Prolyl oligopeptidase (POP) is considered as an important pharmaceutical target for the treatment of numerous diseases. Despite enormous studies on various aspects of POPs structure and function still some of the questions are intriguing like conformational dynamics of the protein and interplay between ligand entry/egress. Here, we have used molecular modeling and docking based approaches to unravel questions like differences in ligand binding affinities in three POP species (porcine, human and A. thaliana). Despite high sequence and structural similarity, they possess different affinities for the ligands. Interestingly, human POP was found to be more specific, selective and incapable of binding to a few planar ligands which showed extrapolation of porcine POP in human context is more complicated. Possible routes for substrate entry and product egress were also investigated by detailed analyses of molecular dynamics (MD) simulations for the three proteins. Trajectory analysis of bound and unbound forms of three species showed differences in conformational dynamics, especially variations in β-propeller pore size, which was found to be hidden by five lysine residues present on blades one and seven. During simulation, β-propeller pore size was increased by ∼2 Å in porcine ligand-bound form which might act as a passage for smaller product movement as free energy barrier was reduced, while there were no significant changes in human and A. thaliana POPs. We also suggest that these differences in pore size could lead to fundamental differences in mode of product egress among three species. This analysis also showed some functionally important residues which can be used further for in vitro mutagenesis and inhibitor design. This study can help us in better understanding of the etiology of POPs in several neurodegenerative diseases.
Citation: Kaushik S, Sowdhamini R (2011) Structural Analysis of Prolyl Oligopeptidases Using Molecular Docking and Dynamics: Insights into Conformational Changes and Ligand Binding. PLoS ONE 6(11): e26251. https://doi.org/10.1371/journal.pone.0026251
Editor: Vladimir N. Uversky, University of South Florida College of Medicine, United States of America
Received: March 28, 2011; Accepted: September 23, 2011; Published: November 23, 2011
Copyright: © 2011 Kaushik, Sowdhamini. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Funding provided by the Department of Biotechnology, India; Wellcome Trust (United Kingdom) for previous funding to RS, and National Centre for Biological Sciences (Tata Institute of Fundamental Research) for infrastructural support. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Serine proteases are the examples where relationship between homology and substrate specificity is still a paradox. Despite high sequence identity in any two proteases, they can be quite specific towards a given macromolecular substrate . Unlike other traditional serine proteases like trypsin and subtilisins, POP cleaves peptides which are smaller than 30 amino acids in length –. Binding of small peptides to POP is essential for many physiological processes and has gained insights as a target for the treatment of numerous disorders like depression, amnesia, schizophrenia, trypanosomiasis, bipolar affective disorder etc –. A recent study also showed lower plasma POP activity in patients of multiple sclerosis . This peptidase has been implicated in neurodegeneration, as well as in the modulation of the inflammatory response . In spite of enormous studies of role of POP in various diseases the precise biological function of protein is still unknown.
POP is a widely distributed enzyme and has been cloned and isolated from several sources –. The X-ray crystal structure of enzyme shows unique domain architecture with a catalytic α/β hydrolase domain and an unusual β-propeller domain. Propeller domain is based on radially arranged seven-fold repeat of four-stranded antiparallel β sheets. In the case of POPs, this domain is considered to be of the “open-velcro” topology, where first and seventh blades are connected only through hydrophobic interactions. The catalytic triad (Ser 554, His 680, and Asp 641) is hidden and located at the interface of two domains. This unique propeller which is absent in other α/β hydrolases, acts as a lid to hide the active site and also as a selectivity or gating filter, thereby allowing only small peptides to reach active site ; despite central inter-domain cavity, that can accommodate bigger ligands. Various experimental studies have suggested concerted movement of propeller and peptidase domains are necessary for enzyme activity .
Evolutionary studies of POP family shows that plant POP diverge before mammalian POP . Phylogenetic analysis showed that POP is the most conserved enzyme in POP family . In animals, POP is widely distributed with high concentration found in the brain, and its involvement in the control of several mammalian peptide hormones signaling pathways have been studied extensively –. As abnormal POP activity is found to be linked with various neurological disorders, for preclinical trials porcine POP is widely studied as model to identify potential, potent and selective inhibitors . Most of the reported inhibitors entered clinical trials , but their use as drug has not been reported yet. Unlike other members of POP family crystal structure of the POP does not explain the possible passage for substrate/product entry/egress therefore substantial conformational changes are expected. Despite the enormous amount of data from both experimental and computational studies, mechanisms of substrate/inhibitor entry and product egress are still unknown , . Recent crystal structure of bacterial POP (S. capsulata) in open form suggests large opening between two domains for substrate/inhibitor entry, while presence of inhibitor shuts the opening . Comparative analysis of the two structures highlights role of inter-domain dynamics. However, from other crystal structures of closed form of mammalian POPs, porcine POP for instance, have various hydrogen bonds that act like a bridge in connecting two domains along with numerous loops from β-propeller domain which behaves like a covering sheath. Sequence analysis of these two POPs suggests salt bridges, present in bacterial POPs that function as a latch for inter-domain opening movement, is not conserved in porcine . This highlights the fact that bacterial POP may not represent a common and unified mechanism for action of every POP enzyme of other species. So, it is also anticipated that different POP species can have different substrate entry or product egress mechanisms . Previous studies have shown that plant POPs are distant members of same family but till now their function in plants is not known. Unavailability of drugs and no successful clinical trials on human has inspired us to carry out this analysis to better understand the differences of POP among different species if any.
In the present study, we have carried out in-depth analysis and comparison of POPs from three different species human, porcine and plant (A. thaliana) in terms of ligand specificity and binding. This comparison was done to better understand the differences and conformational dynamics of the protein. We have focused on two main issues firstly, to what extant extrapolation of porcine POP to human POP is correct and secondly investigating the possible passage for substrate/product entry/egress. To unravel the above mentioned questions we have applied computational based approaches like molecular docking and dynamics. We were interested in studying plant POPs as they are distant members of same family so it would be interesting to see from evolutionary perspectives like how a distant member of same family behaves as compared to newly diverged members. Moreover, few naturally occurring plant molecules are reported as POP inhibitors having therapeutic applications but nothing much is known about their function and presence in plant system –. Three possibilities of substrate entry were considered: (1) Movement of side chains present at β-propeller pore will allow annulus of tunnel to become broad (2) the first and seventh blade will move apart, thereby increasing pore size (3) Hinge-like motion between two domains causing separation, thereby substrate entry . Further comparison of all the reported cavities (Figure 1) present in protein, that includes β-propeller, inter-domain (between two domains reported by Polgar and coworkers, ) and a recently reported cavity (by Tarrago and coworkers found using cryo-electron microscopy which is present just above active site ) was carried out for unrevealing the possible entry/exit mechanism.
Bottom arrow indicates β-propeller cavity which continues till active site and form another smaller cavity near active site shown in red color (Tarrago et al 2009), inter-domain cavity is shown in green color.
Exhaustive examination and analysis of these POP species showed differences in binding affinities of different inhibitors which include differences in binding in porcine and human POPs. Interestingly, human POP was found to be more selective and specific and also showed hindrance in binding to some ligands which binds well to porcine and A. thaliana, this indicates the extrapolation of porcine to human POP will be difficult in all cases. Detailed analyses of molecular dynamics trajectories of bound/unbound form reveals the dynamic conformational changes associated with the ligand binding in three POP species. Striking differences in β-propeller pore size in bound forms of three POP species were noticed. We have also shown that smaller product movement is possible from β-propeller in porcine, while it may differ in other two species.
Alignment and model preparation
Protein sequence of A. thaliana POP was retrieved from TIGR (The Institute of Genomic Research) database . The sequence obtained was subjected to BLAST (Basic Local Alignment Tool) against PDB (Protein Data Bank) to extract information about suitable structural template and to PSIPRED (Protein Structure Prediction Server) for predicting secondary structural elements, respectively 34–36. Sequence alignment was done using CLUSTALW  and Joy4.0 program was used to annotate the alignment using three dimensional structural information of template . DSSP (Database of secondary structure assignment) was employed for the assignment of secondary structure . Crystal structure of porcine POP (PDB ID: 1E5T) was used as a template for the construction of model. Alignment of query and template was considered to build the model using MODELLER (version 9.1, ). A set of 100 models were generated, from which lower energy structure according to DOPE (Discrete optimized protein energy) score was used for further processes. Geometric inaccuracies of the structural model were evaluated by subjecting the model to PDB-ADIT validation server, which validates using PROCHECK –. The structure was further energy minimized with the SYBYL software package (version 7.1) using Tripos forcefield . For first 500 steps minimization was carried out with steepest descent which was followed by 200 iterations of conjugate gradient with distant dependent dielectric constant equal to 1, non-bonded interaction cutoff value of 8 and was terminated at the convergence of 0.05 kcal mol Å−1. The final structure was validated again using the PROCHECK and PROSA that checks for high energy regions of the modeled structure . Structural model was structurally aligned with the experimental structure and rendered using PyMOL.
The set of ligand molecules studied in this work include known inhibitor Z pro prolinal (ZPR) and related molecules like Z-prolyl pyrrolidine, Z-prolyl prolinol, Z-prolyl azetidine, Z-phenyl alanine azetidine, pyrazinone , . Besides this, naturally occurring plant flavonoid inhibitors baicalein  and berberine  were also used which are known porcine POP inhibitors. These ligands were either downloaded from PubChem or were constructed using CHEMDRAW software. Few other ligands like Y-29794 –, UAMC , ono-1603 –, S-17092 –, SUAM-1221 –, and JTP4819 – were also considered for this study. Coordinates were saved and subjected to CORINA for two dimensional to three dimensional conversions. The three dimensional structure of porcine POP (1QFS), which is ligand bound form with Z pro prolinal was downloaded from PDB. This structure was determined using X ray crystallography with resolution of 2 Å. The energy of ligand molecules were minimized for 200 runs using steepest descent followed by 100 runs of conjugate gradient using SYBYL software. Each of the minimization methods were carried out with Tripos forcefield.
In order to carry out the docking simulation, AutoDock 4.0  was used. It is one of the most suitable methods for performing molecular docking of ligand to their macromolecular receptors. Here rigid docking protocol was considered, where both ligands and receptors were kept rigid. The Graphical User Interface program “Autodock tools” was used to prepare, run and analyze the docking simulations. Polar hydrogen's were added into the receptor PDB file for the preparation of protein in docking simulation. Gasteiger charge was assigned and non-polar hydrogens were merged to the ligands. Autodock requires grid maps, and grid must surround the region of interest in the macromolecule. In this study one of the catalytic triad residues Ser 554 was selected as the active site residue. So, the grid was centered on this catalytic active region of the receptor. The grid box size was set at 60, 60 and 60 Å (x, y and z). For blind docking grid size was increased so that entire protein can be accommodated inside grid where grid dimensions were 68, 62 and 82 Å (x, y and z). AutoGrid program was used to produce grid maps. The spacing between grid points was 0.37 Å and 1 Å for active site and blind docking respectively. The Lamarckian Genetic Algorithm (LGA) was used to search for the best conformers. During the docking process, a maximum of 100 conformers were considered for each compound. The population size was set to 150 and the individuals were initialized randomly. Maximum number of energy evaluation was set to 2.5×10−6, maximum number of generations to 27000, maximum number of top individual that automatically survived to 1, with a mutation rate of 0.02 and a crossover rate of 0.80.
Electrostatic charge distribution
The Adaptive Poisson-Boltzmann Solver (APBS) program was employed to calculate the electrostatic charge distribution of porcine and human crystal structure . Electrostatic charge distribution was also mapped on the theoretical structure of A. thaliana and comparison was made with respect to possible cavities present in the protein. PyMOL was used for the visualization of the surface representation.
Molecular Dynamics Calculations
MD calculations at room temperature were carried out using Gromacs  software with gromos96 force field for porcine (1QFS), human (3DDU)  and the model of A. thaliana POP (generated in this work). Proteins were immersed in cubic water box of spce water model with box edges 9 Å from molecule periphery. 21 sodium ions were added to achieve electroneutrality. The system was subjected to pre-equilibration process that involves minimization using steepest descent for 6000 steps followed by position restrained dynamics for 20 ps. MD calculations were performed at constant volume and temperature for both bound and unbound forms each for 20 ns of three POP species. The system was simulated under periodic boundary conditions with cutoffs of 10 Å each for electrostatic and Van der Waals terms. Replicate of the MD runs were also performed using different seed numbers. Snapshots were collected after every 1 ps. Prodrug server was used to obtain drug parameters . Strong/short (distance between nucleophiles <2.8 Å) hydrogen bonds between two domains were calculated using HBPLUS  for the comparison among species.
Alignment and model preparation
For generating structural model of A. thaliana POP BLAST search was performed, using plant POP as a query against PDB database, suggests POP from porcine brain (PDB ID: 1E5T, resolution 1.7 Å) as a potential template for modeling of plant POP. Plant and porcine POPs are definite homologues with sequence identity of 53% (Figure 2a). Alignment was analyzed to check preservation of conserved residues along with catalytic triad residues present towards C-terminal of the sequence. There was good agreement between the secondary structural elements of both sequences. Catalytic triad was also found to be in similar orientations in theoretical structure as in porcine POP (Figure 2b). Overall, 95% of residues were present in allowed region representing a good model quality.
a) alignment between query (At1g20380) and template (PDB ID: 1E5T, porcine) generated by CLUSTALW, sequence identity was found to be 53%. α/β hydrolase domain is cyan colored and β-propeller is shown in yellow color. Active site residue conservation is shown in red color b) homology model of A. thaliana POP generated using porcine POP as a template. Catalytic triad is shown in magenta color.
Electrostatic charge distribution
Electrostatic charge distribution was mapped on all three POP species for finding species specific surface differences. Analysis of electrostatic charge distribution reveals deviations in surface electrostatics at different cavities present in POP proteins of three species (Figure 3). Comparison was done with respect to known cavities present in porcine protein of known structure to study differences in possible initial binding sites of ligand. In order to check whether there are species-specific differences, we have compared electrostatic distribution on the surfaces of human and A. thaliana POP with porcine POP. Intensity of positive potential was found to be varying in the three protein structures. Striking differences were observed among three species: mouth of the β-propeller cavity was found to have more positive potential in human POP as compared to porcine, while A. thaliana POP shows even higher positive potential than other two because of the presence of an extra Arg402 (Figure 3a, Figure S1). These differences were found to be concentrated in loop regions of first and seventh blades of β-propeller. Blade 1 of both porcine and human POP was having high positive surface potential. Further, the comparison of cavity present just above the active site by Tarrago and coworkers  suggests that surrounding residues have higher positive potential in humans, as compared to porcine, while it was found to be near-neutral in A. thaliana (Figure 3b). These differences in the protein electrostatics may lead to difference in the recognition and initial binding of the ligands in three species.
a) down the β-propeller pore view of electrostatic potential. Three species showed differences in electrostatics. A. thaliana POP showed more positive potential; similarly differences were also present in human and porcine POP proteins b) Electrostatic potential of inter-domain cavity and recently reported cavity (Tarrango et al 2009). In A. thaliana later cavity was hidden.
Most of the POP inhibitors contain a proline skeleton (exactly a proline group or proline analogs). The first effective inhibitor which was discovered was Z-pro prolinal. The set of ligand molecules studied in this work include Z-pro prolinal, its related compounds and various other molecules which were tested for preclinical trials (Y-29794, UAMC, ono-1603, S-17092, SUAM-1221, and JTP4819) in other organisms like rat, porcine etc. Molecular docking simulations were conducted using AutoDock4. For all the docking runs, minimum energy conformation was selected.
In order to identify residues potentially involved in ligand binding and to analyze its functional importance, the receptor was docked to different ligands. Initially, rigid docking approach was used. Each docking pose was analyzed and compared in three species. Ligands were found to be docked in different conformations and with varying affinities, as reflected by docking scores (Figure 4). Interestingly, few of the ligands Y-29794, UAMC, ono-1603, S-17092, thioxo and Z-pro prolinol showed some hindrances in binding to human POP as revealed by their positive binding energies (Figure 4a).
a) Rigid docking scores of human, porcine and A. thaliana POP. Some of the ligands showed hindrance in binding to human POP as revealed by their positive energy scores. Different ligands showed difference in binding affinities to three species of POP. b) Flexible docking scores of human, porcine and A. thaliana POP. After inducing flexibility in ligands still two of the ligands were incapable of binding to human POP.
For ligands that showed no evidence of binding (positive energies) using rigid docking approach, docking runs were repeated after subjecting the ligands to flexible docking, where flexibility was induced in ligands alone. This was done in order to provide more optimal interactions between protein and ligand during the docking. Ligands were again found to bind with different affinities in three POPs. Some of the ligands that were showing poor scores in rigid docking were found to bind well when the ligand conformation was permitted to be flexible. However, still two planar ligands Y-29794, UAMC were incapable of binding. This binding incapability suggests difference in selectivity/affinity of ligands for binding to human POP (Figure 4b). Their binding orientation was found to be different than porcine and A. thaliana POP. In Y-29794, change in orientation of binding leads to interaction with more polar residues like Thr481, Ser479, Tyr471and Asn483 which causes hindrance in binding while in porcine smaller residues like Ala and Gly are present. Similarly, UAMC binds in different orientation causing short contact in human POP. Plant based naturally occurring animal POP inhibitors baicalein and berberine were found to bind with high affinities to A. thaliana POP protein too, which shows they are effective ligands but their function in plants is still not known. Other ligands were also found to bind well to A. thaliana POP, but whether their function (inhibitor) is similar in plants is still an area to be explored.
In order to identify possible regions of protein where ligand (Z-pro-prolinal) can bind effectively, blind docking approach was used. Grid was increased to dimensions which can accommodate entire protein inside and then comparisons among three POPs were carried out. In porcine POP, by following the various docking poses, it appears that the ligand can make specific path right from mouth of β-propeller to active site of protein, while it was not very well distributed and found to be present at specific places like end of α/β hydrolase domain and in middle of β-propeller cavity in human (Figure S2). In order to validate whether binding is not random, dummy molecule sucrose was also docked to porcine POP (Figure S3) as a negative control. Sucrose was found to bind all over protein (unspecific binding) while ligand ZPR had more specific binding. In human POP, ligand binding was restricted to some particular regions only, though CASTp  results showed β-propeller cavity is more voluminous in human (11323 Å3) as compared to porcine (10471 Å3). Besides this, binding of Z-pro-prolinal was more feasible in human (binding energy −10.1 to 3.1 kcal/mol), as compared to porcine (binding energy −7.1 to −4.3 kcal/mol) and A. thaliana (binding energy −5.2 to −3.3 kcal/mol) POPs.
All short peptides with an internal proline residue are potential POP substrate. Blind docking of POP with substrate angiotensin IV (V-Y-I-H-P-F) was also carried out to known possible regions of binding . This was found to be similar (data not shown) as when inhibitor Z-pro-prolinal was docked. Binding was restricted to some particular regions in human, unlike porcine POP, but binding feasibility was higher (energy/residue −3.3 kcal/mol in human, −1.9 kcal/mol in porcine). Similarly, N terminal (V-Y-I-H-P) and C terminal (F) products were docked to human and porcine POP. C-terminal product was docked successively after docking N-terminal product. Comparison of percentage docked poses of substrate and product showed some differences with respect to known cavities and domains (Figure S4). It was found that substrate has tendency to bind at particular regions like β-propeller cavity, functionally important loop (residues 192–205) and towards α/β hydrolase domain of protein while products were present at multiple sites which were reported till now including inter-domain cavities reported by Polgar and coworkers  and Tarrago and coworkers , reflecting its multiple egress sites.
Overall, nine simulations were run, with a total duration of ∼0.21 µs. However, rather than running a single long simulation, we explored the effect of presence/absence of ligand (ZPR, each 20 ns). The rationale behind these multiple simulations was to understand differences in dynamics of POP protein in three species, also in the ligand unbound and bound conformations.
Briefly simulations performed with replicates were:
- Six simulations each 20 ns of bound/unbound form of human, porcine and A. thaliana.
- One simulation (20 ns) for unbound form of human when lysines were mutated.
- One 45 ns simulation where ligand was at mouth of β-propeller pore in porcine POP.
Simulation with bound ZPR/unbound forms.
Conformational stability was assessed by the drift of the protein from the crystal structure as measured by root-mean-square deviation (rmsd) of Cα atoms from their initial coordinates (Table 1). During simulation, all the three proteins were found to be stable (Figure 5). Unbound form of A. thaliana showed higher deviations as compared to porcine and human POP, overall β-propeller domain showed higher fluctuations than α/β hydrolase domain (Figure S5). In the ligand-bound form, rmsd of the drug position showed more variability and fluctuations in binding pocket of porcine POP hence more mobility of drug as compared to human and A. thaliana (Figure S6). Rg plots showed a slight decrease (∼0.3 Å) in bound forms of POPs. Flexibility of residues was seen by measuring rmsd during simulation. Like other studies done before (Polgar and coworkers  and Tarrago and coworkers ), we also found that loops present at inter-domain region showed high fluctuations; besides, residues 266–275 of β-propeller blade 4 also showed very high fluctuation, along with other loops connecting β-strands of blade 1, 2, 5 and 7. This, together with the higher positive potential at the propeller mouth of A. thaliana POP, might suggest that the substrate entry is more likely by inter-domain movements.
a) rmsd of bound form of porcine (black), human (red) and A. thaliana (green) POPs b) rmsd of unbound form of porcine, human and A. thaliana POPs. Dotted line indicates rmsd of replicate runs.
Conformational changes during simulation.
Comparison among three POP species was done with respect to known cavities present in structure to understand the possible passage of substrate entry and product egress. Bound form simulations revealed opening and closing of present cavities of proteins in three POP species. In porcine and human bound form POP during initial 4 ns of simulation, inter-domain cavity (Polgar and coworkers ) was present, but after that it was hidden continuously till the end of simulation. Similarly, cavity present in continuation with β-propeller (Tarrago and coworkers ) also followed the same trend. On the other hand, in unbound forms, above mentioned cavities were hidden revealing incapability of these cavities in absorbing a substrate and hence its accessibility to active site. Interestingly, we found that β-propeller pore size showed variations in diameter. During simulation, some species-specific striking differences in pore size of β-propeller in three POPs, especially in bound form were noticed. We discovered that in porcine there was continuous increase in pore diameter size from ∼5.5 Å to ∼7.3 Å during simulation while in human (∼4.5 Å to ∼5.0 Å) and A. thaliana (∼3.0 Å to 4.2 Å) POP there were not high changes in pore size (Figure 6, Figure 7 and Figure S7). Overall, during simulation β-propeller pore size was increased for porcine, while for human POP it does not show any substantial changes. On the other hand, pore size was decreased continuously for A. thaliana POP which shows possibility of some alternate role of β-propeller in plant systems. Pore sizes in unbound forms of POP of human and porcine were smaller in diameter than respective bound forms. Inter-domain distance (Pro34-Thr200) in porcine and human unbound form was increased by 2 Å and 3 Å respectively while A. thaliana was also amenable to such changes. This shows domain opening motion while such high differences were not present in bound forms (Table S1). During simulation distance between blade 1 and 7 (Cys78-Glu397) was decreased in both bound/unbound forms of porcine and human suggesting the incapability of movement of blades of propeller (Table S2).
During simulation of bound form of porcine, human and A. thaliana POP. In porcine diameter of pore was found to be increasing in while human and A. thaliana it was found to be decreasing (Figure S7 for replicate runs).
β-propeller pore size was found to be biggest in porcine, while in human and A. thaliana it was small. Drug (ZPR) is shown in magenta color, active site is colored red.
As both domains are found to be connected via series of hydrogen bonds, we checked how hydrogen bonding interactions are changing during simulation including replicate runs. In unbound form, both porcine and A. thaliana showed similar trend of loosening of interactions between two domains as number of strong hydrogen bonds were found to be reduced (Table S3) in 20th ns structure, while in human unbound form, more strong bonds were formed which showed its dissimilar trend as compared to porcine and A. thaliana. All bound forms were found to be more intact during simulation as shown by increase in number of strong hydrogen bonds during simulation which showed that, in substrate bound condition release of product from opening of two domains is very unlikely. We also noticed changes in strong hydrogen bonds between blade one and seven in the unbound form. Bound form of porcine was more relaxed having less number of strong hydrogen bonds after 20 ns than human and A. thaliana POPs.
Interplay of lysines present on blade one and seven.
During simulation studies of POP we discovered that β-propeller pore was hidden by side chains of lysine residues, three of which are present on blade1 (Lys81, Lys82, Lys84, porcine residue numbering) and two on blade7 (Lys389 and Lys390, porcine residue numbering). Movement of the side chains of these lysines causes fluctuations in β-propeller pore size. Therefore, the conservation of these lysine residues was checked across different species to know if they are present and conserved universally. For this multiple sequence alignment of POPs from different lower to higher organisms (bacteria, archaebacteria, mammals, amphibians, nematodes and plants) was carried out. Interestingly, alignment shows differences in conservation of lysine among different species. From all the species considered for the analysis lysines were totally conserved in mammals and amphibians as shown in Figure 8. In human POP, we observed continuous hindrance in β-propeller pore opening because of side chains of lysines therefore smaller pore size as compared to porcine. This observation leads us to mutate all lysines present on mouth of β-propeller with smaller positively charged histidine. The only intention to mutate lysine with histidine was not to disturb charge distribution on the surface of protein. We found that after mutation pore size was increased to 5.5 Å than wild type (4.5 Å). Further to see dynamics of mutated protein, we subjected the mutant (unbound) to undergo 20 ns simulation like wild type protein. Analysis of trajectory of mutated human POP showed increase in diameter from 5.5 to 8.1 Å (Figure 9). This analysis showed that other species of POP in which lysines are absent and are replaced by smaller amino acids like Ile, Val, Ser etc can have bigger pore size which can lead to differences in substrate entry or product egress. On the other hand if lysines are replaced by longer length amino acids like arginine (e.g. A. thaliana) which in turn can contribute in decreasing or completely hiding the β-propeller pore opening.
Lysines were conserved in all mammalian species and also in amphibian Xenopus. Coloring: Archaebacteria (yellow), mammals (dark grey), amphibian (blue), plants (orange), arthropods (blue, below orange), nematodes (light yellow), bacteria (light grey). Presence of lysines are shown using red color, while absence using green color.
Ligand (ZPR)-interacting residues (present within 5 Å from any one of the ligand atoms) were examined, during the course of simulation of the ligand-bound form, to see if changes in interaction with the ligand happen in comparison to the initial structure. Hydrogen bond interactions of ligand with the protein were compared with porcine POP crystal structure. Surprisingly, neither of the species retained the same binding mode as seen in X-ray structure as after 20 ns orientation of the ligand in three POPs was slightly different than the initial structure (Figure S8). Residues present within 5 Å of the binding site can be functionally important as they play an important role in biological activity of the molecule and also take part in biological interactions hence mapping of these interactions was carried out. Hydrogen bonding interactions with Tyr473, Trp595, and Arg641 were observed. Importantly, these interactions between the ligand and interacting residues were not conserved in three POP species (see Table 2 for details). After 20 ns, numbers of hydrogen bonds were found to decrease in porcine and human and we also noticed few new interacting residues coming within 5 Å region of drug. We also noticed differences in interacting hydrophobic residues among the three POPs. Some of the residues, like Ile-591 and Phe-476, retain highly conserved interactions during the 20 ns simulation in both porcine and A. thaliana POPs, while only Ile590 was conserved in human POP suggesting that rational mutations of the above residues could be carried out to study the functional importance of these residues.
Simulation from docked poses.
Results of blind docking showed the capability of ligand to bind at multiple sites of POP. So, from blind docking, two docked complexes were selected (firstly, when ligand was at the mouth of β-propeller, secondly, when ligand was at the inter-domain cavity), for molecular dynamics studies to better understand the passage for ligand entry/egress. A 45 ns simulation of porcine POP docked complex was performed, where ligand was at the mouth of β-propeller. Distance plot between centre of mass of drug and nucleophilic Ser554-OH was plotted to see drift towards active site (Figure 10). We observed that distance was reduced to more than 1 Å which depicted potentiality of β-propeller in ingesting a ligand. However, no such distance reduction could be observed when the ligand was in one of the other docked pose (like at inter-domain, recently reported cavity) when 20 ns simulation was carried out. In latter case, distance was decreased in initial few picoseconds only and was constant throughout 20 ns run (data not shown). A distance reduction of >1Å in 45 ns simulations may not be considered significant enough and longer length simulations could be carried out in future to understand the theory of ligand ingestion from β-propeller.
a) 0th ns structure (protein: green, drug: red) and 20th ns structure (protein: blue, drug: magenta) were superimposed and distance plot of catalytic Ser-554 hydroxyl group and centre of mass of drug was plotted. Distance was found to decrease by more than 2 Å. B) distance plot.
Smaller product movement.
Cleavage of any small peptide like angiotensin-IV by POP produces two N- and C-terminal products. SLITHER  generates contiguous conformations of ligand across the tunnel. Docking of smaller product in porcine POP (C-terminal) by SLITHER showed that its egress can be through β-propeller pore. Different snapshots generated from molecular dynamics run were used to see movement of C-terminal product from β-propeller pore opening. Further, comparison of free energy profiles was done for different snapshots and it was observed that at 0 ns free energy profile generated showed major barrier at mouth of β-propeller pore while after 20 ns simulated structure having bigger pore size, this barrier was reduced which confirms that smaller product can egress from β-propeller pore (Figure 11). These changes of energetic profiles upon the conformational changes provide an explanation why some conformational changes are beneficial for the entrance or exit of substrate/product.
Differences in structural dynamics among proteins are of key importance which can influence the entire thermodynamics and in turn their biological responses. We have applied molecular docking and dynamics based approaches for exhaustive comparison of prolyl oligopeptidase of three species including porcine, human and A. thaliana. POP is an important therapeutic target and is found to be associated with different disorders due to which hundreds of compounds have been tested for POP inhibitory activity. Porcine POP is a well-known model system on which numerous studies have been carried out. Many of POP inhibitors reached preclinical trial stages but till now drugs are not available for public use. This raises the question that to what extent extrapolation of porcine POP to human POP is correct? Plant POPs are poorly understood distant members of same family which diverged a bit earlier than human and porcine POPs. Therefore, this study was carried out to structurally compare porcine, human and plant POPs in detail at two levels: 1) To identify how ligand binding affinities differ among them 2) To compare their structural dynamics which was seen with both bound and unbound forms to envisage the possible path for substrate entry/product egress and to investigate differences among species if any.
Crystal structure and homology model of prolyl oligopeptidase from different species provides an opportunity to examine the differences in structural scenario of three POP species. These POP species were found to have high structural similarities, although plant POPs are evolved before animal POPs and are distant members of same family. Even though sequence and structural similarity among these species are high, their ligand binding affinities are very different. Further, the level of sequence similarities between the three POPs that we chose to study provides a unique opportunity to perform analysis of amino acid changes but in a structural context. At such high sequence identities, the homology models derived are of high quality (Sali and coworkers , Figure S10) enough to make gross comparisons of spatially proximate residues and conformational changes as evidenced by molecular dynamics. Moreover, by means of a complete threading of a plant POP on porcine POP template, we are able to compare the sequence-specific changes without dependence on the backbone and overall fold of the protein. We also noticed that some of the ligands were incapable of binding to human POP which indicates high specificity and selectivity of human POP. Moreover, ligand binding was found to be more feasible in human POP. Naturally occurring plant based animal inhibitors like berberine, baicalein were observed to bind well to plant POP which showed their binding capacity but it is yet to characterize whether they are plant POP inhibitors or not, and if yes, how they are localized and regulated in plant system? This is still a fascinating area to be explored in detail.
Based on the detail structural analysis of POPs, we suggest that dynamic conformational changes are associated with the ligand binding in plant and animal POPs. Loosening of interactions between β-propeller and α/β hydrolase domains indicates plants and animal POPs also follow similar trend of domain opening motion for substrate entry as previously observed in bacterial POPs. However, our bound form simulations depicted the intactness of two domains revealing the release of product from some alternate pathway. We investigated that in porcine POP, β-propeller pore size increases in diameter by ∼2 Å indicating smaller product release can be from this route as free energy barrier was reduced. But this increase in β-propeller pore size in bound form was not observed in human and A. thaliana. Similarly, analysis of strong hydrogen bond interactions suggests intactness of human POP during both bound and unbound form of simulations. These differences showed difference in free energies for opening motion of two domains in different species. However, we also suspect that these differences in strong hydrogen bonds and β-propeller pore size variation could lead to differences in routes of substrate entry and product release. We found that lysines present on blade one and seven continuously hide the β-propeller pore opening. This effect was more prevalent in human bound form than porcine which leads to broader opening of β-propeller in porcine. It is yet to be figure out that the opening of β-propeller pore is a pH dependent mechanism or not? It is possible that because of the change in pH lysines present on blade one and seven will allow broadening of the annulus of the tunnel and thereby movement of small molecules. In porcine this broad opening of pore acts as a passage for the smaller product movement. But we suspect whether this opening could allow larger product movement? Answering this requires still longer length simulations of both bound and unbound form of proteins. This study also showed that in POPs mode of substrate entry/product egress are via different paths as reported in other POP family members.
Sequence analysis showed that in some species like nematodes, arthropods and even in plants, lysines covering β-propeller pore are absent which would increase diameter of β-propeller pore and eventually can even allow bigger product movement. To our knowledge, this is the first study that reflects role of β-propeller in product movement. As human POP was found to be incapable of binding to few ligands like UAMC and Y-29794, molecular docking with 20th ns structure of human unbound form was re-performed. This showed that after subjecting protein to dynamic state, ligand binding is possible as both the above ligands which showed hindrance in binding were found to bind well due to change in the orientation of the binding as shown in Figure S9 (rigid docking score of UAMC:−5.94 kcal/mol, Y-29794: −4.26 kcal/mol). This depicts that conformational changes are necessary especially at the active site of protein.
This analysis showed differences in conformational dynamics of three POP species with respect to ligand binding and product release. On the whole, the current analysis confirms that the ligand affinities and mode of product egress are different for the plant, porcine and human POPs suggesting that studies with potential drugs have to be carefully interpreted. This work is likely to have implications for the design of inhibitors acting at POP and could contribute to development and maturation of novel leads for neurodegenerative diseases. This analysis will enhance our understanding of evolutionary history of POP protein and will also contribute to understanding of function of the enzyme.
Electrostatic surface potential of β-propeller domain of three species. a) Porcine b) Human c) A. thaliana POPs. The surface between blades 1 and 7 is shown by a black arc for clarity. The blade one of β-propeller is more positively charged in porcine and human as compared to A. thaliana. Presence of an extra Arg402 on mouth of β-propeller pore of A. thaliana makes it more positive.
Blind docking results of three POPs using Z-pro prolinal as an inhibitor. a) porcine b) human c) A. thaliana POP. Inhibitor shown in red color was sampled 100 times.
Blind docking of a dummy molecule. Blind docking was done to porcine POP to validate whether binding of substrate is random or not. Sucrose was found to have affinity all over the protein while substrate (angiotensin) binding was limited to some expected places. Figures below show comparison of binding poses of sucrose and angiotensin.
Percentage docked poses of substrate and product. Docked poses were compared between porcine and human POP a) In porcine POP, out of hundred sampling substrate was not found to bind at inter-domain cavity and also newly reported cavity while product was found to bind all over except inter-domain cavity b) In human POP, substrate has tendency to bind in similar way to porcine POP while product has affinity to bind all over protein.
Backbone (Cα) RMSD of two domains. RMSD of β-propeller (black) and α/β-hydrolase (red) domains separately in porcine POP. RMSD of entire protein (green) is also represented. β-propeller showed higher fluctuations then α/β-hydrolase.
RMSD of drug present in binding pocket during bound form simulation of porcine and human POP. a) drug in porcine POP binding pocket b) drug in human POP binding pocket.
Figure shows β-propeller pore size variations during molecular dynamics replicate runs. Overall, porcine POP shows increase in β-propeller pore size while human, A. thaliana POP these changes were not very high. A. thaliana POP shows slight decrease in pore size.
Superimposed structures of porcine, human and A. thaliana POPs before (0 ns) and after MD (20 ns). Protein shown in green color is 0 ns structure while in blue color is 20 ns structure. Drug (ZPR) is shown in tint (0 ns) and magenta color (20 ns). Figure shows the difference in orientation of drug after 20 ns in all three species.
Docking results showing mode of binding of ligands in 0 ns (green) and 20 ns (blue) structures of human POP. a) UAMC bound to human POP, 0 ns complex is shown green(receptor) and tint (UAMC) color, short contact with Ile-478 is shown in yellow circle b)Y-29794 bound to human POP (b.i) 0 ns structure. (b.ii) 20 ns structure. Residues present in vicinity (4 Å) are shown in yellow color.
Secondary structure alignment of porcine POP (1E5T) used as a template for building the model structure of A. thaliana. It clearly shows the agreement of secondary structure elements between two protein sequences. Yellow and pink color indicates beta-strands and alpha helices respectively.
Distance between two domains. This distance was calculated using Pro34 and Thr200. Distances of replicate runs are shown in brackets.
Distance between blades one and seven during simulation (Cys78 and Glutamine 397). Distances of replicate runs are shown in brackets.
Residues involved in strong hydrogen bond formation. a) Number of strong hydrogen bonds during simulations of bound and unbound forms of porcine, human, A. thaliana b) Strong inter-domain hydrogen bonds in porcine, human, A. thaliana POPs in unbound form. Comparison was done between 0 ns and 20 ns structures c) Strong inter-domain hydrogen bonds in bound form of POP species. Residues in bold indicates same hydrogen bond interactions found during replicate runs.
Conceived and designed the experiments: RS. Performed the experiments: SK. Analyzed the data: SK RS. Contributed reagents/materials/analysis tools: SK. Wrote the paper: SK. Improved the manuscript: RS.
- 1. Serrano SMT, Maroun RC (2005) Snake venom serine proteinases: sequence homology vs. substrate specificity, a paradox to be solved. Toxicon 45(8): 1115–1132.SMT SerranoRC Maroun2005Snake venom serine proteinases: sequence homology vs. substrate specificity, a paradox to be solved.Toxicon45811151132
- 2. Walter R, Shlank H, Glass JD, Schwartz IL, Kerenyi TD (1971) Leucylglycinamide released from oxytocin by human uterine enzyme. Science 173(999): 827–829.R. WalterH. ShlankJD GlassIL SchwartzTD Kerenyi1971Leucylglycinamide released from oxytocin by human uterine enzyme.Science173999827829
- 3. Polgár L (2002) The prolyl oligopeptidase family. Cell Mol Life Sci 59(2): 349–362.L. Polgár2002The prolyl oligopeptidase family.Cell Mol Life Sci592349362
- 4. Walter R, Yoshimoto T (1978) Postproline cleaving enzyme: kinetic studies of size and stereospecificity of its active site. Biochemistry 17(20): 4139–4144.R. WalterT. Yoshimoto1978Postproline cleaving enzyme: kinetic studies of size and stereospecificity of its active site.Biochemistry172041394144
- 5. Maes M, Goossens F, Scharpé S, Meltzer HY, D'Hondt P, et al. (1994) Lower serum prolyl endopeptidase enzyme activity in major depression: further evidence that peptidases play a role in the pathophysiology of depression. Biol Psychiatry 35(8): 545–552.M. MaesF. GoossensS. ScharpéHY MeltzerP. D'Hondt1994Lower serum prolyl endopeptidase enzyme activity in major depression: further evidence that peptidases play a role in the pathophysiology of depression.Biol Psychiatry358545552
- 6. Williams R (2004) Prolyl oligopeptidase and bipolar disorder. Clinical Neuroscience Research 12; 4(3–4): 233–242.R. Williams2004Prolyl oligopeptidase and bipolar disorder.Clinical Neuroscience Research12; 43–4233242
- 7. Maes M, Goossens F, Scharpé S, Calabrese J, Desnyder R, et al. (1995) Alterations in plasma prolyl endopeptidase activity in depression, mania, and schizophrenia: effects of antidepressants, mood stabilizers, and antipsychotic drugs. Psychiatry Res 58(3): 217–225.M. MaesF. GoossensS. ScharpéJ. CalabreseR. Desnyder1995Alterations in plasma prolyl endopeptidase activity in depression, mania, and schizophrenia: effects of antidepressants, mood stabilizers, and antipsychotic drugs.Psychiatry Res583217225
- 8. Tenorio-Laranga J, Coret-Ferrer F, Casanova-Estruch B, Burgal M, García-Horsman JA (2010) Prolyl oligopeptidase is inhibited in relapsing-remitting multiple sclerosis. J Neuroinflammation; 7: 23.J. Tenorio-LarangaF. Coret-FerrerB. Casanova-EstruchM. BurgalJA García-Horsman2010Prolyl oligopeptidase is inhibited in relapsing-remitting multiple sclerosis.J Neuroinflammation;723
- 9. Kanatani A, Yoshimoto T, Kitazono A, Kokubo T, Tsuru D (1993) Prolyl Endopeptidase from Aeromonas hydrophila Cloning, Sequencing and Expression of the Enzyme Gene and Characterization of the Expressed Enzyme. J Biochem (Tokyo) 113: 790–796.A. KanataniT. YoshimotoA. KitazonoT. KokuboD. Tsuru1993Prolyl Endopeptidase from Aeromonas hydrophila Cloning, Sequencing and Expression of the Enzyme Gene and Characterization of the Expressed Enzyme.J Biochem (Tokyo)113790796
- 10. Rennex D, Hemmings BA, Hofsteenge J, Stone SR (1991) cDNA cloning of porcine brain prolyl endopeptidase and identification of the active-site seryl residue. Biochemistry 30(8): 2195–2203.D. RennexBA HemmingsJ. HofsteengeSR Stone1991cDNA cloning of porcine brain prolyl endopeptidase and identification of the active-site seryl residue.Biochemistry30821952203
- 11. Shirasawa Y, Osawa T, Hirashima A (1994) Molecular cloning and characterization of prolyl endopeptidase from human T cells. J Biochem 115(4): 724–729.Y. ShirasawaT. OsawaA. Hirashima1994Molecular cloning and characterization of prolyl endopeptidase from human T cells.J Biochem1154724729
- 12. Vanhoof G, Goossens F, Hendriks L, De Meester I, Hendriks D, et al. (1994) Cloning and sequence analysis of the gene encoding human lymphocyte prolyl endopeptidase. Gene 149(2): 363–366.G. VanhoofF. GoossensL. HendriksI. De MeesterD. Hendriks1994Cloning and sequence analysis of the gene encoding human lymphocyte prolyl endopeptidase.Gene1492363366
- 13. Ishino T, Ohtsuki S, Homma KI, Natori S (1998) cDNA cloning of mouse prolyl endopeptidase and its involvement in DNA synthesis by Swiss 3T3 cells. J Biochem 123(3): 540–545.T. IshinoS. OhtsukiKI HommaS. Natori1998cDNA cloning of mouse prolyl endopeptidase and its involvement in DNA synthesis by Swiss 3T3 cells.J Biochem1233540545
- 14. Yoshimoto T, Miyazaki K, Haraguchi N, Kitazono A, Kabashima T, et al. (1997) Cloning and expression of the cDNA encoding prolyl oligopeptidase (prolyl endopeptidase) from bovine brain. Biol Pharm Bull 20(10): 1047–1050.T. YoshimotoK. MiyazakiN. HaraguchiA. KitazonoT. Kabashima1997Cloning and expression of the cDNA encoding prolyl oligopeptidase (prolyl endopeptidase) from bovine brain.Biol Pharm Bull201010471050
- 15. Ohtsuki S, Homma K, Kurata S, Natori S (1997) Molecular cloning of cDNA for Sarcophaga prolyl endopeptidase and characterization of the recombinant enzyme produced by an E. coli expression system. Insect Biochem Mol Biol 27(4): 337–343.S. OhtsukiK. HommaS. KurataS. Natori1997Molecular cloning of cDNA for Sarcophaga prolyl endopeptidase and characterization of the recombinant enzyme produced by an E. coli expression system.Insect Biochem Mol Biol274337343
- 16. Fülöp V, Böcskei Z, Polgár L (1998) Prolyl oligopeptidase: an unusual beta-propeller domain regulates proteolysis. Cell 94(2): 161–170.V. FülöpZ. BöcskeiL. Polgár1998Prolyl oligopeptidase: an unusual beta-propeller domain regulates proteolysis.Cell942161170
- 17. Szeltner Z, Rea D, Juhász T, Renner V, Fülöp V, et al. (2004) Concerted structural changes in the peptidase and the propeller domains of prolyl oligopeptidase are required for substrate binding. J Mol Biol 340(3): 627–637.Z. SzeltnerD. ReaT. JuhászV. RennerV. Fülöp2004Concerted structural changes in the peptidase and the propeller domains of prolyl oligopeptidase are required for substrate binding.J Mol Biol3403627637
- 18. Venäläinen JI, Juvonen RO, Männistö PT (2004) Evolutionary relationships of the prolyl oligopeptidase family enzymes. Eur J Biochem 271(13): 2705–2715.JI VenäläinenRO JuvonenPT Männistö2004Evolutionary relationships of the prolyl oligopeptidase family enzymes.Eur J Biochem2711327052715
- 19. Goossens F, De Meester I, Vanhoof G, Scharpé S (1996) Distribution of prolyl oligopeptidase in human peripheral tissues and body fluids. Eur J Clin Chem Clin Biochem 34(1): 17–22.F. GoossensI. De MeesterG. VanhoofS. Scharpé1996Distribution of prolyl oligopeptidase in human peripheral tissues and body fluids.Eur J Clin Chem Clin Biochem3411722
- 20. Cunningham DF, O'Connor B (1997) Proline specific peptidases. Biochimica et Biophysica Acta 1343: 160–186.DF CunninghamB. O'Connor1997Proline specific peptidases.Biochimica et Biophysica Acta1343160186
- 21. Williams RSB, Cheng L, Mudge AW, Harwood AJ (2002) A common mechanism of action for three mood-stabilizing drugs. Nature 417(6886): 292–295.RSB WilliamsL. ChengAW MudgeAJ Harwood2002A common mechanism of action for three mood-stabilizing drugs.Nature4176886292295
- 22. Atta-ur-Rahman , Akhtar MN, Choudhary MI, Tsuda Y, Yasin A, et al. (2005) New diterpene isopimara-7,15-dien-19-oic acid and its prolyl endopeptidase inhibitory activity. Nat Prod Res 19(1): 13–22.Atta-ur-RahmanMN AkhtarMI ChoudharyY. TsudaA. Yasin2005New diterpene isopimara-7,15-dien-19-oic acid and its prolyl endopeptidase inhibitory activity.Nat Prod Res1911322
- 23. Tarragó T, Sabidó E, Kogan MJ, de Oliveira E, Giralt E (2005) Primary structure, recombinant expression and homology modelling of human brain prolyl oligopeptidase, an important therapeutic target in the treatment of neuropsychiatric diseases. J Pept Sci 11(5): 283–287.T. TarragóE. SabidóMJ KoganE. de OliveiraE. Giralt2005Primary structure, recombinant expression and homology modelling of human brain prolyl oligopeptidase, an important therapeutic target in the treatment of neuropsychiatric diseases.J Pept Sci115283287
- 24. Tarragó T, Kichik N, Claasen B, Prades R, Teixidó M, et al. (2008) Baicalin, a prodrug able to reach the CNS, is a prolyl oligopeptidase inhibitor. Bioorg Med Chem 16(15): 7516–7524.T. TarragóN. KichikB. ClaasenR. PradesM. Teixidó2008Baicalin, a prodrug able to reach the CNS, is a prolyl oligopeptidase inhibitor.Bioorg Med Chem161575167524
- 25. Tarrago T, Kichik N, Seguí J, Giralt E (2007) The natural product berberine is a human prolyl oligopeptidase inhibitor. ChemMedChem 2(3): 354–359.T. TarragoN. KichikJ. SeguíE. Giralt2007The natural product berberine is a human prolyl oligopeptidase inhibitor.ChemMedChem23354359
- 26. Lawandi J, Gerber-Lemaire S, Juillerat-Jeanneret L, Moitessier N (2010) Inhibitors of prolyl oligopeptidases for the therapy of human diseases: defining diseases and inhibitors. J Med Chem 53(9): 3423–3438.J. LawandiS. Gerber-LemaireL. Juillerat-JeanneretN. Moitessier2010Inhibitors of prolyl oligopeptidases for the therapy of human diseases: defining diseases and inhibitors.J Med Chem53934233438
- 27. Cacabelos R, Alvarez A, Lombardi V, Fernández-Novoa L, Corzo L, et al. (2000) Pharmacological treatment of Alzheimer disease: from psychotropic drugs and cholinesterase inhibitors to pharmacogenomics. Drugs Today 36(7): 415–499.R. CacabelosA. AlvarezV. LombardiL. Fernández-NovoaL. Corzo2000Pharmacological treatment of Alzheimer disease: from psychotropic drugs and cholinesterase inhibitors to pharmacogenomics.Drugs Today367415499
- 28. Fuxreiter M, Magyar C, Juhász T, Szeltner Z, Polgár L, et al. (2005) Flexibility of prolyl oligopeptidase: molecular dynamics and molecular framework analysis of the potential substrate pathways. Proteins 60(3): 504–512.M. FuxreiterC. MagyarT. JuhászZ. SzeltnerL. Polgár2005Flexibility of prolyl oligopeptidase: molecular dynamics and molecular framework analysis of the potential substrate pathways.Proteins603504512
- 29. Kaszuba K, Róg T, St Pierre J, Männistö PT, Karttunen M, et al. (2009) Molecular dynamics study of prolyl oligopeptidase with inhibitor in binding cavity. SAR and QSAR in Environmental Research 20(7): 595.K. KaszubaT. RógJ. St PierrePT MännistöM. Karttunen2009Molecular dynamics study of prolyl oligopeptidase with inhibitor in binding cavity.SAR and QSAR in Environmental Research207595
- 30. Shan L, Mathews II, Khosla C (2005) Structural and mechanistic analysis of two prolyl endopeptidases: Role of interdomain dynamics in catalysis and specificity. Proc Natl Acad Sci U S A 102(10): 3599–3604.L. ShanII MathewsC. Khosla2005Structural and mechanistic analysis of two prolyl endopeptidases: Role of interdomain dynamics in catalysis and specificity.Proc Natl Acad Sci U S A1021035993604
- 31. Rea D, Fülöp V (2006) Structure-Function Properties of Prolyl Oligopeptidase Family Enzymes. CBB 44(3): 349–365.D. ReaV. Fülöp2006Structure-Function Properties of Prolyl Oligopeptidase Family Enzymes.CBB443349365
- 32. Tarragó T, Martín-Benito J, Sabidó E, Claasen B, Madurga S, et al. (2009) A new side opening on prolyl oligopeptidase revealed by electron microscopy. FEBS Letters 583(20): 3344–3348.T. TarragóJ. Martín-BenitoE. SabidóB. ClaasenS. Madurga2009A new side opening on prolyl oligopeptidase revealed by electron microscopy.FEBS Letters5832033443348
- 33. TIGR: The Institute for Genomic Research. TIGR: The Institute for Genomic Research.Available: http://plantta.jcvi.org. Accessed 2008 Aug 10. Available: http://plantta.jcvi.org. Accessed 2008 Aug 10.
- 34. Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, et al. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25(17): 3389–3402.SF AltschulTL MaddenAA SchäfferJ. ZhangZ. Zhang1997Gapped BLAST and PSI-BLAST: a new generation of protein database search programs.Nucleic Acids Res251733893402
- 35. McGuffin LJ, Bryson K, Jones DT (2000) The PSIPRED protein structure prediction server. Bioinformatics 16(4): 404–405.LJ McGuffinK. BrysonDT Jones2000The PSIPRED protein structure prediction server.Bioinformatics164404405
- 36. Bernstein FC, Koetzle TF, Williams GJ, Meyer EF, Brice MD, et al. (1977) The Protein Data Bank: a computer-based archival file for macromolecular structures. J Mol Biol 112(3): 535–542.FC BernsteinTF KoetzleGJ WilliamsEF MeyerMD Brice1977The Protein Data Bank: a computer-based archival file for macromolecular structures.J Mol Biol1123535542
- 37. Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22(22): 4673–4680.JD ThompsonDG HigginsTJ Gibson1994CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice.Nucleic Acids Res222246734680
- 38. Mizuguchi K, Deane CM, Blundell TL, Johnson MS, Overington JP (1998) JOY: protein sequence-structure representation and analysis. Bioinformatics 14(7): 617–623.K. MizuguchiCM DeaneTL BlundellMS JohnsonJP Overington1998JOY: protein sequence-structure representation and analysis.Bioinformatics147617623
- 39. Kabsch W, Sander C (1983) Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 22(12): 2577–2637.W. KabschC. Sander1983Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features.Biopolymers221225772637
- 40. Sali A, Blundell TL (1993) Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol 234(3): 779–815.A. SaliTL Blundell1993Comparative protein modelling by satisfaction of spatial restraints.J Mol Biol2343779815
- 41. Westbrook J, Feng Z, Jain S, Bhat TN, Thanki N, et al. (2002) The Protein Data Bank: unifying the archive. Nucleic Acids Res 30(1): 245–248.J. WestbrookZ. FengS. JainTN BhatN. Thanki2002The Protein Data Bank: unifying the archive.Nucleic Acids Res301245248
- 42. Laskowski RA, MacArthur MW, Moss DS, Thornton JM (1993) PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Crystallogr 26(2): 283–291.RA LaskowskiMW MacArthurDS MossJM Thornton1993PROCHECK: a program to check the stereochemical quality of protein structures.J Appl Crystallogr262283291
- 43. Sybyl website. Sybyl website.Available: http://www.cmbi.kun.nl/samsam/tutorials/studs/sybyl/index.html. Accessed 2008 Dec 3. Available: http://www.cmbi.kun.nl/samsam/tutorials/studs/sybyl/index.html. Accessed 2008 Dec 3.
- 44. Wiederstein M, Sippl MJ (2007) ProSA-web: interactive web service for the recognition of errors in three-dimensional structures of proteins. Nucleic Acids Res 35: W407–W410.M. WiedersteinMJ Sippl2007ProSA-web: interactive web service for the recognition of errors in three-dimensional structures of proteins.Nucleic Acids Res35W407W410
- 45. Wilk S, Orlowski M (1983) Inhibition of rabbit brain prolyl endopeptidase by n-benzyloxycarbonyl-prolyl-prolinal, a transition state aldehyde inhibitor. J Neurochem 41(1): 69–75.S. WilkM. Orlowski1983Inhibition of rabbit brain prolyl endopeptidase by n-benzyloxycarbonyl-prolyl-prolinal, a transition state aldehyde inhibitor.J Neurochem4116975
- 46. Kato A, Fukunari A, Sakai Y, Nakajima T (1997) Prevention of amyloid-like deposition by a selective prolyl endopeptidase inhibitor, Y-29794, in senescence-accelerated mouse. J Pharmacol Exp Ther 283(1): 328–335.A. KatoA. FukunariY. SakaiT. Nakajima1997Prevention of amyloid-like deposition by a selective prolyl endopeptidase inhibitor, Y-29794, in senescence-accelerated mouse.J Pharmacol Exp Ther2831328335
- 47. Li J, Wilk E, Wilk S (1996) Inhibition of prolyl oligopeptidase by Fmoc-aminoacylpyrrolidine-2-nitriles. J Neurochem 66(5): 2105–2112.J. LiE. WilkS. Wilk1996Inhibition of prolyl oligopeptidase by Fmoc-aminoacylpyrrolidine-2-nitriles.J Neurochem66521052112
- 48. Ono Y, Kato A, Maeda J, Ohe T, Nakajima (1992) Y-29794-a non-peptide prolyl endopeptidase inhibitor that can penetrate into the brain. Neuroscience Letters 141(2): 156–160.Y. OnoA. KatoJ. MaedaT. OheNakajima1992Y-29794-a non-peptide prolyl endopeptidase inhibitor that can penetrate into the brain.Neuroscience Letters1412156160
- 49. Brandt I, Gerard M, Sergeant K, Devreese B, Baekelandt V, et al. (2008) Prolyl oligopeptidase stimulates the aggregation of R-synuclein. Peptides 29: 1472–1478.I. BrandtM. GerardK. SergeantB. DevreeseV. Baekelandt2008Prolyl oligopeptidase stimulates the aggregation of R-synuclein.Peptides2914721478
- 50. Katsube N, Sunaga K, Aishita H, Chuang DM, Ishitani R (1999) ONO-1603, a potential antidementia drug, delays age-induced apoptosis and suppresses overexpression of glyceraldehyde-3-phosphate dehydrogenase in cultured central nervous system neurons. J Pharmacol Exp Ther 288(1): 6–13.N. KatsubeK. SunagaH. AishitaDM ChuangR. Ishitani1999ONO-1603, a potential antidementia drug, delays age-induced apoptosis and suppresses overexpression of glyceraldehyde-3-phosphate dehydrogenase in cultured central nervous system neurons.J Pharmacol Exp Ther2881613
- 51. Katsube N, Ishitani R (2006) A Review of the Neurotrophic and Neuroprotective Properties of ONO-1603: Comparison with Those of Tetrahydroaminoacridine, an Antidementia Drug. CNS Drug Reviews 6(1): 21–34.N. KatsubeR. Ishitani2006A Review of the Neurotrophic and Neuroprotective Properties of ONO-1603: Comparison with Those of Tetrahydroaminoacridine, an Antidementia Drug.CNS Drug Reviews612134
- 52. Petit A, Barelli H, Morain P, Checler F (2000) Novel proline endopeptidase inhibitors do not modify Aβ40/42 formation and degradation by human cells expressing wild-type and Swedish mutated β-amyloid precursor protein. Br J Pharmacol 130(7): 1613–1617.A. PetitH. BarelliP. MorainF. Checler2000Novel proline endopeptidase inhibitors do not modify Aβ40/42 formation and degradation by human cells expressing wild-type and Swedish mutated β-amyloid precursor protein.Br J Pharmacol130716131617
- 53. Schneider JS, Giardiniere M, Morain P (2002) Effects of the prolyl endopeptidase inhibitor S 17092 on cognitive deficits in chronic low dose MPTP-treated monkeys. Neuropsychopharmacology 26(2): 176–182.JS SchneiderM. GiardiniereP. Morain2002Effects of the prolyl endopeptidase inhibitor S 17092 on cognitive deficits in chronic low dose MPTP-treated monkeys.Neuropsychopharmacology262176182
- 54. Tarragó T, Masdeu C, Gómez E, Isambert N, Lavilla R, et al. (2008) Benzimidazolium salts as small, nonpeptidic and BBB-permeable human prolyl oligopeptidase inhibitors. ChemMedChem 3(10): 1558–1565.T. TarragóC. MasdeuE. GómezN. IsambertR. Lavilla2008Benzimidazolium salts as small, nonpeptidic and BBB-permeable human prolyl oligopeptidase inhibitors.ChemMedChem31015581565
- 55. Morain P, Lestage P, De Nanteuil G, Jochemsen R, Robin J, et al. (2002) S 17092: a prolyl endopeptidase inhibitor as a potential therapeutic drug for memory impairment. Preclinical and clinical studies. CNS Drug Rev 8(1): 31–52.P. MorainP. LestageG. De NanteuilR. JochemsenJ. Robin2002S 17092: a prolyl endopeptidase inhibitor as a potential therapeutic drug for memory impairment. Preclinical and clinical studies.CNS Drug Rev813152
- 56. Morain P, Robin JL, De Nanteuil G, Jochemsen R, Heidet V, et al. (2000) Pharmacodynamic and pharmacokinetic profile of S 17092, a new orally active prolyl endopeptidase inhibitor, in elderly healthy volunteers. A phase I study. Br J Clin Pharmacol 50(4): 350–359.P. MorainJL RobinG. De NanteuilR. JochemsenV. Heidet2000Pharmacodynamic and pharmacokinetic profile of S 17092, a new orally active prolyl endopeptidase inhibitor, in elderly healthy volunteers. A phase I study.Br J Clin Pharmacol504350359
- 57. Portevin B, Benoist A, Rémond G, Hervé Y, Vincent M, et al. (1996) New Prolyl Endopeptidase Inhibitors: In Vitro and in Vivo Activities of Azabicyclo[2.2.2]octane, Azabicyclo[2.2.1]heptane, and Perhydroindole Derivatives. J Med Chem 39(12): 2379–2391.B. PortevinA. BenoistG. RémondY. HervéM. Vincent1996New Prolyl Endopeptidase Inhibitors: In Vitro and in Vivo Activities of Azabicyclo[2.2.2]octane, Azabicyclo[2.2.1]heptane, and Perhydroindole Derivatives.J Med Chem391223792391
- 58. Yoshimoto T, Tsuru D, Yamamoto N, Ikezawa R, Furukawa S (1991) Structure activity relationship of inhibitors specific for prolyl endopeptidase. Agric Biol Chem 55(1): 37–43.T. YoshimotoD. TsuruN. YamamotoR. IkezawaS. Furukawa1991Structure activity relationship of inhibitors specific for prolyl endopeptidase.Agric Biol Chem5513743
- 59. Arai H, Nishioka H, Niwa S, Yamanaka T, Tanaka Y, et al. (1993) Synthesis of prolyl endopeptidase inhibitors and evaluation of their structure-activity relationships: in vitro inhibition of prolyl endopeptidase from canine brain. Chem Pharm Bull 41(9): 1583–1588.H. AraiH. NishiokaS. NiwaT. YamanakaY. Tanaka1993Synthesis of prolyl endopeptidase inhibitors and evaluation of their structure-activity relationships: in vitro inhibition of prolyl endopeptidase from canine brain.Chem Pharm Bull41915831588
- 60. Wallén EAA, Christiaans JAM, Saario SM, Forsberg MM, Venäläinen JI, et al. (2002) 4-Phenylbutanoyl-2(S)-acylpyrrolidines and 4-phenylbutanoyl-L-prolyl-2(S)-acylpyrrolidines as prolyl oligopeptidase inhibitors. Bioorg Med Chem 10(7): 2199–2206.EAA WallénJAM ChristiaansSM SaarioMM ForsbergJI Venäläinen20024-Phenylbutanoyl-2(S)-acylpyrrolidines and 4-phenylbutanoyl-L-prolyl-2(S)-acylpyrrolidines as prolyl oligopeptidase inhibitors.Bioorg Med Chem10721992206
- 61. Atack JR, Suman-Chauhan N, Dawson G, Kulagowski JJ (1991) In vitro and in vivo inhibition of prolyl endopeptidase. Eur J Pharmacol 205: 157–163.JR AtackN. Suman-ChauhanG. DawsonJJ Kulagowski1991In vitro and in vivo inhibition of prolyl endopeptidase.Eur J Pharmacol205157163
- 62. Shinoda M, Toide K, Ohsawa I, Kohsaka S (1997) Specific inhibitor for prolyl endopeptidase suppresses the generation of amyloid beta protein in NG108-15 cells. Biochem Biophys Res Commun 235(3): 641–645.M. ShinodaK. ToideI. OhsawaS. Kohsaka1997Specific inhibitor for prolyl endopeptidase suppresses the generation of amyloid beta protein in NG108-15 cells.Biochem Biophys Res Commun2353641645
- 63. Toide K, Okamiya K, Iwamoto Y, Kato T (1995) Effect of a novel prolyl endopeptidase inhibitor, JTP-4819, on prolyl endopeptidase activity and substance P- and arginine-vasopressin-like immunoreactivity in the brains of aged rats. J Neurochem 65(1): 234–240.K. ToideK. OkamiyaY. IwamotoT. Kato1995Effect of a novel prolyl endopeptidase inhibitor, JTP-4819, on prolyl endopeptidase activity and substance P- and arginine-vasopressin-like immunoreactivity in the brains of aged rats.J Neurochem651234240
- 64. Toide K, Shinoda M, Fujiwara T, Iwamoto Y (1997) Effect of a novel prolyl endopeptidase inhibitor, JTP-4819, on spatial memory and central cholinergic neurons in aged rats. Pharmacol Biochem Behav 56(3): 427–434.K. ToideM. ShinodaT. FujiwaraY. Iwamoto1997Effect of a novel prolyl endopeptidase inhibitor, JTP-4819, on spatial memory and central cholinergic neurons in aged rats.Pharmacol Biochem Behav563427434
- 65. Toide K, Iwamoto Y, Fujiwara T, Abe H (1995) JTP-4819: a novel prolyl endopeptidase inhibitor with potential as a cognitive enhancer. J Pharmacol Exp Ther 274(3): 1370–1378.K. ToideY. IwamotoT. FujiwaraH. Abe1995JTP-4819: a novel prolyl endopeptidase inhibitor with potential as a cognitive enhancer.J Pharmacol Exp Ther274313701378
- 66. Toide K, Fujiwara T, Iwamoto Y, Shinoda M, Okamiya K, et al. (1996) Effect of a novel prolyl endopeptidase inhibitor, JTP-4819, on neuropeptide metabolism in the rat brain. Naunyn Schmiedebergs Arch Pharmacol 353(3): 355–362.K. ToideT. FujiwaraY. IwamotoM. ShinodaK. Okamiya1996Effect of a novel prolyl endopeptidase inhibitor, JTP-4819, on neuropeptide metabolism in the rat brain.Naunyn Schmiedebergs Arch Pharmacol3533355362
- 67. Toide K, Shinoda M, Iwamoto Y, Fujiwara T, Okamiya K, et al. (1997) A novel prolyl endopeptidase inhibitor, JTP-4819, with potential for treating Alzheimer's disease. Behav Brain Res 83(1–2): 147–151.K. ToideM. ShinodaY. IwamotoT. FujiwaraK. Okamiya1997A novel prolyl endopeptidase inhibitor, JTP-4819, with potential for treating Alzheimer's disease.Behav Brain Res831–2147151
- 68. Morris G, Goodsell D, Halliday R, Huey R, Hart W, et al. (1998) Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J Comput. Chem 19(14): 1639–1662.G. MorrisD. GoodsellR. HallidayR. HueyW. Hart1998Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J Comput.Chem191416391662
- 69. McDonald IK, Thornton JM (1994) Satisfying hydrogen bonding potential in proteins. J Mol Biol 238(5): 777–793.IK McDonaldJM Thornton1994Satisfying hydrogen bonding potential in proteins.J Mol Biol2385777793
- 70. Holst M, Saied F (1993) Multigrid solution of the Poisson-Boltzmann equation. J Comput Chem 14: 105–113.M. HolstF. Saied1993Multigrid solution of the Poisson-Boltzmann equation.J Comput Chem14105113
- 71. Lindahl E, Hess B, van der Spoel D (2001) GROMACS 3.0: a package for molecular simulation and trajectory analysis. J Mol Mod 7(8): 306–317.E. LindahlB. HessD. van der Spoel2001GROMACS 3.0: a package for molecular simulation and trajectory analysis.J Mol Mod78306317
- 72. Haffner CD, Diaz CJ, Miller AB, Reid RA, Madauss KP, et al. (2008) Pyrrolidinyl pyridone and pyrazinone analogues as potent inhibitors of prolyl oligopeptidase (POP). Bioorg Med Chem Lett 18(15): 4360–4363.CD HaffnerCJ DiazAB MillerRA ReidKP Madauss2008Pyrrolidinyl pyridone and pyrazinone analogues as potent inhibitors of prolyl oligopeptidase (POP).Bioorg Med Chem Lett181543604363
- 73. Schüttelkopf AW, van Aalten DMF (2004) PRODRG: a tool for high-throughput crystallography of protein-ligand complexes. Acta Crystallogr D Biol Crystallogr 60(Pt 8): 1355–1363.AW SchüttelkopfDMF van Aalten2004PRODRG: a tool for high-throughput crystallography of protein-ligand complexes.Acta Crystallogr D Biol Crystallogr60Pt 813551363
- 74. Dundas J, Ouyang Z, Tseng J, Binkowski A, Turpaz Y, et al. (2006) CASTp: computed atlas of surface topography of proteins with structural and topographical mapping of functionally annotated residues. Nucleic Acids Res 34: W116–W118.J. DundasZ. OuyangJ. TsengA. BinkowskiY. Turpaz2006CASTp: computed atlas of surface topography of proteins with structural and topographical mapping of functionally annotated residues.Nucleic Acids Res34W116W118
- 75. García-Horsman JA, Männistö PT, Venäläinen JI (2007) On the role of prolyl oligopeptidase in health and disease. Neuropeptides 41(1): 1–24.JA García-HorsmanPT MännistöJI Venäläinen2007On the role of prolyl oligopeptidase in health and disease.Neuropeptides411124
- 76. Lee P, Kuo K, Chu P, Liu EM, Lin J (2009) SLITHER: a web server for generating contiguous conformations of substrate molecules entering into deep active sites of proteins or migrating through channels in membrane transporters. Nucleic Acids Res 37: W559–W564.P. LeeK. KuoP. ChuEM LiuJ. Lin2009SLITHER: a web server for generating contiguous conformations of substrate molecules entering into deep active sites of proteins or migrating through channels in membrane transporters.Nucleic Acids Res37W559W564
- 77. Eswar N, John B, Mirkovic N, Fiser A, Ilyin VA, et al. (2003) Tools for comparative protein structure modeling and analysis. Nucleic Acids Res 31: 3375–3380.N. EswarB. JohnN. MirkovicA. FiserVA Ilyin2003Tools for comparative protein structure modeling and analysis.Nucleic Acids Res3133753380