The TyrR Transcription Factor Regulates the Divergent akr-ipdC Operons of Enterobacter cloacae UW5

The TyrR transcription factor regulates genes involved in the uptake and biosynthesis of aromatic amino acids in Enterobacteriaceae. Genes may be positively or negatively regulated depending on the presence or absence of each aromatic amino acid, all three of which function as cofactors for TyrR. In this report we detail the transcriptional control of two divergently transcribed genes, akr and ipdC, by TyrR, elucidated by promoter fusion expression assays and electrophoretic mobility shift assays to assess protein-DNA interactions. Expression of both genes was shown to be controlled by TyrR via interactions with two TyrR boxes located within the akr-ipdC intergenic region. Expression of ipdC required TyrR bound to the proximal strong box, and is strongly induced by phenylalanine, and to a lesser extent by tryptophan and tyrosine. Down-regulation of akr was reliant on interactions with the weak box, and may also require a second, as yet unidentified protein for further repression. Tyrosine enhanced repression of akr. Electrophoretic mobility shift assays demonstrated that TyrR interacts with both the strong and weak boxes, and that binding of the weak box in vitro requires an intact adjacent strong box. While the strong box shows a high degree of conservation with the TyrR binding site consensus sequence, the weak box has atypical spacing of the two half sites comprising the palindromic arms. Site-directed mutagenesis demonstrated sequence-specific interaction between TyrR and the weak box. This is the first report of TyrR-controlled expression of two divergent protein-coding genes, transcribed from independent promoters. Moreover, the identification of a predicted aldo-keto reductase as a member of the TyrR regulon further extends the function of the TyrR regulon.


Introduction
The TyrR transcription factor controls the expression of genes involved in the uptake and metabolism of aromatic amino acids. In Escherichia coli K12, the TyrR regulon consists of eight operons and nine genes [1,2]. Regulation of each transcriptional unit is either positive or negative, depending on the position of TyrR binding sites, known as TyrR boxes, in the promoter branched chain and aromatic amino acid degradation and components of the glyoxylate shunt of the tricarboxylic acid cycle [24].
Previously, we showed that in the rhizobacterium Enterobacter cloacae UW5, the TyrR regulon includes the ipdC gene encoding the enzyme indolepyruvate decarboxylase which converts indole-3-pyruvate to indole-3-acetaldehyde, a key step in the indole-3-acetic acid (IAA) biosynthetic pathway [25]. Binding of TyrR to a sequence with high similarity to the consensus binding sequence is required for activation of ipdC expression and IAA production. This strong TyrR box is centered directly between ipdC and a divergently transcribed gene, akr, encoding a putative aldo-keto reductase (Fig. 1A). In addition, a second sequence that resembles a TyrR box was found upstream of the akr coding sequence (Fig. 1A and B). This second site contains palindromic arms with high identity to those of the consensus TyrR binding sequence but with non-canonical spacing between them (hereafter referred to as the weak TyrR box). In this study, we determined that akr is a member of the TyrR regulon and that TyrR binds to both predicted sites in the akr-ipdC intergenic region. Using reporter gene expression assays, site-directed mutagenesis and electrophoretic mobility shift assays (EMSAs), we characterize the roles of the strong and weak TyrR boxes in the activation and repression of ipdC and akr, respectively. Genes that are co-regulated are often functionally related and understanding how akr is controlled by TyrR is a first step towards elucidating its possible role in production of IAA, an important signaling molecule in bacterial-plant interactions [26].

Bacterial strains and culture conditions
The bacterial strains and plasmids used in this study are listed in Table 1. E. coli and E. cloacae were routinely grown at 37°C and 30°C, respectively, in either Luria Bertani (LB) broth or M9 minimal media (Difco). Media was supplemented with antibiotics when required at final concentrations of 100 μg/mL ampicillin, 25 μg/mL kanamycin, 20 μg/mL gentamicin or 5 μg/mL tetracycline. L-Aromatic amino acid supplements to the growth media were supplied at a final concentration of 1 mM when indicated. Transcription start site identification 5'-Rapid amplification of cDNA ends (RACE) was used to identify the transcription start sites of akr and ipdC. Overnight cultures of E. cloacae UW5 were inoculated 1:100 into fresh M9 media with or without 1 mM of each aromatic amino acid (L-tryptophan, L-phenylalanine, Ltyrosine). Cultures were grown to mid-logarithmic (OD 600 = 0.6, 8 h) and stationary (OD 600 = 2.5, 48 h) phase, at which point 1 mL of culture was transferred to a microfuge tube containing 0.5 mL RNAprotect (Qiagen), and vortexed for~30 s. Cells were then centrifuged, supernatant discarded and the pellet stored at −20°C. Total bacterial RNA was isolated using the RNeasy Mini Kit (Qiagen). RNA quality was assessed by running a sample on a 2% agarose gel, and RNA quantity and purity was measured spectrophotometrically at 260 nm using a SpectraMax M5 plate reader (Molecular Devices). A total of 2 μg RNA was treated with amplification grade DNase I (Invitrogen) to remove any contaminating DNA. cDNA synthesis was performed using Superscript III reverse transcriptase (Invitrogen) according to the manufacturer's recommendations with primers IPIF and IPDC-GSP1 (Table 2) for akr and ipdC, respectively. cDNA was purified using Qiagen's PCR purification kit and end tailed with dCTP using terminal transferase (New England Biolabs) in a reaction containing 1X TdT buffer, 2.5 mM CoCl 2 and 10 mM dCTP. Reactions were incubated at 95°C for two min and cooled to room temperature before addition of 20 U terminal transferase and further incubation at 37°C for 20 min. Tailing reactions were stopped by heating to 65°C for 10 min.
Ten microliters of tailed cDNA were then used in a 50 μL touchdown PCR with an abridged anchored primer (AAP) and gene specific primers, AKR-GSP2 and IPDC-GSP2 (Table 2) for akr and ipdC, respectively. Touchdown PCR cycling conditions were as follows: initial denaturation at 94°C for 2 min; 20 cycles of 94°C for 30 s, 58°C −0.5°C per cycle for 30 s, 72°C for 30 s; PCR products were gel purified before cloning. Both akr and ipdC tailed cDNA fragments were ligated into pGEM-T Easy (Promega), generating the plasmids pGEM-IP1 and pGEM-AP1 (Table 1), respectively, and subsequently transformed into E. coli JM109 cells (Promega). Inserts were then sequenced using standard T7 and SP6 primers (Robarts Research Institute, London, Ontario, Canada) to identify the transcription start site.

Construction of reporter gene fusions and promoter mutagenesis
The impact of the TyrR transcription factor on expression of akr and ipdC was examined using plasmid-based reporter gene fusions. The suicide plasmids pJQSIPG and pJQSIAPG contain the ipdC and akr promoters, respectively, fused to the uidA reporter gene encoding β-glucurionidase [25]. The PipdC::uidA and Pakr::uidA fusions were excised as ApaI-XbaI fragments from pJQSIAPG and pJQSIAPG, respectively, and inserted into a similarly digested derivative of the pBBR1MCS-2 broad host range plasmid [30] lacking the lacZ promoter (pBBRL). The lacZ promoter was excised from pBBR1MCS-2 using the restriction enzymes KpnI and NsiI. Remaining 5' overhangs were removed by Klenow digestion followed by re-circularisation by blunt end ligation with T4 DNA ligase (NEB) to generate pBBRL. Following insertion of the PipdC::uidA and Pakr::uidA fusions, the resulting plasmids pBBRLI and pBBRLA, respectively, ( Table 1) were transformed into E. coli JM109 (Promega), purified and introduced into E cloacae UW5 and J35 by electroporation.
The involvement of two predicted TyrR binding sites was examined by introducing point mutations into the binding sites using site-directed ligase-independent (SLIM) PCR mutagenesis [31] with the pBBRLI and pBBRLA plasmids as templates. SLIM-PCR utilises inverse PCR amplification of the template plasmid with two pairs of primers, a tailed pair containing the mutation and a short pair flanking the mutation site. Post-amplification denaturation and reannealing produces a mixed population of products, of which only those containing the targeted mutation are capable of hybridization and therefore re-circularization. Mutations were generated in 25 μL PCR reactions using Phusion High Fidelity DNA polymerase (NEB) and 8% DMSO with the following sets of primer pairs: weak box mutagenesis, WFT2 and WRT2; WFS1 and WRS1; strong box mutagenesis, SFT1 and SRT1; SFS1 and SRS1; weak box conversion to strong box, WFT1 and WRT1; WFS1 and WRS1 ( Table 2). A double TyrR box mutant was generated by introducing mutations in the wild-type strong box sequence of a previously generated weak box mutant plasmid. Cycling conditions for all SLIM-PCR reactions were as follows: 98°C for 3 min; 35 cycles of 98°C for 10 s; 63°C for 15 s; 72°C for 3 min; final elongation at 72°C for 5 min. Following amplification, 25 μL of PCR product was treated with 10 U DpnI at 37°C for 60 min with 5 μL D-buffer (20 mM MgCl 2 , 20 mM Tris-CL, pH 8.0, 5 mM DTT) to remove non-mutated template DNA. The reaction was stopped with 30 μL H-buffer (300 mM NaCl, 50 mM Tris-CL pH 9.0, 20 mM EDTA, pH 8.0) before SLIM-PCR hybridisation at 99°C for 3 min; followed by two cycles of 65°C for 5 min; 30°C for 40 min. Plasmid DNA was purified and transformed into CaCl 2 competent E. coli S17-1 (λpir) cells. Successful mutagenesis was confirmed by sequencing the plasmids (Robarts Research Institute). Following sequence confirmation, plasmids were introduced into wild-type E cloacae UW5 and tyrR null mutant E. cloacae J35 by electroporation. The resulting plasmids are listed in Table 1, with their corresponding mutations in Fig. 1C.

Reporter gene expression assays
ipdC and akr promoter-driven reporter gene expression was quantified using the method of Cowie et al. (2006) [32]. Wild-type E. cloacae UW5 and tyrR null mutant E. cloacae J35 cells containing the reporter gene constructs described above were grown in triplicate from independent colonies overnight in LB broth with appropriate antibiotics, pelleted, and washed twice in 1X M9 salts. Cultures were then diluted 100-fold into 1 mL M9 minimal media in 2.2 mL 96-deep well plates (Eppendorf, Canada), with or without each of the aromatic amino acids. Cultures were grown to mid-logarithmic (OD 600 = 0.6, 8 h) and stationary (OD 600 = 2.5, 48 h) phase, at which time they were assayed for β-glucuronidase activity. Briefly, 20 μL of cell culture were withdrawn to wells of a 96-well plate (Cellstar, Sigma) containing 80 μL of reaction buffer (50 mM sodium phosphate [pH 7], 50 mM dithiothreitol, 1 mM EDTA, 0.0125% sodium dodecyl sulfate) and 0.44 mg/mL p-nitrophenyl β-D-glucuronide (PNPG), and incubated at room temperature until development of a strong yellow precipitate, at which time the reaction was stopped with the addition of 100 μL 1 M Na 2 CO 3 . Absorbance of the reaction products was measured at 405 nm using a SpectraMax M5 plate reader (Molecular Devices) and used to calculate specific promoter activity in Miller units (1000 x OD 405nm )/(time x OD 600nm x culture volume in reaction). Wild-type E. cloacae UW5 and uninoculated medium were included as controls to assess background levels of absorbance at 405 nm.

Electrophoretic mobility shift assays
The previously constructed expression vector pQEtyrR (Table 1) was used to purify TyrR via an N-terminal His 6 tag [25]. E. coli M15 carrying plasmids pREP4, which constitutively produces high levels of lac repressor protein, and pQEtyrR were grown at 37°C in LB to mid logarithmic phase (OD 600 = 0.6, 2.5 h) before induction with 1 mM isopropyl β-Dthiogalactopyranoside (IPTG) and allowed to incubate for an additional four hours. His 6 -TyrR was purified under native conditions using the QIAexpressionist Ni-NTA metal affinity chromatography kit (Qiagen). The native purification protocol was followed according to the manufacturer's instructions with the exception of a modified wash step and elution buffer. Wash buffer 1 contained 50 mM NaH 2 PO 4 , 300 mM NaCl, 10 mM imidazole; wash buffer 2 was as described for wash buffer 1 with 10% glycerol; the elution buffer contained 50 mM NaH 2 PO 4 , 300 mM NaCl, 250 mM imidazole, 8% glycerol. Purified protein was quantified using the Quick Start Bradford protein assay (Bio-Rad), and purity assessed on a 12% SDS-PAGE gel stained with Coomassie brilliant blue G-250.
DNA fragments encompassing the intergenic region of akr and ipdC were amplified with Phusion High Fidelity DNA polymerase (NEB) using the wild-type and mutant promoter fusion plasmids described above as templates. Primers CF1 and DIG-CR1 (Table 2) amplified a 198 bp fragment spanning the full intergenic region (Fig. 1B), while incorporating a digoxigenin moiety (DIG) at one end. Non-DIG labeled primer CR1 was used in conjunction with CF1 to amplify unlabelled wild-type intergenic DNA to be used in competition assays. DNA fragments were separated on a 6% native-PAGE gel and stained with ethidium bromide before target bands were cut out and purified using a PCR purification kit (Qiagen).
DNA was transferred to a positively charged nylon membrane (Roche) using a Trans-Blot SD Semi-Dry cell (Bio-Rad) and visualised using the anti-DIG-Detection system according to the manufacturer's instructions (Roche). Developed membranes were imaged with a Bio-Rad ChemiDoc MP (Bio-Rad).

Statistical analysis
Statistically significant differences in gene expression from each promoter during logarithmic or stationary growth phases were detected by two-way analysis of variance (ANOVA) with post-hoc comparisons by TukeyHSD test using JMP 7 [33]. The genotype (i.e., strains UW5 and J35), promoter mutation, and amino acid treatments were treated as fixed effects, fully crossed with each other when indicated for each test. Statistical differences were defined as p values of <0.05. Unless stated otherwise, all main effects and interactions were significant.

Mapping of the akr and ipdC promoters
To understand how TyrR regulates akr and ipdC, the location of the promoters of both genes relative to the predicted TyrR boxes was first determined. 5'-RACE was used to identify the transcription start sites from which the core promoter sequences of the two genes were inferred. The transcription start site for ipdC was found to be a purine base located 26 nucleotides upstream of the ATG start codon (Fig. 1B). Examination of the sequence upstream of the +1 site revealed candidate −10 and −35 hexamers with strong resemblance to the E. coli sigma-70 binding sequence, and separated by 16 base pairs. The −10 sequence matches the E. coli sigma 70 consensus sequence at four of six nucleotides, while the −35 sequence matches at five out of six nucleotides. Within the 5' untranslated region of ipdC is a strong candidate ribosome binding site. This positions the center of the strong TyrR box 33 base pairs upstream from the center of the ipdC −35 sequence, roughly three turns of the DNA helix.
5'-RACE identified a transcription start site for the akr gene 19 base pairs downstream of the proposed ATG start codon at an adenine residue. This was considered to be an artifact based on the following evidence: (i) the akr coding sequence is conserved among closely related bacteria with annotated genomes, including the start codon, and (ii) the identified adenine was located at the end of five contiguous adenines, and reverse transcriptase is known to pause or prematurely terminate activity when it encounters a poly-adenine tract [34]. The genome wide transcription start sites of the closely-related enterobacterium Salmonella enterica sv. typhimurium SL1344 were identified by RNA sequencing, including those for the homologous ipdC and akr genes [35]. Alignment of the S. enterica SL1344 and E. cloacae UW5 akr-ipdC regions suggested that the E. cloacae UW5 akr transcription start site is a cytosine nucleotide 25 base pairs upstream of the akr start codon (Fig. 1B). The S. enterica ipdC transcription start site identified by Kroger et al. (2012) is consistent with the E. cloacae UW5 ipdC transcription start site identified in this study. Visual inspection of the sequence upstream of the predicted E. cloacae UW5 akr transcription start site revealed a possible −10 element that matches four out of six nucleotides of the sigma-70 consensus sequence [36]. This positions the predicted TyrR weak box to overlap the −10 element which supports the proposal that TyrR functions as a repressor of akr. No definitive −35 hexamer was identified upstream of the −10, although the lack of a −35 element does not preclude RNA polymerase interaction [37]. A candidate ribosome binding site sequence was found seven base pairs upstream of the ATG within the 5'-UTR of akr.
ipdC is positively regulated by TyrR and the aromatic amino acids To study the regulatory effects of TyrR on the akr and ipdC genes, uidA transcriptional fusions were generated using a derivative of the low copy number plasmid pBBR1MCS-2 [30] which lacks the lacZ promoter. Consistent with previous results [25], when wild-type E. cloacae UW5 cells containing the ipdC::uidA fusions were grown and assayed for β-glucuronidase activity in stationary phase, expression from PipdC was 3.5-fold greater than that in the tyrR null mutant E. cloacae J35 strain which lacks a functional TyrR ( Fig. 2A). In the wild-type background, expression from ipdC increased significantly in response to each of the aromatic amino acids, relative to that in M9 minimal medium ( Fig. 2A). Phenylalanine and tryptophan had the strongest effects, increasing ipdC expression four-fold and two-fold, respectively, above M9 minimal media alone in stationary phase. Tyrosine also induced expression 1.5-fold above that in minimal media, however this was less then that previously observed [25]. Expression of ipdC was on average 3.6-fold greater in stationary phase relative to logarithmic phase. This is consistent with the observation that IAA, the product of the reaction catalyzed by IPDC, is not detectable in the culture media of E. cloacae UW5 until stationary phase [25,29].

akr is negatively regulated by TyrR
In contrast to the positive regulatory effects of TyrR on the ipdC promoter, akr expression was negatively regulated by TyrR (Fig. 2B). When expression from akr was measured in the tyrR null mutant E. cloacae J35, a 1.6-fold increase was observed (in both log and stationary phases) compared to expression levels in wild-type E. cloacae UW5, indicating that TyrR normally represses akr. Inclusion of the aromatic amino acids tryptophan and phenylalanine did not result in significant down-regulation of akr, however, the inclusion of tyrosine in the medium resulted in a 1.5-fold decrease in akr expression relative to M9 medium alone during log phase. Transcription from the akr promoter remained at relatively constant levels during both logarithmic and stationary phases of growth when assayed in wild-type E. cloacae UW5.

ipdC activation requires the strong TyrR binding site
To test the involvement of the two predicted binding sites in TyrR mediated activation of ipdC, four promoter variants were constructed via SLIM-PCR mutagenesis (Fig. 1C). These mutations were generated to abolish TyrR binding to the weak and/or strong boxes by changing nucleotides known to be essential for TyrR DNA binding in E. coli [1]. The center of the strong TyrR box is located 33 base pairs upstream of the center of the predicted −35 element of the ipdC promoter, positioning the two sites three turns of the DNA helix apart and on the same face of the DNA. This would allow TyrR to act as a strong type I transcriptional activator in a manner consistent with other genes that are positively regulated by TyrR, such as tyrP [38]. Mutation of the strong box (strains carrying pBBRLIS) resulted in the reduction of ipdC expression to a level similar to that in the tyrR null mutant E. cloacae J35 (Fig. 3A). Similar results were observed from the double weak and strong box mutant promoter (pBBRLIWS). In both logarithmic and stationary phase, inactivation of the weak box alone (pBBRLIW) was found to have little effect on expression of ipdC (Fig. 3A).
To examine the effect of TyrR binding at the weak box on the ipdC promoter, the weak box spacing was adjusted such that each arm of the palindromic TyrR box motif was separated by the ideal six nucleotides (corresponding to the same spacer nucleotides as in the adjacent strong box). The high identity of the weak box palindromic arms to the TyrR consensus sequence, coupled with the optimal spacing would ensure that TyrR is bound with high affinity to both sites [1]. The resulting double strong box promoter (pBBRLISS) increased transcriptional activity 2.1-fold over levels of the wild-type promoter (pBBRLI) (Fig. 3A). Similar expression trends were observed in the presence and absence of aromatic amino acids, although this effect was enhanced in the presence of aromatic amino acids, especially phenylalanine, as expected for a TyrR mediated response (S1 Fig.). Repression of akr is facilitated by the weak TyrR binding site occluding the promoter Based on the predicted transcription start site for akr, the weak TyrR box is situated to overlap the predicted −10 sequence of the akr promoter, suggesting that binding of TyrR to the weak box competes with RNA polymerase for access to the promoter. Mutation of the weak box (pBBRLAW) resulted in a 3.5-and 5.3-fold higher level of expression from the akr promoter compared to the wild-type promoter (pBBRLA) in logarithmic and stationary phases, respectively (Fig. 3B). The effect of the weak box mutation was greater in the tyrR null mutant E. cloacae J35, particularly in stationary phase, which showed a 16-fold increase in akr expression over the wild-type UW5 strain. Inactivation of the strong box alone (pBBRLAS) had no effect on akr expression in either growth phase suggesting that the observed repression is mediated by the weak box (Fig. 3B). When both TyrR boxes were mutated (pBBRLAWS), transcription from the akr promoter increased 7.3-and 11-fold over levels from the wild-type promoter in both wild-type E. cloacae UW5 and the tyrR null mutant E. cloacae J35, respectively, in stationary phase. Conversion of the weak box to a strong box (pBBRLASS) yielded the same level of repression as from the wild-type promoter in wild-type E. cloacae UW5; however, in the tyrR null mutant E. cloacae J35, transcription increased 3.8-and 9.3-fold in log and stationary phases, respectively. The effect of the TyrR box mutations on akr expression was similar in M9 medium with and without aromatic amino acid supplements (S2 Fig.). These data indicate that both the weak box and TyrR are important for akr repression, but that TyrR interaction at the weak box does not sufficiently explain the observed repression.

TyrR binds to two sites in the akr-ipdC intergenic region
The ability of TyrR to bind to specific sites within the intergenic region of akr-ipdC was demonstrated using purified His 6 -TyrR protein. When incubated with purified His 6 -TyrR, a DIGlabeled DNA fragment encompassing the wild-type akr-ipdC intergenic region migrated to a position consistent with a higher molecular weight complex on a native-PAGE gel (complex I), compared to the DNA fragment alone (Fig. 4, lanes 1, 2). Migration of the DNA fragment was His 6 -TyrR concentration dependent, with levels of His 6 -TyrR greater than 87 nM inducing a second DNA shift to complex II (Fig. 4, lane 3, and S3 Fig.). These results suggested that TyrR binds to two sites in the akr-ipdC intergenic region. Pre-incubation of His 6 -TyrR with a 10-fold molar excess of unlabeled DNA probe abolished the band at the complex II location, however complex I was still apparent in the gel (S3 Fig., rightmost lane). There was no apparent increase or decrease in affinity for DNA binding by His 6 -TyrR when EMSAs were repeated in the presence of aromatic amino acids (S3 Fig.).
To determine if His 6 -TyrR binds to the akr-ipdC intergenic region at the two predicted TyrR boxes, EMSAs were performed using the mutated promoter sequences generated by SLIM-PCR as DNA binding templates for His 6 -TyrR. The results show that when the weak box was mutated, His 6 -TyrR still bound the DNA to yield complex I, however complex II was abolished (Fig. 4, lanes 5, 6). Mutations in the strong box resulted in the loss of intense bands representing both complexes, although the weak bands may indicate some low affinity interactions with the intact weak box at higher His 6 -TyrR concentrations (Fig. 4, lanes 8, 9). All interactions were abolished with fragments carrying mutations in both TyrR boxes (Fig. 4, lanes 11,  12). When the DNA fragment with the double strong box was used as a template for EMSAs, all labeled DNA migrated to complex II, even at the lowest concentration of His 6 -TyrR (Fig. 4,  Lanes 14, 15). This indicates that the increased spacing between the palindromic arms of the weak box increased the affinity of TyrR for the weak box and supports strong protein-protein interactions between TyrR dimers bound at both locations on the DNA. The results shown in Fig. 4 demonstrate that TyrR interacts with the two predicted TyrR binding sites within the akr-ipdC intergenic region. While interactions with the akr proximal box are weaker and require TyrR to be bound to the adjacent strong box in vitro, this is not the case in vivo where the strong box has no discernible effect on akr expression. In the latter case, additional regulatory factors may be acting at the promoter.

Discussion
In this report we detail the transcriptional control of two adjacent, divergently transcribed genes, akr and ipdC, by the transcription factor TyrR, elucidated by promoter fusion expression assays and EMSAs to assess protein-DNA interactions. Expression of akr and ipdC was controlled by TyrR interaction with a weak and strong TyrR box, respectively, located within the intergenic region. Direct protein-DNA interactions between TyrR and the strong and weak boxes were shown in vitro using purified TyrR protein and DNA fragments containing the akr-ipdC intergenic region; the interactions were abolished when point mutations were introduced in either or both TyrR boxes. These interactions were strengthened when the sequence of the weak box was altered to match that of a strong box as demonstrated by the shift of the DNA fragment to a higher molecular weight complex when incubated with lower concentrations of TyrR and resolved by EMSA, and by the increased expression from ipdC reporter fusions.
Expression of ipdC is dependent on TyrR for activation; in the tyrR null mutant E. cloacae J35, expression of the PipdC::uidA fusion was abolished (this study; [25]), and production of IAA is also reduced. E. cloacae indolepyruvate decarboxylase, encoded by ipdC, catalyzes the production of IAA [25,39]. In plants, IAA is an important growth hormone, controlling many aspects of growth and morphogenesis such as gravitropism, cell elongation and tissue differentiation [40]. E. cloacae UW5 inhabits the rhizosphere of plants, where it may produce high levels of IAA, and has been shown to positively influence root development [29]. Plant growthpromoting rhizobacteria have been applied in field studies and commercial agriculture to improve crop yields, although results are often inconsistent or improvements only marginal [41]. Identifying the mechanisms underpinning regulation of IAA production in plant growth-promoting rhizobacteria will aid in their application as agricultural phytostimulants, especially considering that IAA is also a virulence factor produced by some plant pathogens and levels of this hormone may be an important factor in plant interactions (recently reviewed by [26]).
Inclusion of any of the aromatic amino acids in the growth media significantly increased ipdC expression in a TyrR dependent manner, with phenylalanine as the strongest inducer. The center of the strong TyrR box is located 33 base pairs upstream from the center of the predicted ipdC −35 element. This would position a TyrR dimer bound at the strong box on the same face of the DNA as RNA polymerase bound to the ipdC promoter, and in an optimal location for interactions with the RNA polymerase α-CTD [1,38]. It is likely that the strong induction of ipdC is due to activation by TyrR complexed with aromatic amino acids, which increases the affinity of TyrR for the DNA and leads to strong activation in promoters with similarly positioned TyrR boxes [21,42]. The promoter of the mtr gene in E. coli has a similarly orientated strong TyrR box centered 42 base pairs upstream of the −35 element. In the presence of phenylalanine, a TyrR dimer bound to this strong box interacts with the RNA polymerase α-CTD enhancing expression 2.8-fold [17,21]. A comparison of the TyrR activation of the E. coli mtr promoter and the E. cloacae UW5 ipdC promoter reveals very similar trends. The mtr gene is activated 2-fold (tryptophan), 2.8-fold (phenylalanine) and 1.7-fold (tyrosine); while ipdC is activated 2-fold (tryptophan), 3.7-fold (phenylalanine) and 1.4-fold (tyrosine) [20].
The newest member of the TyrR regulon, akr, identified in this study, was shown to be down-regulated by TyrR. Little is known regarding the function of the protein encoded by the akr gene. From the amino acid sequence it is predicted to be a member of the aldo-keto reductase (AKR) superfamily of proteins, a diverse group of enzymes found in all domains of life that share a common (α/β) 8 barrel fold and utilize the nicotinamide adenine dinucleotide (NAD(P)H) cofactor to reduce carbonyl compounds [43,44]. AKRs are further divided into families based on amino acid sequence identity (at least 40% identity for family designation). Phylogenetic analysis of the E. cloacae UW5 AKR groups it with E. coli YghZ (data not shown), which is homologous to the AKR6 family that comprises the β-subunit of voltage gated potassium channels [45]. The yghZ gene encodes a methylglyoxal reductase, which has been shown to confer protection against the toxic electrophile methylglyoxal by reducing it to acetol [46], and has been designated ARK14A1 based on amino acid identity. Although the function of the E. cloacae UW5 AKR is unknown, the sequence and organization of the open reading frame is conserved among closely related bacteria that possess the ipdC gene: when ipdC is present in a sequenced enterobacterial genome, it is found adjacent to, and divergently transcribed from, a homologous akr (Fig. 5A). Both the strong and weak TyrR bindings sites are also conserved in these genomes (Fig. 5B). Thus, the common regulation of these two genes by TyrR and their evolutionarily conserved pairing suggest a functional relationship between AKR and IAA biosynthesis.
Down-regulation of akr is dependent on the weak TyrR box which overlaps the predicted −10 element of the akr promoter; however, mutations in the strong TyrR box upstream of the akr promoter had no effect on expression. The observation that expression from the akr mutant weak box promoter was higher than that from the wild-type promoter in the tyrR null mutant E. cloacae J35, when the expectation was that they would be similar, may indicate that the nucleotide changes in the weak box increased the strength of the akr promoter. Alternatively, the increase in expression of the mutant weak box promoter may indicate the involvement of a second regulatory protein. This is supported by the similarly high levels of akr expression in the tyrR null mutant E. cloacae J35 containing the mutant weak or double strong box promoters. In both cases, nucleotide changes were introduced in the weak box sequence (in the case of the double strong box these would not be expected to enhance promoter strength) and may have abolished the binding site for a second regulator. In wild-type E. cloacae UW5 with a double strong box, the increased affinity of TyrR for the weak box maintained repression of akr at wild-type levels.
The observation that akr was down-regulated by TyrR, but not completely repressed may reflect the weak interactions between TyrR and the DNA binding site overlapping the −10 element. A TyrR dimer or hexamer may associate transiently with the weak box allowing RNA polymerase access to the promoter and a low level of expression. This is proposed to be the case with the aroG gene of E. coli, where the TyrR box overlaps the −10 element and competition between the two proteins for DNA binding results in a low but constant level of gene expression [1]. Weak interactions between TyrR and the weak box may be stabilized by proteinprotein interactions between TyrR bound to both boxes. These may be dimer-dimer interactions, as in the case of mtr [17] or the formation of a TyrR hexamer. TyrR is thought to polymerize from a dimer to a hexamer in the presence of tyrosine while bound to a strong box sequence which acts as a nucleation point [7]. The resulting hexamer has a 2-to 5-fold increased affinity for DNA binding, allowing its interaction with adjacent, weaker binding sites [47]. TyrR may polymerize when bound to the strong box in the akr-ipdC intergenic region, increasing its affinity for the adjacent weak box in the presence of tyrosine. This is supported by the requirement for the wild-type strong box for strong TyrR binding to the intergenic region in EMSAs. The observation that mutations in the strong box do not affect akr expression in the reporter gene assays may support the involvement of a second regulatory protein that stabilizes TyrR binding to the weak box in vivo.
The functionality of the weak TyrR box, demonstrated by its necessary role in repression of akr in vivo and by its binding to TyrR in vitro as shown by EMSAs, was one of the most interesting findings of this study. In E. coli, TyrR binds the consensus sequence TGTAAA-N 6 -TTTACA and requires the conserved G-N 14 -C pair [1]. The atypical spacing between the weak box arms in the akr promoter and the double tandem 5' arms suggests two possibilities for the spacing of the G-C pair: G-N 10 -C or G-N 16 -C. There is little information regarding protein binding to DNA sequences with atypical spacing between palindromic half-sites. The consensus binding sequence for PrrA, a global transcription factor controlling~25% of the genome in Rhodobacter sphaeroides [48], comprises two inverted half-sites separated by a spacer region of variable length, ranging from 0 to 10 base pairs [49]. Site directed mutagenesis coupled with promoter fusions and EMSAs revealed that while a five base pair spacer is optimal, PrrA also interacts with sites with an 8 base pair spacer [50].
Genome wide studies of transcription factor interactions with DNA have revealed that interactions with non-canonical sites, which were previously considered artifacts, are more common then previously thought and reflect true protein-DNA interactions. Chromatin immunoprecipitation coupled with microarray analysis (ChIP-chip) performed with the LexA regulator of E. coli demonstrated its interaction in vivo with several promoter regions lacking a discernible LexA DNA binding consensus sequence, while these sites were bound extremely poorly when studied in vitro [51]. ChIP-chip performed with the nitric oxide responsive NsrR regulator of E. coli revealed that the protein can recognize and bind to two types of binding sites: one comprised of two 11 base pair inverted repeats separated by a single nucleotide and the other contains only a single 11 base pair arm, with no discernible second inverted arm [52]. While in some cases, interactions between transcription factors and non-canonical sites have no discernible effect on gene transcription [51,53], we have shown that interaction between TyrR and the weak box results in the down-regulation of akr expression.
Supporting Information S1 Fig. Expression of ipdC in promoter mutants by TyrR in response to aromatic amino acids. Expression from ipdC promoter mutants in M9 minimal medium without amino acid supplements (A) and in the presence of phenylalanine (B), or tyrosine (C) in wild-type E. cloacae UW5 (grey) and tyrR null mutant E. cloacae J35 (black). Cells were assayed for β-glucuronidase activity in both logarithmic and stationary phases of growth. Error bars represent the standard error of the means of three independent replicates. Statistically significant differences of p < 0.05 are indicated by lowercase letters in logarithmic phase, and uppercase letters in stationary phase. Expression from akr promoter mutants in M9 minimal medium without amino acid supplements (A) and in the presence of phenylalanine (B), or tyrosine (C) in wild-type E. cloacae UW5 (grey) and tyrR null mutant E. cloacae J35 (black). Cells were assayed for β-glucuronidase activity in both logarithmic and stationary phases of growth. Error bars represent the standard error of the means of three independent replicates. Statistically significant differences of p < 0.05 are indicated by lowercase letters in logarithmic phase, and uppercase letters in stationary phase.