Cells come into focus
Glowing molecules can be distinguished one at a time.
Molecules, students are often told, are too small to see with a light microscope. But not any more.
Last week, a team in Germany unveiled microscope images of living cells that show details at the scale of individual protein molecules1. And now a group at the Howard Hughes Medical Institute (HHMI) in Ashburn, Virginia, has developed a second way to do much the same thing2.
These techniques should give biologists a much better view of what goes on inside cells at the molecular level. Ordinary light microscopes can't make out much below the scale of a whole cell, and certainly can't see single proteins, which are often just tens of nanometres across; details smaller than about 200 nanometres are fuzzy.
Yet to fully understand how a cell organizes its biochemical processes, researchers need to be able to follow these individual molecules as they go about their work. "What we don't understand is how the proteins interact with one another, and how those interactions drive the cell," says Eric Betzig, leader of the HHMI team at its Janelia Farm Research Campus.
Electron microscopes have keener vision than light microscopes they can see down to scales of just a few nanometres. But the electron beams in these devices quickly fry biological samples, and so living cells can't easily be studied this way. Also, says Betzig, the images are indiscriminate: "You see all sorts of structures, and don't know what's what."
One way of beating the resolution limit of light microscopes is to bring the light source very close to the sample, and to measure the light that is reflected or emitted at similarly close quarters. This approach, called near-field optical microscopy, has increased the resolution to around 50 nanometres.
But the two new techniques use other approaches to pick out even finer details.
Making individual molecules light up isn't difficult in itself, says Stefan Hell of the Max Planck Institute for Biophysical Chemistry in Göttingen. If a molecule is made to be fluorescent, it will glow when light is shone on to it in a way that can be picked out by an ordinary light microscope. 'Light-bulb' labels of fluorescent protein or dyes can be introduced into cells to track other proteins in this way.
But if there are several such glowing molecules close together then they will tend to blur into one another, because the glow is bigger than the molecule itself (just as the light of a star in the sky is much, much wider than the real 'size' of the star in our field of view).
A decade ago, Hell proposed a solution to this. He showed theoretically that one could shine a light on a molecule, and then suppress much of the fluorescence with a different-frequency, donut-shaped beam of light. This would create a ring of darkness around the fluorescence, squeezing the light coming from the sample into a much smaller spot.
It took Hell until last year to show that this technique could truly sharpen a fluorescent spot down to just 60 nanometres or so3. Now he and his co-workers have tightened it to less than 20 nanometres across, and have shown that the method works in living cells1. His team has used the technique to gain new insight into the process of neurotransmission, by following the fate of neurotransmitter molecules4, and has proved that it can be used with 'light-bulb' proteins5.
Betzig's group at the HHMI has taken a different route, which Hell calls "complementary" to his own. Instead of reducing the size of the glowing spots from single fluorescent molecules, they have been able to light up, at random, just a small proportion of the many fluorescent molecules in a sample.
This is done by using an 'activating' light beam, which is so dim that individual photons would randomly switch just a few of the molecules to a fluorescent state. A second beam then stimulates fluorescence in just these molecules.
This means that, even though the individual spots are big, they are sparsely scattered and so unlikely to overlap. The 'activated' molecules stay glowing for long enough to take an image, but eventually switch off again. By taking many such frames, the researchers build up a composite image of the sample, piece by piece. "Eventually you pull out every molecule", says Betzig. The team calls its method photoactivated localization microscopy (PALM).
"It's a powerful and clever idea," says Hell. One drawback of PALM, however, is that it takes a long time typically several hours to take all the snapshots needed for a complete cÐomposite image. At the moment, that limits its value for studying cell processes that change over time.
But the researchers think that they could speed this up both by increasing the number of molecules 'switched on' in each image and by making them brighter, so that each 'exposure' is quicker.
Visit our newsblog to read and post comments about this story.
- Donnert G., et al. Proc. Natl Acad. Sci. USA, 103. 11440 - 11445 (2006).
- Betzig E., et al. Science, doi:10.1126/science.1127344 (2006).
- Westphal V.& Hell S. W., Phys. Rev. Lett., 94. 143903 (2005).
- Willig K. I., et al. Nature, 440. 935 - 939 (2006).
- Willig K. I., et al. Nature Methods, advanced online publication doi:10.1038/nmeth922 (2006).