Mechanism and Catalytic Site Atlas

Introduction

EzMechanism is a web tool able to automatically generate possible catalytic mechanisms for a given enzyme active site and reaction. It is a knowledge-based approach, which works by searching the chemical reaction space available to the enzyme, using a set of catalytic rules inferred from the mechanisms annotated in the M-CSA.

1 - Access to EzMechanism

EzMechanism is only accessible to registered M-CSA users. To register please send an email to ribeiro@ebi.ac.uk with the desired username and an account will be created for you. Automatic registration will be added in the future.

Once registered the user can log in with the provided credentials by clicking on the "Log in" button in the main navigation bar. After login, users will be able to access EzMechanism by clicking on the EzMechanism button which will then be visible.

The main EzMechanism page has a table with all the searches previously created by the user (example in the figure below). For new users, this table will be empty

2 - Choosing a PDB and Creating a New Search

In order to make a mechanism search, a PDB structure of the enzyme must be chosen containing an active site with the catalytic residues, the substrates, and any co-factors necessary for the reaction. It is not necessary to have these exact molecules in the active site of the selected PDB structure, although a model resembling the native system will yield better results:

• If the catalytic residues are mutated in the PDB structure, the UniProt residue will be considered instead. The 3D structure of the non-mutated residue will be created in the active site and its maximum-common-substructure (MCS) to the mutated 3D residue will be used as guide for its correct positioning.
• Substrates and co-factors not already present in the active site may be added by replacing an existing molecule (including water molecules) in the active site with the desired one. Similarly, the MCS is used to guide the positioning of the native molecule on top of the PDB ligand.

PDB structures of enzymes typically do not contain the wild-type catalytic residues together with the complete set of substrates and co-factors, since that state is labile and hence difficult to capture. We plan to allow the submission of user generated models in the future to overcome this limitation.

At the moment, it is the biological assembly of the chosen PDB structure that is used for the calculation.

After choosing an adequate PDB structure the user can type the PDB code in the input box and click on "New From PDB" to generate a new prediction based on this PDB. After clicking, a new line should appear on the table. The user can then click on "input" to proceed to the next stage.

3 - Choosing the Catalytic Residues

The first step in the input page, after choosing the PDB structure, is to select the catalytic residues involved in the reaction. This can be done by selecting the residues in the drop-down list, and by selecting which part of the residue is involved in the reaction (defaults to side chain) followed by clicking on the "+Residue Button". To speed up the residue selection in the the drop-down list, users can also type the name of the residue followed by its residue number.

The drop-down list of residues shows two numbers for each residue. The first is the position of the residue in the PDB chain, while the one after "seq:" is the position of the residue in the protein sequence, as taken from UniProt.

When first selecting a residue, its chain will default to the first one shown in the drop-down list. If it is necessary to change the chain of a particular residue, this can be done by clicking on the appropriate chain in the table of residues. The 3D viewer can be used to make sure all residues are associated with the correct chain (this is particularly important when the active site is situated in the interface of two or more chains). The button "Show Active Center" will focus on the already selected residues to facilitate this task.

4 - Defining the substrates and co-factors

Substrates and co-factors are treated in the same manner in EzMechanism. To add a new one of these molecules click on the "+Substrate/Cofactor" button, which will add a new line to the table above. Then click on the edit button (🖉) to access the edit page (figure below).

In the left section of the edit page the molecule can be defined in three ways:

• The user can draw the molecule using the MarvinJS plugin interface.
• If the ChEBI ID of the molecule is known, it can be inserted in the text input box, followed by clicking on the "From ChEBI" button.
• Molecules can be copied and pasted from other MarvinJS instances.

After the molecule is drawn using any of these methods the user needs to click on "Save Mol".

The name of the molecule can be defined in the input text box at the bottom, followed by clicking on "Save Name".

In order to generate appropriate 3D coordinates for the drawn molecule, it needs to be partially mapped to an existing ligand in the PDB structure. On the right side of the page, the user needs to choose to which ligand in the PDB will the drawn molecule be mapped to. The PDB ligands in the drop-down list will be sorted by similarity to the drawn molecule and distance to the active site, so typically the correct ligand should be located towards the beginning of the list.

After selecting a PDB ligand, the maximum-common-substructure between the drawn molecule and the ligand will be shown as green circle highlights. The 3D viewer in the previous page can also be used to confirm the mapping is correct, by comparing the number and name of the residues in the viewer with the selected ones.

5 - Defining the Overall Chemical Reaction

After defining all the active site molecules, the user needs to specify what is the overall reaction of the enzyme. This is done in the following two MarvinJs boxes (as pictured below), where the one in the left is used to define the reactants configuration of the active site and the one in the right is used to define the products configuration.

These are the steps necessary to define the overall reaction:

• Click on the "New Scheme Draft" button. This will cause all the molecules involved in the reaction, as defined in the steps above, to appear in the first box. The user can then rearrange the molecules in a disposition that mimics the 3D structure of the active site more closely. The position of the molecules in this 2D scheme does not affect the calculation but can help the analysis of the output.
• The protonation states of the molecules in these diagrams is taken into account during the calculation, so the user must protonate or deprotonate atoms as required. This can be accomplished by using the "+" and "-" buttons to change the charge of the atoms. In order to test alternative protonation states, additional calculations must be performed.
• No new molecules should be added in this viewer since these will not be mapped correctly to the catalytic residues or the defined substrates. However, atoms can be added or deleted to the existing molecules.
• After defining the position of the molecules and their protonation states, click on the "Save Scheme To Database" button.
• To define the products of the reaction, first click on the "Copy From Reactants" button under the box on the right. This will make an exact copy of the reactants configuration as previously defined.
• Using the "Delete" and "Bond" buttons of the MarvinJs box (third and fourth buttons in the left vertical bar, respectively) add and delete bonds as necessary to draw the products configuration. Each atom in the reactants configuration (left) must have a corresponding atom in the products configuration (product) so if you notice something is missing first draw it in the left box, and then copy the products configuration again. Do not delete or add atoms directly in the right box.
• After defining the products, click on "Save Scheme to Database". To make sure all atoms are mapped correctly between reactants and products, click in "Check Mapping and Bond Changes". This will create an alert for any problems with the mapping and will highlight the cleaved and new bonds in the 2D diagrams.

6 - Defining the Prediction Parameters

Finally, the user defines some calculation parameters. At the moment, the only exposed parameter is the total number of nodes (configurations) the search algorithm checks against the catalytic rules. A bigger number might be required if the mechanism involves a large number of steps or the pose of the 3D structure is not conducive to catalysis. The default number of configurations to explore is 100.

A text box is provided here where the user can write their own comments about the calculation. This field also appears in the overall table that shows all the calculations for the user, so this might be useful to explain what distinguishes this calculation from others.

7 - Submission and Running Status

To submit the search, click on "Run EzMechanism Search" in the input page or on "Run" in the adequate row in the table of the overall EzMechanism page.

The status of the calculation can be tracked in the EzMechanism page. Running calculations will have a red "button" labelled "Cancel" that can be used to interrupt the calculation. The page needs to be refreshed to update the status of the calculation. The figure below shows all the status the calculation will go through before completion.

After the calculation is finished, if the user wishes to try a different selection of catalytic residues, protonation states, or parameters, this can be done by opening the input page, change anything as desired, and click on "Run" again. This will delete the results of the previous run. If the user wants to try different parameters while keeping the previous results, a new calculation should be created from scratch.

Output Documentation