blog

When you want to know how a protein works (and sometimes, if other avenues have been exhausted, what it does), the usual approach is to solve its atomic structure. You can do this by X-ray crystallography or nuclear magnetic resonance spectroscopy, but whichever you choose (rather, whichever actually works for you) you’re going to need a supply of protein. The most common way to make loads of protein from a eukaryote is to clone the gene behind an appropriate promoter in a suitable plasmid, and stick the plasmid into every scientist’s favourite bacterium, Escherichia coli K12.

And a major problem there is that making proteins in E. coli, while fantastic for producing large amounts of some, cooperative proteins, sometimes just doesn’t work because the foreign protein fails to fold properly, and you’re left with a mass of soggy, mis-folded goo at the bottom of a centrifuge tube.

There are a number of ways to get round this, fortunately, from trying to resurrect protein from inclusion bodies using a chaotroph such as urea, to tricking the protein into folding correctly in the first place. This is obviously preferable, and sometimes can be achieved simply by using a protein from a different species, or by clever engineering of the sequence. Another way, which seemed to seep into scientific consciousness around the end of the ’90s, involves allowing the bacteria to grow at a lower temperature, thus reducing the rate of all cellular processes including translation, which gives, if you’re lucky, the protein you’re interested in time to fold properly: protein translation in bacteria is really fast (up to about 20 amino acids per second), and eukaryotic proteins can get muddled up.

So, when I was making proteins for structural studies I’d try growing the cultures at 37°C, and if I didn’t get soluble protein, try again–usually at 34°C, 30°C, 25°C, and, in extreme cases, 20°C. Not unusually, it would work. However, at low temperatures it’s not just translation that’s slowed down, but all cellular processes. This means that it would take up to two whole days for the bacteria to grow enough to produce sufficient protein that it was worth harvesting them. Ordinarily, at the two higher temperatures, I’d set up the cultures as I went home in the evening and come back in to lots of lovely protein-laden bacteria to harvest in the morning. Waiting the extra day and a night was a real drag, especially if the purification process then took another full day.

Now you can understand why three of our F1000 Members got excited about a paper in J. Mol. Biol. last year, Slowing bacterial translation speed enhances eukaryotic protein folding efficiency^{10.1016/j.jmb.2009.12.042}. Here, rather than simply reducing growth temperature, the authors used a strain of E. coli that had ‘hyper-accurate’ mutations in their ribosomes. These mutations prevent the ribosomes from making as many mistakes during translation as the wild-type, but at the cost of slowing translation to about 5 amino acids/second. The clever thing is that streptomycin, an antibiotic commonly used in cell culture that kills bacteria by causing mistakes in protein synthesis, can relieve this restriction so that translation proceeds at the wild-type rate (with almost wild-type levels of accuracy). So the authors were able to slow down or speed up translation at will, without affecting cell growth and metabolism overall–and found that eukaryotic proteins when translated slowly were more likely to fold correctly.

We believe that our findings provide a general strategy for the production of recombinant proteins that does not rely on the individual manipulation of coding sequences or on the introduction of specific accessory factors.

–Siller et al., 2010.

Whether this will indeed turn out to be a general strategy for the production of recalcitrant proteins depends on other researchers using the method successfully. Let us know if you, or anybody you know, has been able to try it out.