3D drug design
Adapted at <![if !vml]><![endif]> from <![if !vml]><![endif]>
Atrolysin, the venom of the Crotalus atrox, better known as the western diamondback rattlesnake, is very nasty. It is responsible for the largest number of deaths by snake bites in the United States of America per year. The venom, a metalloproteinase, is an enzyme that breaks down the protective coating of blood vessels. This causes serious internal bleeding in the victim, leading to death. Needless to say, a suitable antidote is urgently needed.
In this bioinformatics practicum, we will design such an antidote by looking at the 3D structure of atrolysin in a number of exercises. The spatial structure of a protein is very important for its function. Understanding it is the key to solving problems caused by malformed proteins (like we find in many hereditary diseases) or, like in our case, by unwanted malignant proteins in our bodies. You need some background knowledge of protein structure, amino acids and molecular interactions like hydrogen bonds for this practicum. If you think you do not have sufficient knowledge (yet), you can refer to the theory section for a quick introduction.
If you did the murder at the airport practicum, you have seen the primary structure of the protein: the amino acid sequence. This primary structure leads to a specific tertiary (or 3D) structure, which is very important for the function of the protein. A protein with the wrong primary structure and hence the wrong tertiary structure may work poorly or even not all. However, small changes in primary structure do not necessarily lead to an inactive protein. Some amino acids are more important than others in the protein. Extensive mutation studies and so-called multiple sequence alignments are used to investigate which amino acids are most important, and which ones are less important.
Explain why systematically mutating all the amino acids of a protein, one by one, can resolve which amino acids are important. Answer
In enzymes, the most important amino acids are located in the active site of the protein. This is where the actual enzymatic reaction takes place. Later in this practicum, we will look for the active site of the venom. But let's start with some basic exercises in 3D visualisation.
SRS is a data integration tool. In the SRS 3D version, it includes a graphics program (viewer) used to visualise and manipulate protein models. For any protein sequence, SRS 3D shows all related structures; with one click the sequence can be mapped onto the structure; one more click and the structure is colored by features, e.g., domains, SNPs, or posttranslational modification sites. SRS 3D was first developed at Lion Bioscience, and it was open sourced by BioWisdom in 2008. After the practicum, you will have a chance to look at several other interesting proteins. You can also use the program to make nice pictures for reports and science projects. An elaborate description can be found in the User Manual.
Protein structures are collected in the protein data bank “PDB”. Here we will first look at a small peptide form the human prion protein (that forms plaques leading e.g. to Creutzfeldt-Jakob disease). A structure of this peptide has been studied experimentally. The entry name of this structure in pdb is “1oeh”.
<![if !supportLists]>Š <![endif]>Go to SRS 3D and type “1oeh” into the search field.
<![if !supportLists]>Š <![endif]>This will bring a view of the pdb entry “1OEH”.
<![if !supportLists]>Š <![endif]>Clicking on the image of the peptide will bring up the SRS 3D viewer applet.
You can now see the backbone structure of the peptide with sequence His-Gly-Gly-Gly-Trp-Gly-Gln-Pro in the so-called Ribbon representation. In this different levels of structure are represented as different size ribbons and tubes based on the location of the backbone atoms. This is very useful to get an impression of the overall fold of a protein and the arrangement of secondary structural elements. However, a small peptide like this does not have secondary structural elements. Therefore, we will look at the full atom representation:
<![if !supportLists]>Š <![endif]>Clicking the right mouse button gives you a context menu.<![if !vml]><![endif]>
<![if !supportLists]>Š <![endif]>Choose RepresentationąBall&Stick.
<![if !supportLists]>Š <![endif]>Get the context menu again and choose ColoringąElement
<![if !supportLists]>Š <![endif]>Finally activate the Mouse control help: choose ViewąMouse Controls
You can now see the structure of the peptide in the so-called Ball and Stick representation. Individual atoms are shown as balls, connected by sticks representing the atomic bonds. Below, you can see a 2D representation of the same peptide.
At the bottom of the window you can see the sequence bar. When you click on a residue in the sequence bar, the atoms of this residue will be highlighted. If you push Return, the peptide will zoom so that you can clearly see the residue. This residue now is the “center of rotation”. You can manipulate the protein by holding the mouse buttons and moving your mouse. Try the following:
Left: drag Rotate
Left drag + <shift>: Zoom
Right drag: Translate
The atoms are coloured by atom type. If you click on an atom, extra information about this atom will appear on the left side of the window.
Look up which types of atoms (elements) can be found in proteins, using this list of amino acids.
Look at different atoms in your molecule and find out which elements are coloured red, blue, grey, and white.
The ball-and-stick representation is useful to analyse the placement and connections of the atoms. If you would like to see, how much space the atoms actually occupy, you can look at a surface of the molecule:
<![if !supportLists]>Š <![endif]>Use <Ctrl>+A to select all of the molecule
<![if !supportLists]>Š <![endif]>Use the context menu Calculate ą Accessible Surface
Notice that the molecule looks like a globule without any holes to a solvent molecule.
You now know enough to look at the venom atrolysin, which may be familiar to you if you did the Murder at the airport practicum. Let's start by loading the 3D structure of the poison.
Go to srs3d.org and search for “atrolysin d”. This will initiate a database search for a protein with the name atrolysin d.
In the result list, find two proteins, one with an ambiguous name, and one which clearly says “Atrolysin D” in the description.
Notice the “Structure” column on the right, which contains little icons indicating that structural information is available for this protein. Click on that icon in the line with “P15167”.
The result page shows an overview of your sequence with indications for which regions structural information is available. At the top, you see annotations loaded from the Uniprot database (see the Murder at the airport practicum). If you move your mouse over the bars, you can see more details about the respective region of your protein.
Below, you see a graphics and below that a list of matching structures. The front part of the sequence does not have any structural information. Look at the Uniprot annotations to find out what is the function of this front part.
The color coding indicates the level of similarity of the related structures. Dark green means that the sequence of your query and the protein structure are identical. Select structure “1atl”. On the result page, click on the image to bring up the structure viewer.
You now see your found rattle snake venom sequence mapped onto the structure you selected. In this case, the amino acid sequence of the structure and the query sequence are identical. Hence, all the structure is colored green. In case the structure of your query had not been in the database, you could map the query sequence onto a related structure and yellow and red colors would indicate differences in amino acid type and hence areas where your query sequence might adopt a different structure.
Switch to “First Impression” view (via the context menu). Here, you can see that the structure contains two identical protein chains. Can you identify units of secondary structure?
You can also see that some atoms are not directly bound to the protein chain. What are those atoms?
For now, we want to hide the molecule called “SLE”.
<![if !supportLists]>Š <![endif]>Click on one of its atoms to select.
<![if !supportLists]>Š <![endif]>Use the up arrow (twice) to extend the selection to the whole unit.
<![if !supportLists]>Š <![endif]>Now uses the context menu ą Representation ą visible to toggle the visibility.
We will now look for the active site of the protein. That is where the most important amino acids are located and where the actual chemical reaction takes place. The SRS 3D viewer loads the active site annotation from PDB (the structure database) and Uniprot (the sequence database). Click on a lane called “act_site”.
This greys out most of the molecule and highlights one residue. How would you describe where the active site is located? Why is this type of location typical for enzymes? Answer
Now that we know the active site, we can search for an antidote. To do this, we use the fact that enzymes work accordingly to a lock-and-key principle.
Explain briefly what the lock-and-key principle is. If necessary, draw a picture. Answer
You might know (maybe from experience), that a lock can be made useless if you insert something that does not belong there. The key does not fit anymore and the lock is broken. Our antidote will work according to the same principle. We are looking for a molecule that binds so tightly to the venom that it cannot be released anymore. This will inactivate the poison. You have already seen interactions that cause the protein to fold in a particular way. The binding of the antidote depends on similar interactions. Therefore, one should look for hydrogen bonds, hydrophobic interactions and ionic interactions. Of course, the antidote must also fit in the cavity.
In the last part of this practicum, you will look at a couple of adapted versions of the ligand and decide which version is the best. The best ligand is the one that binds best to the protein. This may be a potent antidote...
Below are four different molecules. The first one (A) shows a basic ligand template, the other drawings (B, C and D) represent the version 1, 2 and 3 of our antidote.
Look at the different drawings of antidotes. Mark the atoms that may be involved in interactions with the protein.
Our structure contains a ligand. Which of these (A-D) has been tested for its fit to our protein?
Hint: In order to answer this question choose “Style ą Binding site” from the context menu. Here, most of the protein is dimmed out, but the binding site residues are shown in “lines” representation and the ligand in “ball and stick”.
How does that ligand interact with the protein?
Which ligand would you use as an antidote? Explain why. Answer
You have reached the end of this practicum, hopefully with the right antidote. That one is really used to treat rattle snake bites. So, in this case bioinformatics was used to help solve a medical problem. However, sometimes all we can do is try to understand how a disease works.
This exercise is based on the work of Bioinformatics@school, but has been modified to use SRS 3D for structure visualisation. We (<![if !vml]><![endif]>) thank the Bioinformatics@school project for their great work!
Bioinformatics@school was developed by the Centre for Molecular and Biomolecular Informatics (CMBI), Radboud University Nijmegen Medical Centre and the Netherlands Bioinformatics Centre (NBIC).