Project Details
Popular science description
Proteins are more than something to eat to gain muscles, they are necessary for the body to function at all. For proteins to have a function it was long believed that the protein needed a well-specified 3D-structure. This view has changed after so-called intrinsically disordered proteins (IDPs) were confirmed to have important biological functions. Estimates also show that they are not a fringe-class of proteins, but encompass almost 30 % of proteins in eukaryotic organisms.
Distinguishing IDPs from other proteins is the very fact that they lack a specific 3D- structure. Instead they exist as ensembles of heterogenous structures. However, this makes them harder to study experimentally, why computer simulations may contribute to the understanding of these proteins.
In many studies, proteins are investigated at low protein concentrations. But in biological contexts, the concentration of macromolecules (such as proteins, fats, carbohydrates, etc) is high. The high concentration may affect proteins to assume other structures than those observed at lower concentrations, thus, studies aiming to investigate biological function should consider this effect.
In this Thesis, a particular protein found in the saliva, Histatin 5, has been studied at high concentrations using scattering methods and computer simulations. Having high concen- tration is an impediment for simulations, as it is more computationally demanding. To get around this problem, so-called coarse-grained models has been utilized, which simplifies structures by putting several particles in groups, which then is treated as one big particle.
Several coarse-grained models were considered together with experimental data at high Histatin 5 protein concentration, and while no model was perfect, the best performing model indicated that concentration effects could be found up to medium-high concentra- tions. At very high concentration, experiments indicated that Histatin 5 to some extent lumped together to form aggregates, which simulations did not indicate. An hypothesis for the experimental behaviour was that Histatin 5 was too small protein for any dramatic effects to be observed, why a new protein was constructed by putting together two Histatin 5 proteins, attaching the C-terminal of the first Histatin 5 protein with the N-terminal of the second Histatin 5 protein. The same experiment was repeated, together with the best- performing model from before, and it was found that for the repeat-protein, the model was worse performing, even at lower protein concentrations. Some experiments pointed to properties in the repeat-protein that could not be modelled with the previous model, why a more powerful but more computationally expensive, non-coarse-grained (atomistic) model, able to model additional properties, was applied, but it was found that even this model had difficulty explaining experimental data. This shows the need to develop models further to work better with this important class of proteins. Experimental behaviour of the repeat-protein was as with Histatin 5, with the exception that the concentration where aggregation would start to be observed was lower than for Histatin 5.
In biological contexts, proteins are not surrounded by copies of themselves, but exist in a heterogenous environment. One should therefore also study proteins together with other sorts of molecules, that contribute to a crowded environment. Such a molecule, that are intended to make an environment crowded at increasing concentration, is aptly called a crowder. Here, Histatin 5 was studied together, each by themselves, together with four other molecules, in various size. For three of these, no effect on Histatin 5 was found at all, while the fourth, largest crowder had, according to experiments where structure was probed, no effect, while other experiments probing diffusion showed a difference when the concentration of crowder was fairly large. Simulations were able to confirm the lack of effect by the crowders concerning structural properties.
How diffusion was affected by high concentrations of Histatin 5 alone was also studied. It was found that diffusion slowed down, which possibly is explained by Histatin 5 lumping together at high concentrations. The effect of temperature and salt at high protein con- centration was also investigated, where the temperature effect was found to be trivial (high temperature yielding faster diffusion), while increasing salt content decreased the diffusion rate. Speculatively, the salt effect was explained by Histatin 5 having a slightly different structure at low salt content, or that salt induces clustering of Histatin 5.
The examination of diffusion as a function of protein concentration also included simu- lations at the atomistic level. To compare experiments with the simulations, assumptions was needed. These were primarily chosen based on previous studies of another protein. With such assumptions, the simulation found too slow diffusion rates as compared with experiment. Other assumptions, based on approximations about the geometry of Histatin 5 would yield more favourable values compared with experiment, but studies on real pro- teins should be deemed as more realistic. The trends of diffusion as a function of protein concentration were similar between experiments and simulation, why the simulation model should be regarded as semi-quantitative.
Finally, a comparison between three different coarse grained models was performed, where it was found that the most advanced model did not perform as well as the simpler models in terms of predicting the overall dimensions of a set of intrinsically disordered proteins, indicating that simpler models are competitive.
Distinguishing IDPs from other proteins is the very fact that they lack a specific 3D- structure. Instead they exist as ensembles of heterogenous structures. However, this makes them harder to study experimentally, why computer simulations may contribute to the understanding of these proteins.
In many studies, proteins are investigated at low protein concentrations. But in biological contexts, the concentration of macromolecules (such as proteins, fats, carbohydrates, etc) is high. The high concentration may affect proteins to assume other structures than those observed at lower concentrations, thus, studies aiming to investigate biological function should consider this effect.
In this Thesis, a particular protein found in the saliva, Histatin 5, has been studied at high concentrations using scattering methods and computer simulations. Having high concen- tration is an impediment for simulations, as it is more computationally demanding. To get around this problem, so-called coarse-grained models has been utilized, which simplifies structures by putting several particles in groups, which then is treated as one big particle.
Several coarse-grained models were considered together with experimental data at high Histatin 5 protein concentration, and while no model was perfect, the best performing model indicated that concentration effects could be found up to medium-high concentra- tions. At very high concentration, experiments indicated that Histatin 5 to some extent lumped together to form aggregates, which simulations did not indicate. An hypothesis for the experimental behaviour was that Histatin 5 was too small protein for any dramatic effects to be observed, why a new protein was constructed by putting together two Histatin 5 proteins, attaching the C-terminal of the first Histatin 5 protein with the N-terminal of the second Histatin 5 protein. The same experiment was repeated, together with the best- performing model from before, and it was found that for the repeat-protein, the model was worse performing, even at lower protein concentrations. Some experiments pointed to properties in the repeat-protein that could not be modelled with the previous model, why a more powerful but more computationally expensive, non-coarse-grained (atomistic) model, able to model additional properties, was applied, but it was found that even this model had difficulty explaining experimental data. This shows the need to develop models further to work better with this important class of proteins. Experimental behaviour of the repeat-protein was as with Histatin 5, with the exception that the concentration where aggregation would start to be observed was lower than for Histatin 5.
In biological contexts, proteins are not surrounded by copies of themselves, but exist in a heterogenous environment. One should therefore also study proteins together with other sorts of molecules, that contribute to a crowded environment. Such a molecule, that are intended to make an environment crowded at increasing concentration, is aptly called a crowder. Here, Histatin 5 was studied together, each by themselves, together with four other molecules, in various size. For three of these, no effect on Histatin 5 was found at all, while the fourth, largest crowder had, according to experiments where structure was probed, no effect, while other experiments probing diffusion showed a difference when the concentration of crowder was fairly large. Simulations were able to confirm the lack of effect by the crowders concerning structural properties.
How diffusion was affected by high concentrations of Histatin 5 alone was also studied. It was found that diffusion slowed down, which possibly is explained by Histatin 5 lumping together at high concentrations. The effect of temperature and salt at high protein con- centration was also investigated, where the temperature effect was found to be trivial (high temperature yielding faster diffusion), while increasing salt content decreased the diffusion rate. Speculatively, the salt effect was explained by Histatin 5 having a slightly different structure at low salt content, or that salt induces clustering of Histatin 5.
The examination of diffusion as a function of protein concentration also included simu- lations at the atomistic level. To compare experiments with the simulations, assumptions was needed. These were primarily chosen based on previous studies of another protein. With such assumptions, the simulation found too slow diffusion rates as compared with experiment. Other assumptions, based on approximations about the geometry of Histatin 5 would yield more favourable values compared with experiment, but studies on real pro- teins should be deemed as more realistic. The trends of diffusion as a function of protein concentration were similar between experiments and simulation, why the simulation model should be regarded as semi-quantitative.
Finally, a comparison between three different coarse grained models was performed, where it was found that the most advanced model did not perform as well as the simpler models in terms of predicting the overall dimensions of a set of intrinsically disordered proteins, indicating that simpler models are competitive.
| Status | Not started |
|---|