Acts on the Transcription Process

Proceedings of the National Academy of Sciences Vol. 68, No. 1, pp. 215-218, January 1971

An Adenosine 3':5'-Cyclic Monophosphate-Binding Protein That Acts on the Transcription Process

L. ERON*, R. ARDITTI*, G. ZUBAYt, S. CONNAWAY*, AND J. R. BECKWITH* * Department of Bacteriology and Immunology, Harvard Medical School, Boston, Massachusetts 02115; and t Department of Biology, Columbia University, New York, N.Y. Communicated by Boris Magasanik, November 13, 1970

ABSTRACT An assay system for the in vitro transcrip- on the origin of the phages as predicted by the Campbell tion of the lac operon is described. A protein factor (CAP) model. This information indicates that there are no bacterial and cyclic AMP, which are essential for lac expression in vivo, also stimulate lac transcription in vitro. genes in common between the two phages other than the lac genes. However, since the X and q80 chromosomes do recom- Until recently, the study of the expression and regulation of bine with each other (7, 10, 11), there is almost certainly some bacterial genes has depended on physiological and genetic genetic homology between the two phage chromosomes. To studies. In the last few years, the detection and purification of test the effect of this homology on RNA-DNA hybridization regulatory proteins has opened up the possibility for studying we have transcribed 080plac and Xplac DNAs with RNA regulatory effects in vitro [e.g., lac (1), X (2), T4 (3), T7 (4)]. polymerase and hybridized the resulting RNA to the separate In the case of bacteriophage genomes, studies on in vitro strands of XDNA and 08ODNA, respectively. The results transcription have concentrated on regions of the phage ge- show that there is only very little homology between the RNA nomes (e.g., early and late), for the most part, rather than spe- made from one phage DNA and the DNA of the other (Table cific genes. In bacterial systems, however, one must be able to 1). This homology never amounts to more than 1% of the detect the RNA transcription of a relatively very limited total counts incorporated into the trichloroacetic acid number of genes, usually those comprising an operon. This (TCA)-precipitable material. We expect, therefore, that means that ideally for in vitro studies, an assay system for essentially the only RNA made from, for instance, 080plac mRNA must be developed that can pick up specifically the which will hybridize to Xplac is lac RNA itself. mRNA species from a single operon. Here, we describe such a For most of our experiments, we have used 080plac DNA system for the lac operon. as a template for transcription, and the separated strands of Among many transducing phages isolated that carry the Xplac for hybridization. In vivo the lac genes (16), and pre- lac operon, two have been described which have no other genes sumably all other genes, are transcribed into a mRNA comple- in common than the lac genes. These two phages, Xplac 5 and mentary to only one of the two DNA strands. We have shown 480plac 1, have been used, in fact, to isolate pure lac DNA (5). Thus, an assay system for specifically detecting lac mRNA Aplac, lac exists, either through the use of the pure lac DNA or through A ____ J y z b2 attA int N immA R use of the two phage DNA molecules, one as a template for Aplac, RNA transcription and the other for RNA-DNA hybridiza- Aplac, I_ tion. In this paper we describe the use of the hybridization procedure for studying in vitro transcription of the lac operon. These 080plac, studies have allowed us to show that a protein factor (CAP) and cyclic AMP (cAMP), which are essential for lac expression A ---- J att80 a i N imm80 R in vivo, act on the transcription process. The study of regula- 80pl zay tory effects in vitro, specifically effects of lac repressor and iac ,80plac1,I promoter mutations, suggest that additional components FIG. 1. Direction of transcription of genes on lac transduc- may be necessary for accurate initiation of lac transcription. ing phages. The arrows indicate direction of transcription and The results also indicate some of the problems in dealing with are placed closest to that strand from which the RNA is trans- in vitro transcription of a DNA preparation that contains cribed. In vitro, the b2 region of X is transcribed from both strands genes other than the genes of interest. (6). A brief description of the origins of the two phages is presented elsewhere (5). The 480pilc behaves like an int- phage, METHODS AND RESULTS suggesting deletion of some genes to the left of N. Although in The probable structures of Xplac 5 and 080plac 1 genomes are the 80plac all y-distal bacterial genes were removed in the con- shown in Fig. 1. These structures are based on genetic data and struction of the phage, there are probably still bacterial genes adjacent to the i gene. We do not know how much of the b2 Abbreviation: CAP, a protein factor required for the synthesis region is intact in Xplac5. For maps of X describing the markers, of fl-galactosidase in vitro whose activity depends on cyclic AMP see refs. 7-9. For convenience, the 4)80 marker notation is identical (cAMP). with that of X. The iinm8O notation refers to o80 immunity and No reprints of this paper will be available in the USA. imm X to X immunity region. 215 Downloaded by guest on September 30, 2021 216 Biochemistry: Eron et al. Proc. Nat. Acad. Sci. USA

that with q80plac, it is the 080plaCH strand and with Xplac, The failure to observe high levels of correct transcription of the XplaCL strand that is complementary to lac mRNA made the lac operon was not unexpected. Evidence from genetic and in vivo (5). One of the criteria, then, for determining whether physiological studies have indicated that for correct tran- the lac genes are being transcribed accurately in vitro is scription of the lac operon, in addition to RNA polymerase, whether the RNA made from 080plac is complementary to cAMP and a protein factor may be necessary (18-23). We the XplaCL strand only; i.e., asymmetric transcription. The have described the partial purification of a protein factor data in line 2 of Table 2 show that transcription of lac genes in (CAP) (22) which is necessary for the synthesis of fl-galactosi- 480plac DNA occurs from both strands, predominantly from dase in a crude in vitro system (15). The activity of this pro- the incorrect strand! This result is not unpredicted if one tein depends on the presence of cAMP. Emmer et al. (23) have considers the structure of the phage genome as outlined in described a similar protein. Our protein preparation binds Fig. 1. Transcription from the 480 "N" gene should be ini- cAMP with a constant of approximately 3 X 105 M. tiated in this system (17) and proceed onto the incorrect When CAP and cyclic AMP are included in the transcription strand of the lac region. [We show elsewhere that an RNA reaction mixture, up to a 6-fold stimulation of transcription termination factor, p, which eliminates transcription of the from the correct strand of the lac region is observed (Table 2, region to the left of N (17), does the same to the incorrect lines 2-5). At the concentrations used, the reaction is still transcription of the lac genes seen here (15).] linear with CAP concentration. At the same time as transcription from the correct strand of lac is stimulated, transcription from the opposite strand is depressed. This depres- TABLE 1. Complementarity of RNA transcribed from one lac sion may be due to competition between RNA polymerase transducing phuge DNA uith the heterologous wild molecules from type DNA transcribing opposite strands. phase The expected stimulation of lac transcription seen with and CAP is not due to a Cpm hybridized to cyclic-AMP generalized stimulation of transcription (total counts incorporated are the same) nor Template XL XH 48OL 480H probably to stimulation of transcription of parts of the 480 of the strand. The 080plac 147 (0.4) 165 (0.4) 10,000 (25) 17,400 (43) portions 480plaCH ratio of counts tran- Xplac ...... 200 (0.5) 305 (0.7) scribed from 080plac which hybridized to 480L and 480H strands was identical whether or not CAP was present (Table The figures in parentheses are the % of total input counts 3). (40,000 cpm). The RNA polymerase used in these experiments includes Phage lysates were prepared as follows: 106-107 phage particles the a protein, which affects the template specificity of the were plated onto TYE plates (12) with M top agar (12) (diluted enzyme (3). The stimulation of lac transcription seen with 20% with distilled water) containing 2 X 108 bacteria of a lac CAP and cyclic AMP is completely dependent on the presence deletion strain (M182). After overnight incubation at 370C, of a factor in the transcription reaction mixture (Table 2, 1-2 ml of broth was added to each plate and left at room tem- lines 1 and 5). perature for an additional hour. The surface of the plates were The specificity of the effect of CAP and cyclic AMP on lac then scraped and the mixture of phage, bacteria, and agar was transcription suggests that transcription is proceeding in vitro treated with chloroform and centrifuged twice at 6000 rpm for 15 min. The as it does in vivo. A further criterion to establish this is to supernatant containing the phage was concentrated demonstrate in a Spinco ultracentrifuge at 22,000 rpm for 2.5 hr. The pellet expected regulatory effects in vitro. We have taken was left overnight at 4VC, covered with Weigle's buffer (120 mM two approaches to this aspect of the study. NaCl-0.5 mM MgSO4-6 mM Tris (pH 7.2-7.4) and 50 mg/liter First, purified lac repressor, provided by T. Platt and of gelatin). The phage suspension was then brought to a volume W. Gilbert, was included in the reaction mixture. No effect of of 6 ml, which was centrifuged twice for 18 hr at 22,000 rpm in a repressor was observed on lac transcription even when it was CsCl solution of density 1.5. The band of purified phage was included in a preincubation mixture without RNA polymerase collected and DNA was extracted with phenol. 100 plates (Table 4). Furthermore, the kinetics of lac transcription were ordinarily gave about 2 X 1013 phage. The detailed procedures for extracting DNA from the phages, for separation of strands (13), and for hybridization (14) are de- TABLE 2. Transcription of the lac region of 480plac DNA scribed in detail elsewhere (15). Purified RNA polymerase containing a-protein was a gift of Dr. R. Burgess. The reaction Additions Cpm* mixture, usually 100 IAl, contained 20 ;tg/ml of DNA, 0.15 mM hybridized to Ratio nucleotide triphosphates with [3H]ATP Schwarz (1 Ci/mol, a CAP cAMP XplaCL XplaCH XpkzcL/XplacH BioResearch), 10 mM MgCl2, 0.15 M KCl, 0.04 M Tris (pH 7.9), 0.1 mM EDTA, 6 mM fl-mercaptoethanol, 0.5 mg/ml of bovine 1 + + 130 131 ... serum albumin (Sigma, grade I), 0.2% glycerol, 3 juM dithio- 2 + ...... 775 4703 0.16 threitol, and 20 Mg/ml of RNA polymerase. The reaction, per- 3 + ... + 757 4597 0.16 formed for 20 min at 370C, was terminated by the addition of 0.1 4 + + ... 1179 4238 0.27 volume of 1 M sodium acetate (pH 5.2) and was extracted twice 5 + + + 3582 2991 1.2 with equal volumes of phenol at 60'C. Hybridization reaction mixtures of 100 ,l contained [3H] RNA saturated with phenol for The reaction mixture was as in Table 1, except that 0.1 M the extraction procedure and 5 ug/ml of DNA separated strands. KC1, 10-4 M cAMP, and CAP factor at a concentration of 25 Hybridization efficiency was 80% at 4 hr. All experiments in #1/100 Ml (15) were used. Purified RNA polymerase without a this and subsequent tables were done at least twice (with highly was obtained from R. Burgess. CAP is a protein of wkwular reproducible results), except for the experiments done with Xplac weight 45,000, which has been purified 200-fold (22) DNA as template, which were performed only once. * Input counts were 45,000 cpm. Downloaded by guest on September 30, 2021 Vol. 68, 1971 Transcription Control in vitro 217

TABLE 3. Effect of CAP on transcription of ¢80 genes TABLE 5. Effect of lac promoter mutations on lac transcription Cpm* hybridized to Template Cpm hybridized to Ratio ck80plac k80L t8OH 08OL/080H Template XplacL XplacH +CAP, +cAMP 9,800 17,300 0.57 480plac (p+) + CAP 3256 3468 -CAP, +cAMP 10,000 17,400 0.58 480plac (p-m8) + CAP 3354 3500 4080plac (prUV5) + CAP 3054 2895 Conditions are as in Tables 1 and 2 with 0.15 M KCl and 10-3 M 4080pkac (prUV5) - CAP 1050 4281 cAMP. Input counts were 40,000 cpm. The promoter-mutant derivatives of 080plac were prepared as follows: 080plac (p+z+) forms blue plaques when plated on a not altered by the inclusion of repressor (Eron, unpublished lawn of a lac-deletion strain on minimal-casamino acid plates con- results). The absence of repressor effect was the same whether taining 5-bromo-4-chloro-3-indolyl-(3-D-galactoside (see ref. 26). or not p was included (Eron, unpublished results). The re- A plate lysate of 080plac (p+z+) was made on a strain carrying a pressor was shown both before and after its use to be highly z- mutation (NG-545) and white plaques from this lysate on the above plates were sought. These white plaques were made by effective in binding to X,80dlac DNA (Platt and Gilbert, phages of the type q80plac (p+zi). These plaques were purified personal communication). and used to make plate lysates on a strain carrying the lac pro- Secondly, 080plac DNAs carrying two types of promoter moter mutation, L8, and on a strain carrying the lac promoter mutations were used to study lac transcription. One promoter mutation, UV5. From each of these lysates, blue plaques on the mutation, L8, results in a 15-fold reduction of P-galactosidase lac deletion strain were detected and purified. (080plac (p-4) synthesis both in vivo (24) and in a crude in vitro system will still form blue plaques.) These presumed 080plac (p-) and (Zubay, unpublished results). The second promoter mutation, 4,80plac (pr) phages were verified by isolating recombinants UV5, is a second-site (within the promoter) revertant of L8 between the phage-carried lac region and the kac region of a strain which makes approximately 50% of the normal levels of carrying a promoter deletion. The recombinants were assayed but has resulted in a loss of the for f3-galactosidase activity and shown to have the expected j#-galactosidase in vivo require- properties. ment for CAP and cyclic AMP both in vivo (25) and in the Conditions are as described in previous tables with 0.1 M KCl crude in vitro system (Zubay, unpublished results). and 10-4 M cAMP. If lac transcription is being initiated as expected in our * Input counts, 40,000 cpm. system, the use of 080plac (L8) DNA should result in a much lower level of correct lac mRNA synthesis and 4080plac (UY5) DNA should result in high levels of lac mRNA synthesis with reversed our system and used Xplac DNA as a template and no CAP plus cAMP requirement. Contrary to these expecta- the separated strands of 080plac DNA for hybridization. The tions, we can observe no difference in the pattern of lac tran- Xplac, because the i-adjacent region is deleted (Fig. 1), should scription among the three [including 080plac-(p+) DNA] be missing the nearby bacterial genes. Here, the RNA made different templates (Table 5). from the correct strand should hybridize to 480placH. Sur- We have considered mainly two explanations for the failure prisingly, we have found that when we use the Xplac DNA as to see regulatory effects of repressor and promoter mutations a template, there is a high level of correct lac transcription in in our system. First, it is possible that all of the lac transcrip- the absence of cAMP and CAP (Table 6). [This result has been tion we are seeing is due to read-through from a nearby, observed in the crude g-galactosidase-synthesizing system similarly oriented, CAP plus cAMP-dependent operon, and also (Zubay, unpublished results).] While we have eliminated that conditions are still not appropriate for transcription neighboring bacterial genes by the deletion, at the same time initiation at the lac promoter. It is known that there are we have brought the lac genes closer to regions of the X genome nearby genes to which lac can be fused which are probably (bM and the N-connected early genes) which are known to be also dependent on CAP and cAMP (Silverstone, personal transcribed under these conditions (6, 17). Thus, preliminary communication; ref. 27). To test this possibility, we have experiments show that in order to detect effects at the lac promoter we will have to eliminate the read-through tran- TABLE 4. Effect of lac repressor on lao transcription TABLE 6. Transcription of lac genes with Xplac as template Template Cpm* hybridized to 080 plac XplacL XplacH Transcription Cpm* hybridized to - repressor 3316 3450 reaction 8OplacH + repressor 3711 3586 080pktCL repressor added 3869 3414 -CAP 5018 534 5 min before polymerase +CAP 3868 537

Conditions are as described in previous tables with 0.1 M KCl, Conditions as described in previous tables with 0.1 M KCl CAP, and 10-3 M cAMP. Repressor was included at a concentra- and 10-' M cAMP. This experiment was done once, but similar tion of 10 /g/ml (approximately 100 molecules per q80plac DNA results have been observed by R. Bloch (personal communication) molecule). The preincubation mixture included CAP and cAMP, and in the in vitro g-galactosidase-synthesizing system (Zubay, but not RNA polymerase. Repressor was added before CAP. unpublished results). * Input counts, 45,000 cpm. * Input counts, 40,000 cpm. Downloaded by guest on September 30, 2021 218 Biochemistry: Eron et al. Proc. Nat. Acad. Sci. USA

scription from X genes. These experiments are in progress. So 4. Summers, W. C., and R. B. Siegel, Nature, 223, 1111 far, then, we are unable to rule out this first explanation. (1969). 5. Shapiro, J., L. MacHattie, L. Eron, G. Ihler, K. Ippen, A second explanation for the absence of regulatory effects and J. Beckwith, Nature, 224, 768 (1969). in this system is that initiation of transcription is at the lac 6. Roberts, J. W., Nature, 223, 480 (1969). promoter, but that conditions are not correct for observing 7. Szpirer, J., R. Thomas, and C. M. Radding, Virology, 37, regulation. Using a crude system for in vitro transcription, 585 (1969). 8. Herskowitz, I., and E. R. Signer, J. Mol. Biol., 47, 545 Ohshima et al. (28) have obtained effects of repressor on the (1970). level of transcription. 9. Kaiser, A. D., and T. Masuda, J. Mol. Biol., 47, 557 (1970). CONCLUSIONS AND DISCUSSION 10. Signer, E. R., Virology, 22, 650 (1964). The results presented here give evidence on the role of CAP 11. Franklin, N., W. F. Dove, and C. Yanofsky, Biochem. and cAMP in gene transcription. While many of the properties Biophys. Res. Commun., 18, 910 (1965). 12. Gottesman, S., and J. R. Beckwith, J. Mol. Biol., 44, 117 of the system are still not consistent with the expectations (1969). from in vivo studies, it is clear that these factors exert their 13. Hradecna, A., and W. Szybalski, Virology, 32, 633 (1967). effects upon transcription. In vivo, in the absence of either 14. Gillespie, D., and S. Spiegelman, J. Mol. Biol., 12, 829 CAP or cAMP, RNA polymerase is incapable of interacting (1965). properly with the lac promoter and most, if not all, other 15. Arditti, R. R., L. Eron, G. Zubay, G. Tocchini-Valentini, S. Connaway, and J. R. Beckwith, Cold Spring Harbor Symp. promoters of inducible genes. CAP and cyclic AMP do not Quant. Biol., 35, in press. appear to act by replacing the a factor of RNA polymerase. 16. Kumar, S., and W. Szybalski, J. Mol. Biol., 40, 145 CAP may be an additional transcription-initiation factor or (1969). an enzyme that in some way modifies RNA polymerase. 17. Roberts, J., Nature, 224, 1168 (1969). The difficulties in analyzing lac transcription even with the 18. Perlman, R. L., and I. Pastan, J. Biol. Chem., 243, 5420 (1969). specific assay system developed for lac RNA transcription 19. Ullman, A., and J. Monod, FEBS Lett., 2, 57 (1968). emphasize the problems in studying the transcription of 20. Silverstone, A. E., B. Magasanik, W. S. Reznikoff, J. H. bacterial operons. The phenomenon of read-through tran- Mille , and J. R. Beckwith, Nature, 221, 1012 (1970). scription, which can be reduced to some extent by the p factor, 21. Perlman, R. L., B. deCrombrugghe, I. Pastan, Nature, is one of probably several processes which interfere with the 223, 810 (1969). is 22. Zubay, G., D. Schwartz, and J. Beckwith, Proc. Nat. analysis in vitro. One possible way to avoid these problems Acad. Sci. USA, 66, 104 (1970). to use pure lac DNA as a template for in vitro transcription. 23. Emmer, M., B. deCrombrugghe, I. Pastan, and R. Perl- This work was supported by grants from the National Science man, Proc. Nat. Acad. Sci. USA, 66, 480 (1970). Foundation, the American Cancer Society, and the Jane Coffin 24. Ippen, K., J. H. Miller, J. G. Scaife, and J. R. Beckwith, Childs Memorial Fund for Medical Research to J. Beckwith. Nature, 217, 825 (1968). We acknowledge the excellent technical assistance of Ronnie 25. Schwartz, D., and J. R. Beckwith, in The Lac Operon, MacGillivray. eds. D. Zipser and J. Beckwith, Cold Spring Harbor Laboratory on Quantitative Biology, Cold Spring Harbor, N.Y., 1970. 1. Gilbert, W., and B. Muller-Hill, Proc. Nat. Acad. Sci. 26. Abelson, J., M. L. Gefter, L. Barnett, A. Landy, R. L. USA, 58, 2415 (1967). Russel, J. D. Smith, J. Mol. Biol., 47, 15 (1970). 2. Ptashne, M., Nature, 214, 232 (1967). 27. Beckwith, J. R., J. Mol. Biol., 8, 427 (1964). 3. Burgess, R. R., A. A. Travers, J. J. Dunn, and E. K. F. 28. Ohshima, Y., T. Horiuchi, Y. Iida, and T. Kameyama, Bautz, Nature, 221, 43 (1969). Molec. Genetics, 106, 307 (1970). Downloaded by guest on September 30, 2021