Thursday, January 23, 2014

Does bulk water get crowded out of cells?

Although terahertz spectroscopy has become an important tool for studying biomolecular hydration, its interpretation is not straightforward. To resolve some of the ambiguities, Robert Donnan and colleagues at Queen Mary College in London have used MD to compute the vibrational density of states for several hydrated proteins of varying size, looking in particular at the distance and timescales probed by THz (O. Suchko et al., J. Phys. Chem. B 117, 16486; 2013 – paper here). They find that for all the cases studied – lysozyme, BPTI, TRP tail and TRP-cage – the hydration layer is 10 Å thick, and displays similar dynamics. Differences in the solvation dynamics for these systems seemed to stem primarily from highly retarded water molecules in the proteins’ interiors.

What dominates the solvation free energies of peptides? One view is that it is the free energy needed to create a cavity in the solvent. But this may be offset to at least some degree by the electrostatic interactions of polar groups with the water, and/or by the van der Waals interactions. To examine this balance, Montgomery Pettitt at the University of Texas at Galveston and colleagues have performed MD calculations for flexible alanine oligomers (H. Kokubo et al., J. Phys. Chem. B 117, 16428; 2013 – paper here). They find that, for rigid peptides the free-energy gains from vdW interactions more than compensate for the cost of cavity formation as the oligomers get longer. But when the solutes are flexible and allowed to collapse, this situation reverses – implying that van der Waals interactions provide a significant driving force for the collapse. It seems not yet clear, however, what role intramolecular interactions play in the collapse, since the fluctuations in that component of the free energy are large.

The role of water in the complexation of DNA-binding agents is examined for the case of the minor-groove binder netropsin by Edwin Lewis of Mississippi State University and colleagues (J. P. Ramos et al., J. Phys. Chem. B 117, 15958; 2013 – paper here). Specifically, they use calorimetry to look at how binding is affected by osmolytes (TEG and betaine), which introduce an osmotic pressure on the hydration water. The results support the earlier idea that there are two distinct binding modes of netropsin, and allow quantification of the water molecules that seemingly hydrate the bound molecule: 31 and 19 molecules for the two cases. Moreover, in the latter case at least one water molecule seems to remain trapped at the binding site, mediating the interaction with netropsin. The addition of the osmolytes, which exert broadly similar effects, has much the same effect on complexation as a reduction in the temperature.

At high concentrations, protein solutions have been found to exhibit liquid-liquid phase separation into solutions of different protein concentration. Such highly concentrated solutions are relevant to some medical conditions, such as sickle-cell anaemia and Alzheimer’s, and also to technological processes such as protein purification and storage. Johannes Möller of the TU Dortmund and colleagues relate this phase behaviour to that in protein solutions at high pressure (J. Möller et al., Phys. Rev. Lett. 112, 028101; 2014 – paper here). Using SAXS from lysozyme solutions, they find that the liquid-liquid phase separation is in fact re-entrant at high pressure. They attribute this behaviour to the effect of pressure on solvent-mediated protein-protein interactions, and conclude that pressure might be used as a means of controlling protein aggregation and crystallization.

Another possible influence on protein dynamics and function at high concentrations is crowding. Kevin Kubarych and his collaborators at the University of Michigan have been studying this matter for some time, and their latest paper (J. T. King et al., JACS 138, 188; 2014 – paper here) reports the interesting observation of a dynamical transition above a certain crowding threshold. Above this limit for lysozyme (induced by the polymeric crowding agent PEG-400, or by protein self-crowding), ultrafast 2D-IR spectroscopy reveals a significant slowdown in hydration dynamics on picosecond timescales. The authors suggest that this is a kind of jamming transition between hydration shells of the protein molecules extending out to 15-20 Å, i.e. to separations of 3-4 nm, which are certainly of the order of those found between macromolecules in cells. In other words, the transition reflects a collective frustration of rearrangements of the overlapping hydration shells. According to this picture, one would anticipate most regions of a cell to be in the “over-crowded” regime, with little “bulk-like” water. The same abrupt dynamical transition was not seen, however, in the authors’ previous studies using the small-molecule crowding agent glycerol, for which the dynamical slowdown was more gradual.

Models of the air-water interface – including models of hydrophobic hydration that invoke an interface of that nature adjacent to the hydrophobic surface – can’t in general easily accommodate a good description of long-wavelength density fluctuations, according to an analysis by Suriyanarayanan Vaikuntanathan and Phillip Geissler at the Lawrence Berkeley National Laboratory (Phys. Rev. Lett. 112, 020603; 2014 – paper here). They show that discretizing the interface, as in a lattice model, effectively suppresses long-range fluctuations, even to the extent of suppressing a roughening transition, above some critical value of the interaction potential. The authors show how one can accommodate the resulting nonlinearities, which for example allows them to describe the shape dependence of interfacial free energies.

Thomas DeCoursey of Rush University and Jonathan Hosler of the University of Mississippi Medical Center offer an intriguing discussion of the current understanding of (as the authors somewhat provocatively put it, the “philosophy of”) voltage-gated proton channels (J. R. Soc. Interface 11, 20130799; 2014 – paper here). The paper includes an overview of such issues as proton hopping along water wires, and mechanisms for proton selectivity (for example, exclusion of alkali metal ions) and for the suppression of proton transport in aquaporins.

Christopher Fennell at Oklahoma State University and colleagues have previously developed a computationally inexpensive water model called the semi-explicit assembly (SEA) model, which does a good job of calculating the solvation free energies of polar and nonpolar solvents (Fennell et al., PNAS 108, 3234; 2011). They now extend the SEA model to a version they call field-SEA, which can handle ions and charged solutes with no additional computational overhead (L. Li et al., J. Phys. Chem. B jp4115139; 2014 – paper here).