Mutations that increase the activity of the promoter of the Escherichia coli melibiose operon improve the binding of MelR, a transcription activator triggered by melibiose.

MelR is an Escherichia coli transcription factor that activates expression of the melAB operon in response to the presence of melibiose in the environment. MelR stimulates transcription initiation at the melAB promoter by binding to four sites centered at positions -120.5, -100.5, -62.5, and -42.5 upstream of the transcript start point. In a previous study, we described a spontaneous mutant that exhibited increased melAB expression. Sequence analysis showed that this mutant carries five consecutive base changes at positions -49, -50, -51, -52, and -53 upstream of the melAB transcript start. Here we show that these changes improve MelR binding to the target site centered at position -42.5 at the melAB promoter and that this improvement is responsible for increased promoter activity. Thus, the activity of the melAB promoter is fixed by the occupation by MelR of a DNA site that overlaps the -35 hexamer: MelR appears to be a typical class II-type transcription activator.

MelR is an Escherichia coli transcription factor that activates expression of the melAB operon in response to the presence of melibiose in the environment. MelR stimulates transcription initiation at the melAB promoter by binding to four sites centered at positions ؊120.5, ؊100.5, ؊62.5, and ؊42.5 upstream of the transcript start point. In a previous study, we described a spontaneous mutant that exhibited increased melAB expression. Sequence analysis showed that this mutant carries five consecutive base changes at positions ؊49, ؊50, ؊51, ؊52, and ؊53 upstream of the melAB transcript start. Here we show that these changes improve MelR binding to the target site centered at position ؊42.5 at the melAB promoter and that this improvement is responsible for increased promoter activity. Thus, the activity of the melAB promoter is fixed by the occupation by MelR of a DNA site that overlaps the ؊35 hexamer: MelR appears to be a typical class II-type transcription activator.
Melibiose metabolism by Escherichia coli is contingent on expression of the melA and melB genes, that are cotranscribed from a single promoter in response to the presence of melibiose in the growth medium: melA encodes an ␣-galactosidase, whereas melB encodes a melibiose transporter (1)(2)(3)(4)(5). An unusual feature of E. coli K12 strains is that they are able to metabolize melibiose at 30°C, but they become progressively less able to metabolize melibiose at temperatures above 37°C, due to thermal lability of the melB melibiose transporter (6). In recent work (7), we have studied the W3133-2 mutant that is able to grow at higher temperatures with melibiose as the sole carbon source. Although we had expected that the melB gene in this mutant would be altered in some way, we found that the W3133-2 mutant carried five consecutive base changes in the melAB promoter. We showed that these changes led to an increase in expression of the melAB operon at all temperatures and that this increase, apparently, could compensate for the thermolability of the melB melibiose transporter (7). The principal objective of the work presented here has been to understand the effects of the base changes in the W3133-2 mutant.
Transcription initiation at the melAB promoter is totally dependent on MelR, a melibiose-triggered transcription activator (8). MelR is encoded by the melR gene, which is located immediately upstream of the melA gene, and transcribed from a promoter that is divergent from the melAB promoter (the transcription start points of the melAB and melR promoters are 237 bases apart; see Fig. 1) (9). MelR is a member of the AraC family of transcription factors and binds to 18-base pair target sequences (10). At the melAB promoter, MelR binds to four such target sequences: site 1Ј, site 1, site 2, and site 2Ј centered at positions Ϫ120.5, Ϫ100.5, Ϫ62.5, and Ϫ42.5, respectively, upstream of the melAB transcript start point (Fig. 1). Recent studies (11) have shown that the occupation of site 2Ј by MelR is the most crucial for the activation of the melAB promoter, but that, of the four target sites, site 2Ј binds MelR most weakly. We noted that the base changes in the melAB promoter in the W3133-2 mutant changed the DNA sequence of MelR binding site 2Ј such that it corresponded better to the tighter binding site 1Ј, site 1, and site 2. This suggested that the increased melAB expression in the W3133-2 mutant could be due to better binding of MelR to site 2Ј. In the work reported here, we prove that this is indeed the case, by studying the binding of purified MelR and purified RNA polymerase holoenzyme (RNAP) 1 to DNA fragments carrying the melAB promoter from both the wild type and W3133-2 mutant strain EXPERIMENTAL PROCEDURES Subcloning the mel Operon Regulatory Region-Plasmids pBM3133, carrying the entire wild type mel operon, and pBM3133-2, carrying the mel operon from the W3133-2 mutant, were described in our previous study (7). EcoRI-HindIII fragments, carrying the wild type or mutant mel operon regulatory region, were derived using polymerase chain reaction (PCR) using pBM3133 or pBM3133-2 as templates. The KK81 fragment, carrying an EcoRI site at position Ϫ312 with respect to the melAB transcript start, and a HindIII site at ϩ36 was generated using the primers 5Ј-GCAGAATTCCGGGCTGCGGGT-3Ј and 5Ј-GCAAAGC-TTGGATGCAGGTCGACGGATCTC-3Ј (see Fig. 1). The shorter JK14 fragment, with an EcoRI site at position Ϫ94 with respect to the melAB transcript start, and a HindIII site at ϩ36, was made with the primers 5Ј-GCAGAATTCGATTTACTGCTGCTTCACG-3Ј and 5Ј-GCAAAGCTT-GGATGCAGGTCGACGGATCTC-3Ј (see Fig. 1). PCR products carrying wild type or W3133-2 mutant sequences were restricted with EcoRI and HindIII and cloned either into plasmid vectors pSR (12) or pRW50 (13).
In Vivo Expression Assays-The KK81 or JK14 EcoRI-HindIII fragments containing the wild type or mutant melAB promoters were cloned into pRW50, a low copy number lac expression vector (13), to generate pmelAB::lac fusions. Recombinant plasmids were propagated in strain WAM131 (11), an mel ϩ strain derived from GM-1 (14). ␤-Galactosidase levels in these cells were measured using the method of Miller (15): cells * This work was supported by grants from the Ministry of Education, Science and Culture of Japan and from the United Kingdom Biotechnology and Biological Sciences Research Council. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
were grown in media either with or without melibiose exactly as in our previous work (5).
Overexpression and Purification of His-tagged MelR-In the histidine-tagged variant used here (MelRH), the carboxyl terminus of the MelR is extended by 7 amino acid residues, including 6 tandem histidines; the full sequence of this extension is (in single-letter code) IHH-HHHH. To construct modified melR encoding the MelRH, we started with plasmid pBM3133 (7), in which the melibiose operon was cloned into the EcoRI site of pBR322. To insert a unique NdeI site at the initiation codon site of the melR gene and to remove the termination codon of melR, we used PCR mutagenesis using the primer melR(N-NdeI) (5Ј-ACATATGAATACAGATACGTTTATG-3Ј) and melR(C) (5Ј-GCCGGGAAACGTCTGGCG-3Ј). The amplified DNA fragment was cloned into the EcoRV site of plasmid pBF4-11, 2 which possesses a DNA sequence encoding 6 tandem histidines after the EcoRV site in pBluescript II SK(ϩ). In the final step, an NdeI-BamHI fragment from the resulting plasmid was ligated to pET30a, a T7 expression vector obtained from Novagen, and digested with NdeI and BamHI. This generated plasmid pETRH, in which the encoded protein acquired the extension of 7 amino acid residues noted above. The construction was organized such that overexpressed product was initiated at the melR AUG initiation codon. A control plasmid, pETR, carrying authentic melR, was constructed using a similar strategy with melR(BamHI) (5Ј-TGGATCCTTAGCCGGGAAACGTCTG-3Ј) as a PCR primer and pBluescript II SK(ϩ) as a vector instead of melR(C) and pBF4-11, respectively. BL21(DE3) cells (from Novagen) harboring pETRH were used to overexpress His-tagged MelR protein. Cells from an overnight culture were diluted 50-fold into fresh M9 medium supplemented with 0.4% glucose, 0.2% casamino acid, and 60 mg/liter of kanamycin (200 ml), and grown at 37°C for 3-4 h. After the optical density at 650 nm reached 0.8, addition of 1 mM isopropyl ␤-D-thiogalactopyranoside allowed induction of MelRH during the following 3-to 4-h incubation. All subsequent work was done at 4°C. After cells were harvested by centrifugation, and they were resuspended 10 ml of sonication buffer (100 mM potassium phosphate, pH 8.3, 200 mM KCl, 10% glycerol, 10 mM 2-mercaptoethanol, 1 mM EDTA) and subjected to lysis by ultrasonication (15 s ϫ 12 times with 10-s intervals). After centrifugation at 22,500 ϫ g for 20 min, inclusion body was resuspended in the same volume of the sonication buffer containing 4% Triton X-100 and recentrifuged at 22,500 ϫ g for 20 min. The inclusion body was washed twice with the sonication buffer, and finally the inclusion body was solubilized in 10 ml of solubilization buffer (100 mM potassium phosphate, 200 mM KCl, 6 M guanidine-HCl, 10% glycerol, 1 mM 2-mercaptoethanol, 1 mM EDTA, pH 8.3). After centrifugation at 22,500 ϫ g for 20 min, the supernatant fraction was kept at Ϫ20°C until use.
For protein purification, we used 0.5-ml bed-volume amounts of Ni 2ϩ -agarose (Qiagen). The resin equilibrated with the solubilization buffer was added to 9.5 ml of the solubilized protein and kept for 30 min with gentle shaking to allow for binding of MelRH. The mixture was then placed in a column, unbound material was collected by gravity elution, and the column was washed with 10 ml of the solubilization buffer (pH 8.0). After nonspecifically bound material had been removed, MelRH was eluted with the solubilization buffer (pH 6.3). The eluted protein was dialyzed against a 50-fold volume of dialysis buffer (100 mM potassium phosphate, pH 7.4, 10% glycerol, 50 mM KCl, 10 mM 2-mercaptoethanol, 1 mM EDTA) four times. The dialyzed protein was concentrated by using Microcon 10 or Centricon 10 (Amicon). The purified protein was stored at Ϫ70°C until use.
In Vitro Transcription Assays-To measure in vitro transcription initiation, we used the pSR plasmid vector (12), which carries the bacteriophage oop transcription terminator just downstream of the EcoRI and HindIII cloning sites. Transcription experiments were performed using the protocols described by Kolb et al. (12). The KK81 or JK14 EcoRI-HindIII fragments containing the wild type or mutant melAB promoters were cloned into pSR such that transcripts initiating at the melAB promoter would run to the oop terminator. The resulting recombinants were purified by cesium chloride centrifugation and used 2 E. Tamai, unpublished data. Filled triangles, sites 1 and 2, which have identical sequences; gray triangles, site 1Ј and the related site R that overlaps the melR promoter; open triangle, weak site 2Ј that is discussed here. The locations of the Ϫ160 and Ϫ10 primers, which were used to generate ET1 probes for the gel retardation assays are shown. B, shows the DNA sequence around MelR binding sites (sites 1Ј, 1, 2, and 2Ј) in the wild type melAB promoter and in the W3133-2 mutant. The consecutive five-nucleotide mutation is shown with bold letters. The underlined sequences between sites 1Ј and 2, and those between sites 1 and 2Ј, are identical. The location of the ET2 and ET3 probes used in the gel retardation assays are also indicated.  After an additional 5-min incubation, 32 P-labeled RNA was analyzed on calibrated polyacrylamide sequencing gels. Transcription initiation at the melAB promoter gives rise to an easily detectable discrete transcript that terminates at the oop terminator, with the control RNA I from the pSR ColE1 origin acting as an internal control. Transcripts were quantified with the aid of a PhosphorImager (Molecular Dynamics, Inc.). Gel Retardation Assays-Starting with plasmid pBM3133 or pBM3133-2 as templates, PCR was used to amplify DNA fragments covering MelR binding sites 1Ј, 1, 2, and 2Ј (ET1) from the wild type melAB promoter or the W3133-2 mutant. To do this, we used the Ϫ160 primer (5Ј-CGAGGGGACTCTCCGAGTATC-3Ј) together with the Ϫ10 primer (5Ј-TAACTTCATCATGGCAGGCGA-3Ј) shown schematically in Fig. 1. Before the PCR reaction, the Ϫ160 primer was end-labeled with 32 P using T4 polynucleotide kinase. The labeled PCR product was used as a probe in gel retardation assays. Incubation mixtures contained 32 P-labeled DNA probe and purified His-tagged MelR in 50 mM Tris-HCl, pH 7.4, 50 mM KCl, 10% glycerol, 0.1 mg/ml sheared salmon sperm DNA, 10 mM melibiose, 1 mM EDTA, and 1 mM dithiothreitol. After incubation for 5 min at 37°C, samples were applied to 5% polyacrylamide gels that were run in standard Tris-borate/EDTA electrophoresis buffer at 150 V. After electrophoresis, gels were fixed, dried, and exposed to x-ray film. Gel retardation assays were also performed with the shorter ET2 and ET3 probes illustrated in Fig. 1. These probes were made by annealing the appropriate complementary oligos after 5Ј-end labeling with 32 P using T4 polynucleotide kinase.
DNase I Footprinting Experiments-These experiments were performed with EcoRI-HindIII fragments carrying the wild type or mutant mel operon regulatory region using previously described protocols (16). Incubations contained 4 -10 nM of the purified fragment that had been specifically radiolabeled at the HindIII end using [␥-32 P]ATP and polynucleotide kinase, and different amounts of purified His-tagged MelR and melibiose as indicated. After DNase I treatment, footprint patterns were analyzed on polyacrylamide sequencing gels that were calibrated with Maxam-Gilbert sequence ladders.

In Vivo Assays with the W3133-2 Mutant melAB
Promoter-In our previous work, we described the W3133-2 mutant that results from five consecutive base changes in the melAB promoter (7). To study the effects of these changes in vitro, we subcloned EcoRI-HindIII DNA fragments carrying the wild type or W3133-2 mutant melAB promoter sequences (Fig. 1). First, using PCR, we amplified the KK81 fragment that carries the entire melAB promoter with all four MelR binding sites (sites 1Ј, 1, 2, and 2Ј) that are known to be essential for optimal melAB promoter activity (10,11). Second, we amplified the shorter JK14 fragment that carries the melAB transcript start site and only the MelR binding sites 2 and 2Ј. In preliminary experiments, we checked the activity of the melAB promoter carried by these fragments. To do this, the different EcoRI-HindIII fragments were cloned into the broad host range lac fusion vector, pRW50, and the resulting plasmids were transformed into WAM131, a ⌬lac mel ϩ strain derived from GM-1 (11,14). Expression of the promoter melAB::lacZ fusions carried by the pRW50 recombinants was measured in transformants grown either with or without the inducer, melibiose. The results, listed in Table I, indicate that, with the longer KK81 fragment, expression from the melAB promoter is clearly induced by melibiose and that the W3133-2 mutant promoter is 8-to 9-fold more active than the wild type promoter. With the shorter JK14 fragment, the removal of MelR binding sites 1Ј and 1 results in a 4-fold reduction in expression of the melAB promoter. However, in contrast, with the W3133-2 mutant, melAB promoter activity is only slightly affected by removal of MelR binding sites 1Ј and 1. Thus, the base changes in the W3133-2 mutant result in a Ͼ20-fold increase in melAB promoter activity in the context of the shorter JK14 fragment. These results show that the effects of the base changes in the W3133-2 mutant are easily detected in short fragments and encouraged us to attempt in vivo studies.
Overexpression and Purification of His-tagged MelR-To perform in vitro study, we needed to purify MelR. However, the published method for purification of MelR required a large amount of medium and a complicated column works (20). Therefore, we employed Ni 2ϩ affinity chromatography for the purification and T7 expression system for overexpression of MelR. First we constructed a plasmid that carries a derivative melR encoding MelR with the carboxyl-terminal hexahistidine tag (MelRH). The derivative melR gene is located downstream from the T7 promoter. To confirm that the MelRH is functional in vivo, a XbaI-EcoRI fragment, which carries a Shine-Dalgarno sequence was subcloned from pETRH or pETR into pMW119, in which expression of melRH or melR is controlled by the lac promoter. A MelR-defective mutant, Y1089 (⌬melR) (19), harboring either one of the constructed plasmids, formed similar red colonies on a MacConkey-melibiose (5 mM) plate (data not shown). This result indicates that the MelRH worked as a transcription activator for the melibiose operon in vivo.
Other results from gel retardation assay and in vitro transcription assay done with MelR or MelRH indicated that MelRH had almost the same DNA binding affinity and transcription activation activity as wild type MelR (data not shown).
For overexpression of MelRH, we used the T7 system. We tested several media and found that M9 medium increased the expression level of our target protein and reduced background proteins. The MelRH protein produced in BL21(DE3)/pETRH after addition of isopropyl ␤-D-thiogalactopyranoside reached about 10 -20% of total cell protein. However, most of the overproduced MelRH protein formed inclusion body. The advantage of this preparation was that the MelRH in the inclusion body was more than 80% pure without any column works. The inclusion body was solubilized with 6 M guanidine buffer and purified using Ni 2ϩ -agarose. The purified protein was renatured by dialysis and then concentrated by ultrafiltration. When concentration of the purified MelRH was high, precipitation occurred during dialysis and most protein bound to filter during ultrafiltration. Our protocol yielded 500 g of Ն95% pure His-tagged MelR from 200 ml of culture. This amount is enough for in vitro analysis. The renatured MelRH showed almost identical transcription activation activity as that ob- served with MelR purified by the method reported previously (19) without denaturation (data not shown).
Activation of the melAB Promoter in Vitro using Purified MelR-It is important to test whether the results obtained in the in vivo system (Table I and Ref. 7) are reproducible in the in vitro system. To do this, we cloned EcoRI-HindIII DNA fragments carrying the wild type or W3133-2 mutant melAB promoter sequences into the pSR vector plasmid, which carries the bacteriophage oop transcription terminator just downstream of the HindIII site (12). The resulting recombinant plasmids were used to monitor MelR-dependent activation at the melAB promoter; i.e. transcripts initiating at the melAB promoter run to the oop terminator and give discrete transcripts that are easy to detect by gel electrophoresis. In addition, because pSR contains a ColE1 replication origin, the RNA I transcript can be used as an internal control to aid quantification of MelR-dependent transcripts. Fig. 2 shows the results from a typical experiment, where purified RNAP, together with purified MelR and melibiose, was preincubated with purified circular pSR plasmid carrying the JK14 or KK81 insert, prior to the addition of labeled nucleoside triphosphates. Each incubation produces the control RNA I transcript, but a longer transcript originating at the melAB transcript start is seen when the preincubation includes purified MelR and melibiose. With both the JK14 and KK81 inserts, more transcript initiating at the melAB promoter is observed with template from the W3133-2 mutant. We have done three sets of the same experiment and two sets of an experiment in which concentration of RNA polymerase was increased to 150 nM. All of the results were almost identical to the result in Fig. 2. From these experiments, we conclude that the effects of the base changes in the W3133-2 mutant can be reproduced, at least partially, in vitro.
Gel Retardation Assays-It had previously been noted that the base changes in the melAB promoter in the W3133-2 mutant changed the DNA sequence of MelR binding site 2Ј such that it corresponded better to the tighter binding sites 1Ј, 1, and 2. Therefore, gel retardation assays were used to measure the binding of purified MelR to the different sites at the wild type or W3133-2 mutant melAB promoter. In the first experiment (Fig. 3A), we measured the binding of MelR to ET1 fragments carrying MelR binding sites 1Ј, 1, 2, and 2Ј (the ET1 fragments were made using PCR with the Ϫ160 and Ϫ10 primers and pBM3133 or pBM3133-2 as templates: see Fig. 1). With the wild type ET1 fragment, four retarded bands are observed, with the most retarded bands appearing at higher MelR concentrations. We attribute these bands to progressive occupation of the four DNA sites for MelR: as expected, the fourth, most retarded band remained quite weak, in agreement with our previous observation that MelR binds weakly to site 2Ј. However, with the ET1 fragment from the W3133-2 mutant, this fourth retarded band appears much stronger, suggesting that the base changes in the mutant have tightened the binding of MelR to site 2Ј. The second experiment (Fig. 3B) with the shorter ET2 fragment, carrying only MelR binding sites 2 and 2Ј, confirmed this interpretation: with the fragment derived from the wild type melAB promoter, two retarded bands are observed, with only a small amount of the more retarded band appearing at higher MelR concentrations. However, with the ET2 fragment from the W3133-2 mutant, this second more retarded band appears much stronger. Finally, the control experiment ( Fig. 3C) with the ET3 fragment carrying only MelR binding site 2 shows just one retarded band that is unaffected by the base changes from the W3133-2 mutant. Note that the version of this fragment derived from the mutant does contain the five base changes (see Fig. 1), but these do not affect MelR binding to site 2. It should also be noted that MelR did not bind to the fragment carrying just site 2Ј (data not shown).
DNase I Footprinting Experiments-To confirm the conclusions from gel retardation assays, we used DNase I footprinting to study the binding of purified MelR to the JK14 fragment carrying the wild type or W3133-2 mutant melAB promoter sequences. The results (Fig. 4) show that, with the wild type melAB promoter sequence, MelR affords clear protection of the segment of DNA corresponding to MelR binding site 2. However, even in the presence of melibiose and higher concentrations of MelR, very poor protection of the segment of DNA corresponding to MelR binding site 2Ј is observed. However, with the JK14 fragment derived from the W3133-2 mutant, clear protection of both site 2 and site 2Ј is seen, even in the absence of melibiose. DISCUSSION From our results, we conclude that, as predicted, the five base changes in the melAB promoter of the W3133-2 mutant improve the binding of MelR to site 2Ј. We suppose that this improvement in MelR binding is responsible for the increased melAB promoter activity in the W3133-2 mutant, although we cannot prove that it is solely responsible. In our previous paper, we noted that the five base changes in the W3133-2 mutant create an extended 28-nucleotide inverted repeat at the melAB promoter (7). We suggested that this inverted repeat might facilitate the formation of a stem-loop structure and that this may be responsible for the increased melAB promoter activity in the W3133-2 mutant. However, our observation that the increased melAB promoter activity is observed even in the absence of MelR binding sites 1Ј and 1 (i.e. with the JK14 fragment, which lacks the extended inverted repeat) proves that this cannot be the case. Indeed, the effects of the five base changes are greater with the JK14 fragment: this appears to be because MelR binding to site 2Ј is aided by MelR binding to sites 1Ј and 1, and thus, in the absence of sites 1 and 1Ј, improvement of site 2Ј has a proportionately greater effect.
Taken together, our results show that the activity of the melAB promoter is limited by the binding of MelR to site 2Ј. Because this site overlaps the Ϫ35 hexamer of the melAB promoter, it appears that MelR, like AraC and many other members of the AraC family, may be a typical "Class II" transcription activator (17). Our results suggest that the principal function of the upstream sites 1Ј and 1 is to help the occupation of site 2Ј. Thus, the requirement of sites 1Ј and 1 for melAB promoter activity is greatly reduced when MelR binding site 2Ј is improved (Table I). It is known that Class II transcription activators function by making a direct contact with Region 4 of the RNAP subunit (18). Thus, the challenge now is to understand the molecular details of the MelR-subunit interactions.