Effector-Repressor Interactions, Binding of a Single Effector Molecule to the Operator-bound TtgR Homodimer Mediates Derepression*

The RND family transporter TtgABC and its cognate repressor TtgR from Pseudomonas putida DOT-T1E were both shown to possess multidrug recognition properties. Structurally unrelated molecules such as chloramphenicol, butyl paraben, 1,3-dihydroxynaphthalene, and several flavonoids are substrates of TtgABC and activate pump expression by binding to the TtgR-operator complex. Isothermal titration calorimetry was employed to determine the thermodynamic parameters for the binding of these molecules to TtgR. Dissociation constants were in the range from 1 to 150 μm, the binding stoichiometry was one effector molecule per dimer of TtgR, and the process was driven by favorable enthalpy changes. Although TtgR exhibits a large multidrug binding profile, the plant-derived compounds phloretin and quercetin were shown to bind with the highest affinity (KD of around 1 μm), in contrast to other effectors (chloramphenicol and aromatic solvents) for which exhibited a more reduced affinity. Structure-function studies of effectors indicate that the presence of aromatic rings as well as hydroxyl groups are determinants for TtgR binding. The binding of TtgR to its operator DNA does not alter the protein effector profile nor the effector binding stoichiometry. Moreover, we demonstrate here for the first time that the binding of a single effector molecule to the DNA-bound TtgR homodimer induces the dissociation of the repressor-operator complex. This provides important insight into the molecular mechanism of effector-mediated derepression.

pumps have recently become the object of special interest, not only because they are a cause of concern in the emergence of multidrug resistant pathogens such as Pseudomonas aeruginosa (8), Escherichia coli (9), Stenotrophomonas maltophilia (10), or Neisseria gonorrhoeae (11,12), but also for their importance in bacterial tolerance to toxic xenobiotics like organic solvents (7). Besides the clinical and ecological relevance, the particular ability of these proteins to interact with a wide range of structurally different molecules is of fundamental interest in understanding protein-ligand specificity (13)(14)(15)(16)(17). Moreover, a number of studies have shown that the expression of some MDR transporters is controlled by transcriptional regulators with the same multidrug recognition properties, such as BmrR of Bacillus subtilis (18), EmrR of E. coli (19), QacR of Staphylococcus aureus (20), and TtgR of Pseudomonas putida DOT-T1E (21). In that sense, the recent determination of the three-dimensional structures of QacR and BmrR in complex with several ligands has provided important insight into the molecular mechanism of multiple drug recognition (22)(23)(24).
However, one of the most intriguing issues concerning bacterial MDR systems is the physiological function they exert in their natural habitat (14,25,26). Some MDR efflux pumps present in soil-living or plant-associated bacteria have been shown to confer resistance to plant defense molecules and seem to play an important role in the efficient colonization of plant-influenced microenvironments by these microorganisms and/or in their virulence, in the case of plant pathogens (27)(28)(29)(30).
P. putida is a non-pathogenic bacterium present in water and soil and able to colonize plant roots and seeds (31)(32)(33). The DOT-T1E strain was isolated for its particularly high resistance to toxic organic solvents (34) and three RND efflux pumps, TtgABC, TtgDEF, and TtgGHI, were found to be essential for this resistance (35). This study will focus on the TtgABC efflux pump and its cognate regulator TtgR. TtgABC has been shown to play an important role in the intrinsic tolerance of P. putida DOT-T1E to organic solvents (35,36). In addition, TtgABC is able to extrude several structurally different antimicrobial compounds, such as antibiotics (chloramphenicol, tetracycline, nalidixic acid, norfloxacin, streptomycin, ampicillin, and cefotaxime), biocides (triclosan), and dyes (ethidium bromide) (35,37,38), and hence confers an MDR phenotype to P. putida DOT-T1E.
TtgR is a member of the TetR family of transcriptional repressors. Typically, these repressors have two functional domains, a highly conserved N-terminal DNA binding domain, and a less conserved C-terminal domain involved in both dimerization and effector binding (39). TtgR is a multidrug binding repressor that negatively controls the transcription of the ttgABC operon as well as its own expression. TtgR operates by an effector-mediated derepression mechanism (21), which implies that effector binding to DNA-bound TtgR causes protein dissociation and subsequently transcriptional activation of ttgABC and ttgR.
As mentioned above, a number of structure-based studies on this type of multidrug binding regulators have provided important insights into ligand-induced conformational changes and protein-ligand contacts. However, only a few studies were aimed at identifying possible natural effectors for these regulators, which would help to ascribe a physiological function to a given MDR system. As P. putida is a very abundant soil bacterium found in close association with plant roots, we hypothesized that the TtgABC/TtgR system might be involved in the resistance to plant defense metabolites. To address this issue, two complementary approaches were used. Initial in vivo studies aimed at exploring the TtgABC-mediated resistance to plant defense metabolites, as well as the role of TtgR in the expression of the ttgABC operon in response to these compounds. Subsequently, an extensive in vitro characterization of TtgR binding to different plant secondary metabolites is carried out using ITC. Most importantly, the binding of different effectors to DNA-bound TtgR is also explored, which is an innovative approach we judge of general relevance for the study of transcriptional repressors.

Bacterial Strains, Plasmids, and Culture Medium
The bacterial strains and plasmids used in this study are shown in Table 1. Bacterial strains were grown in LB medium at 30°C as described before (21) or in 2ϫYT medium (40) for the production of the TtgR protein. When required, antibiotics were added to the culture medium to reach the following final concentrations: rifampicin, 20 g/ml, and tetracycline, 20 g/ml. pED14 is a low-copy number plasmid (2-4 copies per cell) derived from pMP220 (41), which carries a transcriptional fusion of the ttgA promoter to the reporter gene ЈlacZ (21).

␤-Galactosidase Assays
Bacterial cells were inoculated from fresh LB agar plates in LB liquid medium supplemented with the appropriate antibiotics and grown overnight at 30°C with shaking (200 rpm). Cultures were diluted 100fold in 15 ml of fresh LB or in the same medium supplemented with 1 mM of the inducing compounds, then incubated at 30°C with shaking until the culture reached an A 660 of 1.0 when ␤-galactosidase activity was determined according to Miller (42). All the chemicals were used at concentrations below the minimal inhibitory concentration (MIC), therefore they did not affect growth. Results presented are the mean of four different experiments.

Antimicrobial Agent Susceptibility Tests
The MIC of the tested compounds was determined in LB medium by the microtiter broth dilution method (43). Microtiter plate wells containing each 100 l of LB were inoculated with 10 5 colony-forming units/ml and incubated for 16 h. The growth was then analyzed and the MIC corresponded to the minimal concentration at which growth was inhibited by at least by 90%.

Overexpression and Purification of Native TtgR Protein
A 651-bp fragment containing the ttgR gene was amplified by PCR from P. putida DOT-T1E chromosomal DNA using the primers TtgR5ЈNdeI (5Ј-NNNNNNCATATGGTCCGTCGAACCAAAG-3Ј) and TtgR3ЈHindIII (5Ј-NNNNNNAAGCTTCTACTATTTCGCCA-GAGCCGGGCTC-3Ј), which contained NdeI and HindIII restriction sites, respectively (underlined), and the latest included also the ttgR stop codon (in bold). The PCR products were restricted with the above enzymes and subsequently ligated into the pET29a(ϩ) vector (Novagen) previously digested with the same enzymes. The resulting plasmid was termed pWilR2.
For the purification of native TtgR, pWilR2 was transformed into E. coli B834(DE3). The cells were grown at 30°C in 2-liter Erlenmeyer flasks containing 1 liter of 2ϫYT culture medium (40) supplemented with 50 g/ml Km. For TtgR production, protein expression was induced at an A 660 of 0.5-0.6 by adding 1 mM isopropyl ␤-D-1-thiogalactopyranoside. Cells were grown for another 3 h at 20°C and subsequently harvested by centrifugation. The pellet resulting from a 1-liter culture was resuspended in 30 ml of buffer A (10 mM Tris-HCl, pH 6.4, 20 mM NaCl, 5% (v/v) glycerol, 0.1 mM EDTA, and protease inhibitor mixture (Complete, Roche)) and broken by treatment with 20 g/ml of lysozyme and French press. Following centrifugation at 13,000 ϫ g for 40 min, the TtgR protein was predominantly present (more than 80%) in the soluble fraction. The supernatant was loaded onto a Hitrap Heparin HP column (5 ml, Amersham Biosciences) equilibrated with buffer A and eluted with a gradient of 0.02-0.5 M NaCl. Fractions containing TtgR were pooled, concentrated to 4 ml by ultrafiltration using an Amicon 8050 apparatus (Amicon-Millipore), and dialyzed against buffer B (20 mM Tris-HCl, pH 7.5, NaCl 250 mM, 10% (v/v) glycerol, 0.1 mM EDTA, and 1 mM DTT). The sample was then submitted to size exclusion chromatography using a Sephacryl HR-200 column (Amersham Biosciences). Eluted fractions of TtgR were pooled, concentrated using the above equipment, and dialyzed against buffer C (10 mM Tris-HCl, pH 7.5, 250 mM NaCl, 50% glycerol (v/v), 0.1 mM EDTA, and 2 mM DTT) for protein storage at Ϫ70°C (long term storage) or Ϫ20°C (short term storage). Protein concentrations were determined using the Bio-Rad Protein Assay kit.

Electrophoretic Mobility Shift Assay
Electrophoretic mobility shift assays were carried out as described previously (21). The DNA probe was a 189-bp fragment containing the ttgABC-ttgR intergenic region obtained from P. putida DOT-T1E chromosomal DNA by PCR. 1 nM of the radiolabeled probe (ϳ10,000 cpm) was incubated with 0.75 M purified TtgR in 10 l of DNA binding buffer (10 mM Tris-HCl, pH 7.0, 250 mM NaCl, 10 mM Mg acetate, 10 mM KCl, 5% glycerol (v/v), 0.1 mM EDTA, and 1 mM DTT) supplemented with 20 g/ml poly(dI-dC) and 200 g/ml bovine serum albumin. Effectors (prepared in Me 2 SO) were added to the binding reaction at a final concentration of 1 mM. Reactions were incubated for 10 min at 30°C and samples were run on 4.5% (w/v) non-denaturing polyacrylamide gels (Bio-Rad Mini-Protean II) for 2 h at 50 V at room temperature in Tris glycine buffer (25 mM Tris-HCl, pH 8.0, and 200 mM glycine). The resulting gels were analyzed using a Personal FX equipment and QuantityOne software (Bio-Rad).

ITC
Effector Binding to TtgR-Measurements were performed on a VPmicrocalorimeter (MicroCal, Northampton, MA) at 30°C. This instrument is equipped with a 1.437-ml sample cell and a 300-l syringe. The protein was thoroughly dialyzed against effector binding buffer (25 mM Pipes, pH 7.0, 250 mM NaCl, 5% (v/v) glycerol, 10 mM Mg acetate, 10 mM KCl, 0.1 mM EDTA, and 1 mM DTT). The protein concentration was determined using the Bradford assay. Stock solutions of 1-naphthol, 1,3-dihydroxynaphthalene, butyl paraben, naringenin, quercetin, phloretin, apigenin, luteolin, coumestrol, and chloramphenicol were freshly prepared at 0.5 M in Me 2 SO, and subsequently diluted in dialysis buffer. The appropriate amount of Me 2 SO (0.1-0.3% (v/v)) was added to the protein sample in each assay. When indicated, analyses were carried out using 5% (v/v) Me 2 SO to facilitate the effector solubilization. All chemicals were manipulated in glass vessels and effector samples were neither degassed nor filtered, to avoid evaporation or nonspecific binding. Typically, an experiment involved a single 2 l and a series of 4-l injections of effector molecules into the protein solution.
Effector Binding to TtgR-DNA Complex-A 40-bp DNA fragment containing the TtgR operator was used for these assays. It was obtained by assembling two complementary oligonucleotides (5Ј-CAGCAG-TATTTACAAACAACCATGAATGTAAGTATATTCC-3Ј and its complementary oligonucleotide). Annealing was carried out by mixing equimolar amounts (60 M) of each oligonucleotide in annealing buffer (10 mM Tris-HCl, pH 8.0, and 0.5 mM MgCl 2 ). The mixture was incubated at 95°C for 5 min, followed by slow cooling to room temperature. Both, DNA and TtgR protein were subsequently dialyzed in DNA binding buffer (10 mM Tris-HCl, pH 7.0, 250 mM NaCl, 5% (v/v) glycerol, 10 mM Mg acetate, 10 mM KCl, 0.1 mM EDTA, and 1 mM DTT). TtgR (typically 15-20 M) was placed into the ITC cell and titrated by the addition of the 40-bp DNA fragment. When protein saturation was reached, i.e. when measured heat changes corresponded to those observed in the titration of buffer with DNA, the experiment was stopped and the cell content adjusted to a volume of 1.437 ml. The completeness of protein saturation with DNA was verified by native polyacrylamide gel electrophoresis. Subsequently, the DNA-TtgR complex was titrated with effector molecules as described above. The titration of unbound TtgR with effector molecules was carried out on the same day using the same chemicals, exactly the same ligand concentrations and injection volumes, which guarantees that the absence/presence of DNA is the only difference between both assays.
In all cases, the mean enthalpies measured from the injection of the ligand into the buffer were subtracted from raw titration data before data analysis with ORIGIN software (MicroCal). Titration curves were fitted by a non-linear least squares method to a function for the binding of a ligand to a macromolecule (44). From the curve thus fitted, the parameters ⌬H (reaction enthalpy), K A (binding constant, K A ϭ 1/K D ), and n (reaction stoichiometry) were determined. From the values of K A and ⌬H, the changes in free energy (⌬G) and in entropy (⌬S) were calculated with the equation: ⌬G ϭ ϪRT lnK A ϭ ⌬H Ϫ T⌬S, where R is the universal molar gas constant, and T the absolute temperature.

RESULTS
In Vivo Susceptibility of P. putida DOT-T1E to Plant Antimicrobials-As mentioned above, the implication of the TtgABC/TtgR system in the organic solvent tolerance and multidrug resistance of DOT-T1E was previously reported (21,35,37). However, the particularly broad substrate specificity of this efflux system and the multidrug binding capacity of its cognate repressor TtgR raised the question concerning their physiological role(s). Based on the fact that the plant rhizosphere is one of the natural habitats of P. putida, and considering that the AcrB transporter from the plant pathogen Erwinia amylovora, involved in the resistance of this strain to various plant defense metabolites (27), is 60% identical to the TtgB transporter from P. putida DOT-T1E, we investigated the potential role of TtgABC in the resistance to plant-derived antimicrobials. To verify this hypothesis, the MICs of five plant secondary metabolites known to have antibacterial activities (45)(46)(47) were determined for both, P. putida DOT-T1E and its derivative mutant strain DOT-T1E18, deficient in the TtgABC efflux pump (Table 2). To different extents, the wild type strain was susceptible to all of the compounds tested. The susceptibility of DOT-T1E18 to phloretin, quercetin, and naringenin was significantly increased, suggesting that the TtgABC efflux pump confers resistance to these compounds. In contrast, no differences in the susceptibilities to coumestrol and berberine were detected, indicating that these compounds are probably not substrates of this pump. To corroborate this fact and given that antimicrobial activity of plant secondary metabolites is often masked by the concerted action of multiple MDR efflux pumps with similar substrate profiles (47), the MICs were determined in the presence of the MC 207110 inhibitor, known to inhibit certain RND efflux pumps (48,49). In the presence of this inhibitor, the susceptibilities of both strains to phloretin, quercetin, naringenin, and even coumestrol dramatically increased. Interestingly, the DOT-T1E18 mutant still showed increased susceptibilities to these compounds as compared with its parental strain. This indicates, first, that coumestrol is, in addition to phloretin, quercetin, and naringenin, also a substrate of the TtgABC pump and, second, that the TtgABC efflux pump is not (or not completely) inhibited by MC 207110 . This hence suggests the existence of other pump(s) that are impaired by MC 207110 and expel the same compounds, accounting for the high level of resistance of P. putida DOT-T1E to these natural antimicrobials. Furthermore, the low MIC values obtained for coumestrol (0.78 g/ml) or quercetin (4.2 g/ml) are comparable with those found for clinically relevant antibiotics (35,37), which points to the antibacterial potential of these natural compounds against Pseudomonas sp. strains when used in combination with MDR inhibitors. No MIC variation was detected for berberine, indicating that this plant alkaloid is not a TtgABC substrate. Taken together, these data confirm the role of this MDR pump in the high resistance of P. putida DOT-T1E to several plant antimicrobials present in the rhizosphere.

In Vivo Expression of the TtgABC Efflux Pump in Response to Plant Secondary Metabolites and Other Aromatic Compounds-Because
TtgABC was able to expel several plant secondary metabolites, it was hypothesized that these and related aromatic compounds might be able to alter the in vivo expression of the ttgABC operon through its cognate repressor TtgR. A transcriptional fusion of the ttgA promoter to the reporter gene lacZ (pED14 plasmid) was used for this purpose (Fig. 1). Promoter activity was determined both in the wild type (DOT-T1E) and in a TtgR-deficient background (DOT-T1E13). In the wild type, a significant increase of ttgABC expression was obtained in the presence of all the flavonoids tested (quercetin, luteolin, apigenin, naringenin, and genistein), as well as coumestrol (1.5-fold increase) and phloretin, which was the most potent inducer (2.4-fold increase). Other aromatic compounds, such as 1-naphthol and butyl paraben, also promoted a significant increase in efflux pump expression. Interestingly, berberine and naphthalene had no significant effect on ttgABC expression. Chloramphenicol, a known inducer of the ttgABC genes and an effector of the TtgR repressor (21), was used as a positive control. In the mutant DOT-T1E13 strain, no significant increase in gene expression was detected with any of the compounds tested, indicating that in the wild type background, TtgR mediates the increase of ttgABC expression in response to those compounds.
In Vitro Effect of Plant Signal Metabolites on TtgR-Operator Binding-As described previously, the ttgABC operon is expressed at a certain basal level that increases in the presence of antibiotics, such as chloramphenicol or tetracycline. These compounds were shown to promote the dissociation of TtgR from its operator (21). To verify whether the plant secondary metabolites used above exert the same effect, electrophoretic mobility shift assays were carried out with a DNA fragment bearing the TtgR operator, as previously described (21). Experimental conditions were optimized (0.75 M TtgR and 1 mM effector) to permit a semi-quantitative comparison of the effector-mediated derepression (Fig. 2). Plant metabolites phloretin, luteolin, naringenin, genistein, and quercetin, which have significant in vivo effects either on bacterial growth inhibition ( Table 2) or gene expression (Fig. 1), were able to dissociate TtgR from its operator DNA. Interestingly, the effect was more pronounced with phloretin, quercetin, and luteolin than with chloramphenicol, which was used as reference. In the presence of the former compounds an almost complete dissociation of the repressor-operator complex was observed. In contrast, berberine was unable to dissociate TtgR from its operator, which is consistent with the absence of in vivo induction in the presence of this plant alkaloid. Other aromatic compounds that enhanced ttgABC expression in vivo (see above) were also tested in vitro. 1-naphthol and butyl paraben were able to partially release TtgR from its operator, as expected from the in vivo ttgABC expression profile (Fig. 1), whereas no effect was obtained in the case of the organic solvent styrene. Thus, these results confirm the multiple effector binding capacity of TtgR, as it seems to recognize not only different classes of antibiotics, but also aromatic solvents and plant secondary metabolites of very different size and structure.
Calorimetric Analysis of the Interaction of TtgR with Multiple Effectors-The thermodynamic parameters for the interaction of TtgR with a wide range of structurally different effectors were determined by ITC. Initial experiments involved the titration of TtgR with ligands belonging to three different classes: aromatic solvents (styrene, butyl paraben, 1-naphthol, 1,3-dihydroxynaphthalene), antibiotics (chloramphenicol), and plant antimicrobials (phloretin, quercetin, berberine, and naringenin). Typical data exemplified for phloretin and naringenin are shown in Fig. 3. ITC data were analyzed using the "one set of sites" algorithm and satisfactory curve fits were obtained. The corresponding thermodynamic parameters are shown in Table 3. In all cases the titration was characterized by exothermal heat changes, indicating favorable enthalpy changes. With the exception of butyl paraben, entropy changes were unfavorable, which is generally attributed to a loss of conformational flexibility of the protein upon effector binding (50). Dissociation constants were found to be between 1 and 150 M, which are values similar to those obtained for the effectors of other multidrug binding transcriptional regulators (22,23). Phloretin and quercetin were found to bind with the highest affinity with K D values close to 1 M ( Table 3). The titration profiles of these effectors were sigmoidal (exemplified for phloretin in Fig. 3A), thus allowing a precise determination of the effector binding stoichiometry, which was found to be one effector molecule for TtgR dimer. Naringenin, chloramphenicol, and butyl paraben exhibited an intermediate affinity ranging between 10 and 20 M, giving rise to hyperbolic titration profiles (exemplified for naringenin in Fig. 3B). It should be noted here that the curve shape (hyperbolic or sigmoidal) is determined solely by the ligand concentration in the ITC cell and the binding affinity. Both, high ligand concentrations and high affinity, favor sigmoidal curves. Therefore no conclusions concerning the binding mechanism can be draw from differences in the curve shape. The aromatic solvents 1-naphthol and 1,3-dihydroxynaphthalene had only modest affinities (around 150 and 60 M, respectively). It should be noted that the additional hydroxyl group of 1,3-dihydroxynaphthalene as compared with 1-naphthol, led to a 2-fold increase in both the binding affinity to TtgR and the enthalpy term (Table 3), which is consistent with a direct recognition of the hydroxyl group at position 3 of 1,3dihydroxynaphthalene by TtgR.   MARCH 17, 2006 • VOLUME 281 • NUMBER 11

JOURNAL OF BIOLOGICAL CHEMISTRY 7105
Structure-Function Studies-Although the TtgR effector profile was found to be large, the absence of styrene and berberine binding as well as important differences in affinity for the variety of effectors tested prompted subsequent studies that were aimed at exploring the structure-function relationship of TtgR effectors. Compounds selected for analysis were plant secondary metabolites that have a structure related to those of phloretin and naringenin, shown to bind tightly in the initial studies (Table 3). However, an obstacle encountered in this series of ITC assays was the reduced solubility of some of the compounds that did not permit their analysis by ITC using the initial buffer system. This problem was overcome by the addition of 5% Me 2 SO to the buffer, an approach typically used if the affinity of poorly soluble compounds is to be determined by ITC (51).
Initial experiments involved the analysis of the binding of phloretin and naringenin to TtgR in the Me 2 SO containing buffer system (Table  4) and a comparison of the obtained thermodynamic parameters with those obtained in the initial buffer system (Table 3, naringenin and phloretin were the only two effectors with a sufficient solubility to be analyzed in the absence of Me 2 SO). In the presence of Me 2 SO, favorable entropic and enthalpic contributions were observed, whereas the binding was strongly enthalpy driven and counterbalanced by unfavorable entropy in the absence of Me 2 SO. However, for both phloretin and naringenin, enthalpy changes as well as binding constants decreased proportionally following the inclusion of Me 2 SO, which justified the use of this solvent to compare thermodynamic parameters of different effectors.
The thermodynamic mode of binding of all effectors was found to be similar. In all cases binding was driven by moderately favorable enthalpy and entropy changes (Fig. 4, Table 4). As illustrated in Fig. 4, naringenin, luteolin, phloretin, and genistein can be compared with apigenin, which represents a core structure. The most dramatic change in binding parameters, a 4-fold decrease in binding affinity, was observed for the transition of apigenin to naringenin, characterized solely by the reduction of a double bond that makes the molecule non-planar and allows free rotation of the C-ring. The increased flexibility of phloretin, as a consequence of the B-ring opening, was accompanied by a slight increase in the binding affinity (K D around 11 M). Comparing the flavone apigenin with its isomer the isoflavone genistein generated an almost 2-fold decrease in affinity. Only modest changes in binding affinity were observed for coumestrol. However, it is important to note that the additional presence of a hydroxyl group of luteolin with respect to apigenin resulted in a 2-fold increase of the binding constant.
It can be concluded that TtgR recognizes a wide range of plant secondary metabolites with physiologically relevant affinities. Nevertheless, two major determinants in the effector molecule seem to be essential for TtgR recognition, i.e. the presence of at least one aromatic ring, which reflects the hydrophobic nature of the compound, and the presence of hydroxyl groups. The latter appears to be a major feature  because the highest binding affinities were observed for the only two effectors (luteolin and phloretin) with four hydroxyl groups ( Fig. 4 and Table 5). Calorimetric Analysis of the Effector Interaction with TtgR-DNA Complex-As reported previously (21), TtgR represses the ttgABC operon by binding to its operator site located in the ttgR-ttgA intergenic region. Transcriptional activation is achieved by an effector-mediated dissociation of DNA-bound TtgR, as shown in Fig. 2. ITC has thus been employed to study the binding of effector molecules to the DNA-bound form of TtgR to compare effector binding parameters to those obtained with the free protein. Initial experiments intended to find a buffer in which the protein-DNA complex was soluble and stable. The optimal buffer (see "Experimental Procedures") slightly differed from the one used above (Tables 3 and 4). The DNA-TtgR complex (generated as detailed under "Experimental Procedures") was titrated with the four effector molecules with the highest affinity for the unbound protein, namely phloretin, quercetin, naringenin, and chloramphenicol (Table  3). No effector binding to free DNA was observed in control experiments (data not shown).
Titrations of free and DNA bound TtgR with phloretin showed that peak widths of both raw data traces were comparable (Fig. 5). Moreover, aliquots of the ITC cell content were taken before and after the experiments were analyzed by native polyacrylamide gel electrophoresis, which confirmed the absence of free protein at the beginning of the experiment and the absence of DNA-bound protein at the end of the experiment (not shown). These observations are consistent with an immediate dissociation of the protein following effector binding. Under those circumstances, observed enthalpy changes for the effector binding to the protein-DNA complex correspond to heat changes because of effector binding and protein dissociation. Fig. 5 demonstrates unequivocally that the binding of TtgR to its operator DNA does not alter the effector binding stoichiometry of 1 effector molecule per TtgR dimer.
Most interestingly, phloretin was found to bind to free TtgR and to the DNA-bound protein with an affinity of around 1 and 0.5 M, respectively, which was accompanied by a slight increase in the enthalpy change for the binding to the DNA-protein complex ( Table 5). A similar increase in affinity and the negativity of the enthalpy change was observed for quercetin (Table 5). However, this behavior was not a general feature for all effectors, because naringenin bound with comparable affinities to free and DNA-bound protein, whereas chloramphenicol had a slightly reduced affinity for the TtgR-DNA complex (Table 5).

DISCUSSION
The TtgABC efflux pump and its cognate transcriptional regulator TtgR were previously characterized in the framework of both organic solvent tolerance and MDR phenotypes of P. putida DOT-T1E (35)(36)(37). The ttgABC operon showed a relatively high basal level of expression, which was not significantly altered by the presence of organic solvents (37), but was enhanced in response to some of the antibiotics expelled by the pump (21). These two facts raised the question concerning the physiological function of the TtgABC/TtgR system.
Plant defense metabolites such as certain flavonoids (quercetin, genistein) were previously shown to possess antibacterial activity (45)(46)(47), which is because of the inhibition of DNA gyrases (52). In the present work, we demonstrate that this system is involved in the resistance to plant antimicrobials like phloretin, quercetin, naringenin, or coumestrol. Therefore, the TtgABC efflux pump may be a critical element in the competitive colonization of plant roots by P. putida, similarly to other MDR transporters found in soil-and plant-associated bacteria (28 -30). In addition, the fact that a mutant E. amylovora deficient in the MDR efflux pump AcrAB (60% identical to TtgAB) was impaired in its virulence against plants (27), emphasizes our hypothesis.
In vitro results demonstrate that the ttgABC-specific regulator TtgR recognizes a number of plant defense metabolites with high affinity. These compounds were shown to promote the dissociation of the TtgRoperator complex and thus appear to be among the effectors that possess a biologically relevant function for the survival of the bacterium in its natural habitat. ITC assays revealed that TtgR binds to multiple structurally unrelated molecules. Dissociation constants obtained ranged from 1 to 150 M, which are biologically relevant and similar to those obtained for other multiple effector binding repressors of bacterial transporters as QacR (23,53), EmrR (19), or TtgV (54). Although this confirms the particularly large multidrug binding profile of the TtgR repressor, it is interesting to note that the effectors with the high-  Table 4.
est affinity (K D values of around 1 M) were plant-derived molecules (such as phloretin and quercetin), in contrast to the other classes of effectors (chloramphenicol and aromatic solvents), which showed a more reduced affinity. It is worth noting that binding affinity of the antibacterial quercetin is almost 15-fold higher for TtgR than for its target, the bacterial DNA gyrase (K D of around 15 M; Ref. 52). This indicates that the induction of TtgABC occurs at low effector concentrations and acts as a preventive defense mechanism prior to the set-in of the antibacterial activity of the compound. Furthermore, for phloretin and quercetin a direct correlation between their effector binding affinities (Table 3), their capacity to induce the operon in vivo (Fig. 1), and their potential to dissociate the TtgR-operator complex in vitro (Fig. 2) was observed. However, butyl paraben, 1-naphthol, or naringenin were effectors of modest affinity but with in vitro and in vivo effects comparable with those observed for phloretin and quercetin. Naringenin, for example, was shown to cause one of the most pronounced increases in gene expression (Fig. 1), but bound to TtgR with an affinity around 10-fold lower than the most efficient effector (Fig. 3, Table 3). This suggests that effector binding affinity can be critical for TtgR-mediated derepression, but cannot be considered the only factor conditioning its efficiency, as it has been already described for other transcriptional repressors (55).
Analysis of the effector structures suggests that the presence of hydroxyl groups and, at least, one aromatic ring in the effector molecule are important for binding to TtgR (see Table 3 and Fig. 4). A similar duality (complementary between hydrophobic and polar interactions) was observed for the binding of effector molecules to other multidrug binding regulators such as BmrR (22) and QacR (23,24) and might represent a general feature for multidrug recognition by this class of proteins.
ITC studies demonstrate unequivocally that one effector molecule binds to a TtgR dimer (Fig. 3), which is the same stoichiometry reported for QacR (23). In general, the typical and by far the most common binding stoichiometry of small ligands to a protein homodimer is 2 effector molecules per dimer, which is also the stoichiometry described for the TetR regulator with tetracycline (56). The 1:2 stoichiometry observed for the binding of effector molecules to TtgR can thus be regarded as unusual and raises the question whether a similar stoichiometry is observed when the protein is bound to the DNA, which is an essential aspect in understanding the molecular mechanism of effectormediated derepression. ITC assays revealed that one effector molecule bound to the operator-TtgR dimer complex (Fig. 5). Mechanistically, this result indicates that the entry of a single effector molecule into the binding pocket of one of the monomers in the TtgR dimer bound to its operator is sufficient to promote the dissociation of the TtgR-operator complex. The analysis of effector binding to the DNA-bound form of TtgR presented in this work constitutes a novel approach for the functional study of a transcriptional repressor, because it reproduces more closely in vitro the biological process occurring in vivo, namely the binding of an effector molecule to DNA-bound protein. This aspect in the study of transcriptional repressors is yet relatively unexplored, as the majority of the biochemical analyses of repressor-effector interactions are carried out with free protein presuming that those interactions are identical when repressors are bound to their operators. Whereas this can be true for some systems, as we have already reported for the TtgV repressor in a previous analysis (54), our present results establish that significant differences exist between effector binding to free and DNAbound protein (Table 5). Interestingly, for phloretin and quercetin, the two effectors found to bind most tightly to the free protein (Table 3), an increase of 2.1-and 2.5-fold in affinity was observed when TtgR was bound to the DNA (Table 5). In contrast, chloramphenicol bound with lower affinity to the protein-DNA complex than to the free protein (Table 5). Differences in repressor-effector interaction caused by DNA binding have been described recently for the VceR regulator of Vibrio cholerae, another member of the TetR family (57). Authors found that

TABLE 5
Thermodynamic parameters derived from the calorimetric titration of free TtgR and TtgR bound to its operator with different effectors the affinity of VceR for its effector (carbonyl cyanide m-chlorophenylhydrazone) decreased when bound to DNA, and proposed an equilibrium model in which effector and DNA compete for VceR in a mutually exclusive manner. However, this model is not applicable to the system under study because clear evidence has been obtained for the binding of effector to DNA-bound protein.
The effector specificity of TtgR could be the result of an evolutionary process during that P. putida has been exposed to a variety of compounds in the plant rhizosphere, where phloretin and quercetin are abundant plant secondary metabolites that are synthesized and secreted by plants either constitutively or in response to pathogens or stress (58). It is thus likely that TtgR has evolved to detect these compounds at concentrations below their toxicity threshold, so that induction of the efflux pump prevents the accumulation of these antimicrobials and their deleterious effects. The optimization of this property might have been achieved by increasing the affinity of TtgR for these plant antimicrobials, particularly in its DNA-bound form. It is then suggested that multidrug-binding proteins (regulators or transporters) have evolved to acquire a specific natural function and that in some cases recognition and/or expulsion of certain drugs or xenobiotic compounds is an accidentally occurring but intrinsic side effect of these systems.