Retroviral display of antibody fragments; interdomain spacing strongly influences vector infectivity.

Five different single-chain antibody fragments (scFv) against human cell-surface antigens were displayed on murine ecotropic retroviral vectors by fusing them to the Moloney SU envelope glycoprotein. The spacing between the scFv and the SU glycoprotein was varied by fusing the scFv to residue +7 or to residue +1 of Moloney SU and by inserting linker sequences of different lengths between the domains. All of the chimeric envelopes were efficiently incorporated into vector particles and could bind to human cells through their displayed antibody fragments, but did not infect them. The spacing between the scFvs and the SU glycoproteins had no significant effect on the efficiency of envelope expression or viral incorporation and did not affect the binding properties of the chimeric envelopes, nor did it influence the efficiency of targeted gene delivery to human cells by scFv-displaying vectors. However, on murine fibroblasts the infectivity of vectors incorporating the chimeric envelopes was strongly influenced by the length of the interdomain spacer. The titers were very low when the single-chain antibodies were fused through a tripeptide linker to SU residue +7 and were greatly enhanced (up to 10(5)-fold) when they were fused to SU residue +1 through a heptapeptide linker. These results point to the importance of steric interactions between the domains of chimeric envelope glycoproteins and may have implications for retroviral vector design for human gene therapy.

Targeting of retroviral vectors through protease-substrate interactions.

Targetable, injectable vectors would greatly facilitate the development of in vivo therapy strategies. Viral and nonviral vectors can be targeted through ligand-receptor interactions, but protease-substrate interactions have not previously been exploited for vector targeting. Epidermal growth factor (EGF) was fused to a retroviral envelope glycoprotein via a cleavable linker comprising a factor Xa protease recognition signal. Vector particles displaying the cleavable EGF domain could bind to EGF receptors on human cells but did not transfer their genes until they were cleaved by factor Xa protease, whereupon gene delivery proceeded normally. Proteolytic activation of receptor-targeted vectors can therefore provide the basis for a novel two-step targeting strategy that may facilitate efficient targeted in vivo delivery of therapeutic genes.

Improvement of retroviral retargeting by using amino acid spacers between an additional binding domain and the N terminus of Moloney murine leukemia virus SU.

We previously reported a strategy to redirect the retroviral host range by expressing single-chain antibodies (S. J. Russell, R. E. Hawkins, and G. Winter, Nucleic Acids Res. 21:1081-1085, 1993) or ligands (F.-L. Cosset, F. J Morling, Y. Takeuchi, R. A. Weiss, M. K. L. Collins, and S. J. Russell, J. Virol. 69:6314-6322, 1995) at the N terminus of Moloney murine leukemia virus (MoMLV) surface proteins (SU). Although such chimeric envelopes were able to bind the new receptors, the transduction efficiency of retargeted viruses was generally low. We hypothesized that conformational rearrangements of envelope glycoproteins were not optimally triggered following binding, and to overcome these postbinding blocks, we have generated here a set of chimeric MoMLV-derived envelopes targeted to the Ram-1 phosphate transporter in which we have varied the spacing between the Ram-1-binding domain and the MoMLV SU. All of the recombinant envelopes were correctly expressed on virions, and all bound efficiently to Ram-1. However, the interdomain spacing greatly affected the efficiency of gene transfer by retroviral vectors that had bound to Ram-1 via their chimeric envelopes. Optimal interdomain spacing allowed a 100-fold-increased viral transduction via Ram-1 compared to our previous results.

Structure and stability of protein H and the M1 protein from Streptococcus pyogenes. Implications for other surface proteins of gram-positive bacteria.

M proteins and other members of the M protein family, expressed on the surface of Streptococcus pyogenes, bind host proteins such as immunoglobulins, albumin, and fibrinogen. Protein H and the M1 protein are expressed by adjacent genes and both belong to the M protein family. In this work, the structure and stability of these two proteins have been investigated. As judged from sequence analysis and circular dichroism spectroscopy, the proteins are almost entirely in an alpha-helix conformation. The amino acids are arranged in a seven-residue (heptad) repeat pattern along the greater part of the proteins. These observations support the previously accepted model of M proteins as coiled-coil dimers. However, it was also found that the structures of both proteins were thermally unstable; i.e., the content of helix conformation was greatly reduced at 37 degrees C as compared to 25 degrees C or below. Together with previous findings that these proteins appear as monomers at 37 degrees C and dimers at low temperatures, the results suggest that the coiled-coil dimers are unfolded at 37 degrees C. The heptad patterns of protein H and the M1 protein showed a nonoptimal distribution of residues expected for a coiled-coil conformation. This is a possible explanation for the low thermal stability of the proteins. It was also demonstrated that the proteins were stabilized in the presence of the ligands IgG and/or albumin. Protein H and M1 protein show a high degree of sequence similarity in their C-terminal regions, and a fragment from this region displayed a high content of helix conformation, whereas fragments from the nonsimilar N-terminal parts did not adopt any stable folded structure. Thus, the C-terminal parts, which are conserved within the M protein family, may constitute a framework for the formation of the parallel helical coiled-coil structure, and we propose that the less stable N-terminal part may also participate in antiparallel interaction with M proteins on adjacent bacteria. The results suggest that temperature fluctuations in the environment could change the properties of bacterial surface proteins, thereby affecting the molecular interactions between the bacterium and its host.

On the interaction between single chain Fv antibodies and bacterial immunoglobulin-binding proteins.

Using four bacterial immunoglobulin-binding proteins, we have analyzed the binding characteristics of a panel of 34 human single chain Fv antibodies, expressed in E. coli and with known specificity and sequence. Several of the single chain Fv antibodies showed affinity for staphylococcal protein A and peptostreptococcal protein L, but not for the streptococcal proteins G or H. The affinity of the binding was higher for protein L (4.5 and 1.4 x 10(9) M-1) than for protein A (7.7 and 6.7 x 10(8) M-1), using the two single chain Fv antibodies displaying the strongest binding activity to these ligands. The binding was shown to be specific by Western blotting, and the single chain Fv antibodies could be purified from crude bacterial culture media by affinity chromatography on protein L- or A-Sepharose. Protein A, which has affinity for the VH domain of the scFv antibodies, was tested against scFv antibodies containing VH1, VH3, VH4 and VH5 domains, and its binding was restricted to approximately half of the scFv antibodies with a VH3 domain. Protein L, which has affinity for the VL domain, was tested against kappa 1, kappa 4, lambda 1, lambda 2 and lambda 3 domains, and it bound all kappa 1 domains, one lambda 2 and one lambda 3 domain. Comparison of the amino acid sequences of binding and non-binding VL domains demonstrated that amino acid residues crucial to the binding of protein L were distributed over a large area outside the hypervariable antigen-binding regions.

Purification of antibodies using protein L-binding framework structures in the light chain variable domain.

Protein L from the bacterial species Peptostreptococcus magnus binds specifically to the variable domain of Ig light chains, without interfering with the antigen-binding site. In this work a genetically engineered fragment of protein L, including four of the repeated Ig-binding repeat units, was employed for the purification of Ig from various sources. Thus, IgG, IgM, and IgA were purified from human and mouse serum in a single step using protein L-Sepharose affinity chromatography. Moreover, human and mouse monoclonal IgG, IgM, and IgA, and human IgG Fab fragments, as well as a mouse/human chimeric recombinant antibody, could be purified from cultures of hybridoma cells or antibody-producing bacterial cells, with protein L-Sepharose. This was also the case with a humanized mouse antibody, in which mouse hypervariable antigen-binding regions had been introduced into a protein L-binding kappa subtype III human IgG. These experiments demonstrate that it is possible to engineer antibodies and antibody fragments (Fab, Fv) with protein L-binding framework regions, which can then be utilized in a protein L-based purification protocol.

On the interaction between protein L and immunoglobulins of various mammalian species.

Protein L, a cell wall molecule of certain strains of the anaerobic bacterial species Peptostreptococcus magnus, shows high affinity for human immunoglobulin (Ig) light chains. In the present study protein L was tested against a panel of human myeloma proteins of the IgG, IgM, IgA and IgE classes, and strong binding was seen with antibodies carrying kappa light chains. A high degree of specificity for Ig was demonstrated in binding experiments with human plasma proteins. Apart from human Ig, strong protein L-binding activity was also detected in the serum of 12 out of 23 tested additional mammalian species, including other primates and rodents. Subsequent analysis with purified Ig samples demonstrated the binding of protein L to Ig of important laboratory animal species such as the mouse, the rat and the rabbit. The affinity constants for the interactions between protein L and polyclonal IgG of these species were 2.6 x 10(9), 3.9 x 10(8) and 7.4 x 10(7), respectively. In non-human species, the binding of protein L was also found to be mediated through Ig light chains, and the results demonstrate the potential value of protein L as an immunochemical tool.

Protein L from Peptostreptococcus magnus binds to the kappa light chain variable domain.

Protein L is an immunoglobulin light chain-binding protein expressed by some strains of the anaerobic bacterial species Peptostreptococcus magnus. The major variable region subgroups of human kappa and lambda light chains were tested for protein L binding; V kappa I, V kappa III, and V kappa IV bound protein L, whereas no binding occurred with proteins of the V kappa II subgroup or with any lambda light chain subgroups. Studies of the protein L binding capacity of naturally occurring VL fragments, and VL- and CL-related trypsin- and pepsin-derived peptides prepared from a kappa I light chain, localized the site of interaction to the VL domain. The affinity constant for the binding to an isolated V kappa I fragment was comparable to that for the native protein (Ka 0.9 x 10(9) M-1 and Ka 1.5 x 10(9) M-1, respectively). No binding occurred with CL-related fragments. Extensive reduction and alkylation of the V kappa fragment or the native kappa chain resulted in complete loss of protein L binding. Although it is possible, from comparative amino acid sequence data, to identify certain VL-framework region residues that account for the selective binding of protein L by kappa I, kappa III, and kappa IV proteins, our studies indicate that this interaction is essentially dependent upon the tertiary structural integrity of the kappa chain VL domain.

Characterization of monoclonal anti-alpha 1-microglobulin antibodies: binding strength, binding sites, and inhibition of lymphocyte stimulation.

Eleven monoclonal antibodies (MoAb) directed against the immunoregulatory plasma glycoprotein alpha 1-microglobulin were characterized. The MoAb were produced in mice immunized with a mixture of alpha 1-microglobulin homologues from man, guinea pig, rat and rabbit. Using radioimmunoassay, western blotting, affinity chromatography, and Scatchard analysis, the affinities and binding sites of the MoAb were analysed. All antibodies were more or less cross-reactive, but most showed a major specificity for one or two of the alpha 1-microglobulin homologues. None of the antibodies was directed against the carbohydrate moiety of alpha 1-microglobulin. Six of the MoAb had high affinity for the antigen and four of these were directed towards the same part of the molecule though differing in their species specificity. Five showed lower affinity for the antigen and were mainly directed towards epitopes on other parts of the molecule. Only some of the antibodies could block the proliferation of lymphocytes induced by human alpha 1-microglobulin. The blocking efficiency of the different antibodies was similar when tested on the stimulation of human or mouse lymphocytes, suggesting that the same part of the alpha 1-microglobulin molecule is responsible in both species. The magnitude of blocking by the different MoAb was not related to their affinities, emphasizing the importance of where on the alpha 1-microglobulin molecule, rather than how strongly, they bind. The binding of the strongest blocking antibody was shown to be directed to a C-terminal peptide of rat alpha 1-microglobulin, indicating that this part of alpha 1-microglobulin is important for the mitogenic effects. Thus the panel of anti-alpha 1-microglobulin MoAb should be a valuable tool for structural and functional studies of alpha 1-microglobulin.

Antibody response in immunized rabbits measured with bacterial immunoglobulin-binding proteins.

Protein G, an immunoglobulin (Ig)-binding protein isolated from group C or G streptococci, binds to the Fc portion of IgG. Protein L, from the anaerobic bacterium Peptostreptococcus magnus, specifically binds light chains of Ig. In this study, protein G and L were used to measure the production of antibodies in immunized rabbits. Two rabbits were immunized with a mixture of human urinary proteins from a patient with tubular proteinuria, and blood samples were collected regularly from the animals for 6 weeks after the immunization. The antibody levels of the blood samples against six of the proteins in the antigen mixture were then measured by ELISA. Microtiter plates were coated with each of the antigens, incubated with the rabbit serum samples, and the specific antibodies of the IgG class measured by incubation with biotinylated protein G, and antibodies of all Ig classes with biotinylated protein L. Alternatively, Western blotting was employed, where the antibodies which bound to each antigen after separation by SDS-PAGE and transfer to nitrocellulose membranes, were detected by protein G or L. The results showed that antibody production against five of the antigens, albumin, alpha 1 gamma-acid glycoprotein, alpha 1 gamma-microglobulin, Ig light chains, and retinol-binding protein, showed a similar pattern, although the magnitude of the initial IgM response differed somewhat. After 6 weeks, the levels of the protein G-binding antibodies had reached a plateau, while those of protein L-binding antibodies were still increasing. The response to the sixth antigen, beta 2 microglobulin, was considerably different. A dramatic increase of anti-beta 2 gamma-microglobulin antibodies was seen during the 4th week after immunization when protein L was used.