Effect of induction of SOS response on expression of pBR322 genes and on plasmid copy number

PLASMID

22, 163-168 (1989)

Effect of Induction

of SOS Response on Expression and on Plasmid Copy Number

of pBR322

Genes

ELISABETHBERTRAND-BURGGRAF,PASCALEOERTEL,MANFRED SCHNARR, MICHEL DAUNE, AND MICHI?LE GRANGER-SCHNARR Institut de Biologie Moltkulaire et Celiulaire du CNRS, 15 rue Renk Descartes,67084 Strasbourg Ckdex, France Received March 23, 1989; revised August 15, 1989 Several lines of evidence are presented that indicate that the level of tetracycline resistance of

Esherichia coli strains harboring plasmid pBR322 varies according to whether the SOS system of the host bacteria has been induced. These include use of strains in which the SOS system is expressed constitutively (lexA def.), is thermoinducible (recA441) or noninducible (leti ind-), or is highly repressed (multiple copies of led+). Similar induction was observed with the product of another plasmid gene, fl-lactamase. The amounts of extractable plasmid DNA were also increased by SOS induction, and we propose that the SOS-induced increases in levels of tetracycline resistance and p-lactamase activity are due to an increased plasmid copy number. Q 1989 Academic press, IX.

Escherichia coli has a complex response to DNA damaging agents that allows survival under adverse conditions. This “SOS response” comprises a set of at least 17 genes whose products are involved in several functions such as excision repair of damaged DNA, inhibition of cell division, and increase of mutational frequency (for a review see Little and Mount 1982; Walker, 1984). Under usual conditions, all these genes are negatively regulated by a single repressor, the dimeric form of the LexA protein. The LexA repressor is inactivated by a specific cleavage into two fragments after DNA damage by chemicals or radiation (Little, 1983). One ofthese fragments retains substantial DNA binding affinity in vitro (Hurstel et al., 1986, 1988; Bertrand-Burggraf et al., 1987) and in vivo (Little and Hill, 1985), but cannot dimerize. The lack of dimerisation lowers binding and elicits enhanced transcription of SOS genes (Markham et al., 1984). The specific cleavage of LexA is an autocleavage reaction, which is stimulated by an activated form of RecA protein (RecA*) (Slilaty and Little, 1987). Both genes lexA and recA are themselves SOS genes and thus the SOS system is autoregulated. Plasmid pBR322 confers resistance to tetracycline and ampicillin. Here we show that the degree of resistance to tetracycline and the 163

amount of P-lactamase (ampicillin resistance) are correlated with SOS induction. We compared the growth of several strains (DM 1180, DM I 187, and GC32 17, see Table 1) transformed with the plasmid pBR322, as a function of the tetracycline concentration. Figure 1 shows that the level of resistance, assessedby the ability to grow at different levels of tetracycline, is correlated with the type of SOS regulation. 1. The recA441 allele, which results in cleavage of wild-type LexA repressor at high temperature, causes thermoinducible increase in resistance. 2. The lexA5 1 allele, which inactivates the LexA repressor function, causes increased tetracycline resistance at high and low temperatures. 3. The ZexA3allele, which makes repressor noncleavable, counteracts the thermal induction of tetracycline resistance by the recA44I allele. These results suggest that inactivation of the LexA repressor by a mutation within the 1exA gene (DM 1187) or by RecA*-mediated cleavage in a way which mimics SOS induction leads to an important increase in tetracycline resistance of bacteria harboring pBR322. 0147-619X/89

$3.00

Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form resaved.

164

SHORT COMMUNICATIONS

0

: 5

10

15

20

25

J-Jg/ml

5

10

15

20 JJg/m(

30

35

40

35

40

tet

25

30

tet

FIG. 1. Tetracycline resistancelevels: relative optical density (R) of different strains after 3 h of growth in Lmia broth medium with ampicillin (50 rcg/ml) and as a function of the tetracycline concentration. Starting cell density was 0.01 OD unit, and all cells were still growing exponentially after 3 h. R = ODyz (after 3 h)/OD\E’ (after 3 h). The strains(all transformed with pBR322) are (A) DM I 187(led-) (*, a) and DM 1180 (lex,4 ind-) (0,O) at 30 or 42°C; and (B) GC3217 (recA441) (El, W)at 30 or 42°C.

To test the effect of overproduction of the LexA protein, we have compared the resistance to tetracycline of a wild-type strain AB 1157 transformed either with a plasmid producing an active LexA protein (pMG 14) or with a similar plasmid producing an inactive LexA protein (pPO4). The results (Fig. 2)

show that overproduction of LexA is associated with increasedsensitivity to tetracycline. The possibility that LexA directly regulates the tet genehas been investigated by the search of a stable complex between the LexA repressor and the proximal part of the tet gene ineluding its regulatory sequence (the EcoRI-

SHORT TABLE

COMMUNICATIONS

1

Bacterial strain GC3217

recA441, s&41 1

DMl180

as GC32 17 but lexA3 (hid-) as DM1180 but leti (Def)

DM1187

George et al., 1975 Mount, 1977 Mount, 1977

165

coli involves “stable DNA replication” (Witkin and Kogoma, 1984) leading to an increase in the amount of chromosomal DNA per bacterium. Therefore, we have simply estimated the amount of plasmid DNA extracted from minilysates of the different strains (Fig. 3). The results are in favor of a variation in the plasmid copy number since we measured five- to sevenfold more plasmid DNA per DM 1187 bacterium than per DM1180 bacterium without significant changes upon the temperature growth conditions. Furthermore, we observe that GC32 17 bacteria grown at 42°C present three- to fourfold more plasmid DNA than bacteria grown at 30°C. All those strains are $4 (able to divide after DNA damage), and we have also seen by light microscopy that there is no significant variation in the cell size (variations in the plasmid copy number could be simply due to dilution of the control factors of plasmid copy number (Tomizawa and Itoh, 1981; Davidson, 1984 for review). An increase in plasmid copy number was also observed in a wild-type E. coli strain after induction of the SOSgenesby uv irradiation. Figure 4 shows that while the amount of plasmid extracted from a nonirradiated culture is relatively stable, it increasessignificantly after uv irradiation, with a maximum 90 min after irradiation. The initial level is reached again

BumHI fragment of pBR322). As previously shown (Schnarr et al., 1985), specific complexes between LexA and end-labeled DNA fragments can be visualized on polyacrylamide gelsusing the Fried and Crothers methodology ( 1981). Even the weakestSOSoperators (uvrA, uvrB, uvrC) give rise to a well-defined specific complex on those gels (M. Kazmaier, M. Granger-Schnarr, and M. Schnarr, unpublished results). Since no specific complexes (data not shown) are observed, we conclude that LexA does not directly regulate the transcription initiation of the tet gene. We asked further if the bla gene (coding for the P-lactamase), also carried by plasmid pBR322, is subject to the same apparent dependenceon the stateof the SOSsystem.Table 2 shows the relative amount of P-lactamase activity of each transformed strain compared to that of the GC3217 strain grown at 30°C. TABLE 2 The results are similar to those obtained with the tet gene; i.e., the induction of the SOSsysRELATIVE /3-LACTAMASE ACTIVITY OF EACH tem of the host bacteria leads to an increase BACTERIAL STRAIN RELATED TO THAT OF THE GC32 17 STRAIN GROWN AT 30°C (TAKEN AS REFERENCE) of the amount of P-lactamaseactivity. Therefore, we can conclude that the prodRelative pucts of both the tet and the bla genesare presTemperature lactamase ent within the cell in higher amounts when Bacterial strain activity (“C) the SOSsystemof the host bacteria is induced. 30 1 kO.3 Finally, we have checked if these variations GC32 17 +pBR322 42 2.6 + 0.3 can be correlated with variation of the plasmid GC32 17 +pBR322 DM 1187 +pBR322 30 2.9 + 0.3 pBR322 copy number. A precise determina- DM1187 +pBR322 42 3 kO.3 tion of the amount of plasmid DNA per bac- DMl lSO+pBR322 30 0.6 f 0.3 terium is delicate and generally achieved DMI 180+pBR322 42 0.2 + 0.2 through a concomitant determination of plasa The activity of the bla gene of pBR322 carried by the mid and chromosomal DNA. However, this different strains was measured as described by Lindstrom requires that the amount of chromosomal and Nordstrom (1972) and is expressed in arbitrary units DNA remains essentially constant. Unfortu- per number of bacterial cells (expressed in OD units). All nately, the induction of the SOS system in E. measurements were performed in duplicate.

SHORT COMMUNICATIONS

B ’

12

34

5

6

Hours

78910

12

34

5

6

78

910

Hours

FIG. 2. Growth of ABll57 harboring plasmid pPO4 (A) or pMGl4 (B). Plasmid pMG14 was obtained by cloning a CluI restriction fragment of pJL45 (Little et al., 1981) containing the 1exAgene into the CkzI site of pBR322. Plasmid pPO4 is similar to pMG14, but the le.& gene was inactivated by filling in the unique MIUI restriction site. As in Fig. 1, the media are supplemented with 50 &ml ampicillin and with the following concentrations of tetracycline: 0 pg/ml (A), 10 &ml (A), 20 &ml (0) and 30 &ml (W).

after 140 min. Moreover we observed that the magnitude of the effectdependson the uv dose given to the host bacteria (the effect is more pronounced when 40 J me2 is given to the bacteria than when 10 J mm2is given). As mentioned above these measurements are delicate, but repetitive and comparative determinations allow us to suggest that the observedvariation in the expressionof the two genes (bla and tet) carried by pBR322 as a function of the state of induction of the SOS system of the host bacteria might be related 1234 RG. 3. Amount of extractable plasmid DNA, asa func- to a variation in the copy number of the tion of growth temperature. The lanes correspond to lexA plasmid. ind- strain DMl180 [pBR322] grown at 30°C (1) and This result may explain the apparent control 42°C (2) and to strain recA441 GC3217 [pBR322] grown of the uvrC gene by the LexA repressor (Van at 30°C (3) and 42°C (4). All cultures were grown to about Sims et al. (1984)). When the galKgene cloned the same OD,, . The small differenceswere corrected for by loading slightly different volumes of the extracted plas- on a multicopy plasmid was placed under the control of the upstream regulatory sequences mid solutions.

167

SHORT COMMUNICATIONS

0. “t .-([I 0.5. ‘; CI s 0.4., a In 0 7 \

0.3.

2 n

0.2

E

0.1.

30

120

90

60

time after UV-irradiation

130

(mn)

FIG. 4. Relative quantities of supercoiled plasmid pBR322 extracted from AB 1157 cultures at different times after uv irradiation. The growth temperature was 37°C. The uv irradiation was performed in luria broth using 0 (A), 10 (0), and 40 (*) J m-*. Under these conditions the cell survival was 100 and 85% when the bacteria were given 10 and 40 J rnm2,respectively.

of uvrC, an increase in galactokinase activity after induction of the SOS system of the host bacteria was measured. However, no specific complex between LexA and its putative binding site could be observed in vitro (GrangerSchnarr et al., 1986). We suggestthat the effect observed by Van Sluis et al. results from an increased plasmid copy number consecutive to the induction of the SOSsystem of the bacteria. In conclusion, our results show that the expression of the tetracycline and ampicillin genes of pBR322 are modulated by the host SOS system, and that this effect is probably due to an increasedplasmid copy number after induction of the SOS response.The way that one of the constituents of the SOS network might interfere with plasmid replication remains to be elucidated.

REFERENCES BERTRAND-BURGCRAF, E., HURSTEL, S., DAUNE, M., AND SCHNARR,M. (1987). Promoter properties and negative regulation of the uvrA gene by the LexA repressor and its amino-terminal DNA binding domain. J. Mol. Biol. 193, 293-302. DAVIDSON,J. ( 1984).Mechanism of control of DNA rep lication and incompatibility in ColE l-type plasmids. Gene 28, 1- 15. FRIED,M., AND CROTHERS,D. M. (198 1). Equilibria and kinetics of lac repressor-operator interactions by polyacrylamide gel electrophoresis. Nucleic Acids Res. 9, 6505-6525.

GEORGE,J., CASTELLAZZI,M., AND BUTTON,G. (1975). Prophage induction and cell division in E. coli. Mutations sfiA and sfiB restore division in tif and lon strains and permit the expression of mutator properties of tif. Mol. Gen. Genet. 140, 309-332. GRANGER-SCHNARR, M., SCHNARR,M., AND VAN SLUIS, C. A. (1986). In vitro study of the interaction of the LexA repressorand UVIC protein with a uvrC regulatory region. FEBS Lett. 198, 6 l-65. HURSTEL,S., GRANGER-SCHNARR, M., DAUNE,M., AND ACKNOWLEDGMENTS SCHNARR,M. (1986). In vivo binding of LexA repressor We thank Dr. J. Little for the gift of plasmid pJL45, to DNA: Evidence for the involvement of the aminoDr. N. Grant for his help in light microscopy,and Professor terminal domain. EMBO J. 5,793-799. G. Dirheimer and Drs. R. Fuchs and M. Takahashi for HURSTEL,S., GRANGER-SCHNARR, M., AND SCHNARR, stimulating discussions. M. (1988). Contacts between the LexA repressoror its

168

SHORT COMMUNICATIONS

DNA binding domain and the backbone of the recA operator DNA. EMBO J 7,269-215. LINDSTROM,E. B., AND NORDSTROM,K. (1972). Automated method for determination of penicillins, cephalosporins and penicillinases Antimicrob. Agents Chemother. 1, 100-106. LITTLE, J. W. (1983). The SOSsystem: Control of its state by the level of RecA protease. J. Mol. Biol. 167, 791808. LITTLE, J. W., AND HILL, S. A. (1985). Deletions within a hinge region of a specific DNA-binding protein. Proc.

SCHNARR,M., POU~ET,J., GRANGER-SCHNARR, M., AND DAUNE, M. (1985). Large-scalepurification, oligomerization equilibria, and specific interaction of the LexA repressorof E. coli. Biochemistry 24, 28 12-28 18. SLILATY, S. N., AND LITTLE, J. W. (1987). Lysine-156 and Serine-119are required for LexA repressorcleavage: A possible mechanisme. Proc. Natl. Acad. Sci. USA 84, 3987-399 1. TOMIZAWA,J., AND ITOH, T. (198 I). Plasmid ColEl incompatibility determined by interaction of RNA I with primer transcript. Proc. Natl. Acad. Sci. USA 78,6096Natl. Acad. Sci. USA 82,2301-2305. 6100. LITTLE, J. W., AND MOUNT, D. W. (1982). The SOS regVAN SLUIS,C. A., MOOLENAAR,G. F., ANDBACKENDORF, ulatory system of E. coli. Cell 29, 539-558. LITTLE, J. W., MOUNT, D. W., AND YANISCH-PERRON, C. (1984). Regulation of the UVICgene of E. coli K12: Localization and characterization of a damage-inducible C. R. (198 1). Purified LexA protein is a repressorof the promoter. EMBO J. 2, 23 13-23 18. recA and 1exA genes. Proc. Natl. Acad. Sci. USA 78, 4199-4203. WALKER,G. (1984). Mutagenesis and inducible response MARKHAM,B. E., HARPER,J. F., MOUNT,D. W., SANCAR, to deoxyribonucleic acid damage in E. coli. Microbial. G. B., SANCAR,A., RUPP,W. D., KENYON,G. J., AND Rev. 48, 60-93. WALKER, G. C. (1984). Analysis of mRNA synthesis WITKIN, E. M., AND KOGOMA(1984). Involvement of the following induction of the E. coli SOS system. J. Mol. activated form of RecA protein in SOSmutagenesisand Biol. 178, 237-248.

MOUNT, D. W. (1977). A mutant of E. coli showing constitutive expressionof the lysogenic induction and errorprone DNA repair pathways.Proc. Natl. Acad. Sci. USA 74.300-304.

stable DNA replication in E. co/i. Proc. Natl. Acad. Sci. USA 81,7539-7543. Communicated by Douglas E. Berg