Testing the DNA Repair Hypothesis in Bacillus Subtilis and Haemophilus Intuenzae

Evolution of Natural Transformation: Testing the DNA Repair Hypothesis in Bacillus subtilis and Haemophilus intuenzae

Rosemary J. Redfield' Department of Biochemistry, University of British Columbia, Vancouver, BritishColumbia V6T lZ3, Canada Manuscript received July 23, 1992 Accepted for publication December 15, 1992

ABSTRACT The hypothesis that the primary function of bacterial transformation is DNA repair was tested in the naturallytransformable bacteria Bacillus subtilis and Haemophilusinjluenzae by determining whether competence for transformation is regulated by DNA damage. Accordingly, DNA damage was induced by mitomycin C and by ultraviolet radiation at doses that efficiently induced a known damage-inducible gene fusion, and the ability of the damaged cultures to transform was monitored. Experiments were carried out both under conditions wherecells do not normally become competent and under competence-inducing conditions.No induction or enhancement of competence by damage was seen in either organism. These experiments strongly suggest thatthe regulation of competence does not involve a response to DNA damage, andthus that explanations other than DNA repair must be sought for the evolutionaryfunctions of natural transformation systems.

ANY groups of bacteria have evolved mecha- noncompetent cells) are provided with homologous M nisms of natural transformation which enable DNA after the cell's own DNA has been damaged by them to take upfree DNA fragmentsfrom their UV irradiation,the proportion of transformants is environments. If this DNA is homologous and genet- increased amongthe survivors (MICHOD, WOJCIE- ically marked, recombinant pregnancy are often pro- CHOWSKI and HOELZER1988). The favored interpre- duced (STEWARTand CARLSON1986). Most research- tation was that the cells that took up DNA survived ers have assumed thattransformation evolved for damage better than the cells that did not. Subsequent geneticexchange, but several recentreports have experiments have confirmed and extended the obser- addressed a possible role for transformation in DNA vations (WOJCIECHOWSKI, HOELZERand MICHOD repair (MICHOD, WOJCIECHOWSKIand HOELZER1988; 1989; HOELZERand MICHOD 199 l),but several anom- WOJCIECHOWSKI, HOELZERand MICHOD1989; HOEL- alies remain unexplained,and theincreased frequency ZER and MICHOD 199 1). These experiments were sug- gested by the DNA repair hypothesis, which postulates of transformants may be due to causes other than that transforming bacteriacan use the DNA they take repair by recombination with transforming DNA. up for recombinational repair of DNA damage, and Here I report experiments testing a different pre- that the advantages of this repair have been respon- diction of the DNA repair hypothesis: that the devel- sible for the evolution of natural transformation sys- opmental regulation of competence should reflect the tems. This proposal is not unreasonable,because DNA need for DNA repair. I have monitored changes in damage is ubiquitous and can have very high costs: the competence of the two best characterized natu- damage that spans both DNA strands is lethal unless rally transformable bacteria, B. subtilis and Huemophi- repaired and irreparable unless a homologous DNA lus infuenrae, in response to treatment with UV irra- strand is available to recombine across the damaged diation and with mitomycin C. Both of these orga- segment and serve as a template for repair synthesis nisms have SOS-like regulatory systems that induce (KUSHNER 1987). DNA repair pathways in response to DNA damage Previous experiments have tested the prediction (NOTANIand SETLOW 1980; YASBIN, CHEO and that DNA uptake will help cells survive DNA damage BAYLES1991), and in both competence is known to (MICHOD, WOJCIECHOWSKIand HOELZER1988; be tightly regulated (DUBNAU1991; REDFIELD1991). WOJCIECHOWSKI, HOELZERand MICHOD 1989; HOEL- Thus theprerequisite cellular regulatory pathways are ZER and MICHOD 1991). The initial experiments in place: the cells can recognize and respond to DNA showed that when Bacillus subtilis cultures competent damage, and they can turn competence on and off in for transformation (containing both competent and response to metabolic or environmental changes. The ' Present address: Department of Zoology, University of British DNA repair hypothesis predicts that cells will induce Columbia, Vancouver, British Columbia V6T 124 Canada. competence in response to DNA damage.

Genetics 133: 755-761 (April, 1993) 756 R. J. Redfield

MATERIALS AND METHODS IA B 1 B. subtilis strainsand culture conditions: Cellswere grown inDifco antibiotic medium #3 (PAB) and in the competence induction media GM 1 andGM2 supplemented with tryptophan and methionine at 50 rg/ml (BOYLANet al. 1972). Cells were plated on LB or minimal glucose (MG) agar, supplemented with tryptophan and/or methionine where appropriate. Damage and transformation experiments were carried out in strain YB886 (metB5 trpC2 xin-I 0 20 40 Bo SPF) using transforming DNA from strain WBlOl (xin-I Mitomycin C (pghl) UV irradiation (J/mz) spa-). B. subtilis DNA damage protocols:Cells grown to OD600 FIGURE1 .-Induction of the din22 lac2 fusion by DNA damag- = 0.3 in PAB (log-phasecultures), or tostationary phase (To ing agents. Samples were assayed 1 hr after exposure to (A) mito- + 90) in GM1 (pre-competent cultures; BOYLANet al. 1972), mycin C or (B) UV irradiation. Solid squares represent cultures were damaged as follows. growing exponentially in PAB; open squares represent competent Mitomycin C: Cells were incubated for 10 min at room cultures. temperature with the specified concentration of mitomycin C, washed once, and resuspended in the original volume of additional 5 min, and cells were diluted and plated on sBHI PAB or GM2. agar with or without novobiocin at 2.5 pg/ml. W irradiation: Cells were pelleted and resuspended in DNA uptakeassays: 'H-LabeledDNA was prepared Spizizen minimal salts (SPIZIZEN1958) at room temperature. from strain KW26 (thy str") as described (TOMBet al. 1989). Aliquots (1 ml) were irradiated in plastic Petri dishes with 1 ml samples of competent H. inf2uenzae cultures were constant agitation under a small 256 nm UV lamp (intensity incubated with 1 rg of 'H-labeled KW26 DNA and with at adistance of 15 inches was 0.4 J/m'/sec). Cells were then DNase I, as described above. The cells were then pelleted, pelleted and resuspended in the original volume of PAB or washed three times with cold MIV, and the final pellet was GM2. counted in Aquasol. The OD600was measured when each After damage with either agent, 1ml samples were taken sample was taken, and uptake was expressed as Pg DNA/ immediately to tubes with and without transforming DNA OD600. for t = 0 survival and transformation assays, and the re- mainder of each culture was diluted 20-fold in prewarmed RESULTS PAB or 10-fold in prewarmed GM2 and shaken at 37". Samples were taken for transformation assays at intervals B. subtilis over the next 4 hr. How much damage is required to induce damage- B. subtilis assays: Transformation assays: Culture aliquots (1 ml) were incubated with or without 1 rg WBlOl DNA inducible gene expression? To ensure that the dam- at37" for 15 min, 1 rg of DNase I was addedfor an aging agents UV andmitomycin C were used at doses additional 5 min, and cells were diluted and plated on LB sufficient to induce damage-regulated genes, I meas- to determine survival and growth, and on MG plates with ured induction ofP-galactosidase expression from the and without tryptophan and/or methionine to determine damage-inducible lac2 fusion din-22 (LOVE,LYLE and transformation frequencies. /3-Galactosidase assays: Cultures of a YB886 derivative YASBIN 1985), using the same damage protocols that containing the din-ZZ::Tn917-lacZ insertion (LOVE,LYLE would be tested for competence induction. The data and YASBIN 1985), were damaged and incubated as de- in Figure 1 show that exposure to 20 J/m' of UV or scribed above, and &galactosidase activity was assayed in to 6 pg/ml of mitomycin C for 10 min induced maxi- toluenized cells as described by MILLER (1972). mal &galactosidase expression in both competent and influenxu strains and culture conditions:DNA dam- H. low age experiments were carried out using strain BC200 (SET- exponentially growing cells. Survivals were very LOW et al. 1973). Cells were grownin Difco brain heart atthese doses (no more than 2% for YB886;see infusion supplemented with 10 pg/ml hemin and 2 pg/ml Figures 2 and 3). Accordingly, in addition to a maxi- NAD (sBHI), and competence was induced by transfer of mally inducing dose of each agent, two lower doses mid-log cultures to MIV starvation medium (HERRIOTT, were used, one giving at least50% survival of YB886 MEYERand VOGT1970). and about 10% induction of din-22, and one giving H. influenxu DNA damage protocols: These were similar to those used for B. subtilis. All began with exponentially about 20% survival of YB886 and about 50% induc- growing cells at ODco0= 0.2-0.25 in sBHI. After incubation tion of din-22. with mitomycin C, cellswere washed in sBHI (for noncom- Effects of DNA damage on noncompetent B. sub- petent cultures) or MIV (for competent cultures) and resus- tilis cultures: B. subtilis cells do not normally become pended in an equal volume of sBHI or MIV. Cells were UV competent while growing in PABor other rich media. irradiated in MIV, and pelleted and resuspended in the original volume of sBHI or MIV. Samples were then taken The effect of DNA damage on competencewas there- for t = 0 survival and transformation assays, and the re- fore firstexamined by exposingcultures of strain mainder diluted in prewarmed sBHI or MIV and shaken at YB886 growing exponentiallyin PAB to various doses 37". of mitomycin C and UV, and then testing aliquots H.influenzu assays: Transformation assays: One milliliter overthe ensuing 4 hr (Figure 2). Controlcultures of cells was incubated at 37" with 1 pg chromosomal DNA from strain MAP7 [no# str" kanRst8 sp~? nulRuioR; BARCAK were mock-irradiated or incubated for10 min without et al. (1991)l for 15 min, 1 pg of DNase I was added for an mitomycin. To measure the extent of induction of Competence and the SOS Response 757 Mitomycin C UV irradiation Mitomycin C UV irradiation

I,.,.,, 10-2 In

-- LI' ~~ ri sf 6 1M ' &-,m . 1M 2w k.9' Timeafter damage (min) Time after damage (min) FIGURE3.-Effect of DNA damaging agents on competent B. subtilis cultures. Cells at T0+90 inGM1 were damaged at room temperature, diluted tenfold in GM2, and returned toculture (time = 0). Panels A, C and E: cells damaged with mitomycin C: solid Time after damaae (min) circles undamaged, open circles 1 rg/ml, open squares 3 pg/ml, FIGURE2.-Effect of DNA damaging agents on noncompetent open triangles 7.5 Pg/ml. Panels B, D and F: cells damaged with B. subtilis cultures. Cells growing exponentially in PAB were dam- UV irradiation: solid circles undamaged, opencircles 10 J/m', open aged at room temperature, diluted 20-fold in PAB, and returned squares 30 J/m', open triangles 60 J/m'. Panels A-D: YB886. Panels to culture (time = 0). Panels A-F: YB886. Panels G and H:din-22. A and B: cfu on LB. Panels C and D: ratio of cfu on MG + Panels A, C, E, G: cells damaged with mitomycin C: solid circles tryptophan/MG + tryptophan + methionine. The upper sets of undamaged,open circles 1 pg/ml, open squares 3 pg/ml, open curves in Panels C and D are from samples incubated with WBlOl triangles 6 pg/ml. Panels B, D, F and H: cells damaged with UV DNA at the indicated times; the lower sets of curves are from irradiation: solid circles undamaged,open circles 8 J/m*, open samples incubated without DNA. Panels E and F: @-galactoside squares 16 J/m*, open triangles 24 J/m'. Panels A and B: cfu on activity (Miller units) of din22 culturestreated as above. The LB. Panels C, D, E, F: ratio of cfu on MG + tryptophan/MG + following data points in the lower set of curves (-DNA) of Panel C tryptophan + methionine. The shaded regions in panels C and D are upper limits, because no methionine+ (Met+) colonies were indicate the typical range of transformation frequencies of compe- detected from 1 ml of culture; undamaged, 0 min; 6 pg/ml dose, tent cultures. Panels A-D are samples incubated with WBlOl DNA 0, 20 min. at theindicated times. Panels E and F are samples incubated without DNA. Panels G and H: @-galactosidaseactivity (Miller units) of A and B of Figures 2, 3, 4 and 5 have been divided din22 cultures treated as above. The following data points are by the appropriate dilution factor before graphing.) upper limits, because no methionine+ (Met+)colonies were detected from 1 ml of culture. Panel C, 6 pg/ml dose: 50,90, 135 min. Panel The three doses of mitomycin C and of UV used E, 3 pg/ml dose: 20,240 min. Panel E, 6 pg/ml dose: 50, 90 min. reduced survivals at t = 0 to 59%, 15% and 0.9%, and 6995, 21% and 2.0%, respectively. At the end of the damage-inducibleoperons, the din-22 lacZ fusion experiments the undamaged cultures had enteredsta- strain was similarly damaged and its &galactosidase tionaryphase, while the heavily damagedcultures activity was measured over the same time period. were still growing exponentially. Damage-induced 8- Figure 2, A and B, shows the effect of damage on galactosidase activity in the din-22 lacZ fusion strain the ability of cells to form colonies. The initial drop had reached maximal levels and begun to fall by this in colony-forming units (cfu) of the undamaged cul- time (Figure 2, G and H), indicating that 4 hr allowed tures (solid lines and circles) was due to theinitial l/ sufficient time for expression of damage-induced 20 dilution; the lower cfu of damaged cultures indi- changes. cates the lethality of the DNA damage. (Althoughthe The solid lines in Figure 2, C and D, confirm that t = 0 samples were actually taken just before dilution thetransformation frequencies of undamaged cells of the rest of the culture, tofacilitate comparisons of under these conditions were only slightly higher than survival and growth rates thet = 0 cfu values in panels the spontaneous mutation rate toMet' (between 10-6 758 R. J. Redfield Mitomycin C UV irradiation Mitomvcin C UV irradiation 16

Timeafter damage (min) Time after damage (min)

FIGURE4.-Effect of DNA damaging agents on non-competent H. influenrue cultures. BC200 cells growing exponentially in sBHI uool~.... ,.,.,,,!...,...,....I were damaged at room temperature, diluted in fresh sBHI, and -50 0 50 1w 150 250204 0 50 1w 150 200 250 returned to culture at time 0. Panels A, C and E: cells damaged Timeafter damage (min) Time after damage (min) with mitomycin C and diluted fivefold: closed circles undamaged, open circles 0.15 pg/ml, open squares 0.75 &ml, open triangles FIGURE5.-Effect of DNA damaging agents on competent H. 2.0 &ml. Panels B, D and F: cells damaged with UV and diluted infuenzae cultures. BC200 cells at OD600 = 0.25 in sBHI were 1/5: closed circles undamaged, open circles 2 J/m'. open squares 6 damaged at room temperature, diluted in MIV, and returned to J/m', open triangles 24 J/m'. Panels A and B: cfu on sBHI. Panels culture at time0. Panels A, C and E, cells damaged with mitomycin C, D, E and F: ratio of cfu on sBH1 with and without novobiocin. C and diluted threefold:closed circles no damage, opencircles 0.25 Panels C and D: cells incubated with MAP7 DNA at the indicated pg/ml, open squares 0.75 pglml, open triangles 2.0 pg/ml. Panels time. Shaded regions show typical transformation frequencies of B, D and F, cells damaged with UV and diluted fivefold: closed fully competent cultures. Panels E and F: cells incubated without circles no damage, open circles 2.0 J/m', open squares 6.0 J/m', DNA. The data for 0.15 pg/ml mitomycin C and for 24 J/m' UV open triangles 12J/m'. Panels A and B: cfu on sBHI. Panels C and are from a separateexperiment. The following datapoints are D: ratio of cfu on sBH1 with and without novobiocin. Panels E and upper limits because no NovR colonies were detected from one ml F: uptake of 'H-labeled KW26 DNA (as DNase I-resistant cpm), of culture: Panel C, 0.75 pg/mldose: 0, 30, 70 min; 2.0 rg/ml normalized for culture density measured as OD~OO. Thefollowing dose: 0, 30, 70, 135 min. Panel D, 2.0 J/m' dose: 0 min, 6.0 J/m' data points for 0 min samples are upper limits because no NovR dose: 0, 70 min; 24 J/m' dose, 30 min, 70 min. Panel E, 0.75 rg/ colonies were detected from 1 ml of culture. Panel C, undamaged, ml dose: 30, 70 min; 2.0 pg/mldose: 0, 30, 70min. Panel F, 0.25 pg/ml dose, 2.0 pg/ml dose; panel D, all doses. undamaged and 2.0 J/m' doses: 30 min; 6.0 J/m' and 24 J/m' doses: 30, 70 min. (such points are specified in the legend to each figure). and lo" per cfu, Figure 2, E and F). Both mitomycin In some cases sensitivity was also limited by the fre- C and UV caused substantial increases in the frequency of reversion mutations to Met+, however se- quency of Met+ colonies, but this was independent of lection for Trp+ gave similar results (data not shown). exposure to transforming DNA (Figure 2, E and F Although the trpC2 mutation reverts less frequently are drawn to the same scale as Figure 2, C and D), than does metB5 (it is thought to be a small deletion), indicating thatthe increases were dueto damage- it also transforms less efficiently, and the sensitivity induced mutagenesis rather than to competence. In was again limited primarily by the survival. all cases the frequency of Met+ colonies remained less Effects of DNA damage on competent B. subtilis than 1% of thatproduced by competentcultures cultures: The above experiments showed that DNA induced by growth in GM1 and GM2 (shaded bars at damage is not the primary inducer of competence in the tops of Figure 2, C and D; for comparison see B. subtilis. However, the possibility remains that DNA Figure 3, C and D). repair is a secondary function of transformation, and Although damage did not induce detectable com- that in rich medium induction of competence is ac- petence, the sensitivity of these experiments was lim- tively prevented by other regulatory mechanisms. Ac- ited by the density of surviving cells in the cultures: cordingly I tested whether the presence of DNA dam- most of the data points for the highly damaged cul- age might enhance the development of competence tures in Figure 2, C and E, are based on three or under standard inducing conditions (Figure 3). fewer colonies from 1 ml of culture, and several are Maximal competence in B. subtilis is customarily upper limits as no Met+ colonies could be detected induced by growing cells into stationary phase Com petence and the andCompetence SOS Response 759

(“T0+90”) in the minimal salts/glucose/amino acids/ toabout at the onset of stationary phase (RED- yeast extract medium GMl, and diluting them 10- FIELD 1991). Accordingly, BC200 cells were grown in fold into GMl supplemented with Mg2+ and Ca2+ sBHI to mid-log (ODsoo = 0.2), damaged,and re- (GM2), where they reach maximal competence in 1 turnedto sBHI at a fivefold dilution;growth and hr (BOYLANet al. 1972). In the experiment shown in competence were followed for 4 hr (Figure 4). Figure 3, cells at To+90 in GMl wereexposed to The solid lines and circles show that after dilution mitomycin C or UV before dilution into GM2, and the undamaged cells initially grew exponentially and survival, transformation andmutation frequencies then entered stationary phase (Figure 4, A and B). As were followed as before.Undamaged cultures and expected,transformation frequencies during expo- induction of din-22 were again followed under iden- nential growthwere about (Figure 4, C and D), tical conditions. only slightly higher than the mutation rate (Figure 4, Figure 3, A and B, shows the survival and growth E and F), and spontaneous competence peaked at the of the YB886 cultures. The three doses of mitomycin onset of stationary phase, giving transformation fre- C and of UV reduced survivals at t = 0 to 51%, 18% quencies of about 1O-4. and 0.6%, and 77%,27% and 0.6%, respectively. The three doses of mitomycin C and of UV reduced Figure 3, E and F, shows that these doses also induced survivals to 66%, 11%and 1.6%,and 53%, 20%and high levels of P-galactosidase activity. 9%, respectively. These doses did not detectably in- The upper sets of lines in Figure 3, C and D, show crease transformationfrequencies in exponential the frequency of Met+ cfu from samples incubated growth,and reduced the spontaneous competence with transforming DNA; thelower groups of lines are seen atthe end of theexperiment. Although the from the controlsamples incubated without DNA. As expected, undamaged cultures reachedmaximal com- transformation and mutation frequencies appear to petence at 1 hr, with transformation frequencies of be initially increased by DNA damage,these data 3-10%. Damaged cultures showed very similar kinet- points are upperlimits, because no NovRcolonies were ics and transformation frequencies, except that com- obtained from the heavily damaged cultures for the petence development was slightly delayed byheavy first 2 hr of the experiment. All of these values are damage. Competence was decreasing in all cultures far below the fully induced competence levels indi- by the end of the experiment. Although the DNA cated by the shadedbars in Figure 4, C and D, damage did induce some mutations,these did not (compare also to those in Figure 5, C and D), indicat- limit the sensitivity of the experiment; the mutation ing that DNA damagedid not induce or enhance frequencies were always lower than the transforma- competence development in H. injluenzae in rich me- tion frequencies, and the maximum transformation dium. frequencies were always at least 1000-fold higher than Effects of DNA damage on competent H. influen- the mutation frequencies. This experiment thus dem- zae cultures: In H. injluenzae maximum competence onstratesthat DNA damagedoes not enhance the develops after rapidly growing cells are transferred to development of competence in B. subtilis under per- the starvation medium MIV (transformation frequen- missive conditions. cies about lo-*) (HERRIOTT, MEYERand VOGT 1970). H. influenzae Accordingly cells at OD600= 0.2 in sBHI were washed with MIV, exposed to mitomycin C or UV, and The natural transformation system of H. injluenzae washed and resuspended in MIV at a final dilution of typifies those of gram negative bacteria, and appears 1/3 or 1/5. Growth, transformation, andDNA uptake to be unrelated to that of B. subtzlis and other gram positive organisms (STEWARTand CARLSON1986). It by these cultures were followed for 4 hr (Figure 5). is therefore suitable for testing the generality of the As expected, undamaged cells ceased dividing after above results. The BC200 derivative of the standard transferto MIV, and transformationfrequencies H. injluenzae Rd strain was used, because it does not peaked at 70 min. The peak competence of 3 x lo-’ inducea defective prophage in response to DNA was slightly lower than usually seen, presumably due damage and hence is less sensitive to damaging agents to the delays incurred during the damaging andwash- than its parent (SETLOWet al. 1973). Similar results ing steps. The three doses of mitomycin C and of UV were seen with the standard Rd strain KW20 (data reduced survivals to 43%, 22% and 9.7%, and 62%, not presented). 40% and 23%, respectively. The transformation fre- Effects of DNA damage on noncompetent H. in- quencies of the damaged cultures were all lower than fluenzae cultures: The basic design was similar to that those of the controls until the final time point, and as used for the B. subtzlis experiments described above. seen with B. subtilis, the most heavily damaged cul- H. influenzae growing exponentially in the rich me- tures showed some delay in reaching maximal com- dium sBHI exhibits transformation frequenciesof less petence. All cultures showed decreasing competence than 1O-’, but the frequency increases spontaneously by the end of the experiment, confirming that suffi- 760 R. J. Redfield cient time had been allowed for any damage-induced regulated in both H. injluenzae and B. subtilis, and changes in competence. because these regulatory mechanisms evolved to per- Measurement of DNA uptakeprovides an inde- mit transformation only under conditions when it pendent assay of the competence of the entire culture, would benefit the cell, they provide direct windows not just of those cells capable of forming colonies. into transformation's functionsin the natural environ- Figure 5, E and F, shows that DNA damage decreased ments where the regulatory genes evolved. Thus, if uptake of 'H-labeled DNA in parallel with the de- the primary functionof transformation is indeed DNA creases in transformation frequencies. Thus all of the repair, we expect that competence, like other repair evidence shows that, in H. injluenzae as in B. subtilis, pathways, will be induced by DNA damage. DNA damaging agents neither induce nor enhance In investigating whether competence is a compo- the development of competence. nent of the SOS response, it was not possible to test all permutations of damaging agents and culturecon- DISCUSSION ditions, but a reasonablerange has been covered.Two damaging agents were used; both are potentinducers Transformation has usually been assumed to serve of the SOS response, and both can generate lesions the same function as sex, where an evolutionary ad- requiring recombinational repair [reviewed by KUSH- vantage is thought to accrue from the production of NER (1987)l. uv causes mainly single-strand pyrimi- recombinant progeny, enabling populations to better dinephotodimers that can berepaired without an cope with deleterious mutations or changing environ- exogenous template, although the lethality associated ments [see collection of articles in and LEVIN MICHOD with high UV doses may result from other photoprod- (1988)l. Because the advantages of sexual recombi- ucts andstrand breaks, and from replication and nation remain controversial,the evolution of transfor- segregation of strands containing unrepaired lesions mation may best be approached by investigating the (BRASH1988). The bifunctional alkylating agent mi- other ways it might benefit the cell. Recombinational tomycin C reacts with guanine residues to cause dou- repair with transforming DNA is a plausible alterna- ble-strand crosslinks, which are lethal unless repaired tive function, because cells must cope with DNA dam- by recombination. Cells were tested under two culture age much more often than with changing environ- conditions, one wherecompetence did not sponta- ments or deleterious mutations, and because the con- neously develop (to find out whether damage could sequences of damageare more likely to be induce competence) and one where competence de- immediately lethal. velopment had been initiated (to find out whether Previous experiments have attemptedto demon- damage would enhance competence under conditions strate that DNA repair using transforming DNA in- known to permit its development). The hr of obser- creases the survival of B. subtilis cells after DNA 4 damage (MICHOD,WOJCIECHOWSKI and HOELZER vation should be sufficient for any damage-induced 1988; WOJCIECHOWSKI,HOELZER and MICHOD 1989; effects to appear. H. injluenzae can develop compe- HOELZERand MICHOD 1991). However, the effect is tence in as little as 30 min, although maximal induc- expected to be small, because the frequency of suc- tion by starvation medium usually takes 90-100 min. cessful repair will be limited by the amount of DNA Comparable times are not known for B. subtilis, but the cells take up, usually equivalent to only a small high levels of @-galactosidaseactivity from the din-22 fraction of the genome. The enhanced survival seen fusion were induced in all controls. Under conditions in these experimentswas often much larger than could where competence developed, it had usually peaked beattributed to transformational repair, unless a and begun to decrease by the endof each experiment. novel mechanism was postulated that enabled cells to Although the present experiments revealed no in- take up only the DNA fragments that matched their duction of competence by DNA damage, theconverse damaged sequences. doesnot hold: some DNA repair processes are in- The need for caution in interpreting these experi- duced by competence. In H. influenzae both prophage ments is emphasized by recent attempts to reproduce induction and recombination by the RecA-homolog the B. subtilis effects in H. influenzae, where cell sur- Rec-1 are increased in competent cells (BOLLINGand vival can be measured directly rather than inferred SETLOW1969; SETLOWet al. 1973). RecA mediates from the transformation frequency. The presence of generalized recombination as well as regulating DNA transforming DNA after damage with UV or mito- repair, so this induction may reflect its role in proc- mycin C was found to have no effect on the survival essing incoming DNA. In B. subtilis the connection of cultures that had completed competence develop- between competence and the SOS response is much ment (REDFIELD1993; MONCOLD 1992). broader, and most damage-inducible processes and An alternative approach is to infer the function of damage-inducible promoters are induced in undam- transformation systems froman understanding of aged competent cells, possibly by competence-specific their regulatory mechanisms. Competence is tightly activation of the recA and dinR promoters (YASBIN, Competence and the SOS Response 76 1

CHEOand BAYLES199 1 ; RAYMOND-DENISEand GUIL- of Bacillussubtilis defective in synthesis of techoic acid. J. Bacteriol. 110 28 1-290. LEN 1992). We do not know whether this evolved only BRASH,D. E., 1988 UV mutagenic photoproducts in Escherichia to provide competent cells with enhanced recombi- coli and human cells: a molecular genetics perspective on hu- nation capabilities, or whether the associated induc- man skin cancer. Photochem. Photobiol. 48 59-66. tion of other repair processes is in itself adaptive. CHANDLER,M. S., 1992 The gene encoding cyclic AMP receptor Our present knowledge aboutthe regulation of protein is required for competence development in Haemophi- lus influenzae Rd. Proc. Natl. Acad. Sci. USA 89 1626-1630. competence provides some support for an alternative DUBNAU,D., 1991 Genetic competence in Bacillus subtilis. Micro- function for competence, use of DNA as a nutrient. biol. Rev. 55 395-424. In H. injluenzae competence is triggered by addition HERRIOTT,R. M., E. M.MEYER and M. VOGT, 1970 Defined of cyclic AMP (WISE,ALEXANDER and POWERS1973), nongrowth media for stage I1 development of competence in and is completely blocked by mutations in adenylate Haemophilw influenrae. J. Bacteriol. 101: 5 17-524. HOELZER,M., and R. MICHOD, 1991 DNA repair and the evolu- cyclase and in the catabolite regulatory protein CRP tion of transformation in Bacillus subtilis.111. Sex withdamaged (I. DOROCICZand R. REDFIELD,unpublished results; DNA. Genetics 128: 215-223. CHANDLER1992), suggesting participationin a global KUSHNER,S. R., 1987 DNA repair, pp. 1044-1053 in Escherichia carbon-energyregulon (BOTSFORD and HARMAN coli and Salmonella typhimurium. American Society for Micro- B. subtilis biology, Washington D.C. 1992). In competence development is con- LOVE,P. E., M. J. LYLEand R. E. YASBIN, 1985 DNA-damage- trolled by a complex regulatory cascade (DUBNAU inducbile (din)loci are transcriptionally activated in competent 1991), several of whose components also regulate Bacillus subtilis. Proc. Natl. Acad. Sci. USA 826201-6205. starvation-induced genes and other postexponential MICHOD,R. E., and B. R. LEVIN(Editors), 1988 TheEvolution of processes. Sex. Sinauer, Sunderland, Mass. MICHOD, R. E., M. WOJCIECHOWSKIand M. HOELZER,1988 DNA The lack of evidence for regulation of competence repair and the evolution of transformation in the bacterium by DNA damage does not mean that cells do not use Bacillus subtilis. Genetics 118: 31-39. transforming DNA for recombinantional repair, but MILLER,J. H., 1972 Assay of beta-galactosidase, pp. 352-355 in only that any such benefits have been too small to Experiments in Molecular Genetics. Cold Spring Harbor Labo- influence the evolution of the regulatory mechanisms. ratory, Cold Spring Harbor, N.Y. MONCOLD,J. A., 1992 DNA repair and the evolution of transfor- Similarly, the evidence for nutritional regulation does mation in Haemophilus influenzae. Genetics 134: 893-898. notpreclude a benefit from DNA repair or from NOTANI,N. K., and J. K. SETLOW,1980 Inducible repair system production of recombinant genomes. Because most in Haemophilus influenzae unaccompanied by mutation. J. Bac- deleterious mutations and environmental changeswill teriol. 143: 5 16-5 19. RAYMONDDENISE,A., and N. GUILLEN,1992 Expressionof the have their main impact on the cell’s nutritional state, Bacillus subtilis dinR and recA genes after DNA damage and nutritional regulation could reflect a recombinational during competence. J. Bacteriol. 174: 3171-3176. function. Furthermore, because the DNA strand dis- REDFIELD,R. J., 1991 sxy-I, a Haemophilusinfluenzae mutation placed by recombination with transforming DNA is causing greatly enhanced competence. J. Bacteriol. 173: 5612- subsequently degraded, cells can have their genetic 5618. REDFIELD,R. J., 1993 Genes for breakfast: the have-your-cake- cake and eat it, too, benefiting from both the new and-eat-it-too of transformation. J. Hered. (in press). genetic information and all of the nutrients in the SETLOW,J. K., M. E. BOLING,D. P. ALLISONand K. L. BEATTIE, DNA they take up. 1973 Relationship between prophage induction and transformation in Haemophilus influenzae. J. Bacteriol. 115 153-161. I thank the Medical Research Council of Canada and the Cana- SPIZIZEN,J., 1958 Transformation of biochemically deficient dian Institutefor Advanced Research for financial support, P. strains of Bacillussubtilis by deoxyribonucleate. Proc. Natl. WILLIAMSfor technical assistance, M. WOJCIECHOWSKIfor B. subtilis Acad. Sci. USA 44 1072-1078. strains, and J. MONGOLD, C.BERNSTEIN, M. WOJCIECHOWSKIand STEWART,G. J., and C. A. CARLSON,1986 The biology of natural R. MICHODfor helpful discussions. transformation. Annu. Rev. Microbiol. 40 21 1-235. TOMB,J. F., G. J. BARCAK,M. S. CHANDLER,R. J. REDFIELDand H. 0.SMITH, 1989 Transposon mutagenesis, characteriza- LITERATURECITED tion, and cloning of transformation genes of Hacmophilw influenzae Rd. J. Bacteriol. 171: 3796-3802. BARCAK,G. J., M. S. CHANDLER,R. J. REDFIELDand J.-F. TOMB, WEE, E. M., S. ALEXANDERand M. POWERS,1973 Adenosine 1991 Genetic systems in Haemophilus influenzae. Methods En- 3’:5’-cyclic monophosphate as a regulator of bacterial transfor- zymol. 204: 321-342. mation. Proc. Natl. Acad Sci. USA 70 47 1-474. BOLLING,M., and J. SETLOW,1969 Dependence of vegetative WOJCIECHOWSKI, M.,M. HOELZERand R. E. MICHOD,1989 DNA phage recombination among Haemophilw influenzae bacterio- repair and the evolution of transformation in Bacillus subtilis. phage on the host cell. J. Virol. 4 240-243. 11. Role of inducible repair. Genetics 121: 41 1-422. BOTSFORD,J. L., and J. G. HARMAN,1992 Cyclic AMP in prokar- YASBIN, R.E., D. CHEOand K. BAYLES,199 1 The SOB system of yotes. Microbiol. Rev. 56 100-122. Bacillus subtilis: a global regulon involved in DNA repair and BOYLAN,R. J., N. MENDELSON,D.BROOKS and F.E. YOUNG, differentiation. Res. Microbiol. 142 885-892. 1972 Regulation of the bacterial cell wall: analysisof a mutant Communicating editor: J. E. BOYNTON