Internal dynamics in the enzyme barnase

Enzymes are the catalysts that carry out all of the reactions in nature. We have known for over 100 years that the structure of an enzyme has to be matched closely to the structures of the molecules that are reacting ('substrates'), and X-ray and NMR structures have shown how this is achieved in detail for many enzymes (the lock and key hypothesis). However, our attempts to design new enzymes have so far been rather pathetic in comparison with the impressive catalytic ability of real enzymes. The best rationally designed enzymes are at least a million times slower than the real thing. Partly this is because the structure has to be very accurately correct. However, another reason, which we are only just beginning to come to grips with, is that an enzyme is not just a static framework, but it moves constantly, mainly as a result of continual bombardment by solvent molecules. This provides it with a lot of kinetic energy, and it appears that somehow this random thermal kinetic energy is channeled into a few very specific motions in order to help the enzyme perform its catalysis. One of the main ways in which this is achieved is that the 'normal' or resting state of an enzyme is an 'open' state, in which the active site (where the reaction occurs) is not in its optimum configuration. Motion within the enzyme very specifically closes the active site, and is precisely tuned so that only a few percent of enzyme molecules are in this active or 'closed' state at any one time. The substrates bind more tightly to the closed state than the open one, and therefore the presence of substrate pulls almost all of the enzyme molecules over into the more active closed state. This model is a refinement of the induced fit hypothesis, and is called conformational selection. It is not clear why enzymes need to do this. In some cases it is because the substrate cannot get into the closed state, but the more general reason may be that evolution does not want the enzyme to be active unless there are substrates bound, to avoid unwanted reactions.
This proposal aims to understand these motions for a model enzyme called barnase, which digests RNA. We have shown that in barnase there are two different motions required to make the closed state. One of these is a simple bending of the enzyme, like a hinge closing, and is a low-energy and common motion. The other requires several loops around the active site to close up together, rather like the fingers of a hand closing, and cannot occur efficiently unless the hinge is closed already. We have good evidence that this happens, but we need more details in order to understand it properly: we need to know rates, energies and structures, and how these motions are determined by the structure of barnase. Exactly what does it do and how does it do it? The first motion is easy to understand, but the second is not. Once we have understood it, we also want to explain it in ways that everyone can understand.
This is important, because until we understand how enzymes really work, we are to a large extent groping around in the dark, and we are unlikely to be able to build an enzyme that works well. Many scientists think that because we know the structural details of enzymes, we understand them already. This is sadly not true. Science has shown that real progress comes from a proper understanding of the problem, which is what this research aims to produce.

This is an exciting time for enzymology, as we are finally getting to grips with the role of enzyme dynamics in catalysis. One of the barriers has been the lack of any good experimental method for measuring dynamics in the timescale between nanoseconds and milliseconds. We have recently shown that two NMR methods can crack this problem: pressure dependence of chemical shifts, and the dependence of chemical shifts on temperature and denaturant, which fill in the gaps by providing information on timescales around microseconds and tens of nanoseconds respectively. We thus for the first time have the technology to address the full range of rates on a residue-specific level. The results so far are exciting, because they show that barnase has evolved two different motions. One is well populated (ca 10%) and is a straightforward hinge-bending motion, already expected from computer simulations. However, the other only about 2% populated and is a concerted closure of loops around the active site, which we have recently shown can only occur from a hinge-closed state. It also differs in that the closed state is in an energy well. The two motions are therefore coupled, providing insight into how dynamics actually closes the active site: the first motion channels the enzyme into a conformation from which the second motion becomes easier. The second motion only happens about one in 10,000 hinge bends.

This proposal aims to put some detail onto this, by characterising the energy landscape of barnase better, both free and bound to both small and large inhibitors. By mutagenesis we will show how crucial motion is, particularly the loop closure. An important element of the proposal is contained within the Impact Plan, in finding ways to bring these results to a wider audience, both non-specialist academics and the general public, using movies and virtual reality.

In general terms, the beneficiaries of this research include anyone who is interested in how proteins (and in particular enzymes) work, because this work characterizes protein function at a fundamental level. This includes a wide range of academics in the public and private sector, particularly those who work on proteins. The group with the most to gain are those developing novel enzymes, in both private and public labs, because this work will demonstrate clearly that merely knowing the structure is not enough: it is necessary to think about the dynamics as well. A reasonable analogy is to say that merely knowing the anatomy of a hand is not enough: it is necessary to know how the muscles work together to move the hand and fingers if you want for example to carry out surgery, or design bionic hands. The benefit comes from the specific detail on barnase to be generated by this project, but much more from the new mindset that goes with it: imagine trying to produce a bionic hand without knowing how hands move as part of their function.
In science, it takes time for new developments to trickle through the system, and the beneficiaries of this research certainly include undergraduate students of biochemistry and related subjects, via my book and its associated website, which will publicise this work widely. I and my publishers are hoping for wide adoption of the book in North America, and I hope that this will be a successful export of UK expertise.
As explained in detail in the Pathways to Impact statement, an outcome of this grant will be 3D Virtual Reality movies of protein motion. These will be made available to visitors to Sheffield (indeed I hope they will be forced upon them!). I anticipate that this way of teaching will become increasingly popular in the UK, for teaching in schools as well as Universities, and I will be keen to help develop and spread these methods. There is thus the potential for a widespread diffusion of the outcomes into schools and colleges: hopefully this will help to produce a better educated, more enthusiastic new generation, with more of a feeling for the wonder and beauty of science. There is clearly a value to the UK economy here. To ride one of my hobbyhorses for a moment, the study of protein dynamics involves some hard numerical work and an insight into maths: it would be great if an outcome of the work was a greater emphasis on numeracy in science teaching.