Study on the Docking Analysis of Homocitrate Synthase Responsible for Hydrogen Production in Nostoc

In the case of Nitrogenase-based Hydrogen production, inactivation of the uptake hydrogenase (Hup) leads to significant increase in hydrogen production activity but it lasts only for few hours under the combined n itrogen atmosphere supplied by active hydrogenase. The catalytic FeMo cofactor of nitrogenase binds homocitrate, which is required fo r efficient nitrogen fixation. Hence blocking of Homocit rate synthase will be an efficient strategy for the significant hydrogen production. Nostoc punctiforme ATCC29133, a nitrogen fixing cyanobacterium that has the genes encoding only for the homocitrate synthase and not for uptake hydrogenase. Since the structure of homocit rate synthase is not yet elucidated, it was taken for modeling using Modeler9v5. The inhibitors for homocitrate synthase were identified from literature and were retrieved from Pubchem. The inhib itors identified were 1,10-phenanthroline, 2,2’-Dipyridyl, Hydroxylysine, Iodoacetic acid, Oxaly lglycine, Pipecolic acid, Thialysine, and oxalate were docked with the homocitrate synthase using Argus lab to compare their activit ies. However, among these eight inhibitors screened, 1,10-phenanthroline and Oxalylg lycine were found to have best docking score and proved efficient in terms of both binding affinity and strong hydrogen bond interactions.


Role of Cyanobacteria in Fuel Production
Molecular hydrogen, also called the fuel of the future mainly due to its high conversion efficiency, recyclability, and nonpolluting nature, is mostly controlled by either photosynthetic or fermentative organis ms and is produced by more environment friendly and less energy-intensive processes as compared to thermo-chemical and electrochemical processes. Cyanobacteria, (blue-green bacteria or b lue-green algae) belonging to cyanophyceae, are a large and widespread group of photoautotrophic microorganisms, that fix atmospheric dinitrogen (N 2 ) into ammonia (NH 3 ), a form in which the nitrogen is further available for b iological reactions [1]. Although quite uniform in nutritional and metabolic respects, cyanobacteria are a mo rphologically d iverse group with unicellular, filamentous, and colonial forms [2]. It p ossess several en zy mes th at are invo lved in h yd ro gen metabolism: nitrogenase(s) catalyze the production of hydrogen (H 2 ) conco mitantly with the reduction of nitrogen to ammon ia; an uptake hydrogenase that catalyzes the consumption of hydrogen produced by the nitrogenase; and a bidirectional hydrogenase, which has the capacity to both uptake and produce hydrogen ( Fig.1) [3][4]. In Nitrogenase-based hydrogen production, inactivation of uptake hydrogenase (Hup) leads to significant increase in hydrogen production activity [5]. Ho wever, the h igh-level-activity of the Hup mutants lasts only a few hours under air, a circu mstance which seems to be caused by sufficient amount of combined nitrogen supplied by active nitrogenase.
While the uptake hydrogenase is present in all nitrogen fixing strains tested so far, the distribution of bid irect ional enzy me is not universal (may be present in both nitrogen-fixing and non-nitrogen-fixing cyanobacteria). The mo lecular masses indicated for the uptake hydrogenase subunits are mean values calculated fro m the deduced amino acid sequences of Anabaena strain PCC 7120 , Nostoc strain PCC 73102 , and A. variabilis ATCC 29413 , wh ile the values for the subunits of the bidirectional enzy me are based on data exclusively fro m A. variabilis ATCC 29413 (Tamagnini et al., 2002).

Biochemistry of the Hydrogen Producti on
A metallo-cluster called iron-mo lybdenum cofactor (Fe-Mo cofactor) is believed to provide the substrate binding and reduction site for bio logical nitrogen fixat ion. FeMo cofactor is contained within the nitrogenase MoFe protein, and X-ray crystallographic analysis has revealed that it consists of a metal sulfur core[Fe 7 S 9 Mo] and one molecu le of (R)-ho mocit rate. The core is constructed fro m[MoFe 3 S 3 ] and[Fe 4 S 3 ] sub fragments that are connected by three inorganic sulfide bridges located between pairs of Fe atoms from opposing fragments. The catalytic FeMo cofactor of nitrogenase binds homocitrate, which is required for efficient nitrogen fixation [6][7][8][9][10] Ho mocitrate[(R)-2-hydro xy-1,2,4butane tricarbo xylic acid] is coordinated to the Mo atom through its 2-hydro xy and 2-carbo xyl groups [11][12][13]. Although it is not yet known how the substrate interacts with FeMo cofactor during turnover, the presence of six coordinately unsaturated Fe atoms, as well as the attachment of homocitrate to the Mo ato m, has invited speculation about the nature of substrate binding [6]. It has also been proposed that the carboxylate group coordinated to the Mo atom might serve as a leaving group in a mechanism that activates Mo to provide a substrate coordination site [14].
This enzy me was found to be a Zn-containing metalloenzy me and belongs to the family of transferases, specifically those acyltransferases that convert acyl groups into alky l groups on transfer. The R-carbo xylate and R-o xo groups of R-ketoglutarate are perhaps required fo r optimu m binding to coordinate to the active site Zn [18]. However, its activity can be inhibited by some of the metal-chelating and sulfhydryl binding agents such as 1,10-phenanthroline, Oxalylglycine, 2,2'-Dipyridyl, Oxalate and Iodoacetic acid [19][20]. Moreover, in the in vivo condition, the activity of Ho mocitrate synthase was inhibited by Lysine and its analogues such as Hydroxylysine and Thialysine, which had a heterotrophic effect on 2-ketoglutaric binding sites. Perhaps, a second class of affectors found, included 2-aminoadipic acid, p ipecolic acid and dip icolinic acid, which affected the S-acetyl coenzy me A binding sites in Saccharomyces lipolytica [21].

Criterion for Choosing the Organism
Nostoc punctiforme ATCC29133 (Nostocaceae) is a n itrogen-fixing cyanobacterium that grows autotrophically with CO 2 as the carbon source, utilizing an o xygen-producing photosynthetic mechanism for the generation of ATP and reductant. Interesting, in this organism, heterocyst occupies 3-10% of total cells, which differentiate in response to the lack o f n itrogen in the environ ment and acts as the site of nitrogen fixat ion. A latest release of Cyanobase (http://genome.kazusa.or.jp/cyanobase/NPUN), a geno me database for Cyanobacteria, has revealed Nostoc punctiforme ATCC29133 to have the genes encoding solitarily for Homocitrate synthase and not for Uptake hydrogenase. And moreover, the Kyoto Encyclopedia of Genes and Genomes metabolic pathway database also suggested that lysine biosynthesis in Nostoc sp. to proceeds by the diaminopimelic acid pathway, which does not involve homocitrate instead of α-aminoadipic acid pathway, which does involve homocitrate [5]. Hence b locking of Ho mocitrate synthase would be an efficient strategy for the significant hydrogen production [17].

Objecti ve
With this perspective, the present study was aimed to work on cyanobacterial en zy me alteration to enhance the production of Hydrogen. Since, co mparative modeling is a useful for Hydrogen Production in Nostoc punctiforme ATCC29133 tool in bioin formatics to predict the three dimensional structure of an unknown protein and also aids to understand protein function, the present study was aimed in this direction to model and predict the structure of Homocitrate synthase using Modeller 9v5 for the sequence (Accession No: B2J701) of Nostoc punctiforme ATCC29133, wh ich was obtained fro m Un iProt database. Moreover, the interaction between Homocit rate synthase and its potential inhibitors such as 1,10-phenanthroline, 2,2'-Dipyridyl, Hydro xy lysine, Iodoacetic acid, Oxalylg lycine, Pipecolic acid, Thialysine, and Oxalate was also investigated Insilico by the process of docking.

Retrieval of Target Protein Sequence
The protein sequence of homocitrate synthase from Nostoc punctiforme ATCC29133 was obtained fro m the Uniprot database (Accession No: B2J701) (http://www. uniprot.org/uniprot/B2J701). Since the three d imensional structure could not be ascertained in PDB (http://www.rcsb. org/), an attempt was made to determine the 3D structure.

Te mpl ate Identification
The NCBI BLAST was used to identify the temp late for modeling the three dimensional structure of Ho mocitrate synthase of Nostoc punctiforme ATCC29133, which revealed no suitable template identity. Therefore mu ltip le temp lates (3BLE_A, 2CW 6_A, 3BG9_ B) were identified using 3D-JIGSAW (http://bmm.cancerresearchuk.org/~3d jigsaw/).

Model Generation and Vali dation
The three dimensional structure of Ho mocitrate synthase was predicted using MODELLER9V5 (http://www.salilab. org/modeller/) where the script "align2d.py" was emp loyed to perform an align ment between the target and template sequence and the script "model-default.py" was employed to obtain a rough 3D model which was subjected to energy minimizat ion using the steepest descent technique and Conjugate gradient technique to eliminate bad contacts between protein ato ms through Swiss-Pdb Viewer (http://expasy.org/spdbv/). Perhaps, the backbone conformat ion of the rough model was inspected using the Phi/Psi Ramachandran plot obtained in the PROCHECK server (http://nihserver.mbi.ucla.edu/saves/).

Acti ve Site Predicti on
After obtaining the final model, the possible bind ing sites of Ho mocitrate synthase was searched using Q-site Finder (http://www.modelling.leeds.ac.uk/qsitefinder/) which works by binding the hydrophobic (CH 3 ) probes to the protein, and finding clusters of probes with the most favorable binding energy. The higher cav ity cluster was considered and the residues around the cluster were identified as the binding residues using Pymol. Thus, ten binding sites were obtained for Ho mocitrate synthase and since, the amino acid present in the first active site formed the binding pocket forming a hollow sphere arranged accordingly to the binding of Zinc metal was chosen to study interactions at this site with the chosen ligands.

Protein-Ligand Interacti on Studies
Based on the previous studies [18][19][20][21] in microorganisms such as Saccharomyces cerevisiae and S. lipolytica, the following eight compounds viz., 1,10-phenanthroline; 2,2'-Dipyridyl; Hydro xylysine; Iodoacetic acid; Oxalylglycine; Pipecolic acid; Thialysine; and Oxalate were found to effectively inhibit Ho mocitrate synthase, of which 1,10-phenanthroline; 2,2'-Dipyridyl; Oxalylg lycine and Oxalate were metal chelating agents; while Iodoacetic acid belongs to the sulfhydryl-binding agent; Hydroxy lysine and Thialysine belongs to lysine analogues; and Pipecolic acidan affector were chosen as the inhibitory ligands. The 3D structure of the inhibitors along with the Pubchem ID and Molecular weights (http://www.ncbi.nlm.nih.gov/Structure/ index.shtml) were determined (Table 1) in order to dock with homocitrate synthase in Argus lab 4.0 (http://www.arguslab. com/downloads.htm), with the grid resolution of 0.40 Å where flexible ligand docking was done when the ligand was described as a torsion tree and the grids were constructed to overlay the binding site. Ligands root node (group of bonded atoms that do not have rotatable bonds) was placed on a search point in the binding site and a set of diverse and energetically favourable rotations were created. For each rotation, torsions in breadth-first order were constructed and those poses that survive the torsion search was scored. The N-lowest energy poses were retained and the final set of poses that underwent coarse minimizat ion, re-clustering, and ranking was selected. The file o f the receptor and ligand were uploaded in pdb format and mo l format respectively. Docking precision was set to 'Regular precision' and 'flexible ligand docking' mode was employed for each docking run. The best ligand-receptor structure fro m the docked structure was chosen based on binding affinity and strong hydrogen bond interactions.

Homol ogy Modeling of Homocitrate Synthase
Ho mocitrate synthase of Nostoc punctiforme ATCC29133 constituted 377 aminoacids with the molecular weight of 41 kDa. The absence of three dimensional structures in PDB prompted to construct the 3D model, which provided valuable insight into molecular function and also enabled the analyses of its interactions with suitable inhib itors. Among the four conformations generated, the one with the least modeller objective function value was considered to be thermodynamically stable and chosen for further refinement and validation (Fig 2A). The stereochemistry of the constructed model was subjected to energy min imization to assess the quality of the structure. Ramachandran plot for the model showed 84.9% of the residues in the core region, 14.4% residues in the allowed regions and 0.6%, that is two residues in the disallowed region (Fig 3).  Figure 2. Modeled structure of Homocitrate synthase represented with active site regions. A. Cartoon display of helixes is shown in red color, sheets in yellow and coils in green. B. Ten number of active sites were identified and reported, of which the encircled site indicates the highly conserved active site region, which was chosen for ligand interaction X-axis shows Phi degrees range between -180 to +180 degrees.
Y-axis shows Psi degrees range between -180 to +180 degrees.
A, B, L are the most favoured regions shown in red color a, b, l, p are the additional allowed reg ions shown in Yellow color ~a, ~b, ~l,~p are the generously allowed regions shown in pale yellow colo r.
White color shows the disallowed region. Val 147 and Asn 189 are the t wo aminoacids present in disallowed regions.

Docking of Homocitrate Synthase with Potential Inhi bitors
Eight final docked conformations obtained for the different inhib itors were evaluated indiv idually based on the number of hydrogen bonds formed; bond distance between atomic co-ordinates of the active site and inhib itor and the binding affin ity (Fig 4). The docking energy showed that the inhibitors possessed the best binding affinity to the protein.
Although the active site constituted about 26 residues, the ligands interacted with only 14 residues at THR9, ARG12, ASP13, GLU43, HIS92, GLU136, GLU225, A RG160, ARG162, HIS193, HIS195, ASN217, GLU225 and ASN229 respectively. However, among the chosen inhibitors, the best dock was performed by 1,10-Phenanthroline, wh ich showed the energy of -7.67492Kcal/ mo l with two hydrogen bond interactions to the residue GLU at position 136 ( Fig 4A) and Oxalylg lycine which formed 6 hydrogen bonds ( Fig 4E) but with slightly higher binding energy(-6.46383 Kcal/ mo l), than 1,10-Phenanthroline, with good bond distances. Iodoacetic acid and Oxalate may have least inhibition on protein due to the higher binding values when compared to other ligands and perhaps also showed more number of hydrogen bonds such as 5 and 12 respectively with higher bond lengths (Fig 4D and 4H). The hydrogen bond interactions between the eight inhibitors and Homocitrate synthase along with their bond distances and binding affinit ies are shown in Table 2.

Discussions
The present study demonstrates the presence of genes for Ho mocitrate synthase and absence of genes for uptake hydrogenase in Nostoc punctiforme ATCC29133. Hence in this condition, Nostoc sp. produces hydrogen in excess than the other species having the genes for uptake hydrogenase. However, in spite of the production of enhanced hydrogen, the combined nitrogen at mosphere created by the active nitrogenase leads to the decrease in the net hydrogen production [5] reported that disruption of homocitrate synthase gene could lead to decrease in nit rogen fixation act ivity and increase in hydrogen production activity in Klebsiella pneumoniae. Ho mocitrate synthase produces homocitrate that binds to the FeMo cofactor of the nitrogenase, which is required for the efficient nitrogen fixat ion. Similarly, also in Nostoc sp., Ho mocitrate synthase functions in the biosyn-thesis of homocitrate that bind with the FeMo cofactor of nitrogenase but not with lysine synthesis through the alpha-aminoadipic acid pathway as reported by the same co-workers [5]. In this way, ho mocitrate synthase, a Zn containing metalloenzy me that catalyses the condensation of acetyl CoA and alpha-ketoglutarate to yield ho mocitrate and CoA [18] if inhib ited could probably improve the chances of enhancing the hydrogen production.
Hence, the activity of Ho mocitrate synthase was inhibited by metal-chelating reagents (1,10-phenanthroline, Oxalylglycine, Oxalate and 2,2'-Dipyridyl) and sulfhydryl binding reagent, Iodoacetic acid. In some of the yeast species, Homocitrate synthase is also believed to involve in biosynthetic pathway of lysine, in which Ho mocitrate synthase is inhibited by the lysine through feedback mechanism [21]. Since, some of the lysine analogues viz., Hydro xylysine and Thialysine have the complete inhibitory action on homocitrate synthase and in particular on alpha-ketoglutarate binding sites and that certain effector mo lecule viz., pipecolic acid also affected the co-operativity of S-acetyl coenzyme A binding sites [21], they were used as small mo lecules to examine their interaction on the active site of Homocitrate synthase.
On analyzing the results obtained from the docking of metal chelating agents, was revealed to exhib it two hydrogen bonds and only one hydrogen bond respectively by 1,10-phenanthroline and Dipyridyl respectively. But in the case of Oxalate, it formed 12 hydrogen bonds in which some of the bonds showed higher bond length (more than 3Å) that lead to the weaker interactions. In the case of Sulfhydryl binding agent Iodoacetic acid also formed mo re nu mber of hydrogen bonds (five) like Oxalate, which exhib ited the least effect on protein due to its higher docking energy [22]. In the case of Lysine analogues, Hydro xylysine and Th ialysine formed four and three hydrogen bonds respectively, but at the same time had a good inhibition for protein. The affector, Pipecolic acid formed only one hydrogen bond with the binding energy of -6.49758 Kcal/ mo l, which in spite of good binding energy, could not be considered as best because of least inhibitory action on protein due to its highest docking energy and the least number of hydrogen bond formed. However, based on the comparison of the analysis and with relative to b inding energies, 1,10-Phenanthroline could be identified as the best potent ligand with very least binding energy of -7.67492Kcal/ mol than other inhibitors. Since Ho mocitrate synthase is a Zn-containing metalloenzy me, Zn metal is supposed to bind to the amino acids such as cysteine, Histidine, Aspartic acid, Glutamic acid, Asparagine, and Glutamine (given according to priority) [23][24]. Hence according to the least binding energy and the amino acids binding to the zinc metal, 1,10-Phenanthroline was revealed to be an effective interactor o f the p rotein Ho mocitrate synthase. But according to the number of hydrogen bonds formed and the length of the hydrogen bonds and the amino acids binding to the zinc metal, Oxalylglycine is considered to be the best since it formed six hydrogen bonds of which four were bound to the appropriate aminoacids such as GLU43, HIS92, HIS193 and HIS195 but with slightly higher energy (-6.46383 Kcal/ mol) than 1, 10-Phenanthroline with good bond lengths (less than three).

Conclusions
Since Ho mocitrate synthase is a Zn-containing metalloenzy me, it was shown to be effectively inhibited by the metal chelating agents, of which 1,10-Phenanthroline and Oxalylg lycine showed the higher effect of inhibition when compared to other inhib itors under investigation. Both 1,10-Phenanthroline and Oxalylg lycine showed very good binding energy and interactions with the active site residues compared to the other inhib itors. But basically any insilico analysis is either partial or inco mplete if unproved under experimental (in v itro) conditions. Hence, in the direct ion of understanding the inhibitory mechanis m, the work should be taken up to be imp lemented in wet lab which is our fore-front research (proposal applied to funding agency).