Coupling of adenine nucleotide binding and hydrolysis to single- and double-stranded DNA binding determines the topoisomerase activity of reverse gyrase from "Thermotoga maritima"

The aim of this PhD thesis was to functionally characterise the hyperthermophilic topoisomerase IA reverse gyrase from the eubacterium Thermotoga maritima with respect to its unique ATP-dependent positive plasmid supercoiling activity. We set out to apply single molecule FRET to observe conformational changes in the N-terminal helicase-like domain and the C-terminal topoisomerase domain of reverse gyrase during topoisomerase activity. Initially, the protocol for the purification of T. maritima reverse gyrase was improved to yield mg amounts of monomeric enzyme with 95% purity. Requirements for positive plasmid supercoiling by reverse gyrase from T. maritima were defined and optimal reaction conditions were established for the rest of the work. However, no universal nucleotide utilisation pattern for topoisomerase activity of reverse gyrases in general exists and various heterogeneous nucleotides-dependent topoisomerase activities have been reported for reverse gyrases from different organisms. Reverse gyrase from T maritima exhibits no topoisomerase activity in the absence of nucleotides but relaxes plasmids in an ADP- and ADPNP-dependent manner. Reverse gyrase positively supercoils plasmid DNA in the presence of ATP. Surprisingly, hydrolysis of ATPγS efficiently promotes positive supercoiling to a similar extend as ATP hydrolysis. Mutation of the GKT sequence in the Walker A motif of the helicase-like domain renders reverse gyrase inactive for nucleotide hydrolysis and reveals vastly reduced topoisomerase activity. Most interestingly, we observed AMP generation in the presence of short double-stranded DNA substrates and plasmid DNA. During plasmid relaxation, ADP and ADPNP are converted into AMP. The same is true for ATP and ATPγS during positive supercoiling. Possibly, ATP is hydrolysed to AMP by reverse gyrase via intermediately generated ADP and energy may be obtained from ADP cleavage. Further investigating the function of the reverse gyrase domains, we demonstrated that the helicase-like domain is a DNA-stimulated ATPase that harbours all determinants for adenine nucleotide binding and hydrolysis of reverse gyrase. The full-length enzyme shows highly reduced ATPase compared to the isolated helicase-like domain even in the presence of DNA. Consequently, the coupling of DNA binding to ATP hydrolysis is diminished in full-length reverse gyrase and the effect of DNA binding on KM,ATP is much smaller than for the isolated helicase-like domain. Thus, the helicase-like domain of reverse gyrase is a nucleotide-dependent switch. Notably, full-length reverse gyrase from T. maritima binds DNA much more tightly compared to the helicase-like domain and single-stranded DNA binds in a cooperative manner. Artificial 20-, 40- and 60-mer DNA constructs were used to elucidate cooperative binding of reverse gyrase. We show that the binding affinity increases with substrate length. In contrast, the DNA-dependent ATPase activity of T. maritima reverse gyrase decreases with substrate length. Cooperative binding was demonstrated for single-stranded DNA substrates longer than 40 nt and points to possible protein-protein interactions of reverse gyrase that may play a role in the positive supercoiling cycle. In order to observe conformational changes in reverse gyrase with smFRET during the supercoiling cycle, fluorescent labelling is required. However, labelling of reverse gyrase cysteine mutants for smFRET measurements is quite challenging. Reverse gyrase from T. maritima contains eight native cysteines in two putative zinc fingers that have to be intact for positive supercoiling activity. Hence, we improved the conditions accordingly for selective labelling of reverse gyrase cysteine mutants. smFRET measurements indicated that the helicase-like domain might be in a more closed conformation compared to the available crystal structure from A. fulgidus reverse gyrase. Furthermore, the gap between the latch region and the topoisomerase domain is suggested to open during the supercoiling cycle, but the calculated FRET distances correspond to the closed conformation observed in the crystal structure. Conformational changes during the catalytic cycle of reverse gyrase have been predicted for the helicase-like domain and the topoisomerase domain. No conformational changes in the helicase-like domain and the topoisomerase domain are detected with different DNA substrates and adenine nucleotides. Thus, the labelled reverse gyrase mutants may either be inactive under the conditions chosen for smFRET measurements or might have been inactivated by fluorescent labelling. Other positions will be chosen to investigate regions prone to undergo conformational changes during the catalytic cycle without impairing the topoisomerase activity of T. maritima reverse gyrase.