MutSβ exceeds MutSα in dinucleotide loop repair

The five proteins involved in the human mismatch repair (MMR) mechanism to maintain genomic integrity function as heterodimers are MutLα (MLH1+PMS2), MutSα (MSH2+MSH6) and MutSβ (MSH2+MSH3). MMR proteins correct base/base mismatches and small insertion/deletion loops (IDLs) that arise on the newly synthesised strand during DNA replication and recombination. Larger loop structures (⩾5 nt) are believed to require a different combination of repair proteins and hence are not targets of the MMR mechanism (Umar et al, 1998). Approximately 25% of sporadic colon tumours, as well as a number of tumours of endometrium, ovary and some other organs and tissues, are deficient in MMR (Peltomaki, 2003). Moreover, germline mutations in MMR genes predispose to hereditary nonpolyposis colorectal cancer (HNPCC) syndrome/Lynch syndrome. To date, 659 MLH1 (44% of all identified MMR gene variations), 595 MSH2 (39%), 216 MSH6 (14%) and 45 PMS2 (3%) germline variations have been reported in the database (Woods et al, 2007; http://www.med.mun.ca/MMRvariants/). However, no HNPCC predisposing MSH3 mutations have yet been identified. MMR-deficient tumours are strongly associated with microsatellite instability (MSI) (Aaltonen et al, 1993). However, the degree and type of MSI differ from high to low and between mono-, di-, tri- and tetranucleotide instability or elevated microsatellite alterations at selected tetranucleotide repeats (EMAST) (Peltomaki and Vasen, 2004; Plaschke et al, 2004; Haugen et al, 2008) depending on the MMR gene affected. MLH1- and MSH2-deficient tumours are characterised by both mono- and dinucleotide repeat instability, whereas the level of MSI is lower in MSH6-deficient tumours (Bhattacharyya et al, 1995; Papadopoulos et al, 1995). MSH6-deficient cells are unable to repair single base mismatches, whereas they retain proficiency to repair two, three and four base loops (Drummond et al, 1995; Risinger et al, 1996; Umar et al, 1997), thus, causing only mononucleotide repeat instability in tumours (Wagner et al, 2001; Plaschke et al, 2004). Recently, EMAST and also low dinucleotide repeat instability have been associated with MSH3 deficiency both in tumour cell lines and in sporadic colorectal tumours (Haugen et al, 2008). The type of MSI seems to be dependent on the substrate specificities of the MMR protein affected. In human cells, the MMR process is initiated by the binding of the mismatch recognition factor MutSα or MutSβ to the mispair, followed by the initiation of the assembly of the repairosome by MutLα (Constantin et al, 2005; Zhang et al, 2005). MutSβ has a high binding affinity to IDLs but, in contrast, a very low affinity to simple base/base mispairs (Acharya et al, 1996; Palombo et al, 1996), whereas MutSα has been shown to bind and repair both base/base mispairs and IDLs (Drummond et al, 1995; Palombo et al, 1996). Lesion specificity is believed to lie within the MSH3/MSH6-specific sequences, which differ notably (Owen et al, 2009). The process through which ADP–ATP exchange occurs on MSH2 seems to be dependent on the protein it forms a complex with; MSH6 requires ATP stabilisation, whereas MSH3 requires ATP hydrolysis, both of which are dependent on specific lesion binding (Owen et al, 2009). However, findings based on assays analysing the binding properties of these MMR proteins do not yet prove their functional ability to repair the bound mismatches (Ou et al, 2007). In this study, we applied the in vitro MMR assay to analyse the substrate specificities and functionality of MutSα and MutSβ using substrates, GT, IDL1 and IDL2 in three different cell lines. The in vitro MMR assay allows the functional analysis of all different MMR protein complexes and all kinds of missense variations in individual genes in a homologous human MMR system. In this study, the assay was for the first time applied to test the interference of an MSH3 variation with repair efficiency.