The DNA mismatch repair system (MMR) is highly conserved among different organisms like E. coli or H. sapiens. In E. coli the MMR system consists of three main proteins, MutH, MutS and MutL. The MutS protein is responsible for the recognition and binding to mispaired nucleotides and small single strand DNA loops. In eukaryotes the mismatch repair machinery is much more complex. For example, in S. cerevisiae six different MutS (MSH) and 4 MutL (MLH) homologs have been found. In yeast and human it has been demonstrated that MSH2 forms heterodimers with MSH6 (involved in repair of basepair mismatches and 1 nucleotide loops) and MSH3 (involved in repair of small single strand DNA loops).

As in all known MutS homologous proteins certain amino acids in the C-terminal part are highly conserved, we were able to search for homologous sequences by RT-PCR with degenerate oligonucleotides. Using this approach, we isolated two MutS-homologous cDNAs, mus1 and mus2. Sequence comparisons with MutS-homologs from bacteria, yeast and mammals revealed that mus1 is most closely related to MSH2 from yeast (MNL 71, pp. 63-64, Figure1) and Arabidopsis, whereas mus2 is most similar to MSH6 genes. By Southern blotting we found that mus2 is � like mus1 � a single copy gene. The mus2-probe detects on a Northern blot a very weakly expressed transcript with a length of approximately 4.1 kb. With the same probe we isolated a 2.9 kb cDNA clone lacking the 5�- and 3�-ends from a maize seedling cDNA library (kindly supplied by Monika Frey, Technical University of Munich). The sequence analysis of a genomic Lambda clone suggests that the coding region is at least 3.5 kb long and codes for a putative protein of 1185 amino acids. The 5�- and 3�-ends of the cDNA have to be confirmed yet.

Figure 1 shows the alignment of the most highly conserved regions of MSH2 and MSH6 proteins. It turns out that the sequences of the MSH2 proteins are more conserved over the region of the four putative nucleotide binding domains, and especially in the helix-turn-helix motif.

In collaboration with Monika Frey (Technical University of Munich) we RFLP-mapped the mus2 gene by using a recombinant inbred population (Burr et al., Genetics 118:519-526,1988; Burr and Burr, TIG 7: 55-60, 1991). Mus2 maps on chromosome 3S.

With the aim of investigating the biochemical properties of the maize Mus1 and Mus2 proteins, we began to establish an overexpression system in E. coli. By optimization of the expression and purification conditions we were able to produce Mus1 in a soluble form. However, in gel shift experiments with DNA probes containing a mismatch this protein exhibits no mismatch specific binding activity.

In analogy to the properties of MSH2 and MSH6 proteins in yeast and mammals (Palombo et al., Science 268:1912-1914, 1995; Iaccarino et al., EMBO J. 17: 2677-2686, 1998), it seems possible that Mus1 is only functional as a heterodimer with MSH3 or MSH6 proteins.

To this end we have first hints for an interaction between Mus1 and Mus2, as the latter can only be overexpressed in E. coli by coexpressing Mus1. This implies some kind of stabilization of Mus2 by Mus1.

Figure 1. Alignment of the highly conserved C-terminal regions of MSH2 and MSH6 homologous proteins. The maize homologs Mus1 and Mus2 are underlined. The putative nucleotide binding regions are shown in bold letters, the putative helix-turn-helix motif is underlined. atMSH - Arabidopsis thaliana: MSH2 (3914056); hMSH - Homo sapiens: MSH2 (1171032), MSH6 (1082386); yMSH - Saccharomyces cerevisiae: MSH2 (2506880), MSH6 (3024187); mus1 and mus2 - Zea mays (Genbank accession numbers are shown in parentheses)
 

Return to the MNL 73 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage