The structure of the adenovirus capsid. III. Hexon packing determined from electron micrographs of capsid fragments.

The orientation and relative positions of all 240 hexons in the icosahedral outer capsid of adenovirus have been determined. Two types of capsid fragments, obtained after selective disruption of the virion, were analyzed using electron microscopy and image-processing techniques. Planar inverted groups-of-nine, arising from the central region of the capsid facet, were minimally stained to reveal the morphology of restricted regions of their component hexons. Images shown to be related by correspondence analysis were averaged and features of the individual hexon molecule, known from an X-ray crystallographic investigation, were used in their interpretation. The study confirms earlier observations that the hexons in the group-of-nine are distributed on a p3 net, shows that the hexons form a close-packed array using the pseudo-hexagonal shape of the hexon base, and provides their relative positions. Twenty interlocking groups-of-nine account for 180 of the 240 hexons present in the viral capsid. The orientation of the remaining 60 peripentonal hexons was obtained from a rotationally averaged image of a quarter-capsid, a novel viral fragment comprising five complete facets. Each peripentonal hexon forms planar asymmetric interactions with two neighbors in an adjacent group-of-nine so that it lies on an extension of the p3 net. The complete facet thus consists of 12 hexons arranged on a planar p3 net, with a shape that permits interlocking of hexons at the capsid edge. The relative positions of the hexons have been determined to within 5 A using the molecular model, and indicate that the pseudo-hexagonal basal regions are close-packed in a manner that maximizes the hexon-hexon contacts. The results confirm the model proposed earlier for the arrangement of hexons within the adenovirus capsid (Burnett, 1985), and show the power of the inter-disciplinary approach.

Human urokinase contains GalNAc beta (1-4)[Fuc alpha (1-3)]GlcNAc beta (1-2) as a novel terminal element in N-linked carbohydrate chains.

Structural analysis of enzymically released N-linked carbohydrate chains of human urokinase (urinary-type plasminogen activator) by 1H NMR spectroscopy and FAB-MS demonstrated that the N-linked oligosaccharides on the only N-glycosylation site contain diantennary structures with the novel GalNAc beta (1-4) [Fuc alpha (1-3)]GlcNAc beta (1-2) element in the upper or the lower branch.

Unique chemical reactivity of His-21 of CRM-197, a mutated diphtheria toxin.

CRM-197 is a mutated diphtheria toxin (63000 Da) widely used as a carrier protein of conjugated vaccines. Among the 14 histidines of CRM-197, His-21 was found to be modified selectively with iodoacetamide based reagents. This finding suggests a simplified method for the preparation of conjugate vaccines crosslinked to CRM-197. A bifunctional iodoacetamide, N,N'-(2-hydroxy-1,3-propanediyl)-bis-12-iodoacetamidel (I-CH2-CONH-CH2-CH(OH)-CH2-NHCO-CH2-I) (HPBIA), was synthesized and allowed to react with CRM-197. In the alkaline buffer of pH 8.0-8.4, HPBIA was shown to react and intra-bridge His-21 and Lys-24 of CRM-197 sequentially. At lower pH (7.1-7.5) in the phosphate buffer, the reactivity of Lys-24 toward HPBIA was suppressed drastically. Under these conditions, His-21 could be specifically labeled with HPBIA. Initial experiments have demonstrated that HPBIA modified CRM-197 is able to crosslink to a cysteine-containing peptide. These results offer a potential route for improving the homogeneity of CRM-197 based protein-peptide as well as protein-polysaccharide conjugates.

Sialylated carbohydrate chains of recombinant human glycoproteins expressed in Chinese hamster ovary cells contain traces of N-glycolylneuraminic acid.

HPLC analysis of sialic acids released from recombinant variants of human tissue plasminogen activator, human chimeric plasminogen activator, human erythropoietin, and human follitropin, expressed in Chinese hamster ovary cells, demonstrates for each glycoprotein the presence of N-acetylneuraminic acid and N-glycolylneuraminic acid in a ratio of 97:3. Structural analysis by 500 MHz1H-NMR spectroscopy, of the enzymatically released N-linked carbohydrate chains of chimeric plasminogen activator and of erythropoietin, showed that alpha 2-3 linked N-glycolylneuraminic acid can occur in different N-acetyllactosamine type antennary structures.

Fluoroaluminate stimulates phosphorylation of p130 Cas and Fak and increases attachment and spreading of preosteoblastic MC3T3-E1 cells.

Fluoroaluminate is a G-protein activator, it stimulates osteoblastic cells in culture, and is a bone-forming agent in vivo. To elucidate the mechanisms of G-protein-mediated action of fluoroaluminate in osteoblasts, we studied protein tyrosine phosphorylation in the preosteoblastic cell line MC3T3-E1. Fluoroaluminate, lysophosphatidic acid (LPA; an agonist for G-protein-coupled receptor), or adhesion to type I collagen all stimulated phosphorylation of a similar set of proteins, including p130, p120, p110 (previously identified as proline-rich tyrosine kinase 2, Pyk2), and p70. The phosphorylation of these proteins was sensitive to an Src inhibitor, but not to a Gi-protein inactivator, pertussis toxin. By purification/mass spectrometry and by immunodepletion, p130 protein was identified as p130 Cas (Crk-associated protein), a Src substrate and a protein involved in signaling by cell-adhesion receptors, integrins. Phosphorylation of immunoprecipitated p130 Cas increased upon stimulation with fluoroaluminate and with agonists of G-protein-coupled receptors, but not with growth factors. By immunodepletion, the p120 protein was identified as focal adhesion kinase, Fak. The addition of fluoroaluminate during cell attachment to type I collagen further stimulated phosphorylation of p130 Cas and of Fak. Simultaneously, fluoroaluminate increased the number of attached MC3T3-E1 cells and their spreading. These novel aspects of fluoroaluminate action in cell culture may be important for the bone-forming action of fluoroaluminate in vivo.

Monoclonal antibody based enzyme-linked and chemiluminescent assays for the human interleukin-1 receptor antagonist. Application to measure hIL-1ra levels in monocyte cultures and synovial fluids.

Interleukin-1 receptor antagonist (IL-1ra) has the potential to counteract at least part of the biological effects of interleukin-1. The outcome of an inflammatory reaction may therefore be determined by the balance between IL-1 and IL-1ra, rather than by IL-1 alone. We have developed an immunoassay to address this issue as well as to assess the effects of anti-inflammatory agents on the expression of IL-1 and IL-1ra in vitro or in body fluids. Recombinant human IL-1ra was expressed in an E. coli system, purified to homogeneity, and used to derive monoclonal antibodies in mice as well as polyclonal antibodies in rabbits. A sandwich ELISA was constructed with F(ab')2 fragments of a high affinity monoclonal antibody and the rabbit serum as a source of secondary antibody. The assay required no sample treatment to avoid interference by rheumatoid factor. The measuring range was 0.020-2 ng/ml. By labelling a second monoclonal antibody with an acridinium ester, a chemiluminescence assay with a wider measuring range (0.050-15 ng/ml) was generated. In accord with published data, we found that IL-1ra was secreted by human monocytes stimulated with LPS, Zymosan, IL-1 alpha, or human IgG. After an induction phase of ca. 4 hours and depending on the stimulus, IL-1ra accumulated linearly for periods up to 96 h. IL-1ra levels in synovial fluids of 19 patients suffering from various inflammatory joint diseases were compared with the cytokine levels of IL-1 beta, IL-6, IL-8, and TNF-alpha. Highest positive correlations were found with IL-8 and IL-1 beta. In normal blood donors IL-1ra serum levels were 150-800 pg/ml (Median: 387 pg/ml). Owing to its sensitivity and large measuring range the newly developed assays appear to be suitable for measuring IL-1ra in cell cultures as well as in biological fluids.

A hybrid plasminogen activator binds to the u-PA receptor and has a reduced thrombolytic potency in vivo.

Fibrinolytic properties of four hybrids of u-PA and t-PA, all containing the u-PA growth factor domain and binding to recombinant human u-PA receptor expressed in CHO cells, were compared. Highest fibrin stimulation was observed with uK2tPA which when compared to t-PA in the rabbit system, had a considerably prolonged circulatory half-life in vivo. Compared to an equimolar dose of t-PA, 0.4 mg/kg uK2tPA caused a similar consumption of alpha 2-antiplasmin and fibrinogen and a considerably greater prolongation of the ex-vivo blood clotting time. Nevertheless, this dose of uK2tPA was inactive in the jugular vein thrombosis assay. This lack of thrombolytic activity is presumably due to the presence of a functional u-PA growth factor domain, which in binding uK2tPA to cellular blood elements possibly retards its penetration into the blood clot and in this manner could neutralize the potential thrombolytic activity of the t-PA kringle 2 and protease domains in uK2tPA.

Mode of action of the lanthionine-containing peptide antibiotics duramycin, duramycin B and C, and cinnamycin as indirect inhibitors of phospholipase A2.

Effects of the lanthionine-containing peptide antibiotics duramycin, duramycin B, duramycin C and cinnamycin on the activity of phospholipase A2 from six different sources were studied, and their mode of action was investigated. The four antibiotics inhibited potently all tested phospholipases A2, with IC50 values of around 1 microM, using phosphatidylethanolamine or [1-14C]oleate-labelled Escherichia coli, whose phospholipids are rich in phosphatidylethanolamine, as substrates. No inhibition was observed when the substrate was phosphatidylcholine. Binding of the antibiotics to the lipid fraction of E. coli could be demonstrated by co-sedimentation with whole, but not with lipid-depleted E. coli. In addition, preincubation of duramycin B with vesicles of phosphatidylethanolamine, but not those of phosphatidylcholine, prevented the inhibition of phospholipase A2 activity. The interaction of duramycin B and C, but not that of the biologically inactive compounds actagardine and the duramycin B trisulphoxide, with phosphatidylethanolamine was demonstrated using circular dichroism studies. On the other hand, no interaction of duramycin B with phosphatidylcholine could be demonstrated. A strict correlation between the physico-chemical interaction of the studied lantibiotics, demonstrated by circular dichroism spectroscopy, and their inhibition of phospholipase A2 was observed. These results suggest that lanthionine-containing peptide antibiotics inhibit phospholipase A2 indirectly by specifically sequestering the substrate phosphatidylethanolamine. This mode of action is analogous to the one described for the protein lipocortin.

Purification, analysis, and enzymatic activity of recombinant human synovial fluid phospholipase A2 and N-terminal variants.

Recombinant human synovial fluid phospholipase A2 (rPLA2) and several variants with N-terminal sequences modified by addition or deletion of one or two amino acid residues (ala or Met; Des-Asn1, Leu2) have been expressed in mammalian cells and in Escherichia coli, respectively, purified to homogeneity, and characterized. The observed values for the molecular mass of rPLA2 and variants are in complete agreement with the predicted values for a correctly folded structure containing seven disulfide bridges. Moreover, the relative proportions of the various types of secondary structures of the variants of rPLA2, as measured by CD spectroscopy, are similar to that found for native porcine pancreatic PLA2, indicating that the recombinant proteins are correctly folded. Enzymatic activities of rPLA2 with modified N-termini decreased to 1.3-0.005% of the activity of the mature rPLA2, emphasizing a key role of the N-terminus for catalytic activity.

Mass spectrometric characterization of stathmin isoforms separated by 2D PAGE.

In proteome analysis, the determination of the phosphorylation status of proteins and protein isoforms, which have been separated by two-dimensional polyacrylamide gel electrophoresis (2D PAGE), is of prime importance in addition to their identification. In this study, the extent to which such information can be directly extracted from the mass spectrometric data used for identification was evaluated. By searching for metastable peaks which are characteristic for loss of phosphoric acid, the Ser-phosphorylated peptides were identified with a high success rate in reflector matrix-assisted laser desorption/ionization (MALDI) mass maps of in-gel digested proteins. Furthermore, by employing a double enzymatic strategy using trypsin and Glu-C in parallel, improved sequence coverage and additional separation of the potential phosphorylation sites of the isoforms were achieved. The precise location of the modified sites within an identified phosphopeptide was obtained by submitting the corresponding molecular ions directly to nano-electrospray tandem mass spectrometric analysis. In this way the detailed phosphorylation status of six isomers of stathmin separated by 2D PAGE was determined. Two of these six isomers were phosphorylated at all four known sites (serines 15, 24, 37 and 62) and were probably derived from the previously reported alpha and beta forms, which differ by a yet unknown modification. In addition, isomers phosphorylated at serines 15, 24 and 37, serines 24, 37 and 62, serines 24 and 37 and serine 37 only were characterized.

Functional effects of kringle 2 glycosylation in a hybrid plasminogen activator.

uK2t-PA is a hybrid plasminogen activator in which the epidermal growth factor-like domain of the urokinase-type plasminogen activator precedes the kringle 2 and catalytic domains of tissue-type plasminogen activator. The molecules are expressed in Chinese hamster ovary cells in two variant forms, a type II form in which only the protease domain is glycosylated, and a type I form in which both the kringle 2 and the protease domains carry N-acetyllactosamine type glycans. The two forms differed slightly in their affinity for fibrin and fibrinogen, which allowed their separation, but the stimulation of plasminogen activation of the type II form by fibrin was up to eight-fold lower than that of the type II form. The sensitivity to fibrin could be restored by treatment of the type I form with N-glycanase or sialidase. Enzymatic activity vs low molecular weight substrates was not influenced by the glycosylation of kringle 2.

Molecular composition of the adenovirus type 2 virion.

The representation of the different structural polypeptides within the adenovirus virion has been accurately determined, and the particle molecular weight has been derived. A stoichiometric analysis was performed with [35S]methionine as a radiolabel, and analytical sodium dodecyl sulfate-polyacrylamide gel electrophoresis was used to separate the polypeptides. The recently available sequence of the adenovirus type 2 genome was used to determine the number of methionines in each polypeptide. The resulting relative representation was placed on an absolute scale by using the known number of hexon polypeptides per virion. The analysis provides new information on the composition of the vertex region, which has been the subject of some controversy. Penton base was found to be present in 60 copies, distributed as pentamers at each of the 12 vertices. Three fiber monomers were associated with one penton base to form the penton complex. Polypeptide IX was present in 240 copies per virion and 12 copies per group-of-nine hexons, supporting a model proposed earlier for the distribution of this protein. The location of polypeptide IX explains the dissociation of the virus outer capsid into groups-of-nine hexons. The penton base was microheterogeneous, and the relative amounts suggest that the symmetry mismatch, which occurs within the penton complex between base and fiber, is resolved by the synthesis of penton base polypeptides from two closely spaced start codons.

Isolation and crystallization of lambda exonuclease.

lambda Exonuclease is a deoxyribonuclease induced by bacteriophage lambda. Mutations in the structural gene for the protein affect general recombination and indicate a possible function for the enzyme. A large scale isolation procedure was employed to purify enough enzyme from a heat-induced lambda lysogen for X-ray crystallographic analysis. Analytical ultracentrifugation and SDS-polyacrylamide electrophoresis revealed that lambda exonuclease is a tetramer with molecular mass 107,000 Da. Crystallization trials produced morphologically perfect crystals of a size suitable for X-ray diffraction studies. Cubic crystallographic symmetry was indicated by the lack of birefringence when the crystals were inspected with polarized light. X-ray precession photographs indicated that lambda exonuclease crystallizes in a space group of P4(1)32, or its enantiomorph P4(3)32, with 24 tetramers in the unit cell of edge 210 A.

Adenovirus polypeptide IX revealed as capsid cement by difference images from electron microscopy and crystallography.

Particles of adenovirus type 2 (ad2), when disassembled, consistently yield groups-of-nine (GON) hexons, which are the major virion shell component. The location of a minor component (6%) of the GON has been determined using a novel combination of electron microscopy and X-ray crystallography. The Brookhaven Scanning Transmission Electron Microscope (STEM) was used to estimate the distribution of protein in the GON to a resolution of 15-18 A. The relative hexon positions then were determined to within 1 A using a model of the hexon derived from the X-ray crystal structure to search the STEM image. The difference image between the STEM image and a model hexon group reveals individual monomers of polypeptide IX extending along the hexon--hexon interfaces. The distribution confirms our earlier proposal that four trimers of polypeptide IX are embedded in the large cavities in the upper surface of the GON to cement hexons into a highly-stable assembly.

Towards high performance two-dimensional gel electrophoresis using ultrazoom gels.

Proteomics is the analysis of protein expression in cells or tissues, e.g., to study cellular processes at the molecular level. Ultimately, a proteome analysis should encompass most if not all protein species in a biological sample, including those present in low copy numbers. We are developing two-dimensional gel electrophoresis technology by applying narrow pH range ultrazoom gels to enhance resolution and to improve the detection of low abundance proteins. Ultrazoom gels in the acidic pH range allow the detection of proteins down to 300 copies per cell of a B-lymphoma cell line. Protein separation in the alkaline pH range, however, still requires optimization, especially in conjunction with high sample loads.

Interpretation of electron micrographs of adenovirus hexon arrays using a crystallographic molecular model.

Two types of two-dimensional arrays of purified adenovirus type 2 hexon have been obtained and analyzed by Fourier filtration of their electron micrographs. One array contained continuously close-packed hexons, distributed on a hexagonal p3 lattice, with a unit cell dimension of 94 +/- 2 A. The other array contained close-packed hexons with a regular absence, so that rings of six hexons related by sixfold symmetry formed a p6 unit cell. The cell dimension of the hexagonal array was 153 +/- 3 A, with neighboring hexons separated by 88 +/- 2 A. Smaller p6 arrays were also formed by hexons freed from complete virions on the microscope grid by treatment with distilled water. A molecular model of hexon, known from the X-ray crystallographic structure, was used to interpret Fourier-filtered images of the arrays, and to determine the relative orientations of the hexon molecules. The hexon-hexon interaction in the p3 array is that found in the virion facet, whereas that in the p6 array is a planar form of the interaction between peripentonal hexons around the vertex.

The structure of murine interleukin-1 beta at 2.8 A resolution.

The three-dimensional structure of recombinant murine interleukin-1 beta has been solved by X-ray crystallographic techniques to 2.8 A resolution and refined to a crystallographic R factor of 0.192. Although murine interleukin-1 beta crystallizes in the same space group as human interleukin-1 beta with almost identical unit cell dimensions, the packing of the molecules is quite different. The murine interleukin-1 beta structure was solved by molecular replacement using the refined structure of human interleukin-1 beta as trial structure, and found to be related to the human structure by a nearly perfect twofold rotation about the crystallographic y-axis and a 14 degrees rotation about the z-axis, with no translation. The folding of murine interleukin-1 beta is similar to that found for the human variant, consisting of 12 beta strands wrapped around a core of hydrophobic side chains in a tetrahedron-like fashion. Significant differences with respect to the human structure are seen at the N terminus and in 4 of the 11 loops connecting the 12 beta strands.

Neoepitope immunoassay: an assay for human interleukin 1 beta based on an antibody induced conformational change.

A secondary monoclonal antibody (mAb2) was generated by immunization with immune complexes of human IL-1 beta and a primary monoclonal (mAb1). mAb2 bound to a neoepitope on the IL-1 beta/mAb1-complex with a dissociation constant (Kd) of 26 pM but not to uncomplexed IL-1 beta. As assessed by the binding of labeled IL-1 beta and neutralization of bioactivity, mAb2 enhanced the affinity of IL-1 beta to mAb1; Kd-values were 108 pM in absence and 5.4 pM in presence of mAb2. By analyzing a series of mutants of IL-1 where surface loops had been exchanged with the corresponding loops of human IL-1 receptor antagonist protein, a critical region responsible for mAb2 binding was localized to the C-terminal region. In addition to mAb1/IL-1 beta-complexes, mAb2 bound pro-IL-1 beta/mAb1 complexes as well as pro-IL-1 beta suggesting that mAb2 recognized a conformation of IL-1 beta resembling that of pro-IL-1 beta. Using this pair of mAbs, chemiluminescent and enzyme linked assays with detection limits of 2 pg/ml hIL-1 beta have been established.

Characterization and direct glycoform profiling of a hybrid plasminogen activator by matrix-assisted laser desorption and electrospray mass spectrometry: correlation with high-performance liquid chromatographic and nuclear magnetic resonance analyses of the released glycans.

Mass spectrometric analysis of the recombinant hybrid plasminogen activator K2tu-PA at the native glycoprotein and corresponding protein level (high mass) is described. For gross structural characterization of the major glycotypes present, a combination of enzymatic degradation and matrix-assisted laser desorption mass spectrometric analysis of the products proved convenient. In this way, mono- and di-N-glycosylated one- and two-chain molecules, unresolved in the spectra of native material, were identified, with the one-chain type monoglycosylated at Asn247 representing the major component. Actual detection of resolved original glycoforms, or unresolved groups thereof when composed of isobaric species, and their assignments regarding antennicity and degree of sialylation were possible by electrospray mass spectrometry for the major monoglycosylated one-chain species. Desialation also allowed detection of the diglycosylated one-chain species as a minor constituent. The electrospray mass spectrometric results were correlated with structural and quantitative data available from a parallel high-performance liquid chromatographic and 1H nuclear magnetic resonance study performed at the glycan (low mass) level on liberated individual carbohydrate components. The results of the two studies showed full consistency in all respects amenable to evaluation, i.e. excellent agreement for the assignments of antennicity and degree of sialylation of the major glycan components, and good agreement for the determination of proportions in which these were present in the major monoglycosylated glycoprotein. This provided a sound basis for direct glycoform profiling by electrospray mass spectrometry. Typical applications to batch quality control and to structural characterization of non-standard material are shown.