A new method for site-specific protein-DNA crosslinking was developed. There is a need for new, improved and more efficient methods because existing strategies generally give very low yield of crosslinked complexes, or are very unselective. The chosen strategy involved the formation of disulfide bonds between the protein and DNA. This required creation of a thiol at a specific base in the DNA and a cysteine thiol at a specific residue in the protein. Nucleotide analogues with a thiol at the C-5 position of the pyrimidine ring were made as the phosphoramidite derivative. This allowed the thiol-containing analogue to be incorporated into oligonucleotides. The crosslinking strategy was tested by attempting to crosslink the site-specific recombinase Tn5 resolvase to its DNA binding site. An X-ray structure of Tn3 resolvase bound to DNA has been solved, making it a good model on which to test the method. Cysteine mutants of the Tn3 resolvase DNA binding domain were created, at residues predicted to be close to the modifiable thymidine residues at the binding site. The mutations did not abolish binding and recombination by the protein. Binding analysis of the cysteine mutants with the oligonucleotides did indicate that the binding affinity of the proteins was reduced. Some crosslinked complexes to the thiol-containing oligonucleotides were observed, albeit in much lower yield than had originally been hoped for. The stereochemical course of the strand exchange reaction catalysed by Tn3 resolvase was investigated by a strategy involving a phosphorothioate linkage placed at the point of resolvase-catalysed DNA strand cleavage.