Obtaining an electron-density map from X-ray diffraction data can be dif®cult and time-consuming even after the data have been collected, largely because MIR and MAD structure determinations currently require many subjective evaluations of the qualities of trial heavy-atom partial structures before a correct heavy-atom solution is obtained. A set of criteria for evaluating the quality of heavy-atom partial solutions in macromolecular crystallography have been developed. These have allowed the conversion of the crystal structure-solution process into an optimization problem and have allowed its automation. The SOLVE software has been used to solve MAD data sets with as many as 52 selenium sites in the asymmetric unit. The automated structure-solution process developed is a major step towards the fully automated structure-determination, model-building and re®nement procedure which is needed for genomic scale structure determinations.

binary enzyme-substrate (2; throughout the text, the bold numbers refer to species in Fig. The Catalytic Pathway of 1) and enzyme-product (4) complexes have Cytochrome P450cam a t been so characterized (4). Some features of the dioxygen-bound or activated oxygen intermediates, in particular the geometry of the Atomic Resolution six-coordinate low-spin heme, have been deduced from the structure of the ferrous llme ~chlichting,'* Joel Berendzen,' Kelvin Chu,'? Ann M. S t~c k ,~ (FelI+) carbonmonoxy complex (3) of Shelley A. M a v e~,~ David E. enso on,^ Robert M. S~e e t ,~ P450cam (5).However, the binding of carbon Dagmar Ringe,6 Gregory A. ~e t s k o ,~ monoxide to heme is likely to be different in Stephen G. Sligar'~~ a number of important ways from the binding Members o f t h e cytochrome P450 superfamily catalyze t h e addition o f m o -of oxygen ( 6 ) , and regardless, carbon monlecular oxygen t o nonactivated hydrocarbons a t physiological temperature-a oxide is an inhibitor, not a substrate, of reaction t h a t requires high temperature t o proceed i n t h e absence o f a catalyst. P450cam. Hence, the primary evidence for Structures were obtained for three intermediates i n the hydroxylation reaction the structures of the ferrous enzyme-substrate o f camphor b y P450cam w i t h trapping techniques and cryocrystallography. The complex (5), the dioxy intermediate (6), and structure o f t h e ferrous dioxygen adduct o f P450cam was determined w i t h 0.91

beta process ͉ dielectric ͉ hydration ͉ solvent P roteins are dynamic systems that interact strongly with their environment (1). Most texts and publications show proteins in unique conformations and naked, without hydration shell and bulk solvent, while fluctuations are rarely mentioned. The unified model of protein dynamics presented here is a radical departure from this picture. In this model, the protein provides the structure for the biological function, but it is dynamically passive. The fluctuations in the bulk solvent power and control the large-scale motions and shape changes of the protein in a diffusive manner (2-4), whereas the fluctuations in the hydration shell power and control the internal protein motions such as ligand migration (5, 6). The hydration shell consists of Ϸ2 layers of water that surround proteins as shown in Fig. 1 (7-12). Protein functions depend on the degree of hydration, h, defined as the weight ratio of water to protein. Dehydrated proteins do not function. Some proteins begin to work at h Ϸ 0.2 (11) but full function may require h Ͼ 1. The controls exerted by the bulk solvent and the hydration shell are possible because the protein interior is fluid-like (13); the intrinsic viscosity of a protein is small, about like water (14-16). The image of the protein being essentially passive and being slaved to the environment is not an idle speculation. It is based on experiments using myoglobin (Mb) that led to the seminal concepts that underlie the present work: (i) Proteins do not exist in a unique conformation; they can assume a very large number of conformational substates (CS) (17, 18). (ii) The CS can be described by an energy landscape (17). (iii) The landscape is organized in a hierarchy; there are energy valleys within energy valleys within energy valleys (19). The description of the effects of the bulk solvent and the hydration shell is based on these concepts. Because knowledge of the fluctuations in glass-forming liquids and of the energy landscape of proteins is essential for understanding these results, we discuss these topics first.The ␣ and ␤ Processes (20, 21) Glass-forming liquids have two types of equilibrium fluctuations, ␣ and ␤.* One tool to study these fluctuations is dielectric relaxation spectroscopy (21). The sample is placed in a capacitor, a sine-wave voltage U 1 ( ) of frequency is applied, and the resulting current is converted into a voltage U 2 ( ) that characterizes the dielectric spectrum. Our spectra exhibit two prominent peaks that characterize ␣, or primary, and ␤, or secondary, relaxations. The ␣ process describes structural fluctuations. The mechanical Maxwell relation,connects the rate coefficient k ␣ (T) for the ␣ fluctuations to the viscosity (T). Here, G 0 is the infinite-frequency shear modulus that depends only weakly on temperature and on the material.Author contributions: H.F., G.C., J.B., P.W.F., H.J., B.H.M., I.R.S., J.S., and R.D.Y. designed research, performed research, contributed new reagents/analytic tools, analyzed data, and wrote ...

Formation of the chromophore of green fluorescent protein (GFP) depends on the correct folding of the protein. We constructed a "folding reporter" vector, in which a test protein is expressed as an N-terminal fusion with GFP. Using a test panel of 20 proteins, we demonstrated that the fluorescence of Escherichia coli cells expressing such GFP fusions is related to the productive folding of the upstream protein domains expressed alone. We used this fluorescent indicator of protein folding to evolve proteins that are normally prone to aggregation during expression in E. coli into closely related proteins that fold robustly and are fully soluble and functional. This approach to improving protein folding does not require functional assays for the protein of interest and provides a simple route to improving protein folding and expression by directed evolution.

After photodissociation of carbon monoxide bound to myoglobin, the protein relaxes to the deoxy equilibrium structure in a quake-like motion. Investigation of the proteinquake and of related intramolecular equilibrium motions shows that states and motions have a hierarchical glasslike structure.The dynamic aspects of proteins have been studied extensively in recent years and a picture of ever increasing complexity has emerged. To bring some order into the complexity, we have introduced a model that classifies states and motions (1). In the present paper, we describe the model and its experimental basis in more detail. STATES, SUBSTATES, AND MOTIONSWe consider myoglobin (Mb), an oxygen storage protein, consisting of 153 amino acids, with molecular weight of 17,900 and approximate dimensions of2.5 x 4.4 x 4.4 nm (2). Embedded in the protein matrix is a heme group with a central iron atom, which binds small ligands such as dioxygen (02) or carbon monoxide (CO) reversibly. Thus, two states are involved in the function of Mb, deoxyMb and liganded Mb (e.g., MbCO). In the liganded state, the heme is planar and the iron has spin 0 and lies close to the mean heme plane. In the unliganded state, the heme group is domed, the iron has spin 2 and lies =0.5 A away from the mean heme plane, and the globin structure differs somewhat from the liganded one (3).A protein molecule in a particular state can assume a very large number of conformational substates (CS) (4-6). Different substates have the same overall structure, but they differ in details; they perform the same function, but with different rates. equilibrium (10). Return to equilibrium occurs through a proteinquake: the released strain energy is dissipated through waves [phonons (11) or solitons (12)1 and through the propagation of a deformation (2, 3). HIERARCHY OF SUBSTATESThe experiments described in the next section imply that the proteinquake released by photodissociation of MbCO propagates sequentially: Fig. 2, consequently is much more complex than we originally anticipated (4).The valley in the top diagram of Fig. 2a represents one state, say MbCO. MbCO can exist in a large number of conformational substates, CS1, separated by high barriers.Each valley in the first tier is structured into substates (CS2) with smaller barriers. The furcation continues through two more tiers, with decreasing barrier heights. The dynamic

We have used x-ray crystallography to determine the structures of sperm whale myoglobin (Mb) in four different ligation states (unligated, ferric aquomet, oxygenated, and carbonmonoxygenated) to a resolution of better than 1.2 A. Data collection and analysis were performed in as much the same way as possible to reduce model bias in differences between structures. The structural differences among the ligation states are much smaller than previously estimated, with differences of <0.25 A root-mean-square deviation among all atoms. One structural parameter previously thought to vary among the ligation states, the proximal histidine (His-93) azimuthal angle, is nearly identical in all the ferrous complexes, although the tilt of the proximal histidine is different in the unligated form. There are significant differences, however, in the heme geometry, in the position of the heme in the pocket, and in the distal histidine (His-64) conformations. In the CO complex the majority conformation of ligand is at an angle of 18 +/- 3 degrees with respect to the heme plane, with a geometry similar to that seen in encumbered model compounds; this angle is significantly smaller than reported previously by crystallographic studies on monoclinic Mb crystals, but still significantly larger than observed by photoselection. The distal histidine in unligated Mb and in the dioxygenated complex is best described as having two conformations. Two similar conformations are observed in MbCO, in addition to another conformation that has been seen previously in low-pH structures where His-64 is doubly protonated. We suggest that these conformations of the distal histidine correspond to the different conformational substates of MbCO and MbO(2) seen in vibrational spectra. Full-matrix refinement provides uncertainty estimates of important structural parameters. Anisotropic refinement yields information about correlated disorder of atoms; we find that the proximal (F) helix and heme move approximately as rigid bodies, but that the distal (E) helix does not.

Ligand binding to heme proteins is studied by using flash photolysis over wide ranges in time (100 ns-1 ks) and temperature (10-320 K). Below about 200 K in 75% glycerol/water solvent, ligand rebinding occurs from the heme pocket and is nonexponential in time. The kinetics is explained by a distribution, g(H), of the enthalpic barrier of height H between the pocket and the bound state. Above 170 K rebinding slows markedly. Previously we interpreted the slowing as a "matrix process" resulting from the ligand entering the protein matrix before rebinding. Experiments on band III, an inhomogeneously broadened charge-transfer band near 760 nm (approximately 13,000 cm-1) in the photolyzed state (Mb*) of (carbonmonoxy)myoglobin (MbCO), force us to reinterpret the data. Kinetic hole-burning measurements on band III in Mb* establish a relation between the position of a homogeneous component of band III and the barrier H. Since band III is red-shifted by 116 cm-1 in Mb* compared with Mb, the relation implies that the barrier in relaxed Mb is 12 kJ/mol higher than in Mb*. The slowing of the rebinding kinetics above 170 K hence is caused by the relaxation Mb*----Mb, as suggested by Agmon and Hopfield [(1983) J. Chem. Phys. 79, 2042-2053]. This conclusion is supported by a fit to the rebinding data between 160 and 290 K which indicates that the entire distribution g(H) shifts. Above about 200 K, equilibrium fluctuations among conformational substates open pathways for the ligands through the protein matrix and also narrow the rate distribution. The protein relaxations and fluctuations are nonexponential in time and non-Arrhenius in temperature, suggesting a collective nature for these protein motions. The relaxation Mb*----Mb is essentially independent of the solvent viscosity, implying that this motion involves internal parts of the protein. The protein fluctuations responsible for the opening of the pathways, however, depend strongly on the solvent viscosity, suggesting that a large part of the protein participates. While the detailed studies concern MbCO, similar data have been obtained for MbO2 and CO binding to the beta chains of human hemoglobin and hemoglobin Zürich. The results show that protein dynamics is essential for protein function and that the association coefficient for binding from the solvent at physiological temperatures in all these heme proteins is governed by the barrier at the heme.

No abstract