Periodic Reporting for period 1 - EVOCATION (Unravelling the Evolution of Complexes with Ancestral Sequence Reconstruction)
Okres sprawozdawczy: 2022-05-01 do 2024-10-31
Most proteins associate into larger complexes that consist of several polypeptide chains that are held together by non-covalent interactions. Such complexes look intricate: Their interfaces comprise dozens of amino acids that make highly specific contacts. They are also easy to destroy by mutation. This seems to imply that natural selection is necessary to build their intricate interfaces and to subsequently protect them from a hail of destructive mutations. In this view, complexes exist because they provide some useful additional function relative to their individual components functioning on their own. Such a function could for example be allosteric regulation by small molecules that bind in the interface between subunits and then change the biochemical properties at a distal active site. Another advantage might be that the active site is only formed at the interaction site between subunits.
But there are many proteins that associate into complexes where no such advantages are apparent. Many enzymes associate into homomeric complexes consisting of multiple copies of the same protein, which are not allosterically regulated and in which each protein chain has its own active site. Is this a case of scientists simply not having found an advantage yet, or is there perhaps another explanation? Work leading up to this grant showed that the two foundational assumptions of why complexes must have functions to exist and persist are sometimes violated. First, they can be very easy to build. Some complexes apparently evolved in as little as a single historical substitution from a simpler ancestor. Many interfacial residues pre-existed in the necessary state long before the interface evolves, making it possible to make the jump in a single mutation that completes the interface. This is such a small number that interfaces may plausibly evolve through the action of random genetic drift. The second assumption, that protein complexes are so fragile that useless ones they would quickly be destroyed by mutations, is also likely incorrect. Interfaces can become entrenched, such that they become very difficult to lose even if they no longer fulfil any useful function. This happens when interfaces accumulate mutations that are harmless so long as the interface can form, but very harmful if the interface is prevented from forming. An example of this are hydrophobic substitutions: hydrophobic surfaces have to be shielded from water to prevent non-specific aggregation in the cell. Once an interface becomes hydrophobic, any mutation that prevents the interface from forming would expose its hydrophobic surface to water and lead to aggregation. Such mutations would be opposed by purifying selection, even if the interface offers no functional advantage over the individual components functioning on their own.
This grant seeks to explore how adaptive and non-adaptive evolutionary processes shape the kinds of complexes that are formed in a cell. To do this, we retrace the evolution of protein complexes using ancestral sequence reconstruction - a technique that allows us to recreate ancient proteins in the lab and investigate how their interactions developed historically. We are investigating the evolution of homomeric interactions in a central metabolic enzyme - citrate synthase - and test if changes in its self-assembly state are associated with new functional capabilities. We will investigate the evolution of a different type of protein-protein interaction: that of so-called chaperones with their client proteins. Chaperones are proteins that bind their clients transiently during folding and assembly to help them reach their active conformation. Some proteins have become completely dependent on their chaperones, such that they can no longer reach their active conformation without them. We are investigating how and why a particularly important protein became totally reliant on a whole suite of assembly chaperones in the course of its evolutionary history. Our model system is Rubisco, the enzyme that fixes CO2 in the Calvin cycle of cyanobacteria and plants. By re-tracing its history, we aim to learn when it became reliant on a set of assembly chaperones, and whether this reliance led to an improvement of its enzymatic performance. Lastly, we are investigating more generally how likely it is that a specific protein-protein interaction evolves entirely by accident.
But there are many proteins that associate into complexes where no such advantages are apparent. Many enzymes associate into homomeric complexes consisting of multiple copies of the same protein, which are not allosterically regulated and in which each protein chain has its own active site. Is this a case of scientists simply not having found an advantage yet, or is there perhaps another explanation? Work leading up to this grant showed that the two foundational assumptions of why complexes must have functions to exist and persist are sometimes violated. First, they can be very easy to build. Some complexes apparently evolved in as little as a single historical substitution from a simpler ancestor. Many interfacial residues pre-existed in the necessary state long before the interface evolves, making it possible to make the jump in a single mutation that completes the interface. This is such a small number that interfaces may plausibly evolve through the action of random genetic drift. The second assumption, that protein complexes are so fragile that useless ones they would quickly be destroyed by mutations, is also likely incorrect. Interfaces can become entrenched, such that they become very difficult to lose even if they no longer fulfil any useful function. This happens when interfaces accumulate mutations that are harmless so long as the interface can form, but very harmful if the interface is prevented from forming. An example of this are hydrophobic substitutions: hydrophobic surfaces have to be shielded from water to prevent non-specific aggregation in the cell. Once an interface becomes hydrophobic, any mutation that prevents the interface from forming would expose its hydrophobic surface to water and lead to aggregation. Such mutations would be opposed by purifying selection, even if the interface offers no functional advantage over the individual components functioning on their own.
This grant seeks to explore how adaptive and non-adaptive evolutionary processes shape the kinds of complexes that are formed in a cell. To do this, we retrace the evolution of protein complexes using ancestral sequence reconstruction - a technique that allows us to recreate ancient proteins in the lab and investigate how their interactions developed historically. We are investigating the evolution of homomeric interactions in a central metabolic enzyme - citrate synthase - and test if changes in its self-assembly state are associated with new functional capabilities. We will investigate the evolution of a different type of protein-protein interaction: that of so-called chaperones with their client proteins. Chaperones are proteins that bind their clients transiently during folding and assembly to help them reach their active conformation. Some proteins have become completely dependent on their chaperones, such that they can no longer reach their active conformation without them. We are investigating how and why a particularly important protein became totally reliant on a whole suite of assembly chaperones in the course of its evolutionary history. Our model system is Rubisco, the enzyme that fixes CO2 in the Calvin cycle of cyanobacteria and plants. By re-tracing its history, we aim to learn when it became reliant on a set of assembly chaperones, and whether this reliance led to an improvement of its enzymatic performance. Lastly, we are investigating more generally how likely it is that a specific protein-protein interaction evolves entirely by accident.
We have reconstructed the evolution of homomeric interactions along the citrate synthase phylogeny and found a whole menagerie of different kinds of protein complexes that come and go on phylogenetic timescales. One of our most dramatic discoveries is the evolution of a citrate synthase that can assemble into molecular representations of the Sierpinski triangle, a famous fractal pattern. Our citrate synthase is the first natural molecule to form this bizarre pattern and to do so it has to violate several symmetry rules that were thought to be near universal. We were additionally able to show that this bizarre assembly evolved in just a single historical substitution from a an ancestor that made a much smaller and very conventional assembly. To the best of our current knowledge, this assembly may represent exactly the kind of easy-to-gain and functionally inconsequential assembly that we thought might exist.
A second achievement is that we proved for the first time that two proteins learned to specifically interact with each other entirely by chance. We did this by re-tracing how two proteins that now engage in a functional interaction in cyanobacteria evolved the interfaces that make this interaction possible. We could show that the two proteins evolved matching interfaces in completely different organisms, long before a horizontal gene transfer brought them into contact in a cyanobacterium. This is a strong proof that proteins can become highly compatible with each other without natural selection directly acting on a function associated their interaction.
We have also retracted the evolution of Rubsico's chaperone interactions. We have worked out exactly when Rubsico required each of a series of dedicated chaperones that are now essential for its function in plants. Using this approach, we were able to work out exactly which mutations cause Rubisco to become totally reliant on a new chaperone and how this affects its function. We have been able to reverse engineer the evolution of Rubisco's dependence on chaperones and to our astonishment been able to create eukaryotic Rubiscos that require none of the elaborations that are normally necessary to fold and assemble them. These (unpublished) results have the potential to radically change our thinking about what chaperones really are.
A second achievement is that we proved for the first time that two proteins learned to specifically interact with each other entirely by chance. We did this by re-tracing how two proteins that now engage in a functional interaction in cyanobacteria evolved the interfaces that make this interaction possible. We could show that the two proteins evolved matching interfaces in completely different organisms, long before a horizontal gene transfer brought them into contact in a cyanobacterium. This is a strong proof that proteins can become highly compatible with each other without natural selection directly acting on a function associated their interaction.
We have also retracted the evolution of Rubsico's chaperone interactions. We have worked out exactly when Rubsico required each of a series of dedicated chaperones that are now essential for its function in plants. Using this approach, we were able to work out exactly which mutations cause Rubisco to become totally reliant on a new chaperone and how this affects its function. We have been able to reverse engineer the evolution of Rubisco's dependence on chaperones and to our astonishment been able to create eukaryotic Rubiscos that require none of the elaborations that are normally necessary to fold and assemble them. These (unpublished) results have the potential to radically change our thinking about what chaperones really are.
Our achievements to date go significantly beyond the state of the art: we have discovered completely novel kinds of protein complexes (our fractal), investigated the evolution of homomeric interactions on previously unimaginable phylogenetic scales, and performed the most complex ancestral resurrections ever attempted in our project on Rubisco. Our work is already beginning to challenge the idea that every newly identified complex has to have a biological function. It challenges how biochemists should think about establishing causal links between structure and function. It also points the way to novel protein engineering strategies that make use of our evolutionary discoveries.