Sensitive and selective phenol sensing in denitrifying Aromatoleum aromaticum EbN1T

ABSTRACT Aromatoleum aromaticum EbN1T anaerobically degrades phenol, p-cresol, and p-ethylphenol, each via a distinct peripheral pathway. The compound-specific regulation of each pathway is proposed to occur on the transcriptional level in the case of phenol supposedly mediated by the one-component system PheR. To confirm its predicted function, we generated an unmarked, in-frame deletion mutant (ΔpheR). This mutant did not express the ppsA1 gene, which encodes the A1 subunit of phenol-activating phenylphosphate synthase. The expression of ppsA1 was restored by in trans complementation of pheR into the ΔpheR background. The responsiveness to phenol was studied in vivo in benzoate-limited anaerobic cultures by adding, upon benzoate depletion, single defined pulses of phenol (from 100 µM down to 0.1 nM). Time-resolved, targeted transcript profiling by qRT-PCR revealed a response threshold for ppsA1 expression of 30‒50 nM phenol. Notably, ppsA1 expression could not be induced by p-cresol or p-ethylphenol. Conversely, lack of expression was also observed for the additional target genes cmh (p-cresol degradation) and acsA1 (p-ethylphenol degradation) applying phenol or p-ethylphenol as well as phenol or p-cresol as stimuli. Thus, the sensory proteins PheR, PcrS, and EtpR should be highly selective for phenol, p-cresol, and p-ethylphenol, respectively. The implicated incapability of cross-stimulus binding was corroborated by comparing the predicted 3D structural models of the proteins’ sensory domains. While the ligand-binding pockets share the conserved hydroxy group-anchoring histidine and tryptophane, their distal faces in PcrS and EtpR are, compared to PheR, enlarged to accommodate the bulkier methyl (p-cresol) and ethyl group (p-ethylphenol), respectively. IMPORTANCE Aromatic compounds are globally abundant organic molecules with a multitude of natural and anthropogenic sources, underpinning the relevance of their biodegradation. A. aromaticum EbN1T is a well-studied environmental betaproteobacterium specialized on the anaerobic degradation of aromatic compounds. The here studied responsiveness toward phenol in conjunction with the apparent high ligand selectivity (non-promiscuity) of its PheR sensor and those of the related p-cresol (PcrS) and p-ethylphenol (EtpR) sensors are in accord with the substrate-specificity and biochemical distinctiveness of the associated degradation pathways. Furthermore, the present findings advance our general understanding of the substrate-specific regulation of the strain’s remarkable degradation network and of the concentration thresholds below which phenolic compounds become essentially undetectable and as a consequence should escape substantial biodegradation. Furthermore, the findings may inspire biomimetic sensor designs for detecting and quantifying phenolic contaminants in wastewater or environments.

P henolic compounds are widespread constituents of dissolved organic matter in the terrestrial and marine realm, with continuous input from natural as well as anthropo genic sources.They represent major building blocks of abundant biopolymers such as the plant cell wall inherent lignin (1) and are essential for sclerotizing the exoskeletons of insects (2).Phenolic compounds, as significant constituents of fossil energy sources, are released during extraction and processing [e.g., reference (3)].Moreover, they are industrially produced in large quantities and represent bulk starting material for widely used phenolic resin-based plastics (4).Environmental and health concerns arise from their toxicity, carcinogenicity, and biologic resistance particularly in anoxic habitats (5), accounting for the need to fundamentally understand the biodegradation of these compounds, including the involved sensory/regulatory mechanisms.
Against the backdrop of the biochemical specificities of the aforementioned enzymes, their differential abundance and activity profiles (7,14,16), and the predicted sen sory/regulatory proteins encoded in close proximity of the pathways' genes (8,17), it stands to reason that highly selective sensory recognition of phenol, p-cresol, and p-ethylphenol should govern a tight differential transcriptional control of the respective degradation pathways in A. aromaticum EbN1 T (Fig. 1).(i) The pps-ppc gene cluster for anaerobic phenol degradation harbors a σ 54 consensus sequence in its promoter region and is directly flanked by the pheR gene, which encodes a predicted σ 54 -depend ent one-component system (activator).Furthermore, the predicted PheR protein is homologous to the known phenol sensors MopR (from Acinetobacter calcoaceticus), CapR (also called DmpR, from Pseudomonas putida), and PoxR (from Ralstonia eutropha).Their determined crystal structures revealed conserved histidine and tryptophane residues in the N-terminal V4R domain, which anchor via hydrogen bridges the hydroxy group of phenol (18)(19)(20)(21).(ii) The genes (cmh, hbd, and hcl2) assigned to anaerobic p-cresol degradation are intercalated by those for the predicted σ 54 -dependent two-component system PcrSR (activator), while the promotor region upstream of the cmh gene harbors a σ 54 consensus promoter (13).(iii) The genes (acsA1 through emhCF) for anaerobic p-ethylphenol degradation are preceded upstream by a σ 54 consensus promoter and a gene encoding the σ 54 -dependent one-component system EtpR (repressor) (14).Notably, the sensory domains of PcrS and EtpR also contain a vinyl 4 reductase (V4R) domain with the conserved histidine and tryptophane residues, suggesting an analogous anchoring of their respective phenolic ligand.The predicted roles of PcrSR and EtpR as transcrip tional regulators of their respective pathways were recently corroborated by in-frame deletions of their coding genes (13,22).Moreover, the in vivo responsiveness of A. aromaticum EbN1 T toward p-cresol and p-ethylphenol (and its degradation intermediate p-hydroxyacetophenone), respectively, was recently shown to be in the lower nanomolar range (13,23).
To advance the understanding of the regulation of the anaerobic degradation of phenolic compounds in A. aromaticum EbN1 T , the present study pursued the following major aims: (i) substantiate the predicted role of PheR as transcriptional regulator of the "phenol-catabolic" genes by molecular genetic approaches; (ii) determine the in vivo tional control of the "catabolic" gene cluster for phenol is proposed to be mediated by the predicted one-component system PheR.In the case of p-cresol and p-ethylphenol, expression of their "catabolic" gene clusters is governed by the two-component system PcrSR and the one-component system EtpR, respectively (13,22).Functional domains of PheR, PcrSR, and EtpR are color coded, and their coding genes are marked in dark gray."Catabolic" genes selected for transcript responsiveness toward phenol; and (iii) investigate the sensory discrimination between phenol, p-cresol, and p-ethylphenol by differential expression profiling and comparing predicted 3D models of the respective sensory domains.

Generation of ∆pheR and pheR-complemented mutants
To verify the predicted role of the one-component system PheR in controlling the phenol-dependent expression of the "phenol-catabolic" gene cluster (Fig. 1A), a mutant with an unmarked in-frame ΔpheR deletion was generated.In the mutant, only the start and stop codons of the pheR gene were preserved in order to maintain the reading frame (Fig. 2A, marked in green and red, respectively).Accordingly, no amplicon could be observed using the primers specific for pheR in the deletion mutant, compared to the expected 258 bp product in the wild type.Using primers hybridizing up-and downstream of the deleted region, only a 1,115-bp amplicon was generated compared to a 2,846-bp product in the wild type (Fig. 2B).The newly generated ΔpheR mutant was in trans-complemented with the broad-host-range plasmid pBBR1 MCS-2 carrying the pheR gene.For the pheR-complemented mutant, a PCR product of 258 bp was obtained with gene-specific primers, which validates successful complementation with the pheR gene (Fig. 2B).The correctness of both mutants was verified by nucleotide sequencing.
Compared to the wild type, the ΔpheR mutant showed a similar growth behavior with benzoate and corresponding substrate consumption profile (Fig. 2C), indicating that the in-frame deletion does not affect the general growth physiology of A. aromaticum EbN1 T .The pheR-complemented mutant attained the maximal optical density (OD max ) ~5 h later than the other two genotypes, which is most likely due to the adverse effect of kanamycin added for selection reasons to the medium of this mutant.

PheR mediates the expression of "phenol-catabolic" genes
The responsiveness to phenol of the ΔpheR and pheR-complemented mutants, compared to that of the wild type, was studied by the following two-stage experiment.First, non-adapted cells of A. aromaticum EbN1 T were anaerobically grown with a limited supply of benzoate (1 mM), upon the depletion of which (after ~17.2 h of incubation) a single pulse of 100 µM phenol was administered.Samples were retrieved 5 min prior to this pulse (reference) as well as 5, 15, 30, 60, 120, and 180 min after the pulse (test states).Second, for targeted transcript analysis by quantitative reverse transcription-PCR (qRT-PCR), the ppsA1 gene located at the beginning of the "phenol-catabolic" gene cluster (Fig. 1A) was selected.Fold changes of transcript abundances are provided in Table S1 and illustrated in Fig. 2D.The lack of clear ppsA1 expression in the ΔpheR mutant compared to the wild type upon the phenol pulse agrees with its predicted function as phenol-responsive transcriptional activator of the "phenol-catabolic" gene cluster.This is further corroborated by ppsA1 expression exceeding the wild-type levels in the pheR-complemented mutant, which is probably due to the high-expression property of the involved pheR-carrying pBBR1 MCS-2 vector.Similar observations were previously reported in the context of mutant studies on p-cresol-sensing PcrSR (13) and p-ethylphe nol-sensing EtpR (22).

In vivo response threshold toward phenol
The response threshold of A. aromaticum EbN1 T wild type toward phenol was deter mined by applying the same two-stage experimental setup as described above for the mutant characterization.However, here, eight distinct phenol concentrations (from 100 µM down to 0.1 nM) were tested, for each of which an independent growth experiment with three replicate cultures was conducted (reproducibility documented in Fig. S2).Responsiveness was assessed by determining the expression profile of the ppsA1 gene across the eight tested phenol concentrations and six test time points (Fig. 3A; Table S1).Constant levels of ppsA1 transcripts at the reference time point (5 min prior to pulse) across all tested phenol concentrations afforded reliable comparison.A significant increase in ppsA1 expression was observed already 5 min after 100 µM phenol was added, reaching its highest fold change (16-fold) after 120 min (Fig. 3A).The lowest pulse concentration still yielding a detectable transcriptional response was 50 nM, indicating the in vivo response threshold to range between 30 and 50 nM phenol.This threshold compares fairly well to those previously determined with A. aromaticum EbN1 T for other phenolic compounds (13) and various phenylpropanoids (24), which all fall in the lower nanomolar range (Fig. 3B).

Physical considerations for cellular phenol uptake
Considering its uptake from the environment, phenol with its lipophilic character can be assumed to enter cells of A. aromaticum EbN1 T via passive diffusion across the cell envelope.Earlier theoretical considerations (13) suggest that equilibration between extra-and intracellular concentrations of alkylphenols occurs on millisecond time scales and is aided by their lipophilic nature, which enhances the effective permeability coefficient of the lipid bilayer membranes in the envelope by a factor of K M (membrane partition coefficient).Compared to the two alkylated phenols p-cresol and p-ethylphenol studied earlier (13), phenol clearly has the lowest effective diffusion coefficient D eff for passive diffusion across lipid bilayers (Table 1), which implies proportionately longer equilibration times, but still in the millisecond range.The experimentally determined threshold concentration for phenol is distinctly higher (30-50 nM) compared to p-cresol and p-ethylphenol (1-10 nM) (Fig. 3B), which we ascribe to lower binding affinity between phenol and its cognate sensor protein for the lack of other obvious reasons.Assuming that the threshold concentrations solely reflect differences in sensor-ligandbinding affinities, we can provide rough first estimates of the K D values (Table 1), which also depend on the number of copies of a sensor protein in the cell before induction [for details, see reference (13)].For essential proteins, a value of 10 copies per cell was found in E. coli (25), and we assume that of the 10 copies of a sensor protein present, at least 1 copy be bound to its ligand in order to have an effect on transcriptional regulation.This consideration leads to a K D value in the submicromolar range.

Negligible gratuitous induction by phenolic compounds
The chemical similarity of the three phenolic compounds supporting anaerobic growth of A. aromaticum EbN1 T prompts the study of their potential to cross-induce transcription of their non-associated "catabolic" gene clusters.This was tested by applying the same two-stage experimental setup as described above, but adding next to pssA1 also cmh (for p-cresol) and acsA1 (for p-ethylphenol) as target genes and next to phenol also p-cresol and p-ethylphenol as compounds to be tested.Since threshold determination was not relevant in this context, only three concentrations (100, 1, and 0.1 µM) were applied per compound.Overall, the transcript profiles delivered a uniform picture across the nine gene-vs-compound conditions (Fig. 4).Essentially, no expression of the pssA1 gene could be detected when single pulses of p-cresol or p-ethylphenol were administered, except for a faint transcript formation at 100 µM p-ethylphenol recognizable only 60 min from wild type, ΔpheR mutant, and pheR-complemented mutant.Applying the primer pair ΔpheR_F/R a 2,846 bp amplicon was obtained from the wild type, whereas the ΔpheR and complemented mutants both yielded 1,115 bp amplicons.Applying the primer pair pheR_468_F/752_R a 258-bp amplicon was obtained from wild type and pheR-complemented mutant, whereas no amplicon was obtained with the ΔpheR mutant.(C) Anaerobic growth of the ΔpheR and pheR-complemented mutants compared to the wild type under nitrate-reducing conditions with benzoate as sole source of carbon and energy.Upon benzoate depletion, a single pulse with phenol (100 µM) was given to each genotype (dashed lines).(D) Time-resolved transcript profile of the ppsA1 gene in response to phenol, comparing wild type with the ΔpheR and pheR-complemented mutants.The ppsA1 gene encodes the subunit A1 of phenylphosphate synthase, the initial enzyme of anaerobic degradation of phenol (see Fig. 1).qRT-PCR was used to determine relative transcript abundances, with the time point 5 min prior to phenol addition serving as reference.Determined fold changes rely each on three biological and three technical replicates and are provided in Table S1.
after the pulse.Expression of the cmh gene was clearly observed only ~60 min after the 100 µM phenol pulse and was merely close to background noise with p-ethylphenol.However, the essentially congruent cmh transcript abundance profiles at 100 µM phenol observed for ΔpheR mutant and wild type (Fig. S3) preclude a contribution of phenolbound PheR to cmh expression and rather suggest a poor phenol-PcrS interaction only at higher ligand concentrations and after longer incubation time.Finally, low-level  S1.Each fold change relies on three biological and three technical replicates each.
Growth cultures providing samples for the eight studied phenol concentrations as well as for the control (no addition of effector) are documented in Fig. S2.(B) Comparison of the response threshold determined here for phenol with those for other phenolic compounds and phenylpropanoids previously reported (13,23,24).Compound names: 1, phenol; 2, p-cresol; 3, p-ethylphenol; 4, 4-hydroxyacetophenone; expression of the acsA1 gene was only observed with 100 µM phenol or p-cresol.
Considering the high-level expression of the target genes in response to the cognate phenolic compound, the aforementioned rare instances of cross-induced expression levels are ~1-2 orders of magnitude lower (Table S1).Taken together, a comparison of the transcript profiles suggests that the three tested phenolic compounds do not (or only very weakly) serve as gratuitous inducers for the non-cognate "catabolic" gene clusters.This could result from highly selective ligand-binding properties of the three involved sensory domains of the regulatory proteins (PheR, PcrS, and EtpR) as presented in the subsequent section "Predicted 3D structural models of sensory domains." To elaborate on the apparent lack of gratuitous induction by the three phenolic compounds, targeted transcript profiling was further expanded to the bssA and ebdA1 genes of A. aromaticum EbN1 T (Fig. S4A).These encode the catalytic subunits of benzylsuccinate synthase and ethylbenzene dehydrogenase, which are key enzymes of the anaerobic degradation of toluene and ethylbenzene, respectively (Fig. S4B) (28,

29
). Essentially, no significant expression of the bssA and ebdA1 genes was observed, irrespective of the phenolic compound tested.This suggests that the involved sensory proteins TdiS (for toluene) and EdiS (for ethylbenzene), which are part of specific two-component systems (Fig. S5A), are not responsive to the phenolic compounds tested and that the regulatory domains of the PheR, PcrR, and EtpR proteins do not interact with the promoter regions of the "toluene-and ethylbenzene-catabolic" gene clusters.

Predicted 3D structural models of sensory domains
To better understand the apparently high selectivity of the sensory domains of the PheR, PcrS, and EtpR proteins from A. aromaticum EbN1 T , insights into their ligand-bind ing were aimed at.For this purpose, crystal structures of the sensory domains from the known phenol sensors MopR, CapR, and PoxR from other bacteria (18-21) were compared with 3D models of their counterparts in PheR, PcrS, and EtpR, which were generated by applying the recently available AlphaFold tool (30).
The multiple sequence alignment of the sensory domains of the six proteins (Fig. 5A; Fig. S5B) reveals, next to overall high similarities, several conserved residues that are functionally and structurally essential for the ligand-binding pocket: (i) a dyad His and Trp anchors the hydroxy group of phenol via hydrogen bonds, (ii) several hydrophobic residues that stabilize the aromatic ring of phenol via van der Waals interactions, and (iii) three Cys and one Glu residue(s) that are involved in tetrahedral zinc coordination for maintaining structural integrity.The overall fold of the sensory domains of the MopR dimer from Acinetobacter calcoaceticus, as revealed by its crystal structure (2.30 Å resolution) (19), is shown in Fig. 5B, highlighting bound phenol and zinc as well as indicating the secondary structural elements.The structural congruence of the sensory domains of MopR, CapR, PoxR, PheR, PcrS, and EtpR, in particular of the core area, is illustrated by an overlay image of the respective monomers in Fig. 5C.This forms a solid basis to reliably compare the ligand-binding pockets of the six sensory proteins in more detail.
The experimentally determined structure of the phenol-binding pocket from MopR is enlarged in Fig. 6A (upper panel).Special emphasis is placed on the hydroxy groupanchoring Trp 134 and His 106 residues, the aromatic ring stabilizing hydrophobic residues (Phe 99 , Pro 103 , Val 112 , Val 114 , Phe 132 , Tyr 161 , Ala 162 , Tyr 165 , Tyr 176 , and Ile 191 ), the Ser 166 , closing the pocket, as well as the shape of the ligand-accommodating cavity (transparent gray surface).The reliability of AlphaFold structures regarding the prediction of overall folds and binding pockets of the sensory domains becomes evident from their direct comparison with the respective crystal structures of the MopR, CapR, and PoxR dimers (Fig. S6 and S7).The structural congruency particularly applies to the decisive core area but is less pronounced for the functionally less relevant, more loosely structured N-and C-termini.The AlphaFold-predicted binding pockets of PheR, PcrS, and EtpR of A. aromaticum EbN1 T have an overall analogous shape to that of MopR, including the spatial positioning of the anchoring His and Trp residues (Fig. 6A, lower panel).While the ligand cavities of MopR and PheR are almost completely superimposable, in the case of PcrS and EtpR, differently sized spatial expansions are present at the opposite of the anchoring face (Fig. 6B).These expansions aptly accommodate the bulky methyl group of p-cresol (PheR) and ethyl group of p-ethylphenol (EtpR), while corroborating the observed non-cross induction by the three studied phenolic compounds in A. aromati cum EbN1 T .
Furthermore, structural considerations also provide explanations for the apparent non-responsiveness of toluene-specific TdiS and ethylbenzene-specific EdiS sensory proteins toward the three tested phenolic compounds in A. aromaticum EbN1 T .First, the V4R domain, which harbors the conserved, phenol-anchoring His and Trp residues, is absent from the sensory domains of TdiS and EdiS (Fig. S5A).Second, the dimeric structures of the sensory domains of phenol-sensing CapR [6IY8; reference (18)] and toluene-sensing PAS1 [5HWV; reference (31)], both from Pseudomonas putida and in the ligand-bound state, are highly dissimilar (Fig. S5C).

Conclusions
A. aromaticum EbN1 T anaerobically degrades phenol, p-cresol, and p-ethylphenol via distinct peripheral reaction sequences, each comprising specific and, for the most part, biochemically intriguing enzymes.Therefore, it is economic for the cell to avoid the unproductive synthesis of such highly specialized enzymes if the native substrate is not available or present only at very low levels.A. aromaticum EbN1 T ensures this efficiency by means of highly substrate-specific, non-overlapping (non-promiscuous) regulatory circuits on the transcriptional level.The present study reveals that these circuits are controlled directly upfront by sensory/regulatory proteins (PheR, PcrSR, and EtpR), which are specifically tailored to recognize their cognate phenolic substrate based on its characteristic chemical feature.
The determined threshold of A. aromaticum EbN1 T for transcriptional responsiveness toward phenol under the applied cultivation conditions is in the nanomolar range (30-50 nM) and somewhat higher than those previously reported for p-cresol and p-ethyl phenol [1-10 nM; reference (13)].Such low thresholds possibly shed new light on the controlling factors for the persistence of dissolved organic matter as well as for the lower limits of bacterial growth with aromatic compounds.Furthermore, in synopsis with the here studied ligand-selectivity (non-promiscuity) of the involved sensory proteins, it appears unlikely that mixtures of phenolic growth substrates each available at nanomo lar to lower micromolar concentrations could exert an amplifying effect on the sensory process.

Bacterial strain and cultivation conditions
A. aromaticum EbN1 T (DSM 19018 T ) has been maintained in our laboratory since its isolation and was cultivated under nitrate-reducing conditions in defined, ascorbatereduced, and bicarbonate-buffered mineral medium at 28°C as previously described (32).Aromatic growth substrates (benzoate, phenol, p-cresol, and p-ethylphenol) were added from aqueous stock solution sterilized by filtration.All cultivation experiments were started from glycerol stocks of the respective genotype of A. aromaticum EbN1 T , following a multi-step procedure as recently described (13).Main cultures were performed in triplicates using 500 mL flat-bottomed glass bottles (400 mL culture volume) sealed with butyl rubber stoppers under an anoxic N 2 /CO 2 (90:10, vol/vol) atmosphere, supplied with benzoate (1 mM) as sole source of carbon and energy, and inoculated with 2% (vol/vol) of active preculture.Cultures of the pheR-complemented mutant were supplemented with kanamycin (50 µg mL -1 ) as selective marker.Growth was monitored by measuring the optical density of the cultures at 660 nm (OD 660 ) (UVmini-1240; Shimadzu, Duisburg, Germany).Culture samples were retrieved, and substrate pulses were administered with sterile, N 2 -flushed syringes.All chemicals were of analytical grade.

Generation of in-frame ΔpheR deletion mutation
Genomic DNA and plasmids were isolated according to standard protocols as previously described (33).For unmarked in-frame deletion of the pheR gene, a knockout plasmid was constructed on the basis of the suicide plasmid pK19mobsacB (34).The final knockout plasmid contained 1.3 kbp of the upstream and 1.2 kbp of the downstream regions of the pheR gene.Both inserts were cloned simultaneously into pk19mobsacB using the In-Fusion HD cloning plus kit (TaKaRa Bio Inc., Kusatsu, Japan) according to the manufacturer's instructions and using the PstI restriction site.For the cloning reaction, 100 ng of linearized pK19mobsacB and a molar ratio of 6:6:1 of the two inserts relative to the linearized plasmid were used.The respective primers were designed using the In-Fusion cloning primer design tool (TaKaRa Bio Inc.) and are given in Table 2. Homologous regions were amplified by PCR from genomic DNA of A. aromaticum EbN1 T using a high-fidelity polymerase (Phusion; Thermo Fisher Scientific, Dreieich, Germany).The final knockout vector construct maintained the start and stop codons of the pheR gene.The knockout vector was transferred from Escherichia coli S17-1 to A. aromaticum EbN1 T via conjugation as described previously (33).Kanamycin-resistant single-crossover clones were verified by PCR.The single-crossover mutant was transferred into liquid medium (4 mM benzoate and 5 mM acetate) without kanamycin to induce the second crossover event, yielding either the ΔpheR or the wild-type genotype.Cells were plated on solid medium containing the same substrate mixture and 5% (wt/vol) sucrose.For identification of the ΔpheR genotype, colonies were screened by PCR, yielding only an amplicon of 1,115 bp (Fig. 2AB).

Complementation of pheR in trans into the ΔpheR mutant
To achieve in trans expression of the pheR gene in the ΔpheR background, a complemen tation vector based on the broad-host-range vector pBBR1 MCS-2 conferring resistance to kanamycin (35) was generated.For this purpose, a 2.2-kbp fragment containing the pheR gene together with its ribosomal binding site was amplified from genomic DNA of A. aromaticum EbN1 T using a high-fidelity polymerase (Phusion; Thermo Fisher Scientific) and the PheRcompIn_Mcs2XbaI_F/R primer pair, containing an XbaI restriction sites (Table 2).Purified PCR product and vector were digested with the restriction enzyme XbaI and purified.For ligation, digested PCR product and plasmid were mixed in a 4:1 ratio, denatured for 10 min at 65°C, and cooled on ice.Subsequently, T4-ligase (1 U) was added, mixed carefully, and incubated for 140 min at 24°C.The ligation reaction was transformed into chemical competent E. coli HST08.Single colonies grown on selective solid medium were screened and plasmids from positive clones were purified.The vector was transferred through a conjugation from the E. coli S17-1 background via agar-plate mating to the ΔpheR mutant, yielding the pheR-complemented mutant (genotype: ΔpheR, pBBR1 MCS-2 ΩpheR) and verified by PCR (Fig. 2B).

Sequence validation of mutants by Sanger sequencing
For sequence validation of the ΔpheR mutant, a 2.5-kbp region of genomic DNA spanning across the entire up-and downstream regions, including the deletion site, was analyzed via Sanger sequencing as previously described (13).Additionally, a 2.7-kbp region of the complementation vector (ΔpheR, pBBR1 MCS-2 ΩpheR) spanning across the entire insert was sequenced to verify the pheR-complemented mutant.Genomic DNA was prepared according to standard methods (33).The fragments were amplified from genomic DNA and the complementation vector (pBBR1 MCS-2 ΩpheR) via PCR using a high-fidelity polymerase (Phusion; Thermo Fisher Scientific) according to the manufacturer's instructions and using the primer pairs provided in Table S2 (1F and 5R of each).For sequencing, samples were prepared using the BigDye buffer and BigDye terminator sequencing reagent and analyzed using a 3130xl genetic analyzer (Applied Biosystems, Waltham, MA, USA) as described by the manufacturer.

Growth experiments for mutant characterization
To assess the effect of the pheR deletion, the three genotypes (wild type, ΔpheR mutant, and pheR-complemented mutant) of A. aromaticum EbN1 T were grown with benzoate (1 mM).Replicate cultures were sampled in short intervals to monitor growth (OD 660 ) and substrate consumption.After complete depletion of benzoate, a pulse of 100 µM phenol was given.For each experiment and replicate, a sample taken 5 min prior to the pulse served as reference state, while samples retrieved 5, 15, 30, 60, 120, and 180 min after the pulse represented the test states.Five milliliters of samples was immediately added to 10 mL of RNAprotect bacterial reagent (Qiagen, Hilden, Germany), mixed rigorously, incubated for 5 min at room temperature, and then centrifuged (4,000 × g, 10 min, 4°C).Pellets were resuspended in 0.5 mL RNAprotect bacterial reagent, transferred into 2 mL microcentrifuge tubes, and centrifuged (20,000 × g, 10 min, room temperature).Supernatants were discarded, and pellets were shock frozen in liquid N 2 and stored at -80°C until further analyses.

Growth experiments for determining phenol response threshold
The responsiveness of A. aromaticum EbN1 T to various concentrations of phenol was studied with non-adapted cells.Cultivation was performed as described above, with three replicate cultures performed per tested phenol concentration.Upon depletion of the primary growth substrate benzoate (highly reproducible after ~17.2 h of incubation; Fig. S2), a single pulse of phenol at a defined concentration was administered (100 µM, 1 µM, 100 nM, 50 nM, 30 nM, 10 nM, 1 nM, or 0.1 nM).A negative-control experiment without the addition of phenol was performed under the same conditions.Throughout the incubation time, 3 mL samples of the culture broth were retrieved to monitor growth: a 1 mL aliquot for measuring OD 660 and a 2 mL aliquot for determining benzoate depletion.In the case of the latter, the samples were immediately centrifuged (20,000 × g, 10 min, 4°C) and the supernatants were stored at -20°C for subsequent analyses by micro-high-performance liquid chromatography (microHPLC).For transcript profiling, additional 5 mL samples was retrieved at each of the seven time points detailed in above paragraph on "Growth experiments for mutant characterization." Sample treatment and storage was likewise performed.

Growth experiments for assessing sensory selectivity
To determine the transcriptional responsiveness of the ppsA1, cmh, acsA1, bssA, and ebdA1 genes to phenol, p-cresol, and p-ethylphenol, essentially, the same experimental setup was used as described in the two paragraphs above (Fig. 4).Here, however, for each of the separately administered phenolic compounds, only three different concen trations were applied (100 µM, 1 µM, and 100 nM).

Quantification of benzoate by (micro)HPLC
Quantitative depletion profiling of benzoate was achieved using a microHPLC system (UltiMate 3000; Thermo Fisher Scientific, Germering, Bavaria, Germany) as previously described (23).The system was equipped with a reverse-phase C 18 column (Thermo Hypersil Gold; Thermo Fisher Scientific) and a diode array detector (DAD-3000; Thermo Fisher Scientific) and operated at 40°C with a flow rate of 0.

Preparation of total RNA
Preparation of total RNA was performed according to the protocol of Oelmüller et al. (36) and as previously described ( 13), using cells treated with RNAprotect Bacteria Reagent and stored at -80°C.A given cell pellet was resuspended in STE buffer (10 mM Tris-HCl, 1 mM EDTA, 100 mM NaCl, pH 8.3).Then, 20 µL SDS (10%, wt/vol) and 900 µL Roti Aqua-Phenol (CarlRoth, Karlsruhe, Germany) were added; the suspension was incubated at 60°C for 8 min by gently inverting and then centrifuged (20,000 × g, 5 min, room temperature).The resultant aqueous phase was transferred into a 2-mL 5PRIME phase lock gel tube (Quantabio, Beverly, MA, USA).One volume of phenol-chloroformisoamylalcohol (25:24:1) was added, and the tube was gently inverted for 5 min.After centrifugation (20,000 × g, 5 min, room temperature), nucleic acids were precipitated using ice-cold ethanol (96%) during incubation at −80°C for 30 min.After centrifugation (20,000 × g, 30 min, 4°C), the pellet was washed with ice-cold ethanol (75%) and centrifuged again (20,000 × g, 15 min, 4°C).The pellet was dried and suspended in RNase-free water.Subsequently, the sample was digested with DNase I (RNase-free; Qiagen).Complete removal of DNA was confirmed by PCR using genomic DNA of A. aromaticum EbN1 T as a positive control.RNA quality underlying the transcript profiles was controlled using an Experion automated electrophoresis station (Bio-Rad, Hercules, CA, USA), confirming the integrity of rRNAs as well as the ratio between 23S and 16S rRNA.RNA concentrations were determined using a microplate reader (SPECTROstar Nano; BMG Labtech, Ortenberg, Germany).RNA samples were stored at −80°C until further analyses.All chemicals used for RNA preparation were of molecular biology grade.

Transcript profiling by qRT-PCR
Relative expression levels of different target genes in the three genotypes were determined by real-time RT-PCR.For the selected target genes, specific primers (Table 3) were designed using the software Primer3 (version 4.1.0;https://primer3.ut.ee/).cDNA generation and real-time PCR were performed with three technical replicates per RNA preparation applying 150 ng of total RNA and using the Brilliant III ultra-fast SYBR green quantitative reverse transcription-PCR (qRT-PCR) master mix (Agilent, Santa Clara, CA, USA) and the CFX96 real-time system (Bio-Rad).In total, nine measurements were conducted per analyzed time point.The one-tube RT real-time PCR was carried out with one cycle of reverse transcription for 10 min at 50°C and one cycle of PCR initiation for 3 min at 95°C, followed by 40 cycles of 10 s of denaturation at 95°C, 30 s of annealing (primer specific), and 30 s of extension at 60°C, succeeded by real-time detection for 5 s.The gene-specific annealing temperatures were the following: 60°C for acsA1, cmh, ebdA1, and ppsA1; 64°C for bssA.The specificity of accumulated products was verified by melting curve analysis, ranging from 60°C to 90°C in steps of 0.5°C.Differences in relative transcript abundance between reference and test states were calculated as previously described (13) according to the following equation: Ratios = E ΔCT(reference -test) .Primer-specific efficiencies (E) (Table 3) were determined as described previously (13,37).

Prediction of 3D structure models
The crystal structures of the sensory domains of MopR (7VQF and 5KBE), CapR (6IY8), and PoxR (5FRW) were obtained from the RCSB protein database (URL: RCSB.org) (38).The structural models were predicted using the neural network-based model AlphaFold (30) and its related database (39) by applying the prediction based on a homodimer generation.Graphical representation of overall folds and zoom-ins of the binding pockets were generated using Pymol (the PyMOL Molecular Graphics System, version 2.2.0;Schrödinger, New York, NY, USA).Multiple sequence alignment of the sensory domains from selected proteins was generated by ESPript (https://espript.ibcp.fr)(40).

ACKNOWLEDGMENTS
This study was supported by the Deutsche Forschungsgemeinschaft (DFG) within the framework of the research training group Molecular Basis of Sensory Biology (GRK 1885).

FIG 2
FIG 2 Generation of in-frame ΔpheR deletion and pheR-complemented (genotype: ΔpheR, pBBR1 MCS-2 ΩpheR) mutants of A. aromaticum EbN1 T .(A) Scale model of the pheR gene and its 3′-and 5′-flanking regions on the chromosome, displaying the wild type (top) and the ΔpheR mutant (bottom).Chromosomal hybridization locations of the primer pairs ΔpheR_F/R and pheR_468_F/752_R to identify the deletion and complementation genotype are indicated by black (Continued on next page)

FIG 2 (
FIG 2 (Continued) arrows.Expected lengths of the respective PCR products are shown below the scale model.(B) Image of electrophoretically separated PCR products obtained

FIG 3
FIG 3 Transcriptional responsiveness of A. aromaticum EbN1 T .(A) Time-resolved, quantitative transcript profiles of the ppsA1 gene in response to different extracellular concentrations of phenol.The determination of relative transcript abundances was described in the legend to Fig. 2D, and fold changes are tabulated in TableS1.Each fold change relies on three biological and three technical replicates each.

FIG 4
FIG 4 Negligible gratuitous induction by phenolic compounds in A. aromaticum EbN1 T .Comparative, time-resolved, quantitative transcript profiles of the ppsA1, cmh, and acsA1 genes in response to different concentrations of phenol, p-cresol, and p-ethylphenol.Gene names are as described in legend to Fig. 3. Determination of relative transcript abundances was as outlined in legend to Fig. 2D and fold changes are tabulated in TableS1.Each fold change relies on three biological and three technical replicates each.Note the differences in scale bars.

FIG 5
FIG 5 Similarities of the sensory domains from different sensory/regulatory proteins for phenolic compounds.(A) Sequence alignment superposed with the secondary structure elements derived from MopR (Acinetobacter calcoaceticus; PDB 7VQF) and generated by the Espript server.Symbols: red squares, perfect conservation; green and black ovals, van der Waals or hydrogen bond interaction with the phenol ligand; pink ovals, zinc ion-binding motif.Aligned proteins: MopR (PDB 7VQF) from A. calcoaceticus; CapR (PDB 6IY8) from Pseudomonas putida; PoxR (PDB 5FRW) from Cupriavidus necator; PheR (Q5P474), PcrS (Q5P0J1), and EtpR (Q5P8R7) from A. aromaticum EbN1 T .(B) Overall fold of the MopR dimer in cartoon representation: one dimer is highlighted by a color gradient from blue (N-terminus) to red (C-terminus) and the other dimer is shown in gray.The nomenclature used for the secondary structure in panels A and B is from Ray et al. (21).Bound phenol is highlighted by black (carbon) and red (oxygen) spheres and the zinc atoms by pink spheres.(C) Overlay of the sensory domains of the six sensory/regulators proteins.

FIG 6
FIG 6 Shape and spatial expansion of binding pockets of phenolic compounds.(A) Zoom-ins into the binding pockets for phenolic compounds in the sensory domains of MopR [based on crystal structure from A. calcoaceticus; reference (19)] compared to the ones of PheR, PcrS, and EtpR from A. aromaticum EbN1 T based on AlphaFold predicted models.The modeled ligand-accommodating cavity is indicated by a transparent gray surface, while the residues enclosing the cavity are highlighted in balls and sticks.The conserved His and Trp residues anchoring the hydroxy group of the ligand are labeled in red, and the dashes show the hydrogen bonds in the experimental models.Phenol, modeled based on experimental data (MopR), has carbon colored in black and oxygen in red.(B) Comparison of the cavity volume between the experimental model MopR (gray surface) and the AlphaFold generated models of PheR, PcrS, and EtpR shown as transparent colored surfaces.The modeled phenol compounds are in transparent balls and sticks with carbon colored in black/pink and oxygen in red.

TABLE 1
Physical considerations for cellular effector uptake a

9 m 2 s −1 ) P ow K M D eff = D K M (10 −9 m 2 s −1 ) c thresh (nM) N Sub /cell est. K D (N Sen = 10) (µM)
(13)usion coefficient in water(26); P ow , octanol-water partition coefficient(27); K M , membrane partition coefficient, with K M = 0.23 P ow , see Eq. 8b in reference(13); D eff , effective diffusion coefficient in lipid membranes; c thresh , determined in vivo threshold concentration; N Sub , number of copies of substrate molecule per cell in equilibrium; K D , estimated upper bound for equilibrium dissociation constant, given observed threshold responses for transcriptional response, for N Sen = 10 copies of sensor protein per cell.

TABLE 2
Oligonucleotide primers used for mutant construction b IR, intergenic region.

TABLE 3
Oligonucleotide primers used for targeted transcript profiling .W. was supported by the Max Planck Society and would like to thank the Max Planck Institute Marine Microbiology for support.R.R. conceived this study; R.B. conducted the molecular genetic experiments with expert advice from L.W.; R.B., P.L., S.S., and S.H. performed the cultivation experiments and RNA work; F.S. and J.N. performed the nucleotide sequencing; M.W. did calculations; T.W. performed the 3D-structural analyses; R.B. and R.R. wrote the manuscript with input from all authors. T