CovalentDock Cloud

Help

This page offers the general help information to use this web server. For specific problems or reporting possible bugs, please write to us.

Covalent docking vs. non-covalent docking

Conventional protein-ligand docking mainly focuses on the docking between the two molecules through non-covalent interactions. This focus is understandable since the main stream of rational drug design relies on these non-covalent interactions as the mechanism of the functionality. However, drugs adopting covalent binding mechanism can hardly be categorized as mavericks. According to this Nature article, 3 of the 10 top-selling drugs in U.S. in 2009 are covalent inhibitors, not to mention the well-known examples of Aspirin and Penicillin as covalent drugs.

How it works?

Upon job submission, the backend CovalentDock program will first try to identify the electrophilic groups in the ligand and nucleophilic groups in the receptor. If they are both found, the structures will be modified to simulate the covalent linkage formation (and possibly some other minor structure alternation). Then covalent docking will be performed on this receptor-ligand pair, using similar techniques to conventional docking to find the pose with lowest free energy. The major difference is that in addition to the scoring functions to estimate the non-covalent interactions, the CovalentDock also has a special energy term to estimate the energy contribution from covalent linkage formation.

Currently, two types of covalent reactions are supported by CovalentDock. Other types are now under active development and will be merged into the stable version of CovalentDock in near future.

Type-I covalent docking simulates the Michael reaction. The electrophile in the ligand is identified by the Michael acceptor (O=C-C=C), while a Cysteine residue in the receptor is identified as nucleophile. Examples of covalently bound complexes of this type include 3IKA and 2JIV.

Type-II docking models the mechanism of β-lactams. The four-membered β-lactam itself is identified as the electrophile, while the nucleophile in the receptor is identified by a Serine residue. Penicillin (in complex 2JBF and 1GHP) is a prominent example of this type.

The interactions between the electrophile and and nucleophile to form a covalent linkage between the ligand and the receptor are illustrated in the figure above. The energy contribution of each type are estimated by computational simulation and calibrated with empirical data.

For more detailed description on how CovalentDock works and the science behind it, please refer to our journal article (Online | PDF).

Submitting a new job

The “New Job” page is pretty much self-explanatory: by specifying the receptor by an uploaded structure or a valid PDB ID along with the coordinates of the binding site center, and the ligand by an uploaded structure or a valid ZINC ID, new jobs can be easily created.

For receptors, if they are retrieved from PDB, the structure will be cleaned and preprocessed automatically: all HETATMs are removed and hydrogens are added to the structure. In most cases, these procedures will produce a clean receptor structure. However, some PDB structures need further modification and adjustment other than this simple automated procedure. In such cases, please prepare the structures manually and upload it. If the uploading option is in use, the structure will be used as-is without any modification. Thus, please prepare a clean structure with all hydrogens fully added.

CovalentDock is NOT a blind docking protocol thus the coordinates for the binding site center are required. The size of the binding site will be automatically adjusted according to the size of the ligand or to the size of 16Å×16Å×16Å, whichever is larger. In fact, it is required for the receptor to have at least one Cysteine (for Type-I) or Serine (for Type-II) within the binding site. Otherwise, the backend CovalentDock will refuse to start after the preprocessing is done.

For ligands, if they are retrieved from ZINC, the only modification applied is to add hydrogens when necessary, whereas uploaded ligands will be used as-is. Currently the backend CovalentDock supports two types of covalent reactions thus require the ligand to have at least one Michael acceptor (for Type-I) or a β-lactam group (for Type-II). For ligands lacking both covalent binding patterns, the backend CovalentDock will also refuse to start.

Upon successful submission of a new job, a unique token will be given to identify the job. Use this token to check the status of the job and retrieve the results.

The data associated with any job will be kept for only a limited time period. All jobs terminated due to any error will be deleted within 24 hours and all successfully finished jobs will be kept for only 15 days.

Instructions to prepare the structures for upload

For receptors, please remove all the solvent molecules and add all hydrogens. If the protein of a bound complexes is used as the receptor, please do remember to remove the native ligand structure from the complex before uploading.

For ligands, please add all hydrogens and make sure that all the atom types and the bond types of the molecule are correct. The backend CovalentDock depends the correctness of the atom types and bond types of the input. Please do note that we have observed multiple cases where the electrophilic patterns can be detected in original structures but later became invalid once the hydrogens are added. More specifically, make sure none of the tree carbons is aromatic for the Michael acceptors and all corresponding bond types are set correctly.

Check job status and retrieve results

Use the token provided upon successful job submission to check the status of the job.

When the job finishes, the results will be parsed and visualized by Jmol. The receptor will be colored grey, with the binding pocket highlighted blue, while the ligand will be colored white. If there’s a covalent linkage formed in-between, the newly built covalent bond will be colored red and labeled with its bond length.

There is a simple control panel alongside the visualization, as exampled on the left. A drop down menu is included to select from all options if chirality of the ligand produces multiple configurations on the covalent linkage atom. The configuration options are named with the convention "ligand_[bond_<type_id>_<atom_id>_<config_id>]_min". It is possible for the ligand to have multiple chiral atoms affecting the covalent bond formation thus the part within the square bracket may appear more than once. For each configuration, ten independent covalent dockings are performed. The results are clustered based on the structural similarity and then ranked according to the estimated free energy. Fewer clusters usually means better convergence of the result. A list of radio buttons is included in the control panel for each cluster and the selected result structure will be visualized.

The results are also available for downloading. All docking result poses for all configurations due to chirality, along with the receptor structure, are compressed in a .zip file. Use standalone programs like PyMOL or AutoDockTools for further investigation.

Sample results

Check the following two sample results for live examples. Both of them are generate with the sample input parameters on the job submission page.

Michael additions

Receptor: ERK-2 (PDB: 3C9W)
Binding site: CYS-164
Ligand: Hypothemycin (ZINC: 34034438)

Check sample result

Additions to β-lactam

Receptor: β-lactamase (PDB: 1GHP)
Binding site: SER-70
Ligand: Penicillin (ZINC: 3871701)

Check sample result

Privacy

All data submitted to this server will be kept confidential and will be destoried after a period of time: 24 hours for jobs terminated due to errors, and 15 days for successfully finished jobs. The submitted data can only be accessed with a valid token.

Known issues