Like Hubble, but on the inside.

Here I am trying to write a science blog that I hope will interest my friends without using the word 'huskily' taking you right up to the cutting edge of some current science - even if you don't have a scientific background. So if you agree to nod and smile at any strange words I use (they may become friends in time) I think it's only fair if I give you a bit of context. Where is all this wonderful biochemistry happening and why is it so damned exciting? Cells is my answer, cells and proteins. The Nobel committee thought so too, honouring these three chaps for development of green fluorescent protein (GFP). GFP has allowed us to see what goes on inside living cells in astonishing detail, moving us away from diagrams of dried-up dead things to photos and 3D maps in living organisms. I'd like to introduce you to one protein, tubulin, and use it to illustrate cell structure and the very clever imaging technology we use to visualise it

Let's kick off with a bit or revision on the eukaryotic cell, as you are living in several billion of them right now. The text book version looks pretty much like this image I've borrowed from wiki.

There's a membrane (7) enclosing various organelles (9,6), the nucleus (1,2) and several other numbers you can safely ignore for now. And there we have it; life laid bare. Speaking in a generic way, one of these babies is quite capable of living by itself as a single cell organism - yeast for example, or of course making up magnificent beasts such as you and I. But this is all pretty dry stuff and the tubulin I've promised isn't even on Wiki's diagram but looks a bit like number 13 -and forms part of the cytoskeleton.

The cytoskeleton, as its name implies, can be thought of as giving structure to the cell - a kind of fibrous mesh. But it's not a static scaffold, rather it is constantly moving and being rebuilt; responding to changes. How is does this is really remarkable. Textbooks tell us that the cytoskeleton is made up from three types of fiber, microfibres, intermediate filaments and microtubules. I'll even drop the intermediate filaments from the discussion lest it turn into a lecture and just talk about microfibers and microtubules. Don't say I'm not good to you. These 'fibers' are pretty simple, they are composed of repeating units of proteins; actin for microfibers and tubulin for microtubules. A bit like lego:

But even cooler than Nathan Sawaya (lego art work above) is that microtubles are self-organising. Get enough of the 'tubulin bricks' together, give them a shake (under the right conditions) and off they go, zipping up into hollow tubes. It's as if the man above could build himself.

Image shamelessly borrowed from this paper

Ok I fibbed a bit, there are actually two forms of tubulin (alpha and beta) and in a cell there are plenty of other proteins involved, but you get the idea. Microtubules are directional structures in that one end grows faster than the other and therefore they can move about - quite fast too - by growing. Occasionally, perhaps at the behest of other proteins, they suffer catastrophe and unpeel like a banana. As well as being self-organising, microtubules and the cytoskeleton can respond something that might threaten the cell, a poke with a very small stick for example. This paper shows how the cytoskeleton of a plant cell responds to such a threat - within minutes you have serious rearrangement of microfibers, microtubules and even the endoplasmic reticulum (the pink bit in the wiki cartoon, number 5) gets in the act. Their paper is open source so I hope they don't mind me stitching a couple of their images together for brevity. And open source authors rock - so thanks guys! BTW these are plant cells which are very different to animal cells. Still they are eukaryotes too and I will recklessly carry on...

The real threats are pathogens (fungi, bacteria and oomycetes) which often try to penetrate cells (without consent!) to feed and grow to the detriment of the plant. In this case the cytoskeleton seems to organise the defence to just to spot attacked - and very fast too. How can bundles of proteins 'know' when to realign themselves? I don't have an answer to that question; it's still work in progress. I would love to know. Remember that there's no central organising 'brain' in a cell; LegoMan just can't compete.

Here's another image of eukaryotic cells showing the cytoskeleton (green) with the nucleus (blue) and membrane (red).

Well, this looks nothing like the lighting fast revision I put you through, indeed not; actual cells are interesting and beautiful especially when viewed with the magic of GFP and confocal microscopy. Not convinced? Well have a look at more images. The green colour comes, of course, from GFP, a small protein derived from jellyfish. (The baby, however is inexplicable.) With a bit of work one can add GFP as a tag to a favourite protein, in the images above tubulin was tagged, when you shine a blue light at GFP if fluoresces green allowing you to see where your protein is - even in living cells. There are now several colours and variations on original GFP tags (a particularly clever group made brainbow where they created a staggering 90 colours out of three florescent tags to identify individual neurones in living mouse brains. Wonderful stuff but sometimes it's hard luck for the mouse).

As if this wasn't enough, pioneering researchers like Stefan Hell are making nanoscopy feasible. A long-standing dogma had been that light microscopy was fundamentally limited by the wavelength of light. To put that in perspective, tubulin is 8 nm (0.000000008 of a meter) and microtubules 25 nm wide, while the resolution of pretty much all light microscopes until very recently has been an oval of 200 nm by 400-700 nm. This means that if your tagged proteins are closer together than this limit they will look like one blob, not two. Don't get me wrong, confocal (light) microscopy is amazing; it has been simply revolutionary to be able to visualise proteins in living cells. But nanoscopy is jaw-dropping we are now able to visualise single molecules in living cells. There are several techniques being developed (STORM, PALM etc.) but basically they all flip the fluorescent tags on and off on a single molecule basis and thus get around the diffraction limit, achieving resolutions of less than 20 nm by 50-70 nm. I leave you with one last image from a fine open source paper as a taste of things to come. Oh it's actually even better than this, ironically I've ruined the resolution whilst 'borrowing' the image. These are microtubules viewed on the left with confocal microscopy and on the right with nanoscopy. We can now clearly see individual microtubule strands!

I am so excited by these developments because getting around the diffraction limit of light delights the geek in me and there is so much we don't know about how cells and proteins function to create life. We can recognise most of the parts now, but there's no one bit that holds the essence of how that cytoskeleton responds to a poke. It's not a tic-tok mechanism but an emergent property of a complex system filled with precision nanoscale machines. Your cells are infinitely wondrous and now we can see a whole new order of magnitude opening up before us. The next few years will be beautiful.

Pox foxed!

If you are not surprised by evolution sometimes then perhaps - as Bohr so famously said of quantum mechanics - you haven't really understood it. I must admit that I frequently take evolution and selection pressures for granted. So it was delight to feel a frisson of wonder when reading a paper by Elde et al .

When to make proteins - and which ones - is a critical decision for a cell, and one which must be constantly under revision as circumstances change. One protein, eIF2a, is a central regulator of protein synthesis. As eIF2a is an essential part of the machinery which synthesises proteins it makes a good 'checkpoint'. Of course the 'machinery' aka ribosomes, are vastly complex and amazing entities and there are plenty of other regulatory and checkpoint mechanisms - but for now we'll just look at eIF2a.

The activity of eIF2a is controlled is by phosphorylation; the reversible addition of a little phosphate group by a kinase (which is a protein too). So phosphorylation acts like a on/off switch. A kinase called PKR can add phosphate to eIF2a and halt protein synthesis. Here's what these two proteins look like in a cartoon format:

PKR is shown in green on the right hand side and eIF2a in blue on the left. Note the little point of contact between the two a helix sticking out of PKR and the barrel of sheets in eIF2a. The crucial phosphrylation, by the way, is on one of those two strands which cross over the barrel. This imlies that PKR must coil around eIF2alpha to reach there.

PKR helps to defend a cell from viral attack, it is activated by the double stranded RNA found in several types of virus, and then rushes off to phosphorylate eIF2a and shut down protein synthesis. This works brilliantly as a defence because many viruses are nothing more than a bit of genetic material (DNA or RNA) coated with protein. Well, it did work brilliantly until some canny viruses - the pox amongst them - evolved proteins which mimic eIF2a and shut down PKR by binding tightly to the defender. If PKR is out of action then the cell can no longer shut down protein synthesis and thus avoid being hijacked by the virus. The pox virus has such a mimic protin called K3L and, as you can see below, K3L looks quite like eIF2a. The virus protein has only got that little barrel and a bit of a neighboring helix but it is enough to do the job of binding to PKR.

Well Elde and co. figured that PKR has a problem; it must be able to recognise its correct target, eIF2a, but would like to avoid binding to pesky viral mimics such as K3L. The structure of eIF2a can't change easily because it has such an important role in a complex machine - initiating protein synthesis in the ribosome - and therefore must fit snugly like a jigsaw piece. In fact eIF2a is one of the most highly conserved proteins, for example, the amino acid sequence of eIF2a of yeast and humans are astonishingly alike. So the virus in wants to copy eIF2a to fool PKR. Such an interaction is often described as an arms race. Sometimes K3L binds to PKR and the virus 'wins' and smallpox rages across countries, wiping out everyone with PKR proteins that fail to spot the intruder. Many of the survivors will, by necessity, have a variant form of PKR which can discriminate between the correct eIF2 and the dangerous decoy. To survive in the population the virus has to re-adapt the mimic protein. In midst of a smallpox epidemic having the 'right' form of PKR is a big advantage and this is a fundamental bit of evolution, the virus is applying a selection pressure on PKR. It's adapt or die.

If we look over evolutionary time we should be able to figure out which parts of PKR are responsible for recognising the mimics. How can we do this? Well, all proteins change slowly over the generations due to little errors made when copying DNA. For many parts of a protein this change may not be very important; it can still do its job well enough. If the mutation hits an important part though, such as eIF2 binding to a ribosome, well then that unfortunate individual will die (or fail to reproduce; this is the same to evolution) and the mutation will be lost. The important part of the protein - critical amino acids - will remain more unchanged than one would expect by chance alone and such parts of the protein are said to be under 'positive selection pressure'. If an external agent - such as a virus - applies the pressure then, somewhat tactlessly, biologists call this 'purifying selection'.

To find out if such selection occurs and which parts of PKR might be important in the eIF2a vs. K3L story Elde looked at DNA sequences of PKR from 20 primate species which cover about 30 million years. (Don't worry it's very unlikely any primates were killed for this work, a small tissue sample would be more than enough. In fact the information is probably already sitting in genbank .) They found that amino acid residue which line the interface between PKR and eIF2a have been subject to intense episodes of positive selection. In other words PKR in our bodies today still bears the marks from devastating epidemics thousands - or hundreds of thousands - or years ago when ALL variant forms of PKR were wiped out, along with their unlucky host. I find it astonishing that a disease - a little virus - can shape our proteins, indeed shape our evolution. I got kind of dizzy thinking about waves of epidemics and so much death - and yet here we are with hundreds of thousands of proteins making up our bodies, all shaped by various forces, many carrying remains of past evolutionary battles but still we rattle along.