Contributed by Jon Beckwith, May 15, 1987. ABSTRACT. It has been proposed that Glu-181 of the catabolite gene activator protein (CAP) makes direct contact.
Proc. Nati. Acad. Sci. USA Vol. 84, pp. 6083-6087, September 1987 Biochemistry
Role of glutamic acid-181 in DNA-sequence recognition by the catabolite gene activator protein (CAP) of Escherichia coli: Altered DNA-sequence-recognition properties of [Val'81]CAP and [Leu181]CAP ("loss-of-contact" mutations/DNA-protein interaction/helix-turn-helix motif/cAMP receptor protein)
RICHARD H. EBRIGHT*t, ANNIE KOLBt, HENRI BUCt, THOMAS A. KUNKELt, JOSEPH S. KRAKOW§, AND JON BECKWITH* *Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, MA 02115; tUnite de Physicochimie des Macromolecules Biologiques, Institut Pasteur, 25 rue du Dr. Roux, 75724 Paris, cedex 15, France; tLaboratory of Genetics, National Institute of Environmental Health Sciences, P.O. Box 12233, Research Triangle Park, NC 27709; and §Department of Biological Sciences, Hunter College of the City University of New York, New York, NY 10021
Contributed by Jon Beckwith, May 15, 1987
ABSTRACT It has been proposed that Glu-181 of the catabolite gene activator protein (CAP) makes direct contact with certain base pairs of the specific DNA site. We have purified wild-type CAP and two substituted CAP variants, [Val'81]CAP and [Leu'811CAP, and have assessed the DNAsequence-recognition properties in vitro with respect to positions 5, 6, 7, 8, and 16 of the DNA site. The data indicate that [Val'81JCAP and [Leu81J]CAP fail to discriminate between the consensus DNA base pair and the three non-consensus-DNA base pairs at 2-fold-related positions 7 and 16 of the DNA site. In contrast, [Val'81]CAP and [Leu'81]CAP retain the ability to discriminate between different base pairs at positions 5 and 8 of the DNA site. We conclude that Glu-181 of CAP makes a direct contact with 2-fold-related positions 7 and 16 of the DNA site, as proposed previously based on in vivo results. We propose that upon replacement of Glu-181 by valine or leucine, this contact is eliminated and is replaced by no other functional contact. We estimate that the contact by Glu-181 with each position contributes -0.7 kcal/mol to the total CAP-DNA binding free energy. The catabolite gene activator protein (CAP; also referred to as the cAMP receptor protein, CRP), when complexed with cAMP, binds to specific DNA sites at or near promoters, where it stimulates the initiation of RNA synthesis (1-3). The three-dimensional structure of CAP has been determined to 2.9-A resolution based on x-ray diffraction analysis (4). CAP is a dimer of two chemically identical subunits, each consisting of 209 amino acids (4). The consensus DNA site for CAP is 5' AA-TGTGA------TCACA-TT 3', where each hyphen represents any nucleotide, N (ref. 3; Fig. 1); it is 22 base pairs (bp) in length and exhibits 2-fold sequence symmetry. We are interested in determining how the DNA sequence information content at each specified position within the consensus DNA site is recognized by CAP. In previous work (5) using an in vivo genetic selection, three mutations were isolated that alter the DNA-sequence specificity of CAP with respect to 2-fold-related positions 7 and 16 of the DNA site. As characterized in vivo, the mutations conferred two properties: (i) a decreased ability of the mutant CAP, compared with wild-type CAP, to interact with the consensus DNA site; and (ii) an increased ability of the mutant CAP, compared with wild-type CAP, to interact with the substituted DNA sites containing A-T at position 7 (mutation designated lacPJ-L8) or T-A at position 16 (mutaThe publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
tion designated lacPI-L29). The three mutations were found to result in substitution of the same amino acid of CAP-i.e., Glu-181. The amino acid substitutions inferred from DNAnucleotide sequencing were valine, leucine, and lysine. Glu-181 is amino acid 2 of the second a-helix of the a-helixturn-a-helix DNA binding motif of CAP (see refs. 6 and 7). Based on these results, we have proposed that Glu-181 makes direct contact with 2-fold-related positions 7 and 16 of the DNA site (5). Using this proposed contact, in conjunction with the three-dimensional structure of CAP (4), we have described a detailed model for the structure of the CAP-DNA complex (8). In this model, the Glu-181 side-chain carboxylate group of one subunit of the CAP dimer is in H-bonded contact with the cytosine N-4 atom of the G-C base pair at position 7 of the DNA site, and the Glu-181 side-chain carboxylate group of the opposite subunit of the CAP dimer is in H-bonded contact with the cytosine N-4 atom of the C-G base pair at position 16 of the DNA site. Weber and Steitz (9) have set forth a similar, but not identical, detailed model for the structure of the CAP-DNA complex, in this case based on considerations of electrostatic and steric complementarity. The model of Weber and Steitz (9) copredicts the H-bonded contact by Glu-181 with positions 7 and 16 of the DNA site and predicts in addition an H-bonded contact by Glu-181 with positions 6 and 17 of the DNA site. In principle, there are two possible explanations for the increased abilities of [Val18l]CAP, [Leu18l]CAP, and [Lys181]CAP, compared with wild-type CAP, to interact with DNA sites containing A-T at position 7 or T-A at position 16: (i) the presence of energetically favorable interactions between Val-181, Leu-181, and Lys-181 and the base thymine; or (ii) the absence of functional interaction between Val-181, Leu181, and Lys-181 and the base thymine, the increased ability to interact being due to the elimination of a putative energetically unfavorable Glu-181-thymine interaction (8). In our previous report, we strongly favored explanation i (8). In the work reported here, we purified wild-type CAP, [Val'1']CAP, and [Leu81l]CAP to homogeneity and characterized the DNA-sequence-recognition properties in vitro with respect to positions 5, 6, 7, 8, and 16 of the DNA site of the Escherichia coli lacPI promoter. The results provide confirmation of our proposal that Glu-181 makes a direct contact with base pairs 7 and 16 of the DNA site. Further, they show that the increased abilities of [Vall1J]CAP and [Leu181]CAP to interact with DNA sites containing A-T at position 7 and/or T-A at position 16 are due to the elimination of a putative Abbreviations: CAP, catabolite gene activator protein; P, promoter.
Biochemistry: Ebright et al.
energetically unfavorable Glu-181-thymine interaction (explanation ii above).
MATERIALS AND METHODS In this section, DNA sequences are numbered with respect to the start point of the E. coli lacPJ promoter. By this numbering convention, position 1 of the lacPJ DNA site for CAP is designated bp -72 (ref. 10; Fig. 1). Bacteriophage M13mp2 Derivatives. M13mp2, M13mp2lacPJ(-68A), M13mp2-lacPl(-67C), M13mp2-lacPJ(-66A), M13mp2-lacPJ(-66C), M13mp2-lacPJ(-66T), M13mp2-IacP1(-65G), M13mp2-lacPJ(-57T), M13mp2IacP1(-57G), and M13mp2-lacPl(-57A) have been described (11-15). M13mp2-IacPl(-66A;-57T), M13mp2IacPI(-66C;-57G), and M13mp2-lacPl(-66T;-57A) were constructed in the following manner. The 64-bp Pvu II-Alu I fragment corresponding to bp -123 to bp -60 was isolated from each of M13mp2-lacPJ(-66A), M13mp2-lacPJ(-66C), and M13mp2-IacPJ(-66T). The resulting fragments were hybridized with uracil-incorporated (16) single-stranded circular viral DNA isolated from, respectively, M13mp2-lacPl(-57T), M13mp2-lacP1(-57G), and M13mp2-lacP1(-57A). In vitro DNA synthesis reactions were carried out as described (13, 16). The product heteroduplexes were transfected (16), and the desired bacteriophage clones were identified by a very faint blue plaque color. The DNAnucleotide sequence of the lac insert in M13mp2 and in each M13mp2 derivative was verified. DNA Fragment Isolation and Labeling. Double-stranded DNA prepared by the method in ref. 13 from M13mp2 or M13mp2 derivatives was digested with Ava II and EcoRI. The reaction products were separated by polyacrylamide gel electrophoresis, and the 314-bp fragment corresponding to bp -261 to bp +53 was eluted. This fragment was digested with Hha I and BstN 1, and the reaction products were labeled with [a-32P]dATP (Amersham) and DNA polymerase I Klenow fragment (supplied by T. Steitz, Yale University). The products were separated by polyacrylamide gel electrophoresis, and the 42-bp fragment corresponding to bp -81 to bp -40 was eluted and washed extensively. Fragments utilized had a specific activity of 0.2 Bq/fmol to 7 Bq/fmol. DNase I protection analysis confirmed that CAP, [Val181]CAP, and [Leu181]CAP bind to a single DNA site on the 42-bp fragment (data not shown). Purification of CAP. Wild-type CAP, [Val181]CAP, and [Leu181]CAP were purified as described (17) from E. coli strains containing, respectively, plasmids pHA5, pPC181V, and pPC181L. Plasmid pHA5 has been described (18). Plasmids pPC181V and pPC181L were constructed by P. Cossart and B. Gicquel-Sanzey by inserting into the BamHI site of plasmid pBR322 (19) the 3500-bp BamHI-BamHI fragments isolated from phage A Y1079-crp'L29-2 (5) and A Y1079-crp'L8-2 (5). Protein concentrations were determined from absorbance at 278 nm (extinction coefficient = 2.00 x 104 M-'cm-1). The activity of each protein preparation was determined by the titration of a fixed concentration of the protein with increasing concentrations of the 42-bp H/ha I-BstNI DNA fragment containing the wild-type lacPJ DNA site. The data in the tables have been corrected for differences in activity. In Vitro DNA Binding Assays. Reaction mixtures (8,ul; ionic strength = 0.138 M) contained 40 mM Tris HCl (pH 8.0), 100 mM KCl, 6.0 mM NaCI, 10 mM MgCI2, 1.0 mM dithiothreitol, 100 micrograms of bovine serum albumin per ml, 0.20 mM cAMP, 0.3-2 nM labeled DNA fragment, and 0.625 nM, 2.5 nM, or 10 nM CAP or CAP derivative. Samples were incubated 45 min at 22°C. Immediately before loading for electrophoresis, 1.5 ,Al of 50% glycerol and 0.01% xylene cyanol in the same buffer was added. Electrophoresis was
Proc. Natl. Acad. Sci. USA 84 (1987)
performed at 24 V/cm through 7.5% polyacrylamide slab gels (70 mm x 70 mm x 1.5 mm) equilibrated with a buffer containing 89 mM Tris/89 mM borate (pH 8.2), 2.0 mM Na2EDTA, and 0.20 mM cAMP. Electrophoresis was for 25 min at 22°C. Developed gels were dried and autoradiographed on Kodak X-Omat XAR5 film. Appropriately exposed autoradiograms were scanned with a Vernon microdensitometer. RESULTS We have constructed in our experiments a set of 13 derivatives of the IacPJ DNA site for CAP. The set consists of the wild-type lacPJ DNA site and 12 lacPI DNA sites containing either one or two single-base-pair substitutions (sequences in Fig. 1). In vitro analysis of the binding of [Val181]CAP and [Leu181]CAP to this set of derivatives has enabled us (i) to confirm the changes in binding affinity previously deduced based on in vivo results (5) and (ii) to assess quantitatively the changes in binding specificity with respect to base pairs at positions 5, 6, 7, 8, 9, and 10 of the DNA site. In vitro DNA binding was assayed by the gel-retardation assay technique developed by Garner and Revzin (20) and Fried and Crothers (21). The ligand utilized was a 42-bp Hha I-BstNI DNA fragment containing the 22-bp lacPI DNA site for CAP and 10 bp of flanking DNA on each side. Thirteen versions of the DNA fragment were utilized: 1 with the wild-type lacPI DNA site and 12 with lacPJ DNA sites having either one or two substitutions. The results are presented in Tables 1, 2, and 3. Note that the results with [Val18l]CAP and [Leu181]CAP are similar. Wild-Type DNA Site. Wild-type CAP exhibits a dissociation constant, Kd, of 0.21 x 10-9 M for the wild-type DNA site (Table 1; Fig. 2). As compared with wild-type CAP, [Val18l]CAP and [Leu181]CAP exhibit reduced affinities for the wild-type DNA site by factors of 13 and 9, respectively. These factors correspond to reductions of 1.5 kcal/mol and 1.3 kcal/mol in the binding free energy. These data are consistent with in vivo results (5). DNA Site Symmetrically Altered at Positions 7 and 16. Wild-type CAP exhibits no detectable affinity for the three DNA sites symmetrically altered at positions 7 and 16-i.e., the DNA site containing A-T at position 7 and T-A at position 16, the DNA site containing C-G at position 7 and G-C at position 16, and the DNA site containing T-A at position 7 and AT at position 16. Given the presumed sensitivity of our 10
A A - T G T G A T T - A C A C T -70
T A - T G T G A A T - A C A C T -
C A C A - T T ----A G T G T - A A
- - - -T
AC A G T GT C
TC A C T - A T AG T G A - T A
FIG. 1. (Upper) The consensus DNA site for CAP (ref. 3; cf. refs. 2 and 5). The axis of 2-fold sequence symmetry is located between
positions 11 and 12. (Lower) The wild-type IacPI DNA site for CAP [bp -72 to bp -50 with respect to the start point of lacPl (10)]. The substitutions at positions 5, 6, 7, 8, and 16 utilized in this study are indicated beneath the sequence (10-14). The G C -- A-T substitution at position 7 and the C-G -* T-A substitution at position 16 elsewhere are designated L8 and L29, respectively.
Biochemistry: Ebright et al.
Proc. Natl. Acad. Sci. USA 84 (1987)
Table 1. Recognition of DNA sites symmetrically altered at positions 7 and 16 Wild-type CAP [Val'81]CAP [Leu"l]CAP DNA site Kd, M x 109 x X Kd, M M Kdp-/Kdp+ 109 Kd, 109 Kdp-/KdP+ Kdp-/Kdp+ + (7G; 16C)  0.21  2.8 1.9  7A; 16T >200 >950 2.2 0.79 1.6 0.84 7C; 16G >200 2.3 >950 12 6.5 6.3 7T; 16A >200 >950 3.0 1.1 2.5 1.3 " + " denotes the wild-type lacPJ DNA site for CAP; substituted lacPJ DNA sites for CAP are identified by the position and sequence of the substitutions (sequences in Fig. 1). The electrophoretic mobilities of all protein-DNA complexes analyzed were identical. Data from assays where the ratio of protein-bound DNA fragment to free DNA fragment was >20 or 200 x 10-9 M (Table 1; Fig. 2). As compared with wild-type CAP, [Val181]CAP and [Leu181]CAP exhibit at least 31-fold and at least 17-fold increased affinities for the three DNA sites symmetrically altered at positions 7 and 16. These factors correspond to increases of at least 2.0 kcal/mol and 1.7 kcal/mol in the binding free energy. These data also are consistent with in vivo results (ref. 5; R.H.E. and T.A.K., unpublished data). The data are most informative when expressed as the ratio Kdp-/KdP+ (i.e., as the dissociation constant for a substituted DNA site divided by the dissociation constant for the wild-type site). Wild-type CAP discriminates between the wild-type DNA site and the three DNA sites symmetrically altered at positions 7 and 16 by factors of >950 (more than +4.0 kcal/mol). In contrast, [Val181]CAP discriminates between the wild-type DNA site and the three DNA sites symmetrically altered at positions 7 and 16 by a mean factor of only 1.4 (+0.20 kcal/mol; 0 kcal/mol within experimental error). [Leu181]CAP discriminates by a mean factor of only 2.8 (+0.60 kcal/mol). The data indicate that replacement of Glu-181 by valine or leucine results in the essentially complete elimination of specificity between G-C, A-T, C-G, and T-A at positions 7 and 16. DNA Sites Substituted at Position 7 or at Position 16. Wild-type CAP discriminates between the wild-type DNA site and the DNA sites containing single substitutions at position 7 or at position 16 by factors of 12 to 240 (Table 2). The mean factor of discrimination is 110 (+2.8 kcal/mol). CAP
12.0 0 -j