Thyniiiieless Death in Eschericchia Coli: Strain Specificity

JOURNAL oF BACT(ERIOL0OY, Jun1e 1967, P. 1917-1924 Vol. 93, No. 6 Copyright (I 1967 Aicitne;czIIi Society for Microbiology Prititc i/l U.S.A.. Thyniiiieless Death in Eschericchia coli: Strain Specificity

DONALD J. CUMMINGS AND LEE MONDALE Departmzenlt of Microbiology, University cf Colorcukl Medical Ceniter, Delivetr, Coloradlo 80220 Received for publication 8 February 1967

Thymineless death of various ultraviolet (UV)-sensitive strains of Esch1erichia co/i B and K-I 2 was investigated. It was found that E. coli B, Bs- ,, K- 12 /e -21, and possibly K-12 Lon-, all sensitive to UV, were also sensitive to thymine starvation. How- ever, other UV-sensitive strains of E. coli were found to display the typical resistant- type kinetics of thymineless death. The correlation of these results with various other cellular processes suggested that the filament-forming ability of the bacteria might be involved in the mechanism of thymineless death. It was aipparent from the present results that capacity for host-cell reactivation, recombination ability, thymine dimer excision, and probably induction of a defective prophage had little to do with determining sensitivity to thymine deprivation.

Recently, Cummings and Taylor (5) reported deprivation were known: sensitive and resistant. that Esciheric/liil coli B, in addition to being Aside from the influence of this finding in under- sensitive to ultraviolet (UV) irradiation and to standing the phenomenon of thymineless death, mitomycin C (MC) treatment, also displayed an it is necessary to determine whether E. coli B is unusual sensitivity to thymine deprivation. Upon unique in its response to thymine deprivation. being deprived of thymine, E. coli B immediately Hill and Feiner (14) presented in detail the began losing its ability to form colonies and was properties of some UV-sensitive mutants of E. termed sensitive to thymine starvation. In con- coli B, and Howard-Flanders (1, 15, 16; P. trast, thymine-requiring derivatives of E. coli Howard-Flanders and A. J. Clark, per-sonal strains such as B/r, 15 (4, 13), and K-12, were comnmunication) has described many different resistant, in that all exhibited a lag of about 50 UV-sensitive mutants of E. coli K-12, including a min prior to the onset of thymineless death. Some recombination-deficient mutant basically different other UV-sensitive strains of E. coli were also from those mutants previously observed (2, 3). examined, and while these strains were sensitive A number of these bacterial strains have been to UV and to MC, their response to thymine characterized with regard to their response to deprivation was essentially identical to that of thymine starvation, and this communication will the resistant strains. Similarly, Mennigmann present evidence that E. coli B is one of many (24) found that E. coli B, but not some other UV- strains sensitive to thymine starvation. sensitive strains, was much more sensitive to 5- fluorodeoxyuridine (FUDR) than was E. coli MATIAZIALS ANI) METHODS B/r. From these results, it was apparent that Bactericml strains. Thymine-requiring strains of E. sensitivity to one or more lethal agents did not coli were obtained by Ltse of the aminiopterin selec- necessarily correlate with sensitivity to thymine tion technlique (27, 36). In addition to the recombi- deprivation. Of the sensitive strains investigated nation deficient (.Rcc) strain studied previously by Cummings and Taylor (5), one was deficient (JC1569 rec-] and its parent JC1557, RecFb; 3, 5), a in its ability to undergo recombination (3), and second Rec- strain, AB2470 rec-21, was obtained another was unable to excise thymine dimers from P. Howard-Flaniders. This Rec- strain difTers (15). It was concluded, therefore (5), that a single physiologically from JC1569 in that it does not de- defect in acid grade its DNA duriing growth (2; P. Howard-Flanders deoxyribonucleic (DNA) repair (1, and A. J. Clark, perso)lcal communltiolicatioin ). P. Howard- 31, 32; B. H. Rosenberg and D. E. Packer, Abstr. Flanders also supplied tive other UV-sensitive strains Bioplhys. Soc., p. 39, 1967) could not satisfac- of E. coli K-12: ABI899 and 2426, each Lon-- (16); torily account for both sensitivity to thymine- AB1884, ivrC; AB1885, uvr'EB; and AB1886, uvrrA. less death and sensitivity to either UV or MC. Each of the last tlhrec UV-sensitive strains is Linable An important result of these earlier studies was to repair UV-irradiatcd TI bacteriophage and fails that two different types of response to thymine to excise thyminii-e ciiers 15). The wild-type Reck 1917 1918 CUMMINGS AND MONDALE J. BACTLIZIOL.

strain (ABI 157) 1ron3i which these strainis were derived was also obtainecl from 1'. Howard-Flanders. The UV-sensitive muLtants of E. coli B (14), B,-, BS 2, B.,3, BS4, B,,, B,- 1, and Bs,-12, as well as the parent strain E. coli B, were obtainied from R. F. Hill acnd R. B. Sehlow. These bacterial strainis were growni aerobically in La glucose-phosplhate salts minimal mediumn at 37 C with required sUpplements. Chaniges of mediulll were achieved by 800-fold dilu- tion into glucose-phosplhate salts containinlg no thyminle. For unliforiii results. thymiinieless death characteristics of all the bacterial strainis examinied were obtainied in thc lpresenlce of I c% Casaminio Acids (Difco). In some cases, the lag in thymnineless death (especially for the Low- strains) was ofteni proloniged in the absence of Casaminio Acids. This was probably because growth conditions were better in the presence z2 of Casarnino Acids or because intracellular pools LU were more efficiently depleted, as was suggested by ,8 Freifelder (10), who uLsed nucleoribosides, in particular uridine, to decrecLse the lag time. In our hands, 3 the addition of 1uclCeoribosides did inot facilitate thyrmineless deatlh as well as did Casaminlo Acids. Samples were taken at initervals and assayed for viable cells on tryptone (DLifco)-agar plates supplemented B, 12 witlh 30 ,g of thymline per ml. UV irradiationi was carried out as previoLusly de- 2c 30 40 0 6 C 70 80 90 100 110 120 130 scribed (5). M I N UT E S Fio. ]. Tl/nviniel/css cl/atli of Escheriic/iia coli B,. RESULTS Ilie opell circles rcfcr to clati obtainiedi wit/h t/ie pareielt and B foim wh/ic/i c/il t/ie B, stralins (closed circles) were Stralin specificity. Hill Feiner compared derived. As ccili he noted, onily E. coli B anld B,_12 the properties of some UV-sensitive mutants of responi/ec/ to t/hynninc deprivration in a senisitive nialdlier. E. oli B and found that of 12 such mutants Straiins B-1, BS.-, B;, B,-,, B-11 responded hi essen- there were 9 ditrerent phenotypes (14). Some tiall/ t/Ch scine resistant nianner11c- ais did B/i (S). StCali,i were more UV-sensitive than the parent E. co/i B_; was coisis5tenit/ h/ least senisitive strain wit/i B, and most had lost their predilection toward respect bothi to thei ilnitial la,ig ailid to thie iiltiniatte late filament formation alfter UV irradiationi. Figure 1 f' t/lvninlenlss d(eit/i. illustrates the thymineless death chlaraLcteristics of seven of these UV-sensitive mutaLnts of E. coli UV irradiation and all were more UV-sensitive B. As can be seen, E. coli B and Bs-12 have than the parent strain B (14). Of these mutants, essentiallythesamne sensitivity to thymine deprivaL- only B,-3 and B13,12 retcained the propensity' of tion in that the loss of colony-for-ming ability strain B to form filaments after UV irriadiation, commiienices aLlmost immediately. The results ob- only Bs-11 had the same sensitivity as its parent B tained here on the Hill strain B were the same as to crystal violet, and strains B1-1, Bs-3 Bs-, and those obtained previously (5) on an E. coli B BS,-1 no longer had the ability to repair UV- obtained from S. Luria. Except for E. coli Bs-4 treated Ti bacteriophage (14). To minimize the all of the UV-sensitive mutants of strain B re- possibility that particular thymine mutants were sponided to thyminie deprivation only aLfter aL lag examiiiined here, several single-colony isolates from of 38 to 55 mmi. This viariability in the lag time sepairate aminopter-in selections were studied; all may reflect a graduated resistance to thymine had the samie UV sensitivity as the or-iginal starvation or may result fromii differenices in the thymnine-nonrequiring bacteriunm, and all dis- rates of depletion of thymine from intrcacellular platyed the same response to thyminie deprivation. pools. The possibility that these strains were It should be noted that, although rall the results "'leaky' in their requir-ement for thymine was reported here were obtLined witlh the so-called excluded on the basis that only stringent colonies high-thymiine (20 ,g/'ml) requiring strains (27), were isolated acnd that 'leaky" mutaLnts would the saime results were Calso observed with low- imiost likely lead to Can increase in the ultimate thymiiine (2,ug/ml) requi-irig strains of E. coli B, surviving fractioni r(ather than aL pr-oloniged lag Bs, aLnd Bs-11. time. It slhould be emplhasised that aLll these B, In a similaL- mnianner, various UV-sensitive strains were directly isolated from E. coli after strains of E. coli K-12 were compcatred with wild- VOL. 93,1')67 THYMINELESS DEATH IN E. COLI 1919 type K-12. It was found (Fig. 2) that uvrA, B, their glossy, mucoid-like appearance on minimal or C mutants had the same thymineless death medium-agar plates; that is, these strains tend characteristics as did the UV-resistant wild- to form filaments during growth in minimal type parent strain. All these strains lost their medium (16). The function of the Casamino colony-forming ability as a result of thymine Acids in reducing the lag time may be to break deprivation only after a lag of about 55 min. up these filaments. That other events may be This behavior was essentially the same as that occurring during the 20-min lag mnay be indicated obtained previously with E. coli K-12 r-ec-] by the fact that, once death commences in Lon- (5). In contrast, E. coli K-12 rec-21 proved to strains, it does so at a rate faster than that be sensitive to thymine deprivation; this was the noted in any of the other strains examined. It is only UV-sensitive K-12 strain examined which difficult to state the ultimate rate of death with responded to thymine starvation in a manner much assurance, however, since this factor de- similar to E. coli B and BS,12. Two Lon- strains pends strongly on the generation time of the were studied [each isolated by UV irradiation organism. Nevertheless, in the cases observed, from AB1157 (16)], and both showed inter- the generation times of the E. coli B Lon- and mediate sensitivities to thymine starvation by rec-21 strains were about 40, 55, and 70 mm, responding after a lag of about 20 min. In the respectively, and the ultimate exponential death absence of Casamino Acids, this lag was ex- half-times observed were 6 to 7 min, 4 to 5 min, tended to about 32 min. It may well be that these and 11 to 12 min for E. coli B or B_1 2, Lon-, Lon- strains are actually more sensitive to and rec-21 strains, respectively. thymine deprivation than it would appear. A Filamiiient Jorniation. The filament-forming distinguishing characteristic of Lon- strains is ability of the bacterial strains must be evaluated in considering a possible mechanism to account for the phenomenon of thymineless death. E. coli 100 I B, Bs-3 BS-12, and K-12 Lon- are all known to form extensive filaments after UV irradiation (14, 16). Of these, E. coli B and BS-12, and possibly K-12 Lon-, are sensitive to thymineless death. The ability to form filaments can be determined either by direct observation in a light microscope (14) or by the rescue effect of pantoyl lactone a precursor 10 (37), of pantothenic acid. In 100 Fig. 3, the effect of pantoyl lactone on E. coli B can be observed. It can be noted that pantoyl lactone had a marked restorative effect on E. coli B exposed either to thymine starvation or to UV irradiation. Presumably, pantoyl lactone inhibits the formation of filaments (37) and allows the bacteria to develop into visible colonies. For purposes of monitoring the various bacterial strains for their response to pantoyl lactone, only the effect on UV survival was measured. In this regard, E. coli K-12 Lon- is already known to be rescued by pantoyl lactone (18), and the results on the other organisms of interest are presented MI NU T ES in Fig. 4. As can be seen, E. co/i BS-?, and B5-12 are readily rescued by pantoyl lactone. On the other FIG. 2. Thlymineless death antd UV senisitivity of Escherichia coli K-12. Closed symbols refe- to UV hand, pantoyl lactone has little or no reactivating (abscissa 0-4 m1i; rig/it side or diniate) alnd openi action on E. coli K-I 2 r-ec-21. As shown by symbols refer to tlhymineless leath kinetics (abscissa Kneser (18), another type of Rec- similar to 0-130 ni)n; /evft sidle ordlinate); ( 7, V) uv'rA, B, or C; rec-1 was also not reactivated by pantoyl lactone. (D, U) wild-type; (0, 0) rec-21; anld (A, A) Lot-. In our hands, pantoyl lactone actually inhibited The plot C an represeniting uvrA, B, or- is average, growth of the r-ec-21, but not of r-ec-l, uvirB, or ancd 110 significant differenuces wzere observed between these UV-senisitive strains. Of all the K-12 strains, any of the other strains examined. After 72 hr of only rec-21 and possibly the Loli- straiins (the avercige incubation in the presence of 0.6% pantoyl of two strains) w1 ere sen/sitive to thvYnine starvation. lactone, rec-21 failed to produce visible colonies 1920 CUMMINGS AND MONDALE J. BACTERIOL.

LU z LU

M N UTES FIG. 3. Ejfect of pantoyl Icictonie oni the survival oJ MINUTES MINUTES SECONDS Eseheric/ija coli B !LrLia). Opeti sYmbols, plaitedt FIG. 4. EFfiect Ol pacmitovI lalctomie oni the UV survival withouit paitoyl lactonl; closed symnbols, plated wit/i of (,A, A) Esc/ieric/iia coli K-i2 rec-21I (0, *) E. 0.3%" panitoyl lactolie. (0, 0; 0-100 hull t/)sciSsa, coli B, ,; anid ( *) E. coli B_ 12 Open synibols, left side ordinaite) refers to tliymwiineless t/eat/i kinetics platedi without pamtoYl laictonie; closed symbols, alid (A, A; 0-1 mmi abscissa, rig/it sidle ordlinacte) platedi wit/i 0.3':, pnltoyl lactone. E. coli B., amlid B,-12 re- refers to UV inactivatiomi. In e(ich case, Iac- ptilitoN'l sponded ini essenltiallv the same imialmlier, to panitovl tonie restored thIl colomiv-fmrning ability to hoitI/tie lactomie but in Il/i case qi' the r-ec-21 straimi, little, sam/71e extenit. if anl, restorative eL'cfct was observed. Inicreasinig th/i pamitovl lactone comicemltratiomi elihamice(d tIle reactiva-

For this reason, 0.3, lactone was in omi pantoyl used tioii effiect E. coli B, B.,-, anid B, 2, blit little or these experiments. Of the strains examined, E. coli tio effect was observed oli E. coli rec-i, uvrB, or B--. B, Bs-37 and Bs-,.) were reactivated by pantoyl lactone, and no palrticular effect was noted in single cells. It would appear, therefore, that, al- strains E. coli B,,, K-12 rec-i, K-12 rec-21, though E. coli B, B-12, and Lon- all form ex- and K-12 uvrB. As shown by Hill and Feiner tensive filaments and are all sensitive to thymine (14), E. coli B, Bs-3 and Bs-I2 had essentially the starvation, there are strains which are sensitive same filament-forming ability. This ability to to thymine starvation (rec-21) but do not form form extensive filaments after UV irradiation is extensive filaments and there are strains which readily observable in the light microscope (14). do form extensive filanments (Bs 3) but are not Figure 5 shows representative photographs of sensitive to thymine starvation. the three strains found to be sensitive to thymine- Induction. Since thymine starvation is known less death. As was indicated by the results with to induce proplhage (17, 22) and colicin forma- pantoyl lactone, E. coli B and Bs-12 displayed tion (9, 23), it has been suggested (24, 35) that many long filaments 3 hr after sufficient UV ir- thymineless death could be caused by the induc- radiation, whereas E. coli rec-21 showed a few tion of a defective prophage. In support of this short filaments and mostly single cells. With the hypothesis, Frampton and Brinkley (8) studied same technique, E. coli B,_3 gave essentially the four strains of thymine-requiring E. coli 15, same results as B and BS-12, whereas E. coli strains known to produce colicins, and found B/r and E. coli K-12 AB1157, the parent of that, after exposure to UV, extracts from lysed rec-21, had occasional short filaments and mostly cells contained some complete, but mostly V()L. 9)3, 1967 THYMINELESS DEATH IN E. COLI 1921 incomplete, phalge particles of morphology different from any phage previously seen as- sociated with E. coli. None of these particles was found to be infective when E. coli B, C, or 15 was used as indicator strain. We therefore examined concentrated extracts from 1010 to 2 X 1010 cells of E. coli B and B/r, both wild-type and thymine-requi-ing strains, using either MC (28') or tlhymine stairvation to induce a hypo- thetical defective prophage. We were not able to find any convincing evidence with electron micros- copy for the presence of defective phages or phage parts in the lysates. Some substructures somewhalt similar to T-even phage parts were A observed rarely in only a few preparations, but no recognizable phage particles were seen. It was readily shown that induction of a normal phage hald little or no effect on the kinetics of thymineless death (Fig. 6). E. coli K-12 rec-], *ec-21, tnrB, and two wild-type thymine-requiring strains lysogenized with X bacteriophage all responded to thymine deprivation in the same manner as did the nonlysogenic parents. In addition, no significant differences were noted in the kinetics of induction of X bacteriophage by

100 . 3.o...-O-Ac.4

B~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

z U

FIci. 5. Light micrographis of the three strains founid to he selisitie to thi,iniiie starlvatioii: (A) Escheric/hia 10 20 30 40 50 60 70 80 90 100 110 120 130 coli B; (B) E. co/i B,12; (C) E. cO/i K-12 rec-21. Fol- MINUTES iowinig tie procedure Of Hi/I an7d Feiner (14), the bac- Fic. 6. Tihninieless cleathi characteristics of teria were UV-iuactivated to 1% Slrivival, an7d the,, Escheriichia coli K-12 (X). (0) AB1157 Rec+; (A) allowed to grow flr 3 hr oni tr- ptone-thvminine-agar JC1557 Rec+; (7) JC1569 rec-J; (E]) iuvrB; anid plctes. Pictures wvere takeln with a LeitZ phase-coIntrast (0) AB2470 rec-21. Note that little or iio chanige inicroscope oni Polaroid fil/n at a 17taglifficationi of wvere observed compared wit/i the kilietics conItained ahout 700:1. iii Figurie 2. 1922 CUMMINGS AND MONDALE J BAC-n--RlIOL

MC, thymine deprivation, or lheat (19) in E. coli (34) using E. coli B,-1, B;, and B r, all of which K-12 AB1157, rec-21, or iirrB. These studies we now know to have essentially the same re- on lysogenic strains also revealed another physi- sistance to thymine starvation. It mnay also be ological difTerence between r-ec-] and r-ec-21 that the mechanism(s) of repairing lesions in that induction of X in rec-21 was reduced only caused by thymine starvation does not involve 5- to 10-fold, whereas in iec-I induction was re- directly the repair of DNA but rather that the duced at least 1,000-fold (P. Howard-Flanders DNA lesion observed (31) after thymine dep- and A. J. Clark, personial coommuniccition). rivation may be a secondary effect occurring It has also been suggested that ribonucleic after a lethal event elsewhere. Recently, in fact, acid (RNA) synthesis (6, 12, 13), and in par- Rosenberg and Packer (Abstr. Biophys. Soc., ticular messenger RNA synthesis (31), is required p. 39, 1967) reported that abnormal methylation in order that the bacteria suffer thymineless occurs in the absence of thymine. The existence damage. It is difficult to test this hypothesis of two distinct modes of thymineless death may (31) directly, but it is possible to induce protein enable us to determine genetically a site (or sites) synthesis and observe the effect on thymineless controlling sensitivity to thymine deprivation, death. In the case of E. coli B and B/r, no changes and thus specify its relationship to other metabolic were observed in the kinetics of thymineless events. death which occurred during the induction of A major difficulty in obtaining evidence con- either 3-galactosidase (29) or D-serine deaminase cerning the mechanism of thymineless death is (30), or both proteins simultaneously, even that thymine deprivation may interfere with though the specific activity of each enzyme in- many cellular processes. Along with UV and duced increased 4- to 14-fold. In agreement MC, thymine starvation is known to induce with Luzzati (20), who showed a reduction in prophages (17, 22), and hence prophage induc- messenger RNA synthesis during thymineless tion must be considered as a possible cause of death, measurable enzyme induction ceased thymineless death. In a variety of UV-resistant when the bacteria had undergone thymineless or UV-sensitive strains, we were not able to death to the extent of 5 to 10() survival. demonstrate any effect of a normal prophage, such as X, on the kinetics of thymineless death, DIscussSoN but this did not exclude the presence of a de- From a variety of E. coli strains, three and fective prophage (24, 35). There is no doubt that possibly four were found to be sensitive to thy- some thymine-requiring bacterial have been mine deprivation: E. coli B, Bs l2, K-12 rec-21, shown (8) to yield unique defective phage parts. and possibly K-12 Lon-. Although each of these Although we were not able to demonstr'ate that strains was sensitive to UV or MC, the range and such phage parts were induced in E. coli B or type of sensitivity varied greatly. E. coli B, Bs12, B/r, one could argue that here the prophage and K-12 Lon- all form filaments extensively, was so defective that even substructural pieces and hence are rescued by pantoyl lactone; E. were not formed. However, it is not apparent coli B-l2 is unable to repair UV-treated Ti how a defective prophage could be the primary bacteriophage, and K-12 l ec-21 is unable to cause for two different types of thymineless death undergo genetic recombination. On the basis of in organisms derived from a common parent. these results and the results obtained with other In addition, there is little evidence that E. co/i bacteria, host-cell reactivation, recombination B or B/r produce colicins and, in fact, it has been ability, and thymine dimer excision do not shown that E. coli B/r difTers from E. coli 15 appear to be particularly relevant in determining (7, 8) in its ability to synthesize RNA after ex- sensitivity to thymine deprivation. The most posure to UV or X rays. The strains of E. coli common feature seems to be filament-forming 15 in which defective phage parts were induced ability. However, one strain sensitive to thymine showed an increased synthesis of RNA, whereas starvation (rec-21) did not form filaments and E. coli B/r showed an inhibited synthesis of was not rescued by pantoyl lactone, whereas RNA. Similarly, althouglh Mennigmann and another strain (B- 3) was not sensitive to thymine Szybalski (25) detected single straCnd breaks in starvation but did form filaments and was rescued DNA isolated from B. sishtilis, a strain known by pantoyl lactone. It seems likely that there to harbor a prophage, while it Was undergoing are many mechanisms involved in repairing thymineless death, several laboratories (11, 1 3, lesions created in DNA, by whatever means, and 21, 26) including our own were unable to detect that each type is genetically determined and any such physical chalnges in DNA isolated fi-om independent of the other types of repair. In this E. coli B, B/r, or 15. Recently, D. W. Smitlh Cand regard, three levels of repair of thymine dimers P. C. Hanatwalt (Abstr. Biophys. Soc., p. 78, in vivo have been described by Setlow et al. 1967) made the observation that tlivmineless VOL.9'3, 1967 THYMINELESS DEATH IN E. COLI 1923 death occurs in pleuropneumonia-like organisms, 14. HILL, R. F., AND R. R. FEINER. 1963. Furtlher and suggested that it is quite unlikely that this studies of ultraviolet-sensitive mutants of smallest living organism could harbor any type Escherichia coli strain B. J. Gen. Microbiol. of prophage. Hence, an important feature of 35: 105-114. 15. R. P. AND thymineless death is its ubiquitous nature, and HOWARD-FLANDERS, 1),, BOYCE, L. THERIOT. 1966. any proposed mechanism for Tlhree loci in Eschericliiac coli thymineless deaths K-12 that control the excision of should be applicable to a pyrimidine great many organisms dimers and certain other mutagen products which differ in many respects. from DNA. Genetics 53:1119-1136. 16. HOWARD-FLANDERS, P., E. SIMSON, AND L. ACKNOWLEDGMENTS THERIOT. 1964. A locus that conitrols filament The senior author is indebted to C. Dullllanh anid formation and sensitivity to radiation in A. L. Taylor for their assistance in the isolation and Escheerichica coli K-12. Genetics 49:237-246. chiaracterization of the E. coli K- 1 2 r-ec-21 Thiy- strain 17. KORN, D., AND A. WEISSBACH. 1962. Thymine- and for the lysogenization of the various E. coli K-12 less induction in Escherichia coli K-12 (X). strains.>' Biochim. Biophys. Acta 61:775-790. 18. H. 1966. This inivestigation was supported by Public Health KNESER, Absence of K-reactivation in Service researclh grant AI-06472 from the National a recombination-deficient mutant of E. coli. Institute of Allergy anid Infectious Diseases. Biochem. Biophys. Res. Commun. 22:383-387. 19. LIEB, M. 1966. Studies of heat-inducible lambda LITERATURE CITED bacteriophage I. Order of genetic sites and properties of mutant prophages. J. Mol. Biol. I. BOYVCLE R. P., AND P. HOWARD-Flanders. 1964. 16:149--163. Release of ultraviolet light-induced thymine 20. LUZZATI, D. 1966. Effect of thymilne starvation dimers from DNA in E. coli K-12. Proc. NatI. on messenger RNA synthesis in Eschlerichica Acad. Sci. U.S. 51:293-300. coli. J. Bacteriol. 92:1435-1446. 2. CLAIRK, A. J., M. CHANIBERLIN, R. P. BOYCE, AND 21. LUZZATI, D., AND C. REVEL. 1962. Effet l&tlhal 1P. HOWARD-FLANDERS. 1966. Abnormal imieta- de la carenice en thymine: etat de l'acide bolic response to ultraviolet of a light recombi- d6oxyribornuc1eique au cours de cette carence. niation deficient mutant of Escherichia coli Biochim. Bioplhys. Acta 61:305-306. K-12. J. Mol. Biol. 19:442-454. 22. MELECHEN, N. E., AND P'. D. SKAAIR. 1962. The 3. CLARK, A. J., AND A. D. MARCGULIES. 1965. Iso- provocationi of an early step of induction by lationi and characterizationi of recomblination thymine deprivationi. Virology 16:21-29. mutanits of E. coli K- 12. IProc. Natl. Acad. 23. MENNIGMANN, H-D. 1964. Iniduction- ill Escher- U.S. 53:451-459. ichia coli 15 of the colicinogenic factor by 4. COHJEN, S. S., AND H. D. BARNER. 1954. Studies on1 thyinineless death. Bioclhem. Biophys. Res. unllbalaniced growth in Esch/ric/iia coli. Proc. Commun. 16:373-378. Natl. Acad. Sci. U.S. 40:885-893. 24. MENNIGMANN, H-D. 1965. Phageniin-duktioni, 5. CUMMINGS, D. J., AND A. L. TAYLOR. 1966. filament-bildUng und thyminmangel-tod bei Tlhymineless death and its relation to UV sensi- Eschericlhia coli B. Zentr. Bakteriol. Parasitenk. tivity in Esc/hcric/ia coli. I'roc. Natl. Acad. Sci. Abt. I Orig. 198:282-286. U.S. 56:171-176. 25. MENNIGNIANN, H-D., AND W. SZYBALSKI. 1962. 6. DODLDNLY, C. 0. 1966. ReCquiremiienit for RNA Molecular imeclhaniismii of thyminleless death. synithesis for DNA replication in bacteria. Biochem. Biophys. Res. Communi. 9:398-404. Nature 211 :39-41. 26. NAKADA, D., AND F. RYAN. 1961. Replicationi 7. FRANIPTON, E. W. 1964. Syntlhesis of ribOnIuCieic of deoxyribonucleic acid in non-dividinig bac- acid by X-irradiated bacteria. J. Bacteriol. teria. Nature 189:398-399. 87:1369-1376. 27. OKADA, T., K. YANAGISAWA, AND F. J. RYAN. 8. FRANIPION, E. W., AND B. R. BRINKLEY. 1965. 1960. Elective production of tlhymineless Evidence of lysogeny in clerivatives of Evhehr- mutants. NatUre 188:340-341. ic/di coli. J. Bacteriol. 90: 446--452. 28. OTSUJI, N., M. SEKIGUCHI, T. IIJIMA, AND Y. 9. FREDIERICQ, P. 1963. On the natLlre of colicino- TAKAG1. 1959. Inductioll of plhage lormation gellic factors: a review. J. Theoret. Biol. 4:159- in the lysogenic Escl/ichic/i coli K-12 by mito- 165. mycini D. Nature 184:1079-1080. 10. FREIEE-.LDER, D. 1965. TechniquC for starvation of 29. PARDEE, A. B., F. JACOB, AND J. MONOD. 1959. Es /heric/iia coli of thviTline. J. Bacteriol. The genetic control ani(d cytoplasmic expressioni 90 1153--1154. of `inducibilit"' in the synthesis of ,-galacto- i 1. FREIFELDER, D., AND 0. MAAIO()E. 1964. Ercrgy side by Esc/ezrichia coli. J. Mol. Biol. 1:165 178. reqLuirement for thyrnineless death in cells of 30. I)ARDEE, A. B., AND L. S. I)RESTIDGE. 1955. In- E.sc/eri(/ia coli. J. Bacteriol. 88:987-990. duced formation of serine and tlhreoninie deami- 12. GALLANT, J., AND S. R. SUSKIND. 1962. Ribo- Inases. J. Ba2teriol. 70:667-674. nucleic acid syInthesis and thynmineless cleatlh. 31. PAULING, C., AND P. HANAWALT. 1965. Noncon- Biochliiml. Biophys. Acta 55:627-638. servative DNA replication in bacteria after 1 HANAWALT, P. C. 1963. In1volvemI1en1t of syn1thesis thymilne starvationi. Proc. Natl. Acad. Sci. U.S. of RNA in tlhyminieless deatlh. Nature 198 :286. 54:1728-1735. 1924 CUMMINGS AND MONDALE J. BACTILITOL.

32. PETTIJOHN, D., ANI) I'. C. HANAWALT. 1964. on1 the kiIetics of tlhymineless deatlh in Evlchr- Evidence for repair-replication of ultraviolet iC'hicl c-oli. J. MO1. BitOl. 20:21-28. damaged DNA in bacteria. J. Mol. Biol. 9:395- 36. SrACLY, K. A., AND E. SIMSON. 1965. Improved 410. method for the isolation of thyninle-requiring 33. SETLOW, R. B., AND W. L. CARRIER. 1964. The rnutaints of E:chreiclhia coli. J. Bacteriol. disappearance of thymiiine dimers from DNA: 90:554-555. an error-correcting mechanism. Proc. Natl. Acad. Sci. U.S. 51:226-231. 37. VAN DIF PUJTTE, P., C. WESTENBROCK, AND A. 34. SETLOW, R. B., W. L. CARRIER, AND R. W. Roms(CH. 1963. Thc relationship between genie- WILLIANIS. 1966. Three levels of DNA repair conitrolled radiation resistance and filament in E. coli. Science 152:676. formrnation in Eschlerichlii coli. Bioclhilm. Bio- 35. SMITH, B. J. 1966. Some efTects of Bromouracil phys. Acta 76:247-256.