The Catabolic Network of Aromatoleum aromaticum EbN1^T Open Access

Aromatic compounds represent, next to carbohydrates, the second most abundant organic molecules on the planet. In the biosphere they are encountered as products of anaerobic biodegradation [Fischer-Romero et al., 1996], as building blocks of major biomacromolecules [de Leeuw et al., 2006], or as phytohormones [Pichersky and Raguso, 2018]. In the geosphere, aromatic compounds are abundant constituents of kerogen [Vandenbroucke and Largeau, 2007] and crude oil [Marshall and Rodgers, 2008]. Anthropogenic activities including industrial synthesis and fossil energy production provide further sources of aromatic compounds in the environment [ATSDR, 2008]. While aromatic compounds are energy-rich substrates for supporting microbial growth, they are at the same time biochemically challenging due to their chemical stability [Wilkes and Schwarzbauer, 2010] and physiologically adverse on account of their toxicology [Sikkema et al., 1995]. To activate and cleave the aromatic ring, O₂-dependent oxygenase-based reactions and a large variety of intriguing O₂-independent reactions are employed by aerobic and anaerobic microorganisms, respectively [Rabus et al., 2016; Boll et al., 2020; Cheng et al., 2022].

Since transitions between oxia and anoxia occur in many different environments, facultative degradation specialists for aromatic compounds represent keystone members of the microbial communities in these habitats, e.g., the facultative denitrifying betaproteobacterial genera Aromatoleum [Rabus et al., 2019] and Thauera [Anders et al., 1995]. The genus Aromatoleum comprises currently 10 validly described species, with Aromatoleum aromaticum EbN1^T (DSM 19018^T) as the type species and versatile representative [Rabus and Widdel, 1995]. This strain is the first anaerobic hydrocarbon degrader with a completely sequenced genome [Rabus et al., 2005] and is particularly well studied on the physiological, genetic, and OMICS levels [e.g., Büsing et al., 2015a, b; Rabus et al., 2014; Trautwein et al., 2012a, 2012b; Vagts et al., 2020; Wöhlbrand et al., 2007, 2008]. In the case of the catabolic signature property of A. aromaticum EbN1^T, namely the anaerobic degradation of ethylbenzene, the crystal structures of the pathway’s first three enzymes have been determined: ethylbenzene dehydrogenase [Kloer et al., 2006], (S)-1-phenylethanol dehydrogenase [Höffken et al., 2006], and acetophenone carboxylase [Weidenweber et al., 2017]. Most recently, A. aromaticum EbN1^T was established as a model for systems biology-level investigations, including a manual re-annotation of the complete genome, multi-OMICS integration, and a genome-wide metabolic model [Becker et al., 2022a].

Here, we summarize the state of knowledge on the catabolism of A. aromaticum EbN1^T. For this purpose, we integrated an up-to-date manual annotation of its genome with all currently available data from differential proteomics and genetic/biochemical analyses conducted over the past more than 2 decades with A. aromaticum EbN1^T. The functional predictions were supported by pathway-specific literature on related, biochemically well-studied Thauera aromatica K172^T, Aromatoleum evansii KB740^T (formerly Azoarcus evansii KB740^T [Rabus et al., 2019]), Rhodopseudomonas palustris, and other strains, as well as by literature on general catabolic aspects using standard bacteria such as Escherichia coli, Bacillus subtilis, and Pseudomonas species.

Circular representations of the chromosome and plasmid 2 of A. aromaticum EbN1^T (Fig. 1A) illustrate the localization of genes attributed to the substrate-specific degradation pathways for aromatic compounds, aliphatic compounds, and amino acids, to the central pathways for terminal oxidation to CO₂, as well as to the modules for respiratory energy conservation employed under anoxic and oxic conditions. In most cases, the genes constituting the various degradation and respiration modules are arranged in operon-like structures, which for the most part are scattered across the chromosome and the plasmid (2^nd plasmid not shown in Fig. 1A). This reflects the previously reported high plasticity of the genome of A. aromaticum EbN1^T [Rabus et al., 2005].

In total, the degradation and respiration modules are built on 279 and 91 genes, respectively, occupying 7.4% and 2.0% of the genome’s coding space, with 178 and 65 encoded proteins identified by differential proteomics (Fig. 1B). Overall, A. aromaticum EbN1^T realizes its catabolic network with 370 predicted proteins (∼67% identified, ∼9.5% of coding region of its 4.7 Mbp genome). This is in a similar range as recently determined for closely related A. aromaticum pCyN1 (215 predicted proteins, 89% proteomic coverage, ∼6% of coding region of the 4.4 Mbp genome) [Becker et al., 2022b] and deltaproteobacterial, marine, versatile, sulfate-reducing Desulfonema magnum (145 predicted proteins, 90% identified, ∼2.16% of coding region of the 8 Mbp genome) [Schnaars et al., 2021].

An integrative scheme of the degradation network of A. aromaticum EbN1^T is presented in Figure 2, illustrating the complex structure of the network, interlinkages, and central modules. This scheme provides a roadmap to the diverse reaction sequences employed for the degradation of 23/21 aromatic/aliphatic compounds and 3 amino acids to CO₂ (via the TCA cycle; in few cases, requiring preparatory passage through the methylmalonyl-CoA pathway). Summed up, about a dozen anaerobic peripheral degradation routes channel a total of 15 different aromatic compounds into the anaerobic benzoyl-CoA pathway, underpinning its central role in the degradation network. Aerobic degradation of 4 aromatic compounds proceeds via the “box” pathway. Peripheral degradation of aliphatic compounds enters the TCA cycle at the levels of citrate via acetyl-CoA (10 compounds), of 2-oxoglutarate (3 compounds) or of succinate (6 compounds).

The network architecture of its degradation and respiration parts is based on the recent manual re-annotation of the complete genome of A. aromaticum EbN1^T [Becker et al., 2022a]. To facilitate comparison of the catabolic network with previous proteogenomic publications on this strain, the tabulated protein constituents of each individual catabolic module (Fig. 3−42) are presented with new [Becker et al., 2022a] and original locus tags [Rabus et al., 2005]; in addition, references for experimental evidence (proteomic, molecular genetic, and/or biochemical) are indicated.

Toluene represents the best-studied model compound in the context of the anaerobic degradation of aromatic hydrocarbons (alkylbenzenes). The involved reactions and enzymes have been mainly elucidated with T. aro-matica K172^T and Azoarcus sp. strain T [for overview, see Heider et al. [2016a], and the coding genes have been identified in T. aromatica K172^T and A. aromaticum EbN1^T [Leuthner et al., 1998; Leuthner and Heider, 2000; Hermuth et al., 2002; Kube et al., 2004]. The following metabolites, reactions, enzymes, and genes are involved in the degradation pathway (Fig. 3A, B), which proceeds via six steps. (i) In the initial reaction step, the glycyl-radical enzyme benzylsuccinate synthase (BssABC) adds toluene to the double bond of the fumarate cosubstrate, forming (R)-benzylsuccinate, with the stereospecificity elucidated among others by modeling and using [d₃-methyl]toluene combined with NMR-analysis [Leuthner et al., 1998; Krieger et al., 2001; Qiao and Marsh, 2005; Seyhan et al., 2016; Szaleniec and Heider, 2016]. Benzylsuccinate synthase belongs to the growing family of glycyl-radical-containing enzymes. This family includes related enzymes activating other aromatic and aliphatic hydrocarbons as well as well-studied pyruvate formate-lyase and anaerobic ribonucleotide reductase, with the latter two, however, catalyzing very different reactions [Frey, 2001; Strijkstra et al., 2014; Heider et al., 2016a]. EPR and Mössbauer spectroscopic, crystal structure, and biochemical analyses revealed that the heterotrimeric benzylsuccinate synthase harbors the glycyl (storage)/thiyl (catalysis)-radical sites in the catalytic α-subunit (BssA) and FeS clusters in the accessory β- and γ-subunits (BssBC) [Duboc-Toia et al., 2003; Li et al., 2009; Hillberg et al., 2012; Funk et al., 2014]. The glycyl storage radical is assumed to be generated by activase BssD according to the known route involving S-adenosylmethionine (SAM). During catalysis toluene is suggested to be sandwiched between the thiyl-radical and the fumarate cosubstrate resulting in syn-addition [Li and Marsh, 2006; Funk et al., 2015; Szaleniec and Heider, 2016], i.e., cleavage of the benzylic C−H-bond concurrent with C−C-bond formation between the benzyl radical and fumarate. This is analogous to the previously proposed concerted mechanism of anaerobic conversion of n-hexane to (R)-(1-methylpentyl)succinate in Aromatoleum sp. strain HxN1. Concurrent to the homolytic abstraction of the pro-S hydrogen from carbon atom 2 of n-hexane, the new bond between the latter and one of the two carbon atoms of the double bond of fumarate is already formed on the other side of the n-hexane molecule, thus avoiding a free alk-2-yl radical intermediate [Jarling et al., 2012; Wilkes et al., 2016]. (ii) (R)-Benzylsuccinate is then prepared for subsequent β-oxidation by activation to (R)-benzylsuccinyl-CoA catalyzed by a specific succinyl-CoA:(R)-benzylsuccinate CoA transferase (BbsEF) having an α₂β₂ conformation [Leutwein and Heider, 2001]. (iii) (R)-Benzylsuccinyl-CoA is then oxidized by substrate-specific, homotetrameric, and FAD-containing (R)-benzylsuccinyl-CoA dehydrogenase (BbsG) forming (E)-phenylitaconyl-CoA (synonym: 2-(E)-benzylidenesuccinyl-CoA) [Leutwein and Heider, 1999, 2002]. Electrons from this oxidation reaction are channeled via electron transfer flavoproteins (ETF) in conjunction with ETF:quinone oxidoreductase directly into the respiratory quinone pool [Vogt et al., 2019]. (iv) Then, homotrimeric enoyl-CoA hydratase (BbsH) adds water to the double bond of (E)-phenylitaconyl-CoA forming 2-(α-hydroxyphenyl) methylsuccinyl-CoA (synonym: 2-(α-hydroxybenzyl)succinyl-CoA) [Leuthner and Heider, 2000; von Horsten et al., 2022]. (v) The latter is then oxidized by the NAD⁺-dependent, heterotetrameric (S,R)-2-(α-hydroxybenzyl)succinyl-CoA dehydrogenase (BbsCD) yielding (S)-2-benzoylsuccinyl-CoA, according to enzymatic and crystal structure analyses [Leuthner and Heider, 2000; von Horsten et al., 2022]. (vi) Finally, 3-oxoacyl-CoA thiolase (BbsAB), catalyzes the cleavage of (S)-2-benzoylsuccinyl-CoA into the central intermediate benzoyl-CoA and succinyl-CoA, which can be reused for CoA transfer to benzylsuccinate and formation of the fumarate cosubstrate [Leuthner and Heider, 2000]. The activity of this novel thiolase requires both subunits, the thio­-lase family member BbsB and the Zn-finger protein BbsA, jointly positioning the CoA moiety of the substrate into the catalytically productive conformation [Weidenweber et al., 2022].

Due to its highly lipophilic property, toluene does not require a specific uptake system, but rather enters the cell by passive diffusion. Notably, a novel type of efflux system appears to be encoded in proximity to the bbs gene cluster (EbN1_C01076-80), which was specifically detected in toluene-metabolizing cells of A. aromaticum EbN1^T and probably contributes to solvent tolerance [Wöhlbrand et al., 2008].

The genes encoding the protein constituents of the anaerobic degradation of toluene are organized in the two neighboring, albeit opposingly oriented operons bss (benzylsuccinate synthase) and bbs (β-oxidation of benzylsuccinate), which are preceded by the tdiSR genes coding for a presumptively toluene-specific two-component sensory/regulatory system (Fig. 3B) [Leuthner and Heider, 1998]. Differential global profiling of substrate-adapted cells of A. aromaticum EbN1^T revealed specific formation of transcripts/proteins of the bss and bbs operons [Champion et al., 1999; Kühner et al., 2005]. In conjunction with the presence of conserved sequences in the promoter regions of both operons, the latter are proposed to be regulated by TdiSR in a concerted manner [Rabus et al., 2014].

The anaerobic growth with ethylbenzene is the physio­logical signature property of A. aromaticum EbN1^T. (S)-1-Phenylethanol and acetophenone represent intermediates of the degradation pathway and additionally serve as growth-supporting substrates [Rabus and Widdel, 1995; Kühner et al., 2005]. The key enzyme of this degradation pathway, ethylbenzene dehydrogenase, is archetypic for oxygen-independent alkyl chain hydroxylation of various hydrocarbons [Heider et al., 2016b]. The following metabolites, reactions, enzymes, and genes are involved in the degradation pathway (Fig. 4A, B), which proceeds via five steps. (i) Ethylbenzene is stereospecifically hydroxy­lated at its methylene carbon forming (S)-1-phenylethanol by the novel molybdenum/iron-sulfur/heme-containing ethylbenzene dehydrogenase (EbdA1B1C1), which is localized in the periplasm [Rabus and Heider, 1998; Johnson et al., 2001; Kniemeyer and Heider, 2001a]. The crystal structure and characterization of redox centers revealed that the α-subunit (EbdA) harbors the Mo-bisMGD (molybdopterin-guanine-dinucleotide) cofactor and one [4Fe-4S] cluster (FS0), that the β-subunit (EbdB) contains three [4Fe-4S] clusters (FS1−3) and one [3Fe-3S] cluster (FS4), and that the γ-subunit (EbdC) carries the heme b cofactor [Kloer et al., 2006; Hagel et al., 2022]. Hydroxylation of ethylbenzene proceeds via two successive one-electron steps reducing the Mo-bisMGD cofactor (from Mo^VI to Mo^IV), with the electrons sequentially transferred via the FeS clusters to the heme b cofactor. Modeling approaches of the reaction indicated the involvement of two transition states, i.e., radical-type followed by carbocation-type, and that stereospecificity results from markedly faster activation of the pro(S) hydrogen compared to its pro(R) counterpart [Szaleniec et al., 2007, 2010, 2014]. Electrocatalytic activity studies with ethylbenzene dehydrogenase using hydroxymethylferrocenium as artificial electron acceptor combined with digital modeling provided new insights into the kinetics of the catalytic mechanism [Kalimuthu et al., 2015]. With a more applied perspective, comprehensive substrate and inhibitor spectra as well as production of industrial chiral alcohols by ethylbenzene dehydrogenase were investigated [Knack et al., 2012, 2013; Tataruch et al., 2014]. (ii) (S)-1-Phenylethanol is then oxidized by stereoselective (S)-1-phenylethanol dehydrogenase (Ped1) to acetophenone [Kniemeyer and Heider, 2001b]; (R)-1-phenylethanol, which also supports growth of A. aromaticum EbN1^T, is oxidized by a different dehydrogenase. The crystal structure of Ped1 revealed the structural features determining enantioselectivity and steady-state kinetic analysis showed high cooperativity of the enzyme with the (S)-enantiomer [Höffken et al., 2006]. From an applied perspective, heterologously produced Ped1 was shown to reduce a wide range of prochiral ketones and 3-oxo esters to the respective enantiopure secondary alcohols [Dudzik et al., 2015]. Notably, the genome of A. aromaticum EbN1^T harbors a 2^ndebd-ped gene cluster, the function of which remains, however, elusive since the initial report [Rabus et al., 2005]; an unmarked in-frame deletion of ebdC2, which encodes the γ-subunit of this paralogous ethylbenzene dehydrogenase, did not show a phenotype [Wöhlbrand and Rabus, 2009]. (iii) Subsequently, acetophenone is carboxylated by a novel type of carboxylase, acetophonene carboxylase (encoded by the apc1−5 genes), to benzoylacetate, requiring 2 mol ATP per mol acetophonene [Jobst et al., 2010]. During purification, the enzyme dissociates into a heterooctameric Apc(αα′βγ)₂ core complex and Apcε. The crystal structure of Apc(αα′βγ)₂ allowed proposal of the following four-step reaction mechanism [Weidenweber et al., 2017]: First, Apcα binds ATP/acetophenone and Apcα′ binds ATP/bicarbonate leading to a tightened Apcαα′βγ core complex. Second, the kinase activities of Apcαα′ are triggered by Apcε attaching to the core complex, which lead to the formation of phosphoenol acetophenone and carboxyphosphate, respectively. Third, formation of electrophilic CO₂ and its approximation to phosphoenol acetophenone by further tightening of the Apc complex facilitates the carboxylation reaction followed by dephosphorylation to benzoylacetate. Fourth, Apcε detaches, the pro­ducts are released, and the relaxed Apc core complex is restored. (iv) Based on proteogenomic and transcript analyses, benzoylacetate was suggested to be activated to its CoA ester by a predicted benzoylacetate-CoA ligase (Bal) [Rabus et al., 2002; Kühner et al., 2005], which was corroborated by enzyme activity measurements [Muhr et al., 2015]. (v) Finally, benzoylacetyl-CoA is suggested to be thiolytically cleaved by a 3-oxo acyl-CoA thiolase yielding the central intermediate benzoyl-CoA and acetyl-CoA [Rabus et al., 2002]; the identity of the thiolase is unclear at present.

Ethylbenzene does not require a specific uptake system, since it is oxidized in the periplasm. Due to their highly lipophilic properties, neither (S)-1-phenylethanol nor acetophenone requires a specific uptake system, but rather enter the cell by passive diffusion.

The genes encoding the protein constituents of the anaerobic degradation of ethylbenzene are organized in two neighboring operons in A. aromaticum EbN1^T (Fig. 4B): ebd-ped (upper pathway section from ethylbenzene to acetophenone) and apc-bal (lower pathway section from acetophenone to benzoylacetyl-CoA). These two operons flank genes encoding 2 two-component sensory/regulatory systems proposed to be specific for ethylbenzene (ediSR, formerly tcs1/tcr1) and acetophenone (adiSR, formerly tcs2/tcr2), respectively [Rabus et al., 2002]. Differential profiling of transcript and protein patterns [Champion et al., 1999; Kühner et al., 2005] and of enzymatic activities [Muhr et al., 2015] indicated substrate-specific, sequential regulation of the two operons.

The anaerobic degradation of phenol was mainly elucidated with T. aromatica K172^T and found to proceed in the same way in A. aromaticum EbN1^T, based on homology and organization of genes [Breinig et al., 2000; Rabus et al., 2005] and on phenol-specific formation of involved enzyme subunits [Wöhlbrand et al., 2007]. The following metabolites, reactions, and enzymes are involved in the degradation pathway (Fig. 5A), which proceeds via three steps. (i) The initial reaction step is catalyzed by phenylphosphate synthase (PpsA1BC) that transfers the β-phosphate group of MgATP to phenol furnishing phenylphosphate. Similar to phosphoenolpyruvate synthase, a ping-pong mechanism of catalysis is proposed involving binding of diphosphate (PP_i) to the catalytic His-residue in PpsA followed by release of an orthophosphate (P_i) (rendering the reaction unidirectional) and transfer of the retained P_i from the His-residue to phenol [Schmeling et al., 2004; Narmandakh et al., 2006]. (ii) Then, phenylphosphate is converted to 4-hydroxybenzoate by phenylphosphate carboxylase (PpcBDAC) via a reaction resembling a Kolbe-Schmitt carboxylation. The enzyme requires K⁺ and Mg²⁺, is highly oxygen sensitive, and requires tight subunit interactions for achieving phenylphosphate dephosphorylation (by the δ-subunit) with concomitant carboxylation (by αβγ-subunits, constituting the core carboxylase) of the intermediary enzyme-bound phenolate [Lack and Fuchs, 1992; Schühle and Fuchs, 2004; Schmeling and Fuchs, 2009]. Anaerobic conversion of catechol via catechylphosphate to protocatechuate (3,4-dihydroxybenzoate) can also be catalyzed by Pps and Ppc, indicating promiscuity of both enzymes [Ding et al., 2008]. (iii) Further degradation of 4-hydroxybenzoate to the central intermediate benzoyl-CoA is as described in section “4-Hydroxybenzoate” below.

Due to its highly lipophilic property, phenol does not require a specific uptake system, but rather enters the cell by passive diffusion.

The genes coding for Pps and Ppc are organized in an operon-like structure (Fig. 5B), directly preceded by an opposingly oriented gene encoding the predicted, σ⁵⁴-dependent transcriptional regulator PheR demonstrated to control expression of the pps and ppc genes [Breinig et al., 2000; Rabus et al., 2005; Buschen et al. 2023].

First evidence for O₂-independent oxidation of 4-methylphenol at its methyl group arose from studies with Pseudomonas putida [Hopper, 1976; Hopper and Taylor, 1977], which was subsequently confirmed for several species of the genera Thauera and Aromatoleum (most species affiliating with the latter were formerly classified as Azoarcus) [Rudolphi et al., 1991]. The following metabolites, reactions, and enzymes are involved in the degradation pathway (Fig. 6A), which proceeds via three steps. (i) The initial reaction has been intensively studied in Ps. putida. It is an anaerobic hydroxylation of the methyl-group of 4-methylphenol forming 4-hydroxybenzyl alcohol followed by oxidation to 4-hydroxybenzaldehyde. These consecutive reactions are both catalyzed by the flavocytochrome enzyme termed p-cresol methylhydroxylase (PCMH), which is localized in the periplasm [Hopper et al., 1985]. Several structural studies on heterotetrameric α₂β₂ PCHM of Ps. putida indicated that a structural change tunes the spatial flavoprotein-cytochrome interactions for optimized intermolecular electron transfer and supported the following proposed reaction mechanism [Shamala et al., 1986; Mathews et al., 1991; McLendon et al., 1991; Cunane et al., 2000, 2005]: 4-Methylphenol is first oxidized to a quinone methide intermediate, with the released two electrons received by the flavin yielding a fully reduced FAD, wherefrom the electrons are sequentially passed on to the heme cofactor. Water is then assumed to attack (nucleophilically) the methide carbon of the intermediate, resulting in formation of 4-hydroxybenzyl alcohol, which is then oxidized by PCMH to 4-hydroxybenzaldehyde. In A. aromaticum EbN1^T the conversion of 4-methylphenol to 4-hydroxybenzaldehyde is apparently catalyzed by a different albeit also FAD-dependent p-cresol methylhydroxylase (Cmh), which is apparently localized in the cytoplasm (no predicted signal peptide). (ii) 4-Hydroxybenzaldehyde is then oxidized to 4-hydroxybenzoate by 4-hydroxybenzaldehyde dehydrogenase (PchA) [Cronin et al., 1999]; this enzyme is termed Hbd in A. aromaticum EbN1^T. (iii) Further degradation of 4-hydroxybenzoate to the central intermediate benzoyl-CoA is as described in section “4-Hydroxybenzoate” below.

Due to its highly lipophilic property, 4-methylphenol does not require a specific uptake system, but rather enters the cell by passive diffusion [Vagts et al., 2020].

The structural genes pchC and pchF encoding the cytochrome (α) and flavoprotein (β) subunits of PCMH, respectively, and pchA encoding 4-hydroxybenzaldehyde dehydrogenase were first determined in Ps. putida [Kim et al., 1994; Cronin et al., 1999]; the function of the pchX gene additionally present in the pch gene cluster is still unknown. Homologs of the pchCF genes were also found in the genome of A. aromaticum EbN1^T, but assigned to the anaerobic degradation of 4-ethylphenol and renamed as emhCF genes (see below). Transcriptional regulation of the pch operon in Pseudomonas spp. was suggested to be mediated by a one-component regulatory protein (DmpR/PchR) [Pavel et al., 1994; Jõesaar et al., 2010], belonging to the σ⁵⁴-dependent regulators of the XylR/DmpR-family [Tropel and van der Meer, 2004]. By contrast, in A. aromaticum EbN1^T transcription of the cmh-hbd genes (Fig. 6B) is controlled by the σ⁵⁴-dependent two-component sensory/regulatory system PcrSR, as evidenced by unmarked, in-frame deletion mutagenesis [Vagts et al., 2020]. The in vivo responsiveness of A. aromaticum EbN1^T toward p-cresol falls in the nanomolar range (<100 nM), as determined by targeted, quantitative transcript analyses [Vagts et al., 2020].

The anaerobic growth of A. aromaticum EbN1^T with 4-ethylphenol represents a physiological trait, which was not predicted during initial genome annotation and only subsequently discovered by differential proteomics. Likewise, 1-(4-hydroxyphenyl)ethanol and 4-hydroxyacetophenone were recognized as degradation intermediates as well as growth-supporting substrates [Wöhlbrand et al., 2008]. The following metabolites, reactions, enzymes, and genes are involved in the degradation pathway (Fig. 7A, B), which proceeds via five steps as previously proposed based on differential proteomic and metabolite analyses [Wöhlbrand et al., 2008]. (i) In the initial reaction, 4-ethylphenol is hydroxylated at the methylene carbon forming (R)-1-(4-hydroxyphenyl)ethanol catalyzed by 4-ethylphenol methylenehydroxylase (EmhCF, revised from original PchCF annotation) of A. aromaticum EbN1^T. EmhCF is similar to PCMH of Ps. putida NCIMB 9866, which was previously demonstrated to also convert 4-ethylphenol [McIntire et al., 1985]. Thus, intermediary formation of a quinone methide appears plausible, too, as inferred from the redox potential of the flavin cofactor of ethylphenol methylhydroxylase from Ps. putida JD1 [Reeve et al., 1989]. EmhCF is probably localized in the cytoplasm, since the large, flavin cofactor-containing F-subunit is apparently devoid of a signal peptide and only for the small C subunit a Sec-type signal peptide can be predicted. (ii) (R)-1-(4-Hydroxyphenyl)ethanol is then further oxidized to 4-hydroxyacetophenone by (R)-specific 1-(4-hydroxyphenyl)ethanol dehydrogenase (Hped1, formerly ChnA), as demonstrated by unmarked, in-frame mutagenesis and analysis of the crystal structure [Büsing et al., 2015a]. Molecular genetic analysis further demonstrated that the dehydrogenase encoded by the hped2 gene (directly neighboring the hped1 gene) does not serve this function [Büsing et al., 2015a]. Prompted by the biotechnological potential of Hped1, the pH-dependent stability of this enzyme was recently investigated under storage and in-process conditions [Tataruch et al., 2023]. (iii) p-Hydroxyacetophenone is then proposed to be converted to 4-hydroxybenzoylacetate by the carboxylase HacABC (formerly XccABC) [Wöhlbrand et al., 2008], which should first carboxylate its biotin cofactor followed by transcarboxylation of the actual substrate as known from other well-studied biotin-dependent carboxylases [Attwood and Wallace, 2002]. (iv) Further degradation is proposed to proceed via activation of 4-hydroxybenzoylacetate to its CoA ester catalyzed by the acetoacetyl-CoA ligase-like protein AcsA1, followed by thiolytic removal of an acetyl-CoA moiety by the predicted thiolase TioL1 yielding 4-hydroxybenzoyl-CoA. (v) Further degradation of 4-hydroxybenzoyl-CoA to the central intermediate benzoyl-CoA is as described in section “4-Hydroxybenzoate” below. Notably, the proposed degradation pathway from 4-ethylphenol to 4-hydroxy-benzoyl-CoA is analogous to the one described above for ethylbenzene, albeit involving a completely different set of enzymes.

Due to its highly lipophilic property, 4-ethylphenol does not require a specific uptake system, but rather enters the cell by passive diffusion [Vagts et al., 2020]. This can also be envisioned for 1-(4-hydroxyphenyl)ethanol and 4-hydroxyacetophenone. Genes encoding a novel type of efflux system (EbN1_C001730-60) are located upstream of the acsA1 gene. This efflux system is homologous to the one for toluene, but specifically formed in 4-ethylphenol or 4-ethylacetophenone-metabolizing cells of A. aromaticum EbN1^T, and probably contributes to the strain’s solvent tolerance [Wöhlbrand et al., 2008].

The genes encoding all above described enzymes for the anaerobic degradation of 4-ethylphenol are localized in a single large operon-like structure (Fig. 7B) [Wöhlbrand et al., 2008]. Differential abundance and activity profiles demonstrated the substrate-specific formation of these enzymes [Wöhlbrand et al., 2008; Muhr et al., 2015]. The direct genomic neighborhood of the etpR gene (formerly ebA324) in conjunction with the presence of the 12/24 consensus sequence of σ⁵⁴-dependent promotors [Helmann and Chamberlin, 1988] upstream of acsA1 as the 1^st gene of the operon, suggests EtpR as its transcriptional regulator [Wöhlbrand et al., 2008]. In accord, unmarked, in-frame deletion of the etpR gene resulted in loss of the capacity for anaerobic growth with 4-ethylphenol or 4-hydroxyacetophenone, additionally underpinning its capacity to recognize both aromatic compounds [Büsing et al., 2015b]. The in vivo responsiveness of A. aromaticum EbN1^T toward them falls in the lower nanomolar range, as determined by targeted, quantitative transcript analyses [Vagts et al., 2018, 2020].

The aromatic primary amines 2-phenylethylamine and benzylamine were recently recognized to support robust anaerobic growth of A. aromaticum EbN1^T [Schmitt et al., 2019]. While amine dehydrogenases are long known for the degradation of aliphatic and aromatic amines in Pseudomonas spp. [e.g., Durham and Perry, 1978; Iwaki et al., 1983] the findings with A. aromaticum EbN1^T represent the first example of an anaerobic metabolism. Here, the oxidative deamination of 2-phenylethylamine and benzylamine (both Fig. 8A) is catalyzed by two specific periplasmic quinohemoprotein amine dehydrogenases [Schmitt et al., 2019] characterized by an encaged quinoid cofactor as known from Paracoccus denitrificans [Datta et al., 2001; Fujieda et al., 2003] and Ps. putida [Satoh et al., 2002]. 2-Phenylamine is converted to phenylacetaldehyde by 2-phenylethylamine dehydrogenase (2-PEADH) (QhpA1B1C1), which has an α₂β₂γ₂-composition and a rather broad substrate spectrum including also aliphatic amines. Further catabolism branches into the pathway for degradation of phenylalanine (see below). Benzylamine is converted to benzaldehyde by benzylamine dehydrogenase (BAmDH) (QhpA2B1C2), which has an αβγ-composition and a restricted substrate spectrum with only 2-phenylethylamine as additional substrate. Further catabolism branches into the pathway for benzaldehyde degradation (see below). 2-PEADH and BAmDH contain 2 heme c cofactors in their α-subunits and the cysteine tryptophylquinone (CTQ) cofactor. The QhpDEF-paralogs encoded in both qhp operons, are assumed to function as maturation factors. QhpD1 and QhpD2 are S-adenosylmethionine (SAM)-radical enzymes presumably forming the thioether bonds in the γ-subunits [Nakai et al., 2015]. QhpF1 and QhpF2 are membrane-bound proteins possibly involved in exporting the modified γ-subunits to the periplasm [Nakai et al., 2014]. QhpE1 and QhpE2 are predicted serine proteases assumed to cleave the nonstandard signal peptides of the γ-subunits [Nakai et al., 2012]. QhpG is only encoded in the qhp2 operon (BAmDH) and represents a putative FAD-dependent monooxygenase possibly involved in forming the quinoid cofactor in the γ-subunit [Nakai et al., 2014]. The CcmF protein, which is also only encoded in the qhp2 operon (BAmDH), is a predicted membrane-bound cytochrome c maturation protein possibly delivering heme c to the α-subunit in the periplasmic space [Pearce et al., 1998]. Each operon (qhp1 and qhp2) harbors a gene (qhpF1 and qhpF2) encoding the permease of an ABC-type transporter (hybrid with ATPase domain); each permease is assumed to translocate its cognate γ-subunit (qhpC1 and qhpC2) to the periplasm.

Neither of the two aromatic amines requires a dedictated uptake system since their initial oxidative deamination occurs in the periplasm. The unpolar products of this reaction (phenylacetaldehyde and benzaldehyde) can be assumed to enter the cell by passive diffusion.

The genes coding for 2-PEADH and BAmDH and their biogenesis are located in two distinct operon-like structures in the genome of A. aromaticum EbN1^T (Fig. 8B) [Schmitt et al., 2019], resembling the gene organization identified in Pc. denitrificans [Nakai et al., 2014]. Only the qhp2 operon (BAmDH) harbors the qhpR gene encoding an AraC-type regulator, which activates transcription in response to the aromatic amine substrate; notably BAmDH was most strongly induced and active with benzylamine. Whether the qhp1 operon is also under control of QhpR is unclear at present.

The anaerobic degradation of o-phthalate (monomer of industrially produced phthalate esters) was discovered by differential proteomic analyses in members of the genera Aromatoleum and Thauera, and shown to be plasmid-born in A. aromaticum EbN1^T [Ebenau-Jehle et al., 2017]. The following metabolites, reactions, and enzymes are involved in the degradation pathway (Fig. 9A), which proceeds via two steps. (i) o-Phthalate is initially activated to phthaloyl-CoA by the highly specific succinyl-CoA:phthalate CoA transferase (SptAB), with the formed CoA ester intermediate showing an extreme lability [Mergelsberg et al., 2018]. (ii) Phthaloyl-CoA is then decarboxylated by phthaloyl-CoA decarboxylase (Pcd1/2) generating the central intermediate benzoyl-CoA. Pcd of T. chlorobenzoica was demonstrated to have a hexameric structure and to contain prenylated FMN, K⁺ and Fe²⁺ as cofactors. High intracellular concentrations of Pcd in conjunction with the irreversibility of the catalyzed decarboxylation reaction allow for efficient o-phthalate conversion despite the lability of the initial phthaloyl-CoA intermediate [Mergelsberg et al., 2017, 2018]. The sptAB and pcd1 genes form an operon-like structure in the genome of A. aromaticum EbN1^T (Fig. 9B), with the pcd2 gene localized 11.7 kbp upstream of the pcd1 gene.

The sptAB and pcd1 genes are flanked by genes encoding a TRAP transporter (DctMQ and DctP4) and an ABC transporter (EbN1_PII00660/70/80) (Fig. 9B) [Ebenau-Jehle et al., 2017], which could both be involved in the uptake of o-phthalate and possibly further substrates of the Spt-Pcd degradation pathway. Since, however, only the TRAP-transporter is encoded next to the sptAB and pcd genes in T. chlorobenzoica, it may be the obvious candidate for the uptake of o-phthalate.

The spt-pcd genetic neighborhood also harbors a gene coding for an IclR-type one-component sensory/regulatory system (EbN1_PII00710) (Fig. 9B) that could control spt-pcd gene expression in response to o-phthalate, in particular as this gene is also conserved and shows a similar genetic neighborhood in T. chlorobenzoica [Ebenau-Jehle et al., 2017].

The anaerobic degradation of indoleacetate (IAA; synonym: indole-3-acetate), which is also known as the plant hormone auxin, was elucidated in Aromatoleum evansii KB740^T (formerly A. evansii KB740^T) and A. aromaticum EbN1^T. The following metabolites, reactions, and enzymes are involved in the degradation pathway (Fig. 10A), which proceeds via nine steps [Ebenau-Jehle et al., 2012; Schühle et al., 2016, 2021]. (i) In the initial reaction, indoleacetate is activated to its CoA ester by a highly specific indoleacetate-CoA ligase (IaaB). (ii) Indoleacetyl-CoA is then anaerobically hydroxylated to the enol form of 2-oxo-indoleacetyl-CoA by a predicted molybdenum cofactor-containing, heterotrimeric dehydrogenase belonging to the xanthine dehydrogenase family (IaaIJK). (iii) The keto-form of 2-oxo-indoleacetyl-CoA is assumed to be hydrolytically cleaved by an ATP-dependent hydrolase (IaaCE) forming 3′-(2-aminophenyl)succinyl-CoA. (iv) The latter is then isomerized by a phenylsuccinyl-CoA transferase (IaaL) to its regioisomer 2′-(2-aminophenyl)succinyl-CoA, the C-skeleton of which is (v) then rearranged by a methylmalonyl-CoA mutase-like enzyme (IaaGH) to 2-aminobenzylmalonyl-CoA. (vi) The carboxyl group in α-position of 2-aminobenzylmalonyl-CoA is then proposed to be oxidatively removed by a benzylmalonyl-CoA dehydrogenase (IaaF) forming (2-aminophenyl)propenoyl-CoA. The latter is then further catabolized according to a β-oxidation-like reaction sequence via (vii) hydration to 3-hydroxy-3-(2-aminophenyl)-propanoyl-CoA followed by (viii) oxidation to 3-oxo-3-(2-aminophenyl)-propanoyl-CoA catalyzed by a fused enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase (IaaP). (ix) Finally, cleaving off acetyl-CoA by a 3-oxoacyl-CoA thiolase (IaaA) yields 2-aminobenzoyl-CoA. Further degradation of 2-aminobenzoyl-CoA to the central intermediate benzoyl-CoA is as described in section “2-Aminobenzoate” below.

The iaa gene cluster (Fig. 10B) harbors the iaaM gene coding for the periplasmic solute-binding protein of an ABC transporter, potentially being involved in high-affinity uptake of indoleacetate.

All genes encoding proteins for the anaerobic degradation of indoleacetate are organized in a coherent cluster in the genome of A. aromaticum EbN1^T (Fig. 10B). The cluster harbors the iaaQ and iaaR genes encoding transcriptional regulators belonging to the TetR [Cuthbertson and Nodwell, 2013] and GntR families [Rigali et al., 2002], respectively.

The anaerobic degradation of 4-hydroxybenzoate, was mainly elucidated in T. aromatica K172^T and R. palustris [e.g., Breese and Fuchs, 1998; Egland and Harwood, 2000]. The following metabolites, reactions, and enzymes are involved in the degradation pathway (Fig. 11A), which proceeds via two steps. (i) In the initial reaction, 4-hydroxybenzoate is activated to its CoA ester by predicted Mg²⁺/ATP-dependent 4-hydroxybenzoate-CoA ligase (Hcl2), as previously reported for characterized homologs in T. aromatica K172^T and R. palustris [Biegert et al., 1993; Gibson et al., 1994]. Note, that 4-hydroxybenzoyl-CoA is also generated during the anaerobic degradation of 4-methylphenol, 4-ethylphenol, and para-hydroxylated phenylpropanoids. (ii) 4-Hydroxybenzoyl-CoA is then reductively dehydroxylated to the central intermediate benzoyl-CoA by the molybdenum-containing 4-hydroxybenzoyl-CoA reductase (HcrABC) [Gibson et al., 1997; Breese and Fuchs, 1998], which was to date, however, not detected in anaerobic cultures of A. aromaticum EbN1^T by differential proteomics [e.g., Wöhlbrand et al., 2007]. Heterodimeric HcrABC belongs to the xanthine oxidase (XO) family of molybdenum enzymes, possesses an (αβγ)₂ structure, harbors two molybdenum cofactors, four [2Fe-2S] and two [4Fe-4S] clusters as well as two FADs. Ferredoxin serves as electron donor and the active site of Hcr has a sidechain lining that is assumed to stabilize the postulated radical intermediate. Two successive one-electron steps via the molybdenum cofactor (Mo^IV to Mo^VI) to 4-hydroxybenzoyl-CoA according to a Birch-like mechanism are proposed to be involved. Mo^IV is probably regenerated by a two electron and two proton transfer after release of the re-aromatized benzoyl-CoA [Breese and Fuchs, 1998; Boll et al., 2001; Unciuleac et al., 2004a, b; Johannes et al., 2008].

In Ps. putida and Acinetobacter sp. ADP1, PcaK, which belongs to the aromatic acid:H⁺ symporters (AAHS) of the major facilitator superfamily (MFS) [Pao et al., 1998], has been reported to function in uptake of 4-hydroxy-benzoate [Harwood et al., 1994; Pernstich et al., 2014]. The genome of A. aromaticum EbN1^T encodes a (di)hydroxybenzoate transporter (PcaK); however, the pcaK gene is not adjacent to the genes encoding enzymes for anaerobic degradation of 4-hydroxybenzoate (Fig. 11B).

In R. palustris, the HbaR protein, which belongs to the Crp-Fnr superfamily [Körner et al., 2003], activates the transcription of the genes for anaerobic degradation of 4-hydroxybenzoate [Egland and Harwood, 2000]. While the genome of A. aromaticum EbN1^T does not contain a gene unambiguously encoding a HbaR homolog in direct neighborhood to the “catabolic” genes, a TaoR-like transcriptional regulator belonging to the LysR-family is encoded (Fig. 11B) [Maddocks and Oyston, 2008; Rajeev et al., 2019].

Anaerobic degradation of 3-hydroxybenzoate was mainly investigated with T. aromatica K172^T [Laempe et al., 2001] and A. aromaticum EbN1^T [Rabus et al., 2005; Wöhlbrand et al., 2007]. The current differential metabolite, transcript, and proteomic data suggest a pathway deviating from the one hypothesized earlier [Wöhlbrand et al., 2007]. Next to the enzymes encoded in the “3-hydroxybenzoate” gene cluster also Oah from the central benzoyl-CoA pathway is involved in feeding into 3-hydroxypimeloyl-CoA. The following metabolites, reactions, enzymes, and genes are involved in the degradation pathway (Fig. 12A, B), which proceeds via eight steps. (i) Early studies with T. aromatica K172^T revealed inducible formation of a 3-hydroxybenzoate-CoA ligase (Hcl1) activating 3-hydroxybenzoate to 3-hydroxybenzoyl-CoA [Laempe et al., 2001], which was subsequently corroborated by means of differential proteomics with A. aromaticum EbN1^T [Wöhlbrand et al., 2007]. (ii) Then, 3-hydroxybenzoyl-CoA is reductively dearomatized by a paralog of class I BCR (termed HbrADBC) forming 3-hydroxydienoyl-CoA, i.e., the meta-hydroxy group is retained. Please note that the paralogous subunit pairs BcrA/HbrA and BcrD/HbrD share 100% sequence identity. Therefore, abundance profiles reflect the pool size in each case, hindering the assignment of the specific abundance value to each individual paralog. (iii) The 3-hydroxy-dienoyl-CoA is then reduced to 5-hydroxycyclohex-2-ene-1-carbonyl-CoA by a predicted (and experimentally detected) dehydrogenase (EbN1_C03840), which is probably FAD-dependent. (iv) The 5-hydroxy-enoyl-CoA is then hydrated to 2,5-dihydroxycyclohex-1-carbonyl-CoA by a predicted (and experimentally detected) enoyl-CoA hydratase/isomerase (EbN1_C03750 or EbN1_C03770). (v) 2,5-Dihydroxycyclohex-1-carbonyl-CoA is then assumed to be dehydrated to 6-hydroxy-cyclohex-2-ene-1-carbonyl-CoA again by either of the two enoyl-CoA hydratases/isomerases (EbN1_C03750 and EbN1_C03770). (vi) Then water is added to the double bond of 6-hydroxycyclohex-2-ene-1-carbonyl-CoA yielding 2,6-dihydroxycyclohex-1-carbonyl-CoA, a reaction presumably catalyzed again by either of the two enoyl-CoA hydratases/isomerases (EbN1_C03750 and EbN1_C03770) or possibly by another enoyl-CoA hydratase (Dch); however, Dch is a constituent of the anaerobic benzoyl-CoA pathway. Note that the speci­ficity of these two enoyl-CoA hydratases/isomerases (EbN1_C03750 and EbN1_C03770) for substrate positions 2, 5, and 6 subsequently targeted during the three aforementioned reactions cannot be differentiated at present [Becker et al., 2022a]. The hydroxy-group shift from C5 to C6 is reminiscent of the hydroxy-group shift occurring during catabolism of levulinic acid in Ps. putida KT2440, where 4-hydroxyvaleryl-CoA is converted to 3-hydroxyvaleryl-CoA involving 4-phosphovaleryl-CoA as key intermediate [Rand et al., 2017]. Such a mechanism is most likely not involved in A. aromaticum EbN1^T since the key enzyme phosphoryl transferase (LvaAB) is not encoded in its genome. (vii) 2,6-Dihydroxycyclohex-1-carbonyl-CoA is then oxidized to 2-oxo-6-hydroxycyclohex-1-carbonyl-CoA by 6-hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase (EbN1_C03780), which is specifically detected in cells adapted to anaerobic growth with 3-hydroxybenzoate. Alternatively, a similarily specifically detected oxoacyl-CoA reductase (EbN1_C03850) could be involved in this reaction step, but also functions in a possible side route of anaerobic degradation of 3-hydroxybenzoate [Becker et al., 2022a]. (viii) Finally, the ring structure of 2-oxo-6-hydroxycyclohex-1-carbonyl-CoA is hydrolytically cleaved to 3-hydroxy-pimelyl-CoA by 6-oxocyclohex-1-ene-1-carbonyl-CoA hydratase (Oah), recruited from the central benzoyl-CoA pathway.

The chromosomal locus for anaerobic degradation of 3-hydroxybenzoate encodes, next to the catabolic enzymes described above, also two uptake systems. First, in-between EbN1_C03780 and hcl1 the EbN1_C03790 gene is localized, which codes for a predicted MFS transporter. However, only low transcript and protein abundances could be observed [Becker et al., 2022a]. Second, in-between the hcl1 and EbN1_C03840 genes, the tarPMQ genes are localized, which code for a TRAP transporter (lignin-derived aromatic compounds). Notably, transcripts and protein components of this transporter are abundant during anaerobic as well as aerobic growth with 3-hydroxybenzoate [Becker et al., 2022a]. This implies that the TarPMQ system could be responsible for the uptake of this aromatic growth substrate.

Downstream of the EbN1_C03750 gene, the EbN1_C03740 gene is localized, which encodes an TetR-like transcriptional regulator. The latter could thus control transcription of the genes for the uptake and anaerobic degradation of 3-hydroxybenzoate.

The anaerobic degradation of phenylacetate (via phenyl­acetyl-CoA) was elucidated by means of biochemical studies with A. evansii KB740^T and T. aro-matica K172^T. The genes encoding the protein constituents of this pathway cluster together in an operon-like structure and have been identified in the genomes of A. evansii KB740^T and A. aromaticum EbN1^T [Rabus et al., 2005; Wöhlbrand et al., 2007]. The following metabolites, reactions, enzymes, and genes are involved in the degradation pathway (Fig. 13A, B), which proceeds via three steps. (i) The initial step of anaerobic phenylacetate degradation is the activation to phenylacetyl-CoA by phenylacetate-CoA ligase (PadJ), as first demonstrated with T. aromatica K172^T [Mohamed and Fuchs, 1993]. Note that anaerobic degradation of phenylalanine branches in at this point. (ii) Phenylacetyl-CoA is then subject to a four-electron α-oxidation yielding phenylglyoxylate. This reaction is catalyzed by heterotrimeric phenylacetyl-CoA:acceptor oxidoreductase (synonym: phenylacetyl-CoA dehydrogenase) (PadBCD), which was discovered and characterized in T. aromatica K172^T. The enzyme is membrane-bound, presumably uses ubiquinone as an electron acceptor, contains molybdenum as well as iron-sulfur clusters, and has an apparent (αβγ)₂ structure. While the enzyme is remarkably stable against exposure to oxygen, it is not formed during aerobic growth with phenylalanine. Intermediary released phenylglyoxylyl-CoA is proposed to be cleaved by the same enzyme to phenylglyoxylate and CoA. Formation of phenylglyoxylyl-CoA and phenylglyoxylate was demonstrated with enzyme assays provided with [1-¹⁴C] phenylacetyl-CoA. Based on homology analyses, the large subunit (PadB) is suggested to carry the molybdenum cofactor [Schneider and Fuchs, 1998; Rhee and Fuchs, 1999]. Specific formation of PadA by A. aromaticum EbN1^T [Becker et al., 2022a] may hint to a functional role, which is most likely that of a chaperone during maturation of the bis-MGD cofactor of PadBCD, as known from homologous FdhD chaperone of formate dehydrogenase in E. coli [Arnoux et al., 2015; Schwanhold et al., 2018]. In A. aromaticum EbN1^T, the predicted thioesterase EbN1_C31020 (here named PadK) was detected with highest abundances (transcript and protein level) during anaerobic growth with phenylalanine [Becker et al., 2022a]. Moreover, the padK gene is part of the pad gene locus. Taken together, the PadK protein is the more suitable candidate for catalyzing the removal of CoA from phenylglyoxylyl-CoA compared to PadBCD, also considering that the latter does not possess a recognizable domain for such an enzymatic activity. A further alternative thioesterase, based on transcript/protein abundance profiles, could be EbN1_C20160 [Becker et al., 2022a]. (iii) Finally, oxidative decarboxylation of phenylglyoxylate to the central intermediate benzoyl-CoA is achieved by multi-subunit, FeS-containing phenylglyoxylate:NAD⁺ oxidoreductase (PadEFGHI), purified and biochemically characterized in A. evansii KB740^T. The core enzyme is suggested to have an α₂β₂γ₂δ₂ composition with the additional ε₂ units transferring electrons from a small ferredoxin-like subunit of the core enzyme to NAD⁺. The enzyme was found to be specifically formed in cells growing anaerobically with phenylalanine, phenylacetate or phenylglyoxylate. The enzyme was found to be highly oxy­gen-sensitive and not formed in cells grown aerobically with phenylalanine [Hirsch et al., 1998].

The genetic neighborhood of the pad gene cluster provides no hints for a possible uptake system for phenylace­tate.

The padR gene is localized less than 1 kbp upstream of the pad gene clusters (opposite strand). The encoded PadR protein probably functions as transcriptional repressor, since it shares a domain with the PaaX protein, which was previously reported to repress transcription of the paa gene cluster for aerobic degradation of phenylacetate in E. coli [Ferrández et al., 2000].

Anaerobic degradation of 2-aminobenzoate has been investigated in T. aromatica K172^T and A. aromaticum EbN1^T and was suggested to proceed via two steps (Fig. 14A). (i) The degradation is initiated by benzoate-CoA ligase, activating 2-aminobenzoate to its CoA ester (see above), albeit at a substantially lower rate compared to benzoate; the CoA ligase apparently functions anaerobically as well as aerobically [Schühle et al., 2003]. Further degradation then separates into an anaerobic and an aerobic branch. (ii) Under anoxic conditions, 2-aminobenzoyl-CoA is then assumed to be reductively dearomatized by HbrCABD yielding 6-aminocyclohex-1-ene-1-carbonyl-CoA [Wöhlbrand et al., 2007]. The amino group is then spontaneously removed in the presence of water yielding 6-oxocyclohex-1-ene-1-carbonyl-CoA, which represents a downstream intermediate of the central anaerobic benzoyl-CoA pathway.

The gene cluster framed by the hcl1 and hbrCABD genes also encodes a TRAP-type uptake system (dctPM), which could function in import of 2-aminobenzoate (Fig. 14B).

No hints for a possible sensory/regulatory system responsive to 2-aminobenzoate can be inferred from the genetic neighborhood.

Insights into the challenging anaerobic degradation of benzoate by denitrifying bacteria mainly emerged from biochemical investigations using T. aromatica K172^T. Genes coding for proteins involved in this pathway were determined for several bacteria, e.g., R. palustris [Egland et al., 1997], T. aromatica K172^T [Breese et al., 1998; Schühle et al., 2003], Azoarcus sp. strain CIB (needs reclassification as Aromatoleum sp.) [López Barragán et al., 2004], and A. aromaticum EbN1^T [Rabus et al., 2005]. The following metabolites, reactions, enzymes, and genes are involved in the degradation pathway (Fig. 15A, B), which proceeds via five steps. (i) Benzoate is activated to benzoyl-CoA by ATP-dependent benzoate-CoA ligase (Bcl) [Breese et al., 1998]. (ii) The reductive dearomatization of benzoyl-CoA via a Birch-like reaction yielding trans-cyclohex-1,5-diene-1-carbonyl-CoA (dienoyl-CoA) is catalyzed by Mg²⁺/ATP-dependent, FeS cluster-containing, oxygen-sensitive, four-subunit benzoyl-CoA reductase (BcrCBAD). The dienoyl-CoA product (from [¹⁴C]benzoyl-CoA) was identified by TLC. The electrons for ring reduction are donated by reduced ferredoxin (Fxd), which is delivered by a 2-oxoglutarate:ferredoxin oxidoreductase (KorAB) [Boll and Fuchs, 1995; Boll et al., 1997; Dörner and Boll, 2002; Möbitz and Boll, 2002; Thiele et al., 2008]. While fdx and korAB genes are localized in proximity to the bcr genes in T. aromatica K172^T, this is only the case for fdx with A. aromaticum EbN1^T. However, a 2-oxoglutarate:ferredoxin oxidoreductase is also encoded in the genome of A. aromaticum EbN1^T (EbN1_C26200-20); while no differential protein abundance profiles were observed, transcript profiles were 5‒8-fold higher during anaerobic growth with phenylalanine, benzoate, 3-hydroxybenzoate, and 3-(4-hydroxyphenyl)propanoate as compared to the other tested substrate adaptation conditions [Becker et al., 2022a]. Notably, members of the genera Thauera and Aromatoleum possess different subtypes of class I BCR [e.g., Buckel et al., 2014]. (iii) The dienoyl-CoA is then hydrated to 6-hydroxycyclohex-1-ene-1-carbonyl-CoA catalyzed by cyclo­hex-1,5-diene-1-carbonyl-CoA hydratase (Dch). The metabolite identification was based on reactions with [ring-¹³C₆]benzoyl-CoA or [ring-¹⁴C]benzoyl-CoA and the purified enzyme, followed by HPLC analysis and NMR spectroscopy [Laempe et al., 1998; Boll et al., 2000]. (iv) 6-Hydroxycyclohex-1-ene-1-carbonyl-CoA is oxidized to 6-oxocyclohex-1-ene-1-carbonyl-CoA by NAD⁺-dependent 6-hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase (Had). Metabolite identification was based on enzymatic assays with [phenyl-¹⁴C]6-hydroxycyclohex-1-ene-1-carbonyl-CoA coupled to HPLC separation, radioactivity detection, and UV spectra [Laempe et al., 1999]. (v) 6-Oxocyclohex-1-ene-1-carbonyl-CoA is then subject to hydrolytic ring cleavage yielding 3-hydroxypimelyl-CoA catalyzed by 6-oxocyclo­hex-1-ene-1-carbonyl-CoA hydratase (Oah). 3-Hydroxypimeloyl-CoA detection was as described above for 6-oxocyclohex-1-ene-1-carbonyl-CoA [Laempe et al., 1999]. Further degradation proceeds via conventional β-oxidation.

Genes coding for a predicted ABC transporter (EbN1_C30450-90) and for a symporter (benK) are localized directly upstream of the bcl1 gene. The transcripts and protein components of both of these transporters were essentially observed at high abundances during anaerobic and aerobic growth with all studied aromatic substrates. In the case of acetate, the abundance profiles were for the most part markedly lower [Trautwein et al., 2012a; Becker et al., 2022a].

The gene cluster for anaerobic degradation of benzoate is preceded by the bzdR gene. BzdR was previously shown in related Aromatoleum sp. strain CIB to act as transcriptional repressor being liberated from the P_N promoter in the presence of benzoyl-CoA [Durante-Rodríguez et al., 2010]. Notably, BzdR displays high sequence similarity with BoxR of the aerobic box pathway (see section “Aerobic degradation of aromatic compounds, benzoate” below). In the case of Aromatoleum sp. strain CIB, both proteins were shown to synergistically control, in a benzoyl-CoA-dependent manner, the transcription of the genes encoding the enzyme constituents of the anaerobic as well as aerobic benzoyl-CoA pathways [Valderrama et al., 2012].

Anaerobic and aerobic degradation pathways for phenylalanine share the initial deamination followed by shared and specific reactions yielding the joint intermediate phenylacetyl-CoA [Rabus et al., 2005; Wöhlbrand et al., 2007]. The following metabolites, reactions, enzymes, and genes are involved in the degradation pathway (Fig. 16A, B), which proceeds via up to five steps. (i) Initially, the NAD⁺-dependent phenylalanine:2-oxo-glutarate transaminase (TyrB; broad substrate spectrum aromatic-amino-acid transaminase) oxidatively deaminates phenylalanine to phenylpyruvate. This reaction was previously described for A. aromaticum EbN1^T using differential proteomics [Wöhlbrand et al., 2007] and for T. aro­matica K172^T using enzyme activity measurements [Schneider et al., 1997]. The reaction/enzyme functions under anoxic as well as oxic conditions. (ii) Phenylpyruvate is then converted to phenylacetaldehyde via decarboxylation, catalyzed by indolepyruvate decarboxylase (IpdC) [Schneider et al., 1997; Rabus et al., 2005; Wöhlbrand et al., 2007], predicted to have a broad substrate spectrum (e.g., pyruvate, indolepyruvate, and phenylpyruvate). The reaction/enzyme functions under anoxic as well as oxic conditions. (iii) In principle, a direct conversion of phenylpyruvate to phenylacetyl-CoA could be catalyzed by an indolepyruvate:ferredoxin oxidoreductase (IorBA). Previous proteomic and transcriptomic data [Wöhlbrand et al., 2007; Becker et al., 2022a] provide circumstantial evidence for this possibility. However, this enzyme was only detected at low scores, implicating an only minor role, if at all, in anaerobic degradation of phenylalanine in A. aromaticum EbN1^T. (iv) Oxidation of phenylacetaldehyde to phenylacetate can be achieved by two different enzymes: First, NAD(P)⁺-dependent phenyl­acetaldehyde dehydrogenase (Pdh; synonym: StyD according to Beltrametti et al. [1997] and Becker et al. [2022a]), which is, however, detected only with small score values by proteomics. Second, a new tungsten-containing, aldehyde oxidoreductase (AOR; synonym: aldehyde dehydrogenase [NAD⁺]) was suggested to catalyze this reaction and denoted AOR_Aa, in order to differentiate it from the various other prokaryotic AORs, but is apparently rather unspecific as it serves a broad variety of aromatic and aliphatic aldehydes [Wöhlbrand et al., 2007; Debnar-Daumler et al., 2014; Schmitt et al., 2017; Arndt et al., 2019]. AOR_Aa (like class II BCR) belongs to the obligate tungsten-dependent members of the AOR family [Seelmann et al., 2020]. Notably, AOR_Aa can also reduce carboxylic acids and NAD⁺ employing hydrogen as electron donor [Winiarska et al., 2022] and is markedly resistant against exposure to air [Arndt et al., 2019]. The AroA subunit harbors 4 FeS clusters (FS1‒FS4), AorB carries a single FeS cluster (FS0) and the enzymatically active tungsten(W-bis-MPT) cofactor (Wco), AorC contains the FAD cofactor, where NAD⁺ is reduced, and AorD and AorE are assumed to function as maturation factors for the tungsten cofactor. Most recent structural analysis of the heterologously overproduced, purified enzyme by means of cryo-electron microscopy revealed that AOR_Aa forms a heteromultimeric filament (Aor[AB]_nC) from multiples of AorAB proteomers, which are oligomerized starting from a single AorC. While the Wco-containing AorBs locate to the outer face of the filament, the AorAs are centrally positioned, allowing their numerous FeS clusters to be aligned into a “nanowire”-like structure. The latter can be highly charged with electrons and forms a connecting conduit between the multiple Wcos and the single FAD cofactor. Furthermore, these FeS-filaments are shielded from the enzyme's surface, whereby the stability against oxygen exposure is achieved [Winiarska et al., 2023]. Based on these structural insights and QM:MM modeling of the W-bis-MPT cofactor, the following intriguing mechanism was hypothesized: Initially, the aldehyde is activated to the 1,1-geminal alcohol by spontaneous addition of water. Then, the α-hydrogen is abstracted from the carbonyl carbon and bound by one of the oxo ligands of the tungsten cofactor, which is thereby reduced. Finally, a proton is abstracted from the carbocation intermediate of the aldehyde, leading to the formation of the carboxylic acid product. Two one-electron transfers from the reduced tungsten cofactor, via the FS0‒FS4 nanowire to the FAD cofactor, restore its catalytic active, oxidized form [Winiarska et al., 2023]. Furthermore, efficient electrocatalytic oxidation of aldehydes could be demonstrated with AOR_Aa immobilized on glassy carbon working electrodes and applying artificial electron acceptors [Kalimuthu et al., 2023]. Most recently, AOR_Aa was employed as key enzyme for ATP production directly from electricity via the synthetic acid/aldehyde ATP (AAA) cycle [Luo et al., 2023]. (v) Activation of phenylacetate to its CoA ester is performed by two different phenylacetate-CoA ligases employed in the anaerobic (PadJ) and aerobic pathway sections (PaaK1/2), respectively. Noteworthy, PadJ and PaaK1 displayed highest transcript and protein abundance when phenylalanine was provided as substrate, but in both cases during anaerobic as well as aerobic growth. Furthermore, gene expression and protein formation were found for these two CoA ligases also with other tested substrates [Becker et al., 2022a]. The reaction/enzyme functions under anoxic as well as oxic conditions.

Uptake of phenylalanine during anaerobic and aerobic growth is probably mediated by an ABC transporter encoded in proximity to the 1^stpaa gene cluster. Further details are provided in section “Phenylacetate (aerobic)” below.

No genes for potential transcriptional regulators were found in the vicinity of the pat and pdc genes. A LysR-type regulator (EbN1_C28810) is encoded about 2.6 kbp downstream of the aor genes and a two-component sensory/regulatory system (EbN1_C00670-80) directly upstream of the iorBA genes. However, since the “catabolic” genes of the anaerobic and aerobic degradation of phenylalanine are scattered across the genome, it remains unclear at present whether these sensory/regulatory systems are indeed involved in their transcriptional regulation.

Plant-derived 3-phenylpropanoids support anaerobic as well as aerobic growth of A. aromaticum EbN1^T. The respective gene cluster and β-oxidation-type degradation pathway were previously elucidated by combining differential proteogenomics and targeted metabolite analyses [Trautwein et al., 2012b]. This gene cluster is adjacent to the one for anaerobic degradation of benzoate, hinting at a prominent role of 3-phenylpropanoids for the nutrition of strain EbN1^T under anoxic conditions. In the case of 3-(4-hydroxyphenyl)propanoate, the β-oxidation of the acyl-sidechain furnishing the central benzoyl-CoA involves the following metabolites, reactions, enzymes, and genes (Fig. 17A, B), and proceeds via six steps. (i) 3-(4-Hydroxyphenyl)propanoate is first activated by a 3-phenylpropanoate-CoA ligase (synonym: trans-feruloyl-CoA synthase) (PprA) forming 3-(4-hydroxyphenyl)propanoyl-CoA. (ii) 3-(4-Hydroxyphenyl)propanoyl-CoA is then oxidized by a 3-phenylpropanoyl-CoA dehydrogenase (synonym: acyl-CoA dehydrogenase) (PprB) yielding 4-coumaroyl-CoA. (iii) The latter is then hydrated at the double bond of its acyl-CoA sidechain by a 3-phenylpropenoyl-CoA 2,3-hydratase (synonym: enoyl-CoA hydratase/isomerase) (PprC) to 3-hydroxy-3-(4-hydroxyphenyl)propanoyl-CoA. (iv) The latter is then oxidized by a 3-aryl-3-hydroxypropanoyl-CoA dehydrogenase (synonym: 3-hydroxyacyl-CoA dehydrogenase) (PprD) forming 4-hydroxy-benzoylacetyl-CoA. (v) The latter is then thiolytically cleaved by a proposed 4-hydroxy-benzoylacetyl-CoA thiolase (synonym: acetyl-CoA C-acetyltransferase) (PprE) into 4-hydroxybenzoyl-CoA and acetyl-CoA. (vi) Finally, 4-hydroxybenzoyl-CoA is proposed to be reductively dehydroxylated by 4-hydroxybenzoyl-CoA reductase (HcrBAC) to benzoyl-CoA [Brackmann and Fuchs, 1993; Unciuleac et al., 2004b]. The complete enzyme set for conversion of 3-(4-hydroxyphenyl)propanoate to 4-hydroxybenzoyl-CoA (PprACEDB) is specifically formed during aerobic and anaerobic growth with this phenylpropanoid, as was also observed for the respective mRNAs [Trautwein et al., 2012b; Becker et al., 2022a]. On the contrary, specific formation of neither transcripts nor protein components of Hcr was detectable [Wöhlbrand et al., 2007; Becker et al., 2022a]. This pathway has a broad substrate range, since it serves the degradation of seven further growth-supporting phenylpropanoids (cinnamyl alcohol, 3-phenylpropanoate, cinnamate, m- and p-coumarate, caffeate, and 3-(3,4-dihydroxyphenyl)propanoate) and performs co-metabolic transformation of non growth-supporting phenylpropanoids (e.g., o-coumarate, ferulate, and sinapate) [Trautwein et al., 2012b].

The degradation of protocatechuate (synonym: 3,4-dihydroxybenzoate) is assumed to be channeled into the 3-hydroxybenzoate degradation pathway following a hypothesized activation to its CoA ester and reductive dehydroxylation [Trautwein et al., 2012b].

The genes encoding an ABC transporter and a symporter are directly flanked by the gene clusters for degradation of 3-(4-hydroxyphenyl)propanoate (anaerobic and aerobic) and benzoate (anaerobic), suggesting their shared utilization for the uptake of both types of substrates. Notably, a 3-phenylpropanoid-specific periplasmic solute-binding protein (PprF) is encoded in the “3-phenylpropanoid” gene cluster. Furthermore, the distantly encoded TRAP transporter TarPMQ is predicted to serve lignin-derived aromatic compounds, however, no transcript and protein abundance increases were observed in response to the presence of the pathway substrate 3-(4-hydroxyphenyl)propanoate [Becker et al., 2022a].

The pprR gene codes for a predicted transcriptional repressor and has previously been indicated in transcriptional control of the “3-phenylpropanoid” gene cluster [Trautwein et al., 2012b]. Most recent studies in our group proved this function based on unmarked, in-frame deletion mutagenesis and further revealed that catabolite repression and additional transcriptional activation should be involved in regulation of gene expression [Vagts et al., 2021]. The transcriptional response threshold toward 3-phenylpropanoids was recently found to also be in the nanomolar range [Vagts et al., 2021].

Benzyl alcohol is degraded via benzaldehyde to benzoate (Fig. 18A) involving sequential activity of benzyl alcohol dehydrogenase (AdhB) and benzaldehyde dehydrogenase (Ald3) [Wöhlbrand et al., 2007], as also reported for T. aromatica K172^T [Biegert et al., 1995] and Ps. putida [Shaw and Harayama, 1990]. The coding genes in A. aromaticum EbN1^T are distinctly located on the chromosome (Fig. 18B). Further catabolism proceeds via the respective anaerobic versus aerobic pathways for benzoate.

Due to their rather lipophilic properties, benzyl alcohol and benzaldehyde are not expected to require a specific uptake system, but rather enter the cell by passive diffusion.

No hints for specific sensory/regulatory systems can be found in the genetic neighborhood of the “catabolic” genes.

The aerobic degradation of 3-hydroxybenzoate has long been studied in various bacterial genera, such as Klebsiella, Rhodococcus, and Pseudomonas. The gene nomenclature of this pathway is rather diverse. For example, in Ralstonia sp. strain U2 the nag gene designation is used, reflecting the degradation of naphthalene via gentisate [Zhou et al., 2001], contrasted by the gen and nar gene designations in Rhodococcus opacus R7 [Di Gennaro et al., 2010] and Rhodococcus sp. strain NCIMB 12038 [Liu et al., 2011], respectively. We adopt the one used by Xu et al. [2012], which is more intuitive with respect to the degradation pathway presented here: mhb for meta-hydroxybenzoate. The mhb gene cluster from Klebsiella pneumoniae M5a1 is depicted in Figure 19B and compared to its counterpart in A. aromaticum EbN1^T (EbN1_C07550-630). The gene order is rather similar and the encoded enzymes of the pathway share amino acid sequence identities in the range of 35–60%. Notably, a paralogous partial gene cluster exists in the genome of A. aromaticum EbN1^T (EbN1_C32310-90), albeit with significantly lower sequence identities of its encoded proteins (30%) and lack of expression [Becker et al., 2022a]. According to the current understanding, aerobic degradation of 3-hydroxbenzoate involves the following metabolites, reactions, enzymes, and genes (Fig. 19A, B), and proceeds via four steps. (i) Initially, 3-hydroxy-benzoate is converted by an NADH-dependent monooxygenase (MhbM), (synonym: 3-hydroxybenzoate 6-hydroxylase), to gentisate (2,5-dihydroxybenzoate) [e.g., Gao et al., 2005]. Structural and modeling approaches with the homolog NagI from E. coli and Silicibacter pomeroyi implicated gentisate and O₂ to bind in vicinity of the Fe²⁺ atom in the active site [Adams et al., 2006; Chen et al., 2008]. (ii) Ring cleavage of gentisate is then achieved by gentisate 1,2-dioxygenase (MhbD) yielding maleylpyruvate [e.g., Harpel and Lipscomb, 1990]. The homolog NagL was shown to have a dimeric structure, to possess the glutathione S-transferase fold and to attack the carbon atom 2 of the substrate with the glutathione thiolate [Marsh et al., 2008; Fang et al., 2011; Hong et al., 2019]. (iii) Maleyl­pyruvate is then transformed to its isomer fumarylpyruvate by maleylpyruvate isomerase (MhbI) [e.g., Jones and Cooper, 1990]. (iv) Finally, fumarylpyruvate is hydrolytically cleaved into pyruvate and fumarate by the fumarylpyruvate hydrolase (MhbH) [e.g., Bayly et al., 1980], which belongs to the large fumarylacetoacetate hydrolase (FAH) superfamily [Hong et al., 2020].

A putative MFS-type transporter (EbN1_C07630) of unknown substrate specificity is encoded one gene further downstream of the mhb gene cluster on the same strand. Since expression of EbN1_C07630 could not be observed across the tested growth conditions [Becker et al., 2022a], its possible role in the uptake of 3-hydroxybenzoate remains unclear at present.

A LysR-type transcriptional regulator (MhbR) is encoded directly upstream of the mhb gene cluster on the opposite strand. However, no transcripts of mhbR could be detected so far. Two further genes upstream on the same strand, another LysR-type transcriptional regulator (EbN1_C07550) is encoded, which displays low transcript abundances during aerobic growth [Becker et al., 2022a]. At present it is unclear which of the two LysR-type regulator (if at all) is responsible for transcriptional control of the mhb gene cluster.

First insights into the aerobic degradation of phenylacetate (via phenylacetyl-CoA) emerged from molecular genetic studies with Ps. putida and E. coli [Ferrandez et al., 1998; Olivera et al., 1998]. Subsequently, the involved reaction sequence was elucidated using purified enzymes and mass spectrometric- as well as NMR-based analyses of intermediates formed from [U-¹³C₈]phenylacetate [Ismail et al., 2003; Teufel et al., 2010]. This pathway is widespread in bacteria [Teufel et al., 2010] and has also been demonstrated in A. evansii KB740^T [Mohamed et al., 2002]. The unconventional conversion of phenylacetate into two acetyl-CoA and one succinyl-CoA involves the following metabolites, reactions, enzymes, and genes (Fig. 20A, B), and proceeds via eight steps. (i) Unusually for aerobic degradation, phenylacetate is initially activated to phenylacetyl-CoA by AMP-forming phenylacetate-CoA ligase (PaaK1). (ii) Subsequent epoxidation of phenyl­acetyl-CoA forming ring 1,2-epoxy-phenylacetyl-CoA is then catalyzed by a phenylacetyl-CoA 1,2-epoxidase (PaaA1B1C1E1). The PaaA subunit harbors the catalytically relevant di-iron center. The enzyme complex has a composition of PaaA₂B_3–4C₂E₁ or multiples of this. Notably, PaaABCE also performs the reverse reaction, i.e., the reductive removal of the epoxide oxygen forming phenyl­acetyl-CoA, which was suggested to represent a relevant detoxification route for toxic epoxides. Taken together, the PaaA1B1C1E1 enzyme represents a bifunctional monooxygenase/deoxygenase [Teufel et al., 2012]. The role of subunit D is at present unclear, it may be involved in protein maturation. (iii) The epoxide is then isomerized to 2-oxepin-2[3H]-ylideneacetyl-CoA by oxepin-CoA forming ring 1,2-epoxy-phenylacetyl-CoA isomerase (PaaG) [Teufel et al., 2010]. PaaG is a bifunctional enzyme also isomerizing 2,3-dehydroadipyl-CoA (cis to trans) downstream in the degradation pathway. (iv) Ring opening of 2-oxepin-2[3H]-ylideneacetyl-CoA proceeds via two steps catalyzed by the bifunctional fusion protein oxepin-CoA hydrolase/3-oxo-5,6-dehydrosuberyl-CoA semialdehyde dehydrogenase (PaaZ1). First, the C-terminal (R)-specific enoyl-CoA hydrolase domain (PaaZ1-ECH) cleaves the ring forming the highly reactive 3-oxo-5,6-dehydrosuberyl-CoA semialdehyde. Second, the latter is then oxidized by the N-terminal NADP⁺-dependent aldehyde dehydrogenase domain (PaaZ1-ALDH) yielding 3-oxo-5,6-dehydrosuberyl-CoA [Teufel et al., 2011]. (v) The subsequent conversion of 3-oxo-5,6-dehydrosuberyl-CoA corresponds to a conventional β-oxidation reaction sequence as outlined in the following. A 3-oxoacyl-CoA thiolase (PaaJ1, encoded in ∼1 Mbp distance to the 1^stpaa gene cluster and formed during aerobic growth with phenylalanine) then thiolytically removes an acetyl-CoA moiety from 3-oxo-5,6-dehydrosuberyl-CoA yielding in cooperation with PaaG trans-2,3-dehydroadipyl-CoA. PaaJ1 is a bifunctional thiolase also catalyzing the last step of the pathway, see below. (vi) trans-2,3-Dehydroadipyl-CoA is then hydrated by 2,3-dehydroadipyl-CoA hydratase (PaaF) forming 3-hydroxyadipyl-CoA. (vii) The latter is oxidized by NAD⁺-dependent 3-hydroxyadipyl-CoA dehydrogenase (PaaH), which is probably (S)-specific, forming 3-oxoadipyl-CoA. (viii) Finally, 3-oxoadipyl-CoA is thiolytically cleaved to succinyl-CoA and acetyl-CoA by the same 3-oxoacyl-CoA thiolase (PaaJ1), employed before (step v) for 3-oxo-5,6-dehydrosuberyl-CoA processing.

The EbN1_C20250-80 genes are organized in an operon-like structure, located in proximity (downstream) to the 1^stpaa gene cluster, and encoding the components of an ABC-type transporter, i.e., 1 periplasmic solute-binding protein, 2 transmembrane permeases, and 1 cytoplasmic ATP-binding protein. All EbN1_C20250-80 genes were expressed at high level during anaerobic and aerobic growth with phenylalanine and the respective protein products were detected with high meta scores in particular in the membrane protein-enriched fraction. Transcript and protein levels were markedly lower under the other tested substrate conditions, pointing to a specificity of this ABC transporter for phenylalanine [Becker et al., 2022a].

The genome of A. aromaticum EbN1^T harbors two paa gene clusters (EbN1_C20120-230 and EbN1_C32870-950). The 1^stpaa gene cluster was found to be expressed [Becker et al., 2022a] and ends with the paaR gene coding for a predicted transcriptional repressor, as previously demonstrated for E. coli [Ferrández et al., 2000]. The 2^ndpaa gene cluster is preceded by genes encoding a two-component sensory/regulatory system (EbN1_C32850/60). Both regulatory systems were found to be expressed. However, in accord with the 1^stpaa gene cluster being responsible for aerobic degradation of phenylacetyl-CoA, the paaR gene showed highest transcript abundance during growth with phenylalanine (from which phenylacetyl-CoA is derived) [Becker et al., 2022a].

Aerobic degradation of N-heterocyclic compounds such as tryptophan and indole is long known to proceed via 2-aminobenzoate and catechol or gentisate [Cain, 1968; Anderson and Dagley, 1981]. An unconventional pathway for aerobic degradation of 2-aminobenzoate has been discovered to be encoded in the genome of A. evansii KB740^T and was found to proceed via two steps (Fig. 21A). (i) As described above for the anaerobic degradation of 2-aminobenzoate, benzoate-CoA ligase is assumed to activate 2-aminobenzoate to its CoA ester [Schühle et al., 2003]. (ii) Then, 2-aminobenzoyl-CoA is converted by homodimeric, bifunctional 2-aminobenzoyl-CoA monooxygenase/reductase (AbmA) to 2-amino-5-oxocyclohex-1-ene-1-carboxyl-CoA [Langkau et al., 1990; 1995]. The coding genes (amb) were identified in A. evansii KB740^T (Fig. 21B). In A. aromaticum EbN1^T there is fair evidence for the presence of an ambA gene, which is isolated on the genome contrasting the operon-like structure known from of A. evansii KB740^T [Schühle et al., 2001]. Further degradation of 2-amino-5-oxo-cyclohex-1-ene-1-carboxyl-CoA then proceeds via β-oxidation, which in the case of A. aromaticum EbN1^T could be realized by sharing the respective enzymes of the anaerobic degradation of 3-hydroxybenzoate:3-oxoacyl-CoA dehydrogenase (EbN1_C03850, AbmB-like), enoyl-CoA hydratase (EbN1_C03770, AbmC-like), and acyl-CoA dehydrogenase (EbN1_C03840, AbmD-like).

Identification of a gene (abmH) coding for a solute-binding protein of an ABC transporter in A. evansii KB740^T [Schühle et al., 2001] implicated uptake of 2-aminobenzoate via such an active transporter. In A. aromaticum EbN1^T no equivalent gene is present in the vicinity of the abmA gene.

Studies with A. evansii KB740^T revealed coordinated expression of the abm genes in response to 2-aminobenzoate and suggested the transcriptional control to be mediated by the MarR-like regulator AbmF [Schühle et al., 2001]. In A. aromaticum EbN1^T no equivalent gene is present in the vicinity of the abmA gene.

A special pathway for the aerobic degradation of benzoate (Box, benzoate oxidation) was discovered in A. evansii KB740^T by means of differential proteome analysis, revealing the involved proteins to be encoded in an operon-like structure [Gescher et al., 2002]. This box pathway is rather widespread among proteobacteria and has also been studied in Burkholderia xenovorans LB400 [Denef et al., 2005; Bains et al., 2009]. According to the current understanding from mechanistic studies with A. evansii KB740^T, this aerobic degradation pathway involves the following metabolites, reactions, enzymes, and genes (Fig. 22A, B), and proceeds via four steps. (i) Initially, benzoate is activated by an ATP-dependent benzoate-CoA ligase (Bcl2) forming benzoyl-CoA, with the CoA ligase functioning in A. evansii KB740^T under oxic as well as anoxic conditions [Schühle et al., 2003]. It is noteworthy that A. aromaticum EbN1^T apparently employs distinct CoA ligases under these two conditions [Becker et al., 2022a]. (ii) Then, benzoyl-CoA 2,3-epoxidase (BoxBA; synonym: benzoyl-CoA,NADPH:oxygen oxidoreductase (2,3-epoxydizing)) converts benzoyl-CoA into 2,3-epoxybenzoyl-CoA (in tautomeric equilibrium with oxepin, which features a seven-membered ring). The structure of this 2,3-epoxide was inferred in assays with the purified enzyme applying CID-induced MS fragmentation, isotopic resolved MS in the presence of ¹⁸O₂, and ¹³C NMR analysis [Rather et al., 2010]. The [4Fe-4S] cluster-containing BoxA protein is the reducing component, which transfers 2 electrons from NADPH to the epoxidase component BoxB [Rather et al., 2011a]. The di-iron enzyme BoxB upon reduction activates O₂, from which one oxygen atom is then introduced into bound benzoyl-CoA while the other is released as H₂O [Rather et al., 2011b]. The BoxAB enzyme belongs to the enoyl-CoA hydratase/isomerase (crotonase) protein family. (iii) Non-oxygenolytic ring cleavage of 2,3-epoxybenzoyl-CoA is then catalyzed by homodimeric 2,3-epoxy-benzoyl-CoA dihydrolase (BoxC) [Rather et al., 2010]. Regiospecific addition of OH⁻ to ring C2 of the epoxide (or its oxepin tautomer) leads to ring opening. A subsequent second OH⁻ addition at the same C-atom hydrolyzes the ring C2 to release formic acid, which results in the formation of 3,4-dehydroadipyl-CoA semialdehyde. The structure of this product was resolved by ¹³NMR analysis of the respective semicarbazone derivative [Gescher et al., 2005]. It is assumed that the BoxAB and BoxC enzymes are associated in vivo. (iv) 3,4-Dehydroadipyl-CoA semialdehyde is then oxidized by homodimeric, NADP⁺-dependent 3,4-dehydroadipyl-CoA semialdehyde dehydrogenase (BoxD) to cis-3,4-dehydroadipyl-CoA, the structure of which was resolved by interpretation of mass spectrometric fragmentation patterns in conjunction with ¹³NMR analysis [Gescher et al., 2006].

Evidence for the involvement of the 3-oxoadipyl-CoA pathway in the aerobic degradation of benzoate has previously been obtained with A. evansii KB740^T and Bacillus stearothermophilus PK1. Using [ring-¹³C₆]benzoyl-CoA as substrate in cell extracts of both bacterial strains and NMR spectroscopy, Zaar et al. [2001; 2004] were able to identify the following ¹³C-labeled conversion products: cis-3,4-[2,3,4,5,6-¹³C₅]dehydroadipyl-CoA, trans-2,3-[2,3,4,5,6-¹³C₅]dehydroadipyl-CoA, the 3,6-lactone of 3-[2,3,4,5,6-¹³C₅]hydroxyadipyl-CoA, and 3-[2,3,4,5,6-¹³C₅]hydroxyadipyl-CoA. These metabolites were suggested to be arranged in a pathway involving cis-trans isomerization and leading to 3-oxoadipyl-CoA. In the recent study by Becker et al. [2022a], three potential CoA esters forming distinct chromatographic peaks and having the sum formula C₂₇H₄₂N₇O₁₉P₃S were detected in extracts of A. aromaticum EbN1^T grown aerobically with benzoate or 3-(4-hydroxyphenyl)propanoate. The following plausible structural interpretations suggest the involvement of the 3-oxoadipyl-CoA pathway mentioned above: (i) The resolved sum formula (C₂₇H₄₂N₇O₁₉P₃S) corresponds to that of an unsaturated adipyl-CoA. An unambiguous HPLC-MS-based differentiation between cis-3,4-dehydroadipyl-CoA and trans-2,3-dehydroadipyl-CoA is prevented by two opposing chromatographic effects. On the one hand, a double bond in 3,4-position to the CoA ester should lead to a shorter elution time compared to when the double bond is at the 2,3-position. On the other hand, the V-shape of the cis-isomer should increase the retention time compared to the linear shape of the trans isomer. Isomerization of cis-3,4-dehydroadipyl-CoA to trans-2,3-dehydroadipyl-CoA should occur prior to β-oxidation [Ren et al., 2004], since the latter mostly acts on trans-2,3-alkanoyl-CoAs. However, a possible enzyme catalyzing such a cis-trans isomerization, e.g., a Δ³,Δ²-enoyl-CoA isomerase [Yang et al., 1995; Teufel et al., 2010], has not yet been identified in A. aromaticum EbN1^T, or A. evansii KB740^T and B. stearothermophilus PK1. cis-3,4-Didehydroadipyl-CoA was specifically detected in aerobic cultures of A. aromaticum EbN1^T provided with benzoate or 3-(4-hydroxyphenyl)propanoate. trans-2,3-Dehydroadipyl-CoA was also found during growth with phenylalanine [Becker et al., 2022a], where this substance is known to be a degradation intermediate [Teufel et al., 2010]. (ii) The role and origin of 3-hydroxyadipyl-CoA 3,6-lactone is not clear, yet; it might originate from one of the unsaturated adipyl-CoA intermediates [Becker et al., 2022a]. (iii) The 3,6-lactone intermediate was previously assumed to be opened by a lactonase to 3-hydroxyadipyl-CoA [Zaar et al., 2001, 2004; Ren et al., 2004]. Such an enzyme could not be predicted from the genome of A. aromaticum EbN1^T. The EbN1_C15610 gene, which is localized in proximity to the box gene cluster of strain EbN1^T and has an ortholog in the box gene cluster of A. evansii KB740^T, belongs to the α/β-hydrolases, but is predicted to be involved in vitamin K2 (menaquinone) biosynthesis. Alternatively, 3-oxoadipyl-CoA could be formed from 3-hydroxyadipyl-CoA, the hydration product of trans-2,3-dehydroadipyl-CoA. No speci­fic β-oxidation enzymes could be assigned to these two reactions. (iv) As final reaction of the presumptive 3-oxo­adipyl-CoA pathway, a thiolase (BoxE) has been proposed for A. evansii KB740^T [Fuchs et al., 2011] and characterized from Pseudomonas sp. strain B13 [Kaschabek et al., 2002]. Accordingly, a boxE gene is present in the box gene cluster of both A. aromaticum EbN1^T and A. evansii KB740^T. The BoxE protein was identified in A. aromaticum EbN1^T with particularly high score in cells adapted to aerobic growth with benzoate, 3-(4-hydroxyphenyl)propanoate, and phenylalanine [Becker et al., 2022a]. Taken together, the findings described above indicate the aerobic benzoate degradation in A. aromaticum EbN1^T to involve the 3-oxo adipyl-CoA pathway, albeit the reaction sequence from cis-3,4-dehydroadipyl-CoA to 3-oxoadipyl-CoA has to await further enzymatic support.

A solute-binding protein (EbN1_C15620) is encoded directly upstream of the bcl2 gene and is detected under all tested conditions [Becker et al., 2022a]. Since, however, the chromosomal neighborhood of the box gene cluster does not contain genes for an ABC transporter, it remains unclear what system might recruit EbN1_C15620 for the uptake of benzoate. Nevertheless, one may speculate that the ABC transporter encoded next to the genes of the anaerobic benzoyl-CoA pathway may serve this function; it is likewise formed with aromatic substrates during anaerobic as well as aerobic growth [Becker et al., 2022a].

The box gene cluster contains the boxR gene, which is transcribed and translated during aerobic growth with benzoate. Furthermore, boxR transcripts were observed at high abundances also during anaerobic growth with each of the tested aromatic substrates [Becker et al., 2022a]. Note that the homologous BoxR and BzdR proteins are functionally redundant transcriptional repressors, which control the benzoate-dependent transcriptional regulation of the anaerobic and the aerobic benzoyl-CoA pathways, as demonstrated for Aromatoleum sp. strain CIB [Valderrama et al., 2012].

The degradation of cyclohexane carboxylate has been mainly investigated in R. palustris [Küver et al., 1995; Pelletier and Harwood, 2000]. The following metabolites, reactions, enzymes, and genes are involved in the degradation pathway (Fig. 23A, B), which proceeds via five steps. (i) Initially, cyclohexane carboxylate is activated to its CoA ester by an ATP-dependent cyclohexane carboxylate-CoA ligase (AliA). (ii) The cyclohexane carbonyl-CoA formed is then oxidized to cyclohex-1-ene-1-carbonyl-CoA, catalyzed by a FAD-dependent dehydrogenase (AliB). At this point, the degradation pathway continues according to that for anaerobic benzoyl-CoA degradation in R. palustris. (iii) Then, cyclohex-1-ene-1-carbonyl-CoA hydratase (BadK) adds water to the C−C-double bond yielding 2-hydroxy-cyclohex-1-carbonyl-CoA, which should be S-configured according to a recent study on anaerobic degradation of p-cymene in A. aromaticum pCyN1 [Küppers et al., 2019]. (iv) Subsequently, NAD⁺-dependent 2-hydroxy-cyclohexane-1-carbonyl-CoA dehydrogenase (BadH) oxidizes the secondary alcohol forming 2-oxo-cyclohexane-1-carbonyl-CoA. (v) Finally, 2-oxo-cyclohexane-1-carbonyl-CoA hydrolase (BadI) hydrolytically cleaves the alicyclic ring forming pimeloyl-CoA, which is further degraded via conventional β-oxidation.

The genetic context of the ali-bad genes on the chromosome of A. aromaticum EbN1^T does not provide any clue about a possible uptake system for cyclohexane carboxylate. However, given the charged carboxyl group of the compound, it must be assumed that import of cyclohexane carboxylate requires an active uptake system.

The ali-bad genes encoding the enzymes for degradation of cyclohexane carboxylate (chc) are organized in an operon-like structure on the genome of A. aromaticum EbN1^T. This is similar to the gene arrangement known from R. palustris (Fig. 23B). The ali-bad gene cluster also harbors the badR gene coding for a transcriptional repressor of the MarR family [Gupta et al., 2019]. BadR was demonstrated in R. palustris to regulate transcription of the ali-bad gene cluster [Egland and Harwood, 1999] and to abrogate repression upon binding of the degradation intermediate 2-oxo-cyclohexane-1-carbonyl-CoA [Hirakawa et al., 2015].

The degradation of 3-hydroxypropanoate via a reductive route has been elucidated in the anoxygenic phototroph Rhodobacter sphaeroides [Schneider et al., 2012; Asao and Alber, 2013]. The following metabolites, reactions, enzymes, and genes (all present in the genome of A. aromaticum EbN1^T) are involved in the degradation pathway (Fig. 24), which proceeds via three steps. (i) Initially, 3-hydroxypropanoate is activated to its CoA ester by a predicted CoA-ligase (PrpE) homologous to the enzyme from Salmonella typhimurium LT2 [Horswill and Escalante-Semerena, 1999]. (ii) The 3-hydroxypropanoyl-CoA formed is then assumed to be oxidized by elimination of water to acrylyl-CoA; the catalyzing dehydratase in A. aromaticum EbN1^T is assumed to be EbN1_C12250. (iii) Finally, acrylyl-CoA is reduced to propanoyl-CoA by NADPH-dependent acrylyl-CoA reductase (AcuI) belonging to the medium-chain dehydrogenase/reductase (MDR) superfamily [Riveros-Rosas et al., 2003]. Further degradation of propanoyl-CoA can then proceed via the methylmalonyl-CoA pathway to succinyl-CoA and therefrom via the TCA cycle. Acrylate could be channeled into this pathway by a CoA ligase [Schneider et al., 2012]. Evidence for the potential to transform acrylonitrile to acrylate could not be provided from inspecting the genome of A. aromaticum EbN1^T.

No genes for a potential uptake system for 3-hydroxypropanoate and/or acrylate are present in the vicinity of the prpE/acuI genes.

No genes coding for a potential sensory/regulatory system, which could control the expression of the prpE/acuI genes, could be detected in their genomic neighborhood.

The secondary aliphatic alcohols 2-propanol and 2-butanol are metabolized via the same degradation route to acetyl-CoA and propanoyl-CoA, respectively, involving the following three steps (Fig. 25A). (i) A yet to be assigned alcohol dehydrogenase oxidizes the alcohols to the respective ketones, namely 2-propanone (acetone) and 2-butanone. (ii) A subsequent carboxylation at the C1-methyl group yields 2-oxobutanoate and 3-oxopentanoate, respectively. This reaction is catalyzed by an ATP-dependent heterotrimeric (AcxABC) carboxylase that has been intensively studied on the genetic-biochemical [Sluis et al., 1996; Sluis and Ensign, 1997; Sluis et al., 2002; Clark and Ensign, 1999; Boyd et al., 2004; Boyd and Ensign, 2005; Dullius et al., 2011; Schühle and Heider, 2012], proteomic [Wöhlbrand et al., 2007; Oosterkamp et al., 2015] and structural levels [Nocek et al., 2004; Mus et al., 2017] in various bacteria including Alicycliphilus sp., A. aromaticum EbN1^T, Xanthobacter autotrophicus Py2, and Rhodobacter capsulatus B10. (iii) The 3-oxoacids formed could then be activated to the respective CoA esters by the predicted acetate-CoA ligase (AcsA) and/or a succinyl-CoA:3-oxoacid CoA transferase (KctAB), which are both encoded by genes proximal to the acxABC genes. Furthermore, substrate-specific formation of KctAB was previously reported for A. aromaticum EbN1^T [Wöhlbrand et al., 2007]. This activation would allow subsequent thiolytic cleavage, yielding 2 acetyl-CoA per acetone contrasted by 1 acetyl-CoA and 1 propanoyl-CoA per 2-butanone. While terminal oxidation of acetyl-CoA proceeds via the TCA cycle, the methylmalonyl-CoA pathway is required for processing propanoyl-CoA.

No genes for a potential uptake system for the two alcohols and ketones are present in the vicinity of the acx/kct genes. However, it may be assumed that the lipo-philicity of these substrates allows entry into the cell merely by passive diffusion.

The zraR gene is located directly upstream of the acx genes and belongs to the atypical two-component ZraP-SR system. Since this has recently been implicated in various functions including “envelope stress response” [Rome et al., 2018; Choudhary et al., 2020], it is unlikely to control expression of the acx/kct genes. However, the σ⁵⁴-dependent regulator AcxR, which was previously suggested to serve this function [Sluis et al., 2002], is encoded downstream of the kctAB genes.

The short-chained primary alcohols 1-butanol, 1-propanol, and ethanol are most likely degraded via a shared classical route involving oxidation of the hydroxy group (possibly via the aldehyde intermediate) to the respective carboxylate (Fig. 26A) [Kunau et al., 1995; Houten et al., 2016]. The enzymes involved in these degradative pathway have yet to be identified, but are likely recruited from the large pool of alcohol dehydrogenases encoded in the genome of A. aromaticum EbN1^T (Fig. 26B) and belonging to the short- and medium-chain dehydrogenase/reductase (SDR, MDR) superfamilies [Kallberg et al., 2002; Eklund and Ramaswamy, 2008; Jörnvall et al., 2010, 2015]. Further degradation of the carboxylates formed involves activation to the respective CoA esters, followed by conversion via β-oxidation, TCA cycle, and methylmalonyl-CoA pathway (see below).

No genes for a potential uptake system for the three alcohols are present in the vicinity of the adh genes. However, the lipophilicity of these alcohols should facilitate their uptake by passive diffusion.

No genes for a potential sensory/regulatory system are present in the vicinity of the adh genes.

The organic acids butanoate, propanoate, acetate, and C₄-dicarboxylates (succinate, fumarate, and malate) are metabolized via conventional routes (Fig. 27). Following activation to the CoA ester, butanoyl-CoA is split via β-oxidation into two acetyl-CoA moieties. Propanoate is activated by CoA and then processed via the methylmalonyl-CoA pathway (Fig. 36). C₄-dicarboxylates are directly channeled into the canonical TCA cycle (Fig. 35). Acetate degradation is initiated by acetate-CoA ligase forming acetyl-CoA [Starai and Escalante-Semerena, 2004], which is then channeled into the TCA cycle. Notably, A. aromaticum EbN1^T also possesses a phosphotransacetylase and an acetate kinase [Dittrich et al., 2005; Ingram-Smith et al., 2006] that could possibly be used to exploit the energy-rich thioester bond of acetyl-CoA for substrate-level ATP regeneration, yielding acetate as end product.

Uptake of n-butanoate, propanoate, and acetate is assumed to be mediated via proton-driven symporters such as ActP1/2, while that of C₄-dicarboxylates is achieved by a high-affinity TRAP-transporter (DctPQM) [Rabus et al., 1999; Rosa et al., 2018]. These transporters are encoded in the genome of A. aromaticum EbN1^T, albeit not in vicinity of the aforementioned “catabolic” genes.

Extracellular occurrence of C₄-dicarboxylates is re-cognized by specific sensors in various bacteria [Janausch et al., 2002]. In A. aromaticum EbN1^T, the two-component sensory system DctSR is encoded by genes adjacent to the genes encoding the DctPQM uptake systems and known from Rhodobacter spp. to control expression of the dctPQM genes [Hamblin et al., 1993; Sánchez-Ortiz et al., 2021]. Furthermore, benzoate was shown in A. aromaticum EbN1^T to negatively affect C₄-dicarboxylate utilization, apparently by repressing dctSR expression [Trautwein et al., 2012c].

Lactate and pyruvate are degraded via the classical routes, involving lactate dehydrogenase [Jiang et al., 2014] and pyruvate dehydrogenase [Patel et al., 2014], yielding in conjunction acetyl-CoA and CO₂ (Fig. 28). Notably, the genome of A. aromaticum EbN1^T also encodes two sets of proteins (LutA1B1C1 and LutA2B2C2) for an alternative pathway of lactate utilization; in each case, the coding genes are organized in an operon-like structure. Such Lut-systems have been demonstrated in B. subtilis to be involved in lactate utilization and biofilm formation [Chai et al., 2009], in Campylobacter jejuni NCTC 11168 to contribute to lactate utilization [Thomas et al., 2011], and to harbor a novel conserved protein domain (LUD) [Hwang et al., 2013]. Notably, proteomic profiling of chemostat studies with C. jejuni NCTC 11168 revealed elevated abundance of the Lut-system with increasing oxygen concentration (aerobiosis) [Guccione et al., 2017], which could be relevant also for facultative anaerobic A. aromaticum EbN1^T.

No evidence for a known lactate- or pyruvate-specific uptake system could be found during re-annotation of the A. aromaticum EbN1^T genome. In E. coli, lactate/pyruvate uptake was reported to be also achievable by peptide transporters [Hwang et al., 2018], several of such systems are encoded in the genome of A. aromaticum EbN1^T.

The GntR family member LutR is known from B. subtilis to control the expression of the lutABC operon [Chai et al., 2009; Chiu et al., 2014]. While no clear candidate for such a regulator could be predicted from the genome of A. aromaticum EbN1^T, an IclR-family member (EbN1_C23190) is encoded directly next to the lutA2B2C2 operon.

Malonate degradation has been described meanwhile for a variety of bacterial species, e.g., Rhizobium leguminosarum bv. trifolii [An and Kim, 1998; Kim, 2002], R. capsulatus [Dehning and Schink, 1994], R. palustris [Wang et al., 2020], and other proteobacteria [Suvorova et al., 2012]. The degradation pathway involves a malonyl-CoA synthetase (MatB) and malonyl-CoA decarboxylase (MatA), which convert malonate via its CoA ester to acetyl-CoA (Fig. 29). While the genome of A. aromati­cum EbN1^T does not encode a MatB homolog, it harbors a standalone matA gene, which implicates that malonate utilization would require its unspecific activation to the CoA ester by one of the many CoA ligases at the disposal of the strain. At present it is, however, unknown whether A. aromaticum EbN1^T can grow with malonate.

The mat gene cluster in R. trifolii also encodes a potential carrier (MatC) for the uptake of malonate [Lee et al., 2000] and Malomonas rubra employs the Na⁺-driven malonate uptake system MadLM [Schaffitzel et al., 1998]. No candidate for such a system could be predicted from the genome of A. aromaticum EbN1^T. However, one of the numerous TRAP transporters encoded in the genome of A. aromaticum EbN1^T could serve this function, as previously proposed for R. palustris [Wang et al., 2020].

In R. leguminosarum bv. trifolii and other bacteria, the GntR family member MatR was previously reported to control expression of the mat genes [Lee et al., 2000; Suvorova et al., 2012]. No MatR homolog could be predicted from the genome of A. aromaticum EbN1^T; however, directly next to the matB gene, a PAS domain-containing, predicted diguanylate cyclase/phosphodiesterase is encoded.

The analysis of the A. aromaticum EbN1^T genome revealed the presence of several genes whose products are predicted to be involved in formate and hydrogen metabolism (Fig. 30). First, the fdhA2BC genes encode an ortholog of the periplasm-facing, respiratory formate dehydrogenase-N (Fdh-N) known in E. coli to function together with nitrate reductase [Jormakka et al., 2003]. In proximity to the fdhA2BC genes is the fdhD gene, which codes for a sulfurtransferase assumed to be required for active Fdh-N [Arnoux et al., 2015]. Second, the A. aromaticum EbN1^T genome possesses a potential monomeric FdhA1 formate dehydrogenase related to the predicted protein of B. subtilis [Rivolta et al., 1998] and encoded next to the gene for the NADH-binding NuoF component of complex I [Velazquez et al., 2005]. Third, the hyfBCEFGI gene cluster codes for large parts of hydrogen-forming, membrane-localized [NiFe]-hydrogenase HyfA-I of the formate hydrogenlyase (FHL) complex, which is known to disproportionate formate to CO₂ and hydrogen (H₂) and to be evolutionarily linked to complex I from the respiratory chain [Andrews et al., 1997; McDowall et al., 2014; Pinske and Sawers, 2016; Sargent, 2016]. However, the FdhF component responsible for formate oxidation is apparently missing in A. aromaticum EbN1^T. Furthermore, the selABCD genes coding for selenocysteine synthase required for selenocysteine incorporation into Fdh [Böck et al., 1991; Rother et al., 2001; Zorn et al., 2013] are absent in the genome of A. aromaticum EbN1^T. Taken together, despite the described genome-based observations, the potential of A. aromaticum EbN1^T for formate and hydrogen metabolism remains ambiguous at present and requires future research also to assess loss of function or adoption of another role.

In line with the aforementioned ambiguities, no homolog of the well-known formate transporter FocA [Wang et al., 2009] is predicted from the genome of A. aromaticum EbN1^T.

Except for a diguanylate cyclase/phosphodiesterase encoded in the neighborhood of the fdh gene cluster, no other genes for potential sensory/regulatory systems are present in the vicinity of either of the two gene clusters.

Degradation of short-, medium-, and long-chain fatty acids is initiated by CoA ligases followed by the canonical β-oxidation cycle yielding acetyl-CoA moieties (in case of uneven fatty acids also propanoyl-CoA) [Clark and Cronan, 2013] (Fig. 31). The genome of A. aromaticum EbN1^T is rich in genes encoding such enzymes. From the 32 predicted CoA ligases, 19 could be assigned to specific substrates including aforementioned aryl and cyclohexane carboxylates. Moreover, 11 enzymes involved in the β-oxidation cycle are encoded, including acyl-CoA dehydrogenases (e.g., FadE) [Campbell and Cronan, 2002] and multifunctional 3-hydroxyacyl-CoA dehydrogenase/enoyl-CoA hydratase (e.g., FadB) [Spratt et al., 1984].

The uptake of the particularly hydrophobic long-chain fatty acids across the cell envelope of Gram-negative bacteria proceeds in a first step via the β-barrel structured FadL, which forms a hydrophobic channel through the outer membrane [van den Berg et al., 2004; Hearn et al., 2009]. Flipping of the fatty acids across the cytoplasmic membrane is then achieved by the membrane localized acyl-CoA ligase [Azizan et al., 1999]. The genome of A. aromaticum EbN1^T encodes outer membrane localized FadL and a variety of acyl-CoA ligases.

In E. coli, expression of the fad genes is controlled by the transcriptional repressor FadR [Cronan, 2021], which is, however, not encoded in the genome of A. aromaticum EbN1^T.

Degradation of leucine proceeds via a pathway comprising six steps (Fig. 32A). (i) Leucine is initially deaminated by a branched-chain amino acid aminotransferase (IlvE) [Hutson, 2001; Goto et al., 2003] forming 4-methyl-2-oxo-pentanoate. (ii) The latter is then oxidatively decarboxylated by the branched-chain 2-oxoacid dehydrogenase BdkABC [Sokatch et al., 1981; Sykes et al., 1987; Ævarsson et al., 1999] to isovaleryl-CoA, which is also the entry point for isovalerate upon its activation to the CoA ester. (iii) Isovaleryl-CoA is then oxidized to 3-methylcrotonyl-CoA by the dehydrogenase LiuA [Förster-Fromme and Jendrossek, 2008], followed by (iv) ATP-dependent carboxylation to 3-methylglutaconyl-CoA catalyzed by the methylcrotonyl-CoA carboxylase LiuBD [Höschle et al., 2005]. (v) Addition of water to the double bond of 3-methylglutaconyl-CoA by the hydratase LiuC [Bock et al., 2016] furnishes 3-hydroxy-3-methylglutaryl-CoA, which is (vi) finally split into acetoacetate and acetyl-CoA by the lyase LiuE [Chávez-Avilés et al., 2009]. Notably, most of the liu genes of A. aromaticum EbN1^T are organized in an operon-like structure (Fig. 32B, C) as previously described for Pseudomonas aeruginosa PAO1 [Förster-Fromme et al., 2006] and the described pathway is absent in E. coli [Reizer, 2013].

Leucine is probably imported by the branched-chain amino acid ABC transporter Bra, which is encoded by the braGFEDC1C2 gene cluster and belongs to the HAAT transporter family [Hoshino and Kose, 1990; Saier, 2000]. Two further similar ABC transporters (EbN1_C15320/40/50/620 and EbN1_C20250-80) are encoded in the genome of A. aromaticum EbN1^T.

The leucine-responsive regulatory protein Lrp is a well-studied global regulator of metabolism in E. coli and known to control expression of the Liv ABC transporter for branched-chain amino acids including isoleucine [Calvo and Matthews, 1994], which might also be the case in A. aromaticum EbN1^T. The transcriptional activator BkdR was shown in Ps. putida to control expression of the bkd operon encoding aforementioned branched-chain 2-oxoacid dehydrogenase [Madhusudhan et al., 1997]. A homolog of BkdR is encoded in the genome of A. aromaticum EbN1^T. The LiuR regulator known from Ps. aeruginosa [Kazakov et al., 2009] is apparently not encoded in the genome of A. aromaticum EbN1^T.

Degradation of histidine in A. aromaticum EbN1^T essentially follows the hut (histidine utilization) pathway previously reported for Ps. putida and Ps. aeruginosa, involving a series of four hydrolytic reactions (Fig. 33A). (i) Initially, histidine ammonia lyase (HutH) non-oxidatively eliminates the α-amino group from histidine whereby trans-uroconate is formed [Consevage and Phillips, 1990; Schwede et al., 1999]. (ii) The latter is then converted to 4-imidazolone-5-propanoate via addition of water to the double bond, catalyzed by NAD⁺-dependent urocanase (HutU) [Klepp et al., 1990; Lenz and Rétey, 1993; Kessler et al., 2004]. (iii) Then, the zinc-dependent imidazolonepropionase (HutI) hydrolyzes via a nucleophilic attack the C−N-bond of the carbonyl carbon of 4-imidazolone-5-propanoate yielding N-formimidoyl-L-glutamate [Yu et al., 2006; Tyagi et al., 2008]. (iv) In the final reaction, the latter is hydrolytically cleaved to L-glutamate and formamide by formiminoglutamate hydrolase (HutG) [Ouzounis and Kyrpides, 1994]. The genes for all aforementioned enzymes are present in the genome of A. aromaticum EbN1^T. By contrast, the alternative enzyme N-formimino-L-glutamate iminohydrolase (HutF), which deaminates N-formimidoyl-L-glutamate to N-formyl-L-glutamate and ammonia [Fedorov et al., 2015], is not encoded in the genome of A. aromaticum EbN1^T.

In Ps. putida, the secondary transporter HutT is responsible for histidine uptake [Wirtz et al., 2021]. Such a system could not be predicted from the genome of A. aromaticum EbN1^T, thus, histidine uptake may be performed by one of several available ABC transporters for amino acids.

In Pseudomonas spp. transcriptional regulation of the hut operon is controlled by the repressor HutC [Hu et al., 1989; Naren and Zhang, 2020]. No evidence for a hutC homolog could be detected in the genome of A. aromaticum EbN1^T.

Degradation of proline has been elucidated in E. coli and S. typhimurium and shown to involve the peripheral membrane protein proline dehydrogenase PutA performing two consecutive reactions [Surber and Maloy, 1998; Vinod et al., 2002] (Fig. 34A). (i) In the initial reaction, the multifunctional flavoprotein PutA acts as proline dehydrogenase, oxidizing L-proline to (S)-1-pyrroline-5-carboxylate with the released electrons delivered via FAD to the quinone pool. (S)-1-Pyrroline-5-carboxy­late is then assumed to spontaneously hydrolyze to L-glutamate-5-semialdehyde. (ii) The latter is then oxidized to L-glutamate by PutA with NAD⁺ serving as cofactor. The genome of A. aromaticum EbN1^T habors a single putA gene (Fig. 34B, C).

Uptake of proline is performed by the sodium-driven transporter PutP [Olkhova et al., 2011], which in E. coli and S. typhimurium is encoded directly next to the putA gene. Notably, in A. aromaticum EbN1^T, the putA and putP genes are localized at very distant positions in the genome.

Remarkably, the PutA protein additionally functions as a redox-dependent transcriptional repressor, controlling in its active form induction of the put operon in the presence of proline in S. typhimurium [Ostrovsky de Spicer and Maloy, 1993; Muro-Pastor and Maloy, 1995].

The canonical TCA cycle [Cronan and Laporte, 2013] serves two major functions: First, acetyl-CoA generated from the multitude of aforementioned degradation pathways is completely oxidized to CO₂, accompanied by the provision of reduced electron carriers for respiratory energy conservation. Second, precursors for monomer biosynthesis (2-oxoglutarate and oxaloacetate) are generated. The genes encoding all enzymes of the TCA cycle are present in, albeit scattered across, the genome of A. aromaticum EbN1^T (Fig. 35). The TCA cycle is constituted by the following eight enzymes: (i) Citrate synthase (GltA) catalyzes the Claisen condensation (C−C-bond formation) between acetyl-CoA and oxaloacetate forming the citryl-CoA intermediate, which is subsequently hydrolyzed to the tricarboxylic acid citrate and coenzyme A [Wiegand and Remington, 1986; Karpusas et al., 1990; Aleksandrov et al., 2014]. (ii) Aconitase (AcnB) is an FeS-enzyme that catalyzes the dehydration of citrate to cis-aconitate followed by the rehydration to isocitrate [Robbins and Stout, 1989; Lloyd et al., 1999]. (iii) Isocitrate dehydrogenase (Icd1) catalyzes, after a random binding of cosubstrates, the oxidation (with NADP⁺) of isocitrate to the enzyme-bound intermediate oxalosuccinate, which is subsequently decarboxylated to 2-oxoglutarate [Bolduc et al., 1995]. (iv) 2-Oxoglutarate dehydrogenase (SuccAB/Lpd) oxidatively decarboxylates 2-oxoglutarate to succinyl-CoA (NADH-forming). The tripartite enzyme complex is closely related to pyruvate dehydrogenase [de Kok et al., 1998], with E1 (SucA, succinyl-transferring) and E3 (Lpd, dihydrolipoyl dehydrogenase) flexibly tethered to the E2 core (SucB, 2-oxoglutarate dehydrogenase) [Murphy and Jensen, 2005]. (v) Heterodimeric succinyl-CoA ligase (SucCD) converts succinyl-CoA via succinyl-phosphate to succinate, representing the sole substrate-level phosphorylation reaction of the TCA cycle [Wolodko et al., 1994]. (vi) Succinate dehydrogenase (SdhAB2C2D2) oxidizes succinate to fumarate concomitantly feeding the released electrons into the membrane-embedded quinone pool (see section “Respiratory Energy Conservation” below). (vii) Fumarase (Fum) catalyzes the reversible hydration of fumarate via an aci-carboxylate intermediate to S-malate [Weaver et al., 1995; Weaver, 2005]. (viii) Finally, NAD⁺-dependent malate dehydrogenase (Mdh) oxidizes malate to oxaloacetate, completing the TCA cycle; notably, Mdh has a markedly stricter substrate range than functionally related lactate dehydrogenase [Hall and Banaszak, 1993; Boernke et al., 1995].

The methylmalonyl-CoA pathway is required for the further degradation of propanoyl-CoA generated during degradation of uneven fatty acids, 1-propanol, and propanoate [Halarnkar and Blomquist, 1989]. The pathway proceeds via three steps (Fig. 36A): (i) In the initial reaction, propanoyl-CoA is carboxylated to (2S)-methylmalonyl-CoA by the ATP- and biotin-dependent, heteromeric propanoyl-CoA carboxylase (PccAB). The α-subunit of the enzyme harbors the domains for biotin carboxylase and biotin carboxyl carrier protein, while the β-subunit provides the activity for the carboxyl transfer [Huang et al., 2010]. (ii) In the subsequent reaction, (2S)-methylmalonyl-CoA is epimerized by EbN1_C21340 to its (2R)-form. (iii) The latter is finally converted to succinyl-CoA via a C-skeleton rearrangement catalyzed by B₁₂-dependent methylmalonyl-CoA mutase (Sbm) [Ludwig and Matthews, 1997]. The coding genes for all three enzymes of this pathway are present in the genome of A. aromaticum EbN1^T (Fig. 36B, C).

As a facultative anaerobic bacterium, A. aromaticum EbN1^T can perform respiratory energy conservation either under anoxic conditions via denitrification or under oxic conditions via oxygen (O₂) respiration. Electrons are channeled from the NADH pool into the respiratory electron transport chain via complex I, which in concert with complex II reduces the membrane-embedded quinone pool, while complex III subsequently reduces the periplasmic cytochrome c pool. Reduced quinone and cytochrome c pools then deliver electrons for the terminal reduction of NO₃⁻ to N₂ (via the denitrification apparatus) or of O₂ to H₂O (via complex IV). In both cases, transmembrane proton motive force (pmf) is generated, which in turn is exploited by the F₁F₀-type ATP synthase for regeneration of ATP [Price and Driessen, 2010; Unden and Dünnwald, 2013; Vercellino and Sazanov, 2022]. Considering beyond that the presence of ETF:Q OR, Rnf complex, and tetrathionate reductase, the respiratory electron transfer capacities of A. aromaticum EbN1^T are even broader.

The respiratory complex I represents the main entry point of respiratory energy conservation. By regenerating NAD⁺, complex I receives electrons at a low redox potential (E° = −320 mV), which it transports to ubiquinone coupled to energy conservation by means of proton pumping. NADH:ubiquinone oxidoreductase is a large multi-subunit membrane protein complex encoded by the nuoA−N genes, organized in a compact cluster in the genome of A. aromaticum EbN1^T, as known from other bacteria [Friedrich, 1998]. Complex I represents the largest protein complex of the respiratory chain and harbors an intricate electron transfer path from flavin mononucleotide via seven FeS clusters to the binding site of ubiquinone [Efremov and Sazanov, 2011; Berrisford et al., 2016]. Recent cryo-EM analysis suggests that complex I of E. coli conducts an intriguing coupling between the proton uptake into the Q-cavity (NuoH) and the proton translocations to the periplasm via NuoL, NuoM, and NuoN [Kolata and Efremov, 2021]. In addition to complex I, A. aromaticum EbN1^T also possesses a non-proton pumping type II NADH dehydrogenase (Ndh2, encoded by ndh), which is known from E. coli to be localized as homodimer to the cytoplasmic face of the membrane and to directly feed electrons into the ubiquinone pool [Heikal et al., 2014; Salewski et al., 2016]. E. coli shifts between complex I and Ndh2, which differ in efficiency and rate, to optimize growth. This shift is mediated by the global regulator FNR, which represses the expression of the ndh gene in response to anoxic conditions [Green and Guest, 1994; Meng et al., 1997; Unden et al., 2002; Friedrich and Pohl, 2013]. For facultative anaerobic A. aromaticum EbN1^T, such a scenario may also be envisioned.

Succinate dehydrogenase (synonym: succinate: quinone oxidoreductase [SQR]) and electron transfer flavoprotein:quinone oxidoreductase (ETF:Q OR) both feed electrons from catabolic oxidation reactions directly into the transmembrane electron transport chain at the level of ubiquinone. Succinate dehydrogenase is a multisubunit membrane associated enzyme and part of the canonical TCA cycle. It transfers the electrons liberated by oxidizing succinate to fumarate via its FAD and 3 FeS clusters to ubiquinone. Biogenesis of the flavinylated catalytic SdhA subunit requires the chaperone SdhE. The FeS-protein SdhB establishes the interaction between SdhA and the membrane proteins SdhCD that connect to the electron acceptor ubiquinone [Cecchini et al., 2002; Ruprecht et al., 2009; Maher et al., 2018]. In A. aromaticum EbN1^T, next to the formed SdhAB2C2D2 encoded in an operon-like structure, also a 2^nd gene set (sdhB1C1D1) is present in the genome, albeit not expressed under the hitherto tested conditions.

The electron transfer flavoprotein of ETF:Q OR generally serves as primary electron acceptor for acyl-CoA dehydrogenases (β-oxidation of fatty acids) and catabolism of some amino acids in bacteria and mitochondria [Ghisla and Thorpe, 2004]. ETF has recently also been shown in A. aromaticum EbN1^T to receive electrons from (R)-2-benzylsuccinyl-CoA dehydrogenase (BbsG) involved in the anaerobic degradation of toluene (Fig. 3) [Vogt et al., 2019]. The ETF:Q OR from human mitochondria was shown by crystal structural analysis to represent a single integral membrane protein with binding domains for FAD, the [4Fe-4S] cluster, and ubiquinone [Zhang et al., 2006]. Recent comparative genome analysis focusing on the ETF:Q OR family revealed that its members are functionally diverse and widespread, and that electron bifurcating members are characterized by the presence of 2 instead of only a single FAD [Garcia Costas et al., 2017]. The ETF:Q OR encoded in the genome of A. aromaticum EbN1^T apparently carries 1 FAD (according to homology and domain structure) and should therefore belong to the non-bifurcating members of this interesting protein family.

The membrane integral complex III (PetABC) transfers electrons from the reduced quinone-pool to the periplasmic cytochrome c (Fig. 37A). Concomitantly, protons are pumped via the Q-cycle across the cytoplasmic membrane, thereby contributing to the generation of pmf. Complex III has a dimeric structure, with each monomer consisting of the following three subunits: PetA represents the Rieske FeS protein with a [2Fe-2S] cluster, PetB the cyctochrome b harboring hemes b_L and b_H, and PetC the cytochrome c with heme c₁ as cofactor. Structural studies of the Q-cycle mechanism suggest electron bifurcation from ubiquinol (QH₂) into an inner membrane low potential chain via hemes b_L and b_H back to ubiquinone (Q) and a high potential chain via the [2Fe-2S] cluster and cytochrome c₁ to the periplasmic cytochrome c [Hunte et al., 2003; Xia et al., 2013; Esser et al., 2019]. Complex III is suggested to form a supercomplex with complex IV [Steimle et al., 2021]. The petABC genes form an operon-like structure in the genome of A. aromaticum EbN1^T and their protein products could be detected (Fig. 37B, C).

The genome of A. aromaticum EbN1^T encodes two variants of an ion-pumping reversible NADH:ferredoxin oxidoreductase termed Rnf complex (for Rhodobacter nitrogen fixation) (Fig. 38). The Rnf complex was first described as energy coupling oxidoreductase in conjunction with N₂-fixation in R. capsulatus [Schmehl et al., 1993; Kumagai et al., 1997]. The Rnf complex pumps Na⁺ and possibly in some cases also H⁺, conserving the energy of excess reduced ferredoxins produced during anaerobic conversion reactions [Biegel et al., 2011; Buckel and Thauer, 2018]. The topology of the Rnf complex from Vibrio cholerae [Hreha et al., 2015] and Clostridium tetanomorphum [Vitt et al., 2022] revealed the electron transfer path from ferredoxin-binding cytoplasmic RnfB via transmembrane RnfAED and periplasmic RnfG to NAD-binding cytoplasmic RnfC.

Denitrification viz. the dissimilatory reduction of nitrate (NO₃⁻) via nitrite (NO₂⁻), nitric oxide (NO), and nitrous oxide (N₂O) to molecular nitrogen (N₂) is long studied, was comprehensively reviewed by Zumft [1997], and more recently suggested to be performed by an enzyme supercomplex termed the nitrate respirasome [Borrero-de Acuña et al., 2016]. Enzymes and genes associated with denitrification in A. aromaticum EbN1^T are depicted in Figure 39.

Since A. aromaticum EbN1^T does not possess genes for periplasmic (napABC) [Stewart et al., 2002], but rather those for a cytoplasm-facing, membrane-localized nitrate reductase (narGHI) reducing NO₃⁻ to NO₂⁻ [Cole and Richardson, 2013], an efficient uptake system for nitrate is required for functional denitrification. Such systems are encoded in the genome of A. aromaticum EbN1^T by the narK1 and narK2 genes, belonging to the MFS superfamily of transporters [Pao et al., 1998]. Previous studies, which were based on the genomic context of their encoding genes, as well as on biochemical and physiological evidence, indicate that these two transporters serve different physiological functions [Wood et al., 2002; Sharma et al., 2006; Goddard et al., 2017]. NarK1 is suggested to represent a NO₃⁻/H⁺-symporter for nitrate uptake in the context of assimilatory reduction of NO₃⁻ to NH₄⁺, where all involved proteins are localized in the cytoplasm and the product NH₄⁺ is widely used for anabolic purposes. On the other hand, NarK2 is suggested to function as a NO₃⁻/NO₂⁻-antiporter associated with dissimilatory nitrate reduction, where NO₂⁻ formed by NarGHI in the cytoplasm has to be exported into the periplasm, where it is reduced to N₂ by the concerted activity of nitrite, nitric oxide, and nitrous oxide reductases (see paragraphs below). The intriguing mechanism of NO₃⁻/NO₂⁻-exchange by the NarK2 antiporter was elucidated by structural and mechanistic studies: (i) The five transmembrane α-helices create two domains above and below a rotation axis in the membrane. Conformational change of these domains establishes alternating access (“rocker-switch”) for the two substrates. (ii) The channel for substrate translocation is positively charged while lacking protonable residues. (iii) Differentiation between NO₃⁻ and NO₂⁻ is achieved by compound-specific hydrogen bond networks [Andrade and Einsle, 2013; Yan et al., 2013; Zheng et al., 2013; Fukuda et al., 2015].

The narK1K2 genes form an operon-like structure together with the narGHJI genes encoding dissimilatory nitrate reductase. In Ps. aeruginosa, transcriptional control of the narK1K2GHJI operon in response to anoxia and presence of nitrate was shown to be excerted at multiple levels by the oxygen regulator Anr, the N-oxide regulator Dnr, and the nitrate-sensing two-component system NarXL [Kolesnikow et al., 1992; Bauer et al., 1999; Härtig et al., 1999; Vollack et al., 1999; Schreiber et al., 2007]. Considering the similar gene organization and availability of regulators, a similar regulatory scenario can be envisioned for A. aromaticum EbN1^T.

The membrane bound dissimilatory nitrate reductase NarGHI of A. aromaticum EbN1^T is encoded in a single operon-like structure (together with NarK1K2, see above paragraph) as previously described for other denitrifying bacteria [Blasco et al., 1990; Berks et al., 1995; Philippot et al., 1997]. Nitrate reductase couples the oxidation of the reduced quinone pool (QH₂) to the two-electron reduction of NO₃⁻ and NO₂⁻ with concomitant export of protons whereby it contributes to the generation of pmf. The intricate electron transfer path from ubiquinol to NO₃⁻ is assumed to proceed via the following three major steps [Bertero et al., 2003, 2005; Jormakka et al., 2004; Fedor et al., 2014a, b; Coelho and Romão, 2015]: (i) The membrane anchored subunit NarI provides the periplasm oriented Q-site (Q_D) for binding and stepwise oxidation of ubiquinol. While the two electrons from QH₂ oxidation are consecutively accepted by distal heme b_D, the two dissociated protons are likewise consecutively released to the periplasm. Electron relay in NarI then proceeds from heme b_D to the proximal heme b_P and therefrom to the FS4 cluster of NarH. (ii) The electron transfer subunit NarH is sandwiched between NarI and NarG. It harbors one [3Fe-4S] cluster (FS4) and three [4Fe-4S] clusters (FS1−3), which together establish the connecting conduit for transferring the electrons from heme b_P of NarI to the final [4Fe-4S] cluster (FS0) of NarG. (iii) The catalytic subunit NarG carries the FS0 cluster that delivers electrons to the Mo-bisMGD, where reduction of NO₃⁻ to NO₂⁻ finally occurs. The chaperon NarJ is also encoded in the narGHJI gene cluster and is required for the molybdenum cofactor assembly in the catalytic subunit NarG [Blasco et al., 1998]; NarJ is specific for its protein substrate and belongs to the NarJ family of redox enzyme maturation proteins [Chan et al., 2014; Bay et al., 2015]. Transcriptional regulation of the narGHJI genes is assumed to involve the same complex sensory/regulatory circuits as described above for narK1K2.

Denitrifying bacteria perform the single-electron reduction of NO₂⁻ to NO by cytochrome cd₁- (NirS) or copper-containing (NirK) dissimilatory nitrite reductases [Jüngst et al., 1991; Ye et al., 1993] of which A. aro-maticum EbN1^T possesses only the former (single nirS gene). Also members of the related genus Thauera harbor nirS genes [Song and Ward, 2003]. In accord with its periplasmic location, NirS carries a Sec-leader peptide. Structural and mechanistic analysis of homodimeric cd₁ nitrite reductases from Ps. aeruginosa revealed the following electron path from periplasmic reduced cytochrome c to NO₂⁻: (i) The heme c carrying domain of NirS serves as electron entry point. (ii) Electron transfer occurs via an intriguing intramolecular path from reduced heme c to oxidized heme d₁. (iii) The catalytic, reduced heme d₁ has a high affinity for NO₂⁻ and its reduction product NO dissociates exceptionally fast [Wherland et al., 2005; Rinaldo et al., 2011; Klünemann and Blankenfeldt, 2020]. In Ps. aeruginosa, expression of the nirS gene was shown to occur only under low oxygen or anoxic conditions, mediated by the NO-sensing, FNR-type regulator DNR (dissimilatory nitrate respiration regulator) [Kuroki et al., 2014].

The membrane integral, heterodimeric nitric oxide reductase (NorBC) performs the intriguing two-electron reduction of NO via hyponitrite to N₂O (2 NO + 2 e⁻ → ⁻ON = NO⁻; ⁻ON = NO⁻ + 2 H⁺ → N₂O + H₂O) and is encoded by the colocalizing norBC genes [Zumft et al., 1994], which is also the case for A. aromaticum EbN1^T. Depending on the type of primary electron donor used, three types of NOR are distinguished [Hino et al., 2012]: cNOR receives electrons from reduced periplasmic cyto-chrome c, qNOR interacts directly with QH₂ for electron transfer, and qCu_ANOR can be replenished by both electron donors. According to highest sequence similarities with well-studied cNorBCs from Ps. stutzeri [Zumft et al., 1994] and Ps. aeruginosa [Hino et al., 2010], the NorCB of A. aromaticum EbN1^T can be classified as a cNOR-type. Based on structural and mechanistic studies, the following electron path is currently assumed for cNOR [Flock et al., 2006; Hino et al., 2010, 2012; Shiro, 2012; Shiro et al., 2012]. The heme c cofactor at the periplasmic face of NorC receives an electron from reduced periplasmic cytochrome c₅₅₁ and delivers it across the NorCB interface to the heme b cofactor of NorB. From there, the electron is transferred to heme b₃, which together with the non-heme Fe_B constitutes the binuclear center of the catalytic site for NO reduction in the NorB subunit. While cNOR is phylogenetically closely related to oxygen-reducing ccb₃ oxidase, it is by contrast non-electrogenic and the protons for conversion of hyponitrite to N₂O are obtained from the periplasm. The NorQ and NorD accessory proteins, which are required for the insertion of the non-heme Fe_B cofactor into the NorB subunit [Kahle et al., 2018], are encoded by genes located in proximity to the norCB genes. The nor operon also contains the dnr gene coding for the NO-responsive Dnr regulator, which controls expression of this operon in Ps. stutzeri [Vollack and Zumft, 2001] and Dinoroseobacter shibae [Ebert et al., 2017]. Such a scenario can also be envisoned for A. aromaticum EbN1^T.

The final step of denitrification is catalyzed by the soluble periplasmic nitrous oxide reductase (NosZ), which converts N₂O in a two-electron reduction to N₂ (N₂O + 2 H⁺ + 2 e⁻ → N₂ + H₂O) [Zumft, 1997]. The enzymatic and crystal structural analysis of NosZ from various Pseudomonas and Paracoccus spp. yielded the following current understanding of the catalytic mechanism [Brown et al., 2000; Haltia et al., 2003; Rasmussen et al., 2005; Pomowski et al., 2011; Carreira et al., 2017]. The multicopper enzyme NosZ contains a mixed-valent Cu_A site (receiving electrons from periplasmic cytochrome c₅₅₁ or transmembrane NosR) and a tetranuclear Cu_Z center. The two subunits of homodimeric NosZ are head-to-tail arranged such that the electron storage Cu_A site from one subunit gets close enough to the Cu_Z center of the other subunit for electron transfer. After entering the active site through a hydrophobic channel, the N₂O substrate is bound in a side-on manner to the [4Cu:2S] (or [4Cu:1S]) cluster of the Cu_Z (or Cu_Z*) centre and reduced via two consecutive one-electron steps. While the product N₂ is released via the hydrophobic channel, the by-product H₂O remains in the hydrophilic part of the active site. Maturation of the NosZ and NosR proteins is a multi-facetted process. Trafficking of copper in the periplasm from the copper chaperone NosL via NosD to NosZ is facilitated by ATP-energized (NosF) conformational changes of the membrane spanning NosDFY maturation machinery [Zhang et al., 2019; Prasser et al., 2021; Müller et al., 2022]. The transmembrane FeS flavoprotein NosR is maturated by the flavinyl transferase ApbE and is assumed to transfer electrons from the reduced quinone pool to NosZ [Wunsch and Zumft, 2005; Zhang et al., 2017, 2019]. In A. aromaticum EbN1^T, the nosZRDFYL genes colocalize in an operon-like structure in proximity to the nar gene cluster (Fig. 39B). Expression of the nos gene cluster may be regulated in response to O₂ and NO by FNR-type regulators as previously reported for P. denitrificans [Bergaust et al., 2012]. In the case of A. aromaticum EbN1^T, semi-inhibitory concentrations of toluene and ethylbenzene, respectively, cause an impairment of denitrification. Concomitantly the flux of acetyl-CoA is re-routed from the TCA cycle toward increasing formation of poly(3-hydroxybutyrate), presumptively as a means to keep alkylbenzene degradation operative despite impaired denitrification [Trautwein et al., 2008; Rabus et al., 2014].

The genome of A. aromaticum EbN1^T contains the tsdAB genes potentially encoding a bifunctional thiosulfate dehydrogenase/tetrathionate reductase belonging to the widespread family of diheme c-type cytochromes (Fig. 40) [Denkmann et al., 2012; Kurth et al., 2016]. The tsdAB genes in A. aromaticum EbN1^T are framed by genes encoding conserved proteins of unknown function. The TsdA(B) enzyme is assumed to oxidatively couple 2 molecules of thiosulfate (S₂O₃²⁻) in the periplasm to tetrathionate (S₄O₆²⁻) according to the following reaction: 2 S−SO₃²⁻ → ⁻O₃S−S−S−SO₃⁻ + 2 e⁻. The released electrons should then be channeled, e.g., via cytochrome c, into the electron transfer of oxygen respiration, presumably at the level of cytochrome cbb₃ (see below) [Kurth et al., 2016]. An intriguing mechanism has been proposed for TsdA from C. jejuni, wherein the dissociation of a heme ligand allows the catalytic Cys-residue to react with the substrate as well as the nearby heme [Jenner et al., 2019, 2022]. The potential of A. aromaticum EbN1^T to utilize thiosulfate as electron donor for aerobic or anaerobic growth has not been investigated so far.

Moreover, the genome of A. aromaticum EbN1^T harbors the complete ttrABC gene cluster coding for a tetrathionate reductase, which is neighbored by the ttrRS genes encoding an associated two-component sensory/regulatory system (Fig. 40), as previously reported for S. typhimurium [Hensel et al., 1999; Hinsley and Berks, 2002; James et al., 2013]. The molybdenum-dependent enzyme [Hinojosa-Leon et al., 1986] catalyzes the reduction to tetrathionate to thiosulfate (⁻O₃S−S−S−SO₃⁻ + 2 e⁻ → 2 S−SO₃²⁻) in the periplasm with required electrons delivered from the quinol-pool. Export of the TrtAB enzyme via the Tat-pathway [Frain et al., 2019] is supported by the presence of a canonical twin-arginine motive containing signal peptide in TtrA and an atypical one in TtrB (one arginine missing) as well as by the interdependency of TtrAB during export [Hinsley et al., 2001; James et al., 2013]. The functions of the three subunits of tetrathionate reductase are follows: TtrA represents the catalytic subunit of the reductase, carrying a Mo-bisMGD cofactor and a FeS cluster; TtrB employs four FeS clusters for electron transfer; and transmembrane TtrC is assumed to serve as anchor as well as quinol dehydrogenase. Notably, TtrC belongs to the so-called NrfD-like subunits, which were recently also proposed to possess a proton-conducting pathway [Calisto and Pereira, 2021]. Expression of the ttrABC operon is controlled by the TtrRS system in response to the presence of tetrathionate and by the global regulator Fnr under anoxic conditions [Price-Carter et al., 2001].

In accord with the changing oxygen concentrations A. aromaticum EbN1^T is very likely encountering in its natural environment, the genome of this bacterium encodes for a rather large number of alternative terminal oxidases (complex IV, quinol/cytochrome c:O₂ oxidoreductase) that should differ by their affinities toward O₂ (Cox for high pO₂ vs. Cco and Cyd for low pO₂) as known from e.g., Ps. aeruginosa [Arai et al., 2014] and that belong to the superfamily of cytochrome c oxidases [Musser and Chan, 1998]. Their common function is to couple the oxi­dation of an electron donor (reduced cytochrome c or ubiquinol) to the four-electron reduction of O₂ to water, while concomitantly contributing by proton-pumping to the generation of pmf (Fig. 41). Evidence for the presence of a supercomplex between complexes III and IV, which facilitates electron transfer possibly by improved diffusion of cytochrome c, was initially provided on the basis of copurification [Berry and Trumpower, 1985] and blue-native PAGE [Schägger and Pfeiffer, 2000], and more recently substantiated by means of cryo-EM [Sun et al., 2018; Rathore et al., 2019; Moe et al., 2021].

The genome of A. aromaticum EbN1^T encodes in total four cox gene clusters for these heme-copper quinol oxidases, one of which (coxB3A2GC) contains more than the coxAB genes encoding the core enzyme (subunits I and II). The structure and mechanism particularly of the latter have been studied in various organisms [Iwata et al., 1995; Ostermeier et al., 1997; Lyons et al., 2012; Xu et al., 2020], revealing the following insights into the core enzyme. Subunit I (CoxA) carries a low spin heme a and a high-spin binuclear center (heme a₃−Cu_B or heme bo₃−Cu_B); at heme a₃ or bo₃ oxygen is bound and reduced. In subunit II (CoxB), the binuclear Cu_A center serves as primary acceptors of electrons delivered by reduced cytochrome c. The enzyme possesses two proton transfer pathways, one for the reduction of O₂ to H₂O in subunit I and the other for proton pumping through subunit II (up to 4 per O₂ reduced). While expression of the cox genes is apparently not controlled by Fnr [van Spanning et al., 1997; Otten et al., 2001], it is positively correlated with starvation conditions [Osamura et al., 2017].

The genome of A. aromaticum EbN1^T harbors a single operon-like structure (ccoN1O1Q1P1G1H1) and a single ccoN gene coding for a cbb₃ oxidase. Cytochrome cbb₃ oxidase is predominantly found in proteobacteria [Pitcher et al., 2002] and can achieve a proton-pumping efficiency of 1 H⁺/e⁻ [Rauhamäki et al., 2012]. The electron path proceeds from reduced soluble cytochrome c through the multisubunit CcoNOP via three heme c cofactors to heme b and from there to heme b₃ and Cu_B; the latter two provide the electrons for final reduction of O₂, which is coupled to proton pumping [Buschmann et al., 2010]. Cytochrome c mediated electron transfer between complex III and cbb₃ oxidase (complex IV) was recently shown by cryo-EM to involve super-complex formation between the two complexes [Steimle et al., 2021]. The high O₂-affinity of cbb₃ oxidase from Pseudomonas stutzeri indicates this oxidase to play a central role under microaerophilic conditions [Pitcher et al., 2002; Xie et al., 2014].

The genome of A. aromaticum EbN1^T contains a single set of cydAB genes coding for a cytochrome bd-type quinol oxidase. CydAB has a high affinity toward O₂ and its genes are therefore expressed at low pO₂, i.e., under O₂-limiting conditions [Govantes et al., 2000; Alexeeva et al., 2002]. Structural studies [Borisov et al., 2011; Safarian et al., 2016; Forte et al., 2017; Grund et al., 2021] showed that CydAB harbors three heme cofactors (b₅₅₈, b₅₉₅, and d) tightly aligned for direct electron relay from QH₂ to O₂. While oxygen diffuses freely to heme d, the protons are transferred there from the cytoplasm through a dedicated path in the protein. In contrast to heme-copper oxygen reductases, CydAB is not a proton pump.

The F₁F₀-ATP synthase is a unique macromolecular machine that transforms the electrochemical energy (pmf) via mechanical energy (subunit rotation) into chemical energy (ADP + P_i → ATP) and has been intensively studied over the past decades [Mao and Weber, 2007; Weber, 2006, 2010; Watanabe et al., 2010; Kühlbrandt, 2019; Sobti et al., 2019; Guo and Rubinstein, 2022]. The enzyme of E. coli consists of the membrane embedded F₀ subcomplex (ab₂c₁₀ composition) and the peripheral F₁ subcomplex (α₃β₃γδε composition). All of these components are encoded in the genome of A. aromaticum EbN1^T (Fig. 42). The principle mechanism of a F₁F₀-ATP synthase is as follows. Pmf energized translocation of protons through a channel between a and c subunits of the F₀ subcomplex drive the rotation of the c₁₀ ring. The latter forms the rotor together with the γ and ε subunits of the F₁ subcomplex. The γ subunit of the rotor penetrates the F₁ cylinder formed by alternating α and β subunits. Rotation of the γ subunit translates into 120° stepwise rotations of this cylinder. Thereby torque is generated, which induces conformational changes in the three catalytic β-subunits yielding the conversion of bound ADP and P_i to ATP. According to the “tri-site” mechanism one catalytic site binds substrate, the second performs the catalysis, and the third releases the product. The peripheral stalk composed of the b₂δ-subunits holds together the F₁ and F₂ subcomplexes during the rotational movements while ensuring elasticity at the same time.

We are grateful to many colleagues for their contributions to the field of anaerobic degradation of aromatic compounds.

The authors have no conflicts of interest to declare.

This study was supported by the Deutsche Forschungsgemeinschaft within the framework of the research training group Chemical Bond Activation (GRK 2226).

R.R. conceived the study; P.B., D.W., M.N.-S., D.S., and R.R. conducted reannotation of the genome; L.W. compiled proteomic data; R.R. wrote the manuscript with input from all authors; and P.B. prepared the figures. All authors have agreed to the final version of the manuscript.