The Protein-Ligand Interaction Profiler (PLIP) web tool allows you to easily analyze non-covalent interactions between biological macromolecules and their ligands based on PDB files. It builds on the PLIP command-line tool, offers a graphical user interface for your input and provides results and visualizations without the need to install the tool yourself. Choose from existing PDB entries by providing a valid 4-letter PDB code or submit a custom PDB file without further preparation of the structure.
PLIP runs per default with proteins as the receptor stucture but nucleic acids can be switched from being considered a ligand to being part of the receptor structure too. In addition PLIP offers a special mode for calculating interactions within one chain of your structure.
After analyzing the complex, the result page offers a comprehensive overview of your results, provides visualizations, and lists all detected non-covalent interactions (hydrogen bonds, water bridges, salt bridges, halogen bonds, hydrophobic interactions, π-stacking, π-cation interactions, and metal complexes) in atom-level detail. Furthermore, the results are available for download in flat text and machine-readable XML format and for visualization purposes as a PyMOL session file (pse). Within the browser, 3D interaction diagrams can be viewed using embedded JSmol applets.
Details about how to use the web tool are provided in the following sections. You can also get to know the tool following the short tutorial. For some additional options only the command line tool provides, see the documentation available on the GitHub page of PLIP.
Tutorial: Analysis of PDB structure 1XDN
This tutorial will walk you through how to search for a protein-ligand complex using the build in search tool, run an analysis with PLIP, and how to view and interpret the results.
Entering the Structure
Open the main page of PLIP. Here, you have two options to choose a structure to be processed, either by uploading a protein-ligand complex in PDB format or by providing a PDB ID. To find the correct PDB ID you can use the build in search tool to find the correct PDB entries without leaving the page. For this quick tutorial, we will use that tool to search for structures of RNA editing ligase. Click on the link with the little magnifier icon below the input field for PDB IDs. Type RNA editing ligase into the search field (A) and click on SEARCH. As shown in the result list, the PDB entry 1XDN (RNA editing ligase 1 from Trypanosoma brucei) is the only match containing a structure with a ligand for this search (B). Click on 1XDN in the result list, the ID will be automatically copied into the PDB ID search field. Now, click on the ANALYZE button to start the analysis. Once your calculation has finished, the results will load automatically.
On the results page, PLIP lists all interaction data, grouped by ligand type (Small Molecules, DNA/RNA, Ions, Polymers), and eventually single complexes. The given structure contains only one binding site which has the identifier ATP-A-501 (C). You get the information that the ligand is a composite ligand of the small molecule ATP and a magnesium ion - it is also marked as biologically relevant with a little star ().
PLIP has detected mainly hydrogen bonds, water bridges and salt bridges for this complex. There is, however, one prominent π-stacking interaction and a complexation of the magnesium ion between water and ATP. A rendered image gives you an overview on the contacts. You can check the details for each single interaction in the tables (D), which are placed to the right or below the visualization, depending on your screen size.
To see which amino acid participates in the stacking, scroll down to the table for π-Stacking. You can collapse and expand each table to focus on the interactions that interest you most by clicking the title (a little arrow and the interaction type) above the respective table. An explanation for each attribute in the table heads is given in the mouseover text. You can also sort the tables by clicking the head of the column you want to sort the entries by.
In this case, for the π-Stacking a phenylalanine residue makes contact with an aromatic ring of ATP adenine. As indicated in the title of the table and in the legend of the rendered image, π-stacking interactions are displayed by green dashed lines. Click on the picture to start a JSmol applet, allowing you to rotate the structure and giving you more information on interactions and labels to quickly identify and explore interactions in the structure. In case you need a higher-resolution picture than provided or want to show the protein-ligand complex from a different angle, you can download the corresponding PyMOL session file using the button below the image (E).
Generating Figures for Talks or Publications
Let's prepare an image of the ligase-ATP complex with focus on the stacking interaction. Download the PyMOL session file and double-click on it or open PyMOL and choose File -> Open. Click and hold the left mouse button and move the molecule until you get a good view on the stacking (F). Let's say we are not interested in water bridges for our purpose. You can hide any interaction type via the named groups on the right side menu. To hide all water bridges, expand Interactions, then click on WaterBridges (G). Hiding the water molecules can be done via accessing .Other and then (H)ide everything for the (Water) selection (H). To render an image, it is recommended to ray-trace (type ray and press Enter) before exporting from PyMOL. To save the file, choose File -> Save Image as -> PNG.
To provide input to run an analysis with PLIP web you have two options. The first is to provide a PDB ID (A). If you know a structure from the Protein Data Bank (PDB) you want to analyze, just type in the PDB ID of the corresponding PDB entry. PLIP web automatically checks whether the given entry is obsolete and replaces the query with the up-to-date structure in that case.
If you do not know the PDB ID, PLIP also allows you to find a PDB entry without leaving the PLIP page. Simply search with structure title, ligand names (including synonyms), or Enzyme Commission Numbers using the integrated search tool, the PDB Wizard. It is reachable by clicking the link with the magnifier icon below the PDB ID entry field (see below for details).
The second option to provide input for PLIP is a file upload (B). Instead of processing a structure from PDB, you can run PLIP on your own protein-ligand complex in PDB format, generated e.g. from previous docking or molecular dynamics analyses. Simply drop your file in the designated area or browse for it by clicking the button.
Simply clicking ANALYZE will now start your calculations assuming you entered a protein-ligand complex and search for interactions based on that. If you instead expand the Advanced Options you can switch to different modes of PLIP, fine tune some settings and provide more information.
(C) PLIP treats nucleic acids as potential ligands in default mode. If you provided a nucleic acid instead of a protein, or if your structure contains both but you want to consider the nucleic acid as part of the receptor along with the protein, you can switch to the nucleic acid receptor mode. Note for command line PLIP users: The
--dnareceptor flag in the command line tool currently behaves a little different than the web tool and does not consider the protein as potential receptor for all interactions anymore in that mode.
(D) In Default mode PLIP detects interactions between ligands and the receptor structure (any protein and additionally, if set to be part of the receptor, any nucleic acids in your structure).
The Intra mode allows you to instead profile interactions only within a chain of your structure. Other chains and possible ligands are not considered for interactions in this mode. Provide the identifier of the chain you are interested in in the text field.
The Inter/Peptide mode can be choosen to analyze inter-chain interactions, i.e. interactions between the chains or strands of your receptor structure. Provide the identifier(s) of the chain or chains you want to investigate separated by spaces and/or commas. Those chains will subsequently be treated as ligands by PLIP. Other ligands will be ignored in this mode. Additionally this mode allows analyzing interactions with peptides that are otherwise not detected by PLIP as ligands automatically due to them being deposited as ATOM entries (instead of HETATM entries) in the pdb file.
Please note that computing interactions within and between chains gets very time consuming for larger numbers of atoms and PLIP web kills jobs automatically after 5 minutes. Usually the web tool is able to handle chains with up to ~3800 atoms. For larger chains you can use the PLIP command-line tool.
(E) PLIP usually excludes modified residues from its calculations and will list them as excluded ligands to let you know. If you want to investigate a structure, where a contained modified residues is of interest, you can tick this option to change that behaviour.
(F) Often pdb files contain more than just one model or conformation. This option allows to choose on which one the analysis is done. If the model you specify is not available in the file provided or downloaded from PDB, model 1 will be used instead. You will be notified on the result page if this was the case.
(G) The default thresholds used by PLIP to detect interactions are based on literature and many years of experience. However, sometimes it might still be necessary to adapt some of them - this option is given here. The thresholds are named according to their constants in the plip command line tool, you can find these along with some hints on the respective source in the PLIP config file. Mouseovers will provide more information on the thresholds, and their default values are given while you enter no other values yourself. You can also find a table further down on this page. Check the thresholds you want to adapt and enter your values. Distances have to be no smaller than 0.5 Å and no larger than 10 Å. Angles take values between 0° and 180°. You can provide up to two decimal places, please use a period as decimal separator. Unchecking a threshold will reset the value to its default. You can also reset all values at once and uncheck all thresholds using the Reset-all button.
(H) Finally you are given the option to provide a job name and email address. A job name can be helpful for you to identify your jobs if you run several in parallel. If you enter an email address, you will receive a notification as soon as your job has been finished, including a link to the result page.
Integrated Search Tool: The PDB Wizard
To search for complexes in the online PDB archive, type any combination of protein/ligand names, terms from the structure's title or Enzyme Commission Number into the search field (I). The suggestion feature will help you to select search terms. Click the Search button (J) to look for entries in the PDB matching your query.
PLIP shows the total number of matches and a list of all hits (K) ranked by importance in relation to the given search terms. In the hit list, the title of the structure as well as HET IDs of ligands in the structure are listed. Matches of search terms are highlighted. Click on a PDB ID to copy it into the search field.
Input File Requirements
PLIP should work for every structure in valid PDB format. As the tool makes use of OpenBabel for processing the files, one option to check the validity is to do a simple conversion with OpenBabel using
babel -i pdb <your_file.pdb> -o pdb any_name.pdb. If the file can be converted without major errors, it should be ready for use with PLIP. In case of non-standard ligand names or missing chains PLIP performs small fixes to the input file.
A In the sidebar of the result page, PLIP web gives an overview of found binding sites. Ligands are grouped by type (Small Molecules, Polymer, DNA/RNA, Ions, and combinations thereof). The binding site identifier consists of the PDB ligand identifier, the chain and the residue assigned to the ligand in the input PDB file. A ligand synonym is given in brackets for most of the compounds in the PDB.
As a separate category, all ligands with no interactions are listed at the bottom. For all entries you can click on the binding site identifier to automatically scroll to the detailed summary of the found interactions (see below). To give a better orientation on the page, the binding sites currently visible in detail on your screen are highlighted in green. In the case of composite ligands (e.g. polysaccharides with separate HET IDs for the subunits, or DNA/RNA), the term Composite Ligand is indicated in brackets. The binding site details for those entries lists all members which have been grouped together.
B Below the summary, you can download result files containing interaction data for all ligands in the structure for further processing or inspection. The text file contains the tables as seen on the result page in a human-readable format. For further computational analysis, you might choose the XML file, which is easily parsable.
C If you ran your analysis on a PDB ID, the bottom of the sidebar contains a link to the structure's page in the PDB. Below are also a link to information on how to cite PLIP and the option to return to the start page to run another analysis.
Binding Site Details
D Binding site sections are grouped by ligand type and ligand and are titled with the binding site identifier (see above). Sometimes that identifier is followed by a little star () to indicate that the ligand is binding specifically and is thus biologically relevant as defined by the BioLiP database. Complexes with composite ligands contain detailed information about the ligand members in red directly below the name.
E For each macromolecule-ligand complex in the structure, PLIP shows a rendered interaction diagram. The ligand is shown in orange, the macromolecule residues in blue color. For intra-chain interactions the chain is shown in orange. Non-covalent interactions are indicated by dashed or solid lines as indicated in the legend on the right side. To show an interactive 3D visualization diagram of the interactions in your browser, click on the preview image to open a JSMol applet.
F You can download the rendered image in high resolution as well as the PyMOL session file which was used to render the image. In the session file, all structural elements and single interactions are conveniently grouped, allowing you to switch on and off specific interaction types. You can inspect the complex or process it further to generate publication-ready images in a few steps. We recommend using PyMOL 2.3.0 or higher. In earlier versions, you might see warnings upon loading the file and encounter missing elements in the visualization or different color schemes. In the case of a red background, you should be able to restore the original color by using
set bg_rgb, [1.0, 1.0, 1.0].
G For each interaction type, a table lists the contacts in atomic-level detail. Each entry contains detailed information on protein/nucleic acid residues, participating ligand atoms and geometry of the interaction (e.g. distance of interacting atoms). Depending on the type, there is additional information available (e.g. offset of aromatic rings for π-stacking). A detailed explanation for each attribute can be shown by a mouseover on the column head. All atom and residue numbering are in accordance with the numbering in the corresponding PDB file given as input.
The tables are shown to the right of the picture if the screen is big enough or below if that is not the case. Each table can be collapsed and expanded again to allow focusing on specific interactions. Clicking a column head will sorte the tables entries according to that column.
PLIP uses a rule-based system for detection of non-covalent interactions between protein/nucleic acid residues and ligands or within a chain. Information on chemical groups able to participate in a specific interaction (e.g. requirements for hydrogen bond donors) and interaction geometry (e.g. distance and angle thresholds) from literature are used to detect characteristics of non-covalent interactions between contacting atoms of protein/nucleic acid and ligands or within a chain. For each binding site, the algorithm searches first for atoms or atom groups in the complex which could possibly be partner in specific interactions. In the second step, geometric rules are applied to match groups in forming an interaction.
Detection and Filtering of Ligands
Previous to the detection step for the interactions, PLIP extracts all ligands contained in the structure. Modified amino acids are identified and excluded using MODRES entries of the PDB files. Additionally, we use the BioLiP list of possible artifacts to remove ligands which are in this list and appear 15 times or more in a structure. Just a few compounds are currently excluded, being listed in the PLIP config file here.
Preparation of StructuresPolar hydrogens are added to the structure and alternative conformations/models/positions removed. Missing chains are assigned to ligands and non-standard ligand names (with special characters) altered to LIG.
Detection of Possible Interacting Groups
Binding Site Atoms
The binding site distance cutoff is determined by adding up BS_DIST to the maximum extent to the ligand (maximum distance of a ligand atom to ligand centroid). All atoms belonging to the macromolecule within this distance cutoff to any binding site atoms are counted as belonging to the binding site.
An atom is classified as hydrophobic if it is a carbon and has only carbon or hydrogen atoms as neighbours.
OpenBabel is used to identify rings (SSSR perception) and their aromaticity. In cases where no aromaticity is reported by OpenBabel, the ring is checked for planarity. To this end, the normals of each atom in the ring to its neighbors is calculated. The angle between each pair of normals has to be less than AROMATIC_PLANARITY. If this holds true, the ring is also considered as aromatic.
Hydrogen Bond Donors and Acceptors
OpenBabel is used to identify hydrogen bond donor and acceptor atoms. Halogen atoms are excluded from this group and treated separately (see below).
The detection of charged groups is only exhaustive for the binding site, not the ligands. For proteins, positive charges are attributed to the side chain nitrogens of Arginine, Histidine and Lysine. Negative charged are assigned to the carboxyl groups in Aspartic Acid and Glutamic Acid. In ligands, positive charges are assigned to quaterny ammonium groups, tertiary amines (assuming the nitrogen could pick up a hydrogen and thus get charged), sulfonium and guanidine groups. Negative charges are reported for phosphate, sulfonate, sulfonic acid and carboxylate.
Halogen bonds donors and acceptors
Assuming that halogen atoms are not present in proteins (unless they are artificially modified), halogen bond donors are searched for only in ligands. All fluorine, chlorine, bromide or iodine atoms connected to a carbon atom qualify as donors. Halogen bond acceptors in proteins are all carbon, phosphor or sulphur atoms connected to oxygen, phosphor, nitrogen or sulfur.
Water atoms are assigned to a ligand-binding site complex if their oxygen atoms are within a certain cutoff to the ligand. The cutoff is determined by adding up BS_DIST to the maximum extent to the ligand (maximum distance of a ligand atom to ligand centroid). This means the farthest distance of a ligand to a water atom is BS_DIST.
Detection of Interactions
For an overview on geometric cutoffs and their default values used for the prediction of interactions, see also the table at the bottom. The values of these thresholds can be adapted for single jobs in both, the PLIP command line tool and the PLIP web tool. If you regularly run jobs with custom thresholds you might prefer to use the command line tool and adapt the default values to your needs.
As hydrophobic interactions result from entropic changes rather than attractive forces between atoms, there are no clear geometries of hydrophobic association. The observed attraction between hydrophobic atoms decays exponentionally with the distance between them. A generous cutoff was chosen, identifying a prime set of hydrophobic interactions between all pairs of hydrophobic atoms within a distance of HYDROPH_DIST_MAX.
Since the number of hydrophobic interactions with such an one-step approach can easily surpass all other interaction types combined, it may strongly influence subsequent evaluation or applications as interaction fingerprinting. To overcome this problem, the number of hydrophobic interactions is reduced in several steps. First, hydrophobic interactions between rings interacting via π-stacking are removed. This is done because stacking already involves hydrophobic interactions. Second, two clustering steps are applied. If a ligand atom interacts with several binding site atoms in the same residue, only the interaction with the closest distance is kept. Subsequently, the set of hydrophobic interactions is checked from the opposite perspective: if a protein atom interacts with several neighboring ligand atoms, just the interaction with the closest distance is kept. Together, these reduction steps help to report only the most representative hydrophobic interactions.
A hydrogen bond between a hydrogen bond donor and acceptor is reported if several geometric requirements are fulfilled. The distance has to be less than HBOND_DIST_MAX and the angle at the donor group (D-H...A) above HBOND_DON_ANGLE_MIN.
Since salt bridges involve purely electrostatic interactions as well as hydrogen bonds, it is not meaningful to report both interaction types between the same groups. Thus, hydrogen bonds between atoms are removed if they belong to groups that already form a salt bridge to that atom. As a general rule, a hydrogen bond donor can take part in only one hydrogen bond, while acceptor atoms can be partners in multiple hydrogen bonds (e.g. bifurcated hydrogen bonds). For multiple possible hydrogen bonds from one donor, only the contact with the donor angle closer to 180 ° is kept.
π-Stacking for two aromatic rings is reported whenever their centers are within a distance of PISTACK_DIST_MAX, the angle deviates no more than PISTACK_ANG_DEV from the optimal angle of 90 ° for T-stacking or 180 ° for P-stacking. Additionally, each ring center is projected onto the opposite ring plane. The distance between the other ring center and the projected point (i.e. the offset) has to be less than PISTACK_OFFSET_MAX. This value corresponds approximately to the radius of benzene + 0.6 Å.
π-Cation interactions are reported for each pairing of a positive charge and an aromatic ring if the distance between the charge center and the aromatic ring center is less than PICATION_DIST_MAX. In the case of a putative π-cation interaction with a tertiary amine of the ligand, an additional angle criterion is applied (see documentation in the source code).
Whenever two centers of opposite charges come within a distance of SALTBRIDGE_DIST_MAX, a salt bridge is reported. In contrast to hydrogen bonds, there are no additonal geometric restrictions.
While residues can be bridged by more than one water molecule, for the prediction in this script the only case considered is one water molecule bridging ligand and protein atoms via hydrogen bonding. The water molecule has to be positioned between hydrogen bond donor/acceptor pairs of ligand and protein with distances of the water oxygen within WATER_BRIDGE_MINDIST and WATER_BRIDGE_MAXDIST to the corresponding polar atoms of the donor or acceptor groups. If a constellation with a water atom fulfils these requirements, two angles are checked. The angle ω between the acceptor atom, the water oxygen and donor hydrogen has to be within WATER_BRIDGE_OMEGA_MIN and WATER_BRIDGE_OMEGA_MAX. Additionally, the angle θ between the water oxygen, the donor hydrogen and the donor atom has to be larger than WATER_BRIDGE_THETA_MIN.
Similar to standard hydrogen bonds, a water molecule is only allowed to participate as donor in two hydrogen bonds (two hydrogen atoms as donors). In the case of more than two possible hydrogen bonds for a water molecule as donor, only the two contacts with a water angle closest to 110 ° are kept
Halogen bonds are reported for each pairing of halogen bond acceptor and donor group having a distance of less than HALOGEN_DIST_MAX and angles at the donor and acceptor group of HALOGEN_DON_ANGLE and HALOGEN_ACC_ANGLE with a deviation of no more than HALOGEN_ANGLE_DEV
For metal complexes, PLIP considers metal ions from a set of more than 50 species (see PLIP config for more details). Possible interacting groups in the protein are sidechains of cystein (S), histidine (N), asparagine, glutamic acid, serin, threonin, and tyrosin (all O), as well as all main chain oxygens. In ligands, following groups are considered for metal complexation: alcohols, phenolates, carboxylates, phosphoryls, thiolates, imidazoles, pyrroles, and the iron-sulfur cluster as a special constellation. For one metal ions, all groups with a maximum distance of METAL_DIST_MAX to the ligand are considered for the complex. After assigning all target groups to one metal ions, the resulting set of angles of the complex is compared with known sets of angles from common coordination geometries (linear , trigonal planar , trigonal pyramidal , tetrahedral , square planar , trigonal bipyramidal , square pyramidal , and octahedral ). The best fit with the least difference in observed targets is chosen as an estimated geometry and targets superfluous to the constellation are removed.
Currently Unsupported Interaction Types
- Covalent bonds
- Weak hydrogen bonds involving carbon atoms
- Halogen-Water-Hydrogen Bridges
- Water bridges of higher degree (bridging over more than one water molecule)
Thresholds Used for Detection Steps in PLIP
|BS_DIST||7.5 Å||Cutoff for determination of binding site atoms|
|AROMATIC_PLANARITY||5.0 °||Cutoff for planarity criterion in aromatic ring detection|
|HYDROPH_DIST_MAX||4.0 Å||Max. distance of carbon atoms for a hydrophobic interaction|
|HBOND_DIST_MAX||4.1 Å||Max. distance between acceptor and donor in hydrogen bonds|
|HBOND_DON_ANGLE_MIN||100 °||Min. angle at the hydrogen bond donor (D-H...A)|
|PISTACK_DIST_MAX||5.5 Å||Max. distance between ring centers for stacking|
|PISTACK_ANG_DEV||30 °||Max. deviation from optimum angle for stacking|
|PISTACK_OFFSET_MAX||2.0 Å||Max. offset between aromatic ring centers for stacking|
|PICATION_DIST_MAX||6.0 Å||Max. distance between charge and aromatic ring centers|
|SALTBRIDGE_DIST_MAX||5.5 Å||Distance between two centers of charges in saltbridges|
|HALOGEN_DIST_MAX||4.0 Å||Max. distance between oxygen and halogen|
|HALOGEN_ACC_ANGLE||120 °||Optimal halogen bond acceptor angle|
|HALOGEN_DON_ANGLE||165 °||Optimal halogen bond donor angle|
|HALOGEN_ANGLE_DEV||30 °||Max. deviation from optimal halogen bond angle|
|WATER_BRIDGE_MINDIST||2.5 Å||Min. distance between water oxygen and polar atom|
|WATER_BRIDGE_MAXDIST||4.1 Å||Max. distance between water oxygen and polar atom|
|WATER_BRIDGE_OMEGA_MIN||71 °||Min. angle between acceptor, water oxygen and donor hydrogen|
|WATER_BRIDGE_OMEGA_MAX||140 °||Max. angle between acceptor, water oxygen and donor hydrogen|
|WATER_BRIDGE_THETA_MIN||100 °||Min. angle between water oxygen, donor atom and hydrogen|
|METAL_DIST_MAX||3.0 Å||Max. distance between metal ion and interacting atom|