Identification of the Subunit of cAMP Receptor Protein (CRP) That Functionally Interacts with CytR in CRP-CytR-mediated Transcriptional Repression*

At promoters of the Escherichia coli CytR regulon, the cAMP receptor protein (CRP) interacts with the repressor CytR to form transcriptionally inactive CRP-CytR-promoter or (CRP) 2 -CytR-promoter complexes. Here, us- ing “oriented heterodimer” analysis, we show that only one subunit of the CRP dimer, the subunit proximal to CytR, functionally interacts with CytR in CRP-CytR-promoter and (CRP) 2 -CytR-promoter complexes. Our re- sults provide information about the architecture of CRP-CytR-promoter and (CRP) 2 -CytR-promoter com- plexes and rule out the proposal that masking of activating region 2 of CRP is responsible for the transcriptional inactivity of the complexes. The and also to

The Escherichia coli cAMP receptor protein (CRP 1 ; also referred to as catabolite gene activator protein, or CAP) is a global transcriptional regulator. CRP serves as a transcriptional activator at numerous promoters and also can serve as a transcriptional co-activator, repressor, or co-repressor (reviewed in Ref. 1). CRP binds to a 22-bp 2-fold-symmetric DNA site located in or near each CRP-regulated promoter. CRP is a dimer of two identical subunits, each consisting of 209 amino acids. The crystallographic structure of the CRP-DNA complex shows that one subunit of CRP interacts with one half of the DNA site, and the other subunit of CRP interacts with the other half of the DNA site, in an approximately 2-fold-symmetric fashion (2,3).
Activation of transcription at simple CRP-dependent promoters (i.e. promoters that require only CRP for activation) involves direct protein-protein contact between CRP and RNA polymerase (RNAP). At promoters that have a DNA site for CRP located upstream of the DNA site for RNAP (class I CRP-dependent promoters), a determinant consisting of amino acids 156 -164 of CRP ("activating region 1" (AR1)) makes direct contact with the RNAP ␣-subunit C-terminal domain (␣CTD) (1,4,5). At CRP-dependent promoters that have the DNA site for CRP centered near position Ϫ42 (class II CRP-dependent promoters), AR1 interacts with the ␣CTD, and a second determinant, consisting of amino acids 19, 21, and 101 of CRP ("activating region 2" (AR2)), interacts with the RNAP ␣-subunit N-terminal domain (␣NTD) (1,6,7).
"Oriented heterodimer" analysis (in which CRP heterodimers having one subunit with a mutant activating region and one subunit with a wild-type activating region are constructed, oriented on promoter DNA, and analyzed) has defined, for each class of CRP-dependent promoter, which subunit of the CRP dimer functionally presents each activating region (7)(8)(9)(10). At class I CRP-dependent promoters, AR1 is functionally presented by the promoter-proximal subunit of the CRP dimer (9,10). At class II CRP-dependent promoters, AR1 is functionally presented by the promoter-distal subunit, and AR2 is functionally presented by the promoter-proximal subunit (7,8,10).
Promoters of the CytR regulon contain DNA sites for both CRP and the repressor CytR (11). At these promoters, CRP interacts alternatively, and mutually exclusively, with RNAP or with CytR (12). Under conditions in which CytR is inactive (in the presence of cytidine), CRP interacts with RNAP to form transcriptionally active CRP-RNAP-promoter complexes, making the same interactions as at the above-described class I and class II CRP-dependent promoters. Under conditions in which CytR is active (in the absence of cytidine), CRP interacts with CytR to form transcriptionally inactive CRP-CytR-promoter complexes.
The mechanism by which formation of CRP-CytR-promoter and (CRP) 2 -CytR-promoter complexes results in transcrip-tional repression remains to be established. Three (not mutually exclusive) models seem possible. (i) CytR may interfere with interaction between AR1 of CRP and ␣CTD of RNAP (by occupying the DNA segment adjacent to CRP, preventing access of ␣CTD to this DNA segment). (ii) CytR may interfere  with interaction between AR2 of CRP and ␣NTD of RNAP (by interacting with a determinant on CRP immediately adjacent to AR2, preventing access of ␣NTD to AR2). (iii) CytR may interfere with interaction between RNAP and the core promoter. As a first step to distinguish among these models, we have used oriented heterodimer analysis (9) to determine which subunit of the CRP dimer functionally interacts with CytR in CRP-CytR-mediated transcriptional repression and to relate the identity of this subunit to the identities of the subunits that functionally present AR1 and AR2 in CRP-mediated transcriptional activation.

EXPERIMENTAL PROCEDURES
Strains-E. coli strains used are listed in Table I. Strain SØ2929crpE181V was constructed by replacing the SalI-EcoRI bla segment of plasmid pTAC3590 (24) with the SalI-EcoRI crpE181V segment of plasmid pKLP4C-E181V (constructed as in Ref. 19), introducing the resulting minicircles into strain SØ2929/pBHK491, and isolating Km R Ap S transformants at 42°C (method of Refs. 24 and 25).
Proteins-CRP and CRP derivatives were prepared using cAMP affinity chromatography (27). CytR proteins were prepared as described (17,23). RNAP was purified by metal ion affinity chromatography followed by ion exchange chromatography (7).
In Vivo Transcription Assays-Cells were grown in AB minimal medium (28) containing 2 g/ml thiamine, 0.5% glycerol, 0.05% casamino acids, 25 g/ml spectinomycin, and 25 g/ml ampicillin. For experiments with CytR⌬9 -49, also 10 g/ml chloramphenicol and 10 M isopropyl-␤-D-thiogalactopyranoside were included in the growth medium (Table III and Fig. 3c). Samples of exponential phase cultures (200 l) were transferred to microtiter plates, and A 450 was measured using an iEMS Reader (Labsystems). Aliquots (100 l) were transferred to new microtiter plate wells containing 50 l of permeabilization buffer (100 mM Tris-HCl, pH 7.8, 32 mM sodium phosphate, 8 mM dithiothreitol, 8 mM CDTA, 200 g/ml polymyxin B, and 4% Triton X-100), and cells were lysed by incubating for 1 h at 30°C. Reactions were initiated by the addition of 50 l of 4 mg/ml o-nitrophenyl-␤-D-galactoside, and absorbance at 414 nm was monitored as a function of time (maintaining temperature at 30°C). ␤-Galactosidase activities were determined from plots of absorbance versus time as follows: activity ϭ 1000 ϫ ((slope/ fraction of cells in sample)/A 450 ).
Repression is expressed as decrease in activity (in percent) by the addition of CytR or CytR⌬9 -49 and is calculated as follows: 100 ϫ (activity ϪCytR Ϫ activity ϩCytR )/activity ϪCytR .
In Vitro Transcription Assays-CRP subunit exchange was carried out as described (9). Subunit exchange reaction mixtures (190 l) contained the following: The reaction product [␣-32 P]ApUpU was resolved by paper chromatography in water/saturated ammonium sulfate/isopropyl alcohol (18:80:2, v/v/v) and was quantified by Cerenkov counting or Phos-phorImager analysis (Molecular Dynamics, Inc., Sunnyvale, CA). Data represent means of four independent measurements. Repression is expressed as decrease in product (in percent) by the addition of CytR or CytR⌬9 -49 and is calculated as follows: 100 ϫ (product ϪCytR Ϫ product ϩCytR )/product ϪCytR .

Construction of Promoter Derivatives
To analyze CRP-CytR-promoter complex formation, we constructed a set of three promoter derivatives, starting from the well characterized class II CRP-dependent promoter CC(Ϫ41.5) (1, 6, 26) (Fig. 1a). Each of the three promoter derivatives has a DNA site for CRP centered at position Ϫ41.5 and a consensus DNA site for CytR (29,30) centered at position Ϫ63.5. The first has a DNA site for CRP with two consensus half-sites (O-CC); FIG. 1. Promoter sequences. a, promoters with one DNA site for CRP and one DNA site for CytR. Sequences from ϩ118 to Ϫ49 are derived from the CC(Ϫ41.5) promoter (26). The sequence from Ϫ50 to Ϫ72 is from the deoP2 isolate C13-32 (30). The CytR operator is underlined, and the DNA half-sites for CRP are shown in boldface letters. b, promoters with two DNA sites for CRP and one DNA site for CytR. Sequences from ϩ118 to Ϫ72 are as in a. A consensus DNA site for CRP is centered at position Ϫ94.5 to enable (CRP) 2 -CytR-promoter complex formation. To analyze (CRP) 2 -CytR-promoter complex formation, we constructed an analogous set of three promoter derivatives, each having DNA sites for CRP centered at positions Ϫ94.5 and Ϫ41.5 and a consensus DNA site for CytR (29,30) (Table III and  Oriented Heterodimers CRP-CytR-Promoter Complex Formation-To determine which subunit of the CRP dimer functionally interacts with CytR in CRP-CytR-promoter complex formation (i.e. to determine which subunit of the CRP dimer functionally presents the RR) we performed oriented heterodimer analysis using promoter derivatives O-XC and O-CX (Fig. 2).
We constructed and analyzed CRP heterodimers having one subunit with a functional RR and relaxed DNA-binding specificity and one subunit with a nonfunctional RR and wild-type DNA-binding specificity. For the subunit with a functional RR and relaxed DNA binding specificity, we used the [Val 181 ]CRP subunit, which has a substitution within the helix-turn-helix DNA binding motif of CRP that permits efficient recognition of both the consensus DNA half-sites (C in O-XC and O-CX) and of DNA half-sites with the nonconsensus base pair A:T at position 7 (X in O-XC and O-CX) (7-10, 31, 32). For the subunit with a nonfunctional RR and wild-type DNA binding specificity, we used the [Ala 13 ]CRP subunit (Table III 13 ]CRP subunit (and the nonfunctional RR) in the CytR-proximal DNA half-site at O-CX (Fig. 2a)). ]CRP homodimer, which served as a further control with a functional RR in each subunit; data not shown). We conclude that the CytR-proximal subunit of CRP functionally presents RR in CRP-CytR-promoter complex formation.
(CRP) 2 -CytR-Promoter Complex Formation-To determine which CRP subunits functionally interact with CytR in (CRP) 2 -CytR-promoter complex formation, we performed analogous oriented heterodimer experiments using promoter derivatives CX-O-XC and XC-O-CX (Fig. 3). Because CytR⌬9 -49 is more defective in [Ala 13 ]CRP-mediated transcriptional repression of CC-O-CC compared with wild-type CytR (Table III) proximal subunit, functionally interacts with CytR in CRP-CytR-promoter complex formation at the O-CC promoter (Fig.  2). Our results further establish that only one subunit of each CRP dimer, the CytR-proximal subunit of each CRP dimer, functionally interacts with CytR in (CRP) 2 -CytR-promoter complex formation at the CC-O-CC promoter (Fig. 3). Our results support the schematic models in Fig. 4 and structural models in Fig. 5.
Spatial Relationship between Determinants for Transcription Activation and the Determinant for CRP-CytR Interaction-In the absence of CytR, the O-CC promoter is a simple class II CRP-dependent promoter (Fig. 4a). At such a promoter, CRP activates transcription through two sets of interactions: (i) protein-protein interaction between AR1 of the upstream subunit of the CRP dimer and RNAP ␣CTD, which interacts with the DNA segment immediately upstream of the CRP dimer and (ii) protein-protein interaction between AR2 of the downstream subunit of the CRP dimer and RNAP ␣NTD (Fig. 4a; Refs. 1, 6, and 7). Our results indicate that, at such a promoter, CytR interacts with the subunit of CRP that functionally presents AR1 but does not interact with the subunit of CRP that functionally presents AR2 (compare locations of blue, red, and yellow determinants in Figs. 4, a and b, and 5a).
In the absence of CytR, the CC-O-CC promoter is a compound class I/class II CRP-dependent promoter (Fig. 4c). At such a promoter, CRP activates transcription through three sets of interactions: (i) protein-protein interaction between AR1 of the downstream subunit of the CRP dimer in the Ϫ94 region and one copy of RNAP ␣CTD, which interacts with the DNA segment immediately downstream of this CRP dimer; (ii) protein-protein interaction between AR1 of the upstream subunit of the CRP dimer in the Ϫ42 region and the other copy of RNAP ␣CTD, which interacts with the DNA segment immediately upstream of this CRP dimer; and (iii) protein-protein interaction between AR2 of the downstream subunit of the CRP dimer Each functional AR1 is shown as a red circle, each functional AR2 as a yellow pentagon, and each functional RR as a blue square (with determinants on the front and rear faces of CRP as filled and open symbols, respectively). In a and c, the functional AR2 is not visible but would located beneath the N in ␣NTD. In b and d, determinants proposed to be masked by CytR (i.e. each functional AR1 and each interaction site for ␣CTD on DNA) are marked (ϫ). ␣CTD, ␣NTD, ␤, ␤Ј, and denote, respectively, the RNAP ␣-subunit C-terminal domain, the RNAP ␣-subunit N-terminal domain, and the RNAP ␤-, ␤Ј-, and 70subunits. ␣CTD is an independently folded module and is connected to ␣NTD, and thus to the remainder of RNAP, through an unstructured, flexible linker; alternative positioning of ␣CTD is facilitated by the linker and by bending of the intervening DNA (1,4). in the Ϫ42 region and RNAP ␣NTD (1). Our results indicate that, at such a promoter, CytR interacts with the subunit of CRP that functionally presents AR1 in each CRP dimer but does not interact with the subunit of CRP that functionally presents AR2 (compare locations of blue, red, and yellow determinants in Figs. 4, c and d, and 5b).
Implications for Mechanism of Repression by CytR-Our results indicate that, at both O-CC and CC-O-CC, CytR interacts with the CRP subunit that functionally presents AR1 (Figs. 4,  b and d, and 5). Molecular modeling indicates that interaction between CytR and the CRP subunit that functionally presents AR1 is likely to interfere with interaction between AR1 and RNAP ␣CTD, both by obstructing access to AR1 and by obstructing access to the DNA segment adjacent to AR1 (Fig. 5). We propose that CytR inhibits transcription, at least in part, by interfering with interaction between AR1 and RNAP ␣CTD.
In contrast, our results indicate that CytR does not interact with the CRP subunit that functionally presents AR2 (Figs. 4,  b and d, and 5). We propose that CytR does not inhibit transcription by interfering with interactions between AR2 and RNAP ␣NTD.
Prospect-Our results provide new information about the architecture of CRP-CytR-promoter and (CRP) 2 -CytR-promoter complexes and constrain possible mechanisms for transcription repression by CytR. Important objectives for further work include determination of the subunit of CytR that functionally interacts with CRP and determination whether repression by CytR is solely due to interference with interaction between CRP and RNAP ("anti-activation") or also involves interference, directly or indirectly, with interactions between RNAP and the core promoter ("direct repression").