The structure of bovine F1-ATPase complexed with the peptide antibiotic efrapeptin.

In the previously determined structure of mitochondrial F1-ATPase determined with crystals grown in the presence of adenylyl-imidodiphosphate (AMP-PNP) and ADP, the three catalytic beta-subunits have different conformations and nucleotide occupancies. AMP-PNP and ADP are bound to subunits beta TP and beta DP, respectively, and the third beta-subunit (beta E) has no bound nucleotide. The efrapeptins are a closely related family of modified linear peptides containing 15 amino acids that inhibit both ATP synthesis and hydrolysis by binding to the F1 catalytic domain of F1F0-ATP synthase. In crystals of F1-ATPase grown in the presence of both nucleotides and inhibitor, efrapeptin is bound to a unique site in the central cavity of the enzyme. Its binding is associated with small structural changes in side chains of F1-ATPase around the binding pocket. Efrapeptin makes hydrophobic contacts with the alpha-helical structure in the gamma-subunit, which traverses the cavity, and with subunit beta E and the two adjacent alpha-subunits. Two intermolecular hydrogen bonds could also form. Intramolecular hydrogen bonds probably help to stabilize efrapeptin's two domains (residues 1-6 and 9-15, respectively), which are connected by a flexible region (beta Ala-7 and Gly-8). Efrapeptin appears to inhibit F1-ATPase by blocking the conversion of subunit beta E to a nucleotide binding conformation, as would be required by an enzyme mechanism involving cyclic interconversion of catalytic sites.

Structure of the catalytic fragment of poly(AD-ribose) polymerase from chicken.

The crystal structures of the catalytic fragment of chicken poly(ADP-ribose) polymerase [NAD+ ADP-ribosyltransferase; NAD+:poly(adenosine-diphosphate-D-ribosyl)-acceptor ADP-D-ribosyltransferase, EC 2.4.2.30] with and without a nicotinamide-analogue inhibitor have been elucidated. Because this enzyme is involved in the regulation of DNA repair, its inhibitors are of interest for cancer therapy. The inhibitor shows the nicotinamide site and also suggests the adenosine site. The enzyme is structurally related to bacterial ADP-ribosylating toxins but contains an additional alpha-helical domain that is suggested to relay the activation signal issued on binding to damaged DNA.

Mutation of Pro-258 in transmembrane domain 6 constitutively activates the G protein-coupled alpha-factor receptor.

The alpha-factor pheromone receptor stimulates MATa yeast cells to undergo conjugation. The receptor contains seven transmembrane domains that function in ligand binding and in transducing a signal to the cytoplasmic receptor sequences to mediate G protein activation. A genetic screen was used to isolate receptor mutations that constitutively signal in the absence of alpha-factor. The Pro-258-->Leu (P258L) mutation caused constitutive receptor signaling that was equivalent to about 45% of the maximum level observed in wild-type cells stimulated with alpha-factor. Mutations of both Pro-258 and the adjacent Ser-259 to Leu increased constitutive signaling to > or = 90% of the maximum level. Since Pro-258 occurs in the central portion of transmembrane domain 6, and since proline residues are expected to cause a kink in alpha-helical domains, the P258L mutation is predicted to alter the structure of transmembrane domain 6. The P258L mutation did not result in a global distortion of receptor structure because alpha-factor bound to the mutant receptors with high affinity and induced even higher levels of signaling. These results suggest that sequences surrounding Pro-258 may be involved in ligand activation of the receptor. Conformational changes in transmembrane domain 6 may effect a change in the adjacent sequences in the third intracellular loop that are thought to function in G protein activation. Greater than 90% of all G protein-coupled receptors contain a proline residue at a similar position in transmembrane domain 6, suggesting that this aspect of receptor activation may be conserved in other receptors.

Minimizing a binding domain from protein A.

We present a systematic approach to minimizing the Z-domain of protein A, a three-helix bundle (59 residues total) that binds tightly (Kd = 10 nM) to the Fc portion of an immunoglobin IgG1. Despite the fact that all the contacts seen in the x-ray structure of the complex with the IgG are derived from residues in the first two helices, when helix 3 is deleted, binding affinity is reduced > 10(5)-fold (Kd > 1 mM). By using structure-based design and phage display methods, we have iteratively improved the stability and binding affinity for a two-helix derivative, 33 residues in length, such that it binds IgG1, with a Kd of 43 nM. This was accomplished by stepwise selection of random mutations from three regions of the truncated Z-peptide: the 4 hydrophobic residues from helix 1 and helix 2 that contacted helix 3 (the exoface), followed by 5 residues between helix 1 and helix 2 (the intraface), and lastly by 19 residues at or near the interface that interacts with Fc (the interface). As selected mutations from each region were compiled (12 in total), they led to progressive increases in affinity for IgG, and concomitant increases in alpha-helical content reflecting increased stabilization of the two-helix scaffold. Thus, by sequential increases in the stability of the structure and improvements in the quality of the intermolecular contacts, one can reduce larger binding domains to smaller ones. Such mini-protein binding domains are more amenable to synthetic chemistry and thus may be useful starting points for the design of smaller organic mimics. Smaller binding motifs also provide simplified and more tractable models for understanding determinants of protein function and stability.

Structure and conformational changes of DNA topoisomerase II visualized by electron microscopy.

Type II DNA topoisomerases, which create a transient gate in duplex DNA and transfer a second duplex DNA through this gate, are essential for topological transformations of DNA in prokaryotic and eukaryotic cells and are of interest not only from a mechanistic perspective but also because they are targets of agents for anticancer and antimicrobial chemotherapy. Here we describe the structure of the molecule of human topoisomerase II [DNA topoisomerase (ATP-hydrolyzing), EC 5.99.1.3] as seen by scanning transmission electron microscopy. A globular approximately 90-angstrom diameter core is connected by linkers to two approximately 50-angstrom domains, which were shown by comparison with genetically truncated Saccharomyces cerevisiae topoisomerase II to contain the N-terminal region of the approximately 170-kDa subunits and that are seen in different orientations. When the ATP-binding site is occupied by a nonhydrolyzable ATP analog, a quite different structure is seen that results from a major conformational change and consists of two domains approximately 90 angstrom and approximately 60 angstrom in diameter connected by a linker, and in which the N-terminal domains have interacted. About two-thirds of the molecules show an approximately 25 A tunnel in the apical part of the large domain, and the remainder contain an internal cavity approximately 30 A wide in the large domain close to the linker region. We propose that structural rearrangements lead to this displacement of an internal tunnel. The tunnel is likely to represent the channel through which one DNA duplex, after capture in the clamp formed by the N-terminal domains, is transferred across the interface between the enzyme's subunits. These images are consistent with biochemical observations and provide a structural basis for understanding the reaction of topoisomerase II.

Three-dimensional structure of human protein kinase C interacting protein 1, a member of the HIT family of proteins.

The three-dimensional structure of protein kinase C interacting protein 1 (PKCI-1) has been solved to high resolution by x-ray crystallography using single isomorphous replacement with anomalous scattering. The gene encoding human PKCI-1 was cloned from a cDNA library by using a partial sequence obtained from interactions identified in the yeast two-hybrid system between PKCI-1 and the regulatory domain of protein kinase C-beta. The PKCI-1 protein was expressed in Pichia pastoris as a dimer of two 13.7-kDa polypeptides. PKCI-1 is a member of the HIT family of proteins, shown by sequence identity to be conserved in a broad range of organisms including mycoplasma, plants, and humans. Despite the ubiquity of this protein sequence in nature, no distinct function has been shown for the protein product in vitro or in vivo. The PKCI-1 protomer has an alpha+beta meander fold containing a five-stranded antiparallel sheet and two helices. Two protomers come together to form a 10-stranded antiparallel sheet with extensive contacts between a helix and carboxy terminal amino acids of a protomer with the corresponding amino acids in the other protomer. PKCI-1 has been shown to interact specifically with zinc. The three-dimensional structure has been solved in the presence and absence of zinc and in two crystal forms. The structure of human PKCI-1 provides a model of this family of proteins which suggests a stable fold conserved throughout nature.

Structure and function in rhodopsin: correct folding and misfolding in point mutants at and in proximity to the site of the retinitis pigmentosa mutation Leu-125-->Arg in the transmembrane helix C.

L125R is a mutation in the transmembrane helix C of rhodopsin that is associated with autosomal dominant retinitis pigmentosa. To probe the orientation of the helix and its packing in the transmembrane domain, we have prepared and studied the mutations E122R, I123R, A124R, S127R, L125F, and L125A at, and in proximity to, the above mutation site. Like L125R, the opsin expressed in COS-1 cells from E122R did not bind 11-cis-retinal, whereas those from I123R and S127R formed the rhodopsin chromophore partially. A124R opsin formed the rhodopsin chromophore (lambda max 495 nm) in the dark, but the metarhodopsin II formed on illumination decayed about 6.5 times faster than that of the wild type and was defective in transducin activation. The mutant opsins from L125F and L125A bound 11-cis-retinal only partially, and in both cases, the mixtures of the proteins produced were separated into retinal-binding and non-retinal-binding (misfolded) fractions. The purified mutant rhodopsin from L125F showed lambda max at 500 nm, whereas that from L125A showed lambda max at 503 nm. The mutant rhodopsin L125F showed abnormal bleaching behavior and both mutants on illumination showed destabilized metarhodopsin II species and reduced transducin activation. Because previous results have indicated that misfolding in rhodopsin is due to the formation of a disulfide bond other than the normal disulfide bond between Cys-110 and Cys-187 in the intradiscal domain, we conclude from the misfolding in mutants L125F and L125A that the folding in vivo in the transmembrane domain is coupled to that in the intradiscal domain.

Crystal structure of the T state of allosteric yeast chorismate mutase and comparison with the R state.

The crystal structure of the tyrosine-bound T state of allosteric yeast Saccharomyces cerevisiae chorismate mutase was solved by molecular replacement at a resolution of 2.8 angstroms using a monomer of the R-state structure as the search model. The allosteric inhibitor tyrosine was found to bind in the T state at the same binding site as the allosteric activator tryptophan binds in the R state, thus defining one regulatory binding site for each monomer. Activation by tryptophan is caused by the larger steric size of its side chain, thereby pushing apart the allosteric domain of one monomer and helix H8 of the catalytic domain of the other monomer. Inhibition is caused by polar contacts of tyrosine with Arg-75 and Arg-76 of one monomer and with Gly-141, Ser-142, and Thr-145 of the other monomer, thereby bringing the allosteric and catalytic domains closer together. The allosteric transition includes an 8 degree rotation of each of the two catalytic domains relative to the allosteric domains of each monomer (domain closure). Alternatively, this transition can be described as a 15 degree rotation of the catalytic domains of the dimer relative to each other.

Swiveling-domain mechanism for enzymatic phosphotransfer between remote reaction sites.

The crystal structure of pyruvate phosphate dikinase, a histidyl multiphosphotransfer enzyme that synthesizes adenosine triphosphate, reveals a three-domain molecule in which the phosphohistidine domain is flanked by the nucleotide and the phosphoenolpyruvate/pyruvate domains, with the two substrate binding sites approximately 45 angstroms apart. The modes of substrate binding have been deduced by analogy to D-Ala-D-Ala ligase and to pyruvate kinase. Coupling between the two remote active sites is facilitated by two conformational states of the phosphohistidine domain. While the crystal structure represents the state of interaction with the nucleotide, the second state is achieved by swiveling around two flexible peptide linkers. This dramatic conformational transition brings the phosphocarrier residue in close proximity to phosphoenolpyruvate/pyruvate. The swiveling-domain paradigm provides an effective mechanism for communication in complex multidomain/multiactive site proteins.

The crystal structure of human glycosylation-inhibiting factor is a trimeric barrel with three 6-stranded beta-sheets.

Glycosylation-inhibiting factor (GIF) is a cytokine that is involved in the regulation of IgE synthesis. The crystal structure of recombinant human GIF was determined by the multiple isomorphous replacement method. The structure was refined to an R factor of 0.168 at 1.9 angstrom resolution. The overall structure is seen to consist of three interconnected subunits forming a barrel with three 6-stranded beta-sheets on the inside and six alpha-helices on the outside. There is a 5-angstrom-diameter "hole" through the middle of the barrel. The barrel structure of GIF in part resembles other "trefoil" cytokines such as interleukin 1 and fibroblast growth factor. Each subunit has a new class of alpha + beta sandwich structure consisting of two beta-alpha-beta motifs. These beta-alpha-beta motifs are related by a pseudo-twofold axis and resemble both interleukin 8 and the peptide binding domain of major histocompatibility complex protein, although the topology of the polypeptide chain is quite different.

The wing of the enhancer-binding domain of Mu phage transposase is flexible and is essential for efficient transposition.

A tetramer of the Mu transposase (MuA) pairs the recombination sites, cleaves the donor DNA, and joins these ends to a target DNA by strand transfer. Juxtaposition of the recombination sites is accomplished by the assembly of a stable synaptic complex of MuA protein and Mu DNA. This initial critical step is facilitated by the transient binding of the N-terminal domain of MuA to an enhancer DNA element within the Mu genome (called the internal activation sequence, IAS). Recently we solved the three-dimensional solution structure of the enhancer-binding domain of Mu phage transposase (residues 1-76, MuA76) and proposed a model for its interaction with the IAS element. Site-directed mutagenesis coupled with an in vitro transposition assay has been used to assess the validity of the model. We have identified five residues on the surface of MuA that are crucial for stable synaptic complex formation but dispensable for subsequent events in transposition. These mutations are located in the loop (wing) structure and recognition helix of the MuA76 domain of the transposase and do not seriously perturb the structure of the domain. Furthermore, in order to understand the dynamic behavior of the MuA76 domain prior to stable synaptic complex formation, we have measured heteronuclear 15N relaxation rates for the unbound MuA76 domain. In the DNA free state the backbone atoms of the helix-turn-helix motif are generally immobilized whereas the residues in the wing are highly flexible on the pico- to nanosecond time scale. Together these studies define the surface of MuA required for enhancement of transposition in vitro and suggest that a flexible loop in the MuA protein required for DNA recognition may become structurally ordered only upon DNA binding.

Hybrid restriction enzymes: zinc finger fusions to Fok I cleavage domain.

A long-term goal in the field of restriction-modification enzymes has been to generate restriction endonucleases with novel sequence specificities by mutating or engineering existing enzymes. This will avoid the increasingly arduous task of extensive screening of bacteria and other microorganisms for new enzymes. Here, we report the deliberate creation of novel site-specific endonucleases by linking two different zinc finger proteins to the cleavage domain of Fok I endonuclease. Both fusion proteins are active and under optimal conditions cleave DNA in a sequence-specific manner. Thus, the modular structure of Fok I endonuclease and the zinc finger motifs makes it possible to create "artificial" nucleases that will cut DNA near a predetermined site. This opens the way to generate many new enzymes with tailor-made sequence specificities desirable for various applications.

Determining RNA solution structure by segmental isotopic labeling and NMR: application to Caenorhabditis elegans spliced leader RNA 1.

Recent developments in multidimensional heteronuclear NMR spectroscopy and large-scale synthesis of uniformly 13C- and 15N-labeled oligonucleotides have greatly improved the prospects for determination of the solution structure of RNA. However, there are circumstances in which it may be advantageous to label only a segment of the entire RNA chain. For example, in a larger RNA molecule the structural question of interest may reside in a localized domain. Labeling only the corresponding nucleotides simplifies the spectrum and resonance assignments because one can filter proton spectra for coupling to 13C and 15N. Another example is in resolving alternative secondary structure models that are indistinguishable in imino proton connectivities. Here we report a general method for enzymatic synthesis of quantities of segmentally labeled RNA molecules required for NMR spectroscopy. We use the method to distinguish definitively two competing secondary structure models for the 5' half of Caenorhabditis elegans spliced leader RNA by comparison of the two-dimensional [15N] 1H heteronuclear multiple quantum correlation spectrum of the uniformly labeled sample with that of a segmentally labeled sample. The method requires relatively small samples; solutions in the 200-300 microM concentration range, with a total of 30 nmol or approximately 40 micrograms of RNA in approximately 150 microliters, give strong NMR signals in a short accumulation time. The method can be adapted to label an internal segment of a larger RNA chain for study of localized structural problems. This definitive approach provides an alternative to the more common enzymatic and chemical footprinting methods for determination of RNA secondary structure.

Three distinct structural environments of a transmembrane domain in the inwardly rectifying potassium channel ROMK1 defined by perturbation.

To probe the protein environment of an ion channel, we have perturbed the structure of a transmembrane domain by substituting side chains with those of two different sizes by using site-specific mutagenesis. We have used Trp and Ala as a high- and a low-impact perturbation probe, respectively, to replace each of 18 consecutive residues within the putative second transmembrane segment, M2, of an inwardly rectifying potassium channel, ROMK1. Our rationale is that a change in the channel function as a consequence of these mutations at a particular position will reflect the structural environment of the altered side chain. Each position can then be assigned to one of three classes of environments, as grated by different levels of perturbation: very tolerant (channel functions with both Trp and Ala substitutions), tolerant (function preserved with Ala but not with Trp substitution), and intolerant (either Ala or Trp substitution destroys function). We identify the very tolerant environment as being lipid-facing, tolerant as protein-interior-facing, and intolerant as pore-facing. We observe a strikingly ordered pattern of perturbation of all three environmental classes. This result indicates that M2 is a straight alpha-helix.

A soluble domain of the membrane-anchoring chain of influenza virus hemagglutinin (HA2) folds in Escherichia coli into the low-pH-induced conformation.

The extensive refolding of the membrane-anchoring chain of hemagglutinin (HA) of influenza virus (termed HA2) in cellular endosomes, which initiates viral entry by membrane fusion, suggests that viral HA is meta-stable. HA2 polypeptide residues 38-175 expressed in Escherichia coli are reported here to fold in vivo into a soluble trimer. The structure appears to be the same as the low-pH-induced conformation of viral HA2 by alpha-helical content, thermodynamic stability, protease dissection, electron microscopy, and antibody binding. These results provide evidence that the structure of the low-pH-induced fold of viral HA2 (TBHA2) observed crystallographically is the lowest-energy-state fold of the HA2 polypeptide. They indicate that the HA2 conformation in viral HA before low pH activation of its fusion potential is metastable and suggest that removal of the receptor-binding chain (HA1) is enough to allow HA2 to adopt the stable state. Further, they provide direct evidence that low pH is not required to form the membrane-fusion conformation but acts to make this state kinetically accessible in viral HA.