MutK may be a 32 kDa protein and be widespread among Proteobacteria
D. Moreira
DOI: https://doi.org/10.1046/j.1365-2958.1999.01444.x
1999-07-01
Molecular Microbiology
Abstract:The repair of DNA damage is essential to assure genome maintenance and integrity in all living beings. Many genes with diverse functions are concerned with the repair pathways. Among them, the so-called mutator genes have attracted much attention. The methyl-directed mismatch repair system, which recognizes and resolves incorrectly paired nucleotides that derive both from replication errors (Lahue et al., 1989, Science 245: 160±164) and from mismatched apareaments during recombination processes (Worth et al., 1994, Proc Natl Acad Sci USA 91: 3238± 3241), involves the products of the mutS, mutL, mutH and mutU mutator genes, together with that of dam (for a review, see Modrich, 1991, Annu Rev Genet 25: 229± 253). The very short patch repair also requires the mutS and mutL gene products and those of polA , dcm and vsr (Lieb et al., 1986, Genetics 114: 1041±1060). Recently, a new mutator gene, mutK, has been characterized in the pathogenic g-proteobacterial species Vibrio cholerae (Bhakat et al., 1999, J Bacteriol 181: 879±883). It encodes a protein of 163 amino acids that seems to be involved in methyl-independent repair of DNA mismatches. When cloned and expressed in Escherichia coli, the spontaneous mutation frequency of mutS, mutL, mutU and dam mutants decreases. The mutK-mediated repair appears to constitute an alternative pathway, operative when the major repair pathway is absent. Bhakat et al. (1999, J Bacteriol 181: 879±883) reported the sequence of a V. cholerae genomic fragment containing the mutK gene (accession number Y11983). This gene is located downstream from an open reading frame (ORF2) that encodes an unknown protein and upstream from a putative S-adenosyl-Lmethione-dependent methyltransferase. Bhakat et al. (1999, J Bacteriol 181: 879±883) failed to ®nd any mutK homologue in other bacterial species. However, a thorough search of sequence data from both ®nished and ongoing bacterial genome sequence projects at the NCBI (http:/ /www3.ncbi.nlm.nih.gov/BLAST/ un®nishedgenome.html) and WIT (http:/ /wit.mcs.anl.gov/ WIT/) Blast servers allows the retrieval of mutK homologues in several bacterial species (Fig. 1), including members of the a, b and g subdivisions of the Proteobacteria (expectation values between 10 and 10), Mycobacterium tuberculosis (expectation value of 10) and the radio-resistant bacterium Deinococcus radiodurans (expectation value of 10). It is not found in the complete genomes of several proteobacteria, such as Rickettsia prowazekii, Haemophilus in ̄uenzae and Helicobacter pylori. In E. coli, the mutK homologue maps close to vsr and dcm (sequence accession number AE000287 for this section of the E. coli genome). mutK is also found close to these DNA repair genes in Salmonella typhi, but at different genomic locations in other species (data not shown). Interestingly, all the mutK homologues retrieved are predicted to have an around 140-amino-acid N-terminal extension that is absent from the reported V. cholerae sequence. Consequently, their predicted sizes are around 300 amino acids (i.e. expected molecular weights of <32 kDa instead of 19 kDa). As this additional segment is highly conserved, it seems likely that it may be functionally important. Therefore, I have searched for this sequence in the V. cholerae genome project using the TIGR Blast facility (http:/ /www. tigr.org/cgi-bin/BlastSearch/blast.cgi?organismv_ cholerae). Surprisingly, a clear homologous sequence (expectation value of 10 using the E. coli sequence as query) is found adjacent to the mutK sequence (contig asm826). It is identical to the ORF2 reported by Bhakat et al. (1999, J Bacteriol 181: 879±883). No stop codon is present in the stretch between ORF2 and mutK in the sequence released by the V. cholerae genome project. Therefore, as occurs in the other bacteria, the predicted protein corresponds to one single product derived from the fusion of ORF2 and mutK. The comparison with the DNA sequence reported by Bhakat et al. (1999, J Bacteriol 181: 879±883) shows that both are identical except for the deletion of an adenine at position 538 of the latter. As a consequence of this deletion, a stop codon introduced at position 558±560 is absent in the genome project sequence. As this stop codon is also absent in all other known mutK homologues, a sequencing error may be suspected. Interestingly, these authors found that the expression of the cloned fragment yields two proteins of 14 kDa and 19 kDa that appear to be the products of ORF2 and mutK. However, the gel shown in their report also exhibits an unexpected band of around 31 kDa that could correspond to the intact product of the putative non-truncated mutK gene. All the above strongly suggests that mutK actually consists of the entire region corresponding to the ORF2 and mutK genes, which might have been incorrectly split by Bhakat et al. (1999, J Bacteriol 181: 879±883) report. Nevertheless, these authors detected a putative ribosome binding site (RBS) that could be involved in the translation of the 19 kDa protein and, in fact, a clone lacking the ORF2 Molecular Microbiology (1999) 33(1), 220±222