I meant to post this in the morning but I got side-tracked with the structure of the inhibitor. I think I have it but, boy, I have no clue how it can be produced enzymatically. Oh well, most probably it will be wrong. Many times in the past I thought I had it only to find out that I was wrong.
So let us go to part II of targets to drugs.
I was talking about using the bacteria to produce protein. Sort of like a factory. The only problem is that the biological systems are entirely unpredictable. So when we put in a gene into the bacteria and ask it to make the protein, it entirely depends upon the bacteria. If it likes the gene and the protein, it will make tons of it. If it does not like it, it will not make it. Sometimes, it will make the protein but the protein might be inactive. There are numerous reasons for this. The bacterial proteins and the eukaryotic proteins differ in many respects. One aspect is what is known as post-translational modifications. Proteins are made of amino acids linked together by peptide bonds. After this is made, the cell can add little pieces of decoration like phosphorylation, glycosylation, ADP-ribosylation, sumoylation etc. etc. The bacterial cell hardly ever does this. However, all eukaryotic cells do this bit of extra decoration. Sometimes these decorations are essential for function, and at other times they are not. The only way to know what is essential and what is not essential is to do the experiment. So one takes the gene and puts it into bacteria and ask whether the protein is produced or not. And whether the protein is active or not. And whether the protein can be crystallized or not.
When I was working for a pharmaceutical company, I came across such a case. The protein could be produced in bacteria but could not be crystallized. We had to move to another system to produce the proteins.
Bacteria are the simplest and the easiest. But if the bacteria does not work, we move through the chain- yeast, baculovirus, and mammalian cells. Each one is tougher and more expensive than the previous one- Mammalian cells are the most expensive, and require lots of care, and really not worth it unless one has no options.
Once we have figured what which of the system works the best for us in terms of protein production and activity, we have to move to the next thing.
Discovery of small molecule inhibitors for the protein target can be done in two ways. One can use them for high throughput screening or one can crystallize the protein and do in silico protein-drug modeling.
All proteins are not really amenable for high throughput screening. Basically in this type of screening, you have a protein whose function can be estimated easily. For example, I work with an enzyme that hydrolyzes ATP, the energy currency in a cell. When ATP breaksdown it releases inorganic phosphate which can be estimated by a method known as Fiske-Subbarao method. Therefore, if you have a protein that can be purified in large amounts and its function can be easily estimated then you can put it through a high throughput screening. In such a screening, you test a variety of chemical compounds and see which compound inhibits the activity of the protein. When you find such a compound, you have a hit.
Instead of doing this laboratory thing, you can do in silico method too. For this you need to solve the structure of the protein. This is another bottleneck. As a colleague of mine put it succinctly, proteins do not like to be crystallized. They were not evolved to do that. Therefore, you have to find conditions and coax the protein to crystallize. This can take time. There are some proteins that you cannot crystallize at all. But assuming you can crystallize and have the structure, you can use in silico programs to model small molecule-protein interactions and see which one of them will bind to the protein. When you find such a molecule, you ask the chemists to synthesize it and test it. The problem with in silico modeling is that you are trying to mimic a 3-D molecule in 2-D space. Does not work! There was much hype about this but slowly the realization has sunk in that the protein in space can not be mimicked on a computer screen.
Assuming everything works and we have a hit, we have to move through steps to make sure that the hit can indeed be developed as a drug molecule.
Next lecture!
No comments:
Post a Comment