With the completion of the human genome project, research effords are shifting from DNA to the gene products, the proteins. Only those features of the genetic blueprint DNA, which are eventually translated into proteins, will affect cell phenomena such as aging and desease. Therefore, protein chips are more informative than DNA chips.
Unfortunately, proteins are much more difficult to handle concerning chip technology. For example, while all DNAs can be attached to the chip substrate in the same manner, the binding chemistry for each protein on a chip has to be optimized indiviually. Even more critical and diverse is the interaction of proteins with specific ligands, such as antibodies, other proteins, aptamers or substrat molecules.
One of the methods of choice is the investigation of protein interactions by fluorescence lifetime measurements. To this end, the binding partners are labelled with chromophores with different fluorescence emission maxima. As the chromophores
get in close proximity, the shorter wavelength emitting chromophore (energy -donor) may transfer its excitation energy to the chromophore with red -shifted emission band (energy -acceptor).
This can either be detected by a change of color or by observing a reduction of excited state lifetime of the donor chromophore. However, this method requires labeling of
the proteins to be investigated, which has several disadvantages, for example an alteration of protein properties, losses of material due to the chemical reactions, or an uncertain binding location.
Therefore, we chose a modified approach by taking advantage of the intrinsic fluorescence of proteins. The amino acids tryptophan and tyrosin are known to exhibit a strong fluorescence around 300 - 350 nm upon excitation around 280 nm. Upon protein -protein interaction, this fluorescence may be altered due to energy -transfer, change in protein folding or for other reasons. With protein -nucleic acid or protein -substrate interactions, the change might be even more dramatic.
Most fluorescence decay times are on the order of several picoseconds to several nanoseconds. To measure such decay times, one has to utilize a very short light pulse, usually from a pulsed laser, that excites the sample, and some measuring equipment, that
detects the decay of the resulting emission. The set-up used in our lab is decribed below.
We were able to demonstrate the applicability of the method for multiple protein
pairs on a protein microarray. The first publication just became available
(Proteomics, Wiley -Interscience, early view).
Time correlated single photon counting (TCSPC) is a relatively easy mean to measure
fluorescence decay times. The method is scetched out in the drawing below. The output from a
mode -locked laser is used for excitation of the sample. This type of laser generates 100 fs pulses in the wavelength range 700 - 1100 nm. Due to the physical principle behind mode-locking, the pulse to pulse gap is just 12 ns, which corresponds to the cavity length of the laser. However this time intervall is too short, because the excitation from most
fluorescent molecules does not completely decay within this time. Therefore, the puls to puls distance has to be enlarged by an external puls-picker, which changes it to 250 ns or more, if required.
Since in our experiments, we want toexcite the intrinsic amino acids Tryptophan and Tyrosin, we have to excite the protein in the UV region of the spectrum, around 280 nm. To this end, the output of the Ti:Saphire laser (840 nm) is put through a frequency tripler. The frequency first generates frequency doubled light (420 nm) in a frequency doubling crystal, which is mixed in a second crystal with portions of the original
840 nm light to generate the tripled output at 280 nm.
Some of the light at 420 nm is not consumed by the light tripling process and is used to produce a trigger signal via a fast photodiode (PD).This trigger signal starts a counter at the very time when the excitation pulse arrives at the sample. (To adjust this time, a delay generator, sometimes only a shilded cable is used). A filter with an absorption edge around 315 nm or a monochromator cuts out scattered excitation light at time zero.
Upon excitation, the sample emits a fluorescence photon, which is delayed according to the excited state lifetime of the sample. This photon is detected by a photomultiplier (PMT), and serves as the stop pulse for the counting unit. By averaging over many events, one gets the decay signal. It is crucial to reduce the intensity so much, that you only get a stop event for every 100 to 1000 start pulses, otherwise the photon would statistically arrive too early and the measured decay curve would be artificially shortened. In our set-up, the counting unit is not a separate instrument, but rather a plugin card in the computer.
|
Alternatively, an avalanche diode is sometimes used for single -photon detection. However, the avalanche diode is not sensitive at below 400 nm, so we can only use it for experiments with longer wavelength fluorescent dyes.
As a second set-up, we also use a streakcamera (Hamamatsu C4334) to measure fluorescence lifetimes.
The function of streakcameras is nicely explained
here.
Note that there was a ESF exploratory workshop held in Jena on 1st to 3rd of April 2004:
'Protein Arrays -bridging the gap between physics and biomedicine'