In our lab we study a peculiar type of protein interactions: those between molecular chaperones and their clients. Among other functions, the main role of chaperones is to assist folding of other proteins, thus preventing and reversing aggregation. Since the early 1970s, many different chaperones have been discovered and thoroughly studied. What makes our research so special is the technique we employ: optical tweezers. Here, individual proteins are attached via DNA handles to micrometer-sized beads which are trapped by laser beams. By moving the lasers, the proteins can be stretched, unfolded, relaxed, and so on (see animation below). Despite the potential of the technique, we soon realized that in order to study chaperone-protein interactions in more detail, we needed something more: direct simultaneous visualization of the proteins.
Animation illustrating the principle of protein folding experiments with optical tweezers
The combination of optical tweezers and fluorescence imaging has been around for some years. However, its application to study protein-protein interactions has not been possible so far, due to a seemingly simple technical limitation: it is very challenging to efficiently attach long and strong DNA handles to proteins. In fact, all previous coupling methods suffer from a three-way trade-off, as high-efficiency attachment is limited to either short or weak tethers, but cannot provide both simultaneously.
For the new type of experiments we had in mind, it was crucial to overcome this trade-off. DNA handles need to be very long (around 5000 base pairs) to enable imaging, as otherwise the autofluorescence emission from the close-by beads obscures any relevant signal from the protein complex. At the same time, proteins within complexes are further stabilized, so higher forces (above 40 pN) and consequently stronger tethers are required.
Could we develop a new coupling strategy that would meet all these needs? After many unsatisfactory attempts in the lab (and countless frustrating negative electrophoresis gels), we came up with an unexpected solution that did not involve complex, unexplored chemistries, but rather well-established biochemical techniques.
Scheme of the DNA-protein coupling method. First small oligos are attached to a protein (different coupling chemistries can be used). Next, a DNA handle with a ligation-compatible overhang is generated using nuclease digestion and partial re-synthesis. Finally, the handles are ligated to the oligo-labelled protein. For details refer to the article.
In order to attach long DNA handles with high efficiency, we first optimized an approach in which first small oligos are coupled to the protein and then longer DNA handles are hybridized. Since hybridization provides weak constructs, we needed a way to increase the strength: DNA ligation. We first generated long double-stranded DNA molecules. Next, we used a exonuclease to digest one of the strands. Finally, we employed a special polymerase that can leave behind an overhang. This hybrid DNA tether can then be covalently ligated to the oligos in the protein with high efficiency to provide long and strong constructs.
Illustration of the simultaneous sensing-imaging experiments, showing in perspective the two trapped beads (pink), the DNA handles (gray) attached to a fluorescent protein (green), with fluorescently labelled trigger factor (orange) bound to an unfolded region.
We used this new method to show for the first time the binding of the chaperone trigger factor to an unfolded polypeptide chain in real time (illustrated in the image above). Yet, this demonstration is a mere glimpse of the potential of our new method. In fact, we very recently used the technique to show how ClpB, a bacterial disaggregation chaperone, forcibly pulls polypeptide loops through its central pore to extract them from aggregates, in an article that has just been published in Nature. In the video below we show the simultaneous sensing-imaging experiments used in both papers (from 0:50), check it out!
For more information about the method, check our article here.