3: Workshop: predicting lipophilicity using computational chemistry

Introduction

In this workshop, we are going to build on what you know already about electronegativities, polarity, and hydrophobicity, and make predictions about how our analogues will compare with the initial medicine (something we call our ‘lead compound’).  

In the lab, we are going to test a parameter called lipophilicity. Lipophilicity is a measure of how fat-like a molecule is (think oils, butter…all those things that don’t dissolve in water). We measure lipophilicity by measuring how much of our compound dissolves in a mix of two immiscible solvents – octanol (non-polar or fatty) and water (polar). We call this the partition coefficient or LogP, as the compound will partition or separate between the two solvents depending on how non-polar or polar it is.  

As medicinal chemists, we use LogP to tell us how well a lead will pass through the cell membrane which is made up of lipids, making it a non-polar barrier. In particular, we look to keep the LogP number below 5, following Lipinski's rule of 5.  

About the workshop: 

Work through this workshop at your own pace! It will be easiest to spend a few 1-hour sessions on this content, rather than attempting it all at once.

Session 1: Chemistry knowledge review 

Start by reading the introduction and making a list of definitions of important words or concepts, including those you might not be familiar with. 

The following videos might help you understand some important concepts we will be discussing:  

Session 2: Using Molinspiration to view polarity of molecules

Now that you are familiar with LogP and hydrophobicity, we are going to look at what factors make our molecule more or less hydrophobic/lipophilic. The trend is not entirely straightforward, as many things influence a molecule’s permeability (ability to pass through the cell membrane), including:

1. Fat-likeness or polar/non-polar surface areas; for example, how much of the area around the molecule is polar or non-polar

2. Acidity and basicity of certain functional groups

3. Local dipole moments as driven by electronegative atoms (e.g. O, N, F, Cl, S)

4. Flexibility of the molecule (we won’t go into this, but if you are interested, undertake some independent research!).

Lipinski's rule of 5:

Lipinski's rule of 5 is used as an indicator of whether a drug will be a good for oral administration in humans. It is based on the chemical and physical properties of the proposed drug. A good orally active drug should follow this criteria:

There are of course exceptions to these rules, however they are a useful guide for drug design!

Why do you think it is called the "rule of 5"?

Take aspirin, for example. Aspirin contains 3 main functional groups: a phenyl ring, a carboxylic acid, and a methyl ester (or acetyl group). The phenyl ring is non-polar, but the acid and acetyl group are polar. Aspirin has a LogP of 1.19, meaning it only slightly prefers the octanol phase to water.

Remember

where C is concentration of the compound of interest in solution.

Look at aspirin below, with the functional groups marked out. Beside it is an image showing different coloured dots around the different functional groups. You can see that the dots around the acetyl groups and carboxylic acid are yellow/orange/red, which are the colours used to indicate more polar areas of the molecule. You will be plotting images just like this in this pre-work!

Designing analogues

Now we are going to design some analogues of aspirin. Analogues are compounds that have a very similar structure to aspirin, but have a few small differences. We take our lead and make small changes and test the results to see if the analogue works better or not.

Have a look at the image below. You can see our lead molecule, aspirin, but it has R1 and R2 on the end of where the acid OH and methyl group CH3 should be. We use R to represent the parts of the molecule we will be changing. You can see in the box next to aspirin some suggestions for different groups for the R1 and R2 positions on aspirin.

What we’re going to do in this session is use a computer program to draw in the 2D structure of aspirin with some minor changes. Then we’ll use the program to estimate how much more polar or non-polar these changes make our molecule. We can link the polarity back to how this might affect the molecule crossing the cell membrane into the cytosol and therefore how well our medicine works. The changes we are making can be whatever your heart desires (within reason). The "R groups" are just suggestions, so be creative and find sub-structures of molecules that you might be able to add to aspirin. (Try searching 'R groups' or 'chemical substructures' on Google for ideas).

Now, let’s work through the program slowly…

How to use Molinspiration

Molinspiration is a free online research tool that can predict LogP and overall hydrophobicity. It may look a little old and clunky, but it is pretty reliable for the type of work we are going to do in the lab.

1. CLICK HERE to go to Molinspiration. Then click on 'Free Web Tools for Cheminformatics Community' in the light blue box.

2. Start by building your desired molecule! Try aspirin first. You will have to play around with how to use the drawing tool; it is easiest to build the carbon structure first, and then put in any double bonds and other atoms.

If you get stuck, check out the video below for a basic tutorial:

3. Once you’re happy with the analogue you’ve designed, generate a 3D image by clicking “Galaxy 3D generator”. Pick dotted rendering.

Let's look at some of the "surface colouring" options - 

i. PSA - shows polar surface areas in red, and non-polar surface areas appear grey. What colour would we expect the surfaces surrounding the oxygen atoms to be? 

The surfaces around the oxygen atoms appear red, because oxygen is more electronegative than carbon! 

ii. MLP: is coloured by the lipophilicity potential of each region - essentially shows us how lipophilic or lipophobic the parts of the molecule are. Fat-like (hydrophobic and lipophilic) appears blue/green, and water-like (hydrophilic and lipophobic) appears orange and yellow. 

Class discussion:

Take a closer look at the PSA and MLP surface plots.

1. How large are the polar surface areas, compared to the non-polar surface areas? How do you think this would affect cell membrane permeability?

2. How much of the molecule is fat-like, and how much is water-like? Comment on what this might mean for our LogP value.

3. Where are the electronegative atoms? Where are the dipole moments, and in what directions are they pointing?

Go back to the 2D drawing of your molecule and click on "Calculate Properties". This will calculate a range of properties for your molecule, including LogP, TSPA, and the number of H-bond donors and acceptors in our molecule.

LogP values between 1.35-1.80 are good for oral administration. Remember that a LogP value over 5 means that a compound may not be able to be administered orally.

The TPSA value also provides valuable information for drug design. It is a measure of the polar surface area. Values over 140 mean that the molecule will probably not be very good at passing through the cell membrane.

nON is the amount of hydrogen bond acceptors in the molecule. We have 4xO atoms which can act as hydrogen bond acceptors.

nOHNH is the number of hydrogen bond donors in the molecule. We have 1xOH group which can act as a hydrogen bond donor (the H atom).

What do you think? Does aspirin follow Lipinski's rule of 5? Is this consistent with what you know about aspirin, and how it is administered?

In the synthetic part of the depth study, you will be investigating many different analogues of aspirin and paracetamol.  You can come back to this workshop at any time, and use Molinspiration to predict whether you think the analogues will be good oral drugs!