Multivalent endosome targeting by homodimeric EEA1.

Early endosome autoantigen localization to early endosomes is mediated by a C-terminal region, which includes a calmodulin binding motif, a Rab5 interaction site, and a FYVE domain that selectively binds phosphatidyl inositol 3-phosphate. The crystal structure of the C-terminal region bound to inositol 1,3-bisphosphate reveals an organized, quaternary assembly consisting of a parallel coiled coil and a dyad-symmetric FYVE domain homodimer. Structural and biochemical observations support a multivalent mechanism for endosomal localization in which domain organization, dimerization, and quaternary structure amplify the weak affinity and modest specificity of head group interactions with conserved residues. A unique mode of membrane engagement deduced from the quaternary structure of the C-terminal region provides insight into the structural basis of endosome tethering.

Determinants of the broad recognition of exocytic Rab GTPases by Mss4.

Rab GTPases function as essential regulators of vesicle transport between subcellular compartments of eukaryotic cells. Mss4, an evolutionarily conserved Rab accessory factor, facilitates nucleotide release and binds tightly to the nucleotide-free form of exocytic but not endocytic Rab GTPases. A structure-based mutational analysis of residues that are conserved only in exocytic Rab GTPases reveals three residues that are critical determinants of the broad specificity recognition of exocytic Rab GTPases by Mss4. One of these residues is located at the N-terminus of the switch I region near the nucleotide binding site whereas the other two flank an exposed hydrophobic triad previously implicated in effector recognition. The spatial disposition of these residues with respect to the structure of Rab3A correlates with the dimensions of the elongated Rab interaction epitope in Mss4 and supports a mode of interaction similar to that of other exchange factor-GTPase complexes. The complementarity of the corresponding interaction surfaces suggests a hypothetical structural model for the complex between Mss4 and Rab GTPases.

Structural plasticity of an invariant hydrophobic triad in the switch regions of Rab GTPases is a determinant of effector recognition.

Rab GTPases function as regulatory components of an evolutionarily conserved machinery that mediates docking, priming, and fusion of vesicles with intracellular membranes. We have previously shown that the active conformation of Rab3A is stabilized by a substantial hydrophobic interface between the putative conformational switch regions (Dumas, J. J., Zhu, Z., Connolly, J. L., and Lambright, D. G. (1999) Structure 7, 413-423). A triad of invariant hydrophobic residues at this switch interface (Phe-59, Trp-76, and Tyr-91) represents a major interaction determinant between the switch regions of Rab3A and the Rab3A-specific effector Rabphilin3A (Ostermeier, C., and Brunger, A. T. (1999) Cell 96, 363-374). Here, we report the crystal structure of the active form of Rab5C, a prototypical endocytic Rab GTPase. As is true for Rab3A, the active conformation of Rab5C is stabilized by a hydrophobic interface between the switch regions. However, the conformation of the invariant hydrophobic triad (residues Phe-58, Trp-75, and Tyr-90 in Rab5C) is dramatically altered such that the resulting surface is noncomplementary to the switch interaction epitope of Rabphilin3A. This structural rearrangement reflects a set of nonconservative substitutions in the hydrophobic core between the central beta sheet and the alpha2 helix. These observations demonstrate that structural plasticity involving an invariant hydrophobic triad at the switch interface contributes to the mechanism by which effectors recognize distinct Rab subfamilies. Thus, the active conformation of the switch regions conveys information about the identity of a particular Rab GTPase as well as the state of the bound nucleotide.

A helical turn motif in Mss4 is a critical determinant of Rab binding and nucleotide release.

Monomeric Rab GTPases function as ubiquitous regulators of intracellular membrane trafficking. Mss4, an evolutionarily conserved Rab accessory factor, promotes nucleotide release from exocytic but not endocytic Rab GTPases. Here we describe the results of a high-resolution crystallographic and mutational analysis of Mss4. The 1.65 A crystal structure of Mss4 reveals a network of direct and water-mediated interactions that stabilize a partially exposed structural subdomain derived from four highly conserved but nonconsecutive sequence elements. The conserved subdomain contains the invariant cysteine residues required for Zn2+ binding as well as the residues implicated in the interaction with Rab GTPases. A strictly conserved DPhiPhi motif, consisting of an invariant aspartic acid residue (Asp 73) followed by two bulky hydrophobic residues (Met 74 and Phe 75), encodes a prominently exposed 3(10) helical turn in which the backbone is well-ordered but the side chains of the conserved residues are highly exposed and do not engage in intramolecular interactions. Substitution of any of these residues with alanine dramatically impairs nucleotide release activity toward Rab3A, indicating that the DPhiPhi motif is a critical element of the Rab interaction epitope. In particular, mutation of Phe 75 results in a defect as severe as that observed for mutation of Asp 96, which is located near the zinc binding site at the opposite end of the conserved subdomain. Despite severe defects, however, none of the mutant proteins is catalytically dead. Taken together, the results suggest a concerted mechanism in which distal elements of the conserved Rab interaction epitope cooperatively facilitate nucleotide release.

Structural basis of 3-phosphoinositide recognition by pleckstrin homology domains.

Lipid second messengers generated by phosphoinositide (PI) 3-kinases regulate diverse cellular functions through interaction with pleckstrin homology (PH) domains in modular signaling proteins. The PH domain of Grp1, a PI 3-kinase-activated exchange factor for Arf GTPases, selectively binds phosphatidylinositol 3,4,5-trisphosphate with high affinity. We have determined the structure of the Grp1 PH domain in the unliganded form and bound to inositol 1,3,4,5-tetraphosphate. A novel mode of phosphoinositide recognition involving a 20-residue insertion within the beta6/beta7 loop explains the unusually high specificity of the Grp1 PH domain and the promiscuous 3-phosphoinositide binding typical of several PH domains including that of protein kinase B. When compared to other PH domains, general determinants of 3-phosphoinositide recognition and specificity can be deduced.

Structural basis of activation and GTP hydrolysis in Rab proteins.

BACKGROUND: Rab proteins comprise a large family of GTPases that regulate vesicle trafficking. Despite conservation of critical residues involved in nucleotide binding and hydrolysis, Rab proteins exhibit low sequence identity with other GTPases, and the structural basis for Rab function remains poorly characterized. RESULTS: The 2. 0 A crystal structure of GppNHp-bound Rab3A reveals the structural determinants that stabilize the active conformation and regulate GTPase activity. The active conformation is stabilized by extensive hydrophobic contacts between the switch I and switch II regions. Serine residues in the phosphate-binding loop (P loop) and switch I region mediate unexpected interactions with the gamma phosphate of GTP that have not been observed in previous GTPase structures. Residues implicated in the interaction with effectors and regulatory factors map to a common face of the protein. The electrostatic potential at the surface of Rab3A indicates a non-uniform distribution of charged and nonpolar residues. CONCLUSIONS: The major structural determinants of the active conformation involve residues that are conserved throughout the Rab family, indicating a common mode of activation. Novel interactions with the gamma phosphate impose stereochemical constraints on the mechanism of GTP hydrolysis and provide a structural explanation for the large variation of GTPase activity within the Rab family. An asymmetric distribution of charged and nonpolar residues suggests a plausible orientation with respect to vesicle membranes, positioning predominantly hydrophobic surfaces for interaction with membrane-associated effectors and regulatory factors. Thus, the structure of Rab3A establishes a framework for understanding the molecular mechanisms underlying the function of Rab GTPases.

Gs alpha meets its target--shedding light on a key signal transduction event.

The recently determined crystal structure of Gs alpha bound to a catalytically active form of adenylyl cyclase reveals the location of the enzyme's active site and provides the first view of heterotrimeric G protein alpha subunit activating a downstream effector. Comparison with the structure of a catalytically inactive form of adenylyl cyclase suggests a plausible allosteric mechanism whereby the synergistic activators Gs alpha and forskolin stimulate the activity of adenylyl cyclase.

The 2.0 A crystal structure of a heterotrimeric G protein.

The structure of a heterotrimeric G protein reveals the mechanism of the nucleotide-dependent engagement of the alpha and beta gamma subunits that regulates their interaction with receptor and effector molecules. The interaction involves two distinct interfaces and dramatically alters the conformation of the alpha but not of the beta gamma subunits. The location of the known sites for post-translational modification and receptor coupling suggest a plausible orientation with respect to the membrane surface and an activated heptahelical receptor.

Crystal structure of a G-protein beta gamma dimer at 2.1A resolution.

Many signalling cascades use seven-helical transmembrane receptors coupled to heterotrimeric G proteins (G alpha beta gamma) to convert extracellular signals into intracellular responses. Upon nucleotide exchange catalysed by activated receptors, heterotrimers dissociate into GTP-bound G alpha subunits and G beta gamma dimers, either of which can modulate many downstream effectors. Here we use multiwavelength anomalous diffraction data to solve the crystal structure of the beta gamma dimer of the G protein transducin. The beta-subunit is primarily a seven-bladed beta-propeller that is partially encircled by an extended gamma-subunit. The beta-propeller, which contains seven structurally similar WD repeats, defines the stereochemistry of the WD repeat and the probable architecture of all WD-repeat-containing domains. The structure details interactions between G protein beta- and gamma-subunits and highlights regions implicated in effector modulation for the conserved family of G protein beta gamma dimers.

GTPase mechanism of Gproteins from the 1.7-A crystal structure of transducin alpha-GDP-AIF-4.

Aluminium fluoride (AIF-4) activates members of the heterotrimeric G-protein (G alpha beta gamma) family by binding to inactive G alpha.GDP near the site occupied by the gamma-phosphate in G alpha.GTP (ref. 3). Here we describe the crystal structure of transducin alpha.GDP activated with aluminium fluoride (Gt alpha.GDP.AIF-4.H2O) at 1.7 A, a resolution sufficient to establish the coordination geometry of the bound aluminium fluoride as well as the extensive network of direct and water-mediated interactions that stabilize it. These observations are derived from three independent representations in the asymmetric unit, eliminating any chance of drawing conclusions based on stereochemistry imposed by crystal packing. Surprisingly, aluminium fluoride activates G alpha.GDP by binding with a geometry resembling a pentavalent intermediate for GTP hydrolysis. The stabilizing interactions involve not only residues that interact with the gamma-phosphate in Gt alpha.GTP gamma S, but also conserved residues for GTPase activity. Thus the Gt alpha.GDP.AIF-4.H2O structure provides new insight into the mechanism of GTP hydrolysis.

Determination of the carbon monoxide binding constants of myoglobin mutants: comparison of kinetic and equilibrium methods.

The carbon monoxide (CO) binding constants of human myoglobin (Mb) and several single-site mutants have been determined using two different methods. In the kinetic method, which is commonly used for this ligand, the overall association (k(on)) and dissociation (k(off)) rates of CO were measured by flash photolysis and NO replacement, respectively, and the ratio k(on)/k(off) was calculated. In the equilibrium method, the binding constant Keq was measured directly using a thin-layer technique. These two measurements yield similar results for human wild-type Mb but differ significantly for some of the mutants. Possible reasons for these discrepancies are analyzed. A model assuming the presence of interconverting conformers with different association and dissociation rates is considered in light of infrared measurements on the CO stretching frequency in the MbCO forms of the same proteins [Balasubramanian et al. (1993a) Proc. Natl. Acad. Sci, U.S.A. 90, 4718]. It is suggested that in the case of some mutants which exhibit multiple conformations, this model may lead to nonequilibrium kinetics, which could produce the observed discrepancies between the kinetic and equilibrium determinations of the binding constant. These results suggest that both equilibrium and kinetic data should be obtained, even for a monomeric protein such as Mb, before the relative stabilities of mutants can be meaningfully compared.

Structural determinants for activation of the alpha-subunit of a heterotrimeric G protein.

The 1.8 A crystal structure of transducin alpha.GDP, when compared to that of the activated complex with GTP-gamma S, reveals the nature of the conformational changes that occur on activation of a heterotrimeric G-protein alpha-subunit. Structural changes initiated by direct contacts with the terminal phosphate of GTP propagate to regions that have been implicated in effector activation. The changes are distinct from those observed in other members of the GTPase superfamily.

Ultrafast measurements of geminate recombination of NO with site-specific mutants of human myoglobin.

Flash photolysis studies of NO recombination to heme proteins offer a direct probe of protein structural changes on the tens of picoseconds timescale where they can be compared with molecular dynamics simulations. The geminate recombination of NO to site-specific mutants of human myoglobin (Mb) was studied following photodissociation of the MbNO form. Single amino acid changes were introduced at positions Val68, His64, Lys45 and Asp60 because motions of residues at these positions are generally regarded as important for the mechanism of ligand binding. In sharp contrast to the properties of simple porphyrin-NO complexes, the rebinding kinetics are found to be non-exponential for all mutants, even in aqueous solution at 298 K. The Val68 and His64 mutants substantially affect the NO rebinding rates but, surprisingly, so do changes on the protein surface that are further away from the iron. These changes in kinetics occur on a tens of picoseconds timescale, and therefore there is either a fast communication between protein residues over quite long distances or there are subtle differences in protein structure that exert great control over the reaction dynamics. Various models for the rebinding kinetics are evaluated. A model-free approach to data analysis using the maximum entropy method is found to be most useful. This analysis shows that the rate distributions are very different for the mutants, but are generally bimodal.

Anatomy and dynamics of a ligand-binding pathway in myoglobin: the roles of residues 45, 60, 64, and 68.

In order for diatomic ligands to enter and exit myoglobin, there must be substantial displacements of amino acid side chains from their positions in the static X-ray structure. One pathway, involving Arg/Lys45, His64, and Val68, has been studied in greatest detail. In an earlier study (Lambright et al., 1989) we reported the surprising result that mutation of the surface residue Lys45 to arginine lowers the inner barrier to CO rebinding. Until then, it had been thought that this barrier primarily involves interior distal pocket residues such as His64 and Val68. In this report, we present a detailed study of the CO rebinding kinetics in aqueous solution of a series of single- and double-site mutants of human myoglobin at positions 64, 68, 45, and 60. On the basis of the observed kinetics, we propose that the effect of surface residue 45 on the inner barrier can be explained by a chain of interactions between surface and pocket residues. Very large, and in some cases unexpected, changes are observed in the kinetics of recombination and in the partitioning between geminate and bimolecular recombination.

Dynamics of protein relaxation in site-specific mutants of human myoglobin.

We have recently reported spectroscopic evidence for structural relaxation of myoglobin (Mb) following photodissociation of MbCO [Lambright, D. G., Balasubramanian, S., & Boxer, S. G. (1991) Chem. Phys. 158, 249-260]. In this paper we report measurements for a series of single amino acid mutants of human myoglobin on the distal side of the heme pocket (positions 45, 64, and 68) in order to examine specific structural determinants involved in this conformational relaxation and to determine the nature of the coupling between relaxation and the functional process of ligand binding. The kinetics of ligand binding and conformational relaxation were monitored by transient absorption spectroscopy in the Soret spectral region, and the results are analyzed using a four-state ligand binding model. Two principal results emerge: (1) amino acid substitutions in the distal heme pocket affect the kinetics of the nonequilibrium conformational relaxation and (2) the rate of ligand escape from the protein matrix is not significantly perturbed by the distal heme pocket mutations.

Perturbations of the distal heme pocket in human myoglobin mutants probed by infrared spectroscopy of bound CO: correlation with ligand binding kinetics.

The infrared spectra of CO bound to human myoglobin and myoglobin mutants at positions His-64, Val-68, Asp-60, and Lys-45 on the distal side have been measured between 100 and 300 K. Large differences are observed with mutations at His-64 and Val-68 as well as with temperature and pH. Although distal His-64 is found to affect CO bonding, Val-68 also plays a major role. The variations are analyzed qualitatively in terms of a simple model involving steric interaction between the bound CO and the distal residues. A strong correlation is found between the final barrier height to CO recombination and the CO stretch frequency: as compared to wild type, the barrier is smaller in those mutants that have a higher CO stretch frequency (vCO) and vice versa. Possible reasons for this correlation are discussed. It is emphasized that the temperature and pH dependence of both the kinetics and the infrared spectra must be measured to obtain a consistent picture.

CO recombination to human myoglobin mutants in glycerol-water solutions.

The kinetics of CO recombination to site-specific mutants of human myoglobin have been studied by flash photolysis in the temperature range 250-320 K on the nanosecond to second time scale in 75% glycerol at pH 7. The mutants were constructed to examine specific proposals concerning the roles of Lys 45, Asp 60, and Val 68 in the ligand binding process. It is found that ligand recombination is nonexponential for all the mutants and that both the geminate amplitude and rate show large variations. The results are interpreted in terms of specific models connecting the dynamics and structure. It is shown that removal of the charged group at position 45 does not substantially affect the barrier height for escape or entry of the ligand; therefore the breakage of the salt bridge linking Lys 45, Asp 60, and a heme propionate is ruled out as the rate-determining barrier for this process. On the other hand, it is found that the escape barrier decreases roughly as size of the residue at position 68 increases, in the order Ala > Val > Asn > Leu. The residue at position 68 is also a major contributor to the final barrier to rebinding, but the barrier height shows no correlation with residue size and is more dependent on the stereochemistry of the residue. A molecular mechanism for ligand binding that is consistent with the results is discussed, and supporting evidence for this mechanism is examined.

X-ray crystal structure of a recombinant human myoglobin mutant at 2.8 A resolution.

We have grown crystals in trigonal space group P3(2)21 of a mutant human myoglobin, aquomet form, in which lysine at position 45 has been replaced by arginine and cysteine at position 110 has been replaced by alanine. Suitable crystals of native recombinant human myoglobin have not been obtained. We have used the molecular replacement method to determine the X-ray crystal structure of the mutant at 2.8 A resolution. At the present stage of refinement, the crystallographic R-value for the model, with tightly restrained stereochemistry, is 0.158 for 5.0 to 2.8 A data. As expected, the overall structure is quite similar to the sperm whale myoglobin structure. Arginine 45 adopts a well-ordered conformation similar to that found in aquomet sperm whale myoglobin.

Ligand and proton exchange dynamics in recombinant human myoglobin mutants.

Site-specific mutants of human myoglobin have been prepared in which lysine 45 is replaced by arginine (K45R) and aspartate 60 by glutamate (D60E), in order to examine the influence of these residues and their interaction on the dynamics of the protein. These proteins were studied by a variety of methods, including one and two-dimensional proton nuclear magnetic resonance spectroscopy, exchange kinetics for the distal and proximal histidine NH protons as a function of pH in the met cyano forms, flash photolysis of the CO forms, and ligand replacement kinetics. The electronic absorption and proton nuclear magnetic resonance spectra of the CO forms of these proteins are virtually identical, indicating that the structure of the heme pocket is unaltered by these mutations. There are, however, substantial changes in the dynamics of both CO binding and proton exchange for the mutant K45R, whereas the mutant D60E exhibits behavior indistinguishable from the reference human myoglobin. K45R has a faster CO bimolecular recombination rate and slower CO off-rate relative to the reference. The kinetics for CO binding are independent of pH (6.5 to 10) as well as ionic strength (0 to 1 M-NaCl). The exchange rate for the distal histidine NH is substantially lower for K45R than the reference, whereas the proximal histidine NH exchange rate is unaltered. The exchange behavior of the human proteins is similar to that reported for a comparison of the exchange rates for myoglobins having lysine at position 45 with sperm whale myoglobin, which has arginine at this position. This indicates that the differences in exchange rates reflects largely the Lys----Arg substitution. The lack of a simple correlation for the CO kinetics with this substitution means that these are sensitive to other factors as well. Specific kinetic models, whereby substitution of arginine for lysine at position 45 can affect ligand binding dynamics, are outlined. These experiments demonstrate that a relatively conservative change of a surface residue can substantially perturb ligand and proton exchange dynamics in a manner that is not readily predicted from the static structures.

Electrostatic interactions in wild-type and mutant recombinant human myoglobins.

Residue Val68 in human myoglobin has been replaced by Asn, Asp, and Glu with site-directed mutagenesis. Purified proteins were characterized by isoelectric focusing and by absorption, CD, and NMR spectroscopy. These studies demonstrated that Mb is able to tolerate substitution of the buried hydrophobic residue Val68 by Asn, Asp, and Glu. In the metaquo derivatives of the Glu and Asp mutants, the negative charge at residue 68 is stabilized by a favorable Coulombic interaction with the heme iron. In the absence of this interaction, as in the metcyano and ferrous deoxy derivatives, the relatively nonpolar protein interior cannot stabilize an isolated buried negative charge, and the carboxylate is either protonated or stabilized via a salt bridge with the nearby distal histidine. Hence in the Asp and Glu mutant proteins, both reduction and cyanide binding are accompanied by proton uptake by the protein. The apoproteins were prepared and reconstituted with the chlorophyll derivative zinc pyrochlorophyllide a. Absorption and fluorescence spectra were quite similar for wild-type and all mutant proteins reconstituted with this derivative. These results do not support the point charge model for the red shifts observed in the spectra of chlorophylls associated with photosynthetic proteins. From the pH dependence of the absorption spectrum of zinc pyrochlorophyllide a in the Glu mutant, the apparent pKa of the buried glutamate residue was estimated to be 8.9. This increase of 4.4 pH units, over the value for Glu in aqueous solution, provides a measure of the polarity of the protein interior.