For many years it has been known that thymine auxotrophic microorganisms undergo cell death in response to thymine starvation [thymineless death (TLD)]. This effect is unusual in that deprivation of many other nutritional requirements has a biostatic, but not lethal, effect. Studies of numerous microbes have indicated that thymine starvation has both direct and indirect effects. The direct effects involve both single- and double-strand DNA breaks. The former may be repaired effectively, but the latter lead to cell death. DNA damaged by thymine starvation is a substrate for DNA repair processes, in particular recombinational repair. Mutations in recBCD recombinational repair genes increase sensitivity to thymineless death, whereas mutations in RecF repair protein genes enhance the recovery process. This suggests that the RecF repair pathway may be critical to cell death, perhaps because it increases the occurrence of double-strand DNA breaks with unique DNA configurations at lesion sites. Indirect effects in bacteria include elimination of plasmids, loss of transforming ability, filamentation, changes in the pool sizes of various nucleotides and nucleosides and in their excretion, and phage induction. Yeast cells show effects similar to those of bacteria upon thymine starvation, although there are some unique features. The mode of action of certain anticancer drugs and antibiotics is based on the interruption of thymidylate metabolism and provides a major impetus for further studies on TLD. There are similarities between TLD of bacteria and death of eukaryotic cells. Also, bacteria have "survival" genes other than thy (thymidylate synthetase), and this raises the question of whether there is a relationship between the two. A model is presented for a molecular basis of TLD.

The levels of both exonuclease HI (exo m, product of xthA) and hydroperoxidase H (HP-II, product of katE) activity in Escherichia coli were influenced by a functional katF gene. The katF gene product is also necessary for synthesis of HP-Il. Mutations in either katF or xthA, but not katE, result in sensitivity to H202 and near-UV (300-400 nm) radiation. Exo HI, encoded by the xthA locus, recognizes and removes nucleoside 5'-monophosphates near apurinic and apyrimidinic sites in damaged DNA. Extracts of katF mutant strains had little detectable exo HI activity. When a katF+ plasmid was introduced into the katF mutant, exo m activity exceeded wild-type levels. We propose that the katF gene is a trans-acting positive regulator of exo m and HP-II enzymes, both of which are involved in cellular recovery from oxidative damage.

We present evidence showing that rpoS (katF) is a regulator of katG gene transcription in an oxyR-independent manner. Mutation of the rpoS gene in several different Escherichia coli strains caused a significant reduction in catalase HPI activity. In rpoS-delta oxyR double mutants, the level of HPI was considerably lower compared to the delta oxyR parent strain, and was restored when transformed with an rpoS+ plasmid. Overproduction of HPI in oxyR- suppressor strains was greatly diminished after inactivation of the rpoS gene and was accompanied by a substantial increase in sensitivity to menadione. Beta-galactosidase expression from a katG::lacZ promoter was lower in rpoS strains compared to rpoS+ isogenic parents. Several delta oxyR strains had detectable levels of katG transcription that was significantly diminished after rpoS gene inactivation.