At present, it is technically impossible to acquire enough data in an image format using a microscope at microsecond rates, as well as deliver the information required to track single molecules. However, an innovative use of the microscope, used to focus a laser excitation spot to a tiny volume directed anywhere inside the cell, does allow very high data rates to be achieved. By examining molecular processes in one spot, there is no spatial information of course, but combined with all the other molecular imaging techniques, we can build an extremely detailed picture of single molecule behaviours. The basic idea behind this approach (Fluorescence Correlation Spectroscopy; FCS) is illustrated here.

The excitation volume (approximately 1 femtolitre) is 'parked' on a region of interest inside a cell. Single fluorescent molecules in and around that region then diffuse across the excitation volume, and emit photons as they pass. Because the rate molecules diffuse is proportional to their mass, the time in the spot varies with protein size. The formation of a protein-protein complex increases this size, so the molecules slow down. By using highly sensitive single photon counting detectors, with a very high temporal resolution, we can measure every photon emitted by each molecule as it diffuses through the excitation volume.

Images and data copyright LSI Laboratory

We can extract a lot of information about dynamic molecular behaviour from these data. A common analysis is to autocorrelate the data: this involved duplicating the trace, above, then comparing it to itself. Clearly, it is identical to itself at this point. Then, one of the copies of the data is shifted one time point (ie 1 microsecond), and the two traces are compared again. At this point they are slightly less similar to each other. This process is repeated, shifting one trace a single time point, comparing the data, shifting again, until the data have zero correlation. The width of each spike in the photon trace describes how long each molecule spends in the excitation spot, so the number of time-shifts in this analysis that it takes until the traces are completely dissimilar reflects this time course.

Acquiring data from inside living cells at the rate biological processes occur at - often on the microsecond (i.e. 1 million microseconds in 1 second) scale - is very challenging. Combined with the other molecular imaging approaches, our goal is to quantify the interactions, rate of diffusion, location and

The sort of data acquired is shown in the graph to the right. Here, we can see fluctuations as molecules enter the excitation volume (seen as spikes in the photon counts over time). As the key to detecting single molecules using this technique is analysing the fluctuations in this record, it is clear that if the number of molecules in the excitation volume is very high, the effect of a single molecule entering or leaving will be relatively diminished. It is vital therefore that the concentration of the molecules under scrutiny is kept very low - a low nanomolar concentration ensures that around 10 molecules are in the spot at any one time.

Fortunately, this is just about feasible when over-expressing fluorescent proteins in cells - the trace here is from a synapse in a living neurone, showing single molecules diffusing through over a 0.5 sec recording period.

The graph on the left illustrates autocorrelation analysis. This used the same data, acquired from a synapse, as the trace above. The correlation between the data is highest at lower time shifts (0.001 ms), decreasing until they converge to zero at around 100 ms. The half-time of this decay tells us the average rate of diffusion of all the molecules in the record. Normalising the result of the autocorrelation to the variance of the data means that the amplitude of the curve at ~t=0 delivers the number of molecules in the spot at any time.

Introduction        PALM       Single particle tracking PALM        FCS        FLIM        GSD Microscopy

concentration of protein molecules inside cells, thus increasing our understanding of normal cell physiology and hopefully, where and why things go wrong in disease.