In E. coli, viable polA mutants have been isolated that produce a defective gene product with little or no 5′→3′ polymerase activity, but normal 5′→3′ exonuclease activity. However, no polA mutant has been identified that is completely deficient in the 5′→3′ exonuclease activity, while retaining 5′→3′ polymerase activity, of DNA polymerase I. How can these results be explained?