T-Cells: Beyond the Resolution Line

I have a diverse set of research interests - high-end microscopy, immunology, infectious disease, cancer, etc.  Its rare that a paper hits the "awesome" end of the scale in most of those categories, but this week Nature Immunology published a paper that got the nerd senses tingling.  In this tour-de-force, Mark Davis's group uses a new form of microscopy to analyse how T-cells work.

As usual, a bit of background first.

T-cells are the major regulatory cell of our immune system.  The express special receptors, called T cell receptors, which they can use to identify cells which have been infected by bacteria or viruses.  After detecting an infection, some T-cells (called CD4 t-cells, or helper t-cells) initiate and regulate the immune response.  Another type of T-cell (CD8 T-cells, or cytotoxic T-cells) go out and destroy infected cell.

The t-cell receptors (TCRs) themselves are complex things, with multiple parts (see pic on right).  There is the alpha/beta chains that detect the infected cell, and then the CD3 chains and the zeta chains which transmit the signal from the receptor into the cell, and the CD4 (or CD8) co-receptor which helps stabilise the interaction between the TCR and the target cell.  Upon engagement these receptors signal by recruiting proteins from within the cell, including one called Linker of Activated T-Cells (LAT), which acts as a scaffold for the rest of the proteins to bind to.

This paper studies the interactions between LAT and the zeta chain portion of the TCR.

The second cool part of the paper is HOW they looked at the TCR.  Microscopy is plagued with one major issue - there is a distinct resolution (diffraction) limit, below which we cannot resolve.  We've all experienced this ourselves, with our own eyes.  Think of driving at night.  When you see a car far off you see only one headlight (a, image to the left) - its not until the car comes closer that you can see two (c, image to left).  Where the one light becomes two is the resolution limit of your eye; microscopes experience a similar limitation.  Under optimal conditions this limit is 200-300nm, while proteins interact in spaces of 30nm or less, meaning we're lacking about 10X the resolution we need to study protein-interactions.

We scientists have a few tricks to get around this limitation.  This paper uses one of the newer of these tricks, called PALM.  The way this works is you use a photoactivatable dye - basically a florescent marker which needs to be activated by a specific wavelength of light before it becomes fluorescent.  The way PALM works is you use a weak activating beam to activate a small portion of the dye.  You then image the dye using a high-powered laser, and you image until all of the active dye is photobleached (the microscopy version of burning out  light bulb).  The resulting image will be a pattern of dots.  You repeat this process time-and-time again, and then mix the dot "images" together to get a single, complete image.

Normally this wouldn't produce anything other than what you would get if you just activated all the dye and then imaged it - you'd end up with nothing more than a resolution-limited image.  But there's a trick here - known as "math" - which lets us break that resolution limit.  Diffraction-limited dots have a specific shape, as you can see in the image above.  This shape is always the same, and the "tip" of the peak lies exactly over the fluorescent molecule.  So by mapping the peak of each dot, we can "break" the resolution limit and see much finer detail - in the case of this paper, down to 25nm!
So what did they find?

We've known for a while that many of the proteins in our cells membranes are not evenly spread out, but instead float around in little "islands".  The clustering of these little "islands" is often what activates these receptors.  But in many cases - like the T-cell receptor - we didn't know what was in these little islands, or what happened to them when they clustered.  There really were three options:

1. The TCR and signalling components like LAT are in the same islands, and clustering activates them through mass-action.
2. The TCR and signalling components like LAT are in separate islands which come together and mix; activating the receptor by mixing normally separate proteins.
3. The TCR and signalling components like LAT are in separate islands which come together but don't mix; activating the receptor by simply bringing things close together, but without actually mixing.

Davis's group has answered this question.  The first image shows the TCR before (left) and after (middle) activation.  You can see several small islands on the left, that come together into "super islands" on the right.  The right-most image is a control of randomly distributed particles, to show they are looking at islands, not spread-out single molecules.

LAT looks almost the same (image on right), with lots of small islands before activation, and fewer big islands after activation.  And while its not obvious when you compare the TCR image with the LAT image, the pre-activation TCR "islands" do not overlap with the pre-activation LAT "islands".

So that answers the first half of our question, LAT and the TCR are in separate islands before activation.  But do they mix, or are they wall flowers?
How this was demonstrated is hard to explain; but they used two mathematical measures to figure it out; Ripleys K-function, and cross-correlation, both of which measure how well two distributions overlap.  Without going into a lot of boring detail, the TCR and LAT cluster togeather upon activation, but the individual clusters of TCR and LAT remain separate; think of a cookie - both chocolate chips and peanuts are in the cookie, but the chips and nuts remains separate.
So that answers the second half of the question - we've got a couple of wall flowers on our hands.

I'm sure to many this doesn't seem that exciting, but from a biologists point of view this is quite the breakthrough.  PALM fast enough to image living cells, while also being able to resolve down small enough to see these little islands.  And they resolved something that's been a mystery since those "islands" were discovered nearly 30 years ago - how do they interact, do they mix, and what do they contain.  While this is just the tip of the iceberg - hundreds of proteins are known to be in these islands - we finally have to tools to start answering these questions.

Lillemeier, B., Mörtelmaier, M., Forstner, M., Huppa, J., Groves, J., & Davis, M. (2009). TCR and Lat are expressed on separate protein islands on T cell membranes and concatenate during activation Nature Immunology, 11 (1), 90-96 DOI: 10.1038/ni.1832