Oncoprotein E6 is essential for oncogenesis induced by human papillomaviruses (HPVs). The solution structure of HPV16-E6 C-terminal domain reveals a zinc binding fold. A model of full-length E6 is proposed and analyzed in the context of HPV evolution. E6 appears as a chameleon protein combining a conserved structural scaffold with highly variable surfaces participating in generic or specialized HPV functions. We investigated surface residues involved in two specialized activities of high-risk genital HPV E6: p53 tumor suppressor degradation and nucleic acid binding. Screening of E6 surface mutants identified an in vivo p53 degradation-defective mutant that fails to recruit p53 to ubiquitin ligase E6AP and restores high p53 levels in cervical carcinoma cells by competing with endogeneous E6. We also mapped the nucleic acid binding surface of E6, the positive potential of which correlates with genital oncogenicity. E6 structure-function analysis provides new clues for understanding and counteracting the complex pathways of HPV-mediated pathogenesis.

To create a rapid system to test the effect of sequence changes on recombinant antibody binding, we have developed a procedure for producing functional scFv fragments in an Escherichia coli cell-free translation system. Functional antibodies with antigen-binding activity are obtained only if disulfide formation and rearrangement is allowed to take place during the translation reaction. The inclusion of protein disulfide isomerase (PDI) leads to a threefold increase in yield over that obtained in the presence of glutathione redox systems. DsbA had no such effect, indicating that disulfide shuffing, and not net formation, is the crucial yield-limiting step. The addition of the molecular chaperones DnaK and DnaJ increased the amount of soluble protein but not the amount of functional scFv, which appears to be limited entirely by correct disulfide formation. None of these factors significantly influenced total protein synthesis. In the presence of PDI, chaperones, reduced glutathione and oxidized glutathione, 50% of the scFv produced (about 8 micrograms/ml in only 15 min) could be recovered from immobilized antigen.

Antibodies are valuable tools for functional studies in vitro, but their use in living cells remains challenging because they do not naturally cross the cell membrane. Here, we present a simple and highly efficient method for the intracytoplasmic delivery of any antibody into cultured cells. By following the fate of monoclonal antibodies that bind to nuclear antigens, it was possible to image endogenous targets and to show that inhibitory antibodies are able to induce cell growth suppression or cell death. Our electrotransfer system allowed the cancer cells we studied to be transduced without loss of viability and may have applications for a variety of intracellular immuno-interventions.

No abstract

Antibody molecules are able to recognize any antigen with high affinity and specificity. To get insight into the molecular diversity at the source of this functional diversity, we compiled and analyzed a non-redundant aligned collection of 227 structures of antibody-antigen complexes. Free energy of binding of all the residue side chains was quantified by computational alanine scanning, allowing the first large-scale quantitative description of antibody paratopes. This demonstrated that as few as 8 residues among 30 key positions are sufficient to explain 80% of the binding free energy in most complexes. At these positions, the residue distribution is not only different from that of other surface residues but also dependent on the role played by the side chain in the interaction, residues participating in the binding energy being mainly aromatic residues, and Gly or Ser otherwise. To question the generality of these binding characteristics, we isolated an antibody fragment by phage display using a biased synthetic repertoire with only two diversified complementarity-determining regions and solved its structure in complex with its antigen. Despite this restricted diversity, the structure demonstrated that all complementarity-determining regions were involved in the interaction with the antigen and that the rules derived from the natural antibody repertoire apply to this synthetic binder, thus demonstrating the robustness and universality of our results.

Recombinant proteins with cytosolic or nuclear activities are emerging as tools for interfering with cellular functions. Because such tools rely on vehicles for crossing the plasma membrane we developed a protein delivery system consisting in the assembly of pyridylthiourea-grafted polyethylenimine (πPEI) with affinity-purified His-tagged proteins pre-organized onto a nickel-immobilized polymeric guide. The guide was prepared by functionalization of an ornithine polymer with nitrilotriacetic acid groups and shown to bind several His-tagged proteins. Superstructures were visualized by electron and atomic force microscopy using 2 nm His-tagged gold nanoparticles as probes. The whole system efficiently carried the green fluorescent protein, single-chain antibodies or caspase 3, into the cytosol of living cells. Transduction of the protease caspase 3 induced apoptosis in two cancer cell lines, demonstrating that this new protein delivery method could be used to interfere with cellular functions.

We have developed a chloramphenicol resistant derivative of fd phage with which cognate pairs of antibodies and antigens can be selected. The phage genome encodes a fusion of single-chain antibody to the C-terminal domain of glIIp, rendering the phage non-infective. The antigen fused to the N-terminal domains of glIIp is encoded in the same phage genome. Antigen and antibody fusion interact with each other in the periplasm of the phage-producing cell, restoring infectivity. This system has a very low background and will allow simultaneous randomisation of antibody and antigen.

A systematic study has been performed on the relationship between linker length, relative orientation of variable domains, multimerization behaviour and antigen binding activity for single chain Fvs (scFvs) of the tumour-binding antibody B72.3. Thirteen scFv variants with linkers comprising up to six repeats of the motif Gly-Gly-Gly-Gly-Ser were studied. All these scFvs showed a tendency to form dimers or higher molecular weight species, and this tendency decreased with increasing linker length. The dimers and higher molecular weight forms may arise from head to tail intermolecular association of VH and VL domains. For each linker length, scFvs with the organization VL-linker-VH showed greater binding activity than those with the organization VH-linker-VL. In fact, for the latter organization only the variant with a 30 amino acid linker showed good binding activity, suggesting that (i) for B72.3 the C-terminus of VH or the N-terminus of VL makes a structural contribution to antigen binding, and (ii) shorter linkers interfere with this contribution. Antigen binding studies on scFvs should be interpreted with caution because of their tendency to multimerize. Such multimerization can be minimized by using linkers longer than those in common use.