Search

Human mitochondrial Hsp60 (mtHsp60) is a class I chaperonin, 51% identical in sequence to the prototypical E. coli chaperonin GroEL. mtHsp60 maintains the proteome within the mitochondrion and is associated with various neurodegenerative diseases and cancers. The oligomeric assembly of mtHsp60 into heptameric ring structures that enclose a folding chamber only occurs upon addition of ATP and is significantly more labile than that of GroEL, where the only oligomeric species is a tetradecamer. The lability of the mtHsp60 heptamer provides an opportunity to detect and visualize lower-order oligomeric states that may represent intermediates along the assembly/disassembly pathway. Using cryo-electron microscopy we show that, in addition to the fully-formed heptamer and an "inverted" tetradecamer in which the two heptamers associate via their apical domains, thereby blocking protein substrate access, well-defined lower-order oligomeric species, populated at less than 6% of the total particles, are observed. Specifically, we observe open trimers, tetramers, pentamers and hexamers (comprising ∼4% of the total particles) with rigid body rotations from one subunit to the next within ∼1.5-3.5° of that for the heptamer, indicating that these may lie directly on the assembly/disassembly pathway. We also observe a closed-ring hexamer (∼2% of the particles) which may represent an off-pathway species in the assembly/disassembly process in so far that conversion to the mature heptamer would require the closed-ring hexamer to open to accept an additional subunit. Lastly, we observe several classes of tetramers where additional subunits characterized by fuzzy electron density are caught in the act of oligomer extension., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Published by Elsevier Ltd.)

The human immunodeficiency virus type-1 (HIV-1) envelope (Env) proteins that mediate membrane fusion represent a major target for the development of new AIDS therapies. Three classes of Env-mediated membrane fusion inhibitors have been described that specifically target the pre-hairpin intermediate conformation of gp41. Class 2 inhibitors bind to the C-terminal heptad repeat (C-HR) of gp41. The single example of a class 3 inhibitor targets the trimeric N-terminal heptad repeat (N-HR) of gp41 and has been postulated to sequestrate the N-HR of the pre-hairpin intermediate through the formation of fusion incompetent heterotrimers. Here, we show that N(CCG)-gp41, a class 2 inhibitor, and N36(Mut(e,g)), a class 3 inhibitor, synergistically inhibit Env-mediated membrane fusion for several representative HIV-1 strains (X4 and R5) in both a cell fusion assay (with membrane-bound CD4) and an Env-pseudo-typed virus neutralization assay. The mechanistic, as well as potential therapeutic, implications of these observations for HIV-Env-mediated membrane fusion are discussed.

Using ensemble refinement of the third immunoglobulin binding domain (GB3) of streptococcal protein G (a small alpha/beta protein of 56 residues), we demonstrate that backbone (N-H, N-C', Calpha-Halpha, Calpha-C') residual dipolar coupling data in five independent alignment media, generalized order parameters from 15N relaxation data, and B-factors from a high-resolution (1.1A), room temperature crystal structure are entirely consistent with one another within experimental error. The optimal ensemble size representation is between four and eight, as assessed by complete cross-validation of the residual dipolar couplings. Thus, in the case of GB3, all three observables reflect the same low-amplitude anisotropic motions arising from fluctuations in backbone phi/psi torsion angles in the picosecond to nanosecond regime in both solution and crystalline environments, yielding a unified picture of fast, high-probability atomic motions in proteins. An understanding of these motions is crucial for understanding the impact of protein dynamics on protein function, since they provide part of the driving force for triggered conformational changes that occur, for example, upon ligand binding, signal transduction and enzyme catalysis.

Using a human non-immune phage library comprising more than 10(9) functional human antibody specificities in Fab format, we have been able to select a set of eight monoclonal Fabs targeted against diverse epitopes of the ectodomain of gp41 from HIV-1. The antigens used for panning the antibodies comprised two soluble, disulfide-linked, trimeric polypeptides derived from gp41, N(CCG)-gp41 and N35(CCG)-N13. The former comprises an exposed trimeric coiled-coil of the N-helices of gp41 fused in helical phase to the minimal thermostable ectodomain of gp41, while the latter comprises only the trimeric coiled-coil of N-helices. The selected Fabs were probed by Western blot analysis against four antigens: N(CCG)-gp41, N35CCG-N13, N34CCG (a smaller version of N35CCG-N13), and the minimal thermostable ectodomain core of gp41 in its six-helix bundle conformation (6-HB). Three classes of Fabs were found: class A (two Fabs) interact predominantly with the 6-HB; class B (four Fabs) interact with both the 6-HB and the internal trimeric coiled-coil of N-helices; and class C (two Fabs) interact specifically with the internal trimeric coiled-coil of N-helices. The IC50 values for the Fabs, expressed as bivalent mini-antibodies, ranged from 6 microg/ml to 60 microg/ml in a quantitative vaccinia virus-based reporter gene assay for HIV-1 envelope-mediated cell fusion using the envelope from the HIV-1 T tropic strain LAV. The two most potent fusion inhibitors belonged to class B. This panel of Fabs provides a set of useful probes for studying HIV-1 envelope-mediated cell fusion and may serve as a basis for developing Fab-based anti-HIV-1 therapeutics.

The solution structure of a stably phosphorylated form of the cytoplasmic B domain of the mannitol-specific transporter (IIB(Mtl)) of the Escherichia coli phosphotransferase system, containing a mutation of the active site Cys384 to Ser, has been solved by NMR. The strategy employed relies principally on backbone residual dipolar couplings recorded in three different alignment media, supplemented by nuclear Overhauser enhancement data and torsion angle restraints related specifically to the active site loop (residues 383-393). As judged from the dipolar coupling data, the remainder of the structure is unchanged upon phosphorylation within the errors of the coordinates of the previously determined solution structure of unphosphorylated wild-type IIB(Mtl). Thus, only the active site loop was refined. Phosphorylation results in a backbone atomic rms shift of approximately 0.7 angstroms in the active site loop. The resulting conformation is less than 0.5 angstroms away from the equivalent P-loop in both the low and high molecular mass eukaryotic tyrosine phosphatases. 3J(NP) coupling constant measurements using quantitative J-correlation spectroscopy provide a direct demonstration of a hydrogen bond between the phosphoryl group and the backbone amide of Ser391 at position i + 7 from phospho-Ser384, with an approximately linear P-O-H(N) bond angle. The structure also reveals additional hydrogen bonding interactions involving the backbone amides of residues at positions i + 4 and i + 5, and the hydroxyl groups of two serine residues at positions i + 6 and i + 7 that stabilize the phosphoryl group.

The solution structure of monomeric stromal cell-derived factor-1alpha (SDF-1alpha), the natural ligand for the CXCR4 G-coupled receptor, has been solved by multidimensional heteronuclear NMR spectroscopy. The structure has a characteristic chemokine fold and is in excellent agreement with the individual subunits observed in the crystal structures of dimeric SDF-1alpha. Using various peptides derived from the N-terminal extracellular tail of the CXCR4 receptor, we show that the principal determinants of binding reside in the N-terminal 17 residues of CXCR4, with a major contribution from the first six residues. From 15N/1HN chemical shift pertubation studies we show that the interaction surface on SDF-1alpha is formed by the undersurface of the three-stranded antiparallel beta-sheet bounded by the N-terminal loop on one side and the C-terminal helix on the other. This surface overlaps with but is not identical to that mapped on several other chemokines for the binding of equivalent peptides derived from their respective receptors.

The HMG-box domain of the human male sex-determining factor SRY, hSRY(HMG) (comprising residues 57-140 of the full-length sequence), binds DNA sequence-specifically in the minor groove, resulting in substantial DNA bending. The majority of point mutations resulting in 46X,Y sex reversal are located within this domain. One clinical de novo mutation, M64I in the full-length hSRY sequence, which corresponds to M9I in the present hSRY(HMG) construct, acts principally by reducing the extent of DNA bending. To elucidate the structural consequences of the M9I mutation, we have solved the 3D solution structures of wild-type and M9I hSRY(HMG) complexed to a DNA 14mer by NMR, including the use of residual dipolar couplings to derive long-range orientational information. We show that the average bend angle (derived from an ensemble of 400 simulated annealing structures for each complex) is reduced by approximately 13 degrees from 54(+/-2) degrees in the wild-type complex to 41(+/-2) degrees in the M9I complex. The difference in DNA bending can be localized directly to changes in roll and tilt angles in the ApA base-pair step involved in interactions with residue 9 and partial intercalation of Ile13. The larger bend angle in the wild-type complex arises as a direct consequence of steric repulsion of the sugar of the second adenine by the bulky S(delta) atom of Met9, whose position is fixed by a hydrogen bond with the guanidino group of Arg17. In the M9I mutant, this hydrogen bond can no longer occur, and the less bulky C(gamma)m methyl group of Ile9 braces the sugar moieties of the two adenine residues, thereby decreasing the roll and tilt angles at the ApA step by approximately 8 degrees and approximately 5 degrees, respectively, and resulting in an overall difference in bend angle of approximately 13 degrees between the two complexes. To our knowledge, this is one of the first examples where the effects of a clinical mutation involving a protein-DNA complex have been visualized at the atomic level.

The crystal structure of cyanovirin-N (CV-N), a protein with potent antiviral activity, was solved at 1.5 A resolution by molecular replacement using as the search model the solution structure previously determined by NMR. The crystals belong to the space group P3221 with one monomer of CV-N in each asymmetric unit. The primary structure of CV-N contains 101 residues organized in two domains, A (residues 1 to 50) and B (residues 51 to 101), with a high degree of internal sequence and structural similarity. We found that under the conditions of the crystallographic experiments (low pH and 26 % isopropanol), two symmetrically related monomers form a dimer by domain swapping, such that domain A of one monomer interacts with domain B' of its crystallographic symmetry mate and vice versa. Because the two swapped domains are distant from each other, domain swapping does not result in additional intramolecular interactions. Even though one of the protein sample solutions that was used for crystallization clearly contained 100 % monomeric CV-N molecules, as judged by various methods, we were only able to obtain crystals containing domain-swapped dimers. With the exception of the unexpected phenomenon of domain swapping, the crystal structure of CV-N is very similar to the NMR structure, with a root-mean-square deviation of 0.55 A for the main-chain atoms, the best agreement reported to date for structures solved using both techniques., (Copyright 1999 Academic Press.)

The solution structure of a complex between the DNA binding domain of a fungal GATA factor and a 13 base-pair oligonucleotide containing its physiologically relevant CGATAG target sequence has been determined by multidimensional nuclear magnetic resonance spectroscopy. The AREA DNA binding domain, from Aspergillus nidulans, possesses a single Cys2-Cys2 zinc finger module and a basic C-terminal tail, which recognize the CGATAG element via an extensive network of hydrophobic interactions with the bases in the major groove and numerous non-specific contacts along the sugar-phosphate backbone. The zinc finger core of the AREA DNA binding domain has the same global fold as that of the C-terminal DNA binding domain of chicken GATA-1. In contrast to the complex with the DNA binding domain of GATA-1 in which the basic C-terminal tail wraps around the DNA and lies in the minor groove, the structure of complex with the AREA DNA binding domain reveals that the C-terminal tail of the fungal domain runs parallel with the sugar phosphate backbone along the edge of the minor groove. This difference is principally attributed to amino acid substitutions at two positions of the AREA DNA binding domain (Val55, Asn62) relative to that of GATA-1 (Gly55, Lys62). The impact of the different C-terminal tail binding modes on the affinity and specificity of GATA factors is discussed., (Copyright 1998 Academic Press Limited.)

The seemingly innocuous leucine-to-valine mutation at position 22 of the AREA DNA binding domain results in dramatic changes in the in vivo expression profile of genes controlled by this GATA transcription factor. This is associated with a preference of the Leu22-->Val mutant for TGATAG sites over (A/C)GATAG sites. Quantitative gel retardation assays confirm this observation and show that the Leu22-->Val mutant AREA DNA binding domain has a approximately 30-fold lower affinity than the wild-type domain for a 13 base-pair oligonucleotide containing the wild-type CGATAG target. To gain insight into the measured affinity data and further explore sequence specificity of the AREA protein, the solution structure of a complex between the Leu22-->Val mutant AREA DNA binding domain and a 13 base-pair oligonucleotide containing its physiologically relevant TGATAG target sequence has been determined by multidimensional nuclear magnetic resonance spectroscopy. Comparison of this structure with that of the wild-type AREA DNA binding domain complexed to its cognate CGATAG target site shows how subtle changes in amino acid side-chain length and hydrophobic packing can affect affinity and specificity for GATA-containing sequences, and how changes in DNA sequence can be compensated for by changes in protein sequence., (Copyright 1998 Academic Press Limited.)

The MuA transposase of phase Mu is a large modular protein that plays a central role in transposition. We show that the Mu end DNA-binding domain, I beta gamma, which is responsible for binding the DNA attachment sites at each end of the Mu genome, comprises two subdomains, I beta and I gamma, that are structurally autonomous and do not interact with each other in the absence of DNA. The solution structure of the I gamma subdomain has been determined by multidimensional NMR spectroscopy. The structure of I gamma comprises a four helix bundle and, despite the absence of any significant sequence identity, the topology of the first three helices is very similar to that of the homeodomain family of helix-turn-helix DNA-binding proteins. The helix-turn-helix motif of I gamma, however, differs from that of the homeodomains in so far as the loop is longer and the second helix is shorter, reminiscent of that in the POU-specific domain.

The gp41 protein of the human (HIV) and simian (SIV) immunodeficiency viruses is part of the envelope glycoprotein complex gp41/gp120 which plays an essential role in viral infection. We present a multidimensional NMR study on the trimeric 44 kDa soluble ectodomain of SIV gp41 (e-gp41), comprising residues 27 to 149. Despite the large molecular weight and very limited spectral dispersion, complete backbone 1H, 13C, 13CO and 15N assignments have been made using a combination of triple resonance experiments on uniformly 13C/15N and 2H/13C/15N-labeled samples. The secondary structure of SIV e-gp41, derived on the basis of 13C chemical shifts, NH exchange rates, medium range nuclear Overhauser enhancements (NOE), and 3JHNalpha coupling constants, consists of a 49 residue helix at the N terminus (residues 29 to 77) and a 40 residue helix at the C terminus (residues 108 to 147), connected by a 30 residue loop which does not display any of the characteristics of regular secondary structure. The cross-peak intensities of the loop region in scalar correlation experiments suggests that it is more mobile than the core helical regions. The presence, however, of numerous long range NOEs, both intra and inter-subunit, within the loop indicates that it adopts a well-defined structure in which the loops from the three subunits interact with each other. Based on a number of long range intra and inter-subunit NOEs, a topological model is presented for the symmetric SIV e-gp41 trimer in which the N-terminal helices are packed within the protein interior in a parallel trimeric coiled-coil arrangement, while the C-terminal helices are located on the protein exterior, oriented antiparallel to the N-terminal helices., (Copyright 1997 Academic Press Limited.)

The B1 domain of streptococcal protein G interacts with the C-terminal fragment of the heavy chain of immunoglobulin G (IgGFc). The binding site for the protein G domain on the antibody fragment is in close proximity or overlapping with that determined for staphylococcal protein A. The interaction of the B1 domain with IgGFc was investigated by 1H-15N correlation spectroscopy. The major interaction site on the B1 domain comprises parts of beta-strand 3 as well as the alpha-helix. Comparison with the crystal structure of the protein A/IgGFc complex suggests that the mode of interaction in the two complexes is analogous, despite the lack of sequence or structural similarity between two antibody binding proteins.

The effects of the number, precision and accuracy of interproton distance restraints, of direct refinement against nuclear Overhauser enhancement (NOE) intensities and of the description of the non-bonded contacts on the precision and accuracy of a nuclear magnetic resonance (NMR) protein structure determination have been investigated. The model system employed is the 56 residue immunoglobulin G binding domain of streptococcal protein G. This choice was based on the availability of a very high resolution NMR structure (atomic root-mean-square distribution of the ensemble of 60 calculated structures about the mean co-ordinate positions of 0.25 A for the backbone atoms, 0.65 A for all atoms and 0.39 A for all atoms excluding disordered surface side-chains). The experimental NMR data set for this structure determination comprised a total of 1058 experimental restraints of which 854 were approximate interproton distance restraints corresponding to all the structurally useful NOEs observable for this protein. The calculations presented in this paper reveal the following. (1) The number of interproton distance restraints constitutes the single most important determinant of both precision and accuracy. The ensemble precision and accuracy improves significantly as the number of interproton distance restraints is increased to an average of approximately 15 per residue, of which approximately 60% involve unique proton pairs; subsequent additions of interproton distance restraints, however, lead to less dramatic improvements as information redundancy sets in. (2) The ratio of ensemble precision to ensemble accuracy (which ranges from 0.5 to 0.7 for the backbone atoms) is approximately independent both of the number, precision and accuracy of the interproton distance restraints, and of whether the structures are refined against interproton distance restraints or directly against NOE intensities. (3) In an ensemble of structures generated from a large number of loose approximate interproton distance restraints (an average of approximately 15 restraints per residue with approximately 60% involving unique proton pairs), the interproton distance vectors corresponding to the restraints are very well defined with approximately 80% of vectors between unique proton pairs having a standard deviation of < or = 0.1 A. (4) The accuracy of the mean co-ordinates of an ensemble of structures is significantly higher than the average accuracy of the individual structures comprising the ensemble. For an average ensemble precision of > or = 0.6 A, the dependence of the accuracy of the mean co-ordinates on ensemble precision is approximately linear.(ABSTRACT TRUNCATED AT 400 WORDS)

The backbone dynamics of the cytokine interleukin-8, a symmetric homodimer of overall molecular mass 16 kDa, has been investigated at pH 5.2 by means of 15N relaxation measurements using heteronuclear two-dimensional inverse detected 1H-15N spectroscopy. 15N T1, T2 and NOE data were obtained for 66 out of a total of 67 backbone amide groups. The overall correlation time is 9.10(+/- 0.05) ns at 26.6 degrees C. All residues exhibit very rapid motions on a time-scale of < or = 20 ps. These very rapid motions alone can account for the 15N relaxation behaviour of 30 residues. The 15N relaxation data for another 21 residues can only be accounted for by the inclusion of an additional internal motion on a time-scale ranging from 0.5 to 3.5 ns. These residues are clustered at the N and C termini, and in the loop regions connecting elements of secondary structures. Finally, the 15N relaxation data for another 15 residues could only be accounted for by the presence of chemical exchange on a time-scale ranging from approximately 170 ns to 2.25 ms. In addition, the inclusion of chemical exchange improved the fit to the experimental data for 10 of the 30 residues whose 15N relaxation behaviour could be accounted for by very fast motions alone. The residues exhibiting chemical exchange line broadening cluster at the interface of the long C-terminal alpha-helix and the underlying beta-sheet. It is suggested that this clustering is indicative of concerted rather than independent motions in regions of secondary structure, with motion at any one residue being propagated to neighbouring residues in van der Waals contact.

We report the results of a 500 ps molecular dynamics simulation of the cytokine interleukin-1 beta, a protein of 153 amino acids, immersed in a sphere of 3783 bulk water molecules with a radius of 33 A. The simulation reproduces the amplitudes of the fast librational motions of the backbone N-H bonds determined from 15N nuclear magnetic relaxation data, as well as the crystallographic B-factors. Moreover, this study suggests a molecular picture of the nature of the slow internal motions that have been inferred from nuclear magnetic resonance relaxation experiments. These experiments indicated that, in addition to fast motions common to all residues, 32 surface residues exhibit slow motions on the 400 ps to 5 ns time-scale. While the present simulation is not sufficiently long to provide a quantitative description of events on this time-scale, it is long enough to observe several large amplitude transitions that are likely candidates for these slow motions. Specifically, in many of these 32 residues, the N-H groups are hydrogen bonded and infrequent dihedral transitions cause the N-H vectors to jump between states with well-defined orientations. It is shown that the time course of the angular reorientational correlation functions of these residues calculated from the trajectory is a reflection of the random times at which these infrequent jumps happen to have occurred. Thus, while the rate of these transitions cannot be quantified, the simulated decay of these correlation functions is completely consistent with the physical picture in which the N-H vectors, in addition to fast librational motion, undergo large amplitude jumps between conformations stabilized by hydrogen bonds.

The presence of bound water in the solution structure of the IgG binding domain of streptococcal protein G has been investigated by nuclear magnetic resonance using three-dimensional 1H rotating frame Overhauser 1H-15N multiple quantum coherence spectroscopy. The backbone amide protons of three residues, Ala20, Gln32 and Tyr33, are found to be in close proximity to bound water. Examination of the three-dimensional structure of the IgG binding domain indicates that in the vicinity of these three residues there are no backbone groups that do not already participate in hydrogen bonding and there are no suitably placed side-chain groups available for hydrogen bonding with water. As the lifetime of the bound water detected in this nuclear magnetic resonance experiment is greater than about one nanosecond, it is likely that the two bound water molecules participate in a bifurcating hydrogen bonding network comprising a CO-NH hydrogen bonded pair, such that the water molecule accepts a hydrogen bond from the NH proton and donates one to the carbonyl oxygen with the result that the amide proton is involved in a three center hydrogen bond. On the basis of the structure, one water molecule participates in such an interaction with the Ala20(NH)-Met1(CO) hydrogen bonded pair at the beginning of an anti-parallel beta-sheet, and the other with the Tyr33(NH)-Val29(CO) hydrogen bonded pair in the single alpha-helix. The latter, which is external and solvent accessible, is associated with a distortion in the alpha-helix centered around Tyr33 which consists of a significant increase in the CO(i-4)-N(i) and CO(i-4)-NH(i) distances relative to those in the rest of the helix, as well as a significant departure in the phi, psi angles of Tyr33 relative to regular helical geometry. Such solvent induced distortions in alpha-helices have been previously noticed in crystal structures and were postulated as possible folding intermediates for helical structures. The present observation of this phenomenon in solution indicates, however, that these water molecules are tightly bound and represent an integral part of the protein framework.

The solution structure of the ribonuclease H domain of HIV-1 reverse transcriptase has been investigated by three-dimensional double and triple resonance heteronuclear magnetic resonance spectroscopy. The domain studied has 138 residues and comprises residues 427 to 560 of the 66 kDa reverse transcriptase with an additional four residues at the N terminus. Initial studies on the wild-type protein were hindered by severe differential line broadening, presumably due to conformational averaging. Mutation of the single tryptophan residue located in a loop at position 113 (position 535 in the reverse transcriptase sequence) to an alanine resulted in much improved spectral properties with no apparent change in structure. 1H, 15N and 13C backbone resonances were assigned sequentially using a range of three-dimensional double and triple resonance heteronuclear experiments on samples of uniformly (greater than 95%) 15N and 15N/13C-labeled protein, and the secondary structure was elucidated from a qualitative analysis of data derived from three-dimensional 15N- and 13C-edited nuclear Overhauser enhancement spectra. The secondary structure comprises three alpha-helices and five strands arranged in a mixed parallel/antiparallel beta-sheet with a +1, +1, -3x, -1x topology. The C-terminal region from residue 114 onwards appears to be conformationally disordered in solution as evidenced by an almost complete absence of sequential and medium range nuclear Overhauser effects.

The solution structure of interleukin-1 beta determined by nuclear magnetic resonance spectroscopy is compared to three independently solved X-ray structures at 2 A resolution. It is shown that the solution and X-ray structures are very similar, both locally and globally. The atomic root-mean-square (r.m.s.) difference between the solution and X-ray structures is approximately 0.9 A for backbone atoms, approximately 1.5 A for all atoms and approximately 1 A for all atoms of internal residues. The largest differences are confined to some of the loops and turns connecting beta-strands. The atomic r.m.s. distribution of the 32 calculated solution structures about their mean co-ordinate positions (approximately 0.4 A for backbone atoms, approximately 0.8 A for all atoms and approximately 0.5 A for all atoms of internal residues) is approximately the same as the atomic r.m.s. differences between the three X-ray structures, indicating that the positional errors in the atomic co-ordinates determined by the two methods are similar.

The presence of bound water molecules in the solution structure of reduced human thioredoxin has been investigated using three-dimensional 1H rotating frame Overhauser 1H-15N multiple quantum coherence spectroscopy. It is demonstrated that the backbone amide protons of Lys21, Lys39, Lys82, Gly83 and Asn102, as well as the side-chain amide group of Asn102, are in close proximity to bound water molecules. Examination of the high-resolution solution structure of reduced human thioredoxin reveals that these results are best accounted for by four bound water molecules. Subsequent simulated annealing calculations carried out on the basis of interproton distance and hydrogen bonding restraints to the bound water molecules, supplemented by the original set of experimental restraints used in the calculation of the three-dimensional structure of human thioredoxin, permit a more precise localization of the bound water positions. Potential hydrogen bonds to these water molecules are described and a comparison is made to corresponding bound water molecules in the crystal structure of oxidized Escherichia coli thioredoxin.

A comparison of the solution structure of the interleukin-8 dimer determined by nuclear magnetic resonance spectroscopy with that of the 2 A resolution X-ray structure, solved by molecular replacement using the solution structure as a starting model, is presented. At the monomer level the atomic root-mean-square difference between the two structures for residues 7 to 72 is approximately 1.1 A for the backbone atoms, approximately 1.6 A for all atoms, and approximately 1 A for all atoms of the internal residues. There are two main regions of difference in the monomer. In the X-ray structure residues 4 to 6 are well ordered and the charged groups of Glu4 of one subunit and Lys23' of the other are in close enough proximity to form an electrostatic interaction. In contrast, these residues are partially disordered in solution and the electrostatic interaction involving Glu4 is replaced by one between Glu29 of one subunit and Lys23' of the other. In the loop comprising residues 31 to 36, His33 accepts a hydrogen bond from the backbone amide group of Gln8 in the solution structure, but donates a hydrogen bond to the backbone carbonyl group of Glu29 in the X-ray structure. There is also a difference in the quaternary structure with regard to the relative orientation of the two subunits produced by a rigid body rotation about the C2 axis that alters the angle between the central beta-strands (formed by residues 23 to 29 of the 2 subunits) at the dimer interface, without breaking the symmetry. In the solution structure this angle has a value of 168 degrees, while in the X-ray structure the central strands are essentially flat, with an angle of 179 degrees. As a result, the separation between the two anti-parallel helices, which lie at an angle of about 60 degrees to the underlying beta-strands, is decreased from 14.8 A in the solution structure to 11.1 A in the X-ray structure. The quaternary structural difference is related to the different conformations of the N terminus and the 31 to 36 loop, both of which display different interactions with respect to the ends of the central beta-strands in the two structures. These findings indicate that interleukin-8 has the potential to undergo conformational transitions that may be of functional significance.

A low resolution solution structure of the cytokine interleukin-1 beta, a 153 residue protein of molecular weight 17,400, has been determined on the basis of 446 nuclear Overhauser effect (NOE) derived approximate interproton distance restraints involving solely NH, C alpha H and C beta H protons, supplemented by 90 distance restraints for 45 hydrogen bonds, and 79 phi torsion angle restraints. With the exception of 27 C alpha H-C alpha H NOEs, all the NOEs were assigned from a three-dimensional 1H-1H NOE 15N-1H heteronuclear multiple quantum coherence (HMQC) spectrum. The torsion angle restraints were obtained from accurate 3JHN alpha coupling constants measured from a HMQC-J spectrum, while the hydrogen bonds were derived from a qualitative analysis of the NOE, coupling constant and amide exchange data. A total of 20 simulated annealing (SA) structures was computed using the hybrid distance geometry-dynamical simulated annealing method. The solution structure of IL-1 beta comprises 12 beta-strands arranged in three pseudo-symmetrical topological units (each consisting of 5 anti-parallel beta-strands), joined by turns, short loops and long loops. The core of the structure, which is made up of the 12 beta-strands, together with the turns joining strands I and II, strands VIII and IX and strands X and XI, is well determined with a backbone atomic root-mean-square (r.m.s.) distribution about the mean co-ordinate positions of 1.2(+/- 0.1) A. The loop conformations, on the other hand, are poorly determined by the current data. A comparison of the core of the low resolution solution structure of IL-1 beta with that of the X-ray structure indicates that they are similar, with a backbone atomic r.m.s. difference of only 1.5 A between the co-ordinates of the restrained minimized mean of the SA structures and the X-ray structure.

The conformation of NAD+ in the sheep liver sorbitol dehydrogenase-NAD+ binary complex has been investigated using time-dependent proton-proton transferred nuclear Overhauser enhancement measurements to determine interproton distance ratios and distances between bound NAD+ protons. The conformation about both the adenosine and nicotinamide riboside glycosidic bonds is anti, the conformations of the adenosine and nicotinamide ribose rings are C3'-endo and C1'-exo, respectively, and the conformations about the adenosine and nicotinamide riboside C4'-C5' bonds are g+ and t, respectively, similar to those found in complexes of NAD+ with other A type dehydrogenases. In addition, however, the distance data are indicative of an unusual overall conformation of NAD+ in the sorbitol dehydrogenase-NAD+ binary complex, with the planes of the nicotinamide and adenine rings separated by 6 to 8 A and at approximately 120 degrees to each other. This overall conformation differs from the concensus extended conformation found in the NAD+-dehydrogenase complexes crystallized to date, where the planes of the nicotinamide and adenine rings are 12 to 14 A apart and nearly perpendicular to each other.

The effect of specific DNA binding of the cAMP . cAMP receptor protein complex to two DNA fragments (301 and 2685 base-pairs in length) containing the lac operon has been investigated by electron microscopy. It is shown that specific DNA binding of the cAMP . cAMP receptor protein complex induces a kink of 30 to 45 degrees in the DNA with the apex of the kink located at the site of protein attachment. These findings lend direct visual support for the kinking hypothesis based on the observation of anomalous electrophoretic mobility of DNA fragments containing specifically bound cAMP receptor protein.

This study was undertaken in order to test the models of ATP and GTP binding to carp deoxyhaemoglobin proposed by Perutz & Brunori (1982) and to find out why GTP is a more potent allosteric effector than ATP. We have determined the conformations of both nucleoside triphosphates by nuclear magnetic resonance studies and found them to be the same. The purines are in anti conformation about the glycosidic bond that links them to the ribose; the pentose ring is 3'-endo; the P-O5'-C5'-C4' torsion angle lies in the trans domain (180 degrees +/- 20 degrees); the P alpha-O-P beta and P beta-O-P gamma angles are as in the free nucleotides, i.e. the trinucleotide chain is fully extended. Models having this conformation were fitted, first manually and then by energy refinement, to the effector site of an atomic model of human deoxyhaemoglobin in which the side-chains in the NA, EF and H segments had been replaced by those of carp. The results showed the location of the polar groups in carp haemoglobin to be such that (PO4) gamma can accept hydrogen bonds from Val NA1 beta 2 and from Arg H21 beta 1, while (PO4) beta and (PO4) alpha can accept hydrogen bonds from Lys EF6 beta 1 and beta 2. In ATP, the 6-amino group of the purine can donate a hydrogen bond to Glu NA2 beta 1. In GTP, the 2-amino group can donate a hydrogen bond to Glu NA2 beta 1; in addition, Val Na1 beta 1 can donate a hydrogen bond to O2' of the ribose. This additional hydrogen bond may explain why in carp haemoglobin GTP is a stronger allosteric effector than ATP. We have found the influence of the two allosteric effectors on the oxygen affinity of trout IV haemoglobin to be the same, even though the only difference in the lining of the allosteric effector sites lies in the replacement of Glu Na2 beta in carp by Asp in trout IV haemoglobin. Model building then showed that formation of a hydrogen bond between Asp Na2 beta and the 2-amino group of guanine precludes formation of a hydrogen bond between Val NA1 beta and O2' of the ribose or vice versa, which makes the number of hydrogen bonds formed between trout IV haemoglobin and GTP the same as those formed with ATP.

A re-examination of the C-2 histidine proton resonances of haemoglobins A and Cowtown (His HC3(146) beta----Leu) in chloride-free Hepes buffer has shown that all the resonances present in haemoglobin A are present in haemoglobin Cowtown, so that the pKa of His HC3(146) beta cannot be determined by nuclear magnetic resonance in this buffer.

A nuclear magnetic resonance study on a heptadecamer (17-mer) peptide comprising the DNA binding helix F of the cyclic AMP receptor protein of Escherichia coli is presented under solution conditions (viz. 40% (v/v) trifluorethanol) where it adopts an ordered helical structure as judged by circular dichroism. Using a combination of two-dimensional nuclear magnetic resonance techniques, complete resonance assignments are obtained in a sequential manner. From the two-dimensional nuclear Overhauser enhancement spectra, a set of 87 approximate distance restraints is derived and used as the basis for three-dimensional structure determination with a restrained molecular dynamics algorithm in which the interproton distances are incorporated into the total energy function of the system in the form of an additional effective potential term. The convergence properties of this approach are tested by starting from three different initial structures, namely an alpha-helix, a beta-strand and a 3-10 helix. In all three cases, convergence to an alpha-helical structure is achieved with a root mean square difference of less than 3 A for all atoms and less than 2 A for the backbone atoms.

A restrained least-squares refinement of the solution structure of the self-complementary B DNA hexamer 5'd(C-G-T-A-C-G)2 is presented. The structure is refined on the basis of 190 inter-proton distances determined by pre-steady-state nuclear Overhauser enhancement measurements. Two refinements were carried out starting from two initial B DNA structures differing by an overall root-mean-square (r.m.s.) difference of 0.32 A. In both cases, the final r.m.s. difference between the experimental and calculated inter-proton distances was 0.12 A compared to 0.61 A and 0.58 A for the two initial structures. The difference between the two refined structures is small, with an overall r.m.s. difference of 0.16 A, and represents the error in the refined co-ordinates. The refined structures have a B-type conformation with local structural variations in backbone and glycosidic bond torsion angles, and base-pair propellor twist, base roll, base tilt and local helical twist angles.

The solution structures of two human growth hormone releasing factor analogues, 27Leu45Gly-hGHRF(1-45)OH and 27Nle-hGHRF(1-29)NH2, are investigated by means of circular dichroism and nuclear magnetic resonance spectroscopy. Using circular dichroism spectroscopy, it is shown that both peptides adopt ordered structures at low concentrations of trifluoroethanol (approximately 30%). Quantitative analysis of the circular dichroism spectra indicates that the same number of residues, approximately 23 to 25, are in a helical state in both peptides. Using two-dimensional nuclear magnetic resonance methods all proton resonances of the 27Nle-hGHRF(1-29)NH2 fragment are assigned and its secondary structure is determined from a qualitative interpretation of the nuclear Overhauser enhancement data. Two distinctive regions of alpha-helix are present extending from residues 6 to 13 and 16 to 29.

The structure of the ribotrinucleoside diphosphate UpUpC, the codon for phenylalanine, bound to yeast tRNAPhe in solution is elucidated using time-dependent proton-proton transferred nuclear Overhauser enhancement measurements to determine distances between bound ligand protons. The glycosidic bond and ribose conformations are low anti and 3'-endo, respectively, typical of an A-RNA type structure. The main chain torsion angles are all within the range of those expected for A-RNA but small differences from those in conventional A-RNA 11 result in a special structure with a larger rotation per residue (40 to 45 degrees compared to 32.7 degrees in R-RNA 11) and almost perfect stacking of the bases. These two structural features, which are similar to those found in the anticodon triplet of the monoclinic crystal form of tRNAPhe, can account for the known greater stability of the codon-anticodon complex relative to an equivalent double helical RNA trimer with a conventional A-RNA structure.

The applicability of restrained molecular dynamics for the determination of three-dimensional protein structures on the basis of short interproton distances (less than 4 A) that can be realistically determined from nuclear magnetic resonance measurements in solution is assessed. The model system used is the 1.2 A resolution crystal structure of the 46 residue protein crambin, from which a set of 240 approximate distance restraints, divided into three ranges (2.5 +/- 0.5, 3.0+0.5(-1.0) and 4 +/- 1 A), is derived. This interproton distance set comprises 159 short-range ([i-j] less than or equal to 5) and 56 ([i-j] greater than 5) long-range inter-residue distances and 25 intra-residue distances. Restrained molecular dynamics are carried out using a number of different protocols starting from two initial structures: a completely extended beta-strand; and an extended structure with two alpha-helices in the same positions as in the crystal structure (residues 7 to 19, and 23 to 30) and all other residues in the form of extended beta-strands. The root-mean-square (r.m.s.) atomic differences between these two initial structures and the crystal structure are 43 A and 23 A, respectively. It is shown that, provided protocols are used that permit the secondary structure elements to form at least partially prior to folding into a tertiary structure, convergence to the correct final structure, both globally and locally, is achieved. The r.m.s. atomic differences between the converged restrained dynamics structures and the crystal structure range from 1.5 to 2.2 A for the backbone atoms and from 2.0 to 2.8 A for all atoms. The r.m.s. atomic difference between the X-ray structure and the structure obtained by first averaging the co-ordinates of the converged restrained dynamics structures is even smaller: 1.0 A for the backbone atoms and 1.6 A for all atoms. These results provide a measure with which to judge future experimental results on proteins whose crystal structures are unknown. In addition, from an examination of the dynamics trajectories, it is shown that the convergence pathways followed by the various simulations are different.

We have recorded the C-2 proton resonances of the histidines of carbonmonoxyhaemoglobin A and of four abnormal human HbCOs in different buffers and at different concentrations of haemoglobin. Resonance H assigned by Perutz et al. (1985) to His HC3(146) beta, is present at both pH 7.30 and pH 6.90, but somewhat broadened when recorded in 5 to 10% HbCO A in 0.1 M-bis-Tris. The broadening disappears on tenfold dilution of the Hb with bis-Tris and the resonance then stands out sharply. Resonance H is absent at both Hb concentrations in HbCO Cowtown (His HC3(146) beta----Leu). HbCO Fort de France (His CD3(45) alpha----Arg) in 0.1 M-bis-Tris of pH 6.90 has a spectrum similar to that of HbCO A. In the same buffer a resonance marked L by Russu et al. (1982) is absent from the spectrum of Hb Abbruzzo (His H21(143) beta----Arg), whereas resonance H is present. Hb Barcelona contains an additional histidine in position FG1(94) beta; in 0.1 M-bis-Tris buffer of pH 6.90 its resonance is not resolved and resonance H is either shifted or broadened. The resonances of both histidines are resolved in phosphate buffer. At pH 6.90, spectra in 0.1 M-bis-Tris buffer are similar to those previously recorded in 0.2 M-HEPES. Addition of 0.1 M-KCl produces marked changes. Replacement of bis-Tris by 0.2 M-KCl + 0.2 M-phosphate gives rise to a different and much better resolved spectrum.

The solution structure of the self-complementary DNA hexamer 5' d(C-G-T-A-C-G)2 is refined by restrained molecular dynamics in which 192 interproton distances, determined from pre-steady-state nuclear Overhauser enhancement measurements, are incorporated into the total energy of the system in the form of effective potentials. First the method is tested by applying an idealized set of distance restraints taken from classical B-DNA to a simulation starting off from A-DNA and vice versa. It is shown that in both cases the expected transition between A- and B-DNA occurs. Second, a set of restrained molecular dynamics calculations is carried out starting from both A- and B-DNA with the experimental interproton distances for 5' d(C-G-T-A-C-G)2 as restraints. Convergence to the same B-type structure is achieved with the interproton distances equal to the measured values within experimental error. The root-mean-square atomic difference between the two average restrained dynamics structures (less than 1 A) is approximately the same as the root-mean-square fluctuations of the atoms.

The solution structure of the 45-residue plant protein, alpha 1-purothionin, is investigated by nuclear magnetic resonance (n.m.r.) spectroscopy. Using a combination of two-dimensional n.m.r. techniques to demonstrate through-bond and through-space (less than 5 A) connectivities, the 1H n.m.r. spectrum of alpha 1-purothionin is assigned in a sequential manner. The secondary structure elements are then delineated on the basis of a qualitative interpretation of short-range nuclear Overhauser effects (NOE) involving the NH, C alpha H and C beta H protons. There are two helices extending from residues 10 to 19 and 23 to 28, two short beta-strands from residues 3 to 5 and 31 to 34 which form a mini anti-parallel beta-sheet, and five turns. In addition, a number of long-range NOE connectivities are assigned and a low resolution tertiary structure is proposed.

A nuclear magnetic resonance study of the conformation of the tetrapeptide acetyl-Pro-Ala-Pro-Tyr-NH2 bound to porcine pancreatic elastase is presented. From two-dimensional transferred nuclear Overhauser enhancement measurements, a set of 23 approximate distance restraints between pairs of bound ligand protons, indicative of an extended type structure, is derived. The structure of the bound tetrapeptide is then refined from two different starting structures (an extended beta-strand and a polyproline helix) by restrained molecular dynamics, in which the interproton distances are incorporated into the total energy of the system in the form of effective potentials. Convergence to essentially the same average restrained dynamics structures is achieved. The refined structures are then modelled into the active site of elastase by interactive molecular graphics. The determination of the anchor point of the bound tetrapeptide on the enzyme was aided by a simultaneous crystallographic study which, despite the fact that only electron density for a Pro-X dipeptide fragment was visible, enabled both the approximate position and orientation of binding to be determined. It is found that the tetrapeptide is bound in the S' binding site in the reverse orientation found in other serine protease-inhibitor complexes and is stabilized both by hydrogen-bonding and by van der Waals' interactions.

Studies of abnormal and chemically modified haemoglobins indicate that in 0.1 M-NaCl about 40% of the alkaline Bohr effect of human haemoglobin is contributed by the C-terminal histidine HC3(146) beta. In deoxyhaemoglobin, the imidazole of this histidine forms a salt bridge with aspartate FG1(94) beta, in oxyhaemoglobin or carbonmonoxyhaemoglobin it accepts a hydrogen bond from its own NH group instead. Kilmartin et al. (1973) showed that in 0.2 M-NaCl + 0.2 M-phosphate this change of ligation lowered the pKa of the histidine from 8.0 in Hb to 7.1 in HbCO, but Russu et al. (1980) claimed that in bis-Tris buffer without added NaCl its pKa in HbCO dropped no lower than 7.85, and that in this medium the C-terminal histidine made only a negligible contribution to the alkaline Bohr effect. We have compared the histidine resonances of HbCO A with those of three abnormal haemoglobins: HbCO Cowtown (His HC3(146)beta----Leu), HbCO Wood (His FG4(97)beta----Leu) and HbCO Malmø (His FG4(97)beta----Gln). Our results show that the resonance assigned by Russu et al. to His HC3(146)beta in fact belongs to His FG4(97)beta. Although in Hb the pKa of His HC3(146)beta is 8.05 +/- 0.05 independent of ionic strength, in HbCO its pKa drops sharply with diminishing ionic strength, so that in the buffer employed by Russu et al. it has a pKa of 6.2 and makes a contribution to the alkaline Bohr effect that is 57% larger than in the phosphate buffer employed by Kilmartin et al. (1973). In HbCO A, His FG4(97)beta does not contribute to the Bohr effect, but in HbCO from which His HC3(146)beta has been cleaved (HbCO des-His), His FG4(97)beta is in equilibrium between two conformations with different pKa values. This equilibrium varies with ionic strength and pH, and presumably also with degree of ligation of the haem moiety. In HbCO A, His FG4(97)beta has a pKa of 7.8 compared to the pKa value of about 6.6 characteristic of free histidines at the surface of proteins. This high pKa is accounted for by its interaction with the negative pole at the C terminus of helices F and FG. It corresponds to a free energy change of the same order as that observed in the interaction of histidines with carboxylate ions and confirms the strongly dipolar character of alpha-helices, which manifests itself even when they lie on the surface of the protein.

A time-dependent transferred nuclear Overhauser enhancement study of the conformation of the single-stranded DNA 11mer 5'd(A-A-G-T-G-T-G-A-T-A-T) bound to the single-stranded DNA binding protein of Escherichia coli (SSB) is presented. It is shown that the conformation of the bound 11mer is that of a right-handed B-type helix similar to that of the free 11mer. The observation of internucleotide transferred nuclear Overhauser enhancements for every base step excludes the possibility of intercalation by aromatic protein residues. In addition, it is shown that the effective correlation time of the bases (80 ns) corresponds to that of a complex of molecular weight approximately 170,000, containing two SSB tetramers. The sugars, on the other hand, exhibit a shorter effective correlation time (40 ns), indicating the presence of internal motion. This suggests that the bases are anchored to the protein surface, possibly by hydrophobic interactions, whereas the sugar-phosphate groups are directed outwards towards the solvent.