EXPRESSION & MUTAGENESIS OF TUMOUR NECROSIS FACTOR BETA

Abstract

The work described in this thesis covers two areas : 1) achievement of good expression of recombinant human tumour necrosis factor beta (rhTNF- ?) in E.coli and 2) identification of the receptor binding region of TNF- ? by site-directed mutagenesis. Inability to obtain large quantities of rhTNF-? has been a stumbling block to the study of this cytokine. To achieve high expression of full length human TNF-? in E.coli, two sequences coding for human TNF-? based on the natural cDNA were manipulated in pBR322-based plasmids under the control of the E.coli tryptophan promoter. Expression of over 30% total bacterial protein was obtained. Conditions were optimized for production of TNF-? in the soluble fraction by growth at 25°C at pH 6.4. TNF-? was purified from the soluble fraction of E.coli proteins using a DEAE-Sepharose CL6B column in one step. A single band of 18.7 kDa, the expected size of unglycosylated human TNF-?, was obtained on silver staining of an SDS-polyacrylamide gel. Its identity was confirmed by Western blotting with a specific polyclonal antibody to human TNF-? and by protein microsequencing of the amino terminus. The specific activity of TNF-? produced was 3-5 x107 units/mg using the cytotoxic assay on mouse L929 cells. To define regions of the molecule essential for receptor binding and biological activity, single amino acid substitutions were made at predicted hydrophilic loop regions in residues conserved between TNF-? and TNF-?, a related cytokine that binds to the same cellular receptors. Site-directed mutagenesis using the Mandecki method was done using the plasmid which gave high expression of TNF-? and also with phagemids derived from the parent plasmid. Mutants were expressed and screened for cytotoxic activity and receptor binding. Selected mutants were purified. Specific activity was measured and receptor binding studies were performed on the purified mutant proteins. Conservative changes at either of two amino acid positions in TNF-?, aspartic acid-50 and tyrosine-108, gave rise to mutants which lost both biological activity and ability to bind to the TNF receptor. Similar changes in neighbouring residues had no such effect. Cross-linking experiments showed that the inactive mutants were able to form dimers and trimers. Studies on circular dichroism showed that spectra obtained from the mutants were very similar to the parent TNF-?. Hence, the mutations do not interfere with trimerization nor disrupt the overall structure of the molecule. The changes at aspartic acid-50 and tyrosine-108 are therefore likely to be specifically involved in receptor binding . Aspartic acid-50 and tyrosine-108 lie on surface loops which line opposite sides of the cleft between two different subunits in the TNF-? trimer. Evidence from TNF-? mutagenesis shows that biological activity and receptor binding is lost on mutation of homologous surface loops lining the cleft between TNF-? subunits. These two lines of evidence suggest that the TNF receptor interacts with amino acid residues which line the cleft between different subunits of the TNF trimer. As each trimer has three such grooves, it is possible for three TNF receptors to bind at the same time to a TNF trimer, thus suggesting a novel mechanism of effecting receptor clustering.