Effect of lysine modification on the stability and cellular binding of human amyloidogenic light chains.

AL amyloidosis is characterized by the pathologic deposition as fibrils of monoclonal light chains (i.e., Bence Jones proteins [BJPs]) in particular organs and tissues. This phenomenon has been attributed to the presence in amyloidogenic proteins of particular amino acids that cause these molecules to become unstable, as well as post-translational modifications and, in regard to the latter, we have investigated the effect of biotinylation of lysyl residues on cell binding. We utilized an experimental system designed to test if BJPs obtained from patients with AL amyloidosis or, as a control, multiple myeloma (MM), bound human fibroblasts and renal epithelial cells. As documented by fluorescence microscopy and ELISA, the amyloidogenic BJPs, as compared with MM components, bound preferentially and this reactivity increased significantly after chemical modification of their lysyl residues with sulfo-NHS-biotin. Further, based on tryptophan fluorescence and circular dichroism data, it was apparent that their conformation was altered, which we hypothesize exposed a binding site not accessible on the native protein. The results of our studies indicate that post-translational structural modifications of pathologic light chains can enhance their capacity for cellular interaction and thus may contribute to the pathogenesis of AL amyloidosis and multiple myeloma.

Structure and molecular weight of the dynein ATPase.

Dynein has been examined by scanning transmission electron microscopy (STEM). Samples of 30S dynein from tetrahymena cilia were applied to carbon films and either were freeze- dried and examined as unstained, unfixed specimens, or were negatively stained with uranyl sulfate. A totally new image of the dynein molecule was revealed showing three globular heads connected by three separate strands to a common base. Two of the heads appeared to be identical and exhibited a diameter of 10 nm while the third head was somewhat larger (approximately 12 nm). The overall length of the particle was 35 nm. Mass analysis, based upon the integration of electron scattering intensities for unstained particles, gave a molecular weight of 1.95 (+/-)0.24) megadaltons. Mass per unit length analysis was performed using bovine brain microtubules decorated with dynein under conditions where the dynein shows a linear repeat of 24 nm with seven dynein molecules surrounding a microtubule made up of 14 protofilaments. Undecorated microtubules gave a molecular weight per unit length of 21,000+/-1,900 daltons/A, compared to a value of 84,400+/-2,200 daltons/A for the fully decorated microtubules. Taken together, these data give a molecular weight of 2.17 (+/- 0.14) megadaltons per dynein molecule, in agreement with measurements on the isolated particles. Mass analysis of individual globular heads observed in isolated particles gave a molecular weight distribution with a mean of 416+/- 76 kdaltons. These data could also be viewed as the sum of two populations of head with two-thirds of the heads at approximately 400 kdaltons and one-third at approximately 550 kdaltons, although more precise data will be required to distinguish two classes of heads with confidence. The mass of the dynein-microtubule complex as a function of distance from the midline of the particle was analysed to distinguish which end of the dynein molecule was bound to the microtubule. The projected mass distribution was consistent with a model where the three dynein heads were oriented toward the microtubule and clearly not consistent with the opposite orientation. These data indicate that the three globular heads form the ATP-sensitive site in this heterologous dynein-microtubule system and suggest that the rootlike base of the dynein molecule forms the structural attachment site to the A-subfiber of the outer doublet in cilia and flagella. The structure and function of the dynein are dicussed in terms of these new results.

Structure and mass analysis of 14S dynein obtained from Tetrahymena cilia.

Scanning transmission electron microscopic analysis revealed that the 14S fraction of Tetrahymena dynein was of a mixture of two types of particles in approximately equal proportions. The 14S dynein molecules were roughly ellipsoid in shape with approximate axes of 9.5 and 14.5 nm. About half of the particles had tails 20-24-nm long. By the integration of electron scattering intensities, particles with tails had an average mass of 510 kD with a SD of 90 kD. The globular heads of both types of particles had an average mass of 330 kD with a SD of 60 kD. The mass of the tail structure was about 180 kD. By SDS-PAGE, the 14S dynein consisted of two high molecular mass polypeptides above 300 kD that could be distinguished by immunoblot analysis.

Mass distributions of coated vesicles isolated from liver and brain: analysis by scanning transmission electron microscopy.

Populations of coated vesicles purified from bovine brain (BCV) and from rat liver (LCV) have been characterized with respect to the parameters of mass and diameter by analysis of scanning transmission electron micrographs of unstained specimens. Coated vesicles from both sources are heterogeneous, particularly in their masses. The respective distributions, compiled from mass measurements of many individual particles, are complex and markedly different. BCV range from 20 Mdaltons to approximately 100 Mdaltons with a weighted average of 35 Mdaltons: most BCV (80%) lie between 20 and 40 Mdaltons, including peaks at approximately 26 Mdaltons and at approximately 34 Mdaltons. In contrast, LCV masses tend to be substantially higher, ranging from 20 to 220 Mdaltons with a weighted average of 66 Mdaltons. There is a prominent subpopulation at approximately 35 Mdaltons, and 59% of all LCV belong to a broad peak between 50 and 120 Mdaltons. The Kolmogorov-Smirnov distribution-free test was used to affirm the statistical reproducibility of these isolates. BCV diameters vary from 50 to 90 nm, and those of LCV from 50 to 150 nm. Both protein compositions, determined by SDS PAGE, are dominated by clathrin and they are generally similar, except that corresponding secondary bands, notably the clathrin-associated light chains, appear to have lower molecular weights in the case of LCV. From consideration of the joint mass-diameter distribution, it is apparent that coated vesicles of a given diameter vary considerably in mass and that this variation is due primarily to widely differing amounts of material enclosed within the clathrin coat.

Structure of L-A virus: a specialized compartment for the transcription and replication of double-stranded RNA.

The genomes of double-stranded (ds)RNA viruses are never exposed to the cytoplasm but are confined to and replicated from a specialized protein-bound compartment-the viral capsid. We have used cryoelectron microscopy and three-dimensional image reconstruction to study this compartment in the case of L-A, a yeast virus whose capsid consists of 60 asymmetric dimers of Gag protein (76 kD). At 16-A resolution, we distinguish multiple domains in the elongated Gag subunits, whose nonequivalent packing is reflected in subtly different morphologies of the two protomers. Small holes, 10-15 A across, perforate the capsid wall, which functions as a molecular sieve, allowing the exit of transcripts and the influx of metabolites, while retaining dsRNA and excluding degradative enzymes. Scanning transmission electron microscope measurements of mass-per-unit length suggest that L-A RNA is an A-form duplex, and that RNA filaments emanating from disrupted virions often consist of two or more closely associated duplexes. Nuclease protection experiments confirm that the genome is entirely sequestered inside full capsids, but it is packed relatively loosely; in L-A, the center-to-center spacing between duplexes is 40-45 A, compared with 25-30 A in other double-stranded viruses. The looser packing of L-A RNA allows for maneuverability in the crowded capsid interior, in which the genome (in both replication and transcription) must be translocated sequentially past the polymerase immobilized on the inner capsid wall.

Solution and surface effects on plasma fibronectin structure.

As assessed by electron microscopy, the reported shape of the plasma fibronectin molecule ranges from that of a compact particle to an elongated, rod-like structure. In this study, we evaluated the effects of solution and surface conditions on fibronectin shape. Freeze-dried, unstained human plasma fibronectin molecules deposited at pH 7.0-7.4 onto carbon films and examined by scanning transmission electron microscopy appeared relatively compact and pleiomorphic, with approximate average dimensions of 24 nm X 16 nm. Negatively stained molecules also had a similar shape but revealed greater detail in that we observed irregular, yarn-like structures. Glutaraldehyde-induced intramolecular cross-linking did not alter the appearance of plasma fibronectin. Molecules deposited at pH 2.8, pH 9.3, or after succinylation were less compact than those deposited at neutral pH. In contrast, fibronectin molecules sprayed onto mica surfaces at pH 7, rotary shadowed, and examined by transmission electron microscopy were elongated and nodular with a contour length of 120-130 nm. Sedimentation velocity experiments and electron microscopic observations indicate that fibronectin unfolds when it is succinylated, when the ionic strength is raised at pH 7, or when the pH is adjusted to 9.3 or 2.8. Greater unfolding is observed at pH 2.8 at low ionic strength (less than 0.01) compared with material at that pH in 0.15 M NaCl solution. We conclude that (a) the shape assumed by the fibronectin molecule can be strongly affected by solution conditions and by deposition onto certain surfaces; and that (b) the images of fibronectin seen by scanning transmission electron microscopy at neutral pH on carbon film are representative of molecules in physiologic solution.

Epidermal keratin filaments assembled in vitro have masses-per-unit-length that scale according to average subunit mass: structural basis for homologous packing of subunits in intermediate filaments.

We have used scanning transmission electron microscopy to elucidate the question of how intermediate filament (IF) subunits of widely differing mass can all form morphologically similar IF. From scanning transmission electron micrographs, the distributions of mass were determined for three types of epidermal keratin IF reassembled in vitro from mixtures of subunits with substantially different masses, viz., "light" and "heavy" human keratins with [Mr] = 50,000 and 56,000, respectively, and mouse keratins of [Mr] = 63,000. Their principal assembly products were found to average 22, 25, and 29 kdalton/nm, respectively. These densities, which correspond to immature "minimal form" IF (Steven, A. C., J. Wall, J. Hainfeld, and P. M. Steinert, 1982, Proc. Natl. Acad. Sci. USA., 79:3101-3105), are directly proportional to the average subunit masses. The human keratin IF (but not those of mouse) also contained minor amounts (15-20%) of more massive polymers averaging 33 and 35 kdalton/nm, respectively, which probably represent mature IF. Taken together with earlier results on IF of other subclasses, these results indicate that the average linear density of IF scales according to the average mass of their constituent subunits, both for "minimal form" and for mature IF. As underlying mechanism for this homology, we propose that the fundamental building-blocks of all these IF contain a common structural element whose packing within the various IF is likewise conserved and which specifies the overall structure. The variable amounts of mass in the nonconserved moieties account for the observed proportionality. This scheme fits with amino acid sequence data for several IF subunits that have revealed, as a likely candidate for the common element, an essentially conserved alpha-helical domain, contrasting with the highly variable sequences of their non-alpha-helical terminal domains.

Biospecific labeling with undecagold: visualization of the biotin-binding site on avidin.

The biotin-binding site on avidin has been labeled with biotin conjugated to undecagold, an organometallic cluster compound containing 11 gold atoms in a core angestroms in diameter. Examination of unstained specimens by scanning transmission electron microscopy reveals the labeled sites directly, without computational averaging or filtering of the images. This approach should be widely applicable for determining the locations of subunits and functional site in biological macromolecules at a resolution at a resolution in range of 15 angstroms.

Visualization of polymercurimethane-labeled for bacteriophage in the scanning transmission electron microscope.

Each of the 2700 coat proteins of fd bacteriophage was labeled with tetrakis(acetoxymercuri)methane (TAMM) or aquoglycylmethionineplatinum(II). The TAMM-labeled specimens reveal striking bright spots in the scanning transmission electron microscope which arise from clustering. Measurements of mass show increases consistent with the addition of four mercury atoms or one platinum atom, respectively, to each coat protein.

The role of fibrinogen D domain intermolecular association sites in the polymerization of fibrin and fibrinogen Tokyo II (gamma 275 Arg-->Cys).

Intermolecular end-to-middle domain pairing between a thrombin-exposed 'A' polymerization site in the central 'E' domain of fibrin, and a constitutive complementary 'a' site in each outer 'D' domain ('D:E'), is necessary but not alone sufficient for normal fibrin assembly, as judged from previous studies of a congenital dysfibrinogen, Tokyo II (gamma 275 arg-->cys), which showed defective fibrin clot assembly and a normal D:E interaction (Matsuda, M., M. Baba, K. Morimoto, and C. Nakamikawa, 1983. J. Clin. Invest. 72:1034-1041). In addition to the 'a' polymerization site, two other constitutive intermolecular association sites on fibrinogen D domains have been defined: between gamma chain regions containing the carboxy-terminal factor XIIIa crosslinking site ('gamma XL:gamma XL'); and between sites located at the outer ends of each molecule ('D:D') (Mosesson, M. W., K. R. Siebenlist, J. F. Hainfeld, and J. S. Wall, manuscript submitted for publication). We evaluated the function of these sites in Tokyo II fibrinogen, and confirmed that there was a normal fibrin D:E interaction, as determined from a normal fibrin crosslinking rate in the presence of factor XIIIa. We also found a normal gamma XL: gamma XL interaction, as assessed by a normal fibrinogen crosslinking rate. Judging from electron microscopic images, factor XIIIa-crosslinked Tokyo II fibrinogen failed to form elongated double-stranded fibrils like normal fibrinogen. Instead, it formed aggregated disordered collections of molecules, with occasional short fibrillar segments. In addition, Tokyo II fibrin formed an abnormal, extensively branched clot network containing many tapered terminating fibers. These findings indicate that the Tokyo II fibrinogen defect results in a functionally abnormal D:D self-association site, and that a normal D:D site interaction is required, in addition to D:E, for normal fibrin or fibrinogen assembly.

The relationship between the fibrinogen D domain self-association/cross-linking site (gammaXL) and the fibrinogen Dusart abnormality (Aalpha R554C-albumin): clues to thrombophilia in the "Dusart syndrome".

Cross-linking of fibrinogen at its COOH-terminal gamma chain cross-linking site occurs in the presence of factor XIIIa due to self-association at a constitutive D domain site ("gammaXL"). We investigated the contribution of COOH-terminal regions of fibrinogen Aalpha chains to the gammaXL site by comparing the gamma chain cross-linking rate of intact fibrinogen (fraction I-2) with that of plasma fraction I-9, plasmic fraction I-9D, and plasmic fragment D1, which lack COOH-terminal Aalpha chain regions comprising approximately 100, approximately 390, and 413 residues, respectively. The cross-linking rates were I-2 > I-9 > 1-9D = D1, and indicated that the terminal 100 or more Aalpha chain residues enhance gammaXL site association. Fibrinogen Dusart, whose structural abnormality is in the COOH-terminal "alphaC" region of its Aalpha chain (Aalpha R554C-albumin), is associated with thrombophilia ("Dusart Syndrome"), and is characterized functionally by defective fibrin polymerization and clot structure, and reduced plasminogen binding and tPA-induced fibrinolysis. In the presence of XIIIa, the Dusart fibrinogen gamma chain cross-linking rate was about twice that of normal, but was normalized in proteolytic fibrinogen derivatives lacking the Aalpha chain abnormality, as was reduced plasminogen binding. Electron microscopy showed that albumin-bound Dusart fibrinogen "alphaC" regions were located in the vicinity of D domains, rather than at their expected tethered location near the fibrinogen E domain. In addition, there was considerable fibrinogen aggregation that was attributable to increased intermolecular COOH-terminal Aalpha chain associations promoted by untethered Dusart fibrinogen aC domains. We conclude that enhanced Dusart fibrinogen self-assembly is mediated through its abnormal alphaC domains, leads to increased gammaXL self-association and gamma chain cross-linking potential, and contributes to the thrombophilia that characterizes the "Dusart Syndrome."

Structural model of porcine factor VIII and factor VIIIa molecules based on scanning transmission electron microscope (STEM) images and STEM mass analysis.

Porcine plasma factor VIII (fVIII) molecules are heterodimers composed of a 76,000-mol wt light chain (-A3-C1-C2) and a heavy chain ranging in molecular weight from 82,000 (A1-A2) to 166,000 (A1-A2-B). Proteolytic activation of fVIII by thrombin results in fVIIIa heterotrimers lacking B domains (A1, A2, A3-C1-C2). In this study, immunoaffinity purified fVIII was further fractionated by mono S or mono Q chromatography to prepare heterodimers containing a light chain and an A1-A2-B heavy chain (fVIII 166/76) or an A1-A2 heavy chain (fVIII 82/76). Mass analysis of scanning transmission electron microscopic (STEM) images of fVIII 166/76 indicated that heterodimers (mass 237 +/- 20 kD) had irregularly globular core structures 10-12 nm across, and frequently displayed a diffuse, occasionally globular to ovoid satellite structure extending 5-14 nm from the core, and attached to it by a thin stalk. Factor VIII 82/76 molecules (mass 176 +/- 20 kD) had the same core structures as fVIII 166/76 molecules, but lacked the satellite structure. These findings indicate that A1-A2 domains of heavy chains and the light chains of the fVIII procofactor molecule are closely associated and constitute the globular core structure, whereas the B domainal portion of heavy chains comprises the peripheral satellite appendage. Factor VIII core structures commonly displayed a finger-like projection near the origin of the B domainal stalk that was also a consistent feature of the free heavy chains (mass 128-162 kD) found in fVIII 166/76 preparations. Factor VIII light chain monomers (mass, 76 +/- 16 kD) were globular to c-shaped particles 6-8 nm across. These chains commonly possessed a v-shaped projection originating from its middle region, that could also be observed at the periphery of fVIII core molecules. Factor VIIIa preparations contained heterotrimers (mass 162 +/- 13 kD) that had the same dimensions as fVIII core structures, lacked the B domainal appendage, and sometimes possessed the same core features as fVIII molecules. Molecular species corresponding to heterodimers (mass, 128 +/- 13 kD) and unassociated subunit chains (40-100 kD) were also observed in fVIIIa preparations, suggesting that heterotrimers have an appreciable tendency to dissociate, a phenomenon that could explain the decay of fVIIIa activity after thrombin activation of fVIII.

Molecular weights of wheat gamma2-, beta6-, alpha7-, alpha8- and alpha9-gliadins.

The molecular weights of wheat gamma2-, beta6-, alpha7-, alpha8- and alpha9-gliadins were calculated with the aid of a computer technique from sedimentation equilibrium data obtained in an ultracentrifuge equipped with photoelectric scanner. The dissociative solvents, all at pH 3.1 by addition of HCl, included 3 M urea, 0.15 M KCl; 8 M urea, 0.15 M KCl and 6 M guanidine-HCl. The minimum molecular weights for gamma2-, alpha7- and alpha9-gliadins, obtained in 6 M guanidine-HCl, were 34 600, 30 400 and 30 900, respectively. The beta6- and alpha8-gliadins gave minimum molecular weights of 33 000 and 36 900, respectively, in 3 M urea, 0.15 M KCl.

Membrane contents of distinct subpopulations of coated vesicles determined by scanning transmission electron microscopy.

In order to investigate the heterogeneity of clathrin-coated vesicles purified from rat liver, and to quantitate rigorously their membrane contents, we have analyzed scanning transmission electron micrographs of unstained coated vesicles before and after extraction with the non-ionic detergent Triton X-100, as well as of vesicles whose coats had been removed by dialysis against 10 mM or 100 mM Tris (pH 8.2). Their respective distributions of particle masses were thus determined and compared, in light of complementary biochemical quantitations of lipid and protein. Smaller coated particles, 25-45 MDa in mass and 60-80 nm in diameter, lose no mass when extracted with Triton, and disappear when their coats are dissociated. We conclude that they do not contain membrane vesicles, although they have dense, presumably proteinaceous, cores. They may represent particles generated during tissue homogenization or, possibly, a storage form of clathrin. The remaining 70% contain bona fide vesicles: these particles are 75-150 nm in diameter, and their average mass is about 80 MDa, of which 48 MDa is contributed by coat proteins, 10-12 MDa by phospholipid and cholesterol, and 20-22 MDa by vesicle-associated proteins. Their vesicles are of two types: smaller, denser, vesicles that contain substantial amounts of internalized material, and larger, less dense, vesicles that do not. The distinction between them may, in view of other findings, reflect a difference between coated vesicles derived respectively from the Golgi and the plasma membrane.

The molecular structure of agrobacterium VirE2-single stranded DNA complexes involved in nuclear import.

Nuclear import of DNA is a central event in genetic transformation of plant cells by Agrobacterium tumefaciens. Agrobacterium elicits tumors on plant hosts by transporting a single-stranded (ss) copy of the bacterial transferred DNA (T-DNA) from its Ti (tumor-inducing) plasmid into the plant cell nucleus. Presumably, the process of T-DNA nuclear import is mediated by two agrobacterium proteins, VirD2 and VirE2, which are thought to directly associate with the transported T-DNA. Both proteins have been shown to contain functional nuclear localizations signals (NLS). Recently, VirE2 alone has been shown to actively transport ssDNA into the plant cell nucleus. To understand the process of DNA nuclear import, it is important to know the structure of the transport intermediate. To this end, complexes of VirE2 and ssDNA were analyzed by scanning transmission electron microscopy (STEM). This analysis suggests that VirE2 packages ssDNA into semi-rigid, hollow cylindrical filaments with a telephone cord-like coiled structure. The outer diameter of these complexes is too large to enter the nucleus by diffusion but is within the size exclusion limits of the active nuclear import. Detailed mass analysis of VirE2-ssDNA filaments is presented and a structural model is proposed.

Visualization of protein-nucleic acid interactions involved in the in vitro assembly of the Escherichia coli 50 S ribosomal subunit.

Protein-nucleic acid interactions which occur during Escherichia coli 50 S ribosomal subunit assembly between 23 S rRNA, 5 S rRNA and a complete set of 34 L-proteins were monitored by high resolution scanning transmission electron microscopy (STEM). This approach made it possible to visualize and quantitatively analyze conformational changes induced in the rRNAs during E. coli 50 S ribosomal subunit assembly. The reconstituted RNA-protein complexes, the control 23 S rRNA and native 50 S subunits were characterized by their mass and morphology. Association of 23 S rRNA with the first assembly protein, L24, led to the formation of a distinct nucleus of mass ("cluster") on the filamentous and loosely coiled molecule of the 23 S rRNA. This structural feature was preserved and enhanced in 23 S rRNA after its association with the rest of the early assembly proteins, namely L3, L20, L13, L4 and L22. Since the above proteins, with the exception of L3, bind to the 5' end of the 23 S rRNA, the cluster seems to be formed predominantly by interactions of L24, L13, L20, L22 and L4 with this segment of the 23 S rRNA molecule. Association with the rest of the primary binding proteins (L2, L23, L9, L1), which interact with the 3' end of the 23 S rRNA, appears to result in the formation of a second mass center. Binding of additional proteins led to the formation of compact particles with an apparent similarity to the 50 S subunit. However, particles with defined structural features characteristic of the native 50 S subunit requires the interaction of both 23 S rRNA and 5 S rRNA with all of the L-proteins. STEM image analysis demonstrated that 50 S subunit reconstitution proceeds by the immediate folding of the 23 S rRNA into a single mass center followed by the formation of a second mass center. These mass centers merge into one central body, which is gradually enhanced and decorated with structural elements characteristic of the 50 S subunit in the latter stages of assembly.

Protein-induced conformational changes in 16 S ribosomal RNA during the initial assembly steps of the Escherichia coli 30 S ribosomal subunit.

The mechanism of 16 S ribosomal RNA folding into its compact form in the native 30 S ribosomal subunit of Escherichia coli was studied by scanning transmission electron microscopy and circular dichroism spectroscopy. This approach made it possible to visualize and quantitatively analyze the conformational changes induced in 16 S rRNA under various ionic conditions and to characterize the interactions of ribosomal proteins S4, S8, S15, S20, S17 and S7, the six proteins known to bind to 16 S rRNA in the initial assembly steps. 16 S rRNA and the reconstituted RNA-protein core particles were characterized by their mass, morphology, radii of gyration (RG), and the extent and stability of 16 S rRNA secondary structure. The stepwise binding of S4, S8 and S15 led to a corresponding increase of mass and was accompanied by increased folding of 16 S rRNA in the core particles, as evident from the electron micrographs and from the decrease of RG values from 114 A and 91 A. Although the binding of S20, S17 and S7 continued the trend of mass increase, the RG values of these core particles showed a variable trend. While there was a slight increase in the RG value of the S20 core particles to 94 A, the RG value remained unchanged (94 A) with the further addition of S17. With subsequent addition of S7 to the core particles, the RG values showed an increase to 108 A. Association with S7 led to the formation of a globular mass cluster with a diameter of about 115 A and a mass of about 300 kDa. The rest of the mass (about 330 kDa) remained loosely coiled, giving the core particle a "medusa-like" appearance. Morphology of the 16 S rRNA and 16 S rRNA-protein core particles, even those with all six proteins, does not resemble the native 30 S subunit, contrary to what has been reported by others. The circular dichroism spectra of the 16 S rRNA-protein complexes and of free 16 S rRNA indicate a similarity of RNA secondary structure in the core particles with the first four proteins, S4, S8, S15, S20. The circular dichroism melting profiles of these core particles show only insignificant variations, implying no obvious changes in the distribution or the stability of the helical segments of 16 S rRNA. However, subsequent binding of proteins S17 and S7 affected both the extent and the thermal stability of 16 S rRNA secondary structure.(ABSTRACT TRUNCATED AT 400 WORDS)

Structure of haptoglobin and the haptoglobin-hemoglobin complex by electron microscopy.

The human serum protein, haptoglobin, forms a stable, irreversible complex with hemoglobin. Haptoglobin is composed of two H chains, which are connected via two smaller L chains to give a protein of 85,000 Mr. In the complex, each H chain binds an alpha beta dimer of hemoglobin for a total molecular weight of 150,000. The scanning transmission electron microscope has been used to derive new information about the shape and structure of haptoglobin and hemoglobin, and about their relative orientation in the complex. The micrographs of negatively stained images show that haptoglobin has the shape of a barbell with two spherical head groups, which are the H chains. These are connected by a thin filament with a central knob, which corresponds to the L chains. The overall length of the molecule is about 124(+/- 8) A and the interhead distance is 87 (+/- 7) A. In the haptoglobin-hemoglobin complex, the head groups are ellipsoidal and under optimal staining conditions bilobal . Thus, the alpha beta dimers are binding to the H chains, but off the long axis of the barbell by 127 degrees in a trans configuration. This angle considerably restricts the region on the surface of the H chain structure that can contain the hemoglobin binding site. The interhead group distance for complex is 116.5(+/- 6.3) A or 30 A greater than for haptoglobin. The N terminus of the beta chain was located on the trans off-axis configured barbell structure of complex by using a hemoglobin that was crosslinked between the alpha beta dimers in the region of the beta N terminus. The distances and angles that are measured on the micrographs for the native and crosslinked complex molecules permit the directions of two of the alpha beta dimer ellipsoid axes to be assigned. Taken together, these data provide an approximate relative orientation for the binding of the alpha beta dimer to the H chain of haptoglobin.

Structure and assembly of haptoglobin polymers by electron microscopy.

Haptoglobin (Hp) consists of light (L) and heavy (H) chains, the latter of which combine with hemoglobin alpha beta dimers to form a highly stable complex. Human haptoglobin assembles as HL units that occur in two allelic forms; HL1 , which is monovalent, and HL2 , which is divalent. As a result, three phenotypic forms exist in the human population: Hp1-1, the homozygous form in which the monovalent HL1 unit occurs as a dimer; Hp2-2, the homozygous form of the divalent HL2 unit, which gives a series of polymers; and the heterozygous Hp2-1 form, which gives a different series of polymers. We have investigated the structures and assembly properties of these two haptoglobin polymeric series in their complexes with hemoglobin using high-resolution scanning transmission electron microscopy. Polymers of complex are composed of ellipsoidal or bilobal head groups, which are the H alpha beta subunits connected by thin filament-like structures, which are the L chains. Polymers of size up to pentamers can be identified easily by counting the number of head groups in the molecule. Complex 2-1 and complex 2-2 trimers were studied extensively. The differences in detailed morphology show that while the 2-1 trimer is a linear polymer, the 2-2 trimer is a closed circular molecule. The micrograph images suggest that complex 2-2 tetramers and pentamers, and perhaps higher forms may also be cyclic. The structure of the L2 subunit of haptoglobin is shown to be composed of two domains, which may be similar in structure to the single domain of the monovalent L1 chain. The two L2 domains are connected by a hinge that has quite limited flexibility. Using these structural models, assembly characteristics and structural properties of the trimers and tetramers of complex 2-1 and complex 2-2 are described.

Stoichiometry and domainal organization of the long tail-fiber of bacteriophage T4: a hinged viral adhesin.

The long-tail fibers (LTFs) form part of bacteriophage T4's apparatus for host cell recognition and infection, being responsible for its initial attachment to susceptible bacteria. The LTF has two parts, each approximately 70 to 75 nm long; gp34 (140 kDa) forms the proximal half-fiber, while the distal half-fiber is composed of gp37 (109 kDa), gp36(23 kDa) and gp35 (30 kDa). LTFs have long been thought to be dimers of gp34, gp37 and gp36, with one copy of gp35. We have used mass mapping by scanning transmission electron microscopy (STEM), quantitative SDS-PAGE, and computational sequence analysis to study the structures of purified LTFs and half-fibers of both kinds. These data establish that the LTF is, in fact, trimeric, with a stoichiometry of gp34: gp37: gp36: gp35 = 3:3:3:1. Averaged images of stained and unstained molecules resolve the LTF into a linear stack of 17 domains. At the proximal end is a globular domain of approximately 145 kDa that becomes incorporated into the baseplate. It is followed by a rod-like shaft (33 x 4 mm; 151 kDa) which correlates with a cluster of seven quasi repeats, each 34 to 39 residues long. The proximal half-fiber terminates in three globular domains. The distal half-fiber consists of ten globular domains of variable size and spacing, preceding a needle-like end domain (15 x 2.5 nm; 31 kDa). The LTF is rigid apart from hinges between the two most proximal domains, and between the proximal and distal half-fibers. The latter hinge occurs at a site of local non-equivalence (the "kneecap") at which density, correlated with the presence of gp35, bulges asymmetrically out on one side. Several observations indicate that gp34 participates in the sharing of conserved structural modules among coliphage tail-fiber genes to which gp37 was previously noted to subscribe. Two adjacent globular domains in the proximal half-fiber match a pair of domains in the distal half-fiber, and the rod domain in the proximal half-fiber resembles a similar domain in the T4 short tail-fiber (gp12). Finally, possible structures are considered; combining our data with earlier observations, the most likely conformation for most of the LTF is a three-stranded beta-helix.