Thymineless Death and Ultraviolet Sensitivity in Micrococcus radiodurans

Thymine-requiring mutants of Micrococcus radiodurans have been isolated by selection on solid medium containing trimethoprim. Strains requiring either high concentrations of thymine (50 μg/ml) or low concentrations (2 μg/ml) for normal growth were obtained. The Thy⁻ mutant requiring low thymine concentrations has been characterized. It was shown to retain the high ultraviolet light (UV) resistance typical of wild-type M. radiodurans, but it was not resistant to thymineless death. Preliminary exposure of the cells to thymineless conditions resulted in enhanced UV sensitivity, and this interaction occurred under conditions where “unbalanced growth” was inhibited by the addition of chloramphenicol. Upon addition of thymine to deprived cells, UV resistance was gradually restored, and this recovery took place in the absence of protein synthesis. A model is proposed to account for the similarity of thymineless death in bacteria whose deoxyribonucleic acid repair efficiencies differ widely.     

R-Plasmid Transfer to and from Escherichia coli Strains Isolated from Human Fecal Samples

Strains of Escherichia coli recently isolated from human feces were examined for the frequency with which they accept an R factor (R1) from a derepressed fi⁺ strain of E. coli K-12 and transfer it to fecal and laboratory strains. Colicins produced by some of the isolates rapidly killed the other half of the mating pair; therefore, conjugation was conducted by a membrane filtration procedure whereby this effect was minimized. The majority of fecal E. coli isolates accepted the R factor at lower frequencies than K-12 F⁻, varying from 10⁻² per donor cell to undetectable levels. The frequencies with which certain fecal recipients received the R-plasmid were increased when its R⁺ transconjugant was either cured of the R1-plasmid and remated with the fi⁺ strain or backcrossed into the parental strain. The former suggests the loss of an incompatibility plasmid, and the latter suggests the modification of the R1-plasmid deoxyribonucleic acid (DNA). In general, the fecal R⁺E. coli transconjugants were less effective donors for K-12 F⁻ and heterologous fecal strains than was the fi⁺ K-12 strain, whereas the single strain of Citrobacter freundii examined was generally more competent. Passage of the R1-plasmid to strains of salmonellae reached mating frequencies of 10⁻¹ per donor cell when the recipient was a Salmonella typhi previously cured of its resident R-plasmid. However, two recently isolated strains of Salmonella accepted the R1-plasmid from E. coli K-12 R⁺ or the R⁺E. coli transconjugants at frequencies of 5 × 10⁻⁷ or less.     

Excision Repair Properties of Isogenic rec− Mutants of Escherichia coli K-12

We have examined the excision repair properties of isogenic rec⁻ and uvr⁻ strains of Escherichia coli K-12. A recB⁻recC⁻ strain excises dimers at a rate nearly that of the rec⁺ parent, reaching the same extent of excision after a 1-hr postirradiation incubation. recA⁻ and recA⁻recB⁻ strains excise 75 to 80% of the dimers excised by their rec⁺ parent, whereas a uvrB⁻ strain excises no dimers during a 1-hr incubation. The doses of ultraviolet light (254 nm) required to reduce survival to 37% of the original population are 8 ergs/mm² for recA or recA recB mutants, 5 ergs/mm² for the uvrB⁻ strain, 30 ergs/mm² for the recB recC mutant, and 230 ergs/mm² for the wild-type parent. From these data one cannot account for the ultraviolet light sensitivity of rec⁻ strains on the basis of their excision repair properties. We conclude that rec gene products play no significant role in the early steps of excision repair. The assay we have used for excision of thymine dimers is a modification of the Carrier-Setlow technique, and is described in detail in the Appendix to this paper. To show the properties and validity of this method, results of experiments with thymine dimers formed in vitro and in vivo in E. coli K-12 are presented. These results show our method to be reproducible and sensitive to 0.005% of the total radioactive thymine present in thymine-containing dimers.     

Thymineless Death in Escherichia coli 15T− and Recombinants of 15T− and Escherichia coli K-12

Thymineless death was examined in Escherichia coli 15T(-) and recombinants of 15T(-) and E. coli K-12. Those strains that were very sensitive to thymine deprivation were also very sensitive to a variety of inducing agents (mitomycin C, ultraviolet light, hydroxyurea, and nalidixic acid). Those strains that were relatively resistant to thymineless death were also relatively resistant to the inducing agents. After exposure to thymineless death and the inducing agents, sensitive strains lysed, produced colicin, and had phage particles in their lysates. These strains also showed an increase in the 6-methyladenine content of their deoxyribonucleic acid (DNA) and an increase in the DNA methylase activity of their crude extracts under these conditions. None of these effects was noted in the strains relatively resistant to thymineless death and the inducing agents. These data indicate that there are two types of thymineless death. One is represented by the strains that are very sensitive to thymine deprivation and other inducing agents and is secondary to the induction of phage psi. The strains more resistant to thymine deprivation and the other inducing agents undergo a non-phage-mediated thymineless death. The mechanism of this latter process is currently under study.     

Restoration induced by catalase in irradiated microorganisms

1. E. coli, strain K-12, and B. megatherium 899, irradiated in strict but still undefined physiological conditions with certain heavy doses of ultraviolet light, are efficiently restored by catalase, which acts on or fixes itself upon the bacteria in a few minutes. This restoration (C. R.), different from photorestoration, is aided by a little visible light. 2. At 37° the restorability lasts for about 2 hours after UV irradiation; the restored cells begin to divide at the same time as the normal survivors. 3. C. R. is not produced after x-irradiation. 4. B. megatherium Mox and E. coli, strain B/r show little C. R.; E. coli strain B shows none. None of these three strains is lysogenic, whereas the two preceding catalase-restorable strains are. 5. Phage production in the system "K-12 infected with T2 phage" is restored by catalase after UV irradiation, whereas phage production in the system "infected B" is not. 6. With K-12, catalase does not prevent the growth of phage and the lysis induced by UV irradiation (Lwoff's phenomenon). 7. Hypotheses are discussed concerning: (a) the chemical nature of this action of catalase; (b) a possible relation between C. R. and lysogenicity of the sensitive bacteria; (c) the consequences of such chemical restorations on the general problem of cell radiosensitivity.     

Genetic Determinant in Escherichia coli Affecting Thymineless Death and Ultraviolet Sensitivity

We have infected Escherichia coli strain B3 with F′ factors carrying different lengths of the K-12 chromosome. When the F′ factor carries the tsx to purE segment, the resulting hybrid merodiploid shows increased sensitivity to thymineless death, although leaky deoxyribonucleic acid synthesis is unaffected, as well as increased sensitivity to ultraviolet irradiation. The dominant or partially dominant character of the effect indicates that it is not a product of allelic differences between E. coli K and B at either ras or lon.     

Utilization of 5-Bromouracil by Thymineless Bacteria

Several thymineless Escherichia coli strains have been examined for their ability to replicate their deoxyribonucleic acid when bromouracil is substituted for thymine. The procedure we describe was used to identify a thymineless strain with characteristics relatively favorable to its use in bromouracil labeling experiments. In addition, mutants with an “absolute” thymine requirement could be easily distinguished from one with a “leaky” thymine requirement.     

Drug Resistance of Enteric Bacteria VIII. Chromosomal Location of Nontransferable R Factor in Escherichia coli

The R₂₁(TC) factor, obtained by transduction of the R₁₀(TC.CM.SM.SA) factor with phage ε to group E Salmonella, is not transferable by the normal conjugal process. However, when R₂₁(TC)⁺ transductants are infected with the F₁₃ factor, the nontransferable R₂₁(TC) factor acquires transmissibility by conjugation. R₂₁(TC)⁺ conjugants of Escherichia coli K-12, to which only the R₂₁(TC) factor was transmitted by cell-to-cell contact from an F′ R⁺ donor, were still unable to transfer their R₂₁(TC) factor by conjugation. In crosses between Hfr and F⁻E. coli K-12 strains containing R₂₁(TC), the gene responsible for tetracycline resistance was located on the E. coli K-12 chromosome between lac and pro, near lac.     

Mechanism of Caffeine Enhancement of Mutations Induced by Sublethal Ultraviolet Dosages

Certain chemical compounds increase mutation frequency of Escherichia coli B/r significantly when used in conjunction with nonlethal ultraviolet (UV) dosages. Studies were done to elucidate the mechanism of this enhancing mutational effect. Dark survival curves showed that 500 μg of caffeine per ml in the postirradiation medium markedly decreased survival to 60 ergs/mm² of UV in strain B/r. Caffeine did not markedly decrease survival to UV in strain B/r WP-2 hcr⁻. At least 90% of the mutations induced to streptomycin resistance by UV and 85% of those induced by UV with caffeine could be photoreversed. Experiments with thymine analogues suggested that thymine dimerization at the streptomycin locus was the primary premutational photoproduct induced by sublethal UV dosages. Caffeine did not interfere with the photoreversal of induced mutants, indicating that it probably does not bind to the photoreactivating enzyme or to a UV-induced lesion in the DNA. Addition of DNA or irradiated DNA with 500 μg of caffeine per ml resulted in no loss of the caffeine activity. The excision of UV-induced thymine-containing dimers from E. coli B/r T⁻ was investigated in the presence and absence of caffeine. Our results indicated that caffeine prevents excision of thymine dimers, presumably by binding to the excising enzyme. This binding results in an impairment of repair, which produces the increase in mutant numbers.     

Inactivation of bacteria by decay of incorporated radioactive phosphorus

Cultures of Escherichia coli will not grow in media containing very high specific activities of radiophosphorus P³², the inhibition of growth being due to the decay of assimilated P³² atoms. Experiments with a differentially labeled thymineless strain of E. coli show that the P³² disintegrations which occur in the bacterial deoxyribonucleic acid, i.e. in the nucleus, are mainly responsible for the inactivation of the cell. The kinetics with which radioactive bacterial populations are inactivated indicate that the function of several nuclei per bacterial cell must be eliminated by P³² decay before the ability to generate a colony is lost. The efficiency with which each P³² disintegration inactivates the nucleus in which it has occurred is calculated to be 0.02 (at –196°), i.e., similar in magnitude to the killing efficiency of P³² decay in bacteriophages. P³² decay and thymine starvation cooperate in bringing about the death of individuals of the thymineless strain, from which observation it is inferred that "thymineless death" is likewise a nuclear inactivation. The descendants of a non-radioactive bacterial culture grown for several generations in the presence of P³² and the descendants of a radioactive culture grown in the absence of P³² are inactivated by P³² decay in a manner which indicates that the phosphorus atoms of bacterial nuclei are dispersed among the progeny nuclei in their line of descendance.     

Lytic Replication of Coliphage Lambda in Salmonella typhosa Hybrids

Hybrids between Escherichia coli K-12 and Salmonella typhosa which conserved a continuous K-12 chromosomal diploid segment extending from pro through ara to the strA locus were sensitive to plaque formation by wild-type λ. These partially diploid S. typhosa hybrids could be lysogenized with λ and subsequently induced to produce infectious phage particles. When the K-12 genes were segregated from a lysogenic S. typhosa hybrid, phage-productive ability was no longer detectable due to loss of a genetic region necessary for vegetative replication of λ. However, λ prophage was shown to persist in a quiescent state in the S. typhosa hybrid segregant with phage-productive ability being reactivated after replacement of the essential K-12 λ replication region. Low-frequency transduction and high-frequency transduction lysates containing the gal⁺ genes of S. typhosa were prepared by induction of λ-lysogenic S. typhosa hybrids indicating that the attλ site is chromosomally located in S. typhosa in close proximity to the gal locus as in E. coli K-12. After propagation in S. typhosa hybrids, λ was subject to restriction by E. coli K-12 recipients, thus establishing that S. typhosa does not perform the K-12 modification of λ deoxyribonucleic acid. Hybrids of S. typhosa, however, did not restrict λ grown previously on E. coli K-12. The K-12 genetic region required for λ phage production in S. typhosa was located within min 66 to min 72 on the genetic map of the E. coli chromosome. Transfer of an F-merogenote encompassing the 66 to 72 min E. coli chromosomal region to λ-insensitive S. typhosa hybrids enabled them to replicate wild-type λ. The λ-insensitive S. typhosa hybrid, WR4255, which blocks λ replication, can be mutagenized to yield mutant strains sensitive to λvir and λimm434. These WR4255 mutants remained insensitive to plaque formation by wild-type λ.     

Polynucleotide Sequence Divergence Among Strains of Escherichia coli and Closely Related Organisms

Polynucleotide sequence similarity tests were carried out to determine the extent of divergence present in a number of Escherichia coli strains, obtained from diverse human, animal, and laboratory sources, and closely related strains of Shigella, Salmonella, and the Alkalescens-Dispar group. At 60 C, relative reassociation of deoxyribonucleic acid (DNA) from the various strains with E. coli K-12 DNA ranged from 100 to 36%, with the highest level of reassociation found for three strains derived from K-12, and the lowest levels for two “atypical” E. coli strains and S. typhimurium. The change in thermal elution midpoint, which indicates the stability of DNA duplexes, ranged from 0.1 to 14.5 C, with thermal stability closely following the reassociation data. Reassociation experiments performed at 75 C, at which temperature only the more closely related DNA species form stable duplexes, gave similar indications of relatedness. At both temperatures, Alkalescens-Dispar strains showed close relatedness to E. coli, supporting the idea that they should be included in the genus Escherichia. Reciprocal binding experiments with E. coli BB, 02A, and K-12 yielded different reassociation values, suggesting that the genomes of these strains are of different size. The BB genome was calculated to be 9% larger than that of K-12, and that of 02A 9% larger than that of BB. Calculation of genome size for a series of E. coli strains yielded values ranging from 2.29 × 10⁹ to 2.97 × 10⁹ daltons. E. coli strains and closely related organisms were compared by Adansonian analysis for their relatedness to a hypothetical median strain. E. coli 0128a was the most closely related to this median organism. In general, these data compared well with the data from reassociation experiments among E. coli strains. However, anomalous results were obtained in the cases of Shigella flexneri, S. typhimurium, and “atypical” E. coli strains.     

Survival of Shewanella oneidensis MR-1 after UV Radiation Exposure

We systematically investigated the physiological response as well as DNA damage repair and damage tolerance in Shewanella oneidensis MR-1 following UVC, UVB, UVA, and solar light exposure. MR-1 showed the highest UVC sensitivity among Shewanella strains examined, with D₃₇ and D₁₀ values of 5.6 and 16.5% of Escherichia coli K-12 values. Stationary cells did not show an increased UVA resistance compared to exponential-phase cells; instead, they were more sensitive at high UVA dose. UVA-irradiated MR-1 survived better on tryptic soy agar than Luria-Bertani plates regardless of the growth stage. A 20% survival rate of MR-1 was observed following doses of 3.3 J of UVC m⁻², 568 J of UVB m⁻², 25 kJ of UVA m⁻², and 558 J of solar UVB m⁻², respectively. Photoreactivation conferred an increased survival rate to MR-1 of as much as 177- to 365-fold, 11- to 23-fold, and 3- to 10-fold following UVC, UVB, and solar light irradiation, respectively. A significant UV mutability to rifampin resistance was detected in both UVC- and UVB-treated samples, with the mutation frequency in the range of 10⁻⁵ to 10⁻⁶. Unlike in E. coli, the expression levels of the nucleotide excision repair (NER) component genes uvrA, uvrB, and uvrD were not damage inducible in MR-1. Complementation of Pseudomonas aeruginosa UA11079 (uvrA deficient) with uvrA of MR-1 increased the UVC survival of this strain by more than 3 orders of magnitude. Loss of damage inducibility of the NER system appears to contribute to the high sensitivity of this bacterium to UVR as well as to other DNA-damaging agents.     

Temperate Bacteriophage Which Causes the Production of a New Major Outer Membrane Protein by Escherichia coli

Under most conditions of growth, the most abundant protein in the outer membrane of most strains of Escherichia coli is a protein designated as “protein 1” or “matrix protein”. In E. coli B, this protein has been shown to be a single polypeptide with a molecular mass of 36,500 and it may account for more than 50% of the total outer membrane protein. E. coli K-12 contains a very similar, although probably not identical, species of protein 1. Some pathogenic E. coli strains contain very little protein 1 and, in its place, make a protein designated as protein 2 which migrates faster on alkaline polyacrylamide gels containing sodium dodecyl sulfate and which gives a different spectrum of CNBr peptides. An E. coli K-12 strain which had been mated with a pathogenic strain was found to produce protein 2, and a temperate bacteriophage was isolated from this K-12 strain after induction with UV light. This phage, designated as PA-2, is similar in morphology and several other properties to phage lambda. When strains of E. coli K-12 are lysogenized by phage PA-2, they produce protein 2 and very little protein 1. Adsorption to lysogenic strains grown under conditions where they produce little protein 1 and primarily protein 2 is greatly reduced as compared to non-lysogenic strains which produce only protein 1. However, when cultures are grown under conditions of catabolite repression, protein 2 is reduced and protein 1 is increased, and lysogenic and non-lysogenic cultures grown under these conditions exhibit the same rate of adsorption. Phage PA-2 does not adsorb to E. coli B, which appears to have a slightly different protein 1 from K-12. These results suggest that protein 1 is the receptor for PA-2, and that protein 2 is made to reduce the superinfection of lysogens.     

Correlation Between the Expression of an Escherichia coli Cell Surface Protein and the Ability of the Protein to Bind to Lipopolysaccharide

The ompA gene of Escherichia coli codes for a major protein of the outer membrane. When this gene was moved between various unrelated strains (E. coli K-12 and two clinical isolates of E. coli) by transduction, the gene was expressed very poorly. Recombinants carrying “foreign” genes produced no OmpA protein which could be detected on polyacrylamide gels and became resistant to bacteriophage K3, which uses this protein as receptor. The recombinants were sensitive to host-range mutants of K3, indicating a very low level of OmpA protein was produced. When an E. coli K-12 recombinant carrying an unexpressed foreign ompA allele was subjected to two cycles of selection for an OmpA⁺ phenotype, a mutant strain was obtained which was sensitive to K3 and which expressed nearly normal levels of OmpA protein in the outer membrane. This strain carried mutations in the foreign ompA gene, as indicated both by genetic mapping and the alteration of a peptide in the mutant OmpA protein. The ability of the OmpA protein to bind to lipopolysaccharide (LPS) showed similar strain specificity, and the mutant OmpA protein which was expressed in an unrelated host showed enhanced ability to bind LPS from its new host. Thus, cell surface expression of the ompA gene appears to depend upon the ability of the gene product to bind LPS, suggesting that an interaction between the protein and LPS plays an essential role in biosynthesis of this outer membrane protein.     

Anaerobic Transfer of Antibiotic Resistance from Pseudomonas aeruginosa

Pseudomonas aeruginosa, an opportunistic pathogen that often initiates infections from a reservoir in the intestinal tract, may donate or acquire antibiotic resistance in an anaerobic environment. Only by including nitrate and nitrite in media could antibiotic-resistant and -sensitive strains of P. aeruginosa be cultured in a glove box isolator. These anaerobically grown cells remained sensitive to lytic phage isolated from sewage. After incubation with a phage lysate derived from P. aeruginosa 1822, anaerobic transfer of antibiotic resistance to recipients P. aeruginosa PS8EtBr and PS8EtBrR occurred at frequencies of 6.2 × 10⁻⁹ and 5.0 × 10⁻⁸ cells per plaque-forming unit, respectively. In experiments performed outside the isolator, transfer frequencies to PS8EtBr and PS8EtBrR were higher, 1.3 × 10⁻⁷ and 6.5 × 10⁻⁸ cells per plaque-forming unit, respectively. When P. aeruginosa 1822 was incubated aerobically with Escherichia coli B in medium containing nitrate and nitrite, the maximum concentration of carbenicillin-resistant E. coli B reached 25% of the total E. coli B population. This percentage declined to 0.01% of the total E. coli B population when anaerobically grown P. aeruginosa 1822 and E. coli B were combined and incubated in the glove box isolator. The highest concentration of the recipient population converted to antibiotic resistance occurred after 24 h of aerobic incubation, when an initially high donor/recipient ratio (>15) of cells was mixed. These data indicate that transfer of antibiotic resistance either by transduction between Pseudomonas spp. or by conjugation between Pseudomonas sp. and E. coli occurs under strict anaerobic conditions, although at lower frequencies than under aerobic conditions.     

Involvement of Recombination Genes in Growth and Viability of Escherichia coli K-12

We have studied the growth properties of 17 isogenic strains of Escherichia coli K-12 differing only in the recA, recB, recC, and sbcA alleles. We have observed the following. (i) All recombination deficient strains have decreased growth rates and decreased viabilities compared with recombination proficient strains. The large populations of nonviable cells in Rec⁻ cultures may arise by spontaneous lethal sectoring (9). (ii) A recA mutant strain which is entirely recombination deficient and which shows high ultraviolet sensitivity and “reckless” deoxyribonucleic acid (DNA) breakdown has approximately the same growth rate and twice the viability as recB and recC mutant strains which have residual recombination proficiency, moderate ultraviolet sensitivity, and “cautious” DNA breakdown. (iii) Indirectly suppressed (sbcA⁻) recombination proficient (Rec⁺) revertants of recB and recC mutant strains have approximately normal growth rates and are three times as viable as their Rec⁻ ancestors (but not as viable as rec⁺ cells). We suggest the following hypothesis to account for the low viability of Rec⁻E. coli. Single-strand breaks in the DNA duplex, necessary for normal bacterial growth, may be repaired in a Rec⁺ cell. Failure of Rec⁻ cells to repair this normal DNA damage may lead to the observed loss of viability.     

H-NS Regulates DNA Repair in Shigella

We report a new role for H-NS in Shigella spp.: suppression of repair of DNA damage after UV irradiation. H-NS-mediated suppression of virulence gene expression is thermoregulated in Shigella, being functional at 30°C and nonfunctional at 37 to 40°C. We find that H-NS-mediated suppression of DNA repair after UV irradiation is also thermoregulated. Thus, Shigella flexneri M90T, incubated at 37 or 40°C postirradiation, shows up to 30-fold higher survival than when incubated at 30°C postirradiation. The hns mutants BS189 and BS208, both of which lack functional H-NS, show a high rate of survival (no repression) whether incubated at 30 or 40°C postirradiation. Suppression of DNA repair by H-NS is not mediated through genes on the invasion plasmid of S. flexneri M90T, since BS176, cured of plasmid, behaves identically to the parental M90T. Thus, in Shigella the nonfunctionality of H-NS permits enhanced DNA repair at temperatures encountered in the human host. However, pathogenic Escherichia coli strains (enteroinvasive and enterohemorrhagic E. coli) show low survival whether incubated at 30 or 40°C postirradiation. E. coli K-12 shows markedly different behavior; high survival postirradiation at both 30 and 40°C. These K-12 strains were originally selected from E. coli organisms subjected to both UV and X irradiation. Therefore, our data suggest that repair processes, extensively described for laboratory strains of E. coli, require experimental verification in pathogenic strains which were not adapted to irradiation.     

Isolation of Lightning-Competent Soil Bacteria

Artificial transformation is typically performed in the laboratory by using either a chemical (CaCl₂) or an electrical (electroporation) method. However, laboratory-scale lightning has been shown recently to electrotransform Escherichia coli strain DH10B in soil. In this paper, we report on the isolation of two “lightning-competent” soil bacteria after direct electroporation of the Nycodenz bacterial ring extracted from prairie soil in the presence of the pBHCRec plasmid (Tc^r, Sp^r, Sm^r). The electrotransformability of the isolated bacteria was measured both in vitro (by electroporation cuvette) and in situ (by lightning in soil microcosm) and then compared to those of E. coli DH10B and Pseudomonas fluorescens C7R12. The electrotransformation frequencies measured reached 10⁻³ to 10⁻⁴ by electroporation and 10⁻⁴ to 10⁻⁵ by simulated lightning, while no transformation was observed in the absence of electrical current. Two of the isolated lightning-competent soil bacteria were identified as Pseudomonas sp. strains.     

Effect of Thymine Starvation on Deoxyribonucleic Acid Repair Systems of Escherichia coli K-12

Thymine starvation of Escherichia coli K-12 results in greatly increased sensitivity to ultraviolet light (UV). Our studies, using isogenic strains carrying rec and uvr mutations, have shown the following. (i) Common to all strains tested is a change from multihit to single-hit kinetics of survival to UV after 60 min of thymine starvation. However, the limiting slope of UV survival curves decreases in the rec(+)uvr(+) strain and changes very little in several rec mutant strains and one uvrB mutant strain. Thus, when either the rec or uvr system is functioning alone, the limiting slopes of the UV survival curves are relatively unaffected by thymine starvation. (ii) Thymine starvation does not significantly inhibit repair processes carried out by either repair system alone; i.e., host cell reactivation of irradiated phage (carried out by the uvr system), excision of thymine dimers (uvr), or X-ray repair (rec). (iii) In a rec(+)uvr(+) strain, repair appears to be a synergistic rather than additive function of the two systems. However, after thymine starvation, repair capacity is reduced to about the sum of the repair capacities of the independent systems. (iv) The kinetics of thymineless death are not changed by rec and uvr mutations. This indicates that the lesions responsible for thymineless death are not repaired by rec or uvr systems. (v) Withholding thymine from thy rec(+)uvr(+) bacteria not undergoing thymineless death has no effect on UV sensitivity. Under these conditions one sees higher than normal UV resistance in the presence or absence of thymine. This is due to increased repair carried out by the uvr system. To explain these results we postulate that thymine starvation does not inhibit either the rec or uvr repair pathway directly. Rather it appears that thymine starvation results in increased UV sensitivity in part by inhibiting a function which normally carries out efficient coordination of rec and uvr pathways.