-to-date-version is available at https://doi.org/10.1042/BCJ20200299 undergoes conformational change exposing a pentapeptide motif (ETVIV), known as TonB box. The inner membrane associated TonB protein specifically interacts with TonB box and transduces energy harvested from PMF components ExbB/ExbD (6). Initially this unique outer membrane transport system was identified with the infection of phage T1 (7). Soon its role was established in the transport of ferric-enterobactin, vitamin B12 and a number of other nutrients in both pathogenic and non-pathogenic Gram negative bacteria (5).
Lateral transfer of OPH coding organophosphate degradation (opd) genes among soil bacteria is a known phenomenon (8,9). Identical opd genes, identified as part of mobile elements or self-transmissible plasmids, code for OPH in taxonomically unrelated bacteria (10). A strong selection pressure, accumulated in agricultural soils due to repeated and indiscriminate use of chemicals, is believed to have a role in evolution and lateral transfer of opd genes (11,12).
In view of OPH association with TonBDT and inherent relationship with lactonases, we have examined if it has a role in transport of ferric-enterobactin. Our investigations failed to show any enterobactin hydrolase activity to OPH. However, we found existence of high affinity interactions between OPH and Ferric-enterobactin. Further, the wild type Sphingopyxis wildii cells showed better iron uptake and growth over opd negative mutants.
Similar effect was seen in E. coli GS029 cells reconstituted with OPH containing Sf TonBDT system suggesting a role for OPH in iron uptake.

Strains, plasmids and growth conditions
Bacterial strains and plasmids used in this study are listed in table 1. All biochemicals, including desferri-enterobactin (Ent) were procured from Sigma Aldrich, India. Restriction and other enzymes used in gene manipulations were procured from Fermentas, India. The Downloaded from https://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200299/888537/bcj-2020-0299.pdf by guest on 29 July 2020 Biochemical Journal. This is an Accepted Manuscript. You are encouraged to use the Version of Record that, when published, will replace this version. The most up-to-date-version is available at https://doi.org/10.1042/BCJ20200299 basic gene manipulation techniques were performed essentially by following standard procedures described elsewhere (13). The E. coli and Sphingopyxis wildii strains were grown at 37 ○ C and 30 ○ C respectively either in LB or in minimal medium (3). Expression of Sf TonBDT system components were induced in mid log phase grown E. coli Arctic Express cells at 18 0 C. The iron-limiting minimal medium was prepared following procedures described elsewhere (14). When necessary, antibiotics ampicillin (100μg/ml), kanamycin (30μg/ml), polymyxinB (10μg/ml), tetracycline (20μg/ml) or chloromphenicol (20μg/ml) were supplemented to the growth medium. Oligonucleotides used in this study are listed in supplementary table 1.

Enzyme assays
The OPH activity was determined by following standard procedures established in our laboratory (15). Cell fractionation and marker enzyme assays were performed following procedures described elsewhere (2). Nitrate reductase, and glucose-6-phosphate dehydrogenase were used as membrane and cytoplasmic marker enzymes. The purity of membrane and cytoplasmic fractions were established by assaying maker enzymes in subcellular fractions (2). The enterobactin hydrolase assay was performed by following protocols described elsewhere using ferric-enterobactin as substrate (16). Briefly the stock solution was prepared by dissolving 1mg of ferric-enterobactin in 100μl of DMSO and was used to dispense 1mM of ferric-enterobactin into prechilled sterile eppendorf tubes. The test reaction contained pure OPH (1μg) and the control reaction was prepared without adding OPH. The volume of the reaction was then adjusted to 6μl with 50mM sodium phosphate buffer pH 8.0 before incubating the tube at 37 0 C for 15minutes. After incubation the reaction mix was extracted with ethyl acetate and analyzed on TLC plate to detect formation of linear enterobactin (16). Downloaded from https://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200299/888537/bcj-2020-0299.pdf by guest on 29 July 2020 OPH interactions with desferri-enterobactin and ferric-enterobactin.
The OPH interactions with desferri-enterobactin and ferric-enterobactin were studied by performing native PAGE, surface plasmon resonance and fluorescence emission measurement using purified mature form of OPH (OPH) and commercially procured desferri-enterobactin. OPH was purified from E. coli (pUCOPH) cultures grown in 5 liters of terrific broth by following established protocols (17,18).
After electrophoresis the gels were destained and the western blots were performed using anti-OPH antibodies following established procedures (3).

Surface plasmon resonance (SPR)
The SPR analysis was used to study OPH interactions with desferri-enterobactin and ferric-enterobactin. The pure OPH was covalently immobilized by amine coupling on a carboxymethylated dextran sensor chip CM7 (GE Healthcare) (19). The amine coupling was done by taking 50μg/ml of OPH protein as ligand in 10mM sodium acetate buffer (pH 5.0).
The ligand was then injected at a flow rate of 30μl/min with a contact time of 60 seconds.
The process was continued until it captured 5340 response units (RU). The blank (reference) surface was treated in a similar manner using the same buffer prepared by Downloaded from https://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200299/888537/bcj-2020-0299.pdf by guest on 29 July 2020 omitting OPH. Ferric-enterobactin was injected as analyte in a running buffer (PBS, pH 7.4 + 0.005% P20 + DMSO 2%). Buffer optimization was done to correct variations between samples and the solvent (DMSO). Initially titration experiments were performed by injecting Fe III (Ent) 3− prepared at increasing concentrations of 3.12μM, 6.25μM, 12.5μM, 25μM, 50μM, 100μM in optimized running buffer at a flow rate of 30 μl/min. Following each injection, sensor chip surfaces were regenerated with a 30 second injection of 50% DMSO in running buffer and the kinetics were repeated for three times. Data were analyzed using Biacore T200 Evaluation software 2. 0 version and 1:1 Binding Model (GE Healthcare). An affinity curve was generated manually by plotting the equilibrium binding response (R eq ) against analyte concentration to obtain the dissociation constant (K D ) values.

Fluorescence emission measurements
Room temperature fluorescence emission spectra were measured with a LS-55 Spectrofluorimeter (PerkinElmer corporate, USA) equipped with 1.0 cm quartz cells at room temperature with an excitation wavelength of 280 nm, and slit width of 5.0 nm for both excitation and emission. The concentration of OPH, and its variants OPH D301A , OPH D301N were fixed at 1 μM, and the concentrations of desferri-enterobactin and ferric-enterobactin dissolved in 0.1M phosphate buffer (pH 7.4) were varied from 1 to 11 μM. Three independent experiments were performed to obtain statistically significant data. The inner filter effect was corrected for absorption of exciting light and re-absorption of the emitted light. The binding constant was calculated using the maximum fluorescence intensity value at maximum emission wavelength (335nm) as described elsewhere (20).
Downloaded from https://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200299/888537/bcj-2020-0299.pdf by guest on 29 July 2020 Biochemical Journal. This is an Accepted Manuscript. You are encouraged to use the Version of Record that, when published, will replace this version. The most up-to-date-version is available at https://doi.org/10.1042/BCJ20200299

Generation of OPH variants
Site-directed mutagenesis was performed to generate OPH variants by using a Q5 These plasmids were first transformed into E. coli strain, and the expression and stability of the OPH variants were assessed both by performing western blots and measuring OPH activity. The OPH variants were then purified (17) and used for performing further studies.

ExbD N6xHis interactions with OPH variants
Pulldown assays were performed to assess interactions between OPH variants and ExbD N6xHis following procedures optimized in our laboratory (3). Briefly the E. coli GS027 (pGS19) cells were transformed with the expression plasmids pMD301A, pMD301N and pMK82A and co-expression of OPH variants with ExbD N6xHis were induced by adding 0.1mM IPTG.
The clear lysate prepared from the induced cultures was taken to perform pull-down experiments and western blots were performed to detect OPH and ExbD N6xHis (4).

Circular Dichroism (CD) Spectroscopy
CD spectra of OPH and OPH variants (OPH D301A ), (OPH K82A ) at a concentration of 100μg (0.0014 mM) were recorded with a Jasco J-810 spectropolarimeter (JAPAN) using a Quartz cuvette with path-length of 0.2 cm in a Nitrogen atmosphere at 25°C (21)(22)(23). Three scans were performed at a rate of 50nm/min and the spectral data were recorded in the range of 190 to 270nm.

Identification of ferric-enterobactin binding site. i) Investigations on the catalytic site
We initially assumed ferric-enterobactin binds at the catalytic site of OPH and performed targeted docking of ferric-enterobactin using Autodock4.2 (24). The OPH structure was taken from its crystal structure (Pdb Id:1EYW) bound with the substrate analog triethyl phosphate.
The 3D structure of ferric-enterobactin was retrieved from the crystal structure of a siderocalin-ferric-enterobactin complex (Pdb Id:3CMP). The active site of OPH (the receptor) was prepared for docking by removing all the water molecules except the one near the Zn metal ion at the catalytic site. The substrate analogue was also removed. The ferricenterobactin structure was used as ligand during the docking. Thirty runs of docking were Downloaded from https://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200299/888537/bcj-2020-0299.pdf by guest on 29 July 2020 attempted using the genetic algorithm in AutoDock. The docking poses were analyzed using AutoDock Tools. Pymol was used for visualization of the docking poses.

ii) Search for a potential ferric-enterobactin binding site
As the volume of the catalytic site was smaller than the volume of ferric-enterobactin, the catalytic site was not found to be suitable for docking. We have therefore searched for a potential ferric-enterobactin binding site in OPH. Since OPH exists as a homodimer (Pdb Id:1EZ2), we investigated the dimer structure for identification of potential ferricenterobactin binding sites using the tools mentioned above.

Reconstitution of Sf TonBDT system in E. coli GS027
Initially we have deleted both exbD and tonB genes from Arctic Express strains following a P1 transduction method described elsewhere (25). The strain designated as E. coli GS027 was then tested for reconstitution of Sf TonBDT system. Functional TonB dependent transport (TonBDT) system contains an inner membrane associated Ton-complex and a transporter (TBDT) localized in the outer membrane (5). While reconstituting Sf TonBDT system E. coli GS027 cells were transformed with the expression plasmids pGS6 and pGS25 coding Ton-complex components Sf ExbB NFLAG , Sf ExbD CMyc , Sf TonB C6xHis and outer membrane transporter Sf TBDT C6xHis respectively. The subcellular localization of these proteins was then established by performing western blots using tag-specific antibodies. The GS027 (pGS6 + pGS25) cells designated as GS029 were then used to independently transform with expression plasmids coding either wild type OPH (pOPHV400) or its variants preOPH D301A (pPD301A), preOPH D301N (pPD301N) and preOPH K82A (pPK82A). The resulting strains designated as GS030 (pGS6+pGS25+pOPHV400), GS031 (pGS6+pGS25+pPD301A) and GS032 (pGS6+pGS25+pPD301N) contain Sf TonBDT system either with wildtype OPH or with its variants preOPH D301A , preOPH D301N and preOPH K82A .

Uptake of radiolabeled ferric-enterobactin
The E. coli GS027 and its derivatives GS029, GS030, GS031 and GS032 were grown to mid log phase in iron-sufficient medium and the expression of Sf TonBDT components was induced for two hours following standard protocols. The induced cells were harvested and washed twice with the iron-free minimal salt medium. The cell pellet was then resuspended in an equal volume of iron limiting medium. Immediately, the culture flasks were shifted to 18°C and the expression of Sf TonBDT components was induced for 12 hours by adding 1mM IPTG. After induction, the cells were harvested, washed extensively with iron free minimal medium and resuspended in the same culture medium to obtain a cell suspension with an OD 600nm of 1.0. Iron uptake was performed with this cell suspension. To an aliquote of cell suspension radiolabeled ferric-enterobactin (equivalent to 178 μmol of 55 Fe) was added incubated at 37°C for 2 h, with gentle shaking (150 rpm). The cells were harvested, extensively washed with 0.1M LiCl 2 followed by two washes with cold iron-free minimal media. After washing, the cell pellet was dried by keeping the pellet in a dry bath set at 60°C. The dried cells were transferred into 5ml of scintillation fluid and the radioactivity determined in the LSS as described above. Wild-type Arctic Express cells treated in similar manner served as controls. Similar strategy was followed while measuring iron uptake in S.
wildii, except that the cells were grown in a minimal medium that supported the growth of these cells (3,14)

RESULTS
The study is designed to address the physiological role of OPH by taking into consideration its promiscuous lactonase activity and interactions with the inner membrane associated Ton-complex components ExbD and TonB (4). Ton-complex plays a role in the transport of nutrients, especially in the transport of ferric-enterobactin across the outer membrane. We hypothesized that the OPH played a role in the hydrolysis of the trilactone ring of enterobactin (Ent), a critical step required to release the iron from ferric-enterobactin.
We experimentally validated this role by incubating OPH with pure ferric-enterobactin at different time intervals and analyzing the products by TLC to detect linear Ent formed due to lactonase activity of OPH. Instead of finding a discrete spot corresponding to linear version of Ent, we detected a streak that was reproducible and which was not observed in the lane spotted from a reaction mix prepared by omitting the OPH (Supplementary fig. S1). These results have prompted us to undertake further studies to detect existence of specific interactions between desferri-enterobactin / ferric-enterobactin and OPH.

OPH shows stronger binding to ferric-enterobactin than to desferri-enterobactin:
We performed three independent studies to ascertain interactions between OPH and desferrienterobactin / ferric-enterobactin. Initially we performed both native PAGE (Fig. 1) and surface plasmon resonance (SPR) spectroscopy (Fig. 2, Panels A -D) to obtain qualitative and quantitative data on OPH interactions with both desferri-enterobactin and ferricenterobactin (19). These results were then revalidated by performing fluorescence Downloaded from https://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200299/888537/bcj-2020-0299.pdf by guest on 29 July 2020 spectroscopy. Native PAGE has been successfully used to detect interactions between Ent and factor-H binding protein (fHbp) (19). The fHbp-Ent complex moved as an additional faster migrating band on native PAGE (19). While employing similar strategy we incubated pure OPH (10μM) with 100μM of ferric-enterobactin for 10 ( Fig. 1, Panel B interactions found between ferric-enterobactin and OPH appears to be much stronger than desferri-enterobactin.

Active site of OPH has no role in Ent interactions
Initially we assumed involvement of OPH active site in establishing interactions with Ent.
We therefore generated variants of OPH with altered active site by substituting active site aspartate (D301) with amino acids having either light or bulky side chains. These OPH variants, OPH D301A and OPH D301N retained stability, secondary structure like native OPH ( Fig. 3, Panels A & C) and failed to show OPH activity (Fig. 3, Panel B, lanes 2 & 3). We took these triesterase negative OPH variants, OPH D301A , OPH D301N to test their ferricenterobactin binding ability by measuring both SPR and fluorescence emission. Interestingly these two independent studies have shown the existence of ferric-enterobactin binding ability both for OPH D301A and OPH D301N (Fig. 4, Panels A to F). The OPH D301A showed better ferricenterobactin binding efficiency than OPH D301N and its binding efficiency was comparable to the wild type OPH (Fig. 4, Panels A, C & E). The existing data fails to explain reasons to have such a weak affinity between OPH D301N and ferric-enterobactin (Fig. 4,

OPH contains A novel ferric-enterobactin binding site
Since active site mutations failed to influence OPH-Ent interactions, we presumed that there exists an exclusive site in OPH for binding with ferric-enterobactin. In order to investigate this possibility, we performed blind docking of ferric-enterobactin on OPH using AutoDock. We obtained seven different potential ferric-enterobactin binding site possibilities in OPH, of which the PARS server supported only three of them (26). Among the three sites, two were identical to each other, situated on both the monomers related by the dimer symmetry (Fig. 5), with the third seen at the dimer interface. The sites on the two monomers corresponded to the best binding energy (Fig. 5) and these sites were regarded as potential binding sites for ferric-enterobactin that were used for all further studies as mentioned below.
To get a better model of a potential OPH-ferric-enterobactin complex with regard to the binding site as mentioned above we did targeted docking at this site as described in with Ton-components to form a four component Ton-complex (4). Since, the GS030 cells contain necessary Ton-complex components, the plasmid, pOPHV400 coded OPH CAviTag targeted the membrane indicating successful reconstitution of OPH containing Sf TonBDT system in E. coli GS030 cells (Fig. 7, Panel VI, lane M). Since TonBDT system is known to transport ferric-enterobactin, the three strains, GS027, GS029 and GS030, together with the wild type, were grown under iron limiting conditions and growth was monitored. Both GS029 (pGS6+pGS25) and GS030 (pGS6+pGS25+pOPHV400) grew successfully under iron-limiting conditions. However, the strain GS030 expressing Sf TonBDT system with OPH CAviTag showed better growth than the strain GS029 having without OPH CAviTag (Fig. 7, Panel B).

OPH-dependent increase in iron uptake
In order to test if the increased growth in these cells was due to increased iron uptake, we incubated cells acclimatized to the low iron-containing medium for two hours with pure radiolabeled ferric-enterobactin. As expected, no 55 Fe was found in ∆exbD, ∆tonB (GS027) cells (Fig. 7, Panel C, lane 2). Nearly 28,000 picomoles of iron was found in wild-type cells

Both triesterase and Ent binding activities are critical for increased iron uptake
A clear positive influence of OPH was seen on growth and iron uptake of E. coli GS030 (pGS6+pGS25+pOPHV400) cells having Sf TonBDT system with OPH. However, it was not have retained ability to cross-talk with ExbD N6xHis (Fig. 8, panels A & B, row-II, lane 12).
When cell lysates prepared from GS027 (pGS19+pMD301A) co-expressing both OPH D301A and ExbD N6xHis was passed through Ni-NTA magnetic beads OPH D301A got copurified with ExbD N6xHis (Fig. 8 Surprisingly the OPH dependent enhanced iron uptake was not seen in cells having Sf TonBDT system reconstituted with OPH D301A and OPH D301N suggesting that 301 aspartate that contributes for triesterase activity is critical for OPH dependent enhanced iron uptake in E. coli GS031 and GS032 cells (Fig. 7, Panel C lanes 5, 6).
Since OPH dependent increase in iron dependent growth was seen in E. coli cells reconstituted with Sf TonBDT system we wished to find out if OPH has similar influence on the iron uptake in S. fuliginis cells. We have made several unsuccessful attempts to generate opd negative mutant of S. fuliginis following standard procedures (2). However, we were able to obtain opd negative mutant of Sphingopyxis wildii following similar strategies (2).
Identical opd genes exist both in S. fuliginis and S. wildii (9). We therefore used opd negative Despite the observed structural diversity in TBDTs, the energy harnessing and transducing Ton complex of TonBDT is highly conserved. In Gram-negative bacteria it is a ternary complex comprising TonB, ExbB and ExbD in a 1:7:2 ratio (32). As opposed to this established notion, in S. fuliginis we find a four-component TonB complex comprising TonB, ExbB, ExbD and OPH (4). The existence of this quaternary Ton complex appears to be advantageous to the cell as it contributes to better cell growth and more efficient iron uptake ( Fig. 7, Panels B & C). However, the mechanistic details of how OPH contributes to an increased iron uptake are not known.
In the TonBDT systems known to date, the transport of ferric-enterobactin and subsequent release of ferric ions bound to Ent are independent events. In certain cases, the trilactone ring of ferric-enterobactin transported to periplasmic space / cytoplasm is hydrolysed to facilitate the release of ferric iron (37). Although this mechanism is found in certain Gram negative bacteria it is considered to be less efficient to the cell (38). The Ent made in the cell cannot be recycled and the cell has to make a fresh Ent for transport of every single iron atom. Protonation induces huge change in coordination mode of ferricenterobactin besides altering its molecular shape (38). Such molecular events weaken association of ferric ions with Ent facilitating their release at physiological pH (39). Release Downloaded from https://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20200299/888537/bcj-2020-0299.pdf by guest on 29 July 2020 of iron from ferric-enterobactin due to protonation is physiologically advantageous to the cell as it promotes recycling of Ent (38).
However, the precise mechanism by which OPH contributes for the enhanced iron uptake in       The experiments were performed essentially as detailed in Figure 2 for the wild type OPH, with Panels A and B representing the sensograms of OPH D301A and OPH D301N respectively and Panels C & D the derivation of the K D values from the graphs generated from the 'response units' vs the ferric-enterobactin concentration. Spectrofluorimetric assay, performed as for the wild type OPH is represented in Panels E and F for OPH D301A and OPH D301N respectively, with the determination of the binding sites from the Stern-Volmer plots (shown as insets) generated as described in Figure 2. Arrow indicates baseline emission of the ferric-enterobactin. Panel C shows the interacting residues of OPH with ferric-enterobactin at the predicted binding site. Lys82 and Arg85 placed between two of the three catechol rings make hydrogen bond interactions with ferric-enterobactin and the other residues (Asp315, Met314, Phe304, Asp318, Arg319, Arg89 and Pro322) make hydrophobic contacts.
Panels A and B were generated using PyMol 2.0 (Schrodinger, LLC ) and Panel C was generated using LigPlot+ tool (46).