Computer Practical

In silico inhibitor design workshop


The aim of this workshop is to gain a better understanding of how structural biology can be used to inform therapeutic design through in silico drug design.  This workshop will use the muscarinic acetylcholine receptor GPCR as an example for the design of potent inhibitors. The tutorial will be a computer based student-directed learning exercise with a series of questions throughout the tutorial which explore the different areas of drug binding.  It is important that you understand the principles underlying the terms highlighted in bold below and also that you understand the importance of these approaches in drug design and the benefits and limitations of this approach. By the end of the workshop you should be able to assess the binding potential of a range of compounds based from the structure alone and be able to inform and what is required for a potent lead compound that could be further developed into a drug. To set the scene you will find the following resources beneficial and they will provide context for the tutorial. 

The targets we will look are based on these two papers

Liu et al., (2018) Structure-guided development of selective M3 muscarinic acetylcholine receptor antagonists. PNAS 115(47) 12046-12050.

Kruse AC et al., (2013). Activation and allosteric modulation of a muscarinic acetylcholine receptor. Nature 504 101-106.

For a brief introduction to in-silico design please look at the following but do not worry about the exact details its more for an overview.

McPhillie MJ et al., (2015) Computational methods to identify new antibacterial targets. Chem. Biol. Drug Des. 85(1) 22-29.

The three main aims are as follows:

1). To demonstrate how different factors can dictate inhibitor binding and how it is often a balancing act between different properties that generates a potent inhibitor.

2). To demonstrate how in silico approaches are a powerful means to rapidly screen a large library of compounds and why this is a common approach in industry.

3). To demonstrate the limitations of computational modelling and why a structure is not essential to develop a new drug but can facilitate and speed up the development. 


The session will be broken down into three sections. The first will look at the muscarinic receptor structure and analyse what makes a good binding pocket.  Here we will consider factors such as shape complementarity, hydrogen bonding, covalent bonding, electrostatics and the role of point mutations. This will allow you to familiarise yourself with the factors that are important in developing potent inhibitors. In the second part you will analyse a range of modified scaffolds and use in silico methods to predict their binding. In the third part you will use the skills developed to answer questions about a different set of compounds designed against the Enoyl reductase enzyme. 

Hence, by the end of this practical the students should be able to: 

  • Have an understanding of protein structure; 
  • Know what the main factors are that drive inhibitor binding; 
  • Perform a docking experiment to predict binding;
  • Use the information gained form the practical to predict the binding of a range of modified scaffolds;



The drug discovery pipeline can take several years from target identification to the development of a lead compound but this can be significantly enhanced when the 3D structural information of the protein target is known, especially if the protein is bound to an inhibitor allowing the mode of inhibitor binding to be established. For this purpose structure-based drug design (SBDD) utilises prior structural knowledge of the target system to design new inhibitors and can be used to complement HTS methods. Typically, this can either be done via an independent approach where no prior hits have been identified to the target and molecules are designed from scratch (de novo design). Alternatively, this can be carried out via the structural development of a hit identified through a high throughput screen or fragment based approach. Molecular docking, such as virtual high throughput screening (vHTS), can be used to virtually screen thousands of compounds from a database against the desired target and identifies compounds which are predicted to bind with high affinity. Different programs are available to carry out virtual high throughput screens (vHTS) such as GLIDE, GOLD and Autodock which use different algorithms to position the molecules into the binding site and subsequently score those compounds based upon their predicted interactions. The hit compounds identified through vHTS can represent the starting point for drug discovery programs as they are developed into lead-like molecules. SBDD by vHTS has played a pivotal role in close to 20 drugs in clinical use including the peptidomimetic HIV protease inhibitors nelfinavir42, amprenavir43 and lopinavir44 (Figure 1).


Figure 1: Example compounds identified through virtual screening approaches. Nelfinavir, Amprenavir and Lopinavir were developed to treat HIV after hit compounds were initially identified in vHTS screens. Note they have similar central scaffold with modifications decorated on the periphery of the drug.


De novo design is an alternative approach which allows inhibitors to be designed from scratch. The general principle behind the technique is that the inhibitor binding site on the protein is analysed and regions which could form bonding interactions to the inhibitor molecule, such as H-bond acceptor/donor sites or hydrophobic side chains, are identified. Fragments or molecular building blocks are then positioned at these sites to create favourable interactions between the ligand and receptor. These fragments are subsequently linked together to form a larger, drug-like molecule which is predicted to bind to the protein. The fragments can be linked together one at a time or by positioning multiple fragments at the favourable sites and linking them together all at once (Figure 2). 



The first part of the workshop will be done using the graphics program PyMOL. This program has a number of options permitting a range of views of the protein-ligand complex.  The example you will look at is the muscarinic acetyl choline receptor, which you will be familiar with from previous lecture material.  PyMOL can be found in the programs on the start menu. There is an additional “how to” tutorial video in the learning resources on PyMOL.

Task 1. From the BMSC2125 (learning resources>week4:workshop…) folder load the desktop file shape_complementarity.pse within PyMOL (click File > open from the top left of the program).  From here you will see the surface of the binding pocket with Iperoxo bound. The protein is coloured cyan, blue, red and orange for carbon, nitrogen, oxygen and sulphur, respectively, and the inhibitor is coloured yellow, blue and red for carbon, nitrogen and oxygen, respectively. 

Question 1.  Briefly describe the shape of the binding pocket with respect to the bound inhibitor Iperoxo.


Task 2. Point mutations can result in a decrease or increase in inhibitor potency.  Here you will look at some specific mutations within the binding pocket and decide how they may influence inhibitor binding.  You will need to open the file mutations.pse.  You will see a series of structure names displayed along the right-hand side (Native, Trp400Ala, Tyr426Ala etc..).  Each one of these will show the binding pocket with a different mutation present. Select each one by clicking on the grey box and hiding the previous one by clicking on the side.  If you “toggle” ie select the state on and off you will see where the difference is.  Note only one amino acid residue is being changed at a particular position. Go to Display > sequence and this will show a sequence view allowing you to see the residues of interest.

Question 2.  Examine the following mutations in the binding site Trp400Ala, Tyr426Ala, Cys429Ser, Cys429Ala and Cys429Lys. Discuss how each mutation may influence the binding of Iperoxo and rank them from the mutation which will most influence binding to the least.


Task 3. Close down the current PyMOL session (Click File > reinitialise > everything) and load up bonding_patterns.pse or simply close down PyMOL completely and reload from the session file bonding_patterns.pse. This will allow you to see the Iperoxo binding site within the muscarinic acetylcholine receptor. Move around the binding site and look at the different interactions made between the protein and inhibitor. You can go to Display > sequence and this will show a sequence view allowing you to see the residues of interest.  The interactions are also shown in the corresponding paper.

Question 3. Both Asn404 and Trp400 form interactions with the bound inhibitor, what are the interactions and how do they compare? 


Task 4. We will now study the effect of binding a positive allosteric modulator (Ly2119620, Ly2 for short) and how the position of the binding pocket may affect the potency of the inhibitor.  Open up the PyMOL file ly2.pse, you will see the Iperoxo structure in blue and ly2 in magenta. (Note that by clicking the grey boxes in the top right you can show and hide the different structures for clarity).

Question 4. How do the binding sites of Ly2 and Iperoxo differ and how might this effect their Kd and why? (The paper by Kruse et al., Fig 4 & 5 may also help). 


Task 5. Now we will look at the electrostatic surface, and study the role of electrostatics in drug binding. Whilst still using the Ly2.pse session, click on 4MQS-rec to hide this structure and then for the 4MQT file click on “S” (small box at right hand side) and click on surface. You will now see a surface over a protein coloured red for negative charge and blue for positive charge. 

Question 5. What do you notice about the charge of the ly2 binding site?  What might an advantage and disadvantage be of a highly charged inhibitor? 


Task 6.  You have now looked at the role of shape complementarity (question 1 & 2), bonding (question 3 & 4) and location and charge of the binding pocket (question 5 & 6).  With all of these factors in mind have a look at the following compounds which have been designed against the M2 receptor.

Question 6 Give a brief explanation for the predicted binding for each compound and rank them in order of predicted potency (Note A is Iperoxo the compound you have studied throughout the previous exercises).

Part 2.

In the first part of the online workbook you have been looking at the different factors that will affect inhibitor binding.  In the second part you will use this knowledge to do some docking into the M3 muscarinic acetylcholine receptor.  For this you will use ezCADD which has been designed for docking small molecules and high throughput inhibitor screening. There is an introductory video in Minerva to the program which you should watch.

Task 7.  Got to

Do this in google or firefox as it does not work in internet explorer!

This will take you to the ezCADD webserver.  Accept the licence agreement and you will be met with the following screen;


Go to “File” in the top left hand corner and select this option in the drop down menu click on open and select the file “M2receptor.pdb” in the pdb folder. This should result in the following screen (the orientation may be different).


The following commands will be useful;

Left mouse button to move the molecule around. 

Right Mouse button to drag the molecule

Middle button to select and centre the image on a specific part.


Next you will do some docking.  Go to “Applications” at the top of the program screen and choose ezSMDOC.  For the receptor click browse and open the “M2receptor.pdb” which you have loaded as the model structure and for the ligand go to browse and the pdb folder and load the file QNB.pdb (This is the inhibitor shown in Fig.1 in Liu et al., 2018). For the docking box choose “use ligand centre” or enter 7.7991, 0.2598, -3.951.  Select the Vina program and click start.  

You will now have the inhibitor docked over the known structure.  Remember the program does not use the known bound inhibitor it is doing this just based on the pocket size and shape.  There should be several different docking poses shown on the left of the screen, each pose has been scored. Now load using “file>open” the M2boundreceptor.pdb file.  This will show the crystal structure of the bound ligand.  Use the eye symbol to turn show and hide the structures to see the binding mode.

Question 7. How well do the binding poses compare to the known binding pose? Dose the score ranking reflect the most likely binding position? 


Task 8. The paper by Liu et al., 2018 (in the papers folder) is an excellent example of using in silico design to modify an existing inhibitor to improve a particular property.  We will now look at how selective the QNB compound is.  To do this now dock QNB.pdb into the file M3receptor.pdb.

Question 8. Looking at Fig. 1 and Fig. 4 in Liu et al., 2018 how selective is QNB to the M2 and M3 receptor? 


The previous tasks will have shown how the same inhibitor can dock into both the M2 and M3 receptor and may cause significant toxicity effects.  Especially when we consider that the M2R modulates heart rate!  We will now try and dock the newly designed inhibitor called BS46.pdb in the folder into both the M2 and M3 receptors. In the first instance simply compare the two compounds by loading both into PyMOL and looking at Liu et al.

Question 9. How do the two compounds compare? 


Task 10. We will now finish the tutorial by trying to dock the second inhibitor (BS46) into both the M2 and M3 receptor.  Use the methodology for the QNB compound to repeat with BS46.

Question 10. What differences do you see between the binding of the two compounds? 

Post Workshop Task

Although the questions are stated within this workbook, please fill in the answers in the separate answers book which can be submitted through Turnitin

Using the skills developed in the online workshop you are now going to compare and contrast a range of inhibitors towards the enoyl reductase enzyme of both E. coli, Toxoplasma and Plasmodium.  One of the best studied inhibitors of enoyl reductase is triclosan which is a broad spectrum antimicrobial that you will find in toothpaste, mouthwashes and other antibacterial cleaning agents.  Here you will look at how this compound has been developed along with other related compounds in order to re-purpose as an antimalarial compound.


Task 1. Within the ENR folder you will find pdb files for three different structures (E.coli (Ecoli_ENR_tric.pdb), Plasmodium (PlasENR_tric.pdb) and Toxoplasma (Toxo_Tric_like.pdb)).  There is also a PyMOL session file ENR.pse. If you load the session file E. coli ENR has yellow carbons (cyan for the triclosan inhibitor), Plasmodium ENR is in magenta and Toxoplasma is in green.  Using the skills picked up in the workshop answer the following question related to the ENR enzyme.


Question 1. What are the main features of the triclosan binding pocket that contribute to the high affinity for this compound? (5 marks


Question 2. The NAD+ cofactor is required for triclosan binding, how could it contribute to the tight binding and how might this be utilised in the design of a potent inhibitor? (5 marks) 


Question 3. The apicomplexan parasites (Plasmodium and Toxoplasma) contain an Alanine residue at the base of the pocket, whereas the bacterial homologues have a Methionine residue. Looking at the Toxo_Tric_like structure what is the primary difference caused by this mutation in the binding pocket and how does the modified scaffold take advantage of this? (10 marks)


Question 4.  The Tyr156 residue in ENR is essential for triclosan binding, why might this be and why would a replacement to Phe result in a reduction in potency of triclosan? (10 marks)


Question 5. Below is a figure of ENR in complex with a diazaborine residue which is covalently linked to the NAD+ cofactor (shown in dark blue sticks).  What sort of Kd value would you expect, high or low and why? (5 marks).


Question 6 Why might toxicity be a problem with covalently bound drugs? (5 marks) 


Question 7 Below is a series of 10 related compounds to triclosan.  How might the modifications affect the potency of the compounds and which would be selective towards Plasmodium ENR over E. coli ENR?  Use the ENR structures and PyMOL (ENR.pse) to guide how these inhibitors may or may not bind. (25 marks)


Question 8  A key aspect of triclosan is its “slow-tight” binding which is due to the formation  of a helix over the inhibitor as shown in the figure below (residues 236-246, shown in magenta, triclosan in green, NAD+ in orange).  Why might the formation of a helix result in an increase in potency of the small molecule? (5 marks) 


Question 9. What are the typical advantages of on in-silico approach over a High throughput screening approach and what would the limitations of the in-silico approach be? (30 marks) Max 500 words


Hints and Tips

Useful Commands, load the pdb files and “hide” lines and “show” cartoon.  If you click “hide” nonbonded this will remove the waters (red dots). 

If you go to “display> sequence on” you will see the corresponding sequence by clicking on any residue, NAD, TCL, NJ8 you can show the different inhibitors.  Alternatively, you can select one and centre using middle mouse or action button.

If you have show sticks for all residues it us hard to see anything, therefore can use “display>Zoom >8angstrom slice”.  Or “display>clip>12 angstrom slice. This relates to the view so 20 angstrom slice or view will let you see everything within 29 angstrom, 8 angstrom will be a smaller window.


Study Cred Tutor

4.6 (24k+)

Purchase the answer to view it



Click one of our contacts below to chat on WhatsApp

× How can I help you?