Friday, October 22, 2010

Pores for thought

Two papers in Science look at the mechanics of proton conduction in the M2 pore of influenza viruses (F. Hu et al., Science 330, 505-508; M. Sharma et al., Science 330, 509-512; 2010 – papers here and here). Hu et al. say that the key is whether a complex of histidine groups (His37) in the tetrameric pore is or is not in contact with a water chain threading the pore, which changes in a pH-dependent manner. Previously it has been controversial whether the His37 group participates in an active fashion by protonation/deprotonation or merely alters the channel diameter electrostatically. Hu et al. say that the latter does happen, but is accompanied by increased rotational freedom of the His imidazole group to make contact with the ‘water wire’ and relay a hopping proton. Sharma et al. could be said to refine this picture by considering the interactions between all the His37s in the tetramer and their relation to the adjacent Trp41 residues. They say that the proton-conducting state is activated at low pH by ‘unlocking’ of the imidazole group in a way that is gated by Trp41 via pi-cation interactions.

More pores. Not all aquaporins are selective for water only – some will admit glycerol and other alcohols. David Savage at UCSF and colleagues have determined their X-ray crystal structures to figure out what determines the selectivity (D. Savage et al., PNAS 10.1073/pnas.1009864107 – paper here). The details are complex, but the basic features arguably those one might expect: in the region of the pore’s selectivity filter, the energetics of water transport are controlled by channel hydrophilicity, while selectivity for larger molecules such as glycerol is steric, governed by channel width.

Some other protein pores will also allow water and other small molecules to pass, such as the sodium-glucose cotransporter, which permits water flow in the presence of Na+ or glucose. This class of proteins allows passive water flow in the presence of an osmotic Na+ or glucose gradient, but also flow against such a gradient if Na+ or glucose is present. To what extent, then, is such flow passive, and to what extent actively pumped in the presence of these solutes? Michael Grabe at the University of Pittsburgh and colleagues investigate that question using MD simulations (S. Choe et al., Biophys. J. 99, L56-L58; 2010 – paper here). They find that water will pass through the sodium-dependent galactose cotransporter vSGLT both in the absence and the presence of galactose. In the former case the flow is passive (and occurs through the galactose-binding region), but that in the latter case the release of galactose ‘pushes’ water molecules through the pore in the same direction as its exit – as the authors put it, galactose acts as a Brownian piston that rectifies the passive water motion through the pore.

Comparable rectification of proton motion through a proton pumps such as bacteriorhodopin, resulting in a ‘proton diode’, is described by Klaus Gerwert and colleagues at Bochum (S. Wolf et al., Angew. Chem. Int. Ed. 49, 6889-6893; 2010 – paper here). This team draw on their detailed investigations of bR over recent years, and new studies of point mutations, to explain how conformational changes in a few residues can control access between internally bound and external water molecules in such a way as to bias the direction of proton flow in a general mechanism that they suspect might be more general to other proton pumps.

Biological processes involving proton motion are prime candidates for manifesting quantum-mechanical effects. A rather striking instance of this is claimed by George Reiter of the University of Houston and colleagues, who say that the zero-point motions of protons in are entirely responsible for the binding of water to A-DNA (G. Reiter et al., Phys. Rev. Lett. 105, 148101; 2010 – paper here). Changes in hydration seem to be what drive the A-to-B transition in DNA (the A phase forms in dehydrated conditions), and Reiter et al. say that this is accompanied by a change in the zero-point kinetic energy of the protons in the hydrated B-DNA that is sufficient in itself to motivate the transition. Whether the protons concerned are those of water molecules in the hydration shell or those in the DNA’s H-bonds (or, presumably, a bit of both) is not yet clear.

It’s many years ago now that David Tirrell, now at Caltech, developed ways to incorporate fluorinated amino acids into proteins using recombinant DNA technology – a feat popularized with reference to non-stick fried eggs. David has now teamed up with Ahmed Zewail and colleagues to investigate what fluorinated residues do to the hydration of such proteins, using ultrafast time-dependent fluorescent Stokes shift spectroscopy on fluorinated coiled-coil peptides (O.-H. Kwon et al., PNAS 10.1073/pnas.1011569107 – paper here). They find that the hydration dynamics are retarded around fluorinated residues, in marked contrast to an acceleration of the dynamics for the corresponding cases of hydrogenated residues, which are of comparable size.

Some mini-proteins, such as the villin headpiece domains HP35 and HP36, with just 35 and 36 residues, have an unusual ability to fold quickly into a conformation with a securely buried hydrophobic core despite their small size. Takao Yoda of the Nagahama Institute of Bio-science and Technology and colleagues have used simulations to investigate how this happens (T. Yoda et al., Biophys. J. 99, 1637-1644; 2010 – paper here). The cores in the folded state are fully dehydrated, but some water molecules remain therein until the final stages of folding. This is an unusually large system for which the folding process has been followed in detail from a fully extended state in explicit solvent. Meanwhile, David Cerutti at Rutgers and coworkers have simulated the crystal structures of a scorpion toxin protein to probe the performance of different protein and water force-field models in predicting the observed structure (D. S. Cerutti et al., J. Phys. Chem. B 114, 12811-12824; 2010 – paper here). They find that the results are not very sensitive to the water model, and far more dependent on the protein model (FF99SB does best) in terms of correctly predicting the right contacts in the peptide chain.

Whether we can understand the various forces involved in protein folding sufficiently to design peptides and other heteropolymers to adopt specific conformations is of course one of the big challenges not only for protein design but for polymer chemistry more generally. Shekhar Garde and colleagues at RPI have studied how water-mediated interactions might be exploited in this endeavour (S. N. Jamadagni et al., J. Phys. Chem. B jp104924g – paper here). In particular, they explore the sequence space of model polymers containing one or two pairs of charged monomers, in cases where the other monomers are more or less hydrophobic, and at the effect of adding salt to such systems. The results reveal subtle factors at play: for example, ion pairs among hydrophobic monomers can stabilize hairpin conformations over collapsed globules via water-mediated Coulombic interactions, but this depends on where the charged monomers appear along the chain. And adding salt can open up these hairpins, whereas salt stabilizes the globular forms of hydrophobic homopolymers.

Some membrane-binding proteins induce curvature of the membrane (see H. T. McMahon & J. L. Gallop, Nature 438, 590; 2005) – for example, the so-called BAR domain of amphiphysin will remodel lipid vesicles into tubules. Greg Voth and colleagues have performed MD simulations to try to figure out how this works, and they find that, surprisingly, there is a layer of water intervening between BAR and the membrane even when the protein domain is strongly bound (E. Lyman et al., Biophys. J. 99, 1783-1790; 2010 – paper here). This implies that the charged region of BAR is screened from the lipid headgroups, so that the bending mechanism is not electrostatic.

How homogeneous are solutions of denaturing guanidinium salts? Recently, Mason and coworkers have reported evidence that the cations form nanoscale aggregates (e.g. P. E. Mason et al., PNAS 100, 4557; 2003). Using dielectric relaxation spectroscopy, Richard Buchner of the University of Regensburg question this conclusion (J. Hunger et al., J. Phys. Chem. B jp101520h – paper here). In other words, perhaps after all Gdm+ ions can ‘fit’ comfortably into the structure of bulk water without altering it – which would seem to support the ‘direct’ mechanism of denaturation whereby Gdm+ interacts with the protein backbone rather than loosening up the folded state by influencing hydration.

More evidence for the direct intermediation of water in enzyme action: Sason Shaik at the Hebrew University of Jerusalem and colleagues say that a water cluster in the binding site of heme oxygenase participates in its ring-opening degradation of heme groups (W. Lai et al., JACS 132, 12960-12970; 2010 – paper here). The water cluster organizes the substrate into the proper geometry, serves as a proton shuttle, and stabilizes the hydroxyl nucleophile that attacks the ring. And Dongping Zhong at Ohio State University and colleagues say that flavoproteins, which act as redox coenzymes, contain water networks in the active site with fast relaxation dynamics that probably controls the protein flexibility in a functionally relevant manner (C.-W. Chang et al., JACS 132,12741-12747; 2010 – paper here). Meanwhile, Peter Brzezinski at Stockholm University and colleagues offer evidence that water molecules are involved in proton transport through cytochrome c oxidase, which actively pumps protons across a membrane to sustain a proton-motive force for ATP synthesis (H. J. Lee et al., JACS ja107244g – paper here).

Mischa Bonn of the FOM Institute AMOLF in Amsterdam and colleagues say that the hydration region of lipid headgroups in monolayers contains two populations of water molecules, one with bulk-like relaxation and the other that relaxes faster (M. Bonn et al., JACS ja106194u – paper here). Their VSF spectroscopic measurements thus imply the presence of a group of water molecules that are strongly hydrogen-bonded to the headgroups, decoupled from the bulk, and potentially involved in rapid, in-plane proton transfer, as previous studies have suggested for lipid membranes.

How does the hydration of antifreeze proteins differ from that of regular proteins? Ann McDermott and colleagues at Columbia use NMR methods to study this question by investigating the hydration shells of ubiquitin and an ice-binding (type III) antifreeze protein at cryogenic temperatures (-35 C) (A. B. Siemer et al., PNAS 10.1073/pnas.1009369107 – paper here). They find that the ubiquitin hydration shell remains unfrozen and uncoupled to the ice lattice, whereas this is true of only parts of the AFP III shell: as one might expect, the ice-binding interface establishes direct contact with ice.

The ability of sufficiently narrow carbon nanotubes to admit water but exclude ions has been proposed as a basis for desalination technologies. But Daejoong Kim and coworkers at Sogang University in South Korea show that admission or exclusion of ions can be a subtle business. They say that at certain nanotube diameters, strong sodium hydration can lead to a preferential admission of potassium ions over sodium in mixed ionic solutions, while at other diameters sodium can be preferred, or that both ions can be increasingly excluded even as the tube diameter increases (J. J. Cannon et al., J. Phys. Chem. B jp104609d – paper here). So it is conceivable that selective filtration and separation might be achieved. It seems quite conceivable that Ilan Benjamin’s new analysis of the hydration of alkali-metal halides in hydrophobic environments, and the formation of ion pairs (J. Phys. Chem. B jp1050673 – paper here), might be relevant here to the state inside the nanotubes.

Finally, bulk water, and a suggestion that the structural picture here is still not fully resolved. Peter Hamm at the University of Freiburg and colleagues use the theoretical tools developed for the study of complex networks (such as those in social science) to reveal hidden topological aspects of the H-bonded network in MD simulations of liquid water (F. Rao et al., J. Phys. Chem. B jp1060792 – paper here). They say this approach reveals structural inhomogeneities, extending at least to the second hydration shell, that are not evident from methods that focus on a single scalar order parameter such as tetrahedrality. Needless to say, the same technique might be used to look at solute hydration shells.

Friday, September 17, 2010

Water in cavities

How hydration affects ligand binding in protein cavities is a subtle business. Not only is it still imperfectly understood, but it seems possible that it might be hard to generalize about how the various enthalpic and entropic effects of dehydration of the cavity and the ligand balance out. Yet such issues could be central to the rational design of drugs. Andrew McCammon at UCSD and his coworkers have used MD simulations to try to bring some order to the problem (R. Baron et al., JACS 132, 12091-12097; 2010 – paper here). They point out that both positive and negative entropy changes have been reported previously for water entering protein cavities, and that the result probably depends both on the chemistry and the geometry of the cavity. The key question, they rightly say, is: does water have a passive or active role in cavity-ligand recogniton? They investigate that question using an idealized cavity-ligand combination with various permutations of surface charges on both, looking to develop ‘thermodynamic profiles’ of the binding events. And indeed, the signs and magnitudes of the enthalpy and entropy changes prove to vary widely for the different cases (their Figure 6 tells the story). One finding is that, coincidentally, the net free energy change is similar both for binding driven electrostatically and by hydrophobic interactions.

Similar issues are explored by Hongtao Yu and Steven Rick at the University of New Orleans, who calculate the entropy, enthalpy and free-energy changes on transferring a water molecule from the bulk to various types of protein cavity large enough to hold only a single water (J. Phys. Chem. B 10.1021/jp104209w – paper here). They look in particular at the effects of having different numbers of H-bond donors and acceptors in the cavity. It is again not easy to generalize, but it seems that the thermodynamic consequences of H-bond formation are greater than those exerted, via entropic effects, by the cavity size.

More on how denaturants work, this time from Ilja K. Voets at the University of Fribourg and colleagues (J. Phys. Chem. B 10.1021/jp103515b – paper here). They use light scattering, SANS, NMR, IR and UV-Vis spectroscopy to study how the size and flexibility of the lysozyme molecule changes in a mixed water/DMSO solvent at varying compositional ratios. They see three regimes, in which the protein is compact, wholly unfolded, and partially unfolded, and deduce that there are three major factors influencing these conformational changes: changes in water’s H-bonded structure induced by DMSO, the propensity of DMSO to act as an H-bond acceptor, and DMSO’s role as a poor solvent for polar groups and the polypeptide backbone but a good one for apolar sidechains.

Meanwhile, Yi Qin Gao at Texas A&M and colleagues look at urea denaturation (H. Wei et al., J. Phys. Chem. B 10.1021/jp103770y – paper here). Here the big debate has been whether urea acts indirectly, via its effect on water’s H-bonded network, or directly by interactions with the protein backbone. This group proposed previously that it’s a bit of both. Now they take that idea further by looking at urea denaturation of the chicken villin headpiece protein and some of its mutants. They consider in particular how urea affects the breaking of backbone hydrogen bonds and the penetration of water into the hydrophobic core. In both respects, the influence of urea seems again to be both direct and indirect. For example, the presence of urea seems to enhance water penetration of the core (just as it has been found previously to enhance the hydration of the interiors of carbon nanotubes), and that this enhancement is correlated with binding of urea to the protein surface. Moreover, as the protein unfolds, the exposed hydrophilic regions are stabilized by binding of urea. It seems, then, that the denaturing action cannot be explained by a ‘physical effect’ so much as by a ‘narrative’ of the dynamical process.

The mechanism(s) of anti-freeze glycoproteins also continue to be debated, and Martina Havenith at Bochum and her colleagues raise the intriguing idea that they owe their protective role to an ability to perturb the collective dynamics of water over long ranges, retarding them in the extended hydration shell in a way that suppresses freezing (S. Ebbinghaus et al., JACS 10.1021/ja1051632 – paper here). This proposal, based on terahertz spectroscopic data for an anti-freeze glycoprotein from an Antarctic fish, is quite different from any other that I know of, and could be quite different from the way anti-freeze proteins work – unlike AFP, AFGPs are flexible and don’t have a well-defined structure, so probably do not operate by binding to the surface of ice crystals.

One of my finer moments in my former life at Nature was getting Reza Ghadiri’s work on cyclic peptide nanotubes published in 1993 in the face of scepticism during the review process. The work was soon vindicated, and these self-assembling tubes showed evidence of being able to mediate transport of dissolved species through lipid membranes. Jianfen Fan at Soochow University in Suzhou and colleagues have now used MD to investigate the mechanism of water diffusion through such pores (J. Liu et al., J. Phys. Chem. B 10.1021/jp1039207 – paper here). Water molecules form a linear chain threading a hexapeptide channel, but the H-bonded structure is increasingly three-dimensional for wider cyclic molecules. Understanding the transport process could be valuable for, say, the potential use of self-assembling channels like this in water purification and desalination.

The dynamics and evolution of such water ‘filaments’ in pores and other biological systems, such as lipid membranes, are studied by Marek Orzechowski and Markus Meuwly at the University of Basel, using MD simulations (J. Phys. Chem. B 10.1021/jp1051003 – paper here). Their medium is, however, not explicitly a biological one: it is a monolayer of alkylsilica chains about 1 nm thick, like those used in some chromatographic columns; and the solvent is a water/acetonitrile mixture. They find that water filaments, containing typically tens of molecules, form intermittently in the alkyl layer, and can persist for around 1 ns before being dispersed by thermal fluctuations.

The transition of lipid membranes from a gel to a liquid-crystal phase may significantly alter the membrane’s interactions with water and aqueous ions, say Tomasz Róg at the Tampere University of Technology in Finland and coworkers (M. Stepniewski et al., J. Phys. Chem. B 10.1021/jp104739a – paper here). Both phases might exist in vivo, although the LC phase is the usual one. The authors’ simulations show that in the gel phase the lipid headgroups are partially dehydrated and sodium ions cannot penetrate to the interfacial region to bind to the carbonyl groups.

A new method for estimating hydration free energies of organic molecules with an implicit solvent is described by Maxim Federov of the MPI for Mathematics in the Sciences in Leipzig and coworkers (E. L. Ratkova et al., J. Phys. Chem. B 10.1021/jp103955r – paper here). It’s a modification of the RISM model of Chandler and Anderson, which provides a reasonably computationally cheap way of accounting for solvent structure along with some empirical parametrization of the solvation of specific chemical functionalities (alkyl, hydroxy, carbonyl etc.), ‘trained on’ and tested with small organics. No indication of whether it might be extended to macromolecules, although the authors intend to try it on bioactive compounds.

But the method for predicting hydration structures outlined by Karl Freed at Chicago and colleagues is explicitly geared towards proteins (J. J. Virtanen et al., Biophys. J. 99, 1611-1619; 2010 – paper here). They use simulations of the hydration of ubiquitin, lysozyme and myoglobin to calculate electron radial distribution functions for the different atom types, and then show that these can be used to generate electron densities – and from them, hydration structures – for other proteins. The electron distributions can also be used to calculate X-ray scattering intensity, and the authors intend to use this in future work to compare their predictions with experiment.

It’s the absence of water in the core of the ‘alpha-solenoid’ protein importin-beta that gives this spring-like molecule its astonishing elasticity, according to Helmut Grubmüller and colleagues at the MPI for Biophysical Chemistry in Göttingen (C. Kappel et al., Biophys. J. 99, 1596-1603; 2010 – paper here). Specifically, their simulations indicate that the hydrophobic core has a molten-globule-like conformation that governs its mechanical properties. As such, the protein occupies a middle ground between fully folded and intrinsically disordered, showing how what one might call ‘secondary unstructure’ determined by hydrophobicity allows fine-tuning of a biologically relevant physical property.

And not really ‘water in biology’ at all, but a neat experiment on water at a hydrophilic (mica) surface is reported by Jim Heath at Caltech and colleagues (K. Xu et al., Science 329, 1188-1191; 2010 – paper here). They deposit graphene sheets on mica, and observe islands on the surface with the AFM that appear to be water monolayers ‘sealed in’ by the graphene. These often have faceted edges, suggesting that they are ice-like even at room temperature, and they may be nucleated at surface defects. At 90 percent humidity a water monolayer appears to cover the entire mica surface, and a second adlayer may grow patchily on top.

Monday, August 16, 2010

Some hows and whys of protein folding

Why do small proteins have such a wide range of folding times? Hue Sun Chan and colleagues at the University of Toronto make a case, using MD simulations, that a critical factor in folding time is the desolvation barrier (A. Ferguson et al., J. Mol. Biol. 389, 619-636; 2009 – yes, an ‘old’ paper, but one I fear I overlooked at the time; paper here). For simulations of 13 proteins, they find that the folding rates span two orders of magnitude if desolvation barriers are not included, but 4.6 orders with those barriers added, which is closer to the range seen experimentally. Moreover, folding in the presence of these solvation effects becomes more cooperative and more channelled, and at the same time sensitive to the native protein’s topological complexity. In other words, if you want to understand protein folding, you probably need explicit water.

Ionizable groups are in a sense ‘incompatible’ with the hydrophobic interiors of proteins, and destabilize the native state, but are nonetheless sometimes found there. Daniel Isom and colleagues at Johns Hopkins consider why (D. G. Isom et al., PNAS 10.1073/pnas.1004213107 – paper not yet online). They do so via mutagenesis experiments that introduce Glu groups into internal hydrophobic sites in staphylococcal nuclease, and measure their pKa values and the effects on protein stability. The Glu groups are accommodated without any major conformational reorganization, suggesting that the protein interior shields the charges surprisingly well, behaving like a material with high dielectric constant. It’s not clear why, although penetration of water is one possibility. Whatever the reason, the results suggest that proteins have somehow (and for some reason) evolved a substantial stability against the internalization of charged groups (albeit with sufficient intolerance not to incorporate them indiscriminately).

Cryoprotectants such as trehalose are also produced by some organisms as a defence against dehydration rather than freezing. But it’s not clear how this protection works. One popular idea is that the sugar simply replaces water molecules hydrating the polar groups of lipids in membranes, or of proteins, maintaining the biomolecules and their aggregates in a fluid state in the face of dehydration. For membranes, the aim must be to prevent the formation of a gel state during dehydration, since this leads to detrimental leakage upon rehydration. This notion is examined by Roland Faller at UC Davis and colleagues using MD simulations (E. A. Golovina et al., Langmuir 26, 11118-11126; 2010 – paper here). They find that the core tenets of the water-replacement hypothesis – replacement of water with trehalose, avoiding a transition to the membranes’ gel state – are borne out, but that the structure of the membrane is different in the presence of trehalose than when fully hydrated.

The precise nature of (charged) lipid hydration is studied by Tahei Tahara and colleagues at RIKEN’s Advanced Science Institute in Saitama, using vibrational sum-frequency generation (J. A. Mondai et al., JACS 10.1021/ja104327t – paper here). They aimed to resolve a discrepancy between hydrogen-down and hydrogen-up water orientations at the interface reported from earlier experiments and simulations, and find that the former seems to hold for cationic lipids and the latter for anionic ones – which I guess is what one would expect on the most simplistic electrostatic grounds.

Manuel Aguilar and colleagues in Spain and Mexico report on MD simulations of the hydration of the tripeptide Cys-Asn-Ser (C. Soriani-Correa et al., J. Phys. Chem. B 114, 8961-8970 (2010) – paper here). Hydration stabilizes a more extended structure of the tripeptide than in the gas phase, owing to the replacement of intramolecular hydrogen bonds with intermolecular ones to water molecules.

Sorin Lusceac and Michael Vogel at the TU Darmstadt use deuterium NMR to investigate water dynamics in the hydration shell of myoglobin in the region of the 200-220 K dynamical crossover (J. Phys. Chem. B 10.1021/jp103663t – paper here). They observe a gradual change from isotropic to anisotropic rotation as the temperature is lowered from around 230 K. At that temperature the hydration water has access to essentially a continuum of orientational states, while by 165 K it can adopt only a few, maybe two. But this change is gradual: there is no sign of a sharp phase transition around 225 K, which has previously been claimed as the temperature of an abrupt fragile-to-strong transition.

Hydration-water dynamics as a function of protein concentration at ambient temperature are studied by Stephen Meech and coworkers at the University of East Anglia using the ultrafast optical Kerr effect, which can probe picosecond time scales (K. Mazur et al., J. Phys. Chem. B 10.1021/jp106423a – paper here). Below 0.4M peptide concentration (for three dipeptides), the water dynamics are slowed (particularly for hydrophilic peptides) but the water retains a primarily tetrahedral geometry. Above this concentration the dynamics are slower still and the tetrahedral network is perturbed, presumably due to intermolecular H-bonding between the peptides.

In fact, David LeBard and Dmitry Matyushov at Arizona State University argue on the basis of numerical simulations (of three globular proteins) that protein hydration shells have sufficient average orientational ordering among the water molecules to constitute a ferroelectric shell that propagates 3-5 molecular layers into the solvent (J. Phys. Chem. B 114, 9246-9258; 2010 – paper here). They argue that this is consistent with THz dielectric measurements, and say the dynamics are dominated by a slow (nanosecond) component that freezes at the protein’s dynamical transition.

Fluorescent molecules are sometimes used to probe the dynamics of DNA and its hydration sphere. For example, coumarin can be inserted into the double helix in place on an entire base pair, attached to one strand while the other is simply without a base. But as Kristina Furse and Steven Corcelli of the University of Notre Dame point out, this is a non-trivial substitution. So they have used MD to investigate how much this substitution perturbs the native state of DNA and its surrounding water molecules and ions (J. Phys. Chem. B 10/1021/jp105761b – paper here). They find that the effects can be significant – widening of the minor groove, increased flexibility, and increased water mobility. They conclude that this is not a reliable way to study, for example, the highly constrained water in DNA’s minor groove.

Melanin pigments are widely distributed in living organisms – in humans they appear in skin, hair, eyes, brain and liver. Their macromolecular structures are still not fully characterized, but seem to be highly dependent on hydration: water fills the slit-like pore regions between stacked graphitic plates in the pigment aggregates. Maria Grazia Bridelli and Pier Raimondo Crippa at the University of Parma use FTIR spectroscopy to look at this water in melanins under different degrees of hydration (J. Phys. Chem. B 10.1021/jp101833k – paper here). They suggest that the traditional picture of water in melanins being divided into relatively labile and tightly bound fractions is simplistic, and that in fact the distribution of adsorption sites is very heterogeneous, with a large pore size dispersion, and the water environments ranging continuously from highly bound in small pores to more or less bulk-like.

Hua Guo at the University of New Mexico in Albuquerque and coworkers have previously reported a ‘promoted-water’ mechanism, involving an active-site bound water molecule, for the action of carboxypeptidase A (CPA) in proteolysis (D. Xu & H. Guo, JACS 131, 9780; 2009). Now, using quantum MD simulations, they find something similar for the CPA-catalysed cleavage of esters (S. Wu et al., J. Phys. Chem. B 114, 9259-9267; 2010 – paper here). An alternative, ‘anhydrous’ nucleophilic mechanism seems to be ruled out for proteolysis, and the authors say that while it is feasible for esterolysis, it has a considerably higher free-energy barrier than the promoted-water pathway.

Hofmeister effects get subtler the harder you look. The series is reversed, say Pavel Jungwirth and colleagues in Prague and Lund, when ammonium halides are substituted by tetraalkylammonium cations (J. Heyda et al., J. Chem. Phys. B 10.1021/jp101393k – paper here). This effect is predicted by their MD simulations, and confirmed by experiment, and may be rationalized from a consideration of the different hydration structures of the cations.

Dissolved salts seem generally to increase water’s surface tension, and in ways specific to particular anions and cations that mirror the respective Hofmeister effects. Why? Irving Langmuir suggested that there is ion depletion at the interface; now we know that the effects may be subtle, especially at hydrophobic rather than free interfaces. Yan Levin and colleagues in Brazil have a shot at developing a first-principles theory based on an electrostatic approach to calculating Gibbs adsorption isotherms for the ions (A. P. dos Santos et al., Langmuir 26, 10778-10783; 2010 – paper here). They say that ‘kosmotropic’ (I know) anions are depleted at the interface, while chaotropic anions are absorbed (the theory is actually, as far as I can see, silent about the actual hydration structures of the ions). It predicts well the observed trends in surface tensions seen for the corresponding sodium salts.

Amphiphilic proteins tend to segregate to the air-water interface. Berk Hess of the MPI Mainz and colleagues say that the key driving force for small peptides of this type is the dehydration of hydrophobic residues, and that the effect scales linearly with the size of the molecules (O. Engin et al., J. Phys. Chem. B 10.1021/jp1024922 – paper here).

Roumiana Tsenkova at Kobe University and colleagues have proposed that studying water dynamics in biological systems using near-IR spectroscopy can provide a way of monitoring changes in an organism’s biological state – a method they call ‘aquaphotomics’ (see R. Tsenkova, J. Near Infrared Spectrosc. 17,303-313; 2009). They now propose that the technique can identify infection of soybean leaves with soybean mosaic virus in vivo, two weeks before the normal visual signs of infection in the plant (B. Jinendra et al., Biochem. Biophys. Res. Commun. 397, 685-690; 2010 – paper here). Two new NIR bands in the water region turn out to be highly sensitive to infection. The mechanism seems unclear; the authors say only that the virus seems to alter hydration hydrogen-bonded structures in a way that brings the water closer to bulk-like.

Another coarse-grained model for water is presented by Qiang Cui and colleagues at Wisconsin-Madison (Z. Wu et al., J. Phys. Chem. B 10.1021/jp1019763 – paper here). This groups four water molecules into a single site, represented as three electrostatic charges (which approximate the cluster’s dipole and quadrupole moments) and a non-electrostatic ‘soft’ interaction. The model is, however, optimized for the bulk and is considered unlikely to be applicable to ice; one imagines the same might be true of hydration structures in which local cluster geometries are non-bulk-like. A simple, computationally cheap model geared specifically to hydration is offered by Piotr Setny and Martin Zacharias (J. Phys. Chem. B 10.1021/jp102462s – paper here), in which solute-solvent and solvent-solvent interaction energies are calculated in a mean-field approximation on a BCC grid. The model performs well for predicting hydration energies of some drug molecules and for reproducing buried-water distributions in proteins.

I recently came across a nice review article by Felix Sedlmeier, Roland Netz and colleagues at TU Munich on ‘water at polar and nonpolar solid walls (F. Sedlmeier et al., Biointerphases 3(3), FC23-FC39 (2008) – paper here), which looks at what MD simulations have to tell us about statics, dynamics, rheology and so forth. And on this topic, Shu Nie and coworkers at Sandia Labs describe, on the basis of scanning tunnelling microscopy studies, an intriguing interfacial structure for a wetting water layer on Pt(111), in which the ice-like bilayer commonly reported is modified due to the appearance of 5-and 7-membered rings in the first wetting layer (S. Nie et al., Phys. Rev. Lett. 105, 026102; 2010 – paper here).

Tuesday, July 6, 2010

Water channels and flu resistance

What happens to hydrophobic interactions in the presence of charge? Bruce Berne and colleagues at Columbia have used MD simulations to explore that question (L. Wang et al., J. Phys. Chem. B 114, 7294-7301; 2010 – paper here). They find that the binding affinity of a hydrophobic particle to a hydrophobic plate, when it is placed between two such plates, is decreased if the plates are charged. Indeed, with increasing charge density the plates can become hydrophilic-like, expelling the particle from the interplate region.

Allan Friesen and Dmitry Matyushov at Arizona State University have considered the polarity of the interface between water and hydrophobic particles, which they model as hard spheres surrounded by a Lennard-Jones layer ( – paper here). (They call these particles, a little confusingly, ‘cavities’, presumably because they open up cavities in the solvent.) They find that there is a significant increase in the local polarity of the water at the interface, meaning that any charges in the solute particle are significantly screened. Moreover, dipolar relaxation in the first hydration shell is slowed significantly, with a relaxation time of around 50 ps.

Hirofumi Sato and colleagues at Kyoto University have used an approach called multicentre molecular Ornstein-Zernicke equation theory to calculate the hydration structure of bacteriorhodopsin (and a simpler serine coiled coil) (K. Hirano et al., J. Phys. Chem. B 114, 7935-7941; 2010 – paper here). I’ve not come across this method and don’t fully understand it – I’m rather surprised that a first-principles method like this, using the O-Z equation, can be applied to such a complex system. But given that it is 20 years since I last set eyes on the O-Z equation, it is very probable that I’m behind the times. In any event, the authors say that the solvent distribution they calculate agrees well with that found from XRD, and they can pull out the relative strengths of the hydrogen-bonding interactions, for example of the bound water molecules close to the Schiff base of bR.

The efficacy of the antiflu drug amantadine (AMT) is undermined by resistance that has been reported as developing in influenza A. A possible mechanism of resistance is documented by Kunqian Yu of the Shanghai Institute of Materia Medica and colleagues (G. Qiu et al., J. Phys. Chem. B 114, 8487-8493; 2010 – paper here). Their MD and quantum molecular mechanics calculations indicate that AMT binds in the pH-gated proton channel M2, as indeed it was designed to do. Here the drug disrupts a water wire that allows protons to cross the pore, and thereby inhibits its function. But AMT can occupy different positions in the channel, and in the resistant mutant SN13 it binds at a site that does not ‘snip’ the water wire: protons can still get through.

In photosystem II, water acts as a ligand, which is oxidized by a Mn4Ca cluster. It’s not been clear exactly where this water is bound, and that is what Robert Stranger and colleagues at ANU set out to establish using density functional calculations (S. Petrie et al., Angew. Chem. Int. Ed. 49, 4233-4236; 2010 – paper here). They find six waters bound close to the Mn cluster through the catalytic cycle, of which the two substrate waters fit in a cleft between two Mn atoms and the Ca.

The thermodynamics and kinetics of water confined between hydrophobic plates is investigated using MD with a simple monoatomic water model by Limei Xu and Valeria Molinero (J. Phys. Chem. B 114, 7320-7328; 2010 – paper here). They consider the conditions under which drying is induced, and look at the marginal situation in which there are rapid fluctuations between a wet and dry state. One obviously has to ask to what extent the quantitative conclusions here would apply to more sophisticated water potentials, but I’m also left to wonder whether the authors have really made proper contact with the vast earlier literature on liquid-vapour phase transitions in confined systems.

Shiang-Tai Lin at the National Taiwan University and colleagues present a method for calculating the entropy and energy of molecular liquids from the trajectory of MD simulations, which they apply to water using various potentials (S.-T. Lin et al., J. Phys. Chem. B 10.1021/jp103120q – paper here). They show that the technique has rapid convergence: these thermodynamic quantities can be computed from just 10 ps of simulation time.

Evan Williams and colleagues at Berkeley have used infrared photodissociation spectroscopy to investigate water-structuring effects of sulphate ions hydrated within clusters of up to 80 water molecules at 130 K (J. T. O’Brien et al., JACS 10.1021/ja1024113 – paper here). They present these results within the context of ‘structure-making and –breaking’ in the Hofmeister series, although naturally it is an open question to what extent the structures seen here reflect those in bulk solution at ambient temperatures. They say they see the spectral signature of bulk-like water – or more precisely, of free OH groups at the cluster surfaces analogous to those at the bulk surface – for clusters with more than about 43 water molecules, equivalent to the third solvation shell. The authors say that in fact their results imply that, although a sulphate ion ‘patterns’ water molecules ‘to a distance much farther than the first solvation shell’, this does not alter the number and strength of the hydrogen bonds beyond the first shell. But it is not really clear what the consequences for Hofmeister effects are, beyond the suggestion by Williams et al. that experiments that probe only rotational dynamics may miss some of the more subtle ‘patterning’ (presumably ordering) effects evident in this spectroscopic study.

Andrei Sommer and his colleagues have extended their previous investigations of the anticancer effects of red light and green tea (thought to be operating via photo-induced changes in the ordering of interfacial water – see here) (A. P. Sommer et al., Photomed. Laser Surg. 28, 429-430; 2010 – paper here). They find that the two things administered together retard the growth of HeLa cells. The mechanism remains speculative.

MD simulations with explicit water are computationally intensive, for which reason Piotr Setny and Martin Zacharias at the TU Munich have developed a computationally efficient way to model hydration (J. Phys. Chem. B 10.1021/jp102462s – paper here). It is a lattice cellular automaton that calculates solute-solvent and solvent-solvent interaction energies using a mean-field approach, which can calculate whether a particular grid cell at the solute surface is hydrated. It performs well in terms of predicting hydration energies for drug molecules, as well as locating buried water molecules in cavities.

Water molecules passing through carbon nanotubes can be pulled by methane molecules across the potential barriers created by tapering junctions where two tubes of different radii are joined, according to simulations by H. Li at Shandong University in Jinan, China, and colleagues (H. Q. Yu et al., J. Chem. Phys. B 10.1021/jp102810j – paper here). The ‘dragging’ effect is mediated by van der Waals interactions, they say.

Wednesday, June 9, 2010

Drying-induced pore gating?

Morten Jensen and colleagues at D. E. Shaw Research in New York show using MD simulations of a potassium channel that hydrophobic gating – dewetting transitions in the hydrophobic pore interior – appear to be a viable mechanism for these and perhaps many other ion channels (M. Ø. Jensen et al., PNAS 107, 5833-5838; 2010 – paper here). This is one of the most persuasive cases I’ve seen for this very interesting possibility, building as it does on the full atomistic structure of the pore protein.

S. Khodadadi at Akron and colleagues have an interesting paper about the hydration of tRNA (Biophys. J. 98, 1321-1326; 2010 – paper here). They use neutron scattering and dielectric spectroscopy to measure the hydration dynamics of this molecule, and find that these are slower than those for typical proteins, but faster than DNA. This, they say, challenges the ‘slaving’ hypothesis, ‘which assumes that the dynamics of biological macromolecules just follows the dynamics of hydration water’. But is that really what it assumes? People have written various things about this, and I’d have to go back and check the papers they cite, but I’d not understood ‘slaving’ to mean something quite so simple – rather, it implies only that the dynamics of the solute and solvent are interdependent. Certainly, I don’t know that anyone would suggest the dynamics of hydration water are bulk-like and unaffected by the nature of the macromolecule. Still, interesting results.

As Thomas Elsaesser at the Max Born Institute in Berlin point out (L. Szyc et al., J. Phys. Chem. B 10.1021/jp101174q – paper here), the residence times of hydration water molecules around DNA actually show an analogously broad distribution to those around proteins, and fluctuations in the hydrogen-bond network happen on fast (fs to ps) timescales. They have looked at how vibrational energy pumped into the phosphate groups gets redistributed into DNA’s hydration shell, which acts as a heat sink. This happens quickly too: within a few fs, while energy transfer within the DNA molecule is slower (timescales around 20 ps).

Returning to proteins, Stefania Perticaroli and colleagues at Perugia find using depolarized light scattering that the water dynamics in dilute solutions of lysozyme display two distinct timescales: fast (>ps) bulk-like relaxation, and slow (a few ps) due to hydration water (S. Perticaroli et al., J. Phys. Chem. B 10.1021/jp101896f – paper here).

One possible mechanism for the operation of antifreeze proteins is to bind to the surface of ice crystallites and stop them growing further. According to Knight and DeVries, this should also imply that the AFP’s should inhibit the melting of ice. Yeliz Celik at Ohio University and colleagues have presented experimental evidence for this (PNAS 107, 5423-5428; 2010 – paper here). They find that ice can be superheated up to 0.44 C in AFP solutions.

Does the electric field at a charged surface induce an ice-like hydrogen-bonding pattern in water? No, accortding to Tahei Tahara and colleagues at RIKEN’s Advanced Science Institute in Saitama, Japan (S. Nihonyanago et al., JACS 132, 6867-6869; 2010 – paper here). They use a form of vibrational sum frequency generation spectroscopy to look at the water structure in the Gouy-Chapman layer at the surface of a charged lipid monolayer, and say that it looks bulk-like.

Well, by this measure perhaps. But FT-IR and DSC measurements of water confined between lamellar bilayers of AOT surfactant suggest that the water here has three components: some is bulk-like, some closely linked to the surfactant head groups, and a layer about 0.5 nm between the two where the bulk H-bond network is disrupted (E. Prouzet et al., J. Phys. Chem. B 10.1021/jp101176v – paper here).

Another nice example of water in the active site playing a key role in enzymatic activity: Karol Kaszuba at the Tampere University of Technology in Finland and colleagues say that water in the binding site enables the high stereoselectivity of a beta-adrenergic receptor, a potential target for beta-blockers, by mediating H-bonding interactions between different enantiomers of the beta-blocker nebivolol (K. Kaszuba et al., J. Phys. Chem. B 10.1021/jp909971f – paper here).

Many simulations have shown single-file filling of and transport through carbon nanotubes. Now Wim Wenseleers and colleagues at Antwerp claim to have seen this experimentally for the first time (S. Cambré et al., Phys. Rev. Lett. 104, 207401; 2010 – paper here). They see the signature of such behaviour in the splitting of a Raman mode, for nanotubes of diameters down to 0.548 nm. But interestingly, the details of the filling (as revealed by the Raman shift) shows a complex, non-monotonic dependence on diameter, owing to variations in tube chirality.

Martina Havenith at Bochum and her coworkers have been using THz spectroscopy to reveal some interesting new features of hydration. Now she, Dominik Marx and their colleagues have used MD simulations to figure out precisely what manner of intermolecular motions are being probed by this technique (M. Heyden et al., PNAS 10.1073/pnas.0914885107 – not yet online). They conclude that ‘a modification of the hydrogen-bond network, e.g. due to the presence of a solute, is expected to affect vibrational motion and THz absorption intensity at least on a length scale that corresponds to two layers of solvating water molecules’ – and that this spectroscopy particularly probes strongly correlated molecular motions.

Here’s something that should stir up discussion: J. Raúl Grigera and Andres McCarthy in Argentina have conducted MD simulations of the pressure induced denaturation of proteins such as lysozyme and apomyoglobin, and say that they think the unfolding is caused by weakening of hydrophobic interactions owing to a change in water structure (weakening of the H-bonding network) (Biophys. J. 98, 1626-1631; 2010 – paper here). Other studies have tended to emphasize the penetration of water into the hydrophobic interior. Besides my now almost knee-jerk suspicion of explanations invoking ‘water structure’, I am forced to wonder, first, if such changes would be profound enough at the kbar pressures used here, and second, to what extent the hydrophobic interaction is really dependent on the H-bond network as opposed to being a more general solvophobic effect.

Nanobubbles – ah, one day I hope to bring together the various and diverse findings reported about their behaviour. In some studies (e.g. Jin et al., J. Phys. Chem. B 111, 2255; 2007), it has been claimed that nanobubbles will form in mixtures of non-aqueous solvents (that have some amphiphilic properties) with water. That notion is investigated by Xuehua Zhang and coworkers at the University of Melbourne (A. Häbich et al., J. Phys. Chem. B 114, 6962-6967; 2010 – paper here). They find that the behaviour of such mixtures is inconsistent with light scattering by bubbles (for example, there is no change in scattering on degassing), and they attribute the scattering instead to the presence of impurities.

Microbubbles are the focus of a study by Derek Chan, also at Melbourne, and colleagues (I. U. Vakarelski et al., PNAS 10.1073/pnas.1005937107; paper here). They use AFM measurements to look at the factors that influence bubble coalescence, particularly the hydrodynamic interactions between microbubbles. Curiously, they say that the hydrodynamics can induce a dynamic coalescence mode which operates as two bubbles separate.

Going back to January, Jan Swenson and colleagues at Chalmers University of Technology in Sweden claim to have identified a slow relaxation process in water (four orders of magnitude slower than the normal viscosity-related relaxation) due to collective motions of the hydrogen-bonded network (H. Jansson et al., Phys. Rev. Lett. 104, 017802; 2010 – paper here). They say that this type of relaxation has been identified before in mono- and polyalcohols, such as glycerol (R. Bergman et al., J. Chem. Phys. 132, 044504; 2010). The researchers see it in water in measurements of the dielectric response. They can’t really say much yet about what causes it, although the suggested connection to the ‘chain-like’ structures proposed by Huang et al. (PNAS 106, 15214; 2009) is speculative in the extreme.

See also Jan’s recent paper with José Teixeira (J. Chem. Phys. 132, 014508; 2010 – paper here) on the relaxation behaviour of supercooled water through the no-man’s-land between 150 and 235 K. They propose a crossover between cooperative α-relaxation at higher temperatures and ‘local’ β-relaxation at low temperatures.

Also, I don’t believe I mentioned previously a paper by Alenka Luzar and colleagues published last November (Phys. Rev. Lett. 103, 207801; 2009 – paper here) on the dynamics of alignment of a hydrated nanoparticle in an electric field. This process is important for applications such as dielectrophoresis and the electrical control of optical properties. Using MD simulations, the researchers conclude that the torque exerted by a typical experimentally realizable field strength is greater than kT (so alignment is possible even at the nanoscale) and greater than that estimated using continuum methods. Moreover, the alignment times are fast – of the order of a few hundred picoseconds.

Greg Voth and colleagues run a check on the self-consistent charge density functional tight binding (SCC-DFTB) method that has been used for quantum simulations of water in various systems, including some biological ones (C. M. Maupin et al., J. Phys. Chem. B 114, 6922-6931; 2010 – paper here). They look at what the method predicts for hydrated protons, and find that it puts the excess proton in a Zundel ion (H5O2+) in the resting state – unlike some other quantum chemical methods, and in contrast to what experiments have suggested. This presumably raises questions about the validity of the method.

Brad Bauer and Sandeep Patel at the University of Delaware also present a kind of model validation study for different water potentials, looking at how these affect the hydrophobic attraction of two flat plates (J. Phys. Chem. B 10.1021/jp101995d – paper here). They find that while many of the structural and dynamic aspects are the same for all the potentials studied – average density, fluctuations, hydrogen bonding – the potential of mean force for attraction between the plates is reduced when the water is polarizable.

Another validation study for simulations of biomolecules is described by Klaus Liedl at Innsbruck and colleagues (J. Phys. Chem. B 114, 7405; 2010 – paper here). They look at how simulations of the X-ray structure of the protein fXa, a key enzyme in blood coagulation, are affected by different choices of sets of water molecules in the hydration sphere. They conclude that only by judicious placement of water molecules around the protein, using available crystal structure data, will ensure a reasonable sampling of phase space when studying the protein’s dynamics. Otherwise, the simulations may take an unfeasible time for the hydration environment to equilibrate. One can’t, apparently, just plunge the protein into a bulk-like solvent environment and assume that it’ll find its own way to the right hydration structure.

A few more to come, but this is enough for now.

Two announcements of publications:
There is a special volume of the Journal of Electron Spectroscopy and Related Phenomena (177 (2-3), March 2010) devoted to water and hydrogen bonds investigated through inner-shell spectroscopies.
And the long-overdue collection of papers stemming from a conference on ‘water and life’ in Varenna in 2005 is now out:
Water and Life: The Unique Properties of H2O, eds Lynden-Bell, Ruth M., Conway Morris, Simon, Barrow, John D.,Finney, John L., and Harper, Charles L., Jr. Boca Raton, Florida: CRC Press / Taylor & Francis Group, 2010. More details here. I just received my copy, and it looks still relevant despite the long delay in publication.

Wednesday, April 28, 2010

Another catch-up

Apologies for a longer-than usual silence – I’m waiting to resolve some access problems. In the meantime, and before the backlog gets too awesome, I’ll have to work with mostly just abstracts here.

I’ve been meaning to comment for a long time on a study by David Chandler and his colleagues that extends his notion of fluctuation-driven hydrophobic forces (A. J. Patel et al., J. Phys. Chem. B 114, 1632-1637; 2010 – paper here). Since the original Lum, Chandler & Weeks paper on ‘drying-induced attraction’, David has been developing the idea that what we’re dealing with at a hydrophobic surface is not so much a static gas-like layer but a density depletion due to enhanced density fluctuations. Here he and his coworkers use simulations to show that these fluctuations are similar to those at a water-air interface, and that the resulting depletion does seem to drive the hydrophobic attraction between two such surfaces.

Francesco Mallamace and colleagues have reported experimental evidence of the dynamical crossover in supercooled water that they have previously postulated as an explanation of the glass-like transition in supercooled hydrated proteins (F. Mallamace et al., J. Phys. Chem. B 114, 1870-1878; 2010 – paper here). They have used NMR and neutron scattering to look at water confined in a nanotube, water in the hydration layer of lysozyme and water in a methanol mixture. In all cases they see the predicted change in temperature-dependence of viscosity from Arrhenius to non-Arrhenius form, and say that this seems to coincide with the development of an extended H-bonded network.

Yurina Sekine and Tomoko Ikeda-Fukazawa at Meiji University in Japan appear to be proposing another kind of transition in glassy peptide-like polymers at 37 C. They see a shift in the Raman O-H stretching mode of bound water (to poly-N,N-dimethylacrylamide) at this temperature (J. Phys. Chem. B 114, 3419-3425; 2010 – paper here). I’m not sure that I fully underastand what is going on here without seeing the full paper, but the transition seems to be marking a switch between probing the dynamics of the hydration layer I general below 37 C and the waters bound specifically to polar groups above 37 C.

More on protein denaturation. Angel Garcia and colleagues at RPI have used MD simulations to look at the mechanism of urea-induced unfolding of the Trp-cage peptide (D. R. Canchi et al., JACS 132, 2338-2344; 2010 – paper here). Like earlier studies with urea, they find that the denaturation seems to stem from direct interactions between the denaturant and the peptide chain – via electrostatic and van der Walls interactions rather than hydrogen-bonding.

Nohad Gresh and coworkers in Paris find using molecular mechanics simulations that the energetics of docking of inhibitors to a protein called the focal adhesion kinase depends critically on a group of 5-7 structured water molecules at the binding site (B. de Courcy et al., JACS 132, 3312-3320; 2010 – paper here).

Biman Bagchi and colleagues at the Indian Institute of Science in Bangalore say that the behaviour of bound water within the major groove of DNA is different from that within the minor groove (B. Jana et al., J. Phys. Chem. B 114, 3633-3638; 2010 – paper here). Their MD simulations of hydrated ploy-AT and poly-GC show that the minor-groove water has slower dynamics due to greater tetrahedral ordering.

How cryoprotectants such as poly-sugars work is still not clear. Fabio Bruni and colleagues have looked into this using neutron diffraction to study the hydration of the disaccharide trehalose (S. E. Pagnotta et al., J. Phys. Chem. B 114, 4904-4908; 2010 – paper here). One hypothesis has been that the tetrahedral structure of water is strongly modified in the hydration shell of trehalose. But the experiments show little sign of this; indeed, rather few water molecules are hydrogen-bonded to the disaccharide. Another blow for a ‘modified-water-structure’ view.

How about urea? That question has been much debated, but Abdenacer Idrissi at the Université des Sciences et Technologies de Lille and colleagues consider the issue using MD simulations (A. Idrissi et al., J. Phys. Chem. B 114, 4731-4738; 2010 – paper here). They find that as the concentration of urea in solution is increased, the tetrahedrality of water declines in favour of an ‘unstructured’ arrangement. What this means for ‘water structure’ as such is perhaps quite subtle, given the method used to compute ‘tetrahedrality’ (i.e. a comparison of the mutual orientation of a ‘probe’ water molecule and an adjacent ‘tetrahedral’ group of them). To be continued, I’ve no doubt.

And also on hydration of small organic molecules, Richard Saykally and colleagues have used XAS to look at the hydration shells of alanine and sarcosine (the smallest peptoid or ploy-N-substituted glycine) (J. S. Uejio et al., J. Phys. Chem. B 114, 4702-4709; 2010 – paper here). The two are hydrated in rather different ways: the sarcosine XAS spectrum is much less affected by hydration than is alanine, but much more affected by conformational changes.

You’d have thought that the situation of water confined in slit-like pores or between parallel plates would have been exhaustively studied by now. But Yubo Fan and Yi Qin Gao at Texas A&M report MD studies of this geometry, using either hydrophobic plates or alkane monolayers, for relatively large separations (up to 800 Å) (J. Phys. Chem. B 114, 4246-4251; 2010 – paper here). They say that the effects of confinement are evident in the centre of the pore even for such large separations, with the water in the centre being (surprisingly, I think) of somewhat reduced density and more ice-like. This surprises me very much, to the extent that I am sceptical without seeing the full paper (I don’t know what the temperatures are). I’d not expect to see any significant departure from bulk-like water beyond distances of, say, 2 nm or so from the surfaces.

Daisuke Matsuoka and Masayoshi Nakasako in Japan have developed a program for predicting the hydration structures around the hydrophilic surfaces of proteins, based on their crystal structures (J. Phys. Chem. B 114, 4652-4663; 2010 – paper here). This simply sums the hydration distribution functions for each solvent-exposed polar atom. I’d have expected there to be more cooperativity than this would seem to allow, but it seems that the program works well when tested against known structures, e.g. lysozyme, bacteriorhodopsin, aquaporin.

An unusual approach to the hydration of proteins is taken by Vitaly Kocherbitov and Thomas Arnebrant at Malmö University (Langmuir 26, 3918-3922; 2010 – paper here). They adapt a method commonly used to study adsorption at the solid-gas interface, based on a BET-type analysis but allowing for heterogeneity of the surface. This may apply to the case of ‘dry’ proteins exposed to a humid environment, but presumably not to proteins in solution.

Amit Galande and colleagues at SRI International in Virginia report some designed peptides that will fold in solution via intramolecular hydrogen bonds, regardless of competition from solvating water molecules (B. Song et al., Langmuir 132, 4508-4509; 2010 – paper here). I can’t immediately see if there are generic principles here that get the free-energy balance right.

From time to time, efforts are made to find a computationally cheap way to approximate water structure in complex simulations. Kevin Hadley and Clare McCabe at Vanderbilt University suggest one such (J. Phys. Chem. B 114, 4590-4599; 2010 – paper here). They propose a coarse-graining in which four-molecule water clusters can be represented by single ‘beads’ in a simulation.

Sason Shaik and coworkers have extended their investigations of the role of water in heme catalysis (P. Vidossich et al., J. Phys. Chem. B 114, 5161-5169; 2010 – paper here; and D. Fishelovitch et al., J. Phys. Chem. 10.1021/jp101894k – paper here). They have computed the free-energy landscape for the position of the water molecule that provides a crucial hydrogen-bonded bridge between peroxide (complexed to ferryloxo) and a histidine residue in the active site of peroxidase, showing that the ‘reactive configuration’ corresponds to a minority population, albeit one that is relevant on the timescale of catalysis. And they clarify the roles of a water channel in the active site of cytochrome P450, showing how this facilitates proton transport.

In a related vein, the role of water in the active site of ribonuclease H is studied by C. Satheesan Babu and Carmay Lim in Taiwan (JACS 10.1021/ja101494m; paper here). They find that two different binding modes of magnesium ions, which act as cofactors, are distinguished by having a water-rich and water-depleted environment. This might have implications for the design of inhibitors.

Hydration of the head groups of a phosphatidylcholine film is investigated by Yuki Nagata and Shaul Mukamel at UC Irvine using SFG, revealing three distinct environments for the water molecules at the interface (JACS 10.1021/ja100508n; paper here).

A curious paper by Michele Pavanello at the University of Arizona and coworkers looks at how solvation influences hole transport in DNA, which is relevant to the issue of radiation-induced damage (M. Pavanello et al., J. Phys. Chem. B 114, 4416-4423; 2010 – paper here). The study looks at (theoretical) DNA conductivity of a DNA double strand contacted by an STM tip, and finds that hydration slows hole transport significantly.

I don’t really know what to make of a paper by Dariusz Czapiewski and Jan Zielkiewicz at the Gdansk University of Technology on the structures of hydration shells around peptides (J. Phys. Chem. B 114, 4536-4550; 2010 – paper here). As far as I can tell, they use some approximate analytical method to calculate the degree of ‘water ordering’ in the hydration shells, and conclude that it is not very different from the bulk, but is ‘pseudo-rigid’, with strengthened hydrogen bonds. That would surprise me.

I have a few other bits and pieces to add at some stage, but this brings things relatively up to date for now.

Tuesday, February 16, 2010

Denaturants again

Yes, more on denaturants: Shekhar Garde and colleagues at RPI have studied the effects of the common denaturant guanidinium chloride on hydrophobicity using MD simulations (R. Godawat et al., J. Phys. Chem. B jp906976q – paper here). GdmCl acts like simple salts (NaCl and CsCl) in increasing the surface tension of water and decreasing the solubility of small hydrophobic solutes (that is, in this case making it harder to insert a hard sphere into solution). But it also destabilizes the compact state of a hydrophobic polymer. Consistent with earlier studies, it does so via direct (vdW) interaction with the polymer backbone, and not via any indirect effect on ‘water structure’.

Meanwhile, Pannuru Venkatesu at the University of Delhi and colleagues have studied the effect of denaturants (urea, GdnHCl) and osmolytes (TMAO, betaine, sucrose and others) on the activity of an enzyme, specifically alpha-chymotrypsin (P. Attri et al., J. Phys. Chem. B 114, 1471; 2010 – paper here). They measure the stability of the enzyme as reflected in the Gibbs free energy of unfolding changes, and the enthalpy change, on addition of cosolvent. These variables increase with increasing osmolyte concentration, and decrease on addition of denaturant – which is, I guess, what one would anticipate. CD spectroscopy suggests that these effects are manifested via changes in beta-helix stability. The osmolytes do not, however, seem to affect enzyme activity. In contrast, and consistent with the study above, denaturants seem to act by binding to the enzyme surface, and induce sufficient structural disruption to reduce activity virtually to zero.

And Yi Qin Gao and colleagues at Texas A&M have used MD simulations to look at the effects of urea, tetramethyl urea (TMU) and the osmolyte trimethylamine N-oxide (TMAO) on the structure of water and dissolved proteins (H. Wei et al., J. Phys. Chem. B 114, 557; 2010 – paper here). TMAO weakens interactions between amide carbonyls on model peptides and water, while urea and TMU strengthen them. Consistent with previous studies, they find a direct interaction between urea and the peptides (via the carbonyls). But they also find evidence of a role for indirect effects on denaturation, whereby the cosolvents alter the hydrophobic interaction via changes to the structure and dynamics of water. They find that a peptide fragment of a G protein unfolds by step-by-step breaking of its native hydrogen bonds, coupled to the formation of water-carbonyl bonds.

Similar territory is explored by Feng Guo and Joel Friedman at the Albert Einstein College of Medicine, who have used vibronic sideband luminescence spectroscopy of a gadolinium(III) probe ion to explore changes in hydrogen bonding between hydration waters of a protein induced by osmolytes (J. Phys. Chem. B 113, 16632; 2009 – paper here). They say that while urea initially weakens hydrogen bonding in the hydration layer, polyol osmolytes such as trehalose, sucrose and glucose enhance it. But they argue that as the concentration of urea increases, it actually enhances water occupancy within the protein and hydrogen bonding in the hydration layer, and that this is the first step in the urea-induced unfolding process. There is a delicate balance here, they say, between entropic effects that favour water penetration of the protein and enthalpic effects that favour a robust hydrogen-bonded hydration network. If I understand the argument correctly, the latter dominates for osmolytes and accounts for the stabilization of the compact folded structure in that case.

The nature of the hydrated hydrogen ion has been much debated. Christopher Reed and colleagues at UC Riverside have investigated the issue using IR spectroscopy, and argue that the best description is neither an Eigen ion (H9O4+) nor a Zundel ion (H5O2+), but the species H13O6+, containing a delocalized proton in the central O-H-O group (E. S. Stoyanov et al., JACS 10.1021/ja9101826 – paper here).

Greg Voth and Takefumi Yamashita have meanwhile investigated the nature of hydrated protons near lipid membranes, using MS-EVB calculations (J. Phys. Chem. B 114, 592; 2010 – paper here). They are interested in clarifying the proposed proton-antenna effect whereby lipid membranes collect protons and shuttle them by lateral diffusion to membrane proteins such as ATP synthase. They confirm this picture, saying that the effect arises because of the stabilization of the hydrated proton by the lipid phosphate groups: a Zundel-like cation bridges phosphate and carbonyl groups. Diffusion of protons within the interface region is significantly slower than it is in the bulk.

Voth, along with Noam Agmon and Hanning Chen, also has a paper on the kinetics of proton transport in pure water (J. Phys. Chem. B 114, 333; 2010 – paper here). The calculations support the idea that the migration of the proton (in effect, of the centre of excess charge) depends on significant reorganization of (perhaps up to 20!) surrounding water molecules.

Mischa Bonn and colleagues at FOM in the Netherlands have used a microfluidic device to study changes in proton mobility near hydrophobic surface (JACS ja9083094 – paper here). They figured that if water molecule reorientation is, as posited, crucial to rapid proton migration, then the slower reorientation of waters near hydrophobic groups seen previously by Rezus and Bakker (Phys. Rev. Lett. 99, 148301; 2007) should have a significant effect on proton transport in such an environment. It’s a very neat experiment: laminar flow in the device means that fluorescein fluorescence in one half of the microfluidic channel may be quenched by lateral proton transport from the other half down a pH gradient. The proton diffusion decreases by an order of magnitude when the hydrophobe tetramethylurea is added, which is consistent with (if not perhaps definitive support for) the hypothesis.

The diffusion of water molecules at lipid surfaces, meanwhile, has been investigated experimentally by Ravinath Kausik and Songi Han at UCSB, using Overhauser dynamic nuclear polarization of the proton NMR signal (JACS 131, 18254; 2009 – paper here). At this stage this is largely a demonstration of the feasibility of the technique, which they hope will also be applicable to the study of solvent dynamics in the hydration shells of macromolecules.

Further along this paper trail, it is the mobility of water molecules in the hydration shells of peptides that is the subject of a simulation study by Charusita Chakravarty at the Indian Institute of Technology in Delhi and colleagues (M. Agarwal et al., J. Phys. Chem. B 114, 651; 2010 – paper here). They consider two small peptides: a 16-residue beta-hairpin structure, and deca-alanine. They see layering structure in the water extending at least 10 Å from the peptide surfaces, and the energetics and dynamics are significantly perturbed relative to the bulk: the discussion is couched mostly in terms of the ‘tagged potential energy’, said to be equivalent to the binding energy of an individual water molecule at a particular location at a given instant in time. This is typically 10-15 percent lower in the innermost hydration layer than in the bulk. But I’m not entirely clear what that implies: does it make diffusion faster or slower? (Surely the latter, but that’s not obvious from this measure.)

In the previous post I mentioned work by Ronen Zangi challenging the notion of ions as structure-makers and structure-breakers. Martina Havenith at Bochum and her colleagues have now raised further problems for this issue, using THz spectroscopy of salt solutions (D. A. Schmidt et al., JACS 131, 18512 (2009) – paper here). They say that the results suggest all ions can be considered simply as defects in the H-bonded network, and so can’t be regarded as either chaotropes or kosmotropes. It turns out that the data can be understood using an appealingly simple model in which the ions undergo damped harmonic oscillations – ‘rattling’ – within the water network. In other words, at least the fast (sub-picosecond) ion motions are essentially decoupled from the dynamics of the network.

One suggested mechanisms of pressure-induced denaturation is the penetration of water into the hydrophobic core. This notion is investigated in simulations by Takashi Imai and Yuji Sugita at RIKEN in Japan (J. Phys. Chem. B jp909701j – paper here). Using ubiquitin as a model case, they look in particular at competing scenarios: does water first penetrate and force the protein to swell, or are the cavities pre-formed by structural fluctuations and then fill with water? They find support for the latter picture: the influx of water stabilizes a pre-existing metastable structure, and drives a distinct transition to a relatively unfolded state.

Harold Sheraga at Cornell and colleagues have probed the nature of hydrophobic interactions as the size of hydrophobic solutes approaches the nanoscale limit (M. Makowski et al., J. Phys. Chem. B jp907794h – paper here). Specifically, they consider the crossover point of around 1 nm at which Lum, Chandler and Weeks (J. Phys. Chem. B 103, 4570; 1999) predicted that the mechanism of the hydrophobic interaction will change from that of small solutes to that of extended surfaces. They calculate the potentials of mean force for large hydrophobes such as adamantine and C60 in water. When two solvation spheres for such large species overlap as they form a dimmer in solution, the water molecules trapped in the concave ‘cleft’ at the edges of the interaction have restricted motion and decreased entropy that offsets any free energy gains from increased solute-particle contact. It seems that fullerenes like this, while too large to be treated as small hydrophobic solutes, are not yet large enough to be considered macroscopic hydrophobic surfaces.

Todd Sformo has sent me a fascinating paper on cold survival strategies of the Alaskan beetle (J. Exp. Biol. 213, 502; 2010 – paper here). He and his coworkers finds that this bug vitrifies at around –76 C, and by that means it can survive cooling of an amazing –150 C. The cryoprotection that supports vitrification is in this case glycerol.

When I was the editor at Nature for Reza Ghadiri’s paper on peptide nanotubes in 1993 (Nature 366; 324), I had to make the decision as something of an act of faith. I remain deeply glad that I did, for the work has stood the test of time. Now Padmanabhan Balaram and colleagues at the Indian Institute of Science have used peptide nanotubes formed from non-cyclic pentamers to study single-file water wires threading through their hydrophobic central channels in molecular crystals (U. S. Raghavender et al., JACS ja9083978 – paper here). This looks like an attractive model system for investigating the relationship between the water structures and the chemical nature of the wall ‘lining’.

Gene Stanley and his coworkers propose a new way of considering the structure of water that involves characterizing the ‘tetrahedral entropy’ associated with the degree of tetrahedral order (P. Kumar et al., PNAS 10.1073/pnas.0911094106 – paper here). They say that this parameter accounts for the specific heat maximum as the Widom line – where there is a cross-over from non-Arrhenius to Arrhenius dynamical behaviour, in general under conditions of supercooling – is crossed.

Tuesday, February 2, 2010

The post-Christmas glut

Well, it was always going to be this way: after several weeks away from the blog there’s now a big stack of papers to catch up on. The list here is incomplete, but more will follow.

One of the most interesting and important papers in this current stack is a MD study by Ronen Zangi of the notion of structure-making and structure-breaking in Hofmeister effects (J. Phys. Chem. B 114, 643; 2010 – paper here). In short, this study offers little succour for that concept, which has long overstayed its welcome. Ronen looks at the correlation between the propensity of various ions to alter the hydrophobic interaction (and thus to salt in/salt out) and changes in structural and dynamical properties they induce in the solvent. While there is a monotonic relationship between the reduction in hydrophobic interaction and the increase in ‘water structure’ as measured by the partial radial distribution factors, Ronen says that he could not identify ‘one property that can predict the change in the strength of the hydrophobic interacitons’. Nor could such properties predict the transition from salting-in to salting-out behaviour. Changes in dynamics, meanwhile, were induced by changes in the ion-water interaction, and not changes that the ions introduce to the ‘structural ordering’ of the water itself. As a result of all this, it seems that predicting whether a particular ion will induce salting-in or salting-out cannot be done on the basis of the properties of the salt solution alone, in the absence of the solute, and the whole notion of kosmotropes and chaotropes seems misleading. It would be nice to think that this paper will serve to banish those terms, but I suspect they will sadly take rather more dislodging than that.

Shekhar Garde and his colleagues have extended their earlier work on the conformations of polymers at surfaces (S. N. Jamadagni et al., Langmuir 25, 13092; 2009 – paper here). They have previously shown (J. Phys. Chem. B 113, 4093; 2009) that hydrophobic polymers adsorb preferentially at the interface between water and a hydrophobic surface (or air), and that the polymers here have significantly different structure and dynamics to those in the bulk. The present study looks in more detail at what is going on there, considering surfaces with a range of chemistries from hydrophilic to hydrophobic. The ‘test polymers’ are hydrophobic 35-mers, and the surfaces are SAMs with different terminal groups. The preferential adsorption at hydrophobic surfaces seems to be due to changes in water dynamics: the water has greater density fluctuations here and lower free energy of cavity formation, making it more able to solvate hydrophobes. The polymers have greater translational diffusion and conformational flexibility, typically flattening into pancake-like shapes.

I’ve been looking somewhat into the literature on nanobubbles, and it seems increasingly clear that it is very much in a state of flux and probably in need of some sort of snapshot review. How and when do nanobubbles form in the bulk and at surfaces? How long-lived are they, and how do they survive at all? There are many questions, and the answers so far are diverse. Detlef Lohse and his coworkers have a new contribution on the subject (B. M. Borkent et al., Langmuir 26, 260; 2009 – paper here). They try to clarify the shape of nanobubbles on hydrophobic surfaces (HOPG) using AFM, saying that they seem uniformly to have contact angles of about 119 degrees even for radii as small as 20 nm. It seems that some cantilevers can deposit material on the surfaces, making them rougher and altering the contact angle.

Robert Bryant and coworkers at the University of Virginia have used magnetic relaxation dispersion spectroscopy to characterize the dynamics of protons in a protein (BSA) backbone and its hydration water (G. Diakova et al., Biophys. J. 98, 138; 2010 – paper here). They find remarkably constant relaxation behaviour in the protein over a wide frequency range (0.01-300 MHz). Water dynamics contribute significantly to the relaxation on timescales of tens of ns, thanks to some rare, rather highly constrained, perhaps buried, hydration waters.

There’s a fascinating exploration of the various functional roles that protein hydration waters can have by Matteo Ceccarelli abd colleagues at the University of Cagliari in Italy (M. A. Scorciapino et al., JACS ja909822d – paper here). They have looked at myoglobin as a model system using MD, and observe three distinct ways in which waters modify the intrinsic dynamical behaviour of the protein. They can (1) block access to or escape of ligands from a binding site; (2) change internal dynamics by expanding the distances between residues in the manner of a ‘wedge’; (3) assist ligand transport, in effect by ‘washing it away’.

Another lovely example of water molecules playing an active role in an important biological process is provided by Göran Wallin and Johan Åqvist at Uppsala (PNAS pnas.0914192107 – paper here). They show that a water molecule trapped at the active site of peptide bond formation on the ribosome serves in a proton shuttle, while a second water molecule helps to stabilize the negative charge on the substrate.

Erik Sunde and Bertil Halle at Lund show how water proton magnetic relaxation dispersion measurements can provide information on slow protein dynamics by virtue of the exchange between buried and bulk water molecules (JACS 131, 18214 (2009) – paper here). In effect, the internal water molecules serve as a probe of protein motions with a relaxation timescale comparable to the exchange time (typically 0.1 ns to 10 microseconds).

An intriguing demonstration that the chemistry of hydrated ions can be critically dependent on the geometry of the surrounding water network is provided by Rachael Relph at Yale and colleagues using vibrational spectroscopy of clusters (R. A. Relph et al., Science 327, 308 (2010) – paper here). They find that the extent to which NO+ reacts with water to form HONO varies with different numbers and arrangements of hydration water molecules (1-4). It’s an intriguing demonstration of geometrical effects in hydration, though what it can say in general about hydration in bulk solution is less immediately clear to me. There’s a commentary by Katrin Siefermann and Bernd Abel in the same issue (paper here).

Finally, just for fun, I’ve indulged in a little speculation here about the possibility of quasicrystalline water. There is a lot more behind all this that I was not able to include, following discussions with John Finney and Alan Mackay in particular. The upshot is that Alan gives at least some cause to think it may be possible in theory to construct a plausible H-bonded network with a quasicrystalline geometry. Whether one could make it in practice is, of course, quite another matter.