JB Accepts, published online ahead of print on 6 September 2013 J. Bacteriol. doi:10.1128/JB.00837-13 Copyright © 2013, American Society for Microbiology. All Rights Reserved. 1 2 Finely-tuned regulation of the aromatic amine degradation pathway in 3 Escherichia coli 4 5 D o w 6 Ji Zeng and Stephen Spiro* nlo a d 7 e d f r 8 om h 9 Department of Molecular and Cell Biology, The University of Texas at Dallas, 800 W t t p : / 10 Campbell Road, Richardson, Texas 75080 /jb . a 11 sm . o 12 r g / o 13 n A p 14 *Corresponding author: Stephen Spiro, Department of Molecular and Cell Biology, The r il 2 15 University of Texas at Dallas, 800 W Campbell Road, Richardson, Texas 75080, USA. , 2 0 1 16 Tel: +1 972 883 6896; Fax: +1 972 883 2409; E-mail: [email protected] 9 b y 17 g u e 18 s t 19 Running title: Regulation of aromatic amine degradation 20 21 Key words: Escherichia coli, FeaR, phenylethylamine, tyramine 22 1 23 ABSTRACT 24 FeaR is an AraC family regulator that activates transcription of the tynA and feaB genes 25 in Escherichia coli. TynA is a periplasmic topaquinone and copper containing amine 26 oxidase, and FeaB is a cytosolic NAD-linked aldehyde dehydrogenase. 27 Phenylethylamine, tyramine, and dopamine are oxidized by TynA to the corresponding D 28 aldehydes, releasing one equivalent of H2O2 and NH3. The aldehydes can be oxidized to o w n 29 carboxylic acids by FeaB, and (in the case of phenylacetate) can be further degraded to lo a d 30 enter central metabolism. Thus, phenylethylamine can be used as a carbon and nitrogen e d f r 31 source, while tyramine and dopamine can be used only as sources of nitrogen. Using o m 32 genetic, biochemical and computational approaches, we show that the FeaR binding site h t t p : 33 is a TGNCA-N8-AAA motif that occurs in 2 copies in the tynA and feaB promoters. We //jb . a 34 show that the co-activator for FeaR is the product rather than the substrate of the TynA s m . 35 reaction. The feaR gene is up-regulated by carbon or nitrogen limitation, which we or g / 36 propose reflects regulation of feaR by CRP and NAC, respectively. In carbon limited o n A 37 cells grown in the presence of a TynA substrate, tynA and feaB are induced, whereas in p r il 2 38 nitrogen limited cells only the tynA promoter is induced. We propose that tynA and feaB , 2 0 39 expression is finely tuned to provide the FeaB activity that is required for carbon source 1 9 b 40 utilization, and the TynA activity required for nitrogen and carbon source utilization. y g u 41 e s t 42 INTRODUCTION 43 In Escherichia coli, TynA is a periplasmic amine oxidase containing copper and 44 topaquinone cofactors (1). Aromatic amines including phenylethylamine (PEA), tyramine, 45 and dopamine are oxidized by TynA to the corresponding aldehydes, in a reaction that 2 46 releases one equivalent of H2O2 and NH3 (Fig. 1A). Therefore, these monoamines can be 47 used as the sole nitrogen source for growth. The aldehydes are further oxidized to the 48 corresponding carboxylic acids by FeaB, a cytosolic NAD-linked aldehyde 49 dehydrogenase (2). Phenylacetate (PA) can be further degraded to acetyl-CoA and 50 succinyl-CoA, and, therefore, PEA can be utilized as the sole carbon and energy source D 51 (3-6). In K-12 strains of E. coli, the carboxylic acids derived from tyramine and o w n 52 dopamine cannot be further catabolized, so these compounds can act only as nitrogen lo a d 53 sources. Therefore, while TynA activity may allow for both carbon and nitrogen e d f r 54 assimilation, FeaB activity is solely related to carbon and energy metabolism (3). Despite o m 55 the potentially different physiological roles of TynA and FeaB, currently available h t t p : 56 information suggests that their genes are coordinately regulated by the product of the // jb . a 57 linked feaR gene, which is a transcriptional regulator from the AraC family (2, 5, 7, 8). s m . 58 The AraC family includes over 800 members, most of which are thought to be or g / 59 transcriptional activators that function to regulate genes related to carbon metabolism, o n A 60 stress responses, or pathogenesis (9-11). With some exceptions, AraC family members p r il 2 61 are characterized by a conserved C-terminal DNA binding domain (CTD), and a non- , 2 0 62 conserved N-terminal domain (NTD). The non-conserved NTD contains the ligand 1 9 b 63 binding site, and, usually, the dimerization interface (9, 10). AraC family regulators that y g u 64 have been well characterized include AraC, MelR, XylS, RhaR, and RhaS (9, 12-18). e s t 65 FeaR is known to be required for the expression of tynA and feaB (7, 8), but its role and 66 mechanism have not otherwise been characterized. 67 Besides FeaR, there is some evidence that tynA and feaB expression may also be 68 modulated by other transcriptional regulators. We have previously shown that the nitric 3 69 oxide (NO) sensitive repressor NsrR binds to sites in the tynA and feaB promoters and has 70 a small effect on the transcription of these genes (8, 19). In addition, there is evidence 71 that feaR expression may be regulated by PhoB (20), and ArcA (P.J. Kiley, personal 72 communication). 73 In this paper, we use computational, genetic, and biochemical approaches to identify D 74 the FeaR binding site in the tynA and feaB promoter regions. We show that the FeaR o w n 75 CTD can bind to DNA in vitro and can activate the tynA promoter in vivo. In full length lo a d 76 FeaR, the NTD appears to act to inhibit the CTD in the absence of the co-activator. We e d f r 77 show that the expression of feaR is regulated by carbon or nitrogen limitation, and is not o m 78 subject to auto-regulation by the FeaR protein. Overall, we find that tynA expression is h t t p : 79 activated in both carbon and nitrogen limited cells in the presence of a FeaR co-activator, // jb . a 80 while feaB can be activated only during carbon limitation. We also show that the co- s m . 81 activator for FeaR is probably an aldehyde (the substrate for FeaB) rather than an amine or g / 82 (the substrate for TynA). o n A 83 p r il 2 84 MATERIALS AND METHODS , 2 0 85 Bacterial strains, growth media, and culture conditions. The strains and plasmids used 1 9 b 86 in this work are list in Table S2. The methods used to make gene knockouts, and to y g u 87 construct chromosomal promoter-lacZ fusions were described previously (8, 21-23). The e s t 88 glnG::kan and nac::kan mutations (in strain BW25113) were obtained from the Keio 89 collection, and then were transferred to the reporter strain by P1 transduction (21, 24). 90 DNA sequences encoding the CTD, and full-length FeaR were amplified by PCR 91 (primers are listed in Table S1) and ligated into pBAD24 (25). For β-galactosidase assays, 4 92 cultures were grown in rich medium (LB) or in defined medium (26) with the indicated 93 carbon and nitrogen sources. For growth with non-preferred nitrogen sources, ammonium 94 sulfate was substituted with sodium sulfate. For defined medium with PEA as the carbon 95 and nitrogen source (PEA medium), casamino acids (0.05% [wt/vol]) were also added. 96 Growth on PEA is temperature sensitive (6) and is significantly improved by the addition D 97 of casamino acids to growth media. PEA has limited solubility in water, so it was added o w n 98 directly to the bulk medium, which was then sterilized by filtration. Phenylacetaldehyde lo a d 99 (PAL) was solubilized in DMSO prior to addition to growth media. Because PAL is toxic e d f r 100 (and insoluble in aqueous buffers), it was added in 0.1 mM aliquots at 2 hour intervals o m 101 during the growth of cultures. h t t p : 102 Promoter analysis. The 250bp, 150bp, 142bp, 140bp, 133bp, 132bp, 129bp, and 126bp // jb . a 103 DNA fragments upstream of tynA start codon (tynA5-1 to tynA5-8), and 612bp, 250bp, s m . 104 109bp, 96bp, 89bp, 75bp, and 63bp DNA fragments upstream of the feaB start codon or g / 105 (feaB5-1 to feaB5-7) were amplified by PCR. The promoter fragments were cloned into o n A 106 pSTBlue-1 as described previously (22). Promoter fusions to lacZ were constructed in p r il 2 107 pRS415, transferred to λRS45, and integrated into the chromosome as described , 2 0 108 previously (22, 23). Mutations were introduced into the tynA5-1 clone using the 1 9 b 109 Invitrogen QuickChange Site-Directed Mutagenesis Kit, and appropriate mutagenic y g u 110 primers (Table S1). Mutant tynA promoters were fused to lacZ in pRS415 and then e s t 111 transferred to the chromosome (22, 23). 5’ transcription start sites were determined by 5’ 112 RACE, using the TAKARA 5’-full RACE Core Set according to the manufacturer’s 113 directions. The primers used for RACE are listed in Table S2. 5 114 Purification of the FeaR C-Terminal Domain (CTD). The C-terminal domain (CTD) 115 and the linker region of FeaR were identified by sequence alignment of five AraC family 116 proteins (FeaR, AraC, MelR, RhaR, and RhaS) using T-coffee (27). The DNA sequence 117 corresponding to the CTD and linker region was amplified by PCR, and ligated into pET- 118 21a(+) (Novagen) in frame with sequences encoding a C-terminal hexa-histidine tag. The D 119 recombinant plasmid was transformed into E.coli strain BL21(λDE3) for over-expression o w n 120 of the His-tagged CTD. CTDhis was purified using the His GraviTrap kit (GE lo a d 121 Healthcare). Protein concentrations were determined using the 660nm Protein Assay e d f r 122 Reagent (Pierce). o m 123 DNA Binding Assay. 5’ biotin-labeled tynA and control (ytfE) promoters were amplified h t t p : 124 by PCR and gel purified. DNA binding buffer (10 mM Tris, pH 7.5, 100 mM KCl, 1 mM // jb . a 125 DTT, 50 ng/uL poly dI.dC, 100 ng/uL salmon sperm DNA, 5% glycerol, 0.05% NP-40, s m . 126 0.5 mM EDTA, 200 ug/mL BSA) was incubated at the room temperature with or without or g / 127 CTDhis for 1 min (Pierce LightShift Chemiluminescent EMSA Kit). 1 nM biotin-labeled o n A 128 DNA was added to the solution and incubated for 20 min. Protein:DNA complexes were p r il 2 129 then resolved on 8% polyacrylamide gels. The biotin-labeled DNA was transferred to a , 2 0 130 Biodyne B membrane (Pall Corporation), and then detected using the Chemiluminescent 1 9 b 131 Nucleic Acid Detection Module (Pierce). y g u 132 The DNA binding activity of the FeaR CTD was also assayed by fluorescence e s t 133 anisotropy (28, 29). The ROX-labeled 31nt site1 and site2 contain 21nt of the first and 134 second repeat of the FeaR binding site, respectively, flanked by 5nt upstream and 135 downstream of the full length FeaR binding site. Complementary DNA strands were 136 annealed by heating at 950C for 2min in TE buffer, then cooling to 250C (at a rate of 10C 6 137 per min). ROX-labeled DNA fragments (5 nM) were incubated with 3 mL FA buffer (10 138 mM Tris, pH 7.4, 200 mM KCl, 1 mM EDTA, 5% glycerol, 25 ug/mL BSA, 75 ug/mL 139 salmon sperm DNA) for 10min, then CTD (5 nM-1300 nM) was added and the reaction 140 incubated for 2 min. The anisotropy change was measured in a Varian Cary Eclipse 141 Fluorimeter. The binding isotherm was fit to equation 1 (28) using Kaleidagraph D 142 (Synergy Software). o w n 143 ΔA= {ΔAT (CTDnH/Kd nH)/ (1+CTDnH/Kd nH) Equation 1 lo a d 144 Where ΔA is the change in fluorescence anisotropy, ΔAT is the total change in anisotropy, ed f r 145 CTD is the total protein concentration in each point in the titration, Kd is the dissociation o m 146 constant, and nH is the Hill coefficient. h t t p : 147 For competition assays, 345 nM CTD was incubated in 3 mL FA buffer for 5min, then // jb . a 148 5 nM ROX-labeled site1 or site2 was added, and the reaction incubated for a further 10 s m . 149 min. Unlabeled competitor DNAs (16-1500 nM competitors, see Table 2) were added, or g / 150 and the anisotropy change was measured after equilibration for 4min. The data were fit to o n A 151 equation 2 (29). p r il 2 152 Fraction bound = FBmax [1-([competitor] / (IC50 + [competitor]))] Equation 2 , 2 0 153 FBmax is the fraction bound in the absence of competitor, IC50 is the concentration of 19 b 154 competitor required for half-maximal inhibition of binding, and the fraction bound is y g u 155 defined according to Equation 3 (29). e s t 156 Fraction Bound = (ΔA-Afree) / (ΔAT-Afree) Equation 3 157 Where Afree is the anisotropy in the absence of protein. 158 159 7 160 RESULTS 161 Regulation of the feaR promoter In order to study the regulation of feaR, feaB and tynA, 162 the promoters of the three genes were fused to lacZ, and the fusions were transferred to 163 the E. coli MG1655 chromosome (8). β-galactosidase activities were measured in 164 cultures grown in defined media with different carbon and nitrogen sources (8, 26). In D 165 some cases, TynA substrates (PEA or tyramine) were used as the sole source of nitrogen, o w n 166 and/or these were added as inducers to media also containing other nitrogen sources lo a d 167 (Table 1). The feaR promoter showed a basal level of activity in defined medium with e d f r 168 glucose as the sole carbon source and ammonia as the sole nitrogen source (preferred o m 169 medium). The promoter activity increased about 2 fold when the carbon source was h t t p : 170 replaced by glycerol (glycerol medium), and 4-6 fold when the nitrogen source was // jb . a 171 glutamine, alanine, or tyramine (glutamine, alanine or tyramine medium). Addition of s m . 172 tyramine to the preferred medium or deletion of feaR had no effect on feaR promoter or g / 173 activity (Table 1), indicating that there is no autoregulation of feaR expression. Growth o n A 174 with glucose as the carbon source and tyramine as the nitrogen source was not possible, p r il 2 175 perhaps reflecting glucose repression of feaR expression (see below). , 2 0 176 The transcription start site of feaR was determined by Rapid Amplification of 5’ 1 9 b 177 cDNA Ends (RACE). According to the RACE results, feaR transcription initiates from y g u 178 three sites: Pm (m for minor), P1, and P2 that are located 111, 66, and 26 bp upstream of es t 179 the translation initiation codon, respectively (Fig. 1B; Fig. 2A). Based on the frequency 180 of different clones recovered from the RACE procedure, all three sites are used in the 181 preferred medium (6% of clones started at Pm, 38% at P1, and 56% at P2), while P1 is the 182 dominant promoter used in glycerol medium (78% P1, 22% P2), and P2 is the only 8 183 promoter used in glutamine medium and in cells grown on defined medium with PEA as 184 the sole carbon source (PEA medium) (Fig. 2A). A sequence that is a good match to the 185 CRP binding site is centered at 71.5 bp upstream of the P1 promoter, which is suggestive 186 of a class I type activation mechanism by CRP-cAMP (30-32). Regulation of P1 by CRP- 187 cAMP would be consistent with the preferential utilization of this promoter in glycerol D 188 medium. In contrast, the P2 promoter is used preferentially in cells grown on a non- o w n 189 preferred nitrogen source, which is suggestive of regulation by NtrC or NAC (33-36). lo a d 190 Up-regulation of feaR in glutamine medium was abolished in ntrC (glnG) and nac e d f r 191 mutants (Fig. S1), which is consistent with NAC acting as a direct regulator of feaR, o m 192 since nac expression is NtrC dependent(33-37). Accordingly, there is a predicted NAC h t t p : 193 binding site associated with the P2 promoter (Fig. 1B; Fig. 2A), and no predicted binding //jb . a 194 sites for NtrC or σ54. s m . 195 Regulation of the feaB and tynA promoters The feaB promoter showed a relatively low or g / 196 activity unless a substrate for the TynA/FeaB pathway was present in the growth medium o n A 197 (Table 1). Thus, there is not a simple correlation between feaR expression and the activity p r il 2 198 of its target promoter. The likely explanation is that a pathway substrate or intermediate is , 2 0 199 required to act as the co-activator for FeaR. Also, activation of the feaB promoter above 1 9 b 200 its basal level required growth on a non-glucose carbon and energy source (for example, y g u 201 compare preferred medium + tyramine and glycerol medium + tyramine, Table 1). In e s t 202 medium with a non-preferred nitrogen source (glutamine medium + tyramine), feaB 203 activity remained low. Two transcription start sites were mapped 149 and 27 bp upstream 204 of the feaB translation initiation codon, named Pm and P1, respectively. The Pm promoter 205 was used in the preferred medium; while only P1 was used in cells growing on PEA 9 206 medium. This pattern of promoter utilization is consistent with the presence of predicted 207 FeaR and CRP binding sites upstream of the P1 promoter (Fig. 1B; Fig. 2B). 208 Unlike the feaB promoter, the tynA promoter was almost silent under non-inducing 209 conditions. Activation of tynA required the presence of either tyramine or PEA in the 210 medium (tyramine medium, PEA medium, or glycrol medium with tyramine or PEA). D 211 The requirement for an inducer for tynA promoter activity is consistent with our detection o w n 212 of only a single transcription start site, which is associated with FeaR binding sites (Fig. lo a d 213 1C; Fig. 2C). Unlike feaB, tynA activity could be elevated above its basal level by e d f r 214 addition of tyramine to the glycerol medium or glutamine medium (Table 1). Thus, in the o m 215 presence of an inducer, tynA expression is elevated in cells growing on non-preferred h t t p : 216 carbon and nitrogen sources. // jb . a 217 Activity of the feaR promoter was at basal levels under anaerobic conditions in all s m . 218 growth media tested (Table 1). Accordingly, tynA could not be induced by pathway or g / 219 substrates in anaerobic cultures, and feaB promoter activity was consistently lower than o n A 220 observed in aerobic cultures. We conclude that expression of the PEA pathway is shut p r il 2 221 down during anaerobic growth, which can be explained by the recent identification of , 2 0 222 feaR as a target for ArcA regulation (P. J. Kiley, personal communication). 1 9 b 223 Overall, our data show that the feaR gene is up-regulated during growth on non- y g u 224 preferred carbon and nitrogen sources. This increase in feaR expression is not sufficient e s t 225 to activate expression of tynA and feaB, unless a pathway inducer is also present 226 (although a non-physiological increase in FeaR abundance may lead to induction of its 227 targets in the absence of inducer, see Fig. S3). Elevated levels of feaB expression require 228 growth on glycerol and either PEA or tyramine, while tynA can be induced by PEA or 10