Proc. Natl. Acad. Sci. USA Vol. 90, pp. 6081-6085, July 1993 Biochemistry Identification of the activating region of catabolite gene activator protein (CAP): Isolation and characterization of mutants of CAP specifically defective in transcription activation (cyclic AMP receptor protein/RNA polymerase/ac promoter) YUHONG ZHOU, XIAOPING ZHANG, AND RICHARD H. EBRIGHT* Department of Chemistry and Waksman Institute, Rutgers University, New Brunswick, NJ 08855 Communicated by Keith Yamamoto, February 26, 1993 (receivedfor review January 4, 1993)
ABSTRACT We have isolated 21 mutants of catabolite (15). However, that report has been shown to be incorrect gene activator protein (CAP) defective in transcription activa- (16). tion at the lac promoter but not defective in DNA binding. The The objectives of the work in this report were (i) to amino acid substitutions in the mutants map to a single region determine whether CAP contains an activating region and, if of CAP: amino acids 156-162. As assessed in vito, the substi- so, (ii) to determine the location of the activating region. Our tuted CAP variants are nearly completely unable to activate approach to these objectives was to perform random muta- transcription at the lac promoter but bind to DNA with the genesis of the gene encoding CAP, followed by an in vivo same affmity and bend DNA to the same extent as wild-type genetic screen, in order to isolate mutants ofCAP specifically CAP. Our results establish that amino acids 156-162 are defective in transcription activation-i.e., mutants of CAP critical for transcription activation at the lac promoter but not defective in transcription activation but not defective in DNA for DNA binding and DNA bending. In the structure of CAP, binding. We designate such mutants crpPc, where crp denotes amino acids 156-162 are part of a surface loop. We propose the gene encoding CAP, and pc denotes defective in positive that this surface loop makes a direct protein-protein contact control of transcription. with RNA polymerase at the lac promoter. Escherichia coli catabolite gene activator protein (CAP; also MATERIALS AND METHODS referred to as the cAMP receptor protein, CRP) is a struc- Strains and Plasmids. Strains XE64.1 (Acrp39 strA thi) and turally characterized transcription activator protein (re- XE82 (Acrp39 strA thi lacPUVS-OCAP) are described in ref. viewed in refs. 1 and 2). CAP functions by binding, in the 16. Plasmid pYZCRP is described in ref. 17. Plasmid pAG- presence of the allosteric effector cAMP, to specific DNA ICAP was constructed by insertion ofa 39-bp DNA sequence sites located in or upstream of CAP-dependent promoters. containing the consensus DNA site for CAP between the Xba The best-characterized CAP-dependent promoter is the lac I and Sal I sites of plasmid pBend3 (kindly supplied by S. promoter. Adhya; cf. ref. 10). The insertion was made by use of The consensus DNA site for CAP is 5'-AAATGTGAT- site-directed mutagenesis [method of ref. 18, except that CTAGATCACATTT-3' (3). The consensus DNA site is 22 template DNA was prepared in E. coli strain XL1-Blue bp long and exhibits perfect twofold sequence symmetry. The (Stratagene)]. The primer used was 5'-GGGATCCTCTA- three-dimensional structure of CAP has been determined to GAGGGTGCCTAAAATGTGATCTAGATCACATTTAT- 2.5 A resolution by x-ray diffraction analysis (4), and the TGCGTTGTCGACACGCGTAGATC-3'. three-dimensional structure of the CAP-DNA complex has Mutagenesis. Random mutagenesis of the crp gene of been determined to 3.0 A resolution by x-ray diffraction plasmid pYZCRP was performed by use of PCR with Taq analysis (5). CAP is a dimer of two identical subunits, each DNA polymerase as described in ref. 17. The primers used of which is 209 aa long. The CAP-DNA complex is twofold were 5'-CAGTCAGGATGCTACAGTAATACA-3' and 5'- symmetric: one subunit of CAP interacts with one-halfofthe ATCTTCCCCATCGGTGATGTCGGC-3'. DNA site; the other subunit of CAP interacts in a twofold- Purification of CAP and CAP Derivatives. CAP and CAP symmetry-related fashion with the other halfofthe DNA site. derivatives were purified by cAMP affinity chromatography CAP sharply bends DNA in the CAP-DNA complex (6, 7). as described in ref. 19. The fraction of molecules active in The CAP-induced DNA bend is -90° in the CAP-DNA sequence-specific DNA binding (0.3-0.8 for CAP, [Ala- complex in the crystalline state (5) and 80°-180° in the 158]CAP, [Ile-158]CAP, [Leu-159]CAP, and [Arg-159]CAP; CAP-DNA complex in solution (8-12). 0.05 for [Cys-162]CAP) was determined by titration ofa40-bp For a number of prokaryotic and eukaryotic transcription DNA fragment having the consensus DNA site for CAP activator proteins, it has been possible to identify regions of under stoichiometric binding conditions (3). the protein that are essential for transcription activation but Transcription Experiments. Abortive initiation in vitro not essential for DNA binding (reviewed in refs. 13 and 14). transcription experiments were performed as described in Such regions have been designated "activating regions" and ref. 16. Experiments were performed using as templates a have been proposed to participate in direct protein-protein 203-bp EcoRI/EcoRI DNA fragment of plasmid pBR-203-lac interactions with RNA polymerase and/or basal transcrip- and a 314-bp Ava II/EcoRI DNA fragment of bacteriophage tion factors. M13mp2 ICAP. Experiments with CAP and [Ala-158]CAP An early report had suggested that CAP contains an also were performed using as templates a 203-bp EcoRI/ activating region consisting, at least in part, of aa 171 of CAP EcoRI DNA fragment of plasmid pBR-203-lacPUV5 and a
The publication costs of this article were defrayed in part by page charge Abbreviation: CAP, catabolite gene activator protein. payment. This article must therefore be hereby marked "advertisement" *To whom reprint requests should be addressed at: Waksman in accordance with 18 U.S.C. §1734 solely to indicate this fact. Institute, Rutgers University, New Brunswick, NJ 08855-0759. 6081 Downloaded by guest on September 25, 2021 6082 Biochemistry: Zhou et al. Proc. Natl. Acad. Sci. USA 90 (1993) 203-bp EcoRI/EcoRI DNA fragment of plasmid pBR-203- 2 and 3, data for DNA fragments 2 and 4, and data for DNA lacPs (kindly supplied by A. Kolb, Institut Pasteur). fragments 3 and 4). Estimates of apparent DNA bend angles DNA Binding Experiments. Nitrocellulose filter in vitro for the six possible pairwise comparisons of data for DNA DNA binding experiments were performed as described in fragments i and j were in excellent agreement (typically ref. 3 using a 40-bp DNA fragment having the consensus within 5%). Values reported are means ± SE for the six DNA site for CAP [DNA fragment ICAP (3)]. pairwise comparisons of data for DNA fragments i and j. DNA Bending Experiments. Electrophoretic mobility DNA Molecular Modeling. Coordinates for the structure of the bending experiments were performed using four circularly CAP-DNA complex at 3.0 A resolution were obtained from permuted 160-bp DNA fragments having the consensus the Brookhaven Protein Data Bank (accession code 1CGP; DNA site for CAP: an Nru I/Nru I DNA fragment of ref. 5). van der Waals surfaces and solvent accessibilities plasmid pAG-ICAP, an EcoRV/EcoRV DNA fragment were calculated by using the program INSIGHT (Biosym of plasmid pAG-ICAP, an Nhe I/Nhe I DNA fragment of Technologies, San Diego). Amino acid side chains were plasmid pAG-ICAP, and an Mlu I/Mlu I DNA fragment designated as solvent accessible ifin each subunit at least one of plasmid pAG-ICAP. Reaction mixtures (10 ,ul) contained atom was accessible to a probe of radius 1.8 A. 15 nM DNA fragment, 100 nM CAP or CAP derivative, 10 mM Mops-NaOH (pH 7.3), 200 mM NaCl, 0.1 mM dithio- RESULTS threitol, 50 ,ug of bovine serum albumin per ml, and 0.2 mM Isolation of crpPc Mutants. We performed random muta- cAMP. After 1 h at 23°C, 2 ul of 50% (vol/vol) glycerol/ genesis of the crp structural gene by use of PCR with Taq 0.01% xylene cyanol in the same buffer was added, and DNA polymerase (17), and we performed a screen to identify samples were loaded onto 7.5% polyacrylamide gels equili- crpPc mutants. Our screen tested two phenotypes on a single brated in 89 mM Tris borate, pH 8.0/2.0 mM EDTA/0.2 mM indicator agar plate: (i) loss of the ability to activate tran- cAMP. Electrophoresis was performed at 14 V/cm for 2.5 h scription at a CAP-dependent promoter, and (ii) retention of at 23°C. After electrophoresis, gels were immersed in 100 ml the ability to bind to DNA. To test phenotype i, the screen of electrophoresis buffer containing 0.5 ug of ethidium bro- tested the ability to activate transcription at the rbs promoter. mide per ml for 5 min at 23°C and were photographed with UV [The rbs operon consists of the rbsD, rbsA, rbsC, rbsB, and transillumination. rbsK genes and is required for utilization ofribose as a carbon Apparent DNA bend angles were calculated using the source (20). The rbs promoter is a CAP-dependent promoter. following equation (cf. ref. 9): The distance between the center ofthe DNA site for CAP and the transcription start point in the rbs promoter is identical to ui [1 - 2(xi/L)(1 - cosa) + 2(xi/L)2(1 - cosa)]0.5 that in the lac promoter (61.5 bp).] To test phenotype ii, the ,j [1 - 2(xj/L)(1 - cosa) + 2(xj/L)2(1- cosa)]05' screen tested the ability to repress transcription at the lacPUV5-0C" promoter. [The lacPUV5-OCAP promoter is a where A denotes the electrophoretic mobility of the protein- derivative of the activator-independent lacPUV5 promoter DNA complex formed with DNA fragment i, p.j denotes the that has, superimposed, a consensus DNA site for CAP electrophoretic mobility of the protein-DNA complex positioned so that it overlaps the transcription start point (15, formed with DNA fragment j, xi/L denotes the fractional 16). CAP represses transcription at lacPUV5-OCAP, presum- distance between the center ofthe DNA site for CAP and the ably through a steric-exclusion mechanism. Substituted CAP left end of DNA fragment i, xj/L denotes the fractional derivatives that retain the ability to bind to DNA likewise distance between the center of the DNA site for CAP and the repress transcription at lacPUV5-0CAP.j left end of DNA fragmentj, and a denotes the apparent DNA The screen utilized Acrp rbs+ lacPUVS-OCAP strain XE82. bend angle. Use of four DNA fragments permitted six pair- In this strain, ribose utilization is under control of the rbs wise comparisons ofdata for DNA fragments i andj (i.e., data promoter, and lactose utilization is under control of the for DNA fragments 1 and 2, data for DNA fragments 1 and lacPUVS-OC" promoter. Plasmids containing mutagenized 3, data for DNA fragments 1 and 4, data for DNA fragments crp structural gene were introduced by transformation into Table 1. Sequences and phenotypes of the 21 crpPc mutants aa Codon No. of Activation, Repression, Activation/ substitution substitution isolates % % repression Ala-156 Asp GCT GAT 1* 6.0 110 0.055 Thr-158 Ala ACT GCT 8t 23 120 0.19 Thr-158 - Ile ACT ATT Pt 50 120 0.43 His-159 - Leu CAC CTC 1 17 140 0.13 His-159 Arg CAC CGC 1§ 45 120 0.38 Gly-162 Cys GGT TGT 2 9.2 120 0.077 Gly-162 - Asp GGT - GAT 2 1.4 110 0.013 Gly-162 - Ser GGT - AGT 5 16 120 0.14 Activation and repression were measured in vivo as described in ref. 16 and are expressed as percentages of activation and repression by wild-type CAP. *The isolate contained a second amino acid substitution: Ile-112 -* Val (ATT -* GTT). Data for activation and repression are for a single-substitution mutant constructed by site-directed mutagen- esis. tThree isolates contained second amino acid substitutions: Leu-73 - Pro (CTG - CCG), Ile-106 Thr (ATT -* ACT), and Thr-127 -. Ser (ACT -- TCT). 1The isolate contained a second amino acid substitution: Glu-191 Gly (GAA GGA). Data for activation and repression are for a single-substitution mutant constructed by site-directed mutagen- esis. §The isolate contained a second amino acid substitution: Stop -- Lys (TAA -. AAA). Data for activation and repression are for a single-substitution mutant constructed by site-directed mutagen- esis. $One isolate contained a second amino acid substitution: Arg-122 -. Cys (CGT -- TGT). Downloaded by guest on September 25, 2021 Biochemistry: Zhou et al. Proc. Natl. Acad. Sci. USA 90 (1993) 6083 strain XE82, and transformants were plated to double-sugar .A ribose/lactose/tetrazolium/ampicillin indicator agar. CrpPC
clones were identified by colony color. CrpPc clones exhib- 60 ited pink or red colony color (Rbs- Lacd). In contrast, Crp+ and Crp- clones exhibited white colony color (respectively, Rbs+ Lac- and Rbs- Lac+). Thirty-seven independent mutagenesis reactions were per- E 40 a) formed, and a total of 18,000 mutagenized transformants 0D ai) were screened. From these, 24 independent candidate CrpPC Ci C clones were identified. 20 For each candidate CrpPc clone, plasmid DNA was pre- co0) pared and was introduced by transformation into Acrp rbs+ lac+ strain XE64.2 and Acrp rbs+ lacPUV5-OCAP strain E XE82, and transcription activation at the lac promoter and w a.
at aa 162 (Table 1). The substitution at aa 156, one substitution z at aa 158, and one substitution at aa 159 were obtained together with second substitutions outside the region consisting of aa FIG. 1. Transcription activation. (A) Transcription activation at as initiation in vitro 156-162 (see legend to Table 1). In each case, the CrpPc the lac promoter assessed in abortive transcrip- tion experiments. (B) Transcription activation at the lacP(ICAP) phenotype was established not to depend on the substitution promoter, a derivative of the lac promoter having a consensus DNA outside the region consisting of aa 156-162 by construction of site for CAP (16), as assessed in abortive initiation in vitro transcrip- a single-substitution mutant (using site-directed mutagenesis; tion experiments. method described in ref. 18) followed by analysis oftranscrip- tion activation at lac and transcription repression at lacPUV5- tives fail to activate transcription or activate it <2-fold. We OCAP. We conclude that aa 156-162 comprise a region of CAP conclude that the substituted CAP derivatives are com- that is critical for transcription activation at the lac promoter pletely, or nearly completely, defective in transcription ac- but not critical for DNA binding. tivation at lac and lacP(ICAP). For three of the four target aa, we obtained multiple (up to For wild-type CAP and [Ala-1581CAP, we also have per- three) different substitutions. Furthermore, for four of the formed abortive initiation in vitro transcription experiments eight substitutions we obtained multiple (up to inde- eight) using as templates the CAP-independent lacPUV5 promoter pendent isolates. These statistics suggest that our analysis (21) and the partly CAP-independent lacPs promoter (21). approached, or reached, saturation of available targets. That The results indicate that the Thr-158 -+ Ala substitution has is, these statistics suggest that there are few, possibly no, no effect on CAP-independent at lacPUVS and other aa at which a substitution obtainable by a single-base- transcription the effect of the Thr-158 pair change could confer a CrpPc phenotype able to pass the lacPs (data not shown). We infer that We screen. Therefore, we propose that aa 156-162 comprise the Ala substitution is specific to activated transcription. only region of CAP that is critical for transcription activation infer further that the effect of the Thr-158 -* Ala substitution at the lac promoter but not critical for DNA binding. is not due to creation of a new, unfavorable interaction that Purification and Analysis of the Substituted CAP Deriva- prevents association of RNA polymerase with the promoter tives. To confirm and extend the in vivo results, we have and/or initiation by RNA polymerase at the promoter. purified and analyzed in vitro five of the eight substituted DNA Binding. To assess DNA binding, we have performed CAP derivatives: [Ala-158]CAP, [Ile-158]CAP, [Leu-159]- nitrocellulose filter in vitro DNA binding experiments (3) CAP, [Arg-159]CAP, and [Cys-162]CAP. For each of these using a 40-bp DNA fragment having the consensus DNA site substituted CAP derivatives, we have performed quantitative in vitro experiments to assess transcription activation, DNA Table 2. DNA binding: Consensus DNA site binding, and DNA bending. Protein Kobs, X 1010 M-1 Transcription Activation. To assess transcription activa- CAP 7±1 tion, we have performed abortive initiation in vitro transcrip- 8 ± 1 tion experiments (16, 21) using as templates the lac promoter [Ala-158]CAP 8 ± 1 and the lacP(ICAP) promoter, a derivative of the lac pro- [Ile-158]CAP 4 ± 1 moter having a consensus DNA site for CAP (16). The results [Leu-159]CAP 4 ± 1 are presented in Fig. 1. Wild-type CAP activates transcrip- [Arg-159]CAP tion 410-fold at the lac promoter and %20-fold at the lacP- [Cys-162]CAP 4 ± 2 (ICAP) promoter. In contrast, the substituted CAP deriva- Results are expressed as means ± 1 SD. Downloaded by guest on September 25, 2021 6084 Biochemistry: Zhou et al. Proc. Natl. Acad. Sci. USA 90 (1993)
Nrul EcoRV Nhel Miul CAP Nrul EcoRV Nhel Mlul 0 24 60 72 100 160 184 220 232 I
NIrul I_ Nrul FIG. 2. DNA fragments used in DNA bending experiments. Top line shows the 232-bp Nru I/Mlu EcoRV EcoRV I region of plasmid pAG-ICAP. The next four lines show four circularly permuted DNA fragments gen- erated by digestion ofplasmid pAG-ICAP with Nru Nlhel Nhel I, EcoRV, Nhe I, and Mlu I, respectively. Each L _ DNA fragment is 160 bp long and contains the _711M--R~ consensus DNA site for CAP. Distances between Mlul the center ofthe DNA site for CAP and the left end Mlul of the DNA fragment are 100, 76, 40, and 28 bp, __I respectively. for CAP. The results are presented in Table 2. Under equilibrium binding constants identical to that exhibited by standard assay conditions (200 mM NaCl; pH 7.3; 23°C), wild-type CAP. We conclude that the substituted CAP de- wild-type CAP binds to the consensus DNA site with an rivatives are not defective in DNA binding. equilibrium binding constant of 7 x 1010 M-1. Within exper- DNA Bending. To assess DNA bending, we have per- imental error (±2 SD) the substituted CAP derivatives exhibit formed electrophoretic mobility DNA bending experiments (7, 9-11). We have used four circularly permuted 160-bp a b c d e f g h i j k m DNA fragments having the consensus DNA site for CAP (Fig. 2). In each DNA fragment, the center of the DNA site for CAP is located a different fractional distance from the left end of the DNA fragment-i.e., 0.63, 0.48, 0.25, or 0.18. For wild-type CAP and for each substituted CAP derivative, we have formed protein-DNA complexes with each DNA frag- ment and analyzed the electrophoretic mobilities of the protein-DNA complexes. The results are presented in Fig. 3. In the case of wild-type CAP, the electrophoretic mobilities ofthe protein-DNA complexes exhibit a pattern diagnostic of protein-induced DNA bending (cf. refs. 7 and 9-11); i.e., electrophoretic mobility is least for the protein-DNA com- a b c d e f g h j k m plex formed with the DNA fragment having the DNA site for CAP located nearest the center of the DNA fragment (frac- tional position 0.48), and electrophoretic mobility is greatest for the protein-DNA complex formed with the DNA frag- ment having the DNA site for CAP located nearest the end of the DNA fragment (fractional position 0.18). In the case of the substituted CAP derivatives, the electrophoretic mobil- ities ofthe protein-DNA complexes likewise exhibit patterns diagnostic of protein-induced DNA bending. Strikingly, the patterns in the cases of the substituted CAP derivatives are identical to the pattern in the case of wild-type CAP. We conclude that the substituted CAP derivatives are not defec- tive in DNA bending. a b c d e f g h It is possible to calculate the DNA bend angle from electrophoretic mobility data using equations derived by Thompson and Landy (9). The results ofsuch calculations are presented in Table 3. In the case of wild-type CAP, the estimated DNA bend angle is 1210. Within experimental error (±1 SEM), in the cases of the substituted CAP derivatives, the estimated DNA bend angles are identical to that of wild-type CAP. We conclude that the substituted CAP de- rivatives are not defective in DNA bending-neither quali- tatively nor quantitatively. DISCUSSION FIG. 3. DNA bending: consensus DNA site. (Top) Lanes: a, plasmid pAG-ICAP digested with Nru I; b-e, wild-type CAP plus aa 156-162 of CAP Comprise an Activating Region. Our plasmid pAG-ICAP digested with Nru I, EcoRV, Nhe I, and Mlu I, results establish that aa 156, 158, 159, and 162 of CAP are respectively; f-i, [Ala-158]CAP plus plasmid pAG-ICAP digested critical for transcription activation at the lac promoter but not with Nru I, EcoRV, Nhe I, and Mlu I, respectively; j-m, [Ile- critical forDNA binding and not critical for DNA bending. Our 158]CAP plus plasmid pAG-ICAP digested with Nru I, EcoRV, Nhe I, and Mlu I, respectively. (Middle) Lanes: a-e, same as in Top; f-i, [Leu-159]CAP plus plasmid pAG-ICAP digested with Nru I, EcoRV, respectively. In lanes b-m of each panel, the lowest band corre- Nhe I, and Mlu I, respectively; j-m, [Arg-159]CAP plus plasmid sponds to the uncomplexed 160-bp DNA fragment, the intermediate pAG-ICAP digested with Nru I, EcoRV, Nhe I, and Mlu I, respec- band corresponds to the protein-DNA complex formed with the tively. (Bottom) Lanes: a-e, same as in Top; f-i, [Cys-162]CAP plus 160-bp DNA fragment, and the highest band corresponds to the plasmid pAG-ICAP digested with Nru I, EcoRV, Nhe I, and Mlu I, 3068-bp vector DNA fragment. Downloaded by guest on September 25, 2021 Biochemistry: Zhou et al. Proc. Natl. Acad. Sci. USA 90 (1993) 6085 Table 3. DNA bending: Consensus DNA site side-chain interactions required for transcription activation. We suggest that substitutions at aa 156 and 162 affect tran- Protein DNA bend angle, scription activation indirectly-by disrupting the activating CAP 121 ± 1 region, either sterically, electrostatically, or conformationally. [Ala-158]CAP 121 ± 1 A Protein-Protein Contact Between aa 156 and 162 of CAP [Ile-158]CAP 121 ± 1 and RNA Polymerase Is Necessary for Transcription Activa- [Leu-159]CAP 121 ± 1 tion. Our results establish that it is possible to separate [Arg-159]CAP 123 ± 1 transcription activation at the lac promoter by CAP from [Cys-162]CAP 120 ± 1 DNA binding by CAP and DNA bending by CAP. We Results are expressed as means ± 1 SEM. conclude that there is a step absolutely necessary for tran- scription activation in addition to DNA binding and DNA results further establish that it is likely that no part of CAP bending. We conclude further that this additional step in- other than aa 156-162 is critical for transcription activation at volves aa 156-162 of CAP. We propose that this additional the lac promoter but not critical for DNA binding. Our results step is a direct protein-protein contact between aa 156 and results of Bell et al. who have 162 of CAP and RNA polymerase. confirm and extend the (22), Three additional results are consistent with the proposal shown that aa 159 of CAP is critical for transcription activa- that transcription activation at the lac promoter requires a tion, and of Eschenlauer and Reznikoff (23), who have shown direct protein-protein contact between CAP and RNA poly- that aa 162 of CAP is critical for transcription activation. merase: (i) Transcription activation at the lac promoter In the three-dimensional structure of CAP (4) and in the requires that the DNA site for CAP and the DNA site for three-dimensional structure ofthe CAP-DNA complex (5), aa RNA polymerase be on the same face of the DNA helix (24, 156, 158, 159, and 162 are part of a surface loop composed of 25). (ii) CAP and RNA polymerase bind cooperatively to the aa 152-166. aa 156, 158, 159, and 162 form a cluster, with lac promoter (25, 26). Cooperative binding, like transcription dimensions of -13 x -7 A, on the surface of the CAP-DNA activation, requires that the DNA site for CAP and the DNA complex (Fig. 4). These aa are distant from the region ofCAP site for RNA polymerase be on the same face of the DNA involved in DNA binding. In addition, these aa are distant helix (25). (iii) CAP and RNA polymerase interact in solution from the region of CAP thought to be involved in DNA in the absence of promoter DNA (27). This interaction bending. Thus, the location of these aa is consistent with the absolutely requires the activating region of CAP (27). observation that their substitution does not affect DNA bind- ing and DNA bending. The location of these aa is consistent We thank S. Adhya for plasmid pBend3 and A. Kolb for plasmids with the proposal they participate in a direct protein-protein pBR-203-lac, pBR-203-lacPUVS, and pBR-203-lacPs. This work was contact with a molecule of RNA polymerase bound adjacent supported by National Institutes ofHealth Grant GM41376 to R.H.E. to CAP on the same face of the DNA helix (see below). and Merck and Busch predoctoral fellowships to Y.Z. aa 158 and 159 have side chains that are exposed to solvent 1. de Crombrugghe, B., Busby, S. & Buc, H. (1984) Science 224, on the surface of the CAP-DNA complex. Therefore, aa 158 831-838. and 159 are candidates to participate in direct side-chain 2. Reznikoff, W. (1992) J. Bacteriol. 174, 655-658. interactions required for transcription activation. In contrast, 3. Gunasekera, A., Ebright, Y. & Ebright, R. (1992) J. Biol. Chem. 162 not that are to 267, 14713-14720. aa 156 and do have side chains exposed 4. Weber, I. & Steitz, T. (1987) J. Mol. Biol. 198, 311-326. solvent on the surface of the CAP-DNA complex. Therefore, 5. Schultz, S., Shields, G. & Steitz, T. (1991) Science 253, 1001-1007. aa 156 and 162 are not candidates to participate in direct 6. Porschke, D., Hillen, W. & Takahashi, M. (1984) EMBO J. 3, 2873-2878. 7. Wu, H.-W. & Crothers, D. (1984) Nature (London) 308, 509-513. 8. Antosiewicz, J. & Porschke, D. (1988) J. Biomol. Struct. Dyn. 5, 819-837. 9. Thompson, J. & Landy, A. (1988) NucleicAcids Res. 16, 9687-9705. 10. Kim, J., Zweib, C., Wu, C. & Adhya, S. (1989) Gene 85, 15-23. 11. Zinkel, S. & Crothers, D. (1990) Biopolymers 29, 29-38. 12. Heyduk, T. & Lee, J. (1992) Biochemistry 31, 5165-5171. 13. Ptashne, M. (1988) Nature (London) 335, 683-689. 14. Ptashne, M. & Gann, A. (1990) Nature (London) 346, 329-331. 15. Irwin, N. & Ptashne, M. (1987) Proc. Natl. Acad. Sci. USA 84, 8315-8319. 16. Zhang, X., Zhou, Y., Ebright, Y. & Ebright, R. (1992) J. Biol. Chem. 267, 8136-8139. 17. Zhou, Y., Zhang, X. & Ebright, R. (1991) Nucleic Acids Res. 19, 6052. 18. Kunkel, T., Bebenek, K. & McClary, J. (1991) Methods Enzymol. 204, 125-138. 19. Zhang, X., Gunasekera, A., Ebright, Y. & Ebright, R. (1991) J. Biomol. Struct. Dyn. 9, 463-473. 20. Bell, A., Buckel, S., Groarke, J., Hope, J., Kingsley, D. & Hermodson, M. (1986) J. Biol. Chem. 261, 7652-7658. 21. Malan, T. P., Kolb, A., Buc, H. & McClure, W. (1984) J. Mol. Biol. 180, 881-909. 22. Bell, A., Gaston, K., Williams, R., Chapman, K., Kolb, A., Buc, H., Minchin, S., Williams, J. & Busby, S. (1990) Nucleic Acids Res. 18, 7243-7250. 23. Eschenlauer, A. & Reznikoff, W. (1991) J. Bacteriol. 173, 5024-5029. 24. Mandecki, W. & Caruthers, M. (1984) Gene 31, 263-267. FIG. 4. Structure of the CAP-DNA complex showing the acti- 25. Straney, D., Straney, S. & Crothers, D. (1989) J. Mol. Biol. 206, vating region. The crystallographic structure of the CAP-DNA 41-57. complex at 3.0 A resolution is shown (5). CAP is illustrated in blue. 26. Ren, Y. L., Garges, S., Adhya, S. & Krakow, J. (1988) Proc. Natl. DNA and CAP-bound cAMP are illustrated in red. aa at which Acad. Sci. USA 85, 4138-4142. substitutions conferring a CrpPc phenotype were obtained-i.e., aa 27. Heyduk, T., Lee, J., Ebright, Y., Blatter, E., Zhou, Y. & Ebright, 156, 158, 159, and 162-are illustrated in dark blue. R. (1993) Nature (London), in press. Downloaded by guest on September 25, 2021