DNA mismatches - their structure and recognition by MutS

Whilst DNA replication is a highly efficient and accurate process, mispairing and base damage can occur which, if left unchecked, could prove deleterious to the organism. Mismatches arise naturallyfromthree major sources: by misincorporation during replication, from damage to bases and in recombination termediates. Restoration and maintenance of genetic fidelity is clearly of great importance. Several DNA repair pathways have evolved to remove damaged or mispaired bases and restore complementarity to damaged DNA. One such pathway is the MutHLS mismatch repair system of E.coli. In this system, the 97kDa MutS protein recognises and binds to DNA mismatches or short loops of unpaired bases. After recognition by MutS, other components of the pathway act to excise the DNA strand containing the incorrect base and resynthesise afiiUycomplementary strand. The mode of recognition of mismatches by MutS remains unknown, but may involve recognition of local DNA structure and/or dynamics.We have studied the effects of different mismatches on DNA structure in solution by examining how they affect cleavage by both chemical and enzymatic footprinting probes including DNase I, micrococcal nuclease, hydroxyl radicals and MPE-Fe(II), comparing cleavage with an homologous Watson-Crick duplex. It appears that although mismatches often introduce changes in local helix structure and/or thermodynamics, there is no correlation between the affinity of MutS for a particular mismatch and its enzymic and chemical cleavage patterns. It is therefore likely that MutS probing of local helix distortions is not a major factor in recognition.Additionally, we have used band shift assays and footprinting to examine the binding of MutS to synthetic DNA fragments containing each of the eight possible mismatches or an unpaired T (ΔT) in a variety of sequence contexts. The order of affinity determined from band shift assays is ΔT > G·T > G·G > A·A ≈ T·T ≈ T·C > C·A > G·A > C·C > G·C. Footprinting confirms the greater affinity for ΔT than for G·T. The effect of sequence context is such that recognition is best in a polyG:polyC environment, with the observed trend in affinity being GnCn > (GC)n > AnTn > (AT)n. These studies suggest that the binding affinity of MutS is not the sole determinant of repair efficiency. Experiments using base analogues were also performed to investigate the importance of functional groups in recognition.Whilst no detailed structural data on MutS is yet available, a model has been proposed which suggests a multi-domain protein with the N- and C-terminal domains both apparently located in the vicinity of the mismatch. We have used the pTX412 plasmid containing the mutS gene as a PCR template for the cloning of various MutS protein fragments which have been expressed in an E.coli BL21 (DE3) host (three N-terminal fragments of sizes 63.8 kDa, 46.5 kDa and 26.1 kDa and one C-terminal fragment of 28.7 kDa). These fragments were purified from inclusion bodies using 5M Guanidine hydrochloride and refolded, either singly or in combination, by stepped dialysis from 5M to 0M Guanidine hydrochloride. Unfortunately, no mismatch binding activity was detectable using band shift assays.