DNA repair enzymes play a pivotal role in the maintenance of chromosom
e integrity and in the elimination of premutagenic lesions from DNA by
patrolling the genome; nuclear import mechanisms are implicated in mo
lecular carcinogenesis. We have attempted to predict cell trafficking
and the nuclear importation of proteins involved in DNA repair by sequ
ence analysis aimed at identifying karyophilic clusters (arginines, ly
sines, histidines) flanked by the helix breakers proline or glycine th
at could function as nuclear localization signals (NLSs). Most mammali
an proteins that participate in DNA repair pathways seem to possess NL
S peptides. Repair proteins with multiple nuclear signals are the ERCC
6 helicase (eight signals), the XPC protein involved in the repair of
the transcribed strand in active genes (eight strong and seven weak si
gnals), and the Rep3/Duc-1 mismatch repair protein (five strong one we
ak signal). We propose that it is unlikely to identify mutations on th
e genes encoding these proteins resulting in cytoplalsmic retention. H
owever, a number of mammalian DNA repair proteins lack NLS clusters; t
hese proteins include ERCC1, ERCC2 (XPD), mouse RAD51, and the HHR23B/
p58 and HHR23A subunits of XPC. NLS-less S. cerevisiae proteins includ
e both RAD51 and RAD52 that function in the recombination and in the r
epair of double-strand breaks as well as the RAD23 and HRR25 molecules
. We propose that these proteins depend on their complexation with oth
er proteins in the cytoplasm for their nuclear localization. The hMSH2
human mismatch repair protein, linked to the hereditary nonpolyposis
colon cancer gene, has a weak nuclear signal containing two histidines
.