Simulating bacterial membranes
“Using Raijin I simulate a simplified model of the lipids that form bacterial membranes,” explains Dr Poger.
“I take a patch of membrane and change the molecules from one model to the other and see how that affects some of the membrane’s properties.”
Membrane lipids are composed of a ‘head’ and two hydrocarbon ‘tails’.
“The heads have been studied a lot. But some aspects of the tails have not been studied that much,” says Dr Poger.
“What I’m looking at is the chemical modifications that occur along the tails, and how these modifications can change the properties of the membranes.”
These properties include measures such as how tightly packed together in the membrane the lipids will be, and whether the tails will tend to be more straight or curved.
“Also things like how fast the lipids can move – or diffuse – within the membrane,” says Dr Poger.
Such modifications – which can be as simple as an extra CH3 group, or a CH2 bridge joining two tails together – can have significant consequences for human health, says Dr Poger.
“For example, tail modification allows Listeria to be very happy in the fridge,” he says.
“If you think of bacterial membranes as being composed of something butter-like, if you put it in the fridge the membrane will be quite solid, which is not a good thing for bacteria.
“But the thing about Listeria is it has these modifications in the tails so that the ‘butter’ in the membrane is basically staying liquid; it’s happy.”
Membrane lipid tail modifications can also help Salmonella and E. coli survive the effects of stomach acid.
“The stomach is very acidic and can kill nasty bacteria that contaminates your food,” says Dr Poger.
“But thanks to modifications in its membrane tails, Salmonella can resist the acidity of the stomach. It can survive and move to the intestines which are much less acidic and the bacteria are very happy.
“Legionella can also become resistant to some antibiotics when you play around with this modification of the tails.”
Dr Poger’s simulations require enormous compute time.
“Even though on the human scale bacteria are very small systems – we’re talking about micrometre scales – in terms of computation, the model membranes simulated are small nanometre-scaled patches that are still tens of thousands of molecules and hundreds of thousands of atoms,” he says.
“What we do is basically let the atoms move free like they would in real life.
“We accumulate short movies of how these atoms move and for each movie we need to calculate all of the interactions between all of the atoms and that is very computationally expensive.
“A snapshot of a movie just a few picoseconds long can take up to several minutes to calculate and a typical simulation involves millions of frames of half a microsecond long.
“I typically use 16 of Raijin’s CPUs for up to three months straight.”