1997 Oxford University Press
Nucleic Acids Research, 1997, Vol. 25, No. 21
Escherichia coli OxyR modulation of bacteriophage Mu mom expression in dam+ cells can be attributed to its ability to bind hemimethylated Pmom promoter DNA Stanley Hattman* and Weiyong Sun Department of Biology, University of Rochester, Rochester, NY 14627, USA Received June 6, 1997; Revised and Accepted September 5, 1997
ABSTRACT Transcription of the bacteriophage Mu mom operon is strongly repressed by the host OxyR protein in dam– but not dam+ cells. In this work we show that the extent of mom modification is sensitive to the relative levels of the Dam and OxyR proteins and OxyR appears to modulate the level of mom expression even in dam+ cells. In vitro studies demonstrated that OxyR is capable of binding hemimethylated Pmom , although its affinity is reduced slightly compared with unmethylated DNA. Thus, OxyR modulation of mom expression in dam+ cells can be attributed to its ability to bind hemimethylated Pmom DNA, the product of DNA replication. INTRODUCTION The bacteriophage Mu mom gene encodes an unusual DNA modification function (1–4) that protects the viral DNA against a variety of host controlled restriction systems (5). Transcription initiation from Pmom by Escherichia coli σ70 RNA polymerase requires transactivation by a phage-encoded protein, C (6–8), which is also required for transcription of three other late operons (9,10). Surprisingly, activity of the host DNA-[N6-adenine] methyltransferase (Dam), which methylates adenine in the sequence GATC (11,12) is also required for mom expression (1,3,13–15), since transcription of the bacteriophage Mu mom operon is strongly repressed in dam– hosts (13,14). The role of Dam is to methylate three closely spaced sites located just upstream of the C recognition site (3,16,17). This methylation prevents binding of another host protein, OxyR, the repressor of mom transcription (18). OxyR, a member of the LysR family, is a redox-sensitive transcriptional regulator that under oxidative stress induces expression of a set of antioxidant defense genes (19). Consistent with Dam-dependent expression of mom is the fact that OxyR binds to unmethylated, but not fully methylated, Pmom DNA in vitro (18,20). Since most of the naturally occurring enteric strains that Mu infects are dam+, it has been something of an enigma as to why OxyR should play a regulatory role in dam– hosts. Another perplexing observation has been that the degree of mom modification of Mu progeny phage DNA varies as a function
of the mode of virus propagation, namely phage grown via an exogenous infection and lytic cycle is modified to a much lower extent than growth following thermal induction of a Mu lysogen (15). This lower degree of mom modification can be measured either directly by DNA base analysis or indirectly by biological assay of the efficiency of plating (EOP) on restricting versus non-restricting hosts; for the latter, the phage P1 system has been particularly useful, since mom modification protects against P1 restriction. For example, the EOP on K12(P1) versus K12 is 0.01–0.02 for Mu grown lytically in dam+ cells; this value is 0.4–0.7 for phage grown following prophage induction (1). In contrast, phage grown in dam– cells by either method have an EOP of only 10–4–10–3. It is also known that in general bacteriophage virion DNA methylation is usually incomplete, presumably due to packaging of the phage DNA prior to saturation of all potential modification sites, whereas intracellular DNAs are generally fully modified (21). Thus, it was suspected that the difference in mom modification levels for Mu grown via the lytic cycle versus prophage induction could be attributed to some difference in Pmom methylation during these two growth cycles. For example, the temperature shift-up to induce the prophage might affect the methylation pattern by either increasing the rate of DNA methylation or access to sites. Support for the notion of an effect of DNA methylation came from an observation that mom modification of lytically grown Mu increased to the same level as following prophage induction if the host harbored a plasmid expressing a cloned dam+ gene (3). This suggested that higher intracellular levels of Dam increased the extent of mom expression. Interestingly, the requirement for Dam was obviated when the three methylation target GATC sites were deleted or mutated (3,16,17). This led to the prediction that in dam– cells a host-encoded protein binds in this region and represses mom transcription (22). This hypothesis was confirmed when it was shown that the E.coli OxyR protein is the repressor; e.g. in dam– cells an oxyR– mutation suppressed the Dam requirement for mom expression (18). The first hint that OxyR might also influence mom expression in dam+ cells came from the following observations. The EOP of Mu progeny phage produced after prophage induction in a dam+oxyR+ host was 0.5; it was increased to 1.0 if induction was in a dam+oxyR– host (18). In contrast, the EOP was lowered to 8 × 10–3 if the cells harbored a plasmid expressing a cloned oxyR+
*To whom correspondence should be addressed. Tel: +1 716 275 8046; Fax: +1 716 275 2070; Email: [email protected]
4386 Nucleic Acids Research, 1997, Vol. 25, No. 21 gene. Because of the high EOP baseline for induced prophage, a 2-fold increase was not a convincing demonstration that OxyR reduces mom expression in dam+ cells. We have used a more sensitive biological assay involving phage grown by exogenous infection to demonstrate OxyR modulation of mom gene expression in dam+ cells and we also show directly that OxyR is able to bind hemimethylated Pmom DNA in vitro. Thus, OxyR-mediated mom regulation operative in dam+ cells can be attributed to the ability of the protein to bind hemimethylated Pmom DNA, the product of DNA replication. MATERIALS AND METHODS Strains and plasmids Both oxyR– and dam– derivatives of E.coli strain JM83 were prepared by P1vir transduction (20) from appropriate oxyR::kan and dam::Tn9 CmR donors respectively. Kanamycin- and chloramphenicol-resistant transductant clones were picked and screened for the OxyR– phenotype by growth sensitivity in the presence of 1 mM H2O2. The Dam– phenotype was screened by sensitivity to cleavage of cellular DNA by the restriction endonuclease DpnII (New England Biolabs, NEB). Production of Pmom DNA in different methylation states Plasmid pLW4 contains the Pmom promoter region on an ∼220 bp BamHI–EcoRI fragment cloned into pNM480 (23,24). Plasmid pLW4 DNA was isolated from E.coli strain JM83 dam– (20) and digested with BamHI and EcoRI and the Pmom fragment purified from a low melting point (LMP) agarose gel. The DNA was digested with DpnII, dephosphorylated with calf intestinal alkaline phosphatase (Gibco BRL) and end-labeled with T4 polynucleotide kinase and [γ-32P]ATP. Plasmid pLW6 contains the Pmom promoter region on the same ∼220 bp BamHI–EcoRI fragment, but cloned into plasmid pGC2. Plasmid pLW6 DNA was isolated from a dam+ host and methylated to saturation with EcoDam methyltransferase (NEB) in vitro. To generate hemimethylated DNA for restriction nuclease digestion (Fig. 1), 0.1 µg (18 fmol) pLW4 dam– DNA was linearized with BamHI and 32P-end-labeled by filling in with Sequenase (United States Biochemical, USB). Following deproteinization with phenol/chloroform, the DNA was mixed with a 100-fold excess (1.8 pmol) of unlabeled, BamHI-digested pLW6 dam+ DNA and digested with EcoRI. The BamHI–EcoRI Pmom fragments (labeled on the top strand from pLW4) were isolated from an LMP agarose gel and dissolved in 8 µl TE buffer, pH 7.9, in a 500 µl microfuge tube. After applying a top layer of 10 µl mineral oil (ultrapure for use in PCR; USB) the DNA was denatured by boiling for 10 min, vortexed briefly and put on ice. Two microliters of 5× Sequenase buffer (USB; 5× = 200 mM Tris–HCl, pH 7.5, 100 mM MgCl2 and 250 mM NaCl) were added and the DNA annealed by slow cooling from 65C to between 35 and 30C. The same method was used to generate Pmom hybrid DNA with 32P-labeling on the bottom strand, except that the plasmids were first linearized with EcoRI and then digested with BamHI. The hemimethylated Pmom DNA used in binding studies (Fig. 2) was produced by denaturation and annealing of labeled (in one strand) fragments from fully methylated pLW6 with a 100-fold excess of unlabeled, unmethylated, but otherwise
Figure 1. Assay of Dam methylation at the three GATC sites in the mom promoter region. Pmom was on an ∼220 bp EcoRI–BamHI fragment described previously (23,24). Lane 1, plasmid pLW4 DNA was isolated from JM83 dam– cells, digested with DpnII and 32P-end-labeled with T4 kinase; lanes 2–7, heat-denatured and annealed DNA samples, 32P-end-labeled on the unmethylated strand, were digested as indicated. The methylation states of the labeled DNA duplexes are denoted by two symbols; for example –/+ indicates that the DNA was unmethylated on the top strand and methylated on the bottom strand.
identical fragments. Unmethylated Pmom DNA, labeled on one strand, was mixed with a 100-fold excess of unlabeled fragments; half the sample was taken for denaturation and annealing. By using labeled methylated rather than unmethylated fragments to prepare the hemimethylated duplexes, we avoided formation of retarded complexes derived from undenatured labeled fragments. Hemimethylation was monitored by resistance to cleavage with DpnII, which cleaves only unmethylated GATC. As a control for DNA digestibility after heat denaturation and annealing, digestion was also performed with Sau3AI, which is indifferent to adenine methylation in GATC. All samples were then electrophoresed on a non-denaturing 16% polyacrylamide gel. In vitro binding of OxyR to Pmom DNA Intracellular OxyR is in a reduced conformation during growth in the absence of oxidative stress (25), however, in vitro it is in an altered conformation and requires high concentrations of dithiothreitol (DTT) to avoid oxidation. Previously we found that a mutant form, OxyR-C199S, which appears to be ‘locked’ in a reduced conformation (26), is able to repress mom expression both in vivo and in vitro (20), therefore, we used purified OxyR-C199S to study in vitro binding to variously methylated forms of Pmom . Labeled Pmom DNA (300 c.p.m.) was incubated with serial dilutions of purified OxyR-C199S (20) at room temperature (22C) for 45 min in a total of 22 µl binding buffer [25 mM Tris–HCl, pH 7.5, 50 mM KCl, 5 mM MgCl2, 5% (v/v) glycerol, 50 µg/ml double-stranded poly(dI–dC), 50 µg/ml BSA and 1 mM DTT]. As a control, unmethylated DNA without denaturation and annealing treatment was analyzed in a separate binding assay. Samples were electrophoresed on a non-denaturing 6% polyacrylamide gel in low ionic strength electrophoresis
4387 Nucleic Acids Acids Research, Research,1994, 1997,Vol. Vol.22, 25,No. No.121 Nucleic
Figure 2. Binding of OxyR-C199S to Pmom DNA in vitro. DNA samples, with or without heat denaturation and annealing as indicated, were incubated with serially diluted OxyR-C199S and electrophoresed on a non-denaturing 6% polyacrylamide gel. See Materials and Methods for details on labeling and production of the hemimethylated species.
buffer (27) at 5 V/cm for 5 h at 22C. The gel was dried and autoradiographed. RESULTS AND DISCUSSION Modulation of mom expression in dam+ cells In order to have a more sensitive assay for OxyR modulation of mom expression in dam+ cells, we took advantage of the fact that a relatively low level of mom modification is produced during Mu lytic infection (1). Thus, phage Mucts62 was grown lytically by plating it on various host strains and the progeny phage from pooled plaques were serially diluted and plated on K704 (P1) and K704. As shown in Table 1, lytically grown phage from both dam+oxyR– and dam–oxyR– hosts were efficiently modified; i.e. they had a relative EOP within the range 0.4–0.6, which is a value typical for progeny from prophage induction in dam+ cells. Thus, Mu grown in a dam+oxyR+ or dam–oxyR+ host had a 40- or 600-fold lower EOP than Mu grown in the corresponding oxyR– variant. Therefore, loss of OxyR activity by mutation, which abolishes repression in dam– cells, also elevated mom modification of Mu phage grown lytically in dam+ cells (about the same high level EOP was observed in both dam– and dam+ cells deficient in OxyR). It should be noted here that phage λcI857 grown lytically in each of these four strains exhibited a low EOP (10–4) on K704(P1); this rules out the possibility that P1 modification genes were left behind during the P1vir transductions creating the mutant strains (data not shown). Table 1. Relative EOPs of phage Mu grown lytically on various host strains Phage grown on hosta
dam + oxyR +
dam + oxyR –
dam – oxyR –
aPhage Mucts62 was plated on various derivatives of E.coli K12 strain JM83 (see text). bRelative EOP is defined as the titer on K704(P1)/titer on K704.
As noted in the Introduction, increasing the intracellular level of OxyR decreased mom modification of induced prophage in dam+ cells by 50- to 100-fold (18). We have repeated those experiments and found at least a 103-fold inhibition of mom expression (unpublished observation); this may be due to differences in the hosts and/or OxyR plasmids used. The results are consistent with the notion that OxyR modulates mom expression in dam+ cells. Taken together with the fact that mom modification is enhanced by expression of a cloned dam+ gene (3), it appears that the intracellular Dam MTase/OxyR ratio is important in determining the degree of mom expression. However, it is not clear whether an increased OxyR level decreases dam expression or whether OxyR is actually capable of binding hemimethylated Pmom DNA. The former possibility was ruled out by (methylation-sensitive) restriction nuclease analysis of DNA from the plasmid harboring a cloned, expressed oxyR gene (data not shown). Experiments to test OxyR binding to hemimethylated Pmom DNA are described below. Monitoring hemimethylation of artificial Pmom duplexes The role of Dam methylation in regulating mom transcription suggested that binding of OxyR to Pmom depends on the presence and methylation state of the three GATC sites within an upstream OxyR recognition sequence (17). This has indeed been shown, as (oxidized) OxyR can footprint dam– but not dam+ Pmom DNA in vitro (18). In our studies the Pmom fragment isolated from dam– cells could be completely digested by DpnII (Fig. 1, lane 1), indicating that all three GATC sites were unmethylated (28). When a Pmom -containing high copy number plasmid (pLW6) was grown in dam+ cells methylation of the GATC sites was not quite complete (data not shown). Therefore, dam+ Pmom DNA was methylated to completion in vitro with the GATC-specific EcoDam methyltransferase. Hemimethylated Pmom DNA duplexes were then prepared as described in Materials and Methods. These duplexes were resistant to DpnII digestion (lanes 3 and 6), showing that the parental strands derived from pLW6 had been fully methylated. An extremely weak band was also observed in both lanes; these are attributed to the products of self-annealing by the unmethylated DNA (which should occur with 1% probability). Digestion with Sau3AI was complete and no unexpected bands
4388 Nucleic Acids Research, 1997, Vol. 25, No. 21 were detected (lanes 4 and 7), indicating that the DNA had been completely denatured and annealed according to their complementarity and that they were not noticeably damaged by the heat treatment. OxyR binds to hemimethylated DNA in vitro The results of in vitro binding of OxyR-C199S to Pmom DNA, analyzed by gel retardation, are shown in Figure 2; it should be noted here that binding of the (oxidized) wild-type OxyR gave virtually identical results (data not shown). Similar binding profiles were obtained with unmethylated DNA before and after heat denaturation/annealing (Fig. 2A and B), showing further that the strategy to produce hemimethylated DNA yielded intact strands in the annealed duplexes. Binding of hemimethylated DNA [with methylation on the top (+/–) or the bottom (–/+) strand] was barely detectable at 0.5 nM and definitely present at 1.0 nM OxyR-C199S (Fig. 2C and D, lanes 3 and 4). In contrast, at the lower concentration 37% of unmethylated DNA was bound (Fig. 2B, lane 3), indicating that OxyR-C199S–Pmom binding was reduced by virtue of hemimethylation on either strand (+/– DNA gave slightly better binding than –/+ DNA). However, substantial binding with both hemimethylated DNAs (>70%) was achieved when the protein concentration was raised by only 4-fold to 2.1 nM (Fig. 2C and D, lanes 5); comparable levels of unmethylated DNA binding required roughly half the concentration of OxyR-C199S (Fig. 2B, lane 4). Separate analyses showed that OxyR-C199S binding was very weak when both DNA strands were methylated (data not shown). These results clearly demonstrate that in vitro OxyR can bind hemimethylated Pmom DNA with only a slight reduction in affinity compared with unmethylated DNA. This suggests that in vivo binding of OxyR to hemimethylated Pmom DNA (generated by DNA replication) is responsible for the observed modulation of mom expression in dam+ cells. Moreover, the intracellular Dam MTase/OxyR ratio seems to be important in determining the degree of mom expression. This competition between DNA methylation and DNA binding by a regulatory protein may be a model for the control of methylation-sensitive genes in higher eukaryotes. We also tested the possibility that the procedure of thermal induction of Mu prophage at 42C might inactivate OxyR or increase the Dam methylation rate (and lead to GATC methylation saturation in Pmom ); either process would favor expression of the mom modification. However, Mu grown lytically at 42C in oxyR+dam+ cells showed the same EOP on K704(P1) as phage grown lytically at 37C (data not shown). This ruled out a
convenient explanation for the difference between lytic infection and prophage induction in the extent of mom modification of progeny phage. Nevertheless, it would be worthwhile to probe the methylation status of the Pmom GATC sites in prophage (before and after thermal induction) and lytically grown phage DNAs to determine whether there is, in fact, any difference between them. ACKNOWLEDGEMENT This work was supported by a grant from the National Institutes of Health no. GM29227 to S.H. REFERENCES 1 Toussaint,A. (1976) Virology, 70, 17–27. 2 Hattman,S. (1979) J. Virol., 32, 468–475. 3 Hattman,S., Goradia,M., Monaghan,C. and Bukhari,A.I. (1982) Cold Spring Harbor Symp. Quant. Biol., 47, 647–653. 4 Swinton,D., Hattman,S., Crain,P.F., Cheng,C.-S., Smith,D.L. and McCloskey,J.A. (1983) Proc. Natl. Acad. Sci. USA, 80, 7400–7404. 5 Kahmann,R. (1984) Curr. Topics Microbiol. Immunol., 108, 29–47. 6 Bölker,M., Wulczyn,F.G. and Kahmann,R. (1989) J. Bacteriol., 171, 2019–2027. 7 Hattman,S., Ives,J., Margolin,W. and Howe,M.M. (1985) Gene, 39, 71–76. 8 Margolin,W. and Howe,M.M. (1986) Nucleic Acids Res., 14, 4881–4897. 9 Marrs,C.F. (1982) PhD Thesis, University of Wisconsin, Madison, WI. 10 van Meeteren,A. (1980) PhD Thesis, State University of Leiden, The Netherlands. 11 Hattman,S., Brooks,J.E. and Masurekar,M. (1978) J. Mol. Biol., 126, 367–380. 12 Marinus,M.G. and Morris,N.R. (1973) J. Bacteriol., 114, 1143–1150. 13 Hattman,S. (1982) Proc. Natl. Acad. Sci. USA, 79, 5518–5521. 14 Plasterk,R.H.A., Vrieling,H. and van de Putte,P. (1983) Nature, 301, 344–347. 15 Toussaint,A. (1977) J. Virol., 23, 825–826. 16 Kahmann,R. (1983) Cold Spring Harbor Symp. Quant. Biol., 47, 639–646. 17 Seiler,A., Blöcker,H., Frank,R. and Kahmann,R. (1986) EMBO J., 5, 2719–2728. 18 Bölker,M. and Kahmann,R. (1989) EMBO J., 8, 2403–2410. 19 Kullik,I., Stevens,J., Toledano,M.B. and Storz,G. (1995) J. Bacteriol., 177, 1285–1291. 20 Sun,W. and Hattman,S. (1996) Nucleic Acids Res., 24, 4042–4049. 21 Hattman,S. (1977) J. Bacteriol., 129, 1330–1334. 22 Hattman,S. and Ives,J. (1984) Gene, 29, 185–198. 23 Balke,V., Nagaraja,V., Gindlesperger,T. and Hattman,S. (1992) Nucleic Acids Res., 20, 2777–2784. 24 Gindlesperger,T.L. and Hattman,S. (1994) J. Bacteriol., 176, 2885–2891. 25 Storz,G., Tartaglia,L.A. and Ames,B.N. (1990) Science, 248, 189–194. 26 Toledano,M.B., Kullik,I., Trinh,F., Baird,P.T., Schneider,T.D. and Storz,G. (1994) Cell, 78, 897–909. 27 Ausubel,F.M., Brent,R., Kingston,R.E., Moore,D. D., Seidman,J.G., Smith,J.A. and Struhl,K. (1992) Short Protocols in Molecular Biology, 2nd Edn. John Wiley & Sons, New York, NY, pp 12.5–12.7. 28 Lacks,S. and Greenberg,B. (1977) J. Mol. Biol., 114, 153–168.