Visualization of cAMP receptor protein-induced DNA kinking by electron microscopy

Efficient control of raf gene expression by CAP and two Raf repressors that bend DNA in opposite directions

The plasmid-borne raf operon of Escherichia coli encodes proteins involved in the uptake and utilisation of the trisaccharide raffinose. The operon is subject to dual regulation; to negative control by the binding of RafR repressor to twin operators, O1 and O2, and to positive control by the cAMP-binding protein, CAP. We have identified the CAP binding site (CBS) as a 22 bp palindromic sequence with incomplete dyad symmetry by deletion analysis, DNasel footprinting and electrophoretic mobility shift assays (EMSA) of CAP-DNA complexes. The CBS is centred 60.5 bp upstream of the transcription start point and partially overlaps O1. In vivo, CAP increases rafA (alpha-galactosidase) gene expression up to 50-fold. The 28 bp spacing between the centres of CBS and the - 35 box is essential, since insertions of 4, 8, 12 or 16 bp completely eliminated rafA gene expression. In vitro binding studies revealed that the CBS, O1 and O2 sites, can be simultaneously occupied by their cognate proteins. However, no cooperativity between binding of CAP and RafR was detected. EMSA with circularly permuted DNA fragments demonstrated that CAP and RafR proteins bend raf promoter (rafP) DNA by 75 degrees +/- 5 degrees and 95 degrees +/- 5 degrees, respectively, in opposite directions. Among sugar catabolic operons, the compact arrangement of three protein-binding sites, a CBS and two operators bounding the - 35 promoter box, is unique and provides a sensitive and highly efficient device for transcriptional control.

Solution structure of the exocyclic 1,N2-propanodeoxyguanosine adduct opposite deoxyadenosine in a DNA nonamer duplex at pH 8.9. model of pH-dependent conformational transition

The solution structure of the complementary d(C1-A2-T3-G4-X5-G6-T7-A8-C9).d(G10-T11-A12-C13-A14-C15-A16-T17-G18) DNA duplex (designated X.A 9-mer), which contains a 1,N2-propanodeoxyguanosine exocyclic adduct X5 opposite deoxyadenosine A14 at the center, is pH dependent [Kouchakdjian, M., Eisenberg, M., Live, D., Marinelli, E., Grollman, A., & Patel, D.J. (1990) Biochemistry 29, 4456-4465]. In our previous paper [Huang, P., & Eisenberg, M. (1992) Biochemistry 31, 6518-6532] we established the three-dimensional structure of this X.A 9-mer duplex at pH 5.8 by use of restrained molecular dynamics followed by NOE-based back-calculation refinement. The present paper discusses the structure at pH 8.9 and the pH-dependent conformational transition between the structures at pH 5.8 and at pH 8.9. The structure at pH 8.9 is calculated starting from five different conformations. The final structures converge and agree well with the experimental NOE intensities. These structures are essentially B-type DNA (with X5 and A14 in the BII conformation while the other residues are in the most commonly described BI conformation) and display an approximate 27 degrees kink at the center of the helix. At the kink site, X5 is positioned in the major groove with the exocyclic ring directed toward the G6.C13 base pair, unstacked from the flanking base G6 and exposed to the solvent. A14, opposite the lesion, remains stacked with its neighbor C15, but not with C13. The kinked helix can accommodate the rotation of the bulky X5 about its glycosidic bond. We propose here a model for the pH-dependent transition. Our model explains the conformational change, which includes the anti and syn rotation of the bulky adduct around its glycosidic bond, with a minimal energy barrier and with an overall kink of the DNA helix. These new findings, fully consistent with the NMR experimental data, were revealed only after restrained dynamics refinement. Distance-restrained energy minimization by itself was insufficient, as shown by the previous NMR study.

Bending of DNA by transcription factors

An increasing number of transcription factors both from prokaryotic and eukaryotic sources are found to bend the DNA upon binding to their recognition site. Bending can easily be detected by the anomalous electrophoretic behaviour of the DNA-protein complex or by increased cyclization of DNA fragments containing the protein-induced bend. Induction of DNA bending by transcription factors could regulate transcription in various ways. Bending may bring distantly bound transcription factors closer together by facilitating DNA-looping or it could mediate the interaction between transcription factors and the general transcription machinery by formation of large nucleoprotein structures in which the DNA is wrapped around the protein complex. Alternatively, the energy stored in a protein-induced bend could be used to favour formation of an open transcription complex or to dissociate the RNA polymerase in the transition from initiation to elongation. Modification of the bend angles and bending centers, caused by homodimerization or heterodimerization of transcription factors, may well turn out to be an important way to enlarge the range of interactions required for regulation of gene expression.

Specific binding of cyclic-AMP receptor protein to DNA. Effect of the sequence and of the introduction of a nick in the binding site

The binding of Escherichia coli Cyclic AMP Receptor Protein (CRP) to several DNA fragments of about 45 base pairs, bearing the natural lactose or galactose sites, as well as several synthetic related sites, was investigated using fluorescence spectroscopy and gel retardation experiments. The salt dependence of the equilibrium binding constant indicates that CRP makes an identical number of ion pairs with the lac, lacL8 and gal sites although the binding constants are drastically different. However increasing the symmetry of the gal site leads to an increase of the number of ion pairs between the protein and the DNA. A single strand nick was introduced at the centre of a symmetrized gal site and this reduces the binding energy of CRP by about 0.6 Kcal. These results are discussed with respect to the bending constraints imposed on the DNA by the binding of CRP. The results are in agreement with the recently published crystal structure of the CRP complexed with DNA [Schutz, S.C., Shields, G.C. and Steitz, T.A., Science 253, 1001-1007 (1991)] showing that the 90 degrees bending of the DNA in the complex results from two kinks.

A comparison of the DNA bending activities of the DNA binding proteins CRP and TFIID

Protein-induced DNA bending is of importance in the formation of complex nucleoprotein assemblies such as those involved in the initiation of DNA replication or transcription initiation. We have compared the DNA bending characteristics of the Escherichia coli cyclic AMP receptor protein (CRP or CAP), an archetypal DNA bending protein, to those of TFIID, the eukaryotic TATA-element binding transcription factor. By altering the helical phasing between a CRP binding site and the E. coli melR promoter we have mapped a DNA sequence-directed bend in the downstream region of the promoter. This intrinsic DNA bend may be important in the regulation of the melR promoter by CRP in vivo. Gel retardation assays and DNAse I footprinting show that human TFIID binds to the melR promoter - 10 region. Taking advantage of this fact, and using the CRP-induced DNA bend as a standard, we have employed phase sensitive detection to show that the DNA bend angle induced by TFIID is far less than that induced by CRP. Further evidence to support this conclusion comes from a comparison of the relative mobilities of CRP-DNA and TFIID-DNA complexes. These results place limits on the role of any DNA bending induced by TFIID alone in the initiation of transcription.

Quantitative electron microscopic analysis of DNA-protein interactions

Electron microscopy offers a unique potentiality to visualize individual molecules. For the last 30 years it has been used to study the structure and the interactions of various biological macromolecules. The contribution of electron microscopy is important because of its capacity to demonstrate the existence of conformational structures such as kinks, bents, loops, etc., either on naked DNA, or on DNA associated with various proteins or ligands. Increasing interest was given to such observations when it was found that they provide a direct visualization of interacting molecules involved in DNA metabolism and gene regulation. Technical advances in the preparation of the specimens, their observation in the electron microscope, and the image processing by computers have allowed the shifting from qualitative to quantitative analysis, as illustrated by a few examples from our laboratory.

Structural studies of protein-nucleic acid interaction: the sources of sequence-specific binding

Structural studies of DNA-binding proteins and their complexes with DNA have proceeded at an accelerating pace in recent years due to important technical advances in molecular genetics, DNA synthesis, protein crystallography and nuclear magnetic resonance. The last major review on this subject by Pabo & Sauer (1984) summarized the structural and functional studies of the three sequence-specific DNA-binding proteins whose crystal structures were then known, theE. colicatabolite gene activator protein (CAP) (McKay & Steitz, 1981; McKayet al.1982; Weber & Steitz, 1987), acrorepressor from phage λ (Andersonet al.1981), and the DNA-binding proteolytic fragment ofλcIrepressor protein (Pabo & Lewis, 1982) Although crystallographic studies of theE. coli lacrepressor protein were initiated as early as 1971 when it was the only regulatory protein available in sufficient quantities for structural studies (Steitzet al.1974), little was established about the structural aspects of DNA-binding proteins until the structure of CAP was determined in 1980 followed shortly thereafter by the structure ofλcrorepressor and subsequently that of the λ repressor fragment. There are now determined at high resolution the crystal structures of seven prokaryotic gene regulatory proteins or fragments [CAP,λcro,λcIrepressor fragment, 434 repressor fragment (Andersonet al.1987), 434crorepressor (Wolbergeret al.1988),E. coli trprepressor (Schevitzet al.1985),E. coli metrepressor (Raffertyet al.1989)],EcoRI restriction endonuclease (McClarinet al.1986), DNAse I (Suck & Ofner, 1986), the catalytic domain of γδ resolvase (Hatfullet al.1989) and two sequence-independent double-stranded DNA-binding proteins [the Klenow fragment ofE. coliDNA polymerase I (Olliset al.1985) and theE. coliHu protein (Tanakaet al., 1984)].

Synergy between Escherichia coli CAP protein and RNA polymerase in the lac promoter open complex

Characterization of ternary complexes containing an Escherichia coli lac promoter DNA fragment, CAP protein and RNA polymerase, separated on non-denaturing polyacrylamide gels and footprinted in the gel slice, reveals a striking stabilization of CAP against dissociation in the open complex, compared to the CAP-DNA complex lacking RNA polymerase. The stabilization is lost when half a helical turn of DNA is inserted between CAP and polymerase sites, but is partially restored with an 11 base-pair insert; stimulation of transcription parallels the stabilization effect. This behavior suggests a direct protein-protein interaction. Comparison of initiation kinetics for wild-type and a mutant in which the P2 promoter has been inactivated shows that CAP both strengthens binding in the closed complex and accelerates isomerization to the open complex; the latter effect accounts for the bulk of the observed transcriptional activation.

DNA bending induced by specific interaction of decamer binding proteins with immunoglobulin gene control sequences

In order to investigate the properties of specific DNA-binding proteins involved in tissue-specific regulation of immunoglobulin genes, we have analyzed the interaction of nuclear proteins from mouse B-cell hybridomas with promoter and enhancer sequences of a mouse immunoglobulin heavy chain gene. Visualization of specific complexes has shown that protein binding induces a sharp bend at the position of the conserved decamer sequence. After fractionation of nuclear extracts, several sequence-specific DNA binding proteins could be distinguished by UV crosslinking to radioactive synthetic oligonucleotides. Decamer binding factor I (DBF-I) a protein of 100-105 kDa and DBF-II, a family of proteins of 25-35 kDa were purified on specific DNA-affinity columns. Both proteins bend the DNA at the dc sequence as shown by electron microscopy and by gel retardation. These data suggest that one possible function of sequence-specific regulatory proteins may be to locally change the DNA topology, thereby facilitating the interaction of additional transcription factors with the primary complex.

DNA curvature in native and modified EcoRI recognition sites and possible influence upon the endonuclease cleavage reaction

The ligation of a decadeoxynucleotide containing the EcoRI recognition site forms a series of multimers which appear to be curved based on observed anomalous gel migration in polyacrylamide gels. The degree of DNA curvature present in the recognition sequence, based upon the observed migration anomaly, can be altered by modifications to the purine functional groups at the 2- and 6-positions. Deletion of the guanine 2-amino group, occurring in the minor groove of the B-DNA helix, is most effective in increasing the observed DNA curvature. Conversely, the displacement of an amino group from the major groove to the minor groove eliminates curvature. DNA curvature is also modulated by the exocyclic group at the purine 6-position with decreasing curvature observed when changing the amino group to a carbonyl or proton substituent. Differences in the kinetic parameters characterizing the cleavage reaction by the endonuclease for many of the modified sequences are the result of modifications of functional groups in the major groove, which are likely to contact the endonuclease during catalysis. However, with two examples, significant decreases in the observed specificity constant (kcat/Km), characterizing the protein-nucleic acid interaction, cannot be easily explained in terms of such functional group contacts. It is more likely in these cases that the functional group modifications affect the efficiency of the endonuclease-DNA interaction by modulation of the structure of the double-stranded DNA helix. With both examples, modifications have been made to minor groove substituents. The extent of DNA curvature is increased significantly for one and decreased for the other, compared with that observed for the native recognition site. The results suggest that curvature of the DNA helix axis is an intrinsic property of the d(GAATTC) sequence which helps to optimize the protein-nucleic acid interactions observed for the EcoRI restriction endonuclease.

Transcriptional activation of the Klebsiella pneumoniae nitrogenase promoter may involve DNA loop formation

Transcriptional activation of nitrogen fixation genes by NifA in Klebsiella pneumoniae requires an upstream NifA binding site. We now report that the introduction of half turns of the DNA helix into the DNA separating the upstream NifA binding site from the downstream promoter element of the nifH promoter decreases NifA-mediated activation to a greater extent than does the introduction of full helical turns. Reducing the spacing between the upstream and downstream elements of the nifH promoter also results in a promoter down phenotype. Introduction of a tight protein-binding site, the lac operator, between the upstream and downstream promoter elements did not render activation of the nifH promoter sensitive to occupancy of this site by the lac repressor. These findings indicate that NifA-mediated activation of transcription requires that NifA is bound upstream, and to the correct face of the DNA helix, in order to interact with downstream transcription factors. This implies that the interaction is brought about by the formation of a DNA loop between upstream and downstream promoter elements rather than by NifA sliding downstream.

Interaction between the cyclic AMP receptor protein and DNA. Conformational studies

The binding of the cyclic adenosine 3',5' monophosphate receptor protein (CRP or CAP) of Escherichia coli to non-specific DNA and to a specific lac recognition sequence has been investigated by circular dichroism (c.d.) spectroscopy. The effect of cAMP and cGMP on the co-operative non-specific binding was also studied. For the non-specific binding in the absence of cAMP a c.d. change (decrease of the intensity of the positive band with a shift of its maximum to longer wavelength) indicates that the DNA undergoes a conformational change upon CRP binding. This change might reflect the formation of the solenoidal coil previously observed by electron microscopy. The amplitude of the c.d. change increases linearly with the degree of saturation of the DNA and does not depend on the size of the clusters of CRP bound. From the variation of the c.d. effect as a function of the ionic strength, the product K omega (K, the intrinsic binding constant and omega, the co-operativity parameter) could be determined. The number of ion pairs involved in complex formation between CRP and DNA was found to be six to seven. Experiments performed with several DNAs, including the alternating polymers poly[d(A-T)] and poly[d(G-C)], demonstrated that the conformational change does not depend on the DNA sequence. However, in the presence of cAMP the c.d. spectrum of the DNA shows only a small variation upon binding CRP. In contrast, in the presence of cGMP the conformational change of the DNA is similar to that observed when non-liganded CRP binds. For the specific lac operon binding, the c.d. change is different from those observed for non-specific binding in the presence or absence of cAMP. These results emphasize the high variability of the DNA structure upon binding the same protein.

Probing the sequence-specific interaction of the cyclic AMP receptor protein with DNA by site-directed mutagenesis

Mutants in the DNA-binding helix of the cyclic AMP receptor protein (CRP), as well as mutants in a synthetic DNA-binding site derived from the sequence in the lac regulatory region, have been constructed by oligonucleotide-directed mutagenesis, and used to study the effect of selected amino acid substitutions on CRP-mediated transcriptional activity and on sequence-specific DNA binding. It has been shown that mutation of Arg-180 to Lys or Leu abolishes both CRP-mediated expression of beta-galactosidase in vivo and CRP binding of DNA as measured by immunoprecipitation. In contrast, the mutation of Arg-185 to Leu or Lys and the mutation of Lys-188 to Leu does not appear to influence these two parameters significantly. On the DNA side, both substitutions studied, namely the exchange of the G . C base pair in position 2 of the consensus T1G2T3G4A5 motif into an A . T base pair and the exchange of the A . T base pair in position 5 for a G . C base pair, abolish specific binding. Implications of these findings with respect to the present models for specific CRP-DNA recognition are discussed.

Electrostatic calculations and model-building suggest that DNA bound to CAP is sharply bent

Two observations suggest that DNA, upon binding to E. coli catabolite gene activator protein (CAP), is sharply bent by a total angle of at least 100-150 degrees: (1) The electrostatic potential field of CAP shows regions of positive potential that form a ramp on 3 sides of the protein. (2) The DNA binding site size as determined by DNA ethylation interference with binding, (Majors: "Control of the E. coli Lac Operon at the Molecular Level." Ph.D. Thesis, Harvard University, Cambridge, 1977) and by relative affinities of DNA fragments of various lengths (Liu-Johnson et al.: Cell 47:995-1005, 1986) requires severe bending of the DNA to maintain its favorable electrostatic contact with the protein.

Bending of promoter DNA on binding of heat shock transcription factor

The Drosophila heat-shock transcription factor (HSTF) has been shown to bind to three domains of the heat shock protein 70 gene (hsp 70) control region. The most critical of these for transcriptional activation appears to be the one closest to the TATA-homology region. This domain, spanning sequences from -40 to -95, consists of two contiguous HSTF binding sites (sites 1 and 2) that are occupied in a cooperative manner (see Fig. 1). Recent alkylation interference and protection studies suggest a conformational change occurs in the protein-DNA complex at site 1 upon sequential HSTF binding at site 2 (ref. 5). We report here that HSTF binding to a single site or to both contiguous sites results in the introduction of a specific DNA bend within this domain of the hsp 70 promoter.

The catabolite activator protein stabilizes its binding site in the E. coli lactose promoter

The effect of catabolite activator protein, CAP, on the thermal stability of DNA was examined. Site specific binding was studied with a 62 bp DNA restriction fragment containing the primary CAP site of the E. coli lactose (lac) promoter. A 144 bp DNA containing the lac promoter region and a 234 bp DNA from the pBR322 plasmid provided other DNA sites. Thermal denaturation of protein-DNA complexes was carried out in a low ionic strength solvent with 40% dimethyl sulfoxide, DMSO. In this solvent free DNA denatured below the denaturation temperature of CAP. The temperature stability of CAP for site specific binding was monitored using an acrylamide gel electrophoresis assay. Results show that both specific and non-specific CAP binding stabilize duplex DNA. Site specific binding to the 62 bp DNA produced a 13.3 degrees C increase in the transition under conditions where non-specific binding stabilized this DNA by 2-3 degrees C.

[Mode of action of cyclic amp in prokaryotes and eukaryotes, CAP and cAMP-dependent protein kinases]

cAMP is an ubiquitous compound which is involved in the regulation of many biological processes. In bacteria such as E. coli, cAMP mediates the activation of catabolic operons via the CAP protein. The CAP-cAMP complex, whose tridimensional structure has recently been established, binds to the promoter regions of catabolic operons at a specific site, and activates their transcription by inducing RNA polymerase to bind and initiate transcription at the correct site. Various phenomenons including protein-protein interactions or CAP-induced DNA bending or kinking could be involved in the process of forming the open transcription complex. In eukaryotes, cAMP activates cAMP dependent protein kinases which covalently modify proteins by phosphorylation on serine or threonine residues. The catalytically inactive holoenzyme is generally a tetramer containing two regulatory subunits, each capable of binding two molecules of cAMP, and two catalytic subunits. In mammalian cells, two types of cAMP dependent protein kinases (I and II) can be distinguished on the basis of their regulatory subunits; their relative proportion varies from tissue to tissue. Binding of cAMP to the regulatory subunits induces the dissociation of the holoenzyme and releases the free and active catalytic subunits. Phosphorylation of proteins occurs at sequences containing two basic residues in the vicinity of the phosphorylated serine or threonine. A heat-stable protein, present in most eukaryotic cells, specifically interacts with the catalytic subunit and inhibits its activity. The amino-acid sequence of cAMP dependent protein kinases has recently been determined. It is interesting to note that the domains responsible for cAMP binding by the regulatory subunits of mammalian cAMP dependent protein kinases and CAP share important sequence homologies. The same phenomenon is observed concerning the domain responsible for ATP binding to the catalytic subunit of cAMP dependent protein kinases and that of tyrosine-specific protein kinases from oncoviruses. Other eukaryotic proteins such as S-adenosyl-L-homocysteine (SAH) hydrolase are also capable of binding cAMP. The latter is involved in the regulation of S-adenosyl-L-methionine dependent methylations, and its activity could be affected by cAMP. Besides its role as an effector of enzymatic activity via phosphorylation, such as in the regulation of glycogen metabolism, cAMP has recently been shown to activate the transcription of a number of eukaryotic genes. This process probably also involves protein phosphorylation, but its precise mechanism remains to be understood.