R.A. Crowther

Harmonic analysis of electron microscope images with rotational symmetry

Abstract This paper describes a method of analysing images from electron micrographs of biological specimens believed to possess rotational symmetry. An objective analysis of the symmetry is possible because the method, which is computational, produces a rotational power spectrum of the image. We can then combine just those components which are consistent with the previously determined symmetry to produce a filtered image. The method is applied to the base plate of bacteriophage T4 and to discs of tobacco mosaic virus protein. The advantages of this new approach over the well-known Markham rotation technique are discussed.

On the structure of coated vesicles

Abstract Electron micrographs of tilted specimens of coated vesicles show that their coats are based on polyhedral lattices constructed from 12 pentagons plus a variable number of hexagons. We have identified three such structures among the smaller particles, two containing 108 molecules of clathrin and a third containing 84. The coats of larger particles are believed to be constructed on similar principles. This polymorphism enables a variety of vesicles to be accommodated in an economical manner.

Visualization of alpha-helices in tobacco mosaic virus by cryo-electron microscopy

We have used tobacco mosaic virus (TMV) as a test specimen, in order to develop techniques for the analysis of high-resolution structural detail in electron micrographs of biological assemblies with helical symmetry. It has previously been shown that internal details of protein structure can be visualized by processing electron micrographs of unstained specimens of extended two-dimensional crystalline arrays. However, the techniques should in principle be applicable to other periodic specimens, such as assemblies with helical symmetry. We show here that data to spacings better than 10 A can be retrieved from electron images of frozen hydrated TMV. The three-dimensional computed map agrees well with that derived from X-ray diffraction and shows the two pairs of alpha-helices forming the core of the coat subunit, the C alpha-helix and the viral RNA. The results demonstrate that it is possible to determine detailed internal structure in helical particles.

Molecular reorganization in the hexagon to star transition of the baseplate of bacteriophage T4

Abstract The baseplate of bacteriophage T4 is a complex structure containing at least 14 different structural proteins. It undergoes a transition from a hexagonal to a star-shaped configuration during infection of the host bacterial cell. We have used a combination of genetics and image processing of electron micrographs to analyse both the wild-type structure and a series of mutant structures lacking specific gene products. Besides describing the basic anatomy of the hexagon and star configurations, we have been able to locate the products of genes 9, 11 and 12. Gene 9 product occupies a peripheral position in hexagons and stars consistent with its providing a binding site for the long tail fibres. Gene 11 product in the hexagon forms the distal part of the tail pin, which folds out to form the point of the hexagram in the star configuration. Gene 12 product is visualized as an extended 350 A fibre in stars and broken baseplates but appears to have a more compact configuration in hexagons and intact phage. We demonstrate the structural relationship between the hexagonal and starshaped configurations and show how the positions of the specific gene products alter as a result of the structural transition. We suggest a speculative model for the role of gene 9 and gene 12 products in triggering the rearrangement of the baseplate and tail contraction.

Arrangement of the heads of myosin in relaxed thick filaments from tarantula muscle

Thick filaments from leg muscle of tarantula, maintained under relaxing conditions (Mg-ATP and EGTA), were negatively stained and photographed with minimal electron dose. Particles were selected for three-dimensional image reconstruction by general visual appearance and by the strength and symmetry of their optical diffraction patterns, the best of which extend to spacings of 1/5 nm-1. The helical symmetry is such that, on a given layer-line, Bessel function contributions of different orders start to overlap at fairly low resolution and must therefore be separated computationally by combining data from different views. Independent reconstructions agree well and show more detail than previous reconstructions of thick filaments from Limulus and scallop. The strongest feature is a set of four long-pitch right-handed helical ridges (pitch 4 X 43.5 nm) formed by the elongated myosin heads. The long-pitch helices are modulated to give ridges with an axial spacing of 14.5 nm, lying in planes roughly normal to the filament axis and running circumferentially. We suggest that the latter may be formed by the stacking of a subfragment 1 (S1) head from one myosin molecule on an S1 from an axially neighbouring molecule. Internal features in the map indicate an approximate local twofold axis relating the putative heads within a molecule. The heads appear to point in opposite directions along the filament axis and are located very close to the filament backbone. Thus, for the first time, the two heads of the myosin molecule appear to have been visualized in a native thick filament under relaxing conditions.

Structure of tomato bushy stunt virus: II. Comparison of results obtained by electron microscopy and X-ray diffraction

A three-dimensional reconstruction from electron micrographs of tomato bushy stunt virus has been used to determine X-ray phases to 28 A resolution, by analogy with the single isomorphous replacement method of protein crystallography. An electron density map computed from X-ray amplitudes and these phases differs in two important respects from the electron micrograph reconstruction. The exclusion of stain from the 5-fold vertices, previously attributed to the presence of a minor protein, is shown to be an artifact of staining. The other difference involves positive staining of the RNA at the quasi-3-fold positions.

DNA sequence of the tail fibre genes 36 and 37 of bacteriophage T4

Abstract We present here the DNA sequence of the tail fibre genes 36 and 37 of bacteriophage T4. The products of these genes form the major part of the 800 A long distal half tail fibre of the phage. Restriction fragments of the DNA were subcloned in M13 phage and sequenced by the dideoxy sequencing method. A marker rescue technique was developed to allow rapid genetic identification of the particular T4 fragment carried by a given M13 clone, using the many T4 mutants available in the tail fibre region. This ensured little duplication in sequencing the M13 clones, and also allowed us to correlate the sequence with the genetic map. Predicted protein sequences are given for genes 36 and 37. Secondary structure prediction rules, when applied to the gene 37 protein, indicate the likely presence of regions of alternating β-strand and β-turn. This is consistent with previous structural analysis of the distal half fibre, which proposed an antiparallel β-structure running normal to the fibre axis, although the folding appears much less regular than anticipated.

The structure of the groups of nine hexons from adenovirus

Rotational filtering of images from electron micrographs of the groups of nine hexons from adenovirus shows that the so-called hexon has 3-fold symmetry. The structure units comprising the group of nine appear to be arranged on a p3 net.

The distal half of the tail fibre of bacteriophage T4: Rigidly linked domains and cross-β structure

Abstract The tail fibre of bacteriophage T4 is a structure about 1500 A long and 30 to 40 A thick with a kink in the middle dividing it into proximal and distal halves. We have studied the distal half fibre, which is the product of five phage genes. In crude lysates the half fibre forms ordered aggregates, which show characteristic banding patterns when examined in the electron microscope by negative staining. Analysis of the packing of half fibres in the various types of aggregate has enabled us to propose and test a model for the morphology of the half fibre. This consists of a series of rigidly linked stain-excluding globular domains, spaced regularly along the proximal 620 A of the half fibre with a thinner region about 195 A long at the distal tip of the fibre. X-ray diffraction and electron diffraction from unstained clumps of purified aggregates both give a strong ring at a spacing of 1 4.7 A − . In the electron diffraction pattern the ring shows marked orientation in a direction corresponding to the axis of the tail fibre, suggesting that it arises from a cross-β fold, in which the polypeptide chain runs transversely to the axis of the fibre.

Mutants of bacteriophage T4 that produce infective fibreless particles

Abstract In wild type bacteriophage T4 the long tail fibres serve both in the initial attachment of the phage to its host and in the triggering of tail contraction. A two-step model for the control of triggering suggests that particles lacking the product of gene 9, which are also structurally fibreless, might be infective. This is shown to be the case, even though such phage do not plate on restrictive strains of bacteria. However, starting from phage carrying an amber mutation in gene 9 it is easy to isolate additional mutations which, under restrictive conditions, permit fibreless plating (pfp mutations). Three such pfp mutations, having also a temperature-sensitive phenotype, have been isolated and shown to map in genes coding for structural components of the baseplate. The mode of action of these pfp mutations is not clear, though they certainly destabilize the baseplate, thereby making triggering easier. The pfp mutations are effective only when in combination with an amber mutation in gene 9 and not with amber mutations in tail fibre genes, establishing the essentially inhibitory nature of the control of triggering exercised by gene 9 product.