Genetic and biochemical analysis of the synaptic complex of invertase Gin.

The Gin inversion system of bacteriophage Mu requires the formation
of a synaptic complex of unique topology, where the two Gin dimers
bound at the recombination gix sites are interacting to form an
enzymatically active tetramer, which then catalyses the
site-specific recombination reaction. After the assembly of the
synaptic complex the DNA strand cleavage is activated by the
DNA-bending protein FIS bound at the recombinational enhancer
sequence. During reaction the complex undergoes conformational
changes resulting in a site-specific inversion of a DNA segment in
the phage Mu genome. In this thesis the protein interactions in the
synaptic complex were analysed. First, the question on the
interactions between FIS and Gin during formation of the synaptic
complex was addressed. In a genetic test system a mutant fisS14P
has been selected that can rescue the recombination-deficient
phenotype of the mutant Gin H106T. FIS S14P was shown to activate
the Gin H106T mutant in vivo but not in vitro. The possible reasons
are the differences in the in vivo and in vitro conditions, and the
observed altered DNA bending ability of the FIS S14P mutant. The
position of the mutation S14P in the “β-hairpin arm” of the FIS
N-terminus suggests it could directly interact with the hydrophobic
dimerisation interface of Gin around the position H106. Next, the
predictions of the preliminary model of the Gin invertasome
organisation have been verified and the catalytic domains of Gin
were demonstrated indeed to be involved in tetramer formation. To
do this, specific mutations at the proposed synaptic interfaces
were introduced and biochemical studies of different mutants of Gin
invertase affected in their ability to promote synapsis were
performed. It was possible to show that in addition to the already
identified surfaces of the Gin dimer-dimer interactions, comprising
of the αE helix and the flexible loop between the β2 sheet and the
αB helix of Gin, also the αD helix and the loop between αA helix
and β2 sheet are involved in the stabilisation of the Gin tetramer.
Cysteine substitutions placed on these surfaces could be
efficiently cross-linked in the tetramer in the presence of DNA and
FIS, indicating their close proximity in the synapse. Furthermore,
Gin mutants with either increased or decreased tetramerisation
abilities were isolated and characterised, and the effects of these
mutations on recombination were studied. These data led to the
notion that the tetramer structure should be flexible, since all
mutations that stabilise the complex cause inversion deficiency. In
turn, the complexes formed by the hyperactive mutants seem to have
high conformational flexibility, although at the expense of the
loss of specificity. Notably, introduction of substitutions that
stabilise the Gin tetramer also lead to suppression of hyperactive
features. A chimeric recombinase protein, containing the N-terminal
catalytic domain from Gin and the DNA-binding domain of ISXc5
resolvase, was found to form a more stable tetramer complex, than
Gin. The chimera ISXc5G10 is inversion deficient, but can still
catalyse resolution. Again, these observations support the notion
that the stabilisation of the tetramer can strongly impair the
ability to catalyse inversion, but may have less effect on the
resolution activity. The DNA-binding domain of ISXc5G10 chimera was
mutagenised to obtain a protein with an inversion proficient
phenotype, but no mutants of this type could be found, perhaps
because in the chimera not only the DNA binding domain, but the
gross organisation of the protein is different. Thus, according to
the obtained data the Gin dimers bound to the recombination sites
are interacting with each other via catalytic domains and
recombination involves gross reorganisations of contact surfaces.
The obtained results allowed to clearly distinguish between the two
previously proposed mechanistically different models of
recombination (the “subunit exchange” and “static subunits”
models), and favour the “subunit exchange” model. Such a model
serves as a useful working hypothesis for future experiments
dedicated to the detailed understanding of the mechanism of
recombination reaction catalysis by members of the serine
recombinase family.