Interplay between the mechanics of bacteriophage fibers and the strength of virus-host links.

Viral fibers play a central role in many virus infection mechanisms since they recognize the corresponding host and establish a mechanical link to its surface. Specifically, bacteriophages have to anchor to bacteria through the fibers surrounding the tail before starting the viral DNA translocation into the host. The protein gene product (gp) 37 from bacteriophage T4 long tail fibers forms a fibrous parallel homotrimer located at the distal end of the long tail fibers. Biochemical data indicate that, at least, three of these fibers are required for initial host cell interaction but do not reveal why three and no other numbers are required. By using atomic force microscopy, we obtained high-resolution images of gp37 fibers adsorbed on a mica substrate in buffer conditions and probed their local mechanical properties. Our experiments of radial indentation at the nanometer scale provided a radial stiffness of ∼ 0.08 N/m and a breaking force of ∼ 120 pN. In addition, we performed finite element analysis and determined a Young's modulus of ∼ 20 MPa. From these mechanical parameters, we hypothesize that three viral fibers provide enough mechanical strength to prevent a T4 virus from being detached from the bacteria by the viral particle Brownian motion, delivering a biophysical justification for the previous biochemical data.

Crystal structure of a heat and protease-stable part of the bacteriophage T4 short tail fibre.

Adsorption of T4 bacteriophage to the Escherichia coli host cell is mediated by six long and six short tail fibres. After at least three long tail fibres have bound, short tail fibres extend and bind irreversibly to the core region of the host cell lipopolysaccharide (LPS), serving as inextensible stays during penetration of the cell envelope by the tail tube. The short tail fibres consist of a parallel, in-register, trimer of gene product 12 (gp12). The 1.9 A crystal structure of a heat and protease-stable fragment of gp12 reveals three new folds: a central right-handed triple beta-helix, a globular C-terminal domain containing a beta-sandwich and an N-terminal beta-structure reminiscent of but different from the adenovirus triple beta-spiral. The centre of the C-terminal domain shows weak homology to gp11, a trimeric protein connecting the short fibre to the base-plate, suggesting that the trimerisation motifs of gp11 and gp12 are similar. Repeating sequence motifs suggest that the N-terminal beta-structure extends further towards the N terminus and is conserved in the long tail fibre proteins gp34 and gp37.

Identification and crystallisation of a heat- and protease-stable fragment of the bacteriophage T4 short tail fibre.

Irreversible binding of T-even bacteriophages to Escherichia coli is mediated by the short tail fibres, which serve as inextensible stays during DNA injection. Short tail fibres are exceptionally stable elongated trimers of gene product 12 (gp12), a 56 kDa protein. The N-terminal region of gp12 is important for phage attachment, the central region forms a long shaft, while a C-terminal globular region is implicated in binding to the bacterial lipopolysaccharide core. When gp12 was treated with stoichiometric amounts of trypsin or chymotrypsin at 37 degrees C, an N-terminally shortened fragment of 52 kDa resulted. If the protein was incubated at 56 degrees C before trypsin treatment at 37 degrees C, we obtained a stable trimeric fragment of 3 x 33 kDa lacking residues from both the N- and C-termini. Apparently, the protein unfolds partially at 56 degrees C, thereby exposing protease-sensitive sites in the C-terminal region and extra sites in the N-terminal region. Well-diffracting crystals of this fragment could be grown. Our results indicate that gp12 carries a stable central region, consisting of the C-terminal part of the shaft and the attached N-terminal half of the globular region. Implications for structure determination of the gp12 protein and its folding are discussed.

Kinetic analysis of adenovirus fiber binding to its receptor reveals an avidity mechanism for trimeric receptor-ligand interactions.

Most adenoviruses bind to the N-terminal immunoglobulin domain D1 of the coxsackievirus and adenovirus receptor via the head part of their fiber proteins. Three receptor molecules can bind per fiber head. We expressed the D1 domain and the adenovirus type 2 fiber head in bacteria and studied binding interactions by surface plasmon resonance measurements. When receptor domains bind adenovirus fiber independently of each other, the dissociation constant is 20-25 nm. However, when adenovirus fiber binds to receptors immobilized on the sensor chip, a situation better mimicking adenovirus binding to receptors on the cell surface, the dissociation constant was around 1 nm. Kinetic analysis shows that this happens via an avidity mechanism; three identical interactions with high on and off rate constants lead to tight binding of one fiber head to three receptor molecules with a very low overall off rate. The avidity mechanism could be used by other viruses that have multimeric adhesion proteins to attach to target cells. It could also be more general to trimeric receptor-ligand interactions, including those involved in intracellular signaling.

Dimeric structure of the coxsackievirus and adenovirus receptor D1 domain at 1.7 A resolution.

BACKGROUND: The coxsackievirus and adenovirus receptor (CAR) comprises two extracellular immunoglobulin domains, a transmembrane helix and a C-terminal intracellular domain. The amino-terminal immunoglobulin domain (D1) of CAR is necessary and sufficient for adenovirus binding, whereas the site of coxsackievirus attachment has not yet been localized. The normal cellular role of CAR is currently unknown, although CAR was recently proposed to function as a homophilic cell adhesion molecule. RESULTS: The human CAR D1 domain was bacterially expressed and crystallized. The structure was solved by molecular replacement using the structure of CAR D1 bound to the adenovirus type 12 fiber head and refined to 1.7 A resolution, including individual anisotropic temperature factors. The two CAR D1 structures are virtually identical, apart from the BC, C"D, and FG loops that are involved both in fiber head binding and homodimerization in the crystal. Analytical equilibrium ultracentrifugation shows that a dimer also exists in solution, with a dissociation constant of 16 microM. CONCLUSIONS: The CAR D1 domain forms homodimers in the crystal using the same GFCC'C" surface that interacts with the adenovirus fiber head. The homodimer is very similar to the CD2 D1-CD58 D1 heterodimer. CAR D1 also forms dimers in solution with a dissociation constant typical of other cell adhesion complexes. These results are consistent with reports that CAR may function physiologically as a homophilic cell adhesion molecule in the developing mouse brain. Adenovirus may thus have recruited an existing and conserved interaction surface of CAR to use for its own cell attachment.

A triple beta-spiral in the adenovirus fibre shaft reveals a new structural motif for a fibrous protein.

Human adenoviruses are responsible for respiratory, gastroenteric and ocular infections and can serve as gene therapy vectors. They form icosahedral particles with 240 copies of the trimeric hexon protein arranged on the planes and a penton complex at each of the twelve vertices. The penton consists of a pentameric base, implicated in virus internalization, and a protruding trimeric fibre, responsible for receptor attachment. The fibres are homo-trimeric proteins containing an amino-terminal penton base attachment domain, a long, thin central shaft and a carboxy-terminal cell attachment or head domain. The shaft domain contains a repeating sequence motif with an invariant glycine or proline and a conserved pattern of hydrophobic residues. Here we describe the crystal structure at 2.4 A resolution of a recombinant protein containing the four distal repeats of the adenovirus type 2 fibre shaft plus the receptor-binding head domain. The structure reveals a novel triple beta-spiral fibrous fold for the shaft. Implications for folding of fibrous proteins (misfolding of shaft peptides leads to amyloid-like fibrils) and for the design of a new class of artificial, silk-like fibrous materials are discussed.

Structure of the human adenovirus serotype 2 fiber head domain at 1.5 A resolution.

Adenovirus binds to its receptor via the head domain of its fiber protein. We have crystallized the adenovirus serotype 2 (subgroup C) receptor binding domain and solved the structure at 1.5 A resolution by the molecular replacement technique using the known adenovirus type 5 head structure. Included in the high-resolution model are 306 water molecules, five alternative side chain conformations, and individual anisotropic temperature factors for each atom. The overall structure of the serotype 2 head is very similar to its serotype 5 homologue, apart from differences in some of the flexible loops. All but subgroup B adenoviruses are believed to use the recently identified protein CAR (Coxsackievirus and adenovirus receptor) as receptor. By comparison of the two structures and sequence alignment of CAR binding and non-CAR binding serotype fiber heads, we discuss possible receptor binding sites and propose a receptor binding site in a crevice between two monomers on the side of the trimer. The structural basis of the extraordinary stability of the fiber head trimer is also discussed.

The structure of bovine mitochondrial F1-ATPase: an example of rotary catalysis.

There is now compelling evidence in support of a rotary catalytic mechanism in F1-ATPase, and, by extension, in the intact ATP synthase. Although models have been proposed to explain how protein translocation in F0 results in rotation of the gamma-subunit relative to the alpha 3/beta 3 assembly in F1 [22], these are still speculative. It seems likely that a satisfactory explanation of this mechanism will ultimately depend on structural information on the intact ATP synthase.

The ATPase inhibitor protein from bovine heart mitochondria: the minimal inhibitory sequence.

The mitochondrial ATPase inhibitor subunit is a basic protein of 84 amino acids that helps to regulate the activity of F1F0-ATPase. In order to obtain structural information on the mechanism of inhibition, the bovine inhibitor subunit has been expressed in Escherichia coli and purified in high yield. The recombinant protein has a similar inhibitory activity to the inhibitor subunit isolated from bovine mitochondria. Progressive N-terminal and C-terminal deletion mutants of the inhibitor subunit have been produced either by overexpression and purification, or by chemical synthesis. By assaying the truncated proteins for inhibitory activity, the minimal inhibitory sequence of the inhibitor subunit has been defined as consisting of residues 14-47. The immediately adjacent sequences 10-13 and 48-56 help to stabilize the complex between F1F0-ATPase and the inhibitor protein, and residues 1-9 and 57-84 appear to be dispensable. At physiological pH values, the inhibitor subunit is mainly alpha-helical and forms monodisperse aggregates in solution. Smaller inhibitory fragments of the inhibitor protein, such as residues 10-50, seem to have a mainly random coil structure in solution, but they can adopt the correct inhibitory conformation when they from a complex with the ATPase. However, these latter fragments are mainly monomeric in solution, suggesting that the aggregation of the inhibitor subunit in solution may be due to intermolecular alpha-helical coiled-coil formation via the C-terminal region. The noninhibitory peptides consisting of residues 10-40 and 23-84 of the inhibitor protein can bind to F1F0-ATPase, and interfere with inhibition by the intact inhibitor subunit. The noninhibitory fragments of the inhibitor protein consisting of residues 22-46 and 44-84 do not compete with the inhibitor subunit for its binding site on F1F0-ATPase.

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.

The structure of bovine F1-ATPase complexed with the antibiotic inhibitor aurovertin B.

In the structure of bovine mitochondrial F1-ATPase that was previously 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. Adenylyl-imidodiphosphate is bound to one beta-subunit (betaTP), ADP is bound to the second (betaDP), and no nucleotide is bound to the third (betaE). Here we show that the uncompetitive inhibitor aurovertin B binds to bovine F1 at two equivalent sites in betaTP and betaE, in a cleft between the nucleotide binding and C-terminal domains. In betaDP, the aurovertin B pocket is incomplete and is inaccessible to the inhibitor. The aurovertin B bound to betaTP interacts with alpha-Glu399 in the adjacent alphaTP subunit, whereas the aurovertin B bound to betaE is too distant from alphaE to make an equivalent interaction. Both sites encompass betaArg-412, which was shown by mutational studies to be involved in binding aurovertin. Except for minor changes around the aurovertin pockets, the structure of bovine F1-ATPase is the same as determined previously. Aurovertin B appears to act by preventing closure of the catalytic interfaces, which is essential for a catalytic mechanism involving cyclic interconversion of catalytic sites.

ATP synthase from bovine heart mitochondria. In vitro assembly of a stalk complex in the presence of F1-ATPase and in its absence.

Four subunits of the F1F0-ATPase from bovine heart mitochondria have been produced by heterologous over-expression in Escherichia coli. They are the oligomycin sensitivity conferral protein (OSCP), coupling factor 6 (F6) and subunits b and d. Likewise, fragments b', bI, bC, and bM (amino acid residues 79 to 214, 121 to 214, 165 to 214 and 79 to 164, respectively, of subunit b), and fragment d' (subunit d lacking residue 1 to 14) have been produced in abundant quantities by bacterial expression. These subunits, and the fragments of subunits b and d, have been assayed singly and in various combinations by gel-filtration chromatography for their abilities to bind to bovine heart F1-ATPase. Only the OSCP was found to be capable of forming a stable binary complex with F1-ATPase. When fragments b', bI or bC were added to F1-ATPase together with the OSCP, the ternary complexes F1.OSCP.b', F1.OSCP.bI or F1.OSCP.bC were formed, but b', bI and bC appeared to be present in sub-stoichiometric amounts. When F6 was added also, then the stoichiometric quaternary complexes F1.OSCP.b'.F6 and F1.OSCP.bI.F6 were obtained, as was a fourth quaternary complex containing approximately equivalent amounts of F1 and OSCP, and sub-stoichiometric quantities of bC and F6. Finally, three pentameric complexes F1.OSCP.b'.F6.d, F1.OSCP.b'.F6.d' and F1.OSCP.b.F6.d were isolated. In a further series of reconstitution experiments, the binary complexes b'.OSCP and b'.d, the ternary complex b'.d'.F6, and the quaternary complex OSCP.b'.F6.d were obtained. The pre-formed quaternary complex produced a stoichiometric pentameric complex with F1-ATPase. It was shown by S-carboxymethylation of cysteine residues with iodo-[2-14C]acetic acid that bovine F1F0-ATPase and the reconstituted F1.stalk complex, F1.OSCP.b'.d.F6, each contained one copy per complex of subunits b (or b'), OSCP and d, and that the separate stalk complex contained the same three subunits in the approximate molar ratio 1:1:1. The ratio of b to d in purified F0 was 1:1. Finally, it was demonstrated that the binding of the various subunits to F1-ATPase increases the ATP hydrolase activity and diminishes its inactivation by exposure to cold. These assembly experiments help to define some of the inter-subunit interactions in the stalk region of the F1F0-ATPase complex, and they are an essential step forward towards the goal of extending the high-resolution structure of bovine F1-ATPase into the stalk.

Fo membrane domain of ATP synthase from bovine heart mitochondria: purification, subunit composition, and reconstitution with F1-ATPase.

The Fo membrane domain of the F1Fo-ATP synthase complex has been purified from bovine heart mitochondria. The purification procedure involves the removal of peripheral membrane proteins, including F1-ATPase, from submitochondrial particles with guanidine hydrochloride, followed by extraction of Fo and other membrane proteins from the stripped membranes in the presence of the detergent n-dodecyl beta-D-maltoside. Fo was then purified by ion-exchange and dye ligand chromatography in the presence of the same detergent. Approximately 15 mg of pure Fo was recovered from 1.8 g of mitochondrial membrane protein. The purified Fo is a complex of nine different polypeptides. They are subunits a, b, c, d, e, F6, and A6L characterized before in F1Fo-ATPase preparations, and two new hitherto undetected subunits, named f and g. The sequences of subunits f and g have been determined. They are not related significantly to any known protein, but subunit f appears to contain a membrane-spanning alpha-helix. Proteins f and g are also present in approximately stoichiometric amounts in a highly purified preparation of intact F1Fo-ATPase, and so it is concluded that they are authentic subunits of the bovine enzyme with unknown functions. Dibutyltin 3-hydroxyflavone, an inhibitor of F1Fo-ATPase, also binds to the purified Fo in detergent and competes for binding with venturicidin. In the presence of F1 and OSCP, the purified Fo was reassembled into the intact F1Fo-ATPase complex. Therefore, this procedure provides a relatively abundant source of pure and functional Fo that is suitable for structural analysis.

Inherent asymmetry of the structure of F1-ATPase from bovine heart mitochondria at 6.5 A resolution.

ATP synthase, the assembly which makes ATP in mitochondria, chloroplasts and bacteria, uses transmembrane proton gradients generated by respiration or photosynthesis to drive the phosphorylation of ADP. Its membrane domain is joined by a slender stalk to a peripheral catalytic domain, F1-ATPase. This domain is made of five subunits with stoichiometries of 3 alpha: 3 beta: 1 gamma: 1 delta: 1 epsilon, and in bovine mitochondria has a molecular mass of 371,000. We have determined the 3-dimensional structure of bovine mitochondrial F1-ATPase to 6.5 A resolution by X-ray crystallography. It is an approximately spherical globule 110 A in diameter, on a 40 A stem which contains two alpha-helices in a coiled-coil. This stem is presumed to be part of the stalk that connects F1 with the membrane domain in the intact ATP synthase. A pit next to the stem penetrates approximately 35 A into the F1 particle. The stem and the pit are two examples of the many asymmetric features of the structure. The central element in the asymmetry is the longer of the two alpha-helices in the stem, which extends for 90 A through the centre of the assembly and emerges on top into a dimple 15 A deep. Features with threefold and sixfold symmetry, presumed to be parts of homologous alpha and beta subunits, are arranged around the central rod and pit, but the overall structure is asymmetric. The central helix provides a possible mechanism for transmission of conformational changes induced by the proton gradient from the stalk to the catalytic sites of the enzyme.

Crystallization of F1-ATPase from bovine heart mitochondria.

Crystals of the F1-ATPase sector of the ATP synthase complex from bovine heart mitochondria have been grown from solutions containing polyethylene glycol 6000. The crystals diffract to 2.9 A resolution on a laboratory X-ray source. They are orthorhombic and belong to the space group P2(1)2(1)2(1). The unit cell axes are a = 285 A, b = 108 A, c = 140 A. There is one molecule of F1-ATPase in the asymmetric unit.

Kirromycin drastically reduces the affinity of Escherichia coli elongation factor Tu for aminoacyl-tRNA.

We have studied the interaction between EF-Tu-GDP or EF-Tu-GTP in complex with kirromycin or aurodox (N1-methylkirromycin) and aminoacyl-tRNA, N-acetylaminoacyl-tRNA, or deacylated tRNA. Three independent methods were used: zone-interference gel electrophoresis, GTPase stimulation, and fluorescence. All three methods revealed that kirromycin induces a severe drop in the stability of the complex of EF-Tu-GTP and aminoacyl-tRNA of about 3 orders of magnitude. The affinities of EF-Tu-kirromycin-GTP and EF-Tu-kirromycin-GDP for aa-tRNA were found to be of about the same order of magnitude. We conclude that kirromycin and related compounds do not induce a so-called GTP-like conformation of EF-Tu with respect to tRNA binding. The findings shed new light on the mechanism of action of the antibiotic during the elongation cycle. In contrast to indirect evidence previously obtained in our laboratory [Van Noort et al. (1982) EMBO J. 1, 1199-1205; Van Noort et al. (1986) Proc. Natl. Acad. Sci. U.S.A. 71, 4910-4914], we were unable to demonstrate complexes of EF-Tu-aurodox-GTP/GDP with N-acetylaminoacyl-tRNA or deacylated tRNA by direct detection using zone-interference gel electrophoresis. Modification with N-tosyl-L-phenylalanine chloromethyl ketone (TPCK) decreases the affinity of EF-Tu-kirromycin-GTP for aminoacyl-tRNA, just like it does in the absence of the antibiotic.