Motivated by the long-term goal of understanding vectorial biological processes such as proton transport (PT) in biomolecular ion pumps, a number of developments were made to establish combined quantum mechanical/molecular mechanical (QM/MM) methods suitable for studying chemical reactions involving significant charge separation in the condensed phase. These developments were summarized and discussed with representative problems. Specifically, free energy perturbation and boundary potential methods for treating long-range electrostatics were implemented to test the robustness of QM/MM results for protein systems. It was shown that consistent models with sufficient sampling were able to produce quantitatively satisfactory results, such as pK a for titritable groups in the interior of T4-lysozyme, while an inconsistent treatment of electrostatics or lack of sufficient sampling may produce incorrect results. Modifications were made to an approximate density functional theory (SCC-DFTB) to improve the description of proton affinity and hydrogen-bonding, which are crucial for the treatment of PT in polar systems. Test calculations on water autoionization showed clearly that both improvements are necessary for quantitatively reliable results. Finally, the newly established SCC-DFTB/MM-GSBP protocol was used to explore mechanistic issues in carbonic anhydrase (CA). Preliminary results suggest that PT in CA occurs mainly through short water wires containing two water molecules in a thermally activated fashion. Although longer water wires occur with similar frequencies, PT along those pathways, on average, has substantially higher barriers, a result not expected based on previous studies. The fluctuations of water molecules peripheral to the water wire were found to make a larger impact on the PT energetics compared to polar protein residues in the active site, which are largely pre-organized and therefore have less tendency to reorganize during the reaction.

The positions of protons are not available in most high-resolution structural data of biomolecules, thus the identity of proton storage sites in biomolecules that transport proton is generally difficult to determine unambiguously. Using combined quantum mechanical/ molecular mechanical computations, we demonstrate that a pair of conserved glutamate residues (Glu 194/204) bonded by a delocalized proton is the proton release group that has been long sought in the proton pump, bacteriorhodopsin. This model is consistent with all available experimental structural and infrared data for both the wild-type bacteriorhodopsin and several mutants. In particular, the continuum infrared band in the 1,800-to 2,000-cm ؊1 region is shown to arise due to the partially delocalized nature of the proton between the glutamates in the wild-type bacteriorhodopsin; alternations in the flexibility of the glutamates and electrostatic nature of nearby residues in various mutants modulate the degree of proton delocalization and therefore intensity of the continuum band. The strong hydrogen bond between Glu 194/204 also significantly shifts the carboxylate stretches of these residues well <1,700 cm ؊1 , which explains why carboxylate spectral shift was not observed experimentally in the typical >1,700-cm ؊1 region upon proton release. By contrast, simulations with the proton restrained on the nearby water cluster, as proposed by several recent studies [see, for example, Garezarek K, Gerwert K (2006) Functional waters in intraprotein proton transfer monitored by FTIR difference spectroscopy. Nature 439:109], led to significant structural deviations from available X-ray structures. This study establishes a biological function for strong, low-barrier hydrogen bonds.infrared spectroscopy ͉ proton pumping ͉ water cluster ͉ quantum mechanical/molecular mechanical simulations

Motivated by the long-term goal of theoretically analyzing long-range proton transfer (PT) kinetics in biomolecular pumps, researchers made a number of technical developments in the framework of quantum mechanics-molecular mechanics (QM/MM) simulations. A set of collective reaction coordinates is proposed for characterizing the progress of long-range proton transfers; unlike previous suggestions, the new coordinates can describe PT along highly nonlinear three-dimensional pathways. Calculations using a realistic model of carbonic anhydrase demonstrated that adiabatic mapping using these collective coordinates gives reliable energetics and critical geometrical parameters as compared to minimum energy path calculations, which suggests that the new coordinates can be effectively used as reaction coordinate in potential of mean force calculations for long-range PT in complex systems. In addition, the generalized solvent boundary potential was implemented in the QM/MM framework for rectangular geometries, which is useful for studying reactions in membrane systems. The resulting protocol was found to produce water structure in the interior of aquaporin consistent with previous studies including a much larger number of explicit solvent and lipid molecules. The effect of electrostatics for PT through a membrane protein was also illustrated with a simple model channel embedded in different dielectric continuum environments. The encouraging results observed so far suggest that robust theoretical analysis of long-range PT kinetics in biomolecular pumps can soon be realized in a QM/MM framework.

Identifying the group that acts as the proton storage/loading site is a challenging but important problem for understanding the mechanism of proton pumping in biomolecular proton pumps, such as bacteriorhodopsin (bR) and cytochrome c oxidase. Recent experimental studies of bR propelled the idea that the proton storage/release group (PRG) in bR is not an amino acid but a water cluster embedded in the protein. We argue that this idea is at odds with our knowledge of protein electrostatics, since invoking the water cluster as PRG would require the protein to raise the pKa of a hydronium by almost 11 pKa units, which is difficult considering known cases of pKa shifts in proteins. Our recent QM/MM simulations suggested an alternative “intermolecular proton bond” model in which the stored proton is shared between two conserved Glu residues (194 and 204). Here we show that this model leads to microscopic pKa values consistent with available experimental data and the functional requirement of a PRG. Extensive QM/MM simulations also show that, independent of a number of technical issues, such as the influence of QM region size, starting x-ray structure and nuclear quantum effects, the “intermolecular proton bond” model is qualitatively consistent with available spectroscopic data. Potential of mean force calculations show explicitly that the stored proton strongly prefers the pair of Glu residues over the water cluster. The results and analyses help highlight the importance of considering protein electrostatics and provide arguments for why the “intermolecular proton bond” model is likely applicable to PRG in biomolecular proton pumps in general.

As stringent tests for the molecular model and computational protocol, microscopic pK(a) calculations are performed for the key residue, Glu286, in cytochrome c oxidase (CcO) using a combined quantum mechanical/molecular mechanical (QM/MM) potential and a thermodynamic integration protocol. The impact of the number of water molecules in the hydrophobic cavity and protonation state of several key residues (e.g., His334, Cu(B)-bound water, and PRD(a3)) on the computed microscopic pK(a) values of Glu286 has been systematically examined. To help evaluate the systematic errors in the QM/MM-based protocol, microscopic pK(a) calculations have also been carried out for sites in a soluble protein (Asp70 in T4 lysozyme) and a better-characterized membrane protein (Asp85 in bacteriorhodopsin). Overall, the results show a significant degree of internal consistency and reproducibility that support the effectiveness of the computational framework. Although the number of water molecules in the hydrophobic cavity does not greatly influence the computed pK(a) of Glu286, the protonation states of several residues, some of which are rather far away, have more significant impacts. Adopting the standard protonation state for all titratable residues leaves a large net charge on the system and a significantly elevated pK(a) for Glu286, highlighting that any attempt to address the energetics of proton transfers in CcO at a microscopic level should carefully select the protonation state of residues, even those not in the immediate neighborhood of the active site. The calculations indirectly argue against the deprotonation of His334 for the proton pumping process, although further studies that explicitly compute its pK(a) are required for a more conclusive statement. Finally, the deprotonated Glu286 is found to be in a stable water-mediated connection with PRD(a3) for at least several nanoseconds when this presumed pumping site is protonated. This does not support the proposed role of Glu286 as a robust gating valve that prevents proton leakage, although a conclusive statement awaits a more elaborate characterization of the Glu286-PRD(a3) connectivity with free energy simulations and a protonated PRD(a3). The large sets of microscopic simulations performed here have provided useful guidance to the establishment of a meaningful molecular model and effective computational protocol for explicitly analyzing the proton transfer kinetics in CcO, which is required for answering key questions regarding the pumping function of this fascinating and complex system.

A combined quantum mechanical/molecular mechanical (QM/MM) potential function is used in a thermodynamic integration approach to calculate the pK a of residue 66 in two mutants (V66E, V66D) of Staphylococal nuclease relative to solution. Despite the similarity in chemical nature and experimentally measured pK a of the two buried titritable residues, the behaviors of the two mutants and the computed pK a values vary greatly in the simulations. For Glu66, the sidechain is consistently observed to spontaneously flip out from the protein interior during titration, and the overall protein structure remains stable throughout the simulations. The computed pK a shifts using conventional sampling techniques with multiple nanoseconds per λ window (Set A & B) are generally close to the experimental value, therefore indicating that large-scale conformational rearrangements are not as important for V66E as suggested by the recent study of Warshel and co-worker. For Asp66, by contrast, flipping of the shorter sidechain is not sufficient for getting adequate solvent stabilization of the ionized state. As a result, more complex behaviors such as partial unfolding of a nearby β-sheet region is observed, and the computed pK a shift is substantially higher than the experimental value unless Asp66 is biased to adopt the similar configurations as Glu66 in the V66E simulations. Collectively, these studies suggest that the lack of electronic polarization is not expected to be the dominant source of error in microscopic pK a shift calculations, while the need of enhanced sampling is more compelling for predicting the pK a of buried residues. Furthermore, the comparison between V66E and V66D also highlights that the microscopic interpretation of similar apparent pK a values and effective "dielectric constants" of proteins can vary greatly in terms of the residues that make key contributions and the scale of structural/hydration response to titration, the latter of which is difficult to predict a priori. Perturbative analyses of interactions that contribute to the titration free energy point to mutants that can be used to verify the microscopic mechanisms of titration in V66E/ D SNase proteins.

We demonstrate that membrane proteins and phospholipids can self-assemble into polyhedral arrangements suitable for structural analysis. Using the Escherichia coli mechanosensitive channel of small conductance (MscS) as a model protein, we prepared membrane protein polyhedral nanoparticles (MPPNs) with uniform radii of ∼20 nm. Electron cryotomographic analysis established that these MPPNs contain 24 MscS heptamers related by octahedral symmetry. Subsequent single-particle electron cryomicroscopy yielded a reconstruction at ∼1-nm resolution, revealing a conformation closely resembling the nonconducting state. The generality of this approach has been addressed by the successful preparation of MPPNs for two unrelated proteins, the mechanosensitive channel of large conductance and the connexon Cx26, using a recently devised microfluidics-based free interface diffusion system. MPPNs provide not only a starting point for the structural analysis of membrane proteins in a phospholipid environment, but their closed surfaces should facilitate studies in the presence of physiological transmembrane gradients, in addition to potential applications as drug delivery carriers or as templates for inorganic nanoparticle formation.membrane protein structure | proteoliposomes | electron microscopy | microfluidic devices T he functions of many membrane proteins are intimately coupled to the generation, utilization, and/or sensing of transmembrane gradients (1). Despite advances in the structure determination of membrane proteins (2), the high-resolution structural analysis of membrane proteins in a biological membrane is uncommon and in the presence of a functionally relevant gradient remains an asyet unrealized experimental challenge. This stems from the fact that the primary 2D-and 3D ordered specimens used in structural studies of membrane proteins by X-ray crystallography and electron microscopy lack closed membrane surfaces, thus making it impossible to establish physiologically relevant transmembrane gradients.As an alternative, we have been developing methodologies for the self-assembly of lipids and membrane proteins into closed polyhedral structures that can potentially support transmembrane gradients for structural and functional studies. The possibility of generating polyhedral arrangements of membrane proteins in proteoliposomes was motivated by the existence of polyhedral capsids of membrane-enveloped viruses (3, 4), the ability of surfactant mixtures to self-assemble into polyhedral structures (5, 6), and the formation of proteoliposomes from native membranes containing bacteriorhodopsin (7, 8) and light-harvesting complex II (LHCII) (9). Significantly, the high-resolution structure of LHCII was determined from crystals of icosahedral proteoliposomes composed of protein subunits in chloroplast lipids (10). Whereas detergent solubilized membrane proteins and lipid mixtures can self-assemble to form 2D-ordered crystalline sheets or helical tubes favorable for structure determination by electron microscopy (11-14), simple...

The generation of highly curved membranes is essential to cell growth, division and movement. Recent research in the field is focused to answer questions related to the consequences of changes in the topology of the membrane once it is created, broadly termed as membrane curvature sensing. Most probes that are used to study curvature sensing are intact membrane active proteins like DP1/Yop1p, ArfGAP1, BAR domains, and Synaptotagmin-I (Syt1). Taking a cue from nature, we created the cyclic peptide C2BL3C based on the membrane penetration C2B loop 3 of Syt1 via ‘Click’ chemistry. Using a combination of spectroscopic techniques, we investigated the peptide-lipid interactions of this peptide with synthetic phospholipid vesicles and exosomes from rat blood plasma. We found that the macrocycle peptide probe was selective for lipid vesicles with highly curved surfaces (d <100 nm). These results suggested that C2BL3C functions as a selective detector of highly curved phospholipid bilayers.