One of the events that can change the DNA sequence of a cell is site-specific recombination, which rearranges DNA by breaking and rejoining it at two specific sites. Site-specific recombinases, the enzymes that catalyse such processes, have recently become a very popular tool for manipulating DNA in vitro and in vivo, mainly for transgenic studies. These uses of site-specific recombinases normally require the introduction of target DNA containing their recombination site(s) into the organism. Taking that into account, there is currently a great deal of interest in modifying these enzymes so as to recognize natural sequences in eukaryote (e.g. human) genomes, for use in gene therapy approaches. However, these attempts at re-engineering site-specific recombinases to obtain new target specificity have until now yielded quite modest results. The work described here is centred on Tn3 resolvase, a member of the 'serine recombinase' family of site-specific recombinases. We aimed to create a site-specific DNA recombinase that can be designed at will to recognize and act at natural genomic sequences, by combining a recombinase module, based on the Tn3 resolvase's N-terminal domain, with a recognition module which would be an engineered zinc finger DNA- binding domain. Hybrid DNA recombinases were produced comprising the zinc finger DNA-binding domain of the Zif268 mouse transcription factor and the catalytic domain of Tn3 resolvase (a mutant version of, called Rnm)- The two-domain hybrid resolvase (referred to as "Z- resolvase") was found to be able to bind to and specifically recombine DNA at two copies of an artificially-designed target sequence referred to as "Z-site", consisting of the crosssover region of site I of res flanked on each side by a binding motif of Zif268. This design of the enzyme and its target site was further improved by cross-selection of a series of their variants with modifications in the inter-unit spacing region. In the improved configuration, the enzyme efficiently resolved its substrate DNA in vivo in E. coli, as evidenced by change in the colour of transformed colonies on an indicator growth medium and subsequent analysis of the plasmid DNA from the cells. Resolution was found to occur specifically at the target sites. The efficiency of resolution in E. coli was estimated at over 99%. Examination of the enzyme's activity in vitro revealed that at high concentrations it is able to recombine most of its target substrate. It both resolved and inverted the DNA between its two target sites, as well as carried out cleavage and intermolecular recombination of the substrate plasmid. When assayed on DNA with two hybrid sites, each of them being composed of a half-Z site joined to a half of site I, Z-resolvase and complemented each other in resolving the substrate. They did not complement each other on a substrate with a full site I and a full Z-site. The results described above could potentially open the way for creation of 'custom-built' site-specific DNA recombinases, to recombine at any naturally-occurring sequence of interest. If successful, this would have important implications in both medicine and biology.