Interface matters: The stiffness route to stability of a thermophilic tetrameric malate dehydrogenase

Understanding the relation between protein flexibility, stability and function remains one of the most challenging, open questions in biophysical chemistry. For example, proteins need to be flexible to facilitate substrate binding but locally rigid to sustain substrate specificity. Enzymes from microorganisms that thrive at elevated temperatures, also referred to as thermophiles, are a natural study-case to dig into the issue. These proteins are stable and functional at a high temperature regime but generally lack activity at ambient conditions. Therefore, their thermal stability has been correlated to enhanced mechanical rigidity through the so-called corresponding states paradigm introduced years ago by Somero. The generality of this view, however, has been questioned by a number of experimental and computational studies. Computer simulations based on the molecular dynamics technique offer a unique opportunity to explore the correlation among mechanical rigidity and thermophilicity. In our recently published article in PLoS One we consider the specific case of two tetrameric orthologous malate dehydrogenase proteins from two bacteria that grow optimally at different temperatures. For these orthologues, as for other oligomeric proteins, the role of interfacial interactions becomes critical, adding up to the other cohesive forces acting on monomeric proteins. How the protein rigidity/flexibility patterns influence the stability and function of the two molecules is discussed in detail in the paper.

Dehydrogenases are proteins that require a cofactor in order to catalyze a reaction.
The picture above depicts one monomer of a lactate dehydrogenase in a
cofactor-bound (holoprotein) and a cofactor-unbound (apoprotein) form.

Stay wet, stay stable?

Proteins often host water molecules inside buried cavities or superficial clefts. The presence of these molecules was first resolved via x-ray crystallography and their exchanging dynamics with the external solution was deeply investigated by NMR experiments, the interested readers can dig all the work done by B. Halle and collaborators, see the Halle's web page. Molecular Dynamics simulations also shed light on the molecular mechanisms of this exchange, earlier work by Hummer and Garcia [1] and Sterpone, Ceccarelli and Marchi [2], date back to 2000 or so. It was always questioned the contribution of this set of molecules to the stability of the protein fold. Very recently we decided to tackle the problem and relate it to the issue of protein thermal stability [3]. The starting question was: is the extra stability of a thermophilic protein correlated to its internal hydration. For our study case, the pair of homologous G-domain from the mesophilic E. coli and the hyperthermophilic S. solfataricus, the answer is yes, at least a bit. The manuscript is here.

[1] G. Hummer and A.E. Garcia "Water Penetration and Escape in Proteins". Proteins 2000, 38, 261−272.
[2] F. Sterpone, M. Ceccarelli, M. Marchi, "Dynamics of Hydration in Hen Egg White Lysozyme. J. Mol. Biol. 2001, 311, 409−419.
[3] O. Rahaman, M. Kalimeri, S. Melchionna, J. Henin, F. Sterpone "Role of Internal Water on Protein Thermal Stability: The Case of Homologous G Domains ", J. Phys. Chem. B 2014 in press.

Volume fluctuation of reaction intermediates determines reactivity

Fig. 6. in Terazima et al, PNAS (2014) shows that a monoactivated protein decamer expands in volume and consequently dissociates, while a polyactivated decamer shrinks in volume and doesn't dissocitate.

 

Proteins are subjected to thermal fluctuations that occur at all length- and timescales, ranging from bond vibrations to large-scale domain rearrangements. The fluctuations affect the volume of the protein and vice versa. The close relationship between protein conformation and protein activity is a well established concept, while the role of fluctuations in protein activity is yet to be elucidated. A recent experimental study on a blue-light photosensor (TePixD) from the thermophile T. elongatus has shown that the flexibility of reaction intermediates is crucial in the overall reaction kinetics.

In solutions and crystals, TePixD proteins form pentameric rings, which further associate to decameric structures. An equilibrium between the pentamers and decamers is impaired by light illumination, upon which the decamers dissociate to pentamers. If the light-source is then removed, the pentamer-decamer equilibrium is restored within 5 s.

Using a pulse laser-based technique, two intermediate states have been identified in the decamer dissociation upon light illumination. These intermediates are named I1 and I2, and they exhibit different volume compressibilities with respect to the ground state. It has been shown that the change in compressibility, which is proportional to change in volume, is positive and about 10% of the compressibility of the protein, meaning that the intermediate species’ volume is larger as compared to the ground state. The volume change is related to different fluctuation levels. It is presumed that the fluctuations are unequally distributed and that the enhancement of the fluctuations at the interface regions of the decamers facilitates the dissociation.

The light illumination causes a conformational change in the immediate vicinity of the photoreactive center, but this center is not near the protein interfaces that are affected by dissociation and association. Evidently, the protein flexibility is changed by the conformational change around the photoactive center, which causes a greater protein flexibility, which in turn increases the volume of the intermediate species and drives protein dissociation.

This is also confirmed by the fact that high-intensity light, which activates more than one protein in the decamer, produces less pentamers because the volume of these species is decreased as well as the protein fluctuations. Thus, there exists an experimental proof that relates higher protein flexibility to enhanced reactivity.

Coarse-graining can track protein thermal stability?

Molecular simulation based on coarse-grained (CG) models is commonly used to explore large scale molecular motion or folding/unfolding processes. However, it is questionable wether these simplified models are good enough to capture the different thermal-stabilities of proteins.

Here several problems come upfront. Not all CG models are apt to monitor protein unfolding and stability. In some cases external biases are glued on top of the model in order to ensure proteins stability during a simulation; without these biases the structures just do not hold. As consequence the effect of physical/chemical perturbations to the folded structure cannot be appreciated. In other models, the folded state is encoded in native-biases -as in the Go-like model- so one wonder wether the detected unfolding paths that are key to distinguish the stabilities of homologues are "real". Finally, one should always keep in mind that a CG hamiltonian is made up of effective interactions, and, if a rigorous bottom-up graining approach is used, this implies that the model is temperature and concentration and pressure dependent. That is, the thermodynamic reference state is embedded in the interactions. This makes troubles when simulations are performed at different temperatures as in the case of the Replica Exchange Method. Despite the limitations, the possibility to use a CG model to explore the kinetic and thermodynamic stabilities of mesophilic and thermophilic proteins is rather appealing. We have tackled the problem recently, and nice results have been obtained. The paper is out here.  If you have not access, just ask!

Tracking the corresponding state principle

Maria KALIMERI just graduated from Univ Paris VII. Congratulations! Her thesis focused on mechanical properties of thermophilic proteins explored via an innovative framework based on network analysis. Here a bite of her work "Understanding the relation between protein flexibility, stability and function remains one of the most challenging, open questions in biophysical chemistry. For example, proteins need to be flexible to facilitate substrate binding but locally rigid to sustain substrate specificity. Exemplary cases are enzymes from microorganisms that thrive at elevated temperatures, also referred to as thermophiles. These proteins are stable and functional at the high temperature regime but generally lack activity at ambient conditions. Therefore, their thermal stability has been correlated to enhanced mechanical rigidity through the corresponding states paradigm. The generality of this view, however, has been questioned by a number of

experimental and computational studies. In the present study, we employ the gold standard of computational techniques, namely Molecular Dynamics simulations, in order to identify microscopical characteristics that distinguish thermophilic from mesophilic proteins, elaborating in particular on the rigidity paradigm mentioned above." For a more complete reading see here.

Yellowstone Thermophiles

Beautiful pictures. I found an interesting web-site dedicated to thermophilic life in the Yellowstone National Park (US), see here. I quote from the home page "There is so much to know and so much to find out about the extreamophiles in Yellowstone National Park. Since the discovery of Thermus aquaticus in 1969 a universe of interest has surrounded the life forms that survive - even thrive at very high temperatures. At the time that Thomas C. Brock and Hudson Freeze reported the new life form, it was thought that only a few organisms could survive at high temperatures above about 130F. Since its discovery, interest has spread and much has been learned. The scientific community has devoured the subject with relish. Only recently has this knowledge crept into the popular mind. As people learn of the extreme conditions of life at 175F they want to know more. It's fascinating and awe inspiring. It is a concept that is wondrous to contemplate. Questions are asked: "What do they eat?" - "How big are they?" - "Where are they?" - "How do they do it?" - "What do they look like?" The answers are as fascinating as the organisms. I would like to help answer the last question above. As I travel in Yellowstone I am drawn to their colors and patterns and diverse images that they etch on my eye. I take some snapshots and present a few of them here." Enjoy the Gallery!

Antifreeze protein with a heart of ice

The four helical bundle of the large antifreeze protein Maxi is glued in one, not primarily by protein-protein interactions but via an extended network of water pentagons! See picture below. The protein was recently crystalized by Peter Davies and his group and published in Science, 2014. 

This finding points to an important open question: What is the contribution of internal water to the stability gap between different homologous proteins (psychrophilic, mesophilic and thermophilic)?

A physical mechanism behind thermal stability

A little less than a decade ago a study suggested that there are two major physical mechanisms for protein thermal stabilization depending on the evolutionary history of the source organism, a ''structure-based" and a ''sequence-based" one [Berezovsky and Shakhnovich, 2005]. Proteins from organisms that originated in hot environments (therein archaea) have a much more compact structure and hydrophobic core. On the other hand, proteins from organisms that started as mesophiles and later recolonized a hotter environment (therein bacteria) remain structurally similar to mesophilic homologues but present some sequence substitutions that result in a few key interactions in the final fold. 

However, this strict assignment of evolutionary history to these two domains of life, archaea and bacteria, has no solid ground. In fact, more recent studies showed that the ancestors of bacteria were also thermophiles [Boussau et al., 2008; Akanuma et al. 2013]. Yet, the same study had also another dark point. The structure of the hyperthermophilic protein rubredoxin from archaeon Pyrococcus furiosus was found to be more tightly packed (number of contacts per residue) as compared to the rubredoxins from another 3 mesophilic bacteria. This result seems to contradict an earlier H/D experiment suggesting that the flexibility of this protein is typical to that of mesophiles [Hernandez et al., 2000] but the authors dedicate no comment to that. One certain aspect is that structural studies, although they put things in a first informative perspective, they neglect dynamics and are based on X-ray structures resolved at low temperatures. Notably, the effect of temperature on the structure and the magnitude of fluctuations of the exact same protein has been pointed out by short timescale MD simulations [Ergenekan et al., 2005]. Although the dynamics of rubredoxin from Pyrococcus furiosus was quite recently studied in detail using incoherent quasi-elastic neutron scattering in combination with MD [Borreguero et al., 2011], further comparative MD studies between mesophilic and thermophilic rubredoxins might shed light to contradictions such as the above.

Unfolding Under Shear

What does it happen to a protein in fluid shear flow? According to some studies, a protein could be forced to unfold. However there is not agreement on the necessary strength, or more technically, on the magnitude of the shear rate that could cause the unfolding to happen. Jaspe and Hagen for example estimated that only an extraordinary value of the shear rate (10⁷ s^-1) is effective for unfolding globular proteins, read their work in Biophys. J.(2006)  here.

In our group we have recently combined an effective coarse-grained model for simulating protein motion with an engine for considering hydrodynamic interactions. A first glance to this coupling is presented in a recent review in ChemSocRev(2014), see here. We also presented the preliminary results of an investigation aimed to understand how shear flow acts on the stability of proteins. We actually show the unfolding process of a simple β-hairpin peptide under laminar shear flow. We have used a very strong shear rate, 10¹⁰s^-1, and the unfolding occurs in about 10 ns. We are now checking how the unfolding rate changes by decreasing the shear rate.

Fig.9 in Sterpone et al, Chem.Soc.Rev. (2014) DOI:10.1039/C4CS00048J.  Unfolding of β-hairpin under shear. 

There cannot be only one

Is it hydrophobics or electrostatics? Is it in structure or in dynamics? Is it an enhanced rigidity or an increased flexibility of the protein matrix? Maybe the answer is in water?

Relevant scientists might not agree on what factor plays the most important role in increasing thermal stability of (hyper)thermophilic proteins, but they all agree that not only one is overall responsible. The enhanced thermal stability of a thermophilic protein is usually a result of a well-orchestrated symphony of more than one structural and/or dynamical factors. 

Even so, several experiments have demonstrated that, in some cases, single point mutations are capable of increasing the thermal stability of an enzyme. Whenever that is possible, it does come in handy, since a thermophilic enzyme with the desired properties doesn’t always exist or even if it exists it is not trivial to obtain. So we go back to studying how thermophilic proteins are mastering it. After all, it gets down to identifying trends that are immediately applicable for a rational design. 

Such a useful trend was recently presented by H. Gohlke and coworkers advocating for the importance of “qualitative” hydrophobic contacts on protein stability. By qualitative contacts the authors mean - and effectively demonstrate - that it is not the size of clusters of hydrophobic residues that distinguishes (hyper)thermophilic proteins from their mesophilic homologues. It is rather the fact that thermophilic, and even more hyperthermophilic proteins, are enriched in those hydrophobic contacts that have a low (favorable) energy. With this, they achieve in distinguishing thermophilic over mesophilic proteins with a discrimination accuracy of 80%, something that is not achieved as well when they use other energy components such as hydrogen bond energy for example. 

Finally and most importantly, the authors successfully locate weak spots on three different proteins where mutations will lead to an increased thermal stability, as well as non-weak spots that should not be mutated as they already stabilize the protein. Moreover, the computational efficiency with which this can be done makes the method a potentially very useful tool for protein design.

A droplet of water forms a spherical shape,
minimizing contact with the hydrophobic leaf.
Photo taken by tanakawho

Towards new thermostable proteins

ResearchMedia just published an highlight of the project THERMOS. The article "Towards new thermostable proteins" is in the new issue of the magazine International Innovation and can be read here (courtesy of RM). The article presents a nice overview of the project, a short description of what done so far and more importantly the lines of research we are following. Enjoy it! 

Solvation of halophilic proteins

My collaborators at the Institut of Biologie Structurale in Grenoble (FR), D. Madern and E. Girard along with the PhD student R. Tallon, just published an interesting work in Frontier (Microbiology/Extreme Microbiology), see the manuscript here. They report a detailed comparison between the structures of two homologous proteins: the halophilic tetrameric malate from the bacterium Salinibacter ruber (Sr) and the non-halophilic malate from the bacterium Chloroflexus aurantiacus (Ca). The core of the discussion concerns the role of hydration on extreme adaptation and relates to the different surface compositions of the two proteins, and the potential different coupling with the solvent layer.

First, the structures are resolved at very high-resolution. Second, the protein from Ca is resolved with a huge number of hydration molecules surrounding the protein surface and hydrating some internal locations. The presence of this well defined hydration layer around the non-halophilic protein allows to individuate precise closed structures of water, à voir pentagons, surrounding some hydrophobic patches of the surface. On the contrary the x-ray structure of the halophilic protein from the Sr bacterium lacks a well defined hydration layer and no clusters of water were visualized. Therefore the authors concluded that the chemical composition of the surface of the halophilic protein, enriched in negatively charged amino-acids, makes unfavorable for water to create extended closed networks of hydrogen bonds.

Then, and this is more speculative, the authors discuss how the enrichment in  negatively charged amino-acids could play a role for i) solubility in high-salt concentration and ii) salt-in effects observed in halophilic proteins.
Stay tuned on the blog because we are currently investigating how the life of a protein at ambient condition influences the stability of the water hydrogen bonds networks at the protein surface. For the moments, some hints from our past studies: i) role of protein surface on dynamics and structure of interfacial water (see here), ii) water networks at protein surface and protein stability (see here), iii) a molecular vision of protein hydration (see here), iv) proteins compositions and water dynamics (see here).

Halophile bacteria in Lake Natron, Tanzania