The moment we step into an unlit room from sun we are blinded for a couple of seconds. After some time our eyes adjust to the low light and then we can see in the dark room. This is because the light sensors in our eyes, the rhodopsin molecules, go on overdrive in presence of strong light. The importance of the adapter protein recoverin lies in the light-adaptation mechanism. In presence of excess light, when the cellular calcium concentration is high, recoverin binds the amino-terminal helix of rhodopsin kinase. The amino-terminal helix of rhodopsin kinase is required for reactivating rhodopsin so that it can sense light. This plugging action by recoverin on rhodopsin kinase prevents over-excitation of rhodopsin, allowing us to adjust to the low light level. How does recoverin bind rhodopsin kinase? The recoverin structure clearly changes upon binding rhodopsin kinase. Does the structure change due to binding, or, recoverin has a minor conformation which the rhodopsin kinase binds selectively? In order to answer this, the first step is to look at the structures.
In the morning of 10th January, 2016, the Protein Data Bank contains a total of 114741 structures of biomolecules. The majority of this high resolution structural information comes from x-ray crystallography. The other major technique that contributes high resolution structural information is nuclear magnetic resonance (NMR) spectroscopy. This huge number of protein structures point to the power of structural biology. However, some scientists think that the large number can give us a false sense of complacency. They argue that in order to function many proteins need more than one structure. Not only that, they argue further that the structures that are biologically active are invisible to the structural methods because of minor population.
The underlying idea is more than 50 years old; the classic studies on hemoglobin binding oxygen molecules yielded two competing mechanisms, the induced fit and the conformational selection. In the induced fit model, the protein binds a ligand to form an unstable pre-complex, which than goes on to the final stable structure. In case of recoverin this model will predict that the structure changes due to binding rhodopsin kinase. Schematically,
P + L PL P*L
In conformational selection model a minor population of the free protein already exists in the “bound-like” conformation. The ligand binds to this minor conformation selectively. As a result of binding the equilibrium is driven towards the bound conformation. In terms of recoverin and rhodopsin kinase system this model will predict that recoverin has a second conformation and rhodospin kinase binds to this minor conformation selectively. Schematically:
P P* + L P*L
The induced fit mechanism is the popular choice for explaining structural plasticity seen in proteins. But scientists working on kinetics noticed evidence of conformational selection or, in most cases noted that the kinetic data can be explained by both induced fit and conformational selection mechanisms.
This debate has been reopened in recent times with the evidence for previously undetected minor conformations by fluorescence spectroscopy, new technique of x-ray crystallography and NMR spectroscopy. Fluorescence spectroscopy can measure local information about the protein dynamics by measuring interaction between strategically positioned tags. The classical x-ray crystallography is done at cold temperature. Only recently there are techniques that allow x-ray crystallography to be done at room temperature, which allows visualization of the minor conformations.
However, the most powerful technique to study minor conformations is NMR spectroscopy. For example, a typical NMR experiment measures dynamics of the entire protein backbone. NMR spectroscopy manipulates the nuclear spins using radio-frequency pulses in presence of strong (upto 662,500 times earth’s magnetic field) magnetic field. Imagine a bunch of indefatigable runners in a track who run at different but constant speed. Before the race begins, all the runners are in the same position, or, they are “in-phase”, as an NMR spectroscopist would say. As soon as the race starts the runners will start to spread out, or they will start “de-phasing”. Next, after a certain time t, the runners are turned around by 180°. So after exactly time t all the runners will be back at the starting point together. In the NMR experiment the runners are nuclear spins and this phenomenon is called the “spin-echo”. The turning by 180° is performed by the application of a radio-frequency pulse. If we go back to the racing analogy again, if the time t is sufficiently long such that some of the runners have finished the race before being turned-around, then only the small number of runners still running will be back at the starting block. Now if t is decreased in steps, then with our shenanigan more and more runners will be collected at the staring block. And finally, when t is so small that even the fastest runner cannot complete the race before being turned around, we will get back all the runners at the starting block. This is the idea behind the relaxation dispersion experiment in NMR, where increasing frequencies (decreasing t) of 180° pulses are applied on nuclear spins. The resulting “dispersion” profiles can be fit to obtain thermodynamic, kinetic and structural information of the conformations at the same time! Recent developments in NMR methodology and hardware allow us to apply the 180° pulses at high enough frequency so that we can reliably measure functionally relevant dynamics.
So what is the implication if there is a minor conformation detected for some proteins? A group of scientists in Brandeis University in Waltham, MA were interested in looking at the implications for recoverin of having a minor conformation. I was a part of this group led by the Howard Hughes Medical Institute investigator and professor of Biochemistry department, Dorothee Kern. Our NMR results showed that 3% of recoverin exists in a minor conformation, which has lifetime in the order of milliseconds. The NMR results also showed that the minor conformation is very similar to the bound structure of the protein. Interestingly, the protein did not show exchange with the minor conformation when bound to the partner protein. This is a strong hint that the minor state is required for binding. However, the skeptics among us were not completely convinced, arguing that the formation of the minor state can be just an unnecessary excursion in the energy landscape; while the conformational exchange is a necessary condition for conformational selection, it is not a sufficient condition. So with the help of my colleagues I measured the actual rate of binding of recoverin with rhodopsin kinase using stopped-flow fluorescence spectroscopy. At this point we already knew the rate of formation of the minor conformation from NMR. Now if the mechanism underlying the binding is conformational selection, then the maximum rate of binding has to be equal to the rate of formation of the minor state. We found exact correlation of the rate of formation of the minor state measured using NMR with the rate of binding measured using stopped-flow fluorescence. We repeated the whole set of measurements at a different temperature to be doubly sure. Our results unambiguously show that the rate of formation of the minor conformation is the rate determining step of binding. While the results fit the conformational selection model beautifully, is it still possible that the data can be explained by induced fit model? Turns out, the answer is no in the case of recoverin. The thermodynamics is fundamentally different in the two models. In the conformational selection model, the actual binding step is much tighter than the overall binding, because the minor conformation is present in very small amount. In case of induced fit, however, the actual binding step is much weaker than the overall binding, because the exchange in the complex drives the equilibrium strongly towards the final stable complex. We determined the thermodynamics of binding using Isothermal calorimetry and the result rules out the possibility of induced fit binding. Additionally, in the induced fit model the expectation is that the rate of binding can be increased by increasing the ligand concentration. Also from this perspective, recoverin did not show any contribution of induced fit in binding rhodopsin kinase. Till date this is the clearest example of conformational selection mechanism. This works has been published in December in Cell Reports (http://www.cell.com/cell-reports/abstract/S2211-1247(15)01423-0).
Figure: Recoverin (Rv) exists in an equilibrium with a minor population (3%) existing as the active conformation Rv*. The ligand, the amino-terminal helix of rhodopsin kinase, selectively binds Rv* to form the final complex Rv*RK. The rate of formation of the Rv* from Rv (30 s-1) is the rate of complex formation in presence of excess rhodopsin kinase. Reprinted from Chakrabarti et al. (2015) Cell Reports, 14, 32-42.
http://dx.doi.org/10.1016/j.celrep.2015.12.010
Why would anybody need to care about the detailed mechanism of binding of a protein that is not involved in a disease? Can we apply this knowledge to develop a drug molecule? Not immediately. But the importance of the observation that some proteins exist in an inactive major conformation, while the minor conformation is active, cannot be overstated. Because then at least for some proteins, the structure that we know is of the inactive form and is a wrong target for drug development efforts. How common is this mechanism? Future research in more systems, when looking beyond the structure, will tell. For now, the message is that dynamics is very important for structural biology!
About the author: Kalyan Chakrabarti is a postdoc at the Department for NMR-based Structural Biology, Max Planck Institute for Biophysical Chemistry (MPIBPC), Gottingen, Germany. The work described here was carried out in Howard Hughes Medical Institute and Dept of Biochemistry, Brandeis University, Waltham, MA, USA.
About the image: The author in front of the 900 MHz NMR spectrometer in MPIBPC.
Reference: Chakrabarti KS, Agafonov, RV, Pontiggia F, Otten R, Higgins MK, Schertler GF, Oprian DD and Kern D (2016) Conformational selection in a protein-protein interaction revealed by dynamic pathway analysis. Cell Reports, 14, 32-42. doi: 10.1016/j.celrep.2015.12.010.
This work by ClubSciWri is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
