Small colloidal gold conjugated to Fab fragments or to immunoglobulin G as high-resolution labels for electron microscopy: a technical overview.

Fab-colloidal gold labelling in conjunction with negative staining and high-resolution electron microscopy was used for targeting single protein units in regular arrays. These were bacteriophage T4 polyheads with Fab-Au2.5, and a specific antibody binding site on the haemagglutinin polypeptide of influenza virus with Fab-Au3, Fab-Au2.5, and Fab-Au1-2. For the latter, IgG-Au3 was also used. Experimental details are summarized to provide generally applicable methods for the preparation of small gold colloids Fab-Au and of labelling. The putative mechanism of protein-gold complex formation and adsorption to preferred sites on Fab and IgG, most probably to sulphur-rich regions, is discussed. The influence of pH during complex formation was found to be of minor importance in the samples investigated. Reported experimental details and our own experiences suggest that the importance of a protein's pI relative to its optimum gold complexing pH critically depends on the nature of the protein in question rather than being of general importance for protein-gold complex stability.

Electron microscopy of the low pH structure of influenza virus haemagglutinin.

Influenza virus haemagglutinin mediates infection of cells by fusion of viral and endosomal membranes, triggered by low pH which induces a conformational change in the protein. We report studies of this change by electron microscopy, neutron scattering, sedimentation and photon correlation on X-31 (H3N2) haemagglutinin, both intact and bromelain cleaved, in various assemblies. HAs in all preparations showed a thinning at low pH, and a marked elongation which was removed on tryptic digestion, revealing altered features in the remaining stem portion of the molecule. A tentative model of the change is proposed, with reference to the known X-ray structure at neutral pH, in which major changes occur in the stem tertiary structure, while the top portion is only affected in its quaternary structure.

Reversible structure transition in gap junction under Ca++ control seen by high-resolution electron microscopy.

Deoxycholate-extracted rat liver gap junction was studied by high-resolution low-dose electron microscopy. Communicating channels between two adjoining cells supposedly form along the common axis of two apposed hexameric trans-membrane protein assemblies. These double hexamers are often arranged in large plaques on an ordered hexagonal net (8-9 nm lattice constant) and seem able to undergo structural alteration as a possible permeability control mechanism. Calcium is widely reported to uncouple gap junction, and we observed this alteration on exposure to Ca++ down to 10(-4) M concentration. When EGTA was added at matching concentrations, the alteration was reversible several times over one hour, but with considerable variability. It was imaged in the absence of any negative stain to avoid ionic and other complications. The resulting lack of contrast plus low-dose "shot" noise required digital Fourier filtering and reconstruction, but no detail was recovered below 1.8 nm. In other experiments with negative stain at neutral pH, gap junction connexons were apparently locked in the "closed" configuration and no transition could be induced. However, recovery of repeating detail to nearly 1.0 nm was possible, reproducibly showing a fine connective matrix between connexons . Whether this was formed by unfolded portions of the 28,000-dalton gap junction protein is not known, but its existence could explain the observed lattice invariance during the connexon structural transition.

Electron microscopy of influenza haemagglutinin-monoclonal antibody complexes.

The molecular locations of antibody binding sites on the haemagglutinin of influenza virus X-31 were investigated by electron microscopy of haemagglutinin-monoclonal antibody complexes. Evidence was obtained for different sites of binding of different antibodies and direct correspondence was observed between these sites and the locations of antigenic sites A, B, and E (D. C. Wiley, I. A. Wilson, and J. J. Skehel (1981). Nature (London) 289, 373-378) defined by determining the amino acids recognized by the specific antibodies.

Electron microscopic evidence for the axial rotation and inter-domain flexibility of the Fab regions of immunoglobulin G.

The Fab arms of immunoglobulin G (IgG) have long been known to hinge about their joint with the Fc subunit. Using monoclonal antibodies bound to influenza haemagglutinin (HA) as position markers, we now show that these arms can also rotate about their long axis with respect to Fc. We also show that when two IgGs are bound cyclically with two HA molecules, the arms can bend between the variable and constant domains to accommodate bond angle constraint.

Combining accurate defocus with low-dose imaging in high resolution electron microscopy of biological material.

High resolution (less than 2 nm) electron microscopy of biological specimens requires three exacting conditions to be met simultaneously: (a) fine specimen detail must be protected from destruction by the electron beam (low dose), (b) the electron optics must be adjusted to be capable of imaging that detail interpretably (accurate defocus), and (c) a suitable field of interest must be identified. We describe a method encompassing all three with an 80% success rate using only minor modifications to a transmission electron microscope, and no expensive on-line computing.

Evidence for a repeating cross-beta sheet structure in the adenovirus fibre.

The amino acid sequence of the adenovirus fibre protein reveals an approximately repeating motif of 15 residues. A diagonal comparison matrix established that these repeats extended from residue 43 to residue 400 of the 581 residue sequence. Assignment of secondary structure combined with model building showed that each 15-residue segment contained two short beta-strands and two beta-bends, one of which incorporated an extra residue in a beta-bulge of the Gx type. The 44 strands together gave a long (210 A) narrow, amphipathic beta-sheet, which could be stabilised by dimer formation to give the shaft of the fibre. The knob could arise from a dimer of the C-terminal 180 residue segment, predicted to be an 8-10 stranded beta-sandwich. This model is consistent with the electron micrographs of the fibre and it was supported by measurements of c.d. and of electron diffraction from microcrystals. The latter gave a pair of wide angle arcs, corresponding to a repeat of 4.7 A, oriented appropriately for a cross-beta structure. The relation of this structure to globular structures is discussed and a folding pathway is proposed. In its general features the structure resembles that proposed for the tail fibre of bacteriophage T4.

Multiple image integration: a new method in electron microscopy.

An optical 'flicker' method is described for the precise azimuthal and translational co-registration of many noisy but identical molecular images. Starting with a real micrograph of known biological objects showing no visible detail below 4 nm, a lattice of images of individual objects was synthesized by computer and translationally filtered, using real experimental data throughout. Detail was recovered conforming with known structural features of the object down to about 1.5 nm, and rotational analysis showed that the registration accuracy of the lattice elements was better than 0.5 nm on the object. Application to the straightening of real two-dimensional lattices with long-range distortion is discussed.

Transcarboxylase. An electron microscopic study of the enzyme with antibodies to biotin.

Antibodies prepared against biotin linked to serum albumin have been purified by affinity chromatography using lipoyl-Sepharose (Harmon, F. (1980) Anal. Biochem. 103, 58-63). These antibodies have been used to study the structure of the biotin-containing enzyme, transcarboxylase, by examination of complexes with the enzyme using electron microscopy. The antibodies are observed complexed with the outer subunit of transcarboxylase, which contains the biotin, but they also combine with the central subunit which does not contain biotin. Addition of excess biotin does not prevent combination of the antibodies with the central subunit. Such nonspecific combination does not occur with the outer subunit which has been treated with avidin-Sepharose to remove the biotinyl subunit. Large insoluble intermolecular enzyme complexes are not observed, apparently because the divalent antibodies react with the multiple sites on a single transcarboxylase molecule or because the larger complexes are unstable. Combination of the antibodies with the enzyme weakens the intersubunit linkages, thus, when grids are prepared for electron microscopy, there frequently is much greater dissociation of the enzyme which is complexed than there is with the noncomplexed enzyme. The results have confirmed that the biotinyl subunits remain complexed with the outer subunits when transcarboxylase dissociates.

A re-examination of the electron microscopic appearance of pyruvate carboxylase from chicken liver.

Electron microscopic studies of chicken liver pyruvate carboxylase conducted under a variety of conditions show that this enzyme has an overall rhombic appearance and is comprised of four nonspherical subunits. The square planar tetramers originally identified by Valentine et al. (Valentine, R.C., Wrigley, N.G., Scrutton, M.C., Irias, J.J., and Utter, M.F. (1966) Biochemistry 5, 3111-3116) as pyruvate carboxylase have been shown to represent a minor protein contaminant found in many impure preparations of this enzyme. The contaminating protein has been separated from pyruvate carboxylase and further purified. It does not contain biotin and its constituent polypeptides are smaller than those of pyruvate carboxylase. This protein, whose function is not yet identified, shows a strong tendency to aggregate and is highly visible under many conditions of electron microscopy. Several lines of evidence support the thesis that the nonsquare tetramers are pyruvate carboxylase. When essentially homogeneous material is examined with a variety of different negative stains, numbers of intact molecules represent 20 to 70% of the visible protein. These tetramers, like pyruvate carboxylase, are very cold-labile and are protected from dissociation under these conditions by acetyl-CoA, a specific activator of this enzyme. Also, the structures form complexes with avidin and antibiotin antibody and thus, like pyruvate carboxylase, contain biotin.

Electron microscopy of the large form of transcarboxylase with six attached subunits.

In this paper we show that the native form of transcarboxylase may be a species which has six rather than three subunits attached to the central subunit. We have designated this form as the 26 S enzyme. Electron micrographs support the view that the six subunits are attached in sets of three at the opposite faces of the central subunit, in contrast to the 18 S form in which all three subunits appear to be attached only at one face. In addition, evidence is presented that the dissociation of the 26 S to the 18 S form of transcarboxylase occurs with the loss of three subunits exclusively from one face of the central subunit. This result may indicate that the two faces of the central subunit differ structurally or there is negative cooperativity in the dissociation of subunits. The 26 S transcarboxylase, which was made by attachment of subunits to the 18 S enzyme or trypsin-treated 18 S enzyme was shown to have subunits on both faces of the central subunit. Treatment of the 26 S enzyme with carbodimide to cross-link the subunits to the central subunit and thus stabilize the structure resulted in improved electron micrographs. A model of the 26 S form of the enzyme is presented.