Efficacy and safety of a new immunoglobulin G product, Gammaplex(®), in primary immunodeficiency diseases.

This open-label multi-centre study evaluated a new intravenous immunoglobulin, Gammaplex®, in the treatment of 50 patients with primary immunodeficiency and significant hypogammglobulinaemia. Patients treated previously with other intravenous immunoglobulins received Gammaplex® on their same infusion schedule for 1 year; 22 were on a 21-day and 28 on a 28-day regimen (300-800 mg/kg/infusion). There were no serious, acute bacterial infections, whereas six subjects (12·0%) had at least one such infection in the 6 months before enrollment. Forty subjects (80·0%) had at least one non-serious infection; the median number of infective episodes per subject per year was 3·07. Antibiotics were taken by 38 subjects therapeutically and prophylactically by 16 at some time. Fewer than half (46·0%) missed any time off work or school because of infection or other illness. Trough immunoglobulin (Ig)G levels were above 6·00 g/l in all subjects at all assessments after 15 weeks with two exceptions. Overall, 21·2% of infusions were associated with an adverse event up to 72 h after infusion. The frequency of adverse events increased with infusion rate. Headache was the most common product-related adverse event (7·5% of 703 infusions). In conclusion, Gammaplex® is effective in primary immunodeficiency and is well tolerated.

Cytokine conformations: predictive studies.

The amino acid sequences of human and murine haemopoietins have been analysed using algorithms predictive for secondary structure. The results for 19 of these proteins (human and murine interleukins 2, 3, 4, 5, 6, 7 and granulocyte, macrophage and granulocyte macrophage-colony stimulating factors as well as human erythropoietin) suggest that they each contain a 4-alpha-helical bundle, ca 25 A long, as a common conformational feature. The most important predictive indicator was considered to be the occurrence of quasi-repeating sequences of seven amino acids of the form (a-b-c-d-e-f-g), with apolar side chains (usually leucine) lying alternately three and four residues apart in the a and d positions. As with other proteins of known secondary structure this periodicity favours the formation of alpha-helical elements, each with an apolar external strip, which interdigitate closely with one another when tested appropriately. Molecular models based on these putative 4-alpha-helical bundles are presented--with special reference to human granulocyte macrophage-colony stimulating factor. The extent to which such models are consistent with experiments designed to delineate receptor binding sites is discussed.

Location of the dihydroorotase domain within trifunctional hamster dihydroorotate synthetase.

Mammalian dihydroorotase (DHOase, EC 3.5.2.3) is part of a trifunctional protein, dihydroorotate synthetase which catalyzes the first three reactions of de novo pyrimidine biosynthesis. We have subcloned a portion of the cDNA from the plasmid pCAD142 and obtained a nucleotide sequence which extends 2.1 kb in the 5' direction from the sequence encoding the aspartate transcarbamoylase (ATCase) domain at the 3'-end of the cDNA. The DHOase and ATCase domains have been purified from an elastase digest of the trifunctional protein and subjected to amino acid (aa) sequencing from their N termini. The sequence of the N-terminal 24 aa of the DHOase domain has been obtained and aligned with the cDNA sequence. The C-terminal residues of the DHOase domain have been identified as Leu followed by Val which, when taken with partial sequences of the CNBr fragments of this domain, defines the coding sequence of the active, globular DHOase domain released by proteolysis. Prediction of protein secondary structure from the deduced aa sequence showed that the DHOase domain (Mr 37,751) is separated from the C-terminal ATCase domain (Mr 34,323) by a bridging sequence (Mr 12,532) consisting of multiple beta-turns.

Conformational homologies among cytokines: interleukins and colony stimulating factors.

Some 30 cytokine amino acid sequences (mainly interleukins, colony stimulating factors and tumor necrosis factors) have been examined for evidence of secondary structure as well as longer-range interactions of a type likely to lead to stable alpha-helical bundles. Most, though not all, of the cytokines examined have a high predicted alpha-helical content (40-60%) and quasi-repeating heptads containing i/i + 3 apolar periodicities. This major subset of the cytokines is predicted to be characterized by molecules in which 4-alpha-helical bundles with an average length of 25A are the most marked conformational features. Based on these conclusions, we suggest structures for huG-CSF, huGM-CSF and muIL-5 in which defined loop segments at the ends of helical bundles are the most likely sites for binding and recognition by specific cell receptors. As such, they provide a means for testing or refining the three working models we have defined, using currently available methods of site-directed substitution and deletion mutagenesis, as well as synthetic peptides corresponding to the proposed loop sequences and the use of monoclonal antibodies of defined epitopic specificity. The structure arrived at for huGM-CSF is consistent with the limited data currently available concerning the residues which are important for binding and activity.

Synthesis, antiviral activity, and conformational characterization of mouse-human alpha-interferon hybrids.

Reciprocal hybrids were constructed between human and mouse interferons (IFNs), and their antiviral activity was examined on different target cells and compared to the activity of the parental molecules. In addition, we used a number of predictive algorithms on a data base of the available alpha-interferon sequences to propose a working model for the overall conformation of the alpha-interferon molecule that is consistent with the structural predictions. Remarkable conservation within the predicted alpha-helical segments of the interferon molecule was observed. We propose that the observed changes in the activity and specificity of the hybrids obtained are largely due to the sequences present in the loops at the ends of the major helical structures; these are less conserved, contain beta-bends, and are generally hydrophilic and flexible. The data on the constructed mouse-human hybrids have shown that the activity on human cells is contributed by determinants present in the N-terminal 122 amino acids of human IFN, thus implicating one or more loops within this region (e.g. loops 1-12, 25-38, 70-74, and 103-113). The activity on bovine cells appears to be localized mainly in sequence 60-121, implicating the role of loops 70-74 and/or 103-113 of the human IFN molecule. The specificity of mouse IFN for mouse cells is in some or all of the loops (70-74, 103-113, 134-139, and 163-166) in the C-terminal sequence. The proposed working model should provide guidelines for the study of the specificity of action in molecular terms.

Antibody response to the C-terminal peptide sequence in beef myoglobin.

The nature of the surface determinants on the beef myoglobin molecule which direct the distinctive antibody response of sheep, have been further defined. Antisera raised to beef myoglobin in sheep, rabbits and mice have been compared for their ability to recognize the synthetic C-terminal beef myoglobin peptide (140-153), which contains four of the six amino acid substitutions between sheep and beef myoglobins. The antibodies raised in rabbits, directed to the entire surface of beef myoglobin, contain only a minor population directed to the C-terminal sequence: in mice, even fewer antibodies are specific for this sequence. In sheep, however, antibodies to beef myoglobin appear to be directed almost exclusively to topographic domains which include the (140-153) sequence. This specificity is most apparent in "early" antisera in which all antibodies display equal avidity for "native" beef myoglobin and the peptide; further immunization produces antibodies which recognize a larger overlapping set of domains and only 20% of these antibodies have effective avidity for the (140-153) peptide. The antibodies to beef myoglobin raised in sheep comprise two discrete populations. One ("common") population is directed to regions of similarity between the beef and sheep myoglobin molecules, in which the region represented by the C-terminal peptide of beef myoglobin is less important in defining the antibody-binding site and/or affinity, while still being directly involved in the topographic determinant. The other ("non-common") appears to be directed almost exclusively to the C-terminal sequence (140-153) of beef myoglobin. The findings are discussed in relation to our previous findings on the effect of the host species on the nature of the antibody response and in relation to views on the possibility of direct vs indirect effects of evolutionary amino acid substitution on immuno-cross-reactivity among homologous proteins.

Antigenic repeat structures in proteins of Plasmodium falciparum.

The majority of malaria antigens that have been cloned contain short sequence repeats which encode antigenic epitopes that are naturally immunogenic. Synthetic peptides have been used to show that natural antibody responses to a strain-specific Plasmodium falciparum S antigen are largely directed against epitopes encoded in an 11-amino acid sequence that is repeated approximately 100 times in the molecule. A 16-amino acid peptide conjugated to bovine serum albumin induced antibodies specific for the S antigen of the homologous isolate. Synthetic peptides have also been used to confirm the natural immunogenicity of epitopes encoded within two blocks of related repeats in the Ring-infected Erythrocyte Surface Antigen (RESA). A 16-amino acid peptide, comprising four repeats of the tetrameric sequence EENV, induced antibodies reactive with the native molecule. Detailed analyses of these anti-peptide antisera indicate that short sequence repeats express more than one epitope, some of which may cross-react with other repeat structures.

Two large immunogenic and antigenic myoglobin peptides and the effects of cyclisation.

Peptides corresponding to sequences (72-88) and (26-54) of beef myoglobin have been synthesised in their open-chain and cyclised forms (using a disulphide bridge) and tested for their antigenicity and immunogenicity. Antibodies raised to beef myoglobin bound to both peptides but more strongly to the 29-residue than to the 17-residue peptide. Cyclisation increased the antigenicity of the larger peptide. In this form the peptide competed much more strongly than in the uncyclised form for specific antibodies to beef myoglobin. The peptides are immunogenic in mice without being coupled to a protein carrier and produce antibodies which bind to beef myoglobin. Peptide (26-54) is the more immunogenic in producing a larger antibody titre to the parent myoglobin and cyclisation again enhances this property. The findings lend weight to the view that longer peptide sequences might be expected to favour the folded state, therefore binding more strongly to antibodies raised to the native protein and eliciting a population of antibodies which contain a larger proportion specific for that conformation. Cyclisation enhances antigenicity and immunogenicity presumably by decreasing the number of degrees of conformational freedom of a peptide without excluding native-like conformations.

Original antigenic sin and the production of autoantibodies.

Autoantibodies to sheep myoglobin have been raised by priming sheep with beef myoglobin and boosting with sheep myoglobin. The autoantibodies appear to be a subset of those produced when beef myoglobin is used for both priming and boosting. This subset of antibodies is presumably directed to the surface regions which are common to both myoglobins. The antibodies which bind to sheep myoglobin in the 2 types of antisera differ. Those elicited by boosting with beef myoglobin bind better to beef myoglobin than to sheep myoglobin, while those obtained by boosting with sheep myoglobin bind with equal avidity to the 2 myoglobins. It would seem therefore that the boosting immunogen determines which fraction of antibodies is selected from the antibody repertoire established by the priming immunogen. Our results also show that tolerance at the T-cell level can be circumvented by exposing the immune system to a protein closely related to a homologous self protein.

Conformational analysis of cyclized dipeptide models for specific types of beta-bends by two-dimensional nuclear Overhauser spectroscopy.

The solution conformations in DMSO-d6 of the two cyclized dipeptides, cyclo(L-alanyl-L-alanyl-epsilon-aminocaproyl) and cyclo(L-alanyl-D-alanyl-epsilon-aminocaproyl), have been analyzed by means of the two-dimensional nuclear Overhauser effect (2D-NOE). The preferred conformations for the two compounds have been deduced by comparing proton-proton distances, derived from the 2D-NOE data and relaxation-time measurements, with the corresponding distances in several possible computed low-energy conformations. The predominant conformations are a type III bend and a type II bend, respectively, for the two compounds. These conclusions agree with those deduced earlier on the basis of infrared and Raman spectra and circular dichroism measurements.

Antibody response to myoglobins: effect of host species.

Using both direct and competitive binding studies it is demonstrated that antibodies to beef myoglobin raised in sheep are able to distinguish between beef and sheep myoglobins although these two proteins differ by only six of the 153 amino acid residues. By contrast, antibodies to beef myoglobin raised in rabbits, dogs and chickens bind almost equally well to beef and sheep myoglobins. It is also shown that antibodies to beef myoglobin raised in sheep have a lower avidity for beef myoglobin than do antibodies raised in more distantly related species. Furthermore, only 50% of the specific anti-beef myoglobin antibodies isolated from sheep antisera will bind to sheep myoglobins whereas 100% of the specific antibodies isolated from the antisera of the other immunised species will bind to sheep myoglobin. It is suggested that antibodies to beef myoglobin are raised to those surface regions which are topographically altered as a result of sequence differences from the host's own myoglobin. When the host animal is evolutionarily distant these sequence differences are considerable and antibodies are raised to the entire surface of the molecule. However, when the host's myoglobin is very similar in sequence to beef myoglobin (as is the case when using sheep as the host animal) antibodies are made only to surface regions affected by the sequence differences. Some of these antibodies--those to the regions of greatest difference--will bind weakly if at all to sheep myoglobin, while those directed to areas of lesser difference will bind well to sheep myoglobin.

The antigenicity of myoglobin-related peptides synthesised on polyacrylamide and polystyrene resin supports.

Polyacrylamide resins [Atherton et al., Bioorg. Chem. 8, 351-370 (1979)] have been found suitable for solid-phase radioimmunoassay of peptides synthesised on the same supports; they are sufficiently stable during side-chain deprotection and swell sufficiently in aq. media to admit antibody molecules to the sites of peptide attachment. A re-examination of five synthetic peptide sequences corresponding to (15-21), (56-62), (94-99), (113-119) and (145-151) of beef myoglobin analogous to those delineated by Atassi [Immunochemistry 12, 423-438 (1975)] for sperm whale myoglobin shows that they all bind anti-beef myoglobin antibodies raised in rabbits, with binding capacities in the order V = III greater than IV greater than I = II. The resin-bound peptide (72-88) binds such antibodies even more extensively, as do certain sequential variants of peptide V. Other peptides, bound to polyacrylamide or polystyrene resins but unrelated to any of the five sequences and varying in size and amino acid composition and sequence were also tested with various antisera. It was concluded that the antibody binding properties of the 30 or so small peptides (two-seven residues) are dominated by their cationic and/or hydrophobic properties. In small peptides, therefore, antibody binding can be safely interpreted only in terms of general structural properties but not in terms of biological specificity. The latter property becomes assessable only with peptides representing larger areas of antigenic protein surfaces.

Solution conformations of oligomers of alpha-aminoisobutyric acid.

The N-acetyl(Aib)nN'-methylamides (with n = 1, 2 and 3) and the N-acetyl-(Aib)3 methyl ester have been synthesized using an oxazolone procedure. An experimental conformational analysis of this series of oligomers has been carried out in water, DMSO-d6 and CDCl3 using n.m.r. techniques, and in chloroform using i.r. spectroscopy. Deuterium exchange rates of amide protons in DMSO-d6 and the rates of these proton chemical shifts with temperature in water, DMSO-d6 and CDCl3 indicate that the oligomeric N'-methylamides adopt conformations that have no hydrogen bonds when n = 1, one hydrogen bond when n = 2, and two hydrogen bonds when n = 3, and that Ac(Aib)3OMe has a conformation with one hydrogen bond. An analysis of the N-H stretching region of the i.r. spectra of these compounds in CHCl3 also suggests the existence of these conformational states. These data imply that the peptides adopt the 3(10)-helical and not the alpha-helical conformation in solution. This conclusion supports the hypothesis that the Aib residue has asymmetric geometry at the C alpha atom in solution, similar to that reported in the literature for the crystalline state.

Circular dichroism studies of relaxin and insulin peptide chains.

The circular dichroism (CD) spectra of pig relaxin and pig insulin are similar, reflecting the known structural similarities between the two hormones. However, the conformational characteristics of the separate chains of insulin and relaxin show significant differences. The S-sulfo forms of insulin A and B chains and S-sulfo relaxin A chain have CD spectra consistent with largely unordered structures whereas the S-sulfo form of relaxin B-chain has at least 90% beta--structure. This beta-structure may explain the unusual solubility and adsorptive properties of the relaxin B-chain and the poor combination yields with A-chain. The relaxin B-chain changes to a largely unordered conformation if the peptide is shortened at the carboxyl terminus by six amino acid residues. This conformational change has important implications in planning relaxin synthesis strategy. Significant interactions and conformational changes are observed between the oxidized forms of the A and B chains of both relaxin and insulin. In using CD to monitor chain recombination of native relaxin peptides it was found that the spectra obtained after depending on whether the reduced chains are separated or not separated from the reaction mixture prior to reoxidation. Although the spectra differ the biological activity was 20% in both cases. The remainder of the reoxidized but inactive material contains beta-structures which make a greater contribution to the CD spectrum when the chains have been separated and processed than when they are reoxidized in situ.

Antigenic specificity of monoclonal antibodies to human myoglobin.

Two monoclonal antibodies directed against different sites of the human myoglobin molecule have been tested for their cross-reactivities against several myoglobins including seven from mammalian species. The relation between their cross-reactivities and their amino acid sequences had led to a possible localization of two antigenic domains in human myoglobin. Each domain includes residues previously considered not to be directly involved in the antigenic structure of myoglobin. Unlike polyclonal serum antibodies, monoclonal hybridoma antibodies directed to a native protein often fail to bind to supposedly antigenic protein fragments. This is explicable in terms of the concept of antigenic domains. Such domains are numerous and overlapping, each comprising a number of contributory amino acid side chains which need not necessarily include continuous sequences of amino acids and which need not exhibit measurable antigenicity in isolation from the rest of the domain.

Conformations of cyclo(L-alanyl-L-alanyl-epsilon-aminocaproyl) and of cyclo(L-alanyl-D-alanyl-epsilon-aminocaproyl); cyclized dipeptide models for specific types of beta-bends.

Conformational energy calculations indicate that the peptide backbones of the low-energy conformations of the cyclized dipeptide derivatives cyclo (L-alanyl-L-alanyl-epsilon-aminocaproyl) and cyclo (L-alanyl-D-alanyl-epsilon-aminocaproyl) are constrained to form beta-bends of types I + III and II, respectively. Thus, the two compounds can serve as models for the spectroscopic properties of beta-bends of these types. The coupling constants obtained from 1H n.m.r. spectra in DMSO-d6 are consistent with the dihedral angeles of the computed lowest-energy conformations. Differences in 13C chemical shifts between the two compounds can be correlated with differences in shielding by C=O groups in bends of various types. 1H and 13C chemical shifts suggest association of cyclo (L-Ala-L-Ala-Aca) but not of cyclo (L-Ala-D-Ala-Aca) in dimethylsulfoxide. The different tendencies to associate can be explained in terms of the difference in conformation. The circular dichroism spectra of the two compounds are quite different. In methanol, trifluoroethanol and water, the L-Ala-L-Ala derivative has a positive extremum near 190 nm and two negative extrema near 206 and 220 nm, whereas the L-Ala-D-Ala derivative has a positive extremum at about 203 nm and negative extrema at about 187 and 229 nm. The spectra can be used to estimate the contribution of various bend types in a related series of compounds. A normal mode analysis of the vibrations of the computed low-energy conformations was compared with solid state infrared and Raman spectra, in order to determine the predominant conformations. The bend types determined by this comparison fully agree with the predictions of the theoretical computations for both derivatives.