There are two competing models for ligand entry and exit in globins. In the histidine gate model, a channel from the heme iron to the solvent opens by the outward rotation of the His(E7) side chain (7 th residue of the E-helix). In the multiple paths model, ligands diffuse through the protein matrix and exit at multiple points at its surface. Previous workers solved crystal structures to identify channels in myoglobin (Mb) with n-alkyl isocyanides (CNRs), a long flexible ligand that acts as a molecular "Ariadne's thread." When bound to Mb, a CNR points either toward solvent through an opened His(E7) (out conformation) or into the back of the distal pocket (in conformation). To measure the in/out equilibrium in solution, FTIR spectra have been collected for wild-type and mutant MbCNRs. The fraction of CNRs that point in (Fin) is regulated by the distal pocket volume, the freedom of the His(E7) side chain to rotate outward, and an unfavorable hydrophobic effect for CNRs that point outward into solvent. The relative importance of distal histidine flexibility and pocket volume on diatomic ligand binding with Mb has been assessed by correlating F in for bound CNRs with O2 and NO binding parameters. These correlations indicate strongly that: (1) CNRs and diatomic ligands use the histidine gate; (2) the volume of the binding pocket regulates non-covalent ligand capture and covalent bond formation with the iron atom; (3) the "baseball glove" model of ligand binding to Mb applies for all ligands; and (4) CNRs serve as useful transition state analogs for the diatomic ligand binding reactions of all globins.

The co-chaperonin protein 10 (cpn10) is a heptameric ring-shaped protein present in most organisms. It works in conjunction with the chaperonin protein 60 (cpn60) in an ATP-dependent process to assist folding a range of substrate polypeptides. Cpn10 from the hyper-thermophilic bacterium Aquifex aeolicus (Aacpn10) contains a 25-residue C-terminal extension in each monomer, not found in any other cpn10 protein. Both Aacpn10 and a mutant where the tail has been removed (Aacpn10del-25) adopt heptameric structures with similar thermal and chemical stabilities. In addition, the equilibrium and kinetic unfolding/dissociation and refolding/reassembly reactions are not affected by the presence or absence of the tail. The presence of the tail, however, increases the affinity between the subunits in the heptamer and limits the formation of ordered aggregates of the heptamers at high temperatures and high protein concentrations. Comparative studies on mesostable cpn10 heptamers from human mitochondria (hmcpn10) and Escherichia coli (GroES) reveal that the extreme stability of Aacpn10 originates from increased stability of individual monomers. The stability profile (i.e., the correlation between free energy and temperature) for Aacpn10 is shifted upwards (i.e., higher stability at each temperature) and to the right (i.e., maximum stability at higher temperature) as compared to that of GroES. This is the first thermodynamic analysis of how hyper-thermostability is achieved in an oligomeric protein system.

The elucidation of the molecular mechanisms of platelet adhesion and aggregation in response to naturally occurring subendothelial surfaces under flow conditions is important medically in atherosclerosis and mural thrombosis. Type VI collagen is a subendothelial constituent that binds vWF and platelets. The interaction of platelets with type VI collagen and the roles of platelet glycoprotein receptors and vWF were studied under flow conditions using epi-fluorescent video microscopy coupled with digital image processing. We found that surface coverage was less than 6% on collagen VI at a high wall shear rate (1000s$\sp{-1}$), and approximately 60% at a low wall shear rate (100s$\sp{-1}$). The molecular mechanisms involved in low shear platelet binding were studied using monoclonal antibodies to platelet GPIb and GPIIb-IIIa, and polymeric ATA. Anti-GPIIb-IIIa was the most effective in eliminating adhesion (surface coverage, 0.8%), followed by anti-GPIb (4.3%), and ATA (12.6%). Experiments with vWD blood indicated that vWF is involved in platelet adhesion to collagen VI at 100s$\sp{-1}$. Our results suggest a possible role for collagen VI and vWF in platelet adhesion and aggregation in vascular regions with low shear rates. The endothelial cells lining the vasculature present the ultimate biocompatible surface to flowing blood. Upon blood vessel injury the endothelial cells exhibit an acute response by releasing compounds that mediate the local tissue response. Endothelial derived PGI$\sb2$ and NO act as local modulators of platelet function. To incorporate this aspect of a vascular wound into our model of a damaged blood vessel, monolayers of human umbilical vein endothelial cells were cultured on collagen I coated coverslips and subjected to microinjury prior to exposure to flowing blood. To determine the effect of PGI$\sb2$ on platelet function, monolayers of HUVECs were incubated with aspirin to inhibit PGI$\sb2$ formation. The effect of NO on platelet function was also investigated by using L-NMMA to block NO production. Treatment with aspirin resulted in increased platelet adhesion within the wound site, while treatment with L-NMMA did not. This indicates that endothelial PGI$\sb2$ but not NO may act to reduce platelet reactivity within a wound.

The structures of sperm whale myoglobin (Mb) and human hemoglobin (Hb) complexed with methyl, ethyl, n-propyl and n-butyl isocyanide (MNC, ENC, nPNC and nBNC, respectively) were determined by X-ray crystallography. The polar isocyano head groups of the alkyl isocyanides (RNC's) have similar affinities for heme iron, whereas the size and stereochemistry of the alkyl groups cause the different RNC ligands to cross varied steric barriers when entering or exiting the protein. Four Mb structures were determined at a resolution $$ 100$\sp\circ$ rotation about its C$\sb\alpha$-C$\sb\beta$ bond. Relatively immobile side chains of amino acids in the heme pocket allowed the alkyl groups of MNC and ENC to lie in the hydrophobic interior of the heme pocket. The larger alkyl groups of nPNC and nBNC projected into the hydrophilic exterior entrance to the heme pocket. Data for the structures of Hb complexed with MNC, ENC, nPNC and nBNC (pH 6.7) were collected to $\sim$2.2A. In $\alpha$ subunits, the alkyl moieties of the ligands lay approximately parallel to the heme plane. In $\beta$ subunits, the alkyl groups lay at about a 45$\sp\circ$ angle from the heme plane. The alkyl groups of the ligands occupied the hydrophobic interior of the heme pocket in both subunits. Little disorder in His-E7 was observed in either subunit. Ligand binding causes greater tertiary structure changes in the $\alpha$ and $\beta$ subunits than in myoglobin. These changes in tertiary structure in response to isocyanide binding appear to cause the $\alpha$ and $\beta$ subunits to have more of a T-state tertiary conformation than does the oxy complex, and may explain the lesser cooperative effect seen in isocyanide binding compared to oxygen binding.

Escherichia coli adenylate kinase has been studied by X-ray crystallographic and molecular dynamics approaches. The structures of E. coli adenylate kinase complexed with either AMP and AMPPNP, or AMP and ADP are presented. The AMP/AMPPNP complex has been determined to a resolution of 2.0 A, and confirms the location of the AMP and ATP binding sites on the protein. The AMP/ADP complex has been determined to resolution of 2.8 A, and demonstrates the geometry of this inhibitory complex. Molecular dynamics studies of E. coli adenylate kinase have been done using the weighted masses technique. Simulations were conducted using two general protocols (CHARMM and X-PLOR) which differ in the nature of the secondary structural constraints applied. Both protocols were successful in simulating the opening of the two mobile domains of the protein. The X-PLOR based trajectories, using a heterogeneous weighting system, were also able to simulate the closing of the mobile domains, resulting in a more Gaussian distribution of conformations than that seen in the CHARMM based method.

The effect of pH on the binding affinities of the conjugate bases of four different tetrahedral oxyacids to the periplasmic sulfate-binding protein from Salmonella typhimurium has been determined. In light of the highly refined 2 A structure of the complex of the sulfate-binding protein with sulfate, and considering the protonation state and net charge of the various oxyacids, the pH dependence of chromate binding and the extremely low affinity of phosphate are attributable mainly to a lack of hydrogen bond acceptors in the binding site. These studies demonstrate that the binding site of the sulfate-binding protein is stringently designed to tightly bind tetrahedral, fully ionized, oxyacid dianions. Based on the refined 2.0 A structure of the periplasmic sulfate-binding protein from Salmonella typhimurium, twelve site-directed mutants in the E. coli periplasmic sulfate-binding protein were designed to test specific hypotheses regarding protein-ligand complex stabilization and binding mechanism. Mutants at position 42 (H42N, H42G, H42D) demonstrate that the fidelity of the H-bond array to His 42 is more important in stabilizing the protein-SO$\sb{4}\sp{2-}$ complex than the proposed positive charge. The structural and chemical differences between serine and cysteine were examined by the series of mutants S130G, S130A, S130C. Mutations in the interdomain salt-bridges formed in the closed conformation of the protein were shown to affect the sulfate on- and off-rates. The necessity for protein conformational change in sulfate binding and release was tested by introducing a disulfide between the two domains. These studies further illuminate binding protein specificity, complex stability, and flexibility. The relative affinities for various metals which bind to the calcium-binding site of the E. coli periplasmic $\sc {D}$-galactose-binding protein in solution have been determined. In order of affinity the metals are: Ca$\sp{2+}$ $\approx$ Tb$\sp{3+}$ $\approx$ Pb$\sp{2+}$ $>$ Cd$\sp{2+}$ $>$ Sr$\sp{2+}$ $>$ Mg$\sp{2+}$ $\gg$ Mn$\sp{2+}$ $>$ Ba$\sp{2+}$. The results of these solution studies support the hypothesis that for a given metal-binding loop, the ligands provided by the protein, and the cation hydration energy, size, and charge are major factors contributing to binding affinity.

We demonstrate a technique of measuring x-ray (or neutron) scattering with the momentum transfer parallel to the plane of membrane. This technique allows us to investigate the lateral organization of protein and peptide in the membrane. To resolve the question of whether gramicidin (GA) forms lateral aggregates, samples of GA in dilauroylphosphatidycholine (DLPC) bilayers (molar ratio 1:10) were investigated. Very clear scattering signals of GA were obtained, even for the peptide without a heavy atom attached. The experiment showed that the gramicidin channels did not aggregate and were randomly distributed in the membrane. The non-conducting state of gramicidin channel was also investigated. We use a synthetic GA analogue in which the formyl group of natural GA is replaced by a BOC group. The in-plane scattering measurements show that the gramicidin channel closes by dissociation into two monomers, each remains embedded and freely diffuses in its own monolayer.

Picosecond, nanosecond, and bimolecular reactions of ligands with myoglobins containing distal pocket mutations at positions 29, 45, 64, and 68 were examined at 20$\sp\circ$C, in 0.1 M potassium phosphate pH 7.0. The Val$\sp{68}{\to}$Ile mutation hindered intramolecular iron-ligand bond formation, slowing recombination of NO on picosecond time scales and O$\sb2$ on nanosecond time scales. Picosecond NO recombination was enhanced by increasing the size of residue 29. The rates for the major picosecond rebinding phase were 1.8, 2.5, 29, and $\ge$100 ns$\sp{-1}$ for Ala$\sp{29}$, Val$\sp{29}$, Leu$\sp{29}$(native), and Phe$\sp{29}$ myoglobin. In contrast to this trend, the Leu$\sp{29}{\to}$Phe mutation caused a 10-fold decrease in the rate of nanosecond NO recombination. These effects were interpreted in terms of a model for picosecond and nanosecond ligands based on diffusion of ligands within the protein matrix. On picosecond time scales, the photodissociated ligand appears to reside in a space bordered by His$\sp{64}$ and Leu$\sp{29}$ close to the iron atom. On nanosecond time scales, unbound ligands have diffused farther away into a space adjacent to Ile$\sp{107}$. The Leu$\sp{29}{\to}$Phe mutation caused the distal pocket to become more compartmentalized, enhancing picosecond recombination and hindering nanosecond recombination. The Leu$\sp{29}{\to}$Val and Leu$\sp{29}{\to}$Ala substitutions caused picosecond and nanosecond intermediates to become less distinct. The Val$\sp{68}{\to}$Phe mutation enhanced nanosecond O$\sb2$ recombination by reducing the volume available to the unbound ligand. Thus, distal pocket structure plays a key role in the internal kinetic barriers to ligand recombination. Replacement of His$\sp{64}$ with apolar residues facilitated ligand entry into the protein, due to loss of the distal water molecule in deoxymyoglobin. The Val$\sp{68}{\to}$Thr mutation produced the opposite effect, hindering ligand entry by stabilizing the distal H$\sb2$O. No definite information was gained about pathways of ligand entry and escape, but there appears to be a global protein barrier to ligand exit that causes the rate of ligand escape to be $\sim$1 $\times$ 10$\sp7$s$\sp{-1}$, regardless of mutations in the distal pocket.

The incoherent quasi-elastic neutron scattering (IQNS) method is a useful technique to study biomolecular dynamics. The versatility of the method makes possible motional studies of biomolecules in different forms: powder, crystal, and solution; and at different temperatures. Thus, it allows for the investigation of biomolecular dynamics over a wide-range of physical conditions. We have used the IQNS method to study the motions of side chains in trypsin and myoglobin at various D$\sb2$O hydration levels. The scattering spectra S(Q,$\omega$) were measured in constant-Q mode. The protein in powder form exhibits vibrational high-frequency motions, while the protein in solution and in crystals are characterized by diffusive jumps, and high-frequency vibrations. At temperatures below 200K, the S(Q,$\omega$) for these proteins in solution is similar to an harmonic solid. As temperature increases, a transition is seen at 200K, above which the protein becomes more liquid-like with rapid transitions between conformational substates. The diffusion constant D for the side chains is on the order of 10$\sp{-6}$ cm$\sp2$/sec.

Poor conformational sampling is the most significant problem in determining stable molecular conformations and free energy changes in molecular systems. This problem is particularly difficult for large molecules such as proteins for which the dimensionality of the conformational space can be of thousands of degrees of freedom. In this work we describe a new method, collective coordinate expansive spaces, for efficient conformational sampling of molecules. The new method performs a dimensional reduction of the molecular system using principal components analysis in dihedral space. The reduced dimensionality representation is subsequently used for conformational sampling by means of a randomized exploration technique. The new method was tested for conformational sampling of four different molecular systems and compared to existing sampling methods. The initial conformational dimensionality of the model systems ranged from 8 to 228 degrees of freedom. The new method displayed the best conformational coverage for systems larger than 36 degrees of freedom.