Bimolecular Fluorescence Complementation

Just read an article in the last issue of JCB, where the authors used a nifty new technique to investigate when and where certain RNA binding factors associate together.

What's neat, is that this technique, bimolecular fluorescence complementation (or BiFC) works by fusing each half of a fluorescent protein (in this case yellow-fluorescent protein; YFP) to two proteins of interest (in this paper the RNA export factor TAP/NXF1 and the RNA binding factor Y14). To regenerate the intact YFP molecule the two proteins in question must come together. Since the fluorescence can be monitored without fixatives you can mow monitor this potential interaction in living cells.

In the past researchers used a related technique called FRET, or fluorescence energy transfer, to demonstrate the interaction of proteins in live cells. It works by fusing one type of fluorophore (say fluorophore "A") to protein #1 and a second type of fluorophore (say fluorophore "B") to protein #2.

If the two proteins came close enough, when excited, fluorophore A can donate (or transfer) it's energy to fluorophore B.

Stimulate A, measure B.

The problem is that when you stimulate fluorophore A, a fraction of it's emitted light overlaps with the fluorophore B's emission spectrum. In addition when you excite fluorophore A, some of fluorophore B will get excited directly (i.e. not by energy transfer) and emits some light. These sources of non-FRET fluorescence contribute to background or noise, and limits your detection. As any good microscopist knows detection = signal/noise. So the only way to determine the amount of fluorescence caused by FRET, is to perform very precise measurement to get the values for the noise (background fluorescence) and total fluorescence (FRET + background) of the experiment. Often FRET changes the fluorescence by as little as 10% of the total fluorescence ... not a very high signal to noise ratio!

With BiFC, fluorescence will only occur once the two YFP halves come together ... thus reducing the backgroud and increasing the signal to noise ratio. The authors provide data that the two YFP halves don't interact on their own, so BiFC can only result if TAP and Y14 get close to each other. One nice result from these experiments was that although the TAP protein is found primarily at the nuclear rim, the BiFC fluorescence (and hence TAP-Y14 association) was seen in another part of the nucleus, in speckles. When the authors swapped which YFP half was attached to TAP and Y14, BiFC wasn't observed indicating that for BIFC to occur, the two halves of YFP have to be not only close to each other but also correctly aligned. In addition the authors only observed fluorescence when the RNA binding motifs of both TAP and Y14 werefunctional, suggesting that the two proteins associate together when bound to a transcript (i.e. mRNA).

So in summary here's another tool to monitor protein interactions in cells. We'll have to see whether this technique can be used for other protein combinations. OK that's enough flashy science for today ...

{Update 2/18/06}

Here is image 1 from the paper described above. They show BiFC fluorescence, staining against the C-terminal half of YFP (YC), DNA (DAPI) and a color merge. This as a good example of using black and white images to present data. If you wonder why the image looks strange, the authors have inverted the B&W picture (i.e. black=fluorescence, white=no fluorescence). This type of data presentation can often help to display very faint patterns. If you look in "B" Y14 is found in speckles while TAP/NXF is localized to the nuclear rim.

Ref: Ute Schmidt, Karsten Richter, Axel Bernhard Berger, and Peter Lichter, In vivo BiFC analysis of Y14 and NXF1 mRNA export complexes: preferential localization within and around SC35 domains JCB (06) 172:373-381