A Unique Homo-Hexameric Structure of 2-Aminomuconate Deaminase in the Bacterium Pseudomonas species AP–3

The bacterium Pseudomonas species sp. AP–3 is one of several microorganisms that are capable of using 2-aminophenol as its sole source of carbon, nitrogen and energy. Several 2-aminophenol-metabolizing enzymes have pivotal roles in the biodegradation of aniline and its derivatives as environmental pollutants in Pseudomonas. The bacterium Pseudomonas sp. AP–3 recruits a unique 2-aminomuconate deaminase (AmnE) to hydrolyze 2-aminomuconate to 4-oxalocrotonate, and releases ammonia in the modified meta-cleavage pathway by forming various compounds—including acetaldehyde, pyruvic acid, acetyl-CoA, and succinate—that may enter the Krebs cycle. AmnE also belongs to the YjgF/YER057c/UK114 family (also known as the Rid family), which is conserved in all domains of life and prefers structurally homotrimeric forms with diverse functional purposes. To study the mechanism of the modified meta-cleavage pathway in Pseudomonas sp. AP–3, we determined the first crystal structure of AmnE from Pseudomonas sp. AP–3 at 1.75 Å. AmnE forms a unique homohexamer instead of a trimer which is normally adopted by the members of YjgF/YER057c/UK114 family. Based on the structure of the AmnE hexamer, we observed a hydrophobic base composed of six Lp3 loops (residues 122–131) in each of the AmnE protomers that have pivotal roles in the assembly of the hexamer. Eighteen hydrogen bonds formed by the residues Met96, Pro126, and Arg56, which surround the hydrophobic base, allowed the combination of the two trimers into a stable hexamer. The single mutant of AmnE R56A lost the ability to maintain the hexameric conformation, and revealed that the hydrogen bonds between residues Arg56 and Met96 have pivotal roles in the AmnE hexameric assembly.


INTRODUCTION
Many bacteria, including numerous species of Pseudomonas, possess genes for the degradation of aromatic compounds through the meta-cleavage pathway. Pseudomonas species AP-3 (Pseudomonas sp. AP-3) was isolated by growth based on the use of 2-aminophenol as the sole carbon, nitrogen, and energy source (Takenaka et al., 1997). A modified meta-cleavage pathway for 2-aminophenol metabolism has been studied comprehensively. This pathway is similar to that known for the meta-cleavage of catechol, except that one of the hydroxyl groups on the metabolites is replaced by an amino group (Takenaka et al., 1998). The 2-aminophenol is degraded by Pseudomonas sp. AP-3 via 2-aminomuconate-6-semialdehyde to 4-hydroxy-2-oxovalerate, which is further degraded to pyruvate and acetyl-CoA via the modified meta-cleavage pathway (Figure 1). A 13.9 kb region of the AP-3 strain encompasses eight tight clusters of genes designed in the form of amn [A-H]. Each gene in the modified meta-cleavage pathway is similar to the corresponding gene that operated in the meta-cleavage pathway, except for the amnE gene. The amnE gene product AmnE is a deaminase, and hydrolyzes 2-aminomuconate to 4-oxalocrotonate and releases ammonia in the modified meta-cleavage pathway (Takenaka et al., 2000). With the inclusion of the AmnE, the metabolic pathways of 2-aminophenol and its derivatives in Pseudomonas sp. AP-3 consist of seven major enzymes translated from the amn gene cluster (amnA-amnH) into a single operon, including 2-aminophenol 1,6-dioxygenase (AmnBA), 2-aminomuconic 6-semialdehyde dehydrogenase (AmnC), 4-oxalocrotonate decarboxylase (AmnD), 2-aminomuconate deaminase (AmnE), 2-oxopent-4-enoate hydratase (AmnF), 4-hydroxy-2-oxovalerate aldolase (AmnG), and acetaldehyde dehydrogenase (AmnH). Zhongqi H. and Jim C. S. also revealed the mechanism of the deamination reaction in the modified meta-pathway and identified the 2-aminomuconate deaminase acted specifically on the unsaturated α-amino acid 2-aminomuconate in P. pseudoalcaligenes JS45 (He and Spain, 1998). The deduced amino acid sequence of 2-aminomuconate deaminase in Pseudomonas sp. AP-3 shows that AmnE belongs to a member of the YjgF/YER057c/UK114 protein family.
The YjgF/YER057c/UK114 protein family (also known as the Rid family) includes 200 prokaryotic and eukaryotic proteins, and consists of 130 amino acid residues (∼15 kDa) which are found in all three domains of life (Kim et al., 2001). Phylogenetic analysis divided this family into a widely distributed archetypal RidA subfamily, and into seven other subfamilies (Rid1-Rid7). The archetypal RidA subfamily is found in all domains, while the Rid1-7 subfamilies are present only in prokaryotes. These proteins have widespread roles in metabolic processes in nearly every organism, and some species encode multiple members (Niehaus et al., 2015). In bacteria and yeast, protein family members exhibit predefined roles in the biosynthesis of essential compounds. Conversely, the mammalian members involved in the regulation of cellular differentiation inhibit translations by degrading messenger ribonucleic acid (Schmiedeknecht et al., 1996) identified and deposited in the protein data bank (PDB). Many other existing homologs have not been published yet. 31 structures in the RidA subfamily exhibit the preferential adoption of homotrimers and form a distinct inter-subunit cavity that may serve as a small molecular binding site, but with unknown functions. Based on the experimental evidence, the RidA subfamily members have been recently annotated as imine deaminase (Hodge-Hanson and Downs, 2017). Although the 2-aminophenol metabolic pathway and the corresponding catabolic enzymes have been studied extensively, the knowledge of the structural basis of the catabolic enzymes is still lacking. In this study, we demonstrate that AmnE selfassembles in a unique homo-hexameric form that is different from a regular trimer model of the RidA subfamily. The homohexameric structure of AmnE has been determined at a high resolution (equal to 1.75 Å), and constitutes the first protein structure in the metabolic pathway of 2-aminophenol. We analyzed the structure of the AmnE hexamer and explained how the six identical protomers assembled into a stable homohexameric conformation. Meanwhile, we identified that the residues Ala 49 -Pro 59 on the Lp2 loop region (residues 38-60) of AmnE have pivotal roles in the assembly of the hexameric structure. The single mutant of AmnE R56A lost the ability to maintain the hexameric conformation, and thus revealed that the hydrogen bonds between residues Arg 56 and Met 96 have pivotal roles in the AmnE hexameric assembly.

Generation of Wild Type and Correlated Mutant Plasmid Constructs
The gene encoding deaminase AmnE from Pseudomonas sp. AP-3 was synthesized by Sangon Biotech Co., Ltd., China, and was cloned into a pET-28a (GE Healthcare, United States) using a fast cloning method (Li et al., 2011). The single mutants AmnE R56A and AmnE P126A, and the truncated mutant AmnE 49-59 were generated by designing primers that encoded the mutated amino acids. Mutagenetic constructs were confirmed by commercial deoxyribonucleic acid (DNA) sequencing (Sangon Biotech, China). The primers used in this study are listed in Table 1.

Protein Expression and Purification
Wild plasmids of the type AmnE and correlated mutants were transformed into the BL21 (DE3) pLysS strain (TransGen Biotech, China) and grown in Luria-Bertani (LB) medium containing kanamycin (50 µg/ml) and chloromycetin (34 µg/ml). Growing cells were induced at an optical density (OD 600 ) value in  Figure S5A).
Optimization was carried out in 24-well plates by using hanging-drop method. Drops were mixed with 1 µl of protein solution at 10 mg/ml and 1 µl of reservoir solution. The high-quality crystals were obtained in 0.1 M CAPSO [3-(Cyclohexylamino)-2-hydroxy-1-propanesulfonic acid] pH 9.6, 0.2 M MgCl 2 •6 H 2 O, and 30% (w/v) PEG 4000 (Supplementary Figure S5B). The crystals were flash-cooled in liquid nitrogen and the diffraction data were collected in the Shanghai Synchrotron Radiation Facility (SSRF) Beamline BL19U1 (Shanghai, China). The data were processed with HKL2000 software (Otwinowski and Minor, 1997). The best crystal diffracted X-rays and formed a pattern with a high-resolution equal to 1.75 Å. The space group of AmnE is C121, with cell dimensions set at a = 170.871 Å, b = 54.781 Å, c = 134.594 Å (Supplementary Figure S5C), while the AmnE structure was solved by molecular replacement method with the Phaser (Lebedev et al., 2008) in the CCP4 suite (Winn et al., 2011), using the initial model of Pyrococcus horikoshii-PH0854 (PDB: 2DYY). The refined model was further built manually according to the electron density map with COOT (Emsley and Cowtan, 2004), and the qualities of the refined models were assessed with PHENIX (Adams et al., 2010). The final structure was derived to a R work factor of 15.7% (R free factor of 19.5%). All figures were generated by using PyMOL 1 . Data collection, statistics, and structure-refinement statistical outcomes are listed in Table 2.

Dynamic Light Scattering
Measurements were obtained using a Zetasizer µV (Malvern Panalytical, Worcestershire, United Kingdom) with AmnE at a concentration of 10 mg/ml in a buffer solution [50 mM HEPES, 150 mM NaCl, 1 mM DTT, 1 mM EDTA, and 5% (v/v) glycerol, pH 8.0] at 4 • C, 16 • C, 20 • C, and 25 • C, separately. The results were analyzed using software provided by the manufacturer. Experimental errors were estimated based on the estimated standard deviations.

Analytic Size-Exclusion Chromatography
Analytic size-exclusion chromatography (SEC) was adopted to investigate the oligomeric state of AmnE and associated mutants using a Superdex 200 10/300 GL column (GE Healthcare, United States). The column was pre-equilibrated with an analysis buffer [50 mM HEPES, 150 mM NaCl, 1 mM DTT, 1 mM EDTA, and 5% (v/v) glycerol, pH 8.0]. Furthermore, all the samples, at a concentration of 1 mg/ml, were analyzed at a flow rate of 0.3 ml/min at 16 • C. The molecular mass was calculated using the following equation,

Chemical Cross-Linking Assay
AmnE wild type and mutant proteins were purified with the use of a Superdex 75 10/300 GL column (GE Healthcare, is an individual intensity measurement and I(hkl) is the average intensity for all reflections.

CD Spectroscopy
Before CD measurements, all proteins were purified using a Superdex 75 10/300 GL column (GE Healthcare, United States) in buffer [50 mM HEPES, 150 mM NaCl, 1 mM DTT, 1 mM EDTA, 5% (v/v) glycerol, pH 8.0]. The CD spectra were recorded on a Jasco J-715 Spectropolarimeter (JASCO, MD, United States) with the use of three scans on average within the 190-260 nm wavelength range from solutions in 10 mm path length quartz cuvettes at 25 • C. Raw ellipticity data θ obs (in 10 4 millidegrees) were converted to mean residue ellipticity (θ) (in 10 4 millidegrees) using the formula, where θ is the ellipticity in 10 4 millidegrees, l is the path length of the cuvette in cm, MRW is the mean residue weight, and c is the concentration in mg/ml. Data were analyzed with Microsoft Office Excel and Origin 8.

Structural Basis of AmnE Trimer
AmnE adopts the classic homotrimer formed by three protomers like the other members of the YjgF/YER057c/UK114 family ( Figure 2B). The core of the homotrimer is composed of three β-sheets. Each one of these is closed into an empty triangular barrel-type structure and is surrounded by six α-helices on its outer part. The inner surface of the barrel forms a highly hydrophobic environment composed of β3, β4, and β6 strands. Viewed from the top to the bottom of the trimer, one end of the barrel is sealed off by three phenylalanines (Phe 31 ), while the other is sealed by a hydrophobic loop Lp3 from each of the three protomers (Supplementary Figure S7). Residues Leu 32 , Val 86 , Val 87 , Leu 93 , Ile 132 , and Ile 134 on the inner surface of the barrel are conserved in all of the members of the RidA subfamily (Supplementary Figure S8A). The hydrophobic residues are distributed on the inner surface of the barrel to maintain the stability of the homotrimer of AmnE. The single protomeraccessible surface buried in the trimer is 1994.1 Å 2 , and the proportion of the buried area to a protomer' total surface area is 25.89%. On the outer parts of the trimer, the interfaces of the adjacent subunits form three equivalent clefts, i.e., AB, BC, and CA (Figures 3A-C). These clefts are formed by α 4 , β 4 , and β 5 in one subunit, and β 2 , β 3 , and β 6 , loop Lp1, and α 4 in the adjacent subunit. Lp1 covers the cleft and is associated with elevated temperature factors and increased RMSD values among the three protomers. Comparison of the protein-ligand complex structures of hp14.5 (PDB code: 1ONI) and TdcF (PDB code: 2UYN), shows that the Lp1 of AmnE is potentially involved in unknown ligand binding (Supplementary Figure S9). The residues Ser 35 , Gly 36 , Glu 88 , Gln 103 , Arg 117 , and Lys 135 in these three clefts are highly conserved in the YjgF/YER057c/UK114 family ( Figure 3D). The areas of the three clefts (AB, BC, and CA) in AmnE were respectively, calculated and were equal to 760.43, 760.47, and 741.33 Å 2 , respectively. These areas could provide ideal contact surfaces for small molecules with unknown functions. The DALI search from the PDB library lists 26 homologies of the AmnE trimer in YjgF/YER057c/UK114 family ( Table 3). The structural superimposition suggests that the trimer is conserved in YjgF/YER057c/UK114 family proteins. The RMSD value estimated based on 192-290 Cα atoms in these structures varied in the range of 0.67-1.89 Å, with a sequence identity in the range of 23.00-40.51%. The superposing structures revealed that three Lp2 loops from each AmnE protomer extended out far from the main body of the trimer, and formed a unique conformation compared to other homologs in the YjgF/YER057c/UK114 family (Supplementary Figures S6D-F). Accordingly, the shape of the AmnE trimer resembled a three-legged structure.

The Structure of AmnE Hexamer
AmnE exists as a hexamer in solution and not in a trimeric state like the other members of the YjgF/YER057c/UK114 family. We identified the hexameric form of AmnE with the use of sizeexclusion chromatography and cross-linking methods. AmnE was eluted at 13.36 ml on a Superdex 200 size-exclusion column, and its calculated molecular mass was 103.5 kDa, which was close to that of the hexamer. This finding was also supported by crosslinking data. The crystal structure of AmnE was solved family. Sequence alignment of AmnE from Pseudomonas sp. AP-3 with YE07 from Saccharomyces cerevisiae (34.00% sequence identity), TdcF from Escherichia coli (29.00% sequence identity), and hp14.5 from homo sapiens (28.72% sequence identity). Residues labeled with triangles are conserved residues involved in the ligand binding pocket. Sequences were aligned using the tool obtained from the web site (https://www.ebi.ac.uk/Tools/msa/clustalo/) and the alignment was presented using the online ESPript 3.0 server (http://espript.ibcp.fr/ESPript/ESPript/). in the space group C121 with nine protomers (A, B, C, D, E, F, G, H, I) composed of three identical homotrimers (trimers I, II, and III) in one asymmetric unit. One of the AmnE hexamers is organized based on the use of two identical trimers, trimer I (A/B/C) and trimer II (D/E/F) (Figures 2B-F). Trimer III forms a hexamer with the use of trimer III in the neighboring asymmetric unit in the crystal lattice. The quaternary structure of AmnE was analyzed by PDBePISA. The total accessible surface of the two trimers buried in the hexamer is very large and equal to 3954.1 Å 2 . Correspondingly, the proportion of the buried area to the total area of the two trimers is 11.59%. In the AmnE hexameric structure, two trimers (trimer I, timer II) interact in a back-to-back manner through the hydrophobic region composed of six Lp3 loops (Lp3 A , Lp3 B , and Lp3 C , from trimer I, and Lp3 D , Lp3 E , Lp3 F , from trimer II). These loops are used to seal off the bottom of the hydrophobic channel in all of the trimers. Six residues His 123 (His 123 A−F ), located on the loops Lp3, form a hydrophobic base with residues Val 122 A−F , Leu 125 A−F , and Leu 130 A−F present in the AmnE hexamer (Figures 4A,B,D). Outside of the hexamer, six loops (Lp2 A , Lp2 B , and Lp2 C , from trimer I, and Lp2 D , Lp2 E , and Lp2 F , from trimer II) surround the hexamer and extend into the cleft area of the opposite trimer to form a solid football-like shape. Based on the structural analysis, the Lp2 (residues 49-59) was identified to play a pivotal role in the maintenance of the stability of the AmnE hexamer. There are six pivotal residues Arg 56 (Arg 56 A−F ) located on loops Lp2 (Lp2 A -F ) that formed eighteen hydrogen bonds in the hexamer with residues Pro 126 (Pro 126 A−F ) and Met 96 (Met 96 A−F ) ( Figure 4C). To further confirm that the functions of Lp2 and Arg 56 contribute to the maintenance of the conformation of the AmnE hexamer in vitro, the mutant proteins AmnE Lp2 (residues Ala 49 -Pro 59 ) and AmnE R56A were prepared and measured with CD spectroscopy before subsequent experiments (Figures 5A,B). The oligomerization states of the wild type and mutants of AmnE were evaluated by SEC. The calculated molecular mass of the wild-type AmnE and mutant AmnE P126A were respectively equal to 103 and 96.3 kDa, and were close to the theoretical molecular mass of the hexamer (Figures 5C,D). However, AmnE Lp2 (residues Ala 49 -Pro 59 ) and R56A exhibited larger retention volumes, which corresponded to the molecular weight of the AmnE trimer (Figures 5E,F). Meanwhile, a glutaraldehyde crosslinking experiment was conducted to verify the multimeric state of each protein (Figures 5G-I). Primarily and foremost, we confirmed that the AmnE hexamer was in a stable state in solution, and that the residues Arg 56 and Met 96 located on the hexameric interface were two key residues required to maintain the hexameric conformation of AmnE.

DISCUSSION
The 2-aminophenol compounds are intermediates in the biodegradation of nitrobenzene. The benzene ring cleavage and deamination steps are essential in the mineralization of 2aminophenol for the growth of the assimilating bacteria. The metabolic pathway for 2-aminophenol in Pseudomonas sp. AP-3 is referred to as the modified meta-cleavage pathway, in which 2-aminomuconate is transformed into 4-oxalocrotonate by the deaminase AmnE (Takenaka et al., 2000). In this study, we elucidated the first structure of 2-aminomuconate deaminase involved in the 2-aminophenol pathway in Pseudomonas sp. AP-3. AmnE also belongs to a member of the RidA subfamily in the YjgF/YER057c/UK114 family. In this family, proteins are distributed in all domains of life, apparently in the archetypal RidA subfamily and in seven other subfamilies. Members of the RidA subfamily prefer to form the trimeric, barrel-like quaternary structures, and inter-subunit cavities with diverse biological roles. Many of these are poorly understood despite their widespread existence (Sinha et al., 1999). In the PDB library, there are 33 structures of RidA subfamily in total, including one nuclear magnetic resonance (NMR) structure and 32 crystal structures. Most of these structures have not been published or characterized yet by the authors. Therefore, we expect to analyze all of the structures in the RidA subfamily and reveal that there are two primary types of homomultimeric protein complexes, i.e., trimeric and hexameric forms.
The trimeric conformation adopted by the RidA subfamily proteins is structurally conserved, and even the sequence similarity is comparatively low. These proteins adopt homotrimeric forms, and the clefts between the protomeric subunits are suggested to have functional relevance (Volz, 1999;Deaconescu et al., 2002;Deriu et al., 2003). The alignment of the RidA subfamily proteins has identified several residues in AmnE, including Ser 35 , Gly 36 , Glu 88 , Gln 103 , Arg 117 , and Lys 135 , which are highly conserved in clefts located at the subunit interfaces of the trimer (Supplementary Figure S8A). This indicates that these proteins play a similar role in cells. Lambrecht et al. revealed that RidA subfamily proteins exhibit enamine/imine deaminase activity and accelerate the release of ammonia from reactive enamine/imine intermediates (Lambrecht et al., 2010). Parsons et al. identified several compounds that interacted separately with the RidA subfamily protein HI0719 and protein hp14.5 in a pocket located in the clefts of the trimer separately (Parsons et al., 2003;Manjasetty et al., 2004). Therefore, the trimerization of the RidA subfamily proteins is necessary for achieving proper functionality in vivo. However, AmnE adopts homohexameric forms both in crystal structures and in solution states. A DALI search in the PDB library retrieved two similar hexameric structures (PDB code: 3KJK, 2EWC) based on the structure of the AmnE hexamer. These two proteins were identified as Nmb1025 in Neisseria Meningitidis and SPy_2060 in Streptococcus Pyogenes M1 Gas. All three proteins adopted hexameric assembly, which comprised two identical trimers in a back-to-back configuration forming the hydrophobic base in the middle, and six hydrophobic loops in the center of the hexamer (Figures 6A,B,E,F). Meanwhile, 15-21 hydrogen bonds around the hydrophobic base connected the two trimers together and maintained a stable hexameric structure. The hexameric structures of the proteins Nmb1025, SPy_2060, and AmnE, were analyzed by PDBePISA (Supplementary Table S1). All three proteins formed a hydrophobic base in loop Lp3 that was composed of the highly conserved residues Val 122 , Leu 125 , Pro 126 , Pro 128 , and Leu 130 on the AmnE loop Lp3, as verified by sequence alignment in RidA subfamily (Supplementary Figure S8A). In addition to the hydrophobic base in the center of the hexamer, the Nmb1025 hexamer recruited residues Arg 68 , Asp 69 , Glu 96 , and Arg 98 from each of the protomers, form hydrogen bonds ( Figure 6C). The structures of the hypothetical protein Spy_2060 formed five hydrogen bonds based on the Asp 72 , Gln 98 , Glu 100 , and Gly 106 from each of the protomers ( Figure 6D). Therefore, the stable hexamer requires the hydrophobic base with 15-21 hydrogen bonds around it to connect the two trimers together. However, these residues forming the hydrogen bonds in three hexameric structures are not highly conserved in RidA subfamily. This is probably the reason why most of members in RidA subfamily adopt a trimeric form instead of hexameric conformation. We constructed a multigene expression system including amnA, amnB, and amnC genes from Pseudomonas sp. AP-3 and expressed in E. coli to produce 2-aminomuconate in order to test the enzyme activity of hexameric AmnE and trimeric variants due to lack of stable substrate and cells of Pseudomonas sp. AP-3. However, we unable to get the 2-aminomuconate due to its high instability. Therefore, the hexameric structure of AmnE presented is the first step to understanding the molecular  mechanism of the different oligomerization types and their relation to their enzyme activity or substrate specificity in the modified meta-cleavage pathway.

DATA AVAILABILITY
Protein Data Bank Accession Codes: the atomic coordinates and structural factors of AmnE were submitted to the RCSB Protein Data Bank. The accession code is 6IZH.

AUTHOR CONTRIBUTIONS
DS and YA-C conceived the experimental study, designed the experiments, analyzed the data, and wrote the manuscript. HJ and DL designed the fast-cloning protocol and assisted YA-C in the preparation of the expression vectors. YA-C and TH expressed and purified the wild type and mutants of AmnE. YA-C crystallized AmnE. YA-C and YI-C collected the X-ray diffraction data. DS and YI-C determined and analyzed the crystal structures. YA-C measured the multimerization of the AmnE and mutants by cross-linking and gel-filtration experiments. YR, BY, GB, and HD provided the material, helped to design the experiments, and analyzed data. DS reviewed and provided the approval for the submission/publication of the manuscript.