For lagging-strand DNA replication, short RNA primers (blue) are made by RNA primase. These are then extended by DNA polymerase to form Okazaki fragments. When these RNA primers are removed, there is no way to synthesize lagging-strand sequence that is complementary to the small region at the end of the chromosome (which is at least as large as an RNA primer). So, with continuing cell division, sequence is lost from the ends of linear chromosomes. Some blunt-ended daughter molecules are produced by this scheme, irrespective of whether the starting terminus is blunt-ended (a) or has a 3' extension (b).

Why are such natural blunt ends not recognized as DNA damage? One possibility is that this is because unnatural ends have a slightly different chemical structure; for example, radiation-induced blunt ends can have terminal 3' phosphoglycolate residues whereas physiological ends do not. Human cells may have an additional mechanism to ensure that natural chromosome ends are even more different. This involves a degradative pathway which, even in telomerase-deficient human cells, results in 3' extensions for most chromosome ends9. This may be analogous to a degradative pathway that has been described in budding yeast, involving Cdc13p and other proteins10.