Distinctive structural motifs co‐ordinate the catalytic nucleophile and the residues of the oxyanion hole in the alpha/beta‐hydrolase fold enzymes

Abstract

The alpha/beta‐hydrolases (ABH) are among the largest structural families of proteins that are found in nature. Although they vary in their sequence and function, the ABH enzymes use a similar acid–base‐nucleophile catalytic mechanism to catalyze reactions on different substrates. Because ABH enzymes are biocatalysts with a wide range of potential applications, protein engineering has taken advantage of their catalytic versatility to develop enzymes with industrial applications. This study is a comprehensive analysis of 40 ABH enzyme families focusing on two identified substructures: the nucleophile zone and the oxyanion zone, which co‐ordinate the catalytic nucleophile and the residues of the oxyanion hole, and independently reported as critical for the enzymatic activity. We also frequently observed an aromatic cluster near the nucleophile and oxyanion zones, and opposite the ligand‐binding site. The nucleophile zone, the oxyanion zone and the residue cluster enriched in aromatic side chains comprise a three‐dimensional structural organization that shapes the active site of ABH enzymes and plays an important role in the enzymatic function by structurally stabilizing the catalytic nucleophile and the residues of the oxyanion hole. The structural data support the notion that the aromatic cluster can participate in co‐ordination of the catalytic histidine loop, and properly place the catalytic histidine next to the catalytic nucleophile.

Abbreviations

ABH

structural family of alpha/beta‐hydrolases

EstFa_R

recombinant carboxylesterase from Ferroplasma acidiphilum

Nuc

catalytic nucleophile residue

PDB

Protein Data Bank

SCOP database

Structural Classification of Proteins database

Sm_nuc‐2

small residue located two positions before the catalytic nucleophile residue

X_nuc‐1

residue located adjacent (N‐terminal) to the catalytic nucleophile residue

X_oxyI

residues of the oxyanion hole, first position

X_oxyII

residues of the oxyanion hole, second position

X_ozI

residue at position “Oxyanion zone I”

X_ozII

residue at position “Oxyanion zone II”

Introduction

The alpha/beta‐hydrolases (ABH) are a structural family of enzymes that share a common fold structure and a catalytic mechanism, while differing in their degree of sequence similarity and function.1, 2, 3, 4 The ABH enzymes catalyze reactions on different substrates, but some ABH enzymes can also act as transporters, receptors, or have other ancillary functions.5 The ABH‐fold enzymes are structurally stable, and thus have been extensively used in protein engineering.6, 7

The ABH family members maintain a half‐barrel structure shaped by eight, mostly parallel β‐strands (the strand β2 is antiparallel) and six α‐helices (αA‐αF) flanking the β‐sheet. The β‐strands and the α‐helices are organized in an α‐turn‐β supersecondary structure geometry that starts with the αA‐helix followed by strand β4. The catalytic machinery of all ABH enzymes is similar, with the catalytic triad consisting of an acid: aspartic or glutamic acid; a base: histidine; and a nucleophile: serine, cysteine or aspartate. This catalytic triad is found at identical locations within the ABH fold (Fig. 1). In particular, the catalytic acid residue is located on a reverse turn after strand β7 or on a tight turn after strand β6; and the catalytic base is a histidine located on the loop following strand β8.1

Figure 1.

The active site of the ABH‐fold enzymes. Elements of the catalytic machinery are arranged as follows: the catalytic acid (“Catalytic Acid”) is located either after strand β7 (Group A) or strand β6 (Group B); the catalytic base (“Catalytic Base”); and the catalytic nucleophile (“Catalytic Nucleophile”). The oxyanion hole is usually formed by two residues: the Oxyanion I residue (“XoxyI”), adjacent to the catalytic nucleophile, and the Oxyanion II residue (“XoxyII”) in the loop following strand β3. The “Acid zone”, the “Nucleophile zone” and the “Oxyanion zone” designate the positions of three clusters of conserved residues that are linked by hydrogen bonds around the catalytic acid, the catalytic nucleophile and the residues of the oxyanion hole, respectively; the boundaries of each zone are marked by dashed lines. The positions of conserved elements of the zones are shown as: “Water” for structural water molecules; “His” for the conserved histidine of the Oxyanion zone; and “W/Y/F” for the aromatic residues tryptophan, tyrosine or phenylalanine from the aromatic cluster. The positions “Oxyanion zone I” (“XozI”) and “Oxyanion zone II” (“XozII”) are part of the Oxyanion zone

The catalytic nucleophile is the third key player of the charge‐relay system that enables the covalent catalysis by nucleophilic attack. The nucleophile is located at the apex of a sharp turn after strand β5, known as the “nucleophile elbow”, identified by the sequence motif Sm_nuc‐2–X_nuc‐1–Nuc–X_{nuc + 1}–Sm_{nuc + 2} (Sm, small residue; X, any residue; Nuc, nucleophile). The dipeptide Sm_nuc‐2–X_nuc‐1 is located at the end of strand β5 prior to the nucleophile and followed by the dipeptide X_{nuc + 1}–Sm_{nuc + 2} located at the N‐terminal end of the αC‐helix.

The helical axis of the αC‐helix forms an acute angle with respect to strand β5 and the helical N‐terminus is placed close to the C‐terminus of strand β3; this architecture leads to the arrangement of the “oxyanion hole” that stabilizes the tetrahedral intermediate during catalysis.1 The oxyanion hole is situated adjacent to the nucleophile and is mainly shaped by the main‐chain nitrogen atoms of two residues that we term here oxyanion I (X_oxyI) and oxyanion II (X_oxyII). Residue X_oxyI is always located after the nucleophile, at position X_{nuc + 1} of the nucleophile elbow.1 Residue X_oxyII is located after the start of loop_β3‐αA, from which the main‐chain nitrogen atom helps form the oxyanion hole in the majority of ABH enzymes. Sometimes, residue X_oxyII is located at an alternate position and the oxyanion hole is formed by a side chain. For example, residue X_oxyII can be a tyrosine at the start of loop_β3‐αA,8, 9 an aspartic acid after the start of loop_β3‐αA,10 or an arginine emerging from an α‐helix that is located opposite the active site.11 For some ABH enzymes, the oxyanion hole is formed by the main‐chain amide NH group from three residues.12, 13, 14

The nucleophile and the residues forming the oxyanion hole are essential for the enzymatic activity of ABH enzymes. Site‐directed mutagenesis of the nucleophile and residues forming the oxyanion hole results in the loss of enzymatic function, usually with little change in the local arrangement of the catalytic site and without interfering with the binding of the ligand8, 14, 15, 16, 17; however, it is possible that the mutation of residues of the oxyanion hole radically distorts the shape of loop_β3‐αA and hence inactivate the enzyme.18

In addition to the residues of the charge‐relay network, other residues around the active site have been shown to be essential for enzymatic activity. For example, Kiss et al.19 have shown experimentally that a residue located in the vicinity of the oxyanion hole affects the activity of the enzyme acylaminoacyl‐peptidase from Aeropyrum pernix. Specifically, the site‐directed mutation of His367 to alanine, located two residue‐positions before Gly369 of the Oxyanion II, distorted the shape of the oxyanion hole and consequently, peptidase activity was lost.

In our previous work, we studied the structure surrounding the catalytic acid and identified common structural elements that comprise the catalytic acid zone in ABH enzymes.20 We have shown that the catalytic acid zone co‐ordinates the position of the catalytic histidine loop relative to the catalytic acid and histidine residues. We have also shown that the catalytic acid zone is sometimes found at domain‐domain interfaces and residues of the acid zone are directly involved in the formation of multi‐subunit complexes and in protein–protein interactions.

Both the experimental study from Kiss et al.19 and our work on the catalytic acid zone20 suggested that there were additional elements important for the enzymatic activity in the structure around the active site. This motivated us to focus our attention on the architecture around the nucleophile and the oxyanion hole and describe their structural frameworks. In particular, we examined structures for the presence of common structural elements around each key part of the catalytic machinery of the charge‐relay system. Herein, we report the occurrence of unique substructures, which we refer to as the nucleophile zone and oxyanion zone, and discuss their role as supporting structural scaffolds for the enzymatic activity of ABH enzymes.

Results

From the structural point of view, ABH‐fold enzymes maintain a catalytic mechanism, which consists of an acid–base‐nucleophile triad and the oxyanion hole. Here, we have extensively studied all residues that directly surround key elements of the catalytic machinery; specifically, those residues adjacent to the nucleophile and the oxyanion hole, aiming to understand their influence on the activity of the ABH‐fold enzymes. We compared 41 structures, obtained from the Protein Data Bank21 (PDB), representing each of the 40 ABH enzyme families described in the SCOP database.22

This comprehensive analysis revealed structural conservation in three regions that surround the catalytic site: the nucleophile zone near the catalytic nucleophile and Oxyanion I residue X_oxyI; the oxyanion zone near Oxyanion II residue X_oxyII; and a cluster of aromatic residues that often surrounds both zones.

The nucleophile zone

The nucleophile zone is a planar structural organization that is found repeatedly in the ABH enzymes (Table 1). It co‐ordinates two key units of the catalytic mechanism: the catalytic nucleophile and the following residue, which is X_oxyI of the oxyanion hole. Structurally, the nucleophile zone is formed by residues that belong to the nucleophile elbow and the loop_β3‐αA, and its hydrogen‐bond network consists of two conserved, weak hydrogen bonds that connect residues between the nucleophile elbow and the loop_β3‐αA [for the definition of weak hydrogen bonds, see Materials and methods].

Table 1.

Inventory of Interactions within the Nucleophile Zone in 36 ABH‐Fold Families

N	SCOP Family #/ SCOP family name	PDB ID resolution	S nuc	O/Xnuc‐1 – CA/GoxyII‐1	CA/Gnuc + 2 – O/HoxyII‐2	N/XoxyI	N/XoxyII
A	2. Carboxylesterase	3WJ2_A23 1.61 Å	S156	O/D155‐CA/G84 3.1 (2.7)	CA/G158‐O/H83 3.3 (2.7)	N/A157	N/G85
1	1. Acetylcholinesterase‐like	1QE3_A75 1.50 Å	S189	O/E188‐CA/G105 3.2 (2.5)	CA/G191‐O/H104 3.6 (3.2)	N/A190	N/G106
2	2. Carboxylesterase	1LZL_A76 1.30 Å	S160	O/Q159‐CA/G87 3.1 (2.7)	CA/G162‐O/H86 3.2 (2.6)	N/A161	N/G88
3	3. Mycobacterial antigens	1DQZ_A47 1.50 Å	S124	O/L123‐CA/G39 4.1 (3.1)	OG/S126‐O/D38 2.7	N/M125	N/L40
4	4. Hypothetical protein TT1662	1UFO_A77 1.60 Å	S113	O/G112‐CA/G32 3.2 (2.9)	CA/G115‐O/H31 3.4 (2.9)	N/L114	N/L33
5	5. PepX catalytic domain‐like	3PUI_A24 1.53 Å	S117	‐	‐	N/Y118	OH/Y44
6	6. Prolyl oligopeptidase, C‐terminal domain	1H2W_A25 1.39 Å	S554	‐	‐	N/N555	OH/Y473
7	7. DPP4 catalytic domain‐like	1ORV_A26 1.80 Å	S630	‐	‐	N/Y631	OH/Y547
8	8. Serine carboxypeptidase‐like	3SC2_A48 2.20 Å	S146	O/E145‐CA/G52 3.8 (3.2)	CB/A148‐O/N51 4.0 (2.9)	N/Y147	N/G53
9	9. Gastric lipase	1HLG_A46 3.00 Å	S153	O/H152‐CA/G66 3.1 (2.7)	CA/G155‐O/H65 3.3 (2.8)	N/Q154	N/L67
10	10. Proline iminopeptidase‐like	1MTZ_A78 1.80 Å	S105	O/S104‐CA/G36 3.2 (2.8)	CA/G107‐O/H35 3.3 (2.7)	N/Y106	N/G37
11	11. Acetyl xylan esterase‐like	1L7A_A79 1.50 Å	S181	O/G180‐CA/G90 4.0 (3.3)	CA/G183‐O/H89 3.6 (3.4)	N/Q182	N/Y91
12	12. Haloalkane dehalogenase	1MJ5_A80 0.95 Å	D108	O/H107‐CA/G37 4.2 (3.6)	CA/G110‐O/H36 3.4 (3.0)	N/W109	N/N38
13	13. Dienelactone hydrolase	1ZI9_A27 1.50 Å	S123	O/Y122‐CA/D36 3.1 (2.3)	CA/G125‐O/Q35 3.3 (2.8)	N/L124	N/I37
14	14. Carbon–carbon bond hydrolase	2OG1_A81 1.60 Å	S112	O/N111‐CA/G41 3.3 (2.4)	CA/G114‐O/H40 3.2 (2.8)	N/M113	N/G42
15	15. Biotin biosynthesis protein BioH	4ETW_A82 2.05 Å	A82a	O/W81‐CA/G21 3.3 (2.7)	CA/G84‐O/H20 3.3 (2.7)	N/L83	N/W22
16	16. Aclacinomycin methylesterase RdmC	1Q0R_A28 1.45 Å	S102	O/L101‐CA/G31 3.3 (2.7)	CA/G104‐O/M30 3.2 (2.7)	N/M103	N/G32
17	17. Carboxylesterase/lipase	4DIU_A83 2.00 Å	S93	O/L92‐CA/G23 3.3 (2.7)	CA/G95‐O/H22 3.5 (3.0)	N/L94	N/F24
18	18. Epoxide hydrolase	1QO7_A84 1.80 Å	D192	O/G191‐CA/G116 4.4 (3.7)	CA/G194‐O/H115 3.4 (2.9)	N/I193	N/W117
19	19. Haloperoxidase	1BRT_A85 1.50 Å	S98	O/F97‐CA/G31 3.7 (3.1)	CA/G100‐O/H30 3.3 (2.4)	N/T99	N/F32
20	20. Thioesterases	1EI9_A86 2.25 Å	S115	O/F114‐CA/G40 3.2 (2.8)	CA/G117‐O/H39 3.8 (3.2)	N/Q116	N/M41
21	21. Carboxylesterase/thioesterase 1	1FJ2_A87 1.50 Å	S114	O/F113‐CA/G24 3.3 (2.9)	CA/G116‐O/H23 3.4 (2.6)	N/Q115	N/L25
22	22. Ccg1/TafII250‐interacting factor B (Cib)	1IMJ_A88 2.20 Å	S111	O/P110‐CA/G40 4.2 (3.4)	CB/S113‐O/H39 3.6 (2.6)	N/L112	N/I41
23	23. A novel bacterial esterase	1QLW_A89 1.09 Å	S206	O/H205‐CA/G70 3.3 (2.7)	OG/S208‐O/H69 2.9	N/Q207	N/C71
24	24. Lipase	1JFR_A30 1.90 Å	S131	O/H130‐CA/G62 3.3 (3.0)	CA/G133‐O/P61 3.4 (2.7)	N/M132	N/F63
25	25. Fungal lipases	1TCA_A31 1.55 Å	S105	O/W104‐CA/G39 3.5 (2.8)	CA/G107‐O/P38 3.5 (3.1)	N/Q106	N/T40
26	26. Bacterial lipase	1ISP_A90 1.30 Å	S77	O/H76‐CA/G11 3.6 (3.3)	CA/G79‐O/H10 3.4 (2.9)	N/M78	N/I12
27	27. Pancreatic lipase, N‐terminal domain	1BU8_A91 1.80 Å	S152	O/H151‐CA/G76 3.6 (2.8)	CA/G154‐O/H75 3.3 (2.7)	N/L153	N/F77
28	28. Hydroxynitrile lyase‐like	3C6X_A92 1.05 Å	S80	O/E79‐CB/T11 3.5 (2.7)	CA/G82‐O/H10 4.1 (3.7)	N/C81	N/I12
29	29. Thioesterase domain of polypeptide, polyketide and fatty acid synthases	1JMK_C32 1.71 Å	S80	O/Y79‐CA/P26 3.3 (2.3)	CA/G82‐O/P25 3.0 (2.5)	N/A81	N/V27
30	30. Cutinase‐like	1BS9_A29 1.10 Å	S90	O/Y89‐CA/E12 3.4 (2.8)	CA/G92‐O/R11 3.6 (3.1)	N/Q91	N/T13
31	31. YdeN‐like	1UXO_A93 1.80 Å	S71	O/H70‐CA/G10 3.4 (3.1)	CA/G73‐O/H9 3.5 (2.9)	N/L72	N/Y11
32	32. Putative serine hydrolase Ydr428c	1VKH_A94 1.85 Å	S110	O/H109‐CA/G37 3.2 (2.5)	CA/G112‐O/H36 3.5 (3.0)	N/V111	N/G38
33	33. Acylamino‐acid‐releasing enzyme, C‐terminal domain	1VE6_A95 2.10 Å	S445	O/Y444‐CA/G368 3.7 (3.3)	CA/G447‐O/H367 3.5 (3.2)	N/Y446	N/G369
34	34. Hypothetical esterase YJL068C	1PV1_A96 2.30 Å	S161	O/H160‐CA/G57 3.6 (3.2)	CA/G163‐O/S56 3.4 (3.0)	N/M162	N/L58
35	35. Hypothetical protein VC1974	1R3D_A97 1.90 Å	S91	O/Y90‐CA/G23 3.4 (3.0)	CA/G93‐O/H22 3.3 (2.6)	N/L92	N/L24
36	36. Atu1826‐like	2I3D_A98 1.50 Å	S108	O/Y107‐CA/P33 3.5 (2.8)	CA/G110‐O/H32 3.3 (2.9)	N/F109	N/H34
37	37. PHB depolymerase‐like	2D80_A35 1.70 Å	S39	O/L38‐CA/G249 3.1 (2.5)	CA/G41‐O/H248 3.4 (2.6)	N/S40	N/C250
38	38. IroE‐like	2GZS_A11 1.40 Å	S189	‐	‐	N/Y190	NH1/R130
39	40. O‐acetyltransferase	2B61_A99 1.65 Å	S143	O/G142‐CB/A48 3.7 (2.8)	CA/G145‐O/H47 3.3 (2.9)	N/F144	N/L49
40	41. 2,6‐dihydropseudo‐oxynicotine hydrolase‐like	2JBW_A49 2.10 Å	S217	O/R216‐CA/G146 3.6 (3.1)	CA/G219‐O/G145 3.0 (2.7)	N/L218	N/L147

The nucleophile zone coordinates the catalytic nucleophile (Column S_nuc) and the Oxyanion I residue (Column N/X_oxyI), and supports the formation of the oxyanion hole by fixing their relative conformation near the Oxyanion II residue (Column N/X_oxyII). The two weak hydrogen bonds of the nucleophile zone (Columns O/X_nuc‐1–CA/G_oxyII‐1 and CA/G_{nuc + 2}–O/H_oxyII‐2) occur in 36 ABH families, including the carboxylesterase EstFa_R (PDB ID: 3WJ2), which we use as the reference structure (row A). In four families (Rows 5, 6, 7, and 38), the contact network of the nucleophile zone is disrupted. The “‐” designates that the respective hydrogen bond is not formed; values in parentheses show distances to the hydrogen atom of the corresponding hydrogen bond.

^aIn biotin biosynthesis protein BioH (SCOP Family 15), the nucleophile is serine,100, 101 and not alanine as in the mutant with the PDB ID: 4ETW.82 However, the nucleophile side chain is not involved in the formation of the nucleophile zone. It points outwards, and does not interfere with the network of contacts within the zone.

For example, in the carboxylesterase EstFa_R from Ferroplasma acidiphilum 23 (PDB ID: 3WJ2), the shape of the nucleophile zone is defined by residues of the nucleophile elbow Asp155–Ser156–Ala157–Gly158 and by the dipeptide His83–Gly84 at the start of loop_β3‐αA [Fig. 2(A)]. One of the two connecting weak hydrogen bonds is formed between O/Asp155 at position X_nuc‐1 and CA/Gly84 that precedes Gly85 at X_oxyII. The second weak hydrogen bond is formed between CA/Gly158_{nuc + 2} and O/His83_oxyII‐2; this contact anchors the αC‐helix to loop_β3‐αA.

Figure 2.

The nucleophile zone (A), the oxyanion zone (B), and the aromatic cluster (C) in the active site of ABH enzymes. (A) The nucleophile zone in the structure of the carboxylesterase EstFa_R from Ferroplasma acidiphilum (PDB ID: 3WJ2). The zone is shaped by the residues Asp155–Ser156–Ala157–Gly158 of the nucleophile elbow and His83–Gly84 at the start of loop_β3‐αA, and involves two, weak hydrogen bonds, O/Asp155–CA/Gly84 and CA/Gly158–O/His83. The residues Tyr82, Gly85, and Gly154 occupy semi‐conserved, peripheral positions and form four, supplementary hydrogen bonds to the nucleophile zone. “Snuc” designates the catalytic nucleophile residue, which is a serine in the majority of ABH enzymes; main‐chain nitrogen atoms at the oxyanion I (“N/XoxyI” – from Ala157 in EstFa_R) and Oxyanion II (“N/XoxyII” – from Gly85 in EstFa_R) positions form the oxyanion hole. (B) The oxyanion zone in the structure of the carboxylesterase EstFa_R from Ferroplasma acidiphilum (PDB ID: 3WJ2). The zone is formed by: the residues Tyr81–Tyr82–His83–Gly84 at the end of strand β3 and the start of loop_β3‐αA; the Gly90–Asn91–Ile92 at the end of loop_β3‐αA; and Ile113–Glu114 from strand β4. All amino acids except XoxyII, and all contacts except the structural β‐sheet hydrogen bond between Tyr81 and Ile113, belong to the oxyanion zone.“N/XoxyII” designates the main‐chain nitrogen atom of oxyanion II residue Gly85 in EstFa_R; and “Wat409” designates the conserved structural water molecule forming multiple hydrogen bonds in the oxyanion zone of the ABH enzymes. A hydrogen bond is formed between the main‐chain nitrogen atom of the residue at “oxyanion zone I” (“N/XozI”, Ile92 in EstFa_R) and the residue at “Oxyanion zone II” (“OD,OE/XozII”, side‐chain oxygen atom of Glu114 in EstFa_R). (C) Conserved residues at the aromatic cluster near the active site of the carboxylesterase EstFa_R from Ferroplasma acidiphilum (PDB ID: 3WJ2). The residues Tyr81, Tyr95, and Asp155 are located at positions where aromatic amino acids (as well as hydrophobic residues) are frequently found in all ABH enzymes; the side‐chain interactions of the residues at these positions form a “roof”‐like structural arrangement over the β‐sheet. Conservation of an aromatic residue is also observed before or after the residue at the position “oxyanion zone II” (“XozII”, Glu114 in EstFa_R); in EstFa_R, the side‐chain hydroxyl of Tyr115 interacts with the main‐chain oxygen of His83, and the two residues form an aromatic pair below the β‐sheet.“Snuc” and “XoxyII” designate the catalytic nucleophile and oxyanion II residue of the catalytic machinery, and “Wat409” and “XozI” designate the conserved structural water molecule and the residue at position “oxyanion zone I” of the Oxyanion zone.

Besides the residues of the nucleophile zone, there are also flanking residues, which in EstFa_R are Tyr82_oxyII‐3, Gly85_oxyII and Gly154_nuc‐2. X_oxyII‐3, X_oxyII and Sm_nuc‐2 interact with the nucleophile zone and with each other [Fig. 2(A)], and contribute to the local stability. The distances of these peripheral contacts are detailed in the Supplementary material (Table S1).

Unlike most ABH families that have the main and supplementary contacts of the nucleophile zone, four structural representatives are exceptions (PDB IDs: 3PUI,24 1H2W,25 1ORV,26 and 2GZS11 in Table 2), where the entire contact network of the nucleophile zone is disrupted when residue X_oxyII is located at an alternate position in the fold. In these cases, the nucleophile zone is not formed (see Discussion).

Table 2.

Inventory of Interactions within “the Oxyanion Zone” in 29 ABH‐Fold Families

N	SCOP Family #/ SCOP family name	PDB ID resolution	(ND1, CD2)/HoxyII‐2 – O/XoxyII‐1	(ND1, CD2)/HoxyII‐2 – O/XozII‐1	N/XozI‐2 – O/XoxyII‐1	N/XozI – OD/XozII
A	2. Carboxylesterase	3WJ2_A23 1.61 Å	ND1/H83‐O/G84 2.6	CD2/H83‐O/I113 3.9 (3.0)	CA/G90‐O/G84 3.9 (2.9)	N/I92‐OE2/E114 2.8
1	1. Acetylcholinesterase‐like	1QE3_A75 1.50 Å	ND1/H104‐O/G105 2.7	CD2/H104‐O/L134 3.8 (2.7)	CA/G111‐O/G105 4.1 (3.1)	N/G113‐OD1/N135 2.9
2	2. Carboxylesterase	1LZL_A76 1.30 Å	ND1/H86‐O/G87 2.7	CD2/H86‐O/V116 4.0 (3.1)	CA/G93‐O/G87 4.2 (3.1)	N/A95‐OE1/E117 2.8
3	4. Hypothetical protein TT1662	1UFO_A77 1.60 Å	ND1/H31‐O/G32 2.8	CD2/H31‐O/F57 3.5 (2.7)	N/G35‐O/G32 2.9	N/K37‐OD2/D58 2.9
4	9. Gastric lipase	1HLG_A46 3.00 Å	CD2/H65‐O/G66 3.0 (1.9)	ND1/H65‐O/G97 4.2	N/A69‐O/G66 3.1	CB/A71‐O/G97 3.3 (2.5)
5	10. Proline iminopeptidase‐like	1MTZ_A78 1.80 Å	CD2/H35‐O/G36 2.9 (1.9)	ND1/H35‐O/Y61 3.3	N/M40‐O/G36 3.3	N/H42‐OD2/D62 2.9
6	11. Acetyl xylan esterase‐like	1L7A_A79 1.50 Å	ND1/H89‐O/G90 2.7	CD2/H89‐O/M115 3.4 (2.5)	N/A93‐O/G90 3.1	OG/S94‐CD2/L116 3.7 (2.8)
7	12. Haloalkane dehalogenase	1MJ5_A80 0.95 Å	ND1/H36‐O/G37 2.8	CD2/H36‐O/C61 3.2 (2.4)	N/T40‐O/G37 3.0	N/S42‐OD1/D62 2.9
8	13. Dienelactone hydrolase	1ZI9_A27 1.50 Å	NE2/Q35‐O/D36 3.3	OE1/Q35‐N/L63 3.1	N/G39‐O/D36 3.6	CG2/V40‐OD1/D62 4.3 (3.5)
9	14. Carbon–carbon bond hydrolase	2OG1_A81 1.60 Å	ND1/H40‐O/G41 2.8	CD2/H40‐O/K69 3.1 (2.3)	CB/A46‐O/G41 3.4 (3.0)	N/G48‐OD1/D70 2.8
10	15. Biotin biosynthesis protein BioH	4ETW_A82 2.05 Å	ND1/H20‐O/G21 2.8	CD2/H20‐O/V45 3.1 (2.2)	N/L24‐O/G21 3.4	N/A26‐OD1/D46 2.9
11	16. Aclacinomycin methylesterase RdmC	1Q0R_A28 1.45 Å	CB/M30‐O/G31 3.9 (2.9)	‐	N/L34‐O/G31 3.4	N/A36‐OD1/D58 2.8
12	17. Carboxylesterase/ lipase	4DIU_A83 2.00 Å	ND1/H22‐O/G23 2.8	CD2/H22‐O/P48 3.4 (2.4)	N/G26‐O/G23 2.9	OG/S28‐CD1/I49 3.1 (2.0)
13	18. Epoxide hydrolase	1QO7_A84 1.80 Å	CD2/H115‐O/G116 3.0 (1.9)	ND1/H115‐O/P147 3.4	N/G119‐O/G116 3.0	CD1/F121‐OG/S148 3.3 (2.7)
14	19. Haloperoxidase	1BRT_A85 1.50 Å	ND1/H30‐O/G31 2.8	CD2/H30‐O/Y56 3.3 (2.3)	N/L34‐O/G31 3.0	N/G36‐OD1/D57 2.8
15	20. Thioesterases	1EI9_A86 2.25 Å	CD2/H39‐O/G40 3.0 (2.0)	ND1/H39‐O/L70 3.7	N/D43‐O/G40 3.4	N/C45‐OD2/D79 3.0
16	21. Carboxylesterase/ thioesterase 1	1FJ2_A87 1.50 Å	ND1/H23‐O/G24 2.7	CD2/H23‐O/P49 3.6 (2.8)	N/D27‐O/G24 3.5	N/H30‐Wat859‐ ‐CE1/H50 2.8; 3.4 (2.6)
17	22. Ccg1/TafII250‐interacting factor B (Cib)	1IMJ_A88 2.20 Å	ND1/H39‐O/G40 2.7	CD2/H39‐O/I67 3.3 (2.3)	N/F43‐O/G40 3.1	N/S45‐OD1/D68 2.9
18	23. A novel bacterial esterase	1QLW_A89 1.09 Å	ND1/H69‐O/G70 2.8	CD2/H69‐O/I102 3.4 (2.4)	N/L73‐O/G70 4.0	N/G75‐OD1/D103 2.8
19	26. Bacterial lipase	1ISP_A90 1.30 Å	ND1/H10‐O/G11 2.8	CD2/H10‐O/V39 3.3 (2.3)	N/G14‐O/G11 2.9	N/S16‐OD1/D40 2.9
20	27. Pancreatic lipase, N‐terminal domain	1BU8_A91 1.80 Å	CD2/H75‐O/G76 3.2 (2.2)	ND1/H75‐O/V104 3.3	CG1/I78‐O/G76 3.3 (2.3)	CD1/I78‐(CG,CD)/R107 4.0; 4.0
21	28. Hydroxynitrile lyase‐like	3C6X_A92 1.05 Å	ND1/H10‐O/T11 2.7	CD2/H10‐O/L36 3.2 (2.4)	N/H14‐O/T11 3.1	N/A16‐OD1/D37 2.8
22	30. Cutinase‐like	1BS9_A32 1.10 Å	CG/R11‐O/E12 3.6 (2.6)	NH1/R11‐O/I41 3.7	N/A15‐O/E12 3.2	N/G18‐Wat320‐ ND2/N42 3.0; 3.5
23	31. YdeN‐like	1UXO_A29 1.80 Å	ND1/H9‐O/G10 2.7	CD2/H9‐O/L37 3.5 (2.8)	N/A13‐O/G10 3.5	N/S15‐OD1/N38 2.8
24	32. Putative serine hydrolase Ydr428c	1VKH_A94 1.85 Å	ND1/H36‐O/G37 2.8	CD2/H36‐O/I71 3.5 (2.7)	CB/N45‐O/G37 3.0 (2.5)	CD/P47‐OE2/E72 3.2 (2.2)
25	33. Acylamino‐acid‐releasing enzyme, C‐terminal domain	1VE6_A95 2.10 Å	CD2/H367‐O/G368 2.6 (1.6)	ND1/H367‐O/P395 3.9	CB/A372‐O/G368 3.5 (2.6)	N/D374‐OD1/N396 2.7
26	35. Hypothetical protein VC1974	1R3D_A97 1.90 Å	CD2/H22‐O/G23 2.8 (1.7)	ND1/H22‐O/L48 3.3	N/G26‐O/G23 3.0	N/G28‐OD1/D49 2.8
27	36. Atu1826‐like	2I3D_A98 1.50 Å	ND1/H32‐O/P33 2.7	CD2/H32‐O/F63 3.3 (2.3)	CA/G39‐O/P33 3.3 (2.3)	N/M41‐OD1/N64 2.7
28	37. PHB depolymerase‐like	2D80_A35 1.70 Å	ND1/H248‐O/G249 2.8	CD2/H248‐O/P281 3.6 (2.7)	N/Q252‐O/G249 3.1	N/Y254‐NE2/Q282 3.0
29	40. O‐acetyltransferase	2B61_A99 1.65 Å	ND1/H47‐O/A48 2.7	CD2/H47‐O/S85 3.2 (2.4)	N/G51‐O/A48 3.0	CA/D52‐OD1/N86 3.9 (2.9)

Contacts between residues at Positions H_oxyII‐2 and X_oxyII‐1, H_oxyII‐2 and X_ozII‐1, X_ozI‐2 and X_oxyII‐1, X_ozI and X_ozII, are shown. The ABH families that partially or entirely lack the contact network of the oxyanion zone are not shown (SCOP ABH Families 3, 5, 6, 7, 8, 24, 25, 29, 34, 38. and 41). Each “‐” sign designates that the respective hydrogen bond is not formed, while values in parentheses show distances to the hydrogen atom of the corresponding hydrogen bond.

The nucleophile zone and its adjacent structural elements that have been described, altogether serve to optimally arrange the oxyanion hole for the 36 of 40 ABH enzyme families. The hydrogen‐bonding network of the nucleophile zone, together with the peripheral interactions, fix the conformation of the catalytic nucleophile relative to the residues forming the oxyanion hole; the co‐ordination of the second residue of the oxyanion hole also depends on the oxyanion zone that surrounds loop_β3‐αA, as is described in the following section.

The oxyanion zone

Adjacent to the nucleophile zone, there is a planar structural organization that we call the oxyanion zone. The oxyanion zone stabilizes the shape of loop_β3‐αA and co‐ordinates residue X_oxyII. With regard to the structure of the active site, the role of the oxyanion zone is complementary to that of the nucleophile zone with which it overlaps through the dipeptide X_oxyII‐2–X_oxyII‐1. Structurally, the oxyanion zone is formed by residues that belong to strand β3, loop_β3‐αA and strand β4, and it is mainly organized around a conserved histidine residue at the start of loop_β3‐αA; histidine is located two residues before X_oxyII and the imidazole side chain faces the interior of the protein in 26 of 40 ABH families (Columns 4 and 5, Table 2). The hydrogen‐bond network of the oxyanion zone consists of four contacts: two contacts link the imidazole ring with neighboring residues and two contacts are formed among other residues of that zone. In 3 of the 40 ABH families, alternative residues are present having the same conformation as the histidine and the zone is formed; whereas, in 11 of 40 ABH families, the oxyanion zone is not present because the hydrogen‐bonding network is not properly formed due to short side chains or a side chains having a different conformation from histidine.

In EstFa_R23, for example, the oxyanion zone is shaped by three elements: (i) the tetrapeptide Tyr81–Tyr82–His83–Gly84 at the C‐terminus of strand β3 and the start of loop_β3‐αA; (ii) the tripeptide Gly90–Asn91–Ile92 at the C‐terminus of loop_β3‐αA; and (iii) the dipeptide Ile113–Glu114 at strand β4 [Fig. 2(B)]. The conserved histidine at the start of loop_β3‐αA (His83 in EstFa_R) forms two conserved contacts: a hydrogen bond with the adjacent Gly84 (ND1/His83–O/Gly84), and a weak hydrogen bond with Ile113 at strand β4 (CD2/His83–O/Ile113). The Residues Gly84 and Gly90, located at each side of loop_β3‐αA, are linked by a weak hydrogen bond (O/Gly84–CA/Gly90). Lastly, a hydrogen bond N/Ile92–OE2/Glu114 connects the end of loop_β3‐αA with strand β4 at positions that we refer to as “Oxyanion zone I” (X_ozI; Ile92 in EstFa_R) and “Oxyanion zone II” (X_ozII; Glu114 in EstFa_R).

In addition to the geometry described above, we frequently observed a conserved structural water molecule positioned near the side chain of the residue at the start of loop_β3‐αA, and bound by multiple interactions [Fig. 2(B)]. For example, in EstFa_R, the water Wat409 interacts with His83 through a weak hydrogen bond (Wat409–CD2/His83), and with the other residues of the oxyanion zone through standard (Tyr81, His83, Ile113) and weak (Tyr82, Ile92) hydrogen bonds. The interaction distances of the water molecule‐based contact network can be found in the Supplementary material (Table S2).

Besides the conserved histidine at the start of loop_β3‐αA, which is critical for maintaining the shape of the loop, two key residues are highly conserved within the oxyanion zone. Gly84 (Gly_oxyII‐1) in EstFa_R is located between the conserved histidine and X_oxyII. Like histidine, Gly_oxyII‐1 is a structural element shared between the nucleophile and oxyanion zones present in 30 of the 40 ABH families (Table 1). The second conserved residue of the oxyanion zone is an acidic amino acid – aspartate in 13 families and glutamate in 3 families – located at strand β4 at Position X_ozII in 16 of the 29 ABH families having the oxyanion zone (Table 2). Alternatively, position X_ozII can be occupied by an amide – asparagine in 6 families and glutamine in 1 family and they play a similar role in 7 of the 29 ABH families. In contrast to X_ozII, the type of amino acid that occupies position X_ozI is variable since it is the main‐chain nitrogen atom that participates in the conserved interaction between residues X_ozI and X_ozII (Column N/X_ozI–OD/X_ozII in Table 2); this interaction likely offers additional stabilization of the geometry of loop_β3‐αA.

Thus, the oxyanion zone is formed by seven basic elements: (i) a tetrapeptide at the end of strand β3 and the start of loop_β3‐αA; (ii) a tripeptide at the end of loop_β3‐αA; (iii) a dipeptide at strand β4; (iv) a conserved histidine at position X_oxyII‐2; (v) a conserved glycine at position X_oxyII‐1; (vi) a conserved acidic residue at position X_ozII; and (vii) a conserved structural water molecule that is positioned at the center of the zone, near the imidazole ring of the conserved histidine at X_oxyII‐2 [Fig. 2(B)].

All in all, the oxyanion zone, the nucleophile zone and the previously described catalytic acid zone20 are three common structural organizations of the active site of ABH enzymes that lie on the same plane within the β‐sheet and co‐ordinate the residues of the oxyanion hole, the catalytic nucleophile and the catalytic acid residue. The catalytic acid zone interacts with the catalytic histidine loop and co‐ordinates the catalytic histidine next to the catalytic acid residue.20 The results of that study showed that the conservation of the structure of the active site is extended beyond the plane of the β‐sheet. This finding appears to be valid also for the space that surrounds the nucleophile and the oxyanion zones, as described below.

Aromatic residues surround the catalytic nucleophile and residues of the oxyanion hole

The planar structural organization of both the nucleophile zone and the oxyanion zone primarily involve main‐chain interactions, and thus, the overall shapes of the zones are based on the protein fold. Indeed, apart from the residues at positions X_oxyII‐2 and X_ozII, the side chains of other residues of the nucleophile and oxyanion zones are not involved in the formation of the geometry of the two zones. Within the nucleophile and oxyanion zones, or located nearby, there are, however, conserved residues whose side chains interact outside of the plane of the β‐sheet.

For example, in the carboxylesterase EstFa_R23 the Residues X_nuc‐1 (Asp155) from the nucleophile zone and X_oxyII‐4 (Tyr81) from the oxyanion zone interact through a weak hydrogen bond [OD1/Asp155–CE2/Tyr81 in Fig. 2(C)], and their side‐chains lie parallel and above these two zones. A nearby residue, Tyr95, which is located after position X_ozI (Ile92_ozI), has a similar side‐chain conformation and interacts with both X_nuc‐1 (OD1/Asp155–CE2/Tyr95) and X_oxyII‐4 (OD2/Asp155–OH/Tyr95) through weak and conventional hydrogen bonds; the aromatic ring of Tyr95 also forms a CH‐π interaction with the side‐chain CH₂ group of His83, and seems to play a role in stabilizing and directing the side chain of His83 toward the protein interior. Thus, the side chains of all three residues are linked outside the plane of the nucleophile and the oxyanion zones.

As a unit, the hydrogen‐bond network among these three residues (Asp155, Tyr81, and Tyr95 in EstFa_R) creates a “roof”‐like structural arrangement, positioned over the plane of the nucleophile and oxyanion zones. In order to establish whether this is a common aspect among the ABH families, we have compared their equivalent positions with the structure of carboxylesterase EstFa_R: namely, (i) the residue at Position X_nuc‐1 (Asp155), (ii) the residue at Position X_oxyII‐4 (Tyr81), and (iii) the residue (Tyr95) that protrudes from the start of the αA‐helix or the end of loop_β3‐αA and lies above the β‐sheet and within an interaction distance of X_nuc‐1 and X_oxyII‐4 – two positions of the nucleophile and oxyanion zones.

By comparing the active sites of the 40 ABH family representatives, we observed aromatic residues (Trp, Tyr, Phe, His; Table 3): in 22 of 40 ABH families at position X_nuc‐1, in 8 of 40 ABH families at X_oxyII‐4, and in 24 of 40 ABH families at the position equivalent to Tyr95 in EstFa_R. This latter position is occupied by tryptophan in 14 of the 40 ABH families and its side chain has a characteristic perpendicular conformation over the β‐sheet (Column W/Y/F in Table 3). The amino‐acid type at position X_nuc‐1 is more variable than for the other two locations: it is the only position that is occupied by a histidine residue in addition to tryptophan, tyrosine, and phenylalanine (Column X_nuc‐1 in Table 3). Residue X_oxyII‐4 has a lower occupancy of aromatic residues but larger hydrophobic amino acids (leucine, isoleucine, methionine) are present in the “aromatic cluster” in 24 of the 40 ABH families (Column X_oxyII‐4 of Table 3). Together, we observed that the residues from these three positions often cluster – forming a “roof”‐like structural arrangement – frequently enriched in aromatic side chains positioned over the plane of nucleophile and oxyanion zones: 13 ABH families have aromatic pairs and 3 ABH families have aromatic triads but only 5 ABH families do not have any aromatic residue at one of these three positions.

Table 3.

Amino Acid Sets that Form Aromatic Clusters Near the Active Site

N	SCOP Family #/ SCOP family name	PDB ID Resolution	XoxyII	XoxyII‐4	W/Y/F	Xnuc‐1	Snuc	SC/XozII ± 1 – O/HoxyII‐2
A	2. Carboxylesterase	3WJ2_A23 1.61 Å	G85	Y81	Y95	D155	S156	OH/Y115‐O/H83; 2.6
1	1. Acetylcholinesterase‐like	1QE3_A75 1.50 Å	G106	W102	Y118	E188	S189	OH/Y136‐O/H104; 2.7
2	2. Carboxylesterase	1LZL_A76 1.30 Å	G88	W84	S98	Q159	S160	OH/Y118‐O/H86; 2.7
3	3. Mycobacterial antigens	1DQZ_A47 1.50 Å	L40	L36	W49	L123	S124	O/G71‐OH/Y77; 4.0 CE1/Y77‐O/D38; 3.6
4	4. Hypothetical protein TT1662	1UFO_A77 1.60 Å	L33	A29	I40	G112	S113	CB/A59‐π/F88; 3.4 CE2/F88‐O/H31; 3.5
5	5. PepX catalytic domain‐like	3PUI_A24 1.53 Å	OH/Y44	V40	W52	V116	S117	‐
6	6. Prolyl oligopeptidase, C‐terminal domain	1H2W_A25 1.39 Å	OH/Y473	Y471	I480	G553	S554	‐
7	7. DPP4 catalytic domain‐like	1ORV_A26 1.80 Å	OH/Y547	E545	K554	W629	S630	‐
8	8. Serine carboxypeptidase‐like	3SC2_A48 2.20 Å	G53	W49	S58	E145	S146	O/S95‐OH/Y151; 4.4 CE2/Y151‐O/N51; 3.1
9	9. Gastric lipase	1HLG_A46 3.00 Å	L67	L63	W74	H152	S153	OG/S99‐OE1/Q64; 3.7 NE2/Q64‐O/H65; 3.3
10	10. Proline iminopeptidase‐like	1MTZ_A78 1.80 Å	G37	T33	L45	S104	S105	OE1/Q63‐O/H35; 3.0
11	11. Acetyl xylan esterase‐like	1L7A_A79 1.50 Å	Y91	K87	E98	G180	S181	CG2/V117‐NE2/Q182;3.9 NE2/Q182‐O/H89; 2.8
12	12. Haloalkane dehalogenase	1MJ5_A80 0.95 Å	N38	F34	W45	H107	D108	CD1/L63‐O/H36; 3.8
13	13. Dienelactone hydrolase	1ZI9_A27 1.50 Å	I37	I33	M44	Y122	S123	CD1/L63‐CD1/L128; 3.7 CD1/L128‐O/Q35; 5.4
14	14. Carbon–carbon bond hydrolase	2OG1_A81 1.60 Å	G42	M38	Y52	N111	S112	CB/S71‐CE/M113; 4.0 CE/M113‐O/H40; 4.0
15	15. Biotin biosynthesis protein BioH	4ETW_A82 2.05 Å	W22	L18	W29	W81	A82a	CD1/L47‐O/H20; 3.5
16	16. Aclacinomycin methylesterase RdmC	1Q0R_A28 1.45 Å	G32	L28	W39	L101	S102	CE1/H59‐O/M30; 2.9
17	17. Carboxylesterase/lipase	4DIU_A83 2.00 Å	F24	L20	V31	L92	S93	OH/Y50‐O/H22; 3.0
18	18. Epoxide hydrolase	1QO7_A84 1.80 Å	W117	L113	F124	G191	D192	CD1/L149‐O/H115; 3.6
19	19. Haloperoxidase	1BRT_A85 1.50 Å	F32	L28	W39	F97	S98	CD/R58‐O/H30; 3.2
20	20. Thioesterases	1EI9_A86 2.25 Å	M41	I37	M51	F114	S115	CD1/I72‐CZ2/W38; 4.3 CD1/W38‐O/H39; 3.2
21	21. Carboxylesterase/ thioesterase 1	1FJ2_A87 1.50 Å	L25	F21	W32	F113	S114	CB/A51‐π/W67; 3.5 NE1/W67‐O/H23; 2.7
22	22. Ccg1/TafII250‐interacting factor B (Cib)	1IMJ_A88 2.20 Å	I41	L37	W48	P110	S111	CD2/L69‐O/H39; 4.4
23	23. A novel bacterial esterase	1QLW_A89 1.09 Å	C71	L67	W78	H205	S206	NE2/Q104‐O/H69; 3.0
24	24. Lipase	1JFR_A30 1.90 Å	F63	I59	I70	H130	S131	OG1/T89‐NH1/R99; 3.9 NH1/R99‐O/P61; 2.8
25	25. Fungal lipases	1TCA_A31 1.55 Å	T40	L36	F48	W104	S105	CG/P68‐SD/M72; 3.7 CE/M72‐O/P38; 3.2
26	26. Bacterial lipase	1ISP_A90 1.30 Å	I12	M8	F19	H76	S77	CE2/F41‐O/H10; 3.7
27	27. Pancreatic lipase, N‐terminal domain	1BU8_A91 1.80 Å	F77	I73	W85	H151	S152	NE1/W106‐O/H75; 3.0
28	28. Hydroxynitrile lyase‐like	3C6X_A92 1.05 Å	I12	L8	W19	E79	S80	CD1/L38‐O/H10; 3.9
29	29. Thioesterase domain of polypeptide, polyketide and fatty acid synthases	1JMK_C32 1.71 Å	V27	A23	Y34	Y79	S80	CB/F51‐π/F155; 4.3 CE1/F155‐O/P25; 3.3
30	30. Cutinase‐like	1BS9_A29 1.10 Å	T13	G9	S22	Y89	S90	OH/Y43‐O/R11; 2.7
31	31. YdeN‐like	1UXO_A93 1.80 Å	Y11	I7	F20	H70	S71	CG/M39‐π/W50; 3.5 CZ3/W50‐O/H9; 3.5
32	32. Putative serine hydrolase Ydr428c	1VKH_A94 1.85 Å	G38	Y34	F50	H109	S110	OH/Y73‐O/H36; 2.7
33	33. Acylamino‐acid‐releasing enzyme, C‐terminal domain	1VE6_A95 2.10 Å	G369	L365	F381	Y444	S445	OH/Y397‐O/H367; 2.8
34	34. Hypothetical esterase YJL068C	1PV1_A96 2.30 Å	L58	Y54	A65	H160	S161	CG2/T87‐O/S56; 3.7
35	35. Hypothetical protein VC1974	1R3D_A97 1.90 Å	L24	L20	W31	Y90	S91	CD1/L50‐O/H22; 3.5
36	36. Atu1826‐like	2I3D_A98 1.50 Å	H34	I30	V46	Y107	S108	CE2/F65‐O/H32; 3.4
37	37. PHB depolymerase‐like	2D80_A35 1.70 Å	C250	A246	F261	L38	S39	CB/A283‐π/W305; 4.0 NE1/W305‐O/H248; 2.9
38	38. IroE‐like	2GZS_A11 1.40 Å	NH1/R130	M88	V94	H188	S189	‐
40	39. O‐acetyltransferase	2B61_A99 1.65 Å	L49	I45	W65	G142	S143	CG1/V87‐OE1/Q148; 3.4 NE2/Q148‐O/H47; 2.9
41	40. 2,6‐dihydropseudo‐ oxynicotine hydrolase‐like	2JBW_A49 2.10 Å	L147	M143	S154	R216	S217	N/G173‐OE1/Q176; 2.8 NE2/Q176‐O/G145; 3.0

The plane of nucleophile and oxyanion zones in ABH families is sandwiched between aromatic residues. Three specific positions (Columns X_oxyII‐4, W/Y/F, X_nuc‐1), which lie close to the catalytic nucleophile (Column Snuc), and the Oxyanion I residue (Column X_oxyII), are frequently occupied by aromatic amino acids. An aromatic pair is often formed between the residue side chain (SC) that precedes or follows the position “Oxyanion zone II” and the conserved histidine of the oxyanion zone (Column SC/X_{ozII ± 1}–O/H_oxyII‐2). The “‐” sign designates exceptions, where the respective bond is not formed, or where there is no residue at that position that can interact with other residues of the aromatic cluster.

^aAs is shown in Table 1, in biotin biosynthesis protein BioH (SCOP Family 15), the nucleophile is serine,100, 101 and not alanine, as in the mutant with the PDB ID: 4ETW.82 This, however, does not interfere with the formation of the aromatic cluster.

On the opposite side of the β‐sheet we also observed an aromatic residue packing against the imidazole ring of the conserved histidine (His83 in EstFa_R) in 13 of 26 ABH families. The aromatic residue [Tyr115_{ozII ± 1} in EstFa_R, in Fig. 2(C)] is typically located either before or after Residue X_ozII (interaction distances: see Column SC/X_{ozII ± 1}–O/H_oxyII‐2 in Table 3).

In summary, the arrangement of the nucleophile and oxyanion zones and the conserved residues that interact and are located outside of the β‐sheet is such that an aromatic cluster can occur above the two planar zones and directly over the side chain of the conserved histidine (or the equivalent residue) of the oxyanion zone; an aromatic residue also tends to occur “below” the imidazole ring, opposite to and symmetrically to the aromatic cluster. The conservation of these residues clearly shows that the active site structure is extended outside the plane of the β‐sheet, where main elements of the catalytic mechanism are situated.

Discussion

The nucleophile zone and the oxyanion zone are planar, structural organizations that co‐ordinate the catalytic nucleophile and the two residues that form the oxyanion hole. Both zones are located at the catalytic site in the majority of ABH enzymes, while aromatic residues frequently surround the β‐sheet region where the key units of the catalytic machinery are situated. Here, we discuss residues of the zones that have structural and functional meaning for the ABH enzymes.

The nucleophile zone co‐ordinates the catalytic nucleophile and the oxyanion I residue, and aids the optimal arrangement of the oxyanion hole in most ABH enzymes

The nucleophile zone is situated in the core of the active site of ABH enzymes of the overwhelming majority of ABH enzymes, where it directly co‐ordinates the catalytic nucleophile residue and the Oxyanion I residue. Besides the co‐ordination of these two key units of the catalytic machinery, the weak hydrogen bonds of the zone (between Asp155 and Gly84; Gly158 and His83) bind the nucleophile elbow and the loop_β3‐αA. Consequently, it seems likely that the nucleophile zone is required for the optimal arrangement of the oxyanion hole adjacent to the catalytic nucleophile. This optimal arrangement is achieved by placement of the catalytic nucleophile and the Oxyanion I residues of the nucleophile elbow close to the Oxyanion II residue that follows the His–Gly dipeptide on loop_β3‐αA.

The nucleophile zone is only formed when the Oxyanion II residue is located two positions after the start of loop_β3‐αA. The importance of the nucleophile zone for the optimal geometry of the oxyanion hole is further confirmed in that only four ABH families lack important features of the nucleophile zone. In these four ABH families, the Oxyanion II residue is located at an alternate position and employs its side chain to assemble the oxyanion hole. The weak hydrogen bonds of the nucleophile zone are apparently not required for the formation of the oxyanion hole in these cases.

While analyzing structural consequences in the four ABH families that lack the nucleophile zone, we deduced that the residue at the start of loop_β3‐αA – equivalent to His83 in EstFa_R – is a determinant of the local geometry, and the type of amino acid and the side‐chain conformation of the residue at the start of loop_β3‐αA correlates with the existence of the nucleophile zone in the active site of ABH enzymes. The four exceptions fall into two separate cases. In the first case (PDB IDs: 3PUI,24 1H2W,25 1ORV26), the Oxyanion II residue is a tyrosine, which emerges from the start of loop_β3‐αA and provides its hydroxyl group to form the oxyanion hole. This tyrosine occupies the sequence and main‐chain structural position of the conserved histidine (His83 in EstFa_R). In the second case (PDB ID: 2GZS11), the residue that occurs at the start of loop_β3‐αA is an aspartic acid. The aspartic acid at this position helps form the oxyanion hole through a salt bridge with arginine – the oxyanion II residue – which is located on an α‐helix opposite to the active site and distant from the β‐sheet. Thus, in both cases where the nucleophile zone is not formed, the residue at the start of loop_β3‐αA is either the Oxyanion II residue or a residue that forms a salt bridge with it. However, in the remaining 36 ABH enzymes that have the nucleophile zone, the residue at the start of loop_β3‐αA is mostly a histidine, which is critical for the formation of the oxyanion hole, as we describe below.

In addition to the two weak hydrogen bonds important for the nucleophile zone, there are flanking residues (Gly154, Tyr82, an dGly85 in EstFa_R) that structurally support the rigidity of the nucleophile zone by interacting with residues of the zone and with each other. Both the principal and the peripheral contacts of the nucleophile zone are located in the area of the active site of the ABH enzymes and are shown in Figure 2(A). However, their logical division into main and flanking types is based on the role of the nucleophile zone. We consider that the two conserved weak hydrogen bonds of the nucleophile zone actively co‐ordinate the catalytic nucleophile and oxyanion I residues. In contrast, the other contacts, which are shown in Figure 2(A), are situated further away from the catalytic site. They seem to play supportive roles for the local supersecondary structure rather than a straightforward role in the co‐ordination of the key units of the catalytic machinery and the formation of the oxyanion hole. Our assumption for the supplementary structural role of the flanking elements is also supported by the high conservation of glycines at positions Sm_nuc‐2 and Sm_{nuc + 2}. Because of the lack of steric clashes and the existence of the two extra weak hydrogen bonds that can be formed by the CA atoms of both glycine residues [Fig. 2(A)], these residues may help enhance the local stability. With the β‐sheet contact N/Asp155–O/Tyr82 creating an imaginary line of separation, this sets a boundary between the elements of function [nucleophile zone; Fig. 2(A), right] and the flanking elements [Fig. 2(A), left].

The oxyanion zone is organized around a conserved histidine residue, which is relatively remote from the active site, but structurally important for the enzymatic function

In most ABH enzymes, the oxyanion zone of the active site is situated right next to the nucleophile zone. It includes seven structural elements, which are connected through an extended hydrogen‐bonding network that functions to stabilize elements of the catalytic machinery [Fig. 2(B)]. Among the elements of the oxyanion zone, we consider the conserved Residues His_oxyII‐2, Gly_oxyII‐1, and the acidic Residue X_ozII to be critical for the stability of the overall shape of loop_β3‐αA, that eventually leads to the optimal arrangement of the oxyanion hole. By being critical for the stability of loop_β3‐αA, we mean that the hydrogen‐bonding network of the oxyanion zone is committed to co‐ordinating the Oxyanion II residue next to Oxyanion I residue and the catalytic nucleophile, where the latter two residues are co‐ordinated by the nucleophile zone, as we discussed earlier. The structures of the nucleophile and oxyanion zones are linked through the shared dipeptide His_oxyII‐2–Gly_oxyII‐1 that participates in the contact networks of both zones. The association of the two zones through this connecting segment shows that the nucleophile and oxyanion zones have complementary roles in the overall arrangement of the structure of the catalytic site. However, their clear structural separation results in different levels of structural variability, where the nucleophile zone appears to be much more conserved than the oxyanion zone, that may lead to diverse catalytic activities among the ABH families.

His_oxyII‐2 appears to be the central organizing unit of the oxyanion zone. Half of the contacts of the zone are formed around its side chain. The imidazole ring is placed centrally in the area defined by loop_β3‐αA and interacts with the adjacent Residue X_oxyII‐1 and a neighboring residue from Strand β4. Other residues than histidine at the Position X_oxyII‐2 can play a similar structural role when their side‐chain points toward the center of loop_β3‐αA. For example, in our dataset, we have found that the hydrogen‐bonding network of the oxyanion zone is fully maintained when glutamine (PDB ID: 1ZI927), methionine (PDB ID: 1Q0R28), or arginine (PDB ID: 1BS929) appear at the start of loop_β3‐αA (Columns 4 and 5 in Table 2). A proline (PDB IDs: 1JFR,30 1TCA,31 1JMK32) at the same position partially maintains the bonds of the oxyanion zone. Outside the dataset, we observed another variation at this position, in the structure of cholesterol esterase from Candida cylindracea (PDB ID: 1LLF33, fungal lipases family): a phenylalanine occurs in the position of the conserved histidine and maintains all of the contacts of the oxyanion zone.

The conserved histidine at position X_oxyII‐2 was previously suggested to be part of the oxyanion hole, with the dipeptide His_oxyII‐2–Gly_oxyII‐1 termed the “oxyanion pocket”.34 This suggestion was later withdrawn because the histidine does not directly participate in catalysis.35, 36 Here, we show that the conserved histidine (or the equivalent residue) in the middle of oxyanion zone is indirectly involved in the ideal positioning of Oxyanion II to complete the shape of the oxyanion hole, due to the hydrogen bond with the adjacent residue that appears to stabilize the position of the Oxyanion II backbone amide.

The first four residues of loop_β3‐αA – the conserved histidine being the first residue – have been previously mentioned as a possible sequence identifier of different ABH families.37, 38, 39 The four‐residue loop segment is suggested to be involved in substrate binding40 and in the stabilization of the position of the putative hydrolytic water molecule.37 This indicates that the area around the oxyanion zone is actively involved in several aspects of catalytic activity, as mentioned above.

Multiple experimental studies confirm the importance of the conserved histidine near the oxyanion hole, and thus, the oxyanion zone in general. Site‐directed mutagenesis of histidine to alanine resulted in the complete loss of enzymatic function.19, 34, 41, 42 It is suggested that this mutation is responsible for the distortion of the shape of loop_β3‐αA and thus, the oxyanion hole fails to be formed.19, 34, 42 The mutation of histidine to leucine and asparagine had a similar adverse effect on enzymatic activity.43 Experimental studies have also been applied to those residues that have a similar side‐chain conformation as the conserved histidine. The mutation of glutamine to histidine improved the catalytic activity,44 while the mutation of arginine to alanine and lysine again resulted in a loss of function.45 Notably, site‐directed mutagenesis of the residue X_oxyII‐1 that follows the conserved histidine also leads to a loss of activity, for example, mutation of the glycine12, 46 or aspartate at this position.27 Taken together, these mutagenesis studies show that the Position X_oxyII‐2 should preferably be occupied by a histidine, since in many structures, mutating histidine leads to loss of function as does the adjacent residue; in contrast, mutating other residues to histidine can improve the enzymatic activity.

Thus, for the oxyanion zone, the residue at Position X_oxyII‐2 (His83 in EstFa_R) is critical for the geometry of loop_β3‐αA. In those ABH families that do not have the oxyanion zone, the type of amino acid at that position can be part of large‐scale variations of the structure of the active site that serve the enzymatic function,47, 48, 49 or its specific occurrence can contribute to the substrate binding,26 the regulation of pH optimum50 or the formation of the oxyanion hole.8, 9, 11, 51 Nonetheless, regardless the amino acid variation, the residue at Position X_oxyII‐2 also affects the enzymatic activity in the ABH enzymes without the oxyanion zone. Indeed, experimental studies have shown that site‐directed mutation of the residue at the start of loop_β3‐αA can decrease9 or eliminate the enzymatic activity entirely.8, 51

Residue X_ozII enhances the stability of the oxyanion zone

The acidic side chain of X_ozII faces the exterior of the protein, opposite to the conserved histidine that points toward the core of the protein. In many cases, Residue X_ozII is an aspartic acid, and its side‐chain oxygen atom forms a hydrogen bond with the amide of Residue X_ozI located near the end of loop_β3‐αA. This contact seems to stabilize the shape of loop_β3‐αA in the ABH enzymes with the oxyanion zone. Although Position X_ozII is located distantly from the catalytic site, site‐directed mutagenesis of the aspartic acid results in detrimental loss of activity, which suggests that this amino acid plays an important role in maintaining local folding and stability of the enzyme structure.41 A second mutational study showed that this conserved aspartic acid is actively involved in the formation of the protein–protein complex in the case of acyl esterase Aes from Escherichia coli.52

The aromatic residues around the nucleophile and oxyanion zones can affect the stability of the catalytic site

Besides the conservation of the nucleophile and oxyanion zones, we showed that aromatic residues regularly occur around both structural organizations. Particularly, we showed that aromatic residues often cluster over the nucleophile and oxyanion zones [Fig. 2(C)], while aromatic pairs are formed by the conserved histidine and a residue whose side chain lies opposite to the aromatic cluster with respect to the β‐sheet.

The conservation of aromatic residues in the periphery of the catalytic site can prove beneficial for the local structural stability. Several studies have previously analyzed the importance of aromatic‐aromatic interactions around the active site and shown their influence on the enzymatic42, 53, 54, 55, 56, 57 and the thermal58 stability of the protein, on the lining of the binding pocket25 and on the non‐local stabilization of the protein tertiary structure.59

We observed aromatic residues at three specific positions forming the aromatic cluster: the Residue X_nuc‐1 at the C‐terminus of strand β5, the Residue X_oxyII‐4 at the C‐terminus of strand β3 and a residue at the end of loop_β3‐αA or the start of αA‐helix. Our results indicate that the Residue X_nuc‐1, located before the nucleophile, is not a randomly occurring amino acid, but it is very often an aromatic residue. Experimental studies have shown that site‐directed mutation of Residue X_nuc‐1 results in the reduction60 or the elimination of the catalytic efficiency of the enzyme.61 It is suggested that the Residue X_nuc‐1 plays a structural role in the correct geometry of the catalytic site.60, 61, 62 Residue X_oxyII‐4 appears to be the most conserved of the aromatic cluster residues, as it is mostly a hydrophobic or else an aromatic residue.

The third residue of the aromatic cluster does not have a defined sequence position, because its position heavily depends on the length of loop_β3‐αA. Nonetheless, it is always linked with the Residues X_nuc‐1 and X_oxyII‐4. In the majority of ABH families, this position is occupied by an aromatic amino acid, quite often by a tryptophan residue, whose side chain has a characteristic, nearly perpendicular conformation over the β‐sheet, with the indole ring landing opposite to residue Sm_nuc‐2 of the nucleophile elbow. Interestingly enough, this structural phenomenon has been previously observed63 in a structurally similar part of the Clostridium MP flavodoxin64 structure, which also belongs to α/β‐fold class. Experimental studies, that refer to the third residue of the aromatic cluster, have showed that its site‐directed mutation can lead to decreased catalytic activity60, 65, 66 and affect the thermostability of the enzyme.66

Finally, the location of the aromatic cluster suggests that it would be involved in the co‐ordination of the catalytic histidine loop. Two separate studies have mentioned that Residue X_nuc‐1 62 and the third residue of the cluster42 form contacts with aromatic residues that occur at the loop fragment after the catalytic histidine, and thus participate in extended aromatic interactions near the catalytic histidine. This fact is rather interesting, because in our previous study20 we have shown that the catalytic acid zone residues co‐ordinate the loop fragment before the catalytic histidine.

Key residues surrounding the catalytic nucleophile retain their amino‐acid conservation, even in the absence of a canonical catalytic triad

Recently, the crystallographic structures of two ABH proteins were resolved, the Methylketone Synthase1 (MKS1) from Solanum habrochaites 67 and the aminoacrylate hydrolase RutD from Escherichia coli, 68 which have enzymatic activity and use a non‐canonical catalytic triad. MKS1 has an Ala–His–Asn triad in the place of the canonical catalytic triad and the residue Thr18 at the canonical Position X_oxyII‐1 is proposed to play the role of the putative, alternative nucleophile. RutD has a non‐canonical catalytic triad His–His–Asp, with its putative, alternative nucleophile His87 situated at the canonical Position X_nuc‐1. Given the exceptional catalytic mechanisms of MKS1 and RutD, we would like to examine the existence and conservation of the nucleophile zone, the oxyanion zone and the aromatic cluster in ABH enzymes with non‐canonical triads. We were able to identify eight ABH proteins (Entries B‐I in Table 4) that lack a canonical catalytic triad, two of which have a putative, alternative nucleophile (ABH enzymes MKS1 and RutD, Entries B and E respectively), while the rest lack some or all the catalytic residues (Entries C, D, F, G, H, I). These eight new ABH proteins have different degrees of activity, spanning from catalytically active to non‐enzymatic ABH proteins.

Table 4.

Relatedness of ABH Proteins Based on the Amino‐Acid Conservation of Five Key Positions of the Catalytic Site, Surrounding the Catalytic Nucleophile

N	SCOP Family #/ SCOP family name	PDB ID	XoxyII‐2	XoxyII‐1	W/Y/F	XozII	Xnuc‐1
I. “+++++” group
1	9. Gastric lipase	1HLG_A	H65	G66	W74	N98	H152
2	12. Haloalkane dehalogenase	1MJ5_A	H36	G37	W45	D62	H107
3	23. A novel bacterial esterase	1QLW_A	H69	G70	W78	D103	H205
4	26. Bacterial lipase	1ISP_A	H10	G11	F19	D40	H76
5	27. Pancreatic lipase, N‐terminal domain	1BU8_A	H75	G76	W85	D105	H151
6	31. YdeN‐like	1UXO_A	H9	G10	F20	N38	H70
7	32. Putative serine hydrolase Ydr428c	1VKH_A	H36	G37	F50	E72	H109
8	15. Biotin biosynthesis protein BioH	4ETW_A	H20	G21	W29	D46	W81
9	19. Haloperoxidase	1BRT_A	H30	G31	W39	D57	F97
10	21. Carboxylesterase/thioesterase 1	1FJ2_A	H23	G24	W32	H50	F113
11	33. Acylamino‐acid‐releasing enzyme, C‐terminal domain	1VE6_A	H367	G368	F381	N396	Y444
12	35. Hypothetical protein VC1974 3BE8	1R3D_A	H22	G23	W31	D49	Y90
II. “+ − +++” group
B	Methylketone synthase 1	3STT_A67	H17	T18	W26	D44	H86
III. “++++−” group
A	2. Carboxylesterase	3WJ2_A	H83	G84	Y95	E114	D155
13	1. Acetylcholinesterase‐like	1QE3_A	H104	G105	Y118	N135	E188
14	14. Carbon–carbon bond hydrolase	2OG1_A	H40	G41	Y52	D70	N111
15	16. Aclacinomycin methylesterase RdmC	1Q0R_A	M30	G31	W39	D58	L101
16	37. PHB depolymerase‐like	2D80_A	H248	G249	F261	Q282	L38
17	22. Ccg1/TafII250‐interacting factor B (Cib)	1IMJ_A	H39	G40	W48	D68	P110
18	18. Epoxide hydrolase	1QO7_A	H115	G116	F124	S148	G191
IV. “+ − ++−” group
19	28. Hydroxynitrile lyase‐like^	3C6X_A	H10	T11	W19	D37	E79
20	40. O‐acetyltransferase	2B61_A	H47	A48	W65	N86	G142
C	Protein NDRG2	2QMQ_A102	H69	D70	F79	D101	V145
V. “++ − ++” group
21	20. Thioesterases	1EI9_A	H39	G40	M51	D79	F114
VI. “+−−++” group
22	13. Dienelactone hydrolase	1ZI9_A	Q35	D36	M44	D62	Y122
23	30. Cutinase‐like	1BS9_A	R11	E12	S22	N42	Y89
24	36. Atu1826‐like	2I3D_A	H32	P33	V46	N64	Y107
VII. “++ − +−” group
25	2. Carboxylesterase	1LZL_A	H86	G87	S98	E117	Q159
26	10. Proline iminopeptidase‐like	1MTZ_A	H35	G36	L45	D62	S104
27	4. Hypothetical protein TT1662	1UFO_A	H31	G32	I40	D58	G112
D	Neuroligin‐4, X‐linked	3BE8_A103	H173	G174	I185	N202	S253
VIII. “++ − ‐‐” group
28	11. Acetyl xylan esterase‐like	1L7A_A	H89	G90	E98	L116	G180
29	17. Carboxylesterase/lipase	4DIU_A	H22	G23	V31	I49	L92
IX. “−++++” group
E	Putative aminoacrylate hydrolase RutD	3V48_A68	S20	G21	W29	D46	H87
30	25. Fungal lipases	1TCA_A	P38	G39	F48	S67	W104
X. “‐ − +++” group
31	29. Thioesterase domain of polypeptide, polyketide and fatty acid synthases	1JMK_C	P25	P26	Y34	D50	Y79
XI. “‐ − ++−” group
32	5. PepX catalytic domain‐like	3PUI_A	N42	P43	W52	D73	V116
F	39. TTHA1544‐like	2DST_A70	A29	E30	W35	D49	A87
XII. “− + −++” group
33	34. Hypothetical esterase YJL068C	1PV1_A	S56	G57	A65	D86	H160
34	24. Lipase	1JFR_A	P61	G62	I70	D88	H130
G	UPF0255 protein FrsA	4I4C_A104	A200	G201	M209	D228	F271
XIII. “‐‐ − ++” group
35	7. DPP4 catalytic domain‐like	1ORV_A	Y547	A548	K554	D579	W629
XIV. “‐‐ − +−” group
H	Inactive dipeptidyl peptidase 10 (DPP10)	4WJL_A69	D568	E569	L575	D600	K650
XV. “− + − + −” group
36	6. Prolyl oligopeptidase, C‐terminal domain	1H2W_A	Y473	G474	I480	N503	G553
37	8. Serine carboxypeptidase‐like	3SC2_A	N51	G52	S58	D94	E145
38	41. 2,6‐dihydropseudo‐oxynicotine hydrolase‐like	2JBW_A	G145	G146	S154	D172	R216
XVI. “−++ − ‐” group
39	3. Mycobacterial antigens	1DQZ_A	D38	G39	W49	V70	L123
XVII. “‐ − + − ‐” group
I	MPT51/MPB51 antigen	1R88_A105	D66	A67	W77	A98	A144
XVIII. “‐‐‐ − +” group
40	38. IroE‐like	2GZS_A	‐	‐	V94	G118	H188

The active site of ABH proteins is surrounded by five key residues from three structural organizations: the nucleophile zone (Columns X_nuc‐1, X_oxyII‐2, and X_oxyII‐1), the oxyanion zone (Columns X_oxyII‐2, X_oxyII‐1, and X_ozII), and the aromatic cluster (Columns W/Y/F and X_nuc‐1). ABH proteins with canonical catalytic triads (Entries 1–40, A) and ABH proteins with non‐canonical catalytic triads (Entries B‐I) are classified in 18 groups (Groups I–XVIII), based on the amino‐acid type present at the five key positions. Every group is labeled by a ± five‐character sequence, which corresponds to the sequence X_oxyII‐2 − X_oxyII‐1− W/Y/F − X_ozII − X_nuc‐1. Each character of the sequence is marked with “+”, if the corresponding position of ABH protein is occupied by the most conserved amino‐acid according to the scheme (±H)_oxyII‐2 − G_oxyII‐1 − (±Φ)_W/Y/F−(±D/E/N/Q)_ozII−(±H/Φ)_nuc‐1 (“Φ” designates the presence of an aromatic amino acid). Alternatively, the character is marked with “‐”, if this position is occupied by a non‐conserved residue. Thus, Group I (“+++++”) gathers all those ABH proteins that retain the amino‐acid conservation in all five key residues, while Group XIV (“‐‐ − +−”), Group XVII (“‐ − + − ‐”), and Group XVIII (“‐‐‐ − +”) retain the amino‐acid conservation in none‐but‐one position. In the five columns that contain the amino‐acid sets, the “‐” sign designates that the respective position is not occupied by any residue.

To compare the ABH enzymes that have a canonical catalytic triad and the eight exceptional cases that have a non‐canonical catalytic triad with respect to our findings, we have singled‐out five key residues from the nucleophile zone, the oxyanion zone and the aromatic cluster in order to assess how similar or different is the structural environment around the catalytic nucleophile in canonical and non‐canonical settings. We have selected the nucleophile zone Residues X_nuc‐1, X_oxyII‐2, X_oxyII‐1 (from which the first residue is shared with the aromatic cluster, and the latter two are shared with the oxyanion zone), the oxyanion zone Residue X_ozII and the third residue of the aromatic cluster, based on their proximity to the catalytic nucleophile, the experimental data that show their effect in the enzymatic activity, and their high conservation rate. These five residues constitute the structural core around the active site of the ABH enzymes. As we have shown in the Results section, each of these five positions is occupied by conserved and specific amino‐acids: Residue X_nuc‐1 is often aromatic (Trp, Tyr, Phe, and His); Residue X_oxyII‐2 is usually a histidine or an amino‐acid with a similar side‐chain conformation; Residue X_oxyII‐1 is mostly a glycine; Residue X_ozII has an acidic (Asp, Glu) or an amidic side chain (Asn, Gln); and the third residue of the aromatic cluster is clearly aromatic (Trp, Tyr, Phe).

Using the five key positions shown above, we have analyzed the relatedness of ABH families that have canonical catalytic triads (SCOP Families #1–38, 40, 41 in Table 4) with those that have non‐canonical catalytic triads (Entries B‐I in Table 4). It appears that the majority of ABH families that have both canonical and non‐canonical catalytic triads retain the amino‐acid conservation in all (Group I in Table 4) or all‐but‐one (Groups III, V, and IX in Table IV) five key positions; Groups I and III are the largest categories, with the ABH enzymes of the two groups differing in the amino‐acid occupancy of Position X_nuc‐1. The groups that retain the amino‐acid conservation in two or three positions are generally poorly populated (Groups IV, VI, VII, VIII, X, XI, XII, XIII, XV, and XVI in Table 4).

Based on the classification that arose from the comparison of ABH enzymes with a canonical catalytic triad, five of the eight ABH proteins with non‐canonical catalytic triad (Entries C, D, E, F, and G in Table 4) fall in the same categories with ABH enzymes that have a canonical catalytic triad (Groups IV, VII, IX, XI, and XII in Table 4, respectively). Three of the eight additional ABH proteins (Entries B, H, and I) make up new, single‐entry categories (Groups II, XIV, and XVII, respectively). Group II is similar to Group I, while Groups XIV and XVII are very different from Group I with respect to the amino acids that occupy the five key positions. Both Entries H and I, however, are enzymatically inactive proteins that contain incomplete catalytic triads (Table 4). In the case of Entry H, the catalytic triad appears almost intact, except there is a glycine, which occupies the canonical position of the catalytic nucleophile.69 Nonetheless, the mutation of this glycine to serine failed to induce any enzymatic activity, probably due the long side‐chain amino acids (Asp568 and Glu569) at the key Positions X_oxyII‐2 and X_oxyII‐1, that apparently hinder both the ligand binding and optimal formation of the oxyanion hole.69

Thus, while there are ABH fold proteins that have non‐conventional catalytic triads, with some of them being enzymatically active (Entry B in Table 4) and others lacking any catalytic activity (Entries H and I in Table 4), then based on our analysis and grouping, we can conclude that (1) these non‐canonical proteins still have largely the same structural cores as the ABH enzymes with the canonical catalytic triads; and (2) the more the structural core of a protein with a non‐canonical catalytic triad differs from that of Group I (Table 4), the less likely it will retain any catalytic activity.

Conclusions

In this study, we analyzed the structure around the nucleophile elbow and the oxyanion hole in representative structures from 40 ABH families. We have shown that the nucleophile zone, the oxyanion zone and the aromatic cluster comprise a three‐dimensional structural organization that shapes the active site of ABH enzymes and plays an important role in the enzymatic function by structurally stabilizing the catalytic nucleophile and the residues of the oxyanion hole. This finding is complementary to the result from our previous study, where we showed that there is a common structural organization that co‐ordinates the catalytic acid residue and links it with the catalytic histidine. Overall, we were able to confirm that there is a common supporting architecture around each key unit of the catalytic machinery.

Undeniably, Position X_oxyII‐2 at the start of loop_β3‐αA is paramount for the local structure of the catalytic site and the activity of ABH enzymes, and is shared by both the nucleophile and the oxyanion zones. Regardless the type of amino acid, position X_oxyII‐2 is typically occupied by a residue that is suitable for the optimal geometry of the oxyanion hole, either indirectly by stabilizing the arrangement of the oxyanion hole, or directly by providing a side‐chain group to form the oxyanion hole. In addition to that, a conserved acidic residue at Position X_ozII of the oxyanion zone proves to be essential for the formation of the oxyanion hole, although it is located remote from the active site.

Finally, we showed that aromatic residues tend to appear around the nucleophile and the oxyanion zones in most ABH families, often forming an aromatic cluster over and an aromatic pair below the plane of the nucleophile and oxyanion zones. Aromatic residues are shown to enhance the stability of the active site, and thus, we consider that they also affect the stability of the ABH enzymes. The aromatic cluster, which is located close to the active site, connects the nucleophile and the oxyanion zones through interactions between residues that belong to both zones. In addition, there is evidence that the aromatic cluster can be involved in the co‐ordination of the catalytic histidine loop. Further experimental studies on the conserved aromatic residues may provide valuable insight about their role in the enzymatic activity.

Materials and Methods

We have used the Structural Classification Of Proteins (SCOP) database22 to choose representative structures for our dataset. There are 41 ABH families according to SCOP classification, which contain 128 different entries of proteins with the ABH fold. The single entry from the TTHA1544‐like family (SCOP Family #39) is not included in our dataset because it is not a hydrolase.70 From the 127 entries, we have selected 40 representative structures ensuring that they represent each listed ABH family, on the condition that they are mainly unligated and have the highest resolution. In total, we have formed a dataset, which contains 40 structures plus the carboxylestarase EstFa_R from the study of Ohara et al.23 that we use as the reference structure throughout this study. We have obtained all protein structures from Protein Data Bank (PDB).21

The choice of structures to be analyzed: The results of this study are presented in four tables. Each table has a different number of entries, depending on the subset of representative structures that have each conserved structural motif. Specifically, Table 1 describes 40 representative structures from 40 ABH enzyme families (Rows 1–40) plus the carboxylesterase EstFa_R (Row A), with the exception of the single‐entry protein from the TTHA1544‐like family. In Table 1, we show not only the 36 ABH structures and EstFa_R that maintain the contact network of the nucleophile zone but also the four ABH structures that lack all interactions of the nucleophile zone. Table 2 only includes those structures that maintain the contact network of the oxyanion zone and specifically, 29 representative structures from the 40 ABH enzyme families (Rows 1–29) plus the carboxylesterase EstFa_R (Row A). The protein from SCOP Family TTHA1544‐like is not taken into account here as well. Table 3 includes all 40 representative structures (Rows 1–40) from the 40 ABH families (except the TTHA1544‐like family) and the carboxylesterase EstFa_R (Row A).

“Conserved amino acid residues” and “conserved interactions and bonds”: Throughout the text we use this terminology to show amino acids and bonds, which are repeatedly found at the same structural positions and places among many families of ABH proteins. For example, structurally, the oxyanion zone is formed around a conserved histidine residue at the start of loopβ3‐αA, and exactly two residues before XoxyII in 26 of 40 ABH fold families (Columns 4 and 5, Table 2). If two structurally conserved residues or water molecules make an interaction or form a bond, which is also observed in many different ABH families, we call it a conserved bond or interaction. A Structural Motif, common to many ABH protein families, is essentially a set of structurally conserved amino acids joined by similar interactions in a similar way.

For all structural analysis, such as hydrogen bonds, hydrophobic or other types of weak interactions, we have used the BIOVIA (Accelrys) Discovery Studio71 (http://accelrys.com/products/collaborativescience/bioviadiscoverystudio/). We have identified the weak hydrogen bonds from CHO contacts in structures with at least 2.0 Å resolution based on geometrical criteria given in (72) and with distances C…O ≤ 4.1 Å and H…O ≤ 3.0 Å, and the weak hydrogen bonds from CH‐π contacts with distance between 3.4 Å and 6 Å.57

Lastly, for the visualization and analysis of the structural data, we have used Discovery Studio71 and Bodil.73 Figures are drawn with MolScript.74

Supporting information

Appendix S1: Supporting information