Carboxylic Ester Hydrolases in Bacteria: Active Site, Structure, Function and Application

Abstract

Carboxylic ester hydrolases (CEHs), which catalyze the hydrolysis of carboxylic esters to produce alcohol and acid, are identified in three domains of life. In the Protein Data Bank (PDB), 136 crystal structures of bacterial CEHs (424 PDB codes) from 52 genera and metagenome have been reported. In this review, we categorize these structures based on catalytic machinery, structure and substrate specificity to provide a comprehensive understanding of the bacterial CEHs. CEHs use Ser, Asp or water as a nucleophile to drive diverse catalytic machinery. The α/β/α sandwich architecture is most frequently found in CEHs, but 3-solenoid, β-barrel, up-down bundle, α/β/β/α 4-layer sandwich, 6 or 7 propeller and α/β barrel architectures are also found in these CEHs. Most are substrate-specific to various esters with types of head group and lengths of the acyl chain, but some CEHs exhibit peptidase or lactamase activities. CEHs are widely used in industrial applications, and are the objects of research in structure- or mutation-based protein engineering. Structural studies of CEHs are still necessary for understanding their biological roles, identifying their structure-based functions and structure-based engineering and their potential industrial applications.

1. Introduction

Carboxylic ester hydrolases (CEHs, EC 3.1.1.-) are catalysts that hydrolyze linear and cyclic carboxylic ester bonds to produce carboxyl groups (–COOH) and alcohol groups (–OH) at termini. CEHs are found in all living organisms, including vertebrates, insects, fungi, plants, archaea and bacteria. Hydrolysis by CEHs is important for metabolite regulation [1], signal transduction [2], protein synthesis [3], stem elongation [4] and thermal stress response [5]. Offensive and defensive interactions between insects and plants are also mediated by CEHs [6]. Bacteria use CEHs to demolish cell walls and membrane structures for food uptake [7], and to infect hosts [8,9,10]. The functional diversity of CEHs is mediated by substrate specificities on various biomolecules, such as carbohydrates [11], lipids [12], polypeptides [13,14], nucleic acids [15] and other small molecules [16]. Their catalytic reaction is followed by the cooperation of catalytic residues including a classical Ser-His-Asp triad and substrate-binding residues.

In this review, we classified 136 bacterial CEHs deposited in the Protein Data Bank (PDB) [17] based on the source, their Enzyme Commission (EC) number, substrate, localization, active site structure, physiological function and three-dimensional (3D) structure, in order to provide a general overview of bacterial CEHs. In addition, their potency for industrial application is discussed.

2. Sampling of Bacterial CEH Structures in the Protein Data Bank

The structures of 525 bacterial enzymes belonging to the EC 3.1.1.- group were released in the PDB database until 2019 July. The 518 structures were identified using X-ray crystallography (XRC), and seven structures were identified using nuclear magnetic resonance (NMR). Here we only focused upon crystal structures. According to the PDB description (www.RCSB.org), EC numbers are assigned based on UniProtKB, GenBank, the Kyoto Encyclopedia of Genes and Genomes (KEGG), and authors’ descriptions. Among 518 structures, 91 of them, such as metallo-β-lactamase, have been excluded, because they were reported to have other enzymatic roles that did not describe CEH activity. Theoretical structures (e.g., 1A20) are not analyzed. CEHs not annotated with an EC number, such as carbohydrate esterase SmAcE1, are also beyond the scope of this review, although they do have esterase activity [18]. As a result, 424 structures of 136 CEHs from 52 genera and metagenomic (uncultured and unidentified) samples were selected for the analysis, and 136 CEHs were classified based on EC numbers (Table 1). When several crystal structures were available in one enzyme (several PDB codes), the first reported structure was used as a CEH structure. Their sequences containing tag or mutation were analyzed after being recovered to wild type sequences. Exceptionally, the sequences were not recovered when the mutations in the active sites of CEHs induced a gain of function. The majority of the crystal structures were identified in 1.0–3.0 Å resolution, and even extremely high resolution (below 1 Å) cases were also reported (Figure 1A). The major bacterial sources of CEHs in this analysis are Pseudomonas (9.6%), Bacillus (8.8%) and Geobacillus (7.4%), as shown in Figure 1B. Ten CEHs (7.4%) originate in metagenomic samples.

3. Classification of CEHs Based on Substrates

CEHs can be categorized based on substrates, which can be identified by EC number as shown in Table 1. CEHs using lipids as substrates are phospholipases (only PLA₁ [EC 3.1.1.32] and PLA₂ [EC 3.1.1.4]), carboxylesterases (EC 3.1.1.1), lysophospholipases (EC 3.1.1.5) and triacylglycerol lipases (also called true lipases [EC 3.1.1.3]). However, bacterial lipolytic CEHs are not sensitive to alcohol-group substrates, but to acyl chain length. Carboxylesterase (EC 3.1.1.1) also cleaves carboxylic esters, but the length of the acyl chain is much shorter than the substrates of lipases. Enzyme kinetics and lid structure can also be used to distinguish between lipase and carboxylesterase [178]. In general, lipase contains a lid that covers the active site. However, these approaches are controversial, as some carboxylesterases, such as a lid-containing carboxylesterase, contain a lid similar to those found in lipases [72,179,180].

CEHs that use carbohydrates as substrates are called carbohydrate esterases [181,182]. Carbohydrate esterases that are active on xylan, cutin, and pectin are known as acetylxylan esterase (EC 3.1.1.72), cutinase (EC 3.1.1.74) and pectinesterase (EC 3.1.1.11), respectively, and the acyl chain, most often a member of the acetyl group, is removable in the monomeric and polymeric forms of carbohydrates.

4. Classification Based on Localization

CEHs are distributed from extracellular to cytosolic regions. Ninety of 136 CEHs are localized in the cytosolic region, 43 CEHs have signal peptides for secretion and the remaining three CEHs are transmembrane proteins (Figure 2). Outer membrane phospholipase As (OMPLAs) from Escherichia coli (representative PDB code: 1FW2) [107,108,109] and Salmonella typhi (PDB code: 5DQX) span membranes. Autotransporter EstA from Pseudomonas aeruginosa (PDB code: 3KVN) is another outer membrane-spanning protein [50]. Representatively, pectin methylesterase from Dickeya dadantii (PDB code: 2NSP) with a signal sequence at the N-termini, is a representative secretary CEH for the bacterial invasion of plant tissues [115]. LipA from Xanthomonas oryzae (representative PDB code: 3H2G) [28,183], lipase from Geobacillus zalihae (PDB code: 2DSN) [80] and phospholipase A₂ from Streptomyces violaceoruber (PDB code: 1LWB) [184] have been physiologically verified as secreted proteins, as they can be isolated from culture media.

5. Classification Based on the Active Site Residues

Ser-His-Asp in the catalytic triad, which works as a nucleophile, a base and an acid, respectively, is necessary for hydrolysis [189]. In addition to the conventional catalytic triad (Ser-His-Asp), various types of nonconventional triads and dyads have been reported [190]. Gariev et al. classified hydrolases based on components in active sites to produce hierarchical four-digit layers and a web-based database (http://www.enzyme.chem.msu.ru/hcs) [191]. In the hydrolysis reaction, O_γ in Ser, S_γ in Cys, O_γ1 in Thr, O_δ in Asp and O in the water molecule are the nucleophiles, attacking carbonyl carbon in carboxylic ester bonds. The base, usually His residue, deprotonates the nucleophile, and increases the activity of these nucleophiles. The acid stabilizes the position of the base, and assists the function of base to the nucleophile. Along with catalytic residues, the oxyanion hole plays a key role in stabilizing transition states. CEHs can be classified into several groups based on consensus sequences encompassing their active site residues. Here, we divide CEHs into four groups (groups 1 to 4), based on catalytic residues. Each group is divided into sub-groups according to motifs and conserved residues in the catalytic domain. The information of key residues in each group is provided using the representative structures in Figure 3.

5.1. Ser Hydrolases (Group 1)

A sequence analysis of CEHs revealed that many CEHs contain the catalytic triad Ser-His-Asp. We defined group 1 CEHs as those containing the catalytic triad with Ser serving as a nucleophile. Based on motifs containing catalytic Ser, CEHs can be classified into several groups, such as GXSXG, GDSX, SXXK and YTQ/HXSNG groups (underlined residues are nucleophiles). Among them, the GXSXG group is the most common, containing 88 esterases among 136 CEHs.

• Group 1-1

Group 1-1 is the most abundant CEH, and contains the GXSXG motif, along with the AXSXG and GXSXXG variants. The GXSXG motif is localized in the loop region, and forms a catalytic triad with Asp and His in other loops in the C-terminal region. The GXSXXG motif is found in glucuronoyl esterase from Solibacter usitatus (PDB code: 6GRY) [42], carbohydrate esterase 15 from a marine metagenome (PDB code: 6EHN) [41], cocaine esterase from Rhodococcus spp. (representative PDB code: 3I2K) [29], and alpha-amino acid ester hydrolases from Acetobacter pasteurianus (representative PDB code: 2B9V) [148] and Xanthomonas citri (PDB code: 1MPX) [147].

Artificial dienelactone hydrolases were obtained through protein engineering, including the mutation of C123S in the GXCXG motif of carboxymethylenebutenolidase from Pseudomonas knackmussii (representative PDB code: 4U2B) [153] and from Pseudomonas putida (representative PDB codes: 1ZI8) [151]. Introducing the GXSXG motif reportedly enables the production of an artificial dienelactone hydrolase [193]. The other two triad components are most frequently identified as His-Asp by order in CEHs with GXSXG motifs. Exceptionally, Glu is positioned instead of Asp in the following six CEHs: naproxen esterase from Bacillus carboxylesterases cleaving naproxen ester (PDB code: 4CCW) [53], carboxylesterase CesB from Bacillus sp (PDB code: 4CCY) [53], Est1 from Hungatella hathewayi (PDB code: 5A2G) [37], pNB esterase from Bacillus subtilis (PDB code: 1C7I) [19], a putative carboxylesterase from B. subtilis (PDB code: 2R11) and metagenomic Est5 (PDB code: 3FAK) [26]. Their catalytic triad therefore is composed of Ser-His-Glu. In this group, the GGGX, GX and Y motifs, which are located mostly in the N-terminal region of a CEH, are involved in forming the oxyanion hole [1,194]. In the GGGX motif, most oxyanion hole components are positioned at the second Gly and third Gly residues. In CEHs containing the GX or Y motifs, with residue X in the GX motif or Tyr in the Y motif, an oxyanion hole forms with the second X in the active site GXSXG motif. In addition, GGAX (representative PDB code: 4V2I [58] and 3DOH [49]) and GAGX (representative PDB code: 1C7I [19], 5A2G [37] and 4C89 [52]) motifs have also been also reported. Y-motif-containing CEHs have been reported in amino acid ester hydrolases (PDB code: 2B9V [148] and 1MPX [147]), and cocaine esterase (representative PDB code: 3I2K [29]). Catalytic His and Asp/Glu are typically positioned with the 20–30 amino acid gap in the order of Asp-His. However, in chemotaxis methylesterase (CheB) from Salmonella typhimurium (PDB code: 1CHD), catalytic His190 and D286 are positioned in the reverse order, with 95 unique amino acid gaps [13]. Important residues in this group are shown in Figure 3A.

• Group 1-2

Group 1-2 includes the GDSX motif-containing CEHs (called the GDSL family), in which catalytic Ser is localized close to the N-terminus in a hydrolase domain [196]. According to previous analysis, in the GDSL family, sequence consensus blocks (Block I, II, III and V) contain the functionally important residues Ser, Gly, Asn and His, and thus named SGNH hydrolases, as shown in Figure 4. Catalytic Ser is found in Block I, the oxyanion hole components Gly and Asn are located in Blocks II and III, and a general base known as His exists in Block V (Each residue is marked with an asterisk in Figure 4). The general base His and a general acid Asp form a DXXH motif near the C-terminus of the hydrolase domains. As a rare group, xylan esterase from Cellvibrio japonicus (CjCE2A, PDB code: 2WAA) has Asp789 and His791 in a DXH motif (Figure 4) [22].

Moreover, catalytic dyads lacking Asp in the DXXH motif are also identified in lipase from Streptomyces rimosus (PDB code: 5MAL) [39] and esterase from Streptomyces scabies (PDB code: 1ESC) as shown in Figure 4 [20]. In their structures, Asp residues in the catalytic triad are replaced by nonfunctional Asn in the S. rimosus lipase and Trp in the S. scabies esterase, in which both Asn and Trp only stabilize the orientation of the catalytic His instead of playing a role as acids. In phospholipase A₁ from S. albidoflavus (PDB code: 4HYQ), its sequence shows a conserved DXXH motif of Block V, but the 3D structure reveals that the Ser-His dyads form because of the position of Asp in the DXXH motif is not proper for its function as a general acid [141]. In oxyanion hole formation, aryl esterase from Mycobacterium smegmatis (representative PDB code: 2Q0Q) has Ala instead of Gly in Block II, but its function is similar [65]. Important residues in this group are shown in Figure 3B.

• Group 1-3

In group 1-3, the SXXK motif is commonly found in peptidases including β-lactamase [197], and also in family VIII lipases [198,199]. The lipases in family VIII have β-lactamase-like structures, but they usually have carboxylic esterase activity, not β-lactamase activity. Ser71, Lys74 and Tyr160 compose catalytic triads in the EstSRT1, family VIII lipase from the metagenome (PDB code: 5GMX), as in Figure 3C [61]. Another family VIII lipase, EstU1 from the metagenome (PDB code: 4IVK), also has the SXXK motif in which Ser64 and Lys67 form a catalytic triad with Tyr150 [55].

• Group 1-4

Group 1-4 CEHs contain YTQ and HXSNG motifs. OMPLAs from E. coli (representative PDB code: 1FW2) [108] and from S. typhi (PDB code: 5DQX) belong to this group. The YTQ motif is essential for dimerization of OMPLAs in the membrane, and the HXSNG motif is critical for hydrolase activity [200,201]. In E. coli OMPLA, His142 and Ser144 in the HXSNG motif compose a catalytic triad with Asn156, and consecutive Asn145 and Gly146 of the motif are components of an oxyanion hole, as shown in Figure 3D. S. typhi OMPLA also has a Ser164-His162-Asn176 catalytic triad and an oxyanion hole formed by Asn165 and Gly166 with the YTQ motif (residues 112–114).

5.2. Aspartyl Hydrolases (Group 2)

Group 2 CEHs are aspartyl hydrolases containing an Asp-Asp dyad with a nucleophilic Asp and a basic Asp, as described in Figure 3E. Epoxide hydrolases [202], and glycosyl hydrolases [203] belong to Asp hydrolases. The Asp-Asp catalytic dyad is also identified in pectin methylesterases (orange squares in Figure 2). Pectin methylesterase A proteins from Dickeya chrysanthemi (representative PDB code: 1QJV) [114] and D. dadanti (PDB code: 2NSP) [115] contain the GXSXXG motif, although Ser in this motif does not work as a nucleophile. In these enzymes, commonly, Asp199-Asp178 form a catalytic dyad, Gln177 is the oxyanion hole-forming residue, and Arg267 and Trp269 are involved in pectin binding [114,115]. The hydrolysis reaction proceeds metal-independently, and without a nucleophilic water molecule. In a similar manner, in pectin methylesterase (or carbohydrate esterase family VIII) from Yersinia enterocolitica (PDB code: 3UW0), Asp199 and Asp177 work as a nucleophile and a general acid/base, respectively [116]. In this enzyme, Arg264 and Trp266 are functionally conserved as a pectin-binding site, and Gln176 belongs to an oxyanion hole [116].

5.3. Metal-independent Hydrolase with a Nucleophilic Water (Group 3)

Group 3 CEHs are nonconventional hydrolases, lacking typical nucleophilic residues such as Ser/Thr/Cys, but containing a water molecule that functions as a nucleophile. The water molecule is activated by general base His (Group 3-1, 3-3 and 3-4), Asp (Group 3-5) and Gln (Group 3-2) residues without metal coordination or the assistance of other cofactors.

• Group 3-1

In the peptidyl-tRNA hydrolase from Vibrio cholerae (representative PDB code: 4Z86) belonging to Group 3-1 CEHs, Asn14, His24, Asn72, Asp97 and Asn118 are functionally important [136]. His24 and Asp97 work as a general base and an acid, respectively. Asn72 and Asn118 form an oxyanion hole, and two Asp residues at 14 and 118 stabilize tRNA-binding in its active site. As described in Figure 5, catalytic His-Asp (denoted by an asterisk), two oxyanion hole-forming Asn residues (O-marked) and Asn involved in substrate binding (S-marked) are also conserved in other peptidyl-tRNA hydrolases, such as those from Mycobacterium tuberculosis (PDB code: 2Z2I) [124], M. smegmatis (PDB code: 3P2J), E. coli (PDB code: 3VJR, 2PTH, and 3OFV) [123], S. typhimurium (PDB code: 4P7B) [132], Acinetobacter baumannii (PDB code: 4LWQ), Pseudomonas aeruginosa (PDB code: 4FYJ) [131], Francisella tularensis (PDB code: 3NEA) [125], Streptococcus pyogenes (PDB code: 4QT4) [133], Staphylococcus aureus (PDB code: 4YLY) [135] and Burkholderia thailandensis (PDB code: 3V2I) [127]. In the Thermus thermophilus peptidyl-tRNA hydrolase (PDB code: 5ZX8), Arg103 plays a role in substrate binding and oxyanion formation similar to Asn118 in V. cholerae (Figure 5) [138]. Important residues in this group are shown in Figure 3F.

• Group 3-2

Another type of peptidyl-tRNA hydrolases belonging to the RF-1 family is defined as group 3-2 CEHs in this review (Figure 3G). A common feature of RF-1 family peptidyl-tRNA hydrolase is a GGQ motif. Gln in this motif works as a base that stabilizes nucleophilic water, and also plays a key role in interactions with tRNA [204]. The O_ε1 of Gln28 in YaeJ, a peptidyl-tRNA hydrolase from E. coli (PDB code: 4V95) and the nucleophilic water molecule, form a hydrogen bond [134]. The water molecule attacks the carbonyl carbon of the carboxylic ester bond between A76 in tRNA and the peptide.

• Group 3-3

Secreted phospholipase A₂, defined as a group 3-3 CEH, is considered a metal-independent hydrolase using His-Asp as a base-and-acid dyad in the catalytic site to stabilize the nucleophilic water, as shown in Figure 3H [205]. The secreted PLA₂ from S. violaceoruber (PDB code: 1LWB) uses His64 and Asp85 as residues in the catalytic dyad and Tyr68 to encourage substrate-binding [106]. In the Ca²⁺-free form of PLA₂ (PDB code: 1LWB), W260 is the inferred nucleophilic water, attacking sn-1 carbonyl carbon. In the Ca²⁺-binding form of PLA₂ (PDB code: 1KP4), the calcium ion, coordinated by O_δ1 of Asp43, O of Leu44, O_δ2 of Asp65 and three water molecules (W201, W202, and W203), induces a hydrogen bond network in a substrate binding pocket. In this enzyme, the water molecule (W256) is regarded as a nucleophile and attacks the sn-2 carbonyl carbon of the substrate [105,184].

• Group 3-4

6-phosphogluconolactonases from M. smegmatis (representative PDB code: 3OC6) [140] belonging to group 3-4 CEHs contains His151-Glu149 as a catalytic dyad with nucleophilic water (Figure 3I). Similarly, 6-phosphogluconolactonases from T. tuberclosis (PDB code:3ICO) [139] contains His152-Glu150 as a catalytic dyad, and Thermotoga maritima 6-phosphogluconolactonase (PDB code: 1VL1) contains the conserved dyad His138-Asp136.

• Group 3-5

Unlike hydrolases described above, the Asp residue of the enzyme belonging to group 3-5 CEH works as a general base on nucleophilic water (Figure 3J). For example, LigI from Sphingomonas paucimobilis (PDB code: 4D8L, light green with a black outline in Figure 2) belonging to the amidohydrolase superfamily has lactonase activity using Asp248 as a catalytic base without metal–ion coordination [154]. The water molecule forming a hydrogen bond with Asp248 attacks carbonyl carbon in the lactone group as a nucleophile. His31 and His180 form an oxyanion hole, and His33 contributes to the lactonase reaction by stabilizing the tetrahedral intermediate.

5.4. Metal-dependent Metallohydrolases (Group 4)

• Group 4-1

In the group 4-1 CEH, the HXHXDH motif is involved in Zn²⁺-binding [206], and His and Asp residues in this motif mainly coordinate the metal ion that stabilizes a nucleophilic water. In N-acyl homoserine lactone hydrolase from Bacillus thuringiensis (PDB code: 2A7M) [21,166], the HXHXDH motif composed of His104-X-His106-X-Asp108-His109 coordinates two zinc ions with additional Asp191, His169 and His235, as in Figure 3K. The first Zn²⁺ is coordinated by His104, His106 and His169, and the second Zn²⁺ is coordinated by Asp108, His109, Asp191 and His235. According to a known mechanism, one water molecule coordinated by two zinc ions works as a nucleophile, and Tyr194 works as an acid in hydrolysis [24,164]. In 4-pyridoxolactonase from Mesorhizobium loti (PDB code: 4KEP), His96, His98, Asp100 and His101 are HXHXDH motif components. Asp100, His101, Asp207 and His252 coordinate the first Zn²⁺, and His96, His98, His185 and Asp207 coordinate the second Zn²⁺. In contrast to the HXHXDH-containing group 4-1 CEHs, lactonase UlaG from E. coli (representative PDB code: 2WYM) has only one Mn²⁺ at the second Zn²⁺ position, although it has the HXHXDH motif (Figure 6) [23]. HXHXDH motif-containing CEHs are highlighted using the pale blueish-green diamond in Figure 2.

• Group 4-2

This group of CEHs does not contain an HXHXDH motif. For example, the de-O-acetylases family CE4 from Streptomyces lividans (PDB code: 2CC0) contains a single Zn²⁺ coordinated by Asp13, His62 and His66 (Figure 3L). For hydrolysis, the nucleophilic water molecule that binds to Zn²⁺ attacks the carbonyl carbon of substrates, and His62 stabilizes the carbonyl group of the substrate by forming an oxyanion hole [157]. Amidohydrolase from Mycoplasma synoviae (PDB code: 3OVG) does not have an HXHXDH motif, but two Zn²⁺-binding sites. His186 and His214 coordinate the first Zn²⁺, and His24, His26 and Asp272 coordinate the second Zn²⁺. It also has an His26-Asp68 catalytic dyad.

6. Classification Based on Tertiary Structure

The CATH database (for class, homology, architecture, homologous superfamily; http://www.cathdb.info) is designed to predict protein function from structures [207]. In the most recent version of CATH (v4.2.0), a total of 150,885 PDB-deposited structures are classified into four classes, 41 architectures, 1391 topologies and 6119 homologous superfamilies. According to CATH-based analysis, 107 structures of CEHs are classified into an α/β/α sandwich architecture (Figure 7A).

Among the remaining 29 CEH structures, 16 belong to various architectures, but 13 CEHs (representative PDB codes: 4V95 [134], 5AH1 [100], 5GMX [61], 5H3B [62], 5H6B [102], 5UGQ [40], 5XPX, 5YAL [160], 6A12 [103], 6AAE [63], 6EHN [41], 6GRY [42] and 6QGB [177]) are not classified under the CATH database classification. Secreted PLA₂ from S. violaceoruber (PDB code: 1LWB) forms an up–down bundle structure with only helices (Figure 7B). Gluconolactonase from Xanthomonas campestris (PDB code: 3DR2) is composed of six blades in a propeller structure (Figure 7C) [117], and similarly, the tertiary structure of 6-phosphogluconolactonase from Klebsiella pneumoniae (PDB code: 6NAU) is a 7-blade propeller (Figure 7D). E. coli OMPLA (representative PDB code: 1FW2) [108] and S. typhi OMPLA (PDB code: 5DQX) form β-barrel structures, and contain a catalytic triad at a dimeric interface in the β-barrel transmembrane area (Figure 7E). Lactone hydrolases (PDB code: 4KEP, 2WYM and 2A7M) have an α/β/β/α 4-layered fold (Figure 7F). The overall shapes of amidohydrolase LigI from Sphingobium (PDB code: 4D8L) [154], acetyl xylan esterases from S. lividans (PDB code: 2CC0) [157] and amidohydrolase from M. synoviae (PDB code: 3OVG) are α/β barrels (Figure 7G). Apo and substrate-binding forms of pectin methylesterase A from D. chrysanthemi (representative PDB code: 1QJV and 2NSP) [114,115] and Y. enterocolitica (PDB code: 3UW0) [116] are right-handed parallel β helices, called 3-solenoids (Figure 7H).

The CEH structures can be also classified according to the CASTLE database (https://castle.cbe.iastate.edu) in which bacteria, archaea and eukaryote CEHs are clustered into three clans of α/β hydrolase based on the number and order of β-sheets; eight β-sheets arranged with 1-2-4-3-5-6-7-8 order (Clan A, Figure 8A), five β-sheets arranged with 2-1-3-4-5 order (Clan B, Figure 8B) and seven β-sheets arranged with 1-3-2-4-5-6-7 order (Clan C, Figure 8C) β-sheet sequences. In addition, there are two non-α/β hydrolase clans in the CEHs; a six-bladed β-propeller (Clan D) and three α-helix bundle (Clan E) [208]. Initially, the α/β hydrolase fold was described as α helices surrounding eight central β sheets [209], but it was extended to involve variations including a smaller fold composed of five central β sheets and sandwiched α helices [210]. Among 136 CEHs, 46 structures (33.8%) are turned to have an α/β hydrolase fold in Clans A, B and C (Table 2). Only single structure (PDB code: 3DR2) belongs to Clan D, and no structure is assigned to Clan E.

7. Substrate-Structure Connection of CEHs

The development of sequencing technology enables the identification of new enzymes from various organisms, including bacteria, and even from the metagenome [212,213,214,215,216]. Functional annotation of those enzymes has been followed by direct or indirect approaches, such as computational sequence/structure analysis and comparison with characterized enzymes [217]. 3D-Fun [218], MOLMAP descriptors [219], ECAssigner [220], EC-blast [221] and the recently released DeepEC [222] have been developed to find links between the EC number and enzyme structures.

Accuracy in predicting the functional assignment of enzymes has improved, but as Gerlt’s statistical analysis shows, only 0.63% of proteins by computationally automated annotation have been manually assigned to an EC class [223]. The Ferrer group with the Industrial Applications of Marine Enzymes Consortium (INMARE) attempted to predict enzyme-substrate correlation using the results of a high-throughput assay (145 ester hydrolases sequences and 96 substrates) [224]. Based on this analysis, enzyme promiscuity was proposed. It has been consistently reported that one enzyme can be assigned to multiple classes from primary to quaternary orders of EC class [225]. For example, E. coli lysophospholipase L1 (representative PDB code: 1IVN) is assigned to be a lysophospholipase (EC 3.1.1.5), but shows arylesterase (EC 3.1.1.2), palmitoyl-CoA hydrolase (EC 3.1.2.2), acyl-[acyl-carrier-protein] hydrolase (EC 3.1.2.14) and protease (EC 3.4.21.-) activities (Table 1).

8. Physiological Functions of CEHs

Bacterial CEHs participate in various phycological processes, such as signaling pathways, protein synthesis and offensive-defensive responses. Bacteria recognize ligands and respond, moving toward or repelling from ligands through flagellar movement [226]. The response to ligands in surrounding environment conditions occurs using chemotaxis receptors, which are regulated by methyl transferase CheR and hydrolase CheB [14]. CheB removes methyl groups from methylated glutamate residues in the cytoplasm and inactivates receptors. When recognizing high cell density, the Gram-negative bacteria release autoinducers, such as N-acyl homoserine lactone (AHL) derivatives that induce virulence gene expression by self-receptors [227,228,229,230]. To remove autoinducing signals, they release AHL lactonase, inactivating and degrading AHL by cleavage lactone rings [16,231]. From the function of lactonase, the regulation of AHL through lactonases has been suggested for medicinal applications [232]. PLAs, including secreted, cytosolic and membrane-integrated forms, work in nutrient digestion, inflammation and intra-signaling cascades [233,234]. Mono- and diacylglycerol lipases are important for lipid metabolism in bacteria [235]. Short-chain fatty acids produced by CEHs in gut microbiota regulate host signaling and metabolic systems [236,237]. In bacterial ribosomal machinery, CEHs also resolve non-stop translation problems by cleaving peptide- or amino acid-conjugated tRNA [238,239].

9. Industrial Applications of CEHs

CEHs are widely used in industrial fields because their endogenous characteristics, such as the substrate specificity and stability of structures [240,241,242,243]. Moreover, it has been used for green chemistry, since reactions using biocatalysts employ less steps for the chemical synthesis and produce less harmful wastes during the reaction compared to reactions using chemical catalysts. Furthermore, biocatalysts cover wide substrates with fewer unexpected side products. Carbohydrate esterases have been widely used in animal- and plant-oriented biomass degradation, in the production of biofuels such as ethanol [244,245,246] and in coffee fermentation to improve flavor and taste [247]. Lipases or carboxylesterases are useful for providing the flavor of yogurt and cheese [248,249]. Polyethylene terephthalate hydrolases are suggested as potent biocatalysts for waste management [250]. These characteristics of CEHs can be further improved through structure-based engineering and directed evolution [251]. Additionally, enzyme reusability through immobilization methods using nanostructure [252], cross-linking [253], encapsulation [254] and entrapment [255], are also being considered, and the immobilized enzymes make it possible to reduce the high initial costs associated with enzyme preparation [256].

10. Perspectives on Identifying More CEHs and Their Functions

CEHs are one of the largest group of enzymes, comprising lipases, carboxylesterases, carbohydrate esterases, peptidyl-tRNA hydrolases and lactonases. They target not only carboxylic ester bonds, but also amide (EC 3.5.1.-), thioester (EC 3.1.2.-) and peroxide (EC 1.-.-.-) bonds rarely. The promiscuity of CEHs should be considered when identifying enzyme characteristics.

For these purposes, the factors that induce promiscuity should be identified and the range of specificity chosen. Collecting functional and structural genomics data and linking these large datasets should be done systemically. Amin et al. mentioned the importance of motif in structures when defining enzyme function [257], and Kingsley et al. used various kinetic models to confirm that substrate tunnels in enzymes affect substrate specificity [258]. However, in contrast to an abundance of structural information, fewer structures have been properly matched to biochemical data or functional annotation in the PDB. Moreover, much is unknown about orphan reactions, in which the substrate and the products are already known, but the responsible enzymes are not [259,260,261]. All existing information should be gathered and used to fill in the remaining blanks to generate a full understanding at the molecular level and draw a heuristic map of the biochemical universe.

Funding

This research was funded by grant from the National Research Foundation of Korea, grant number 2017M3A9E4078553.

Conflicts of Interest

The authors declare no conflict of interest.

Figures and Tables

Figure 1. Distribution of the resolution from 424 crystals and the genera of the 136 functionally-annotated CEHs from the PDB are described. (A) The distribution of crystal resolution is shown as a histogram, and the number on top of each bar means the number of crystals in the defined range of resolution. (B) The distribution of genera of CEHs is shown using circle graph. The numbers beside the genera indicate the numbers of CEHs in each genus. When groups have fewer than three members, all are assigned to others.

Figure 2. Phylogenetic tree of CEH sequences of all 136 CEHs are aligned and analyzed using a neighbor-joining method and the Jones–Taylor–Thornton (JTT) substitution model. Active site residue-based classification of CEHs is described using a combination of shape and color. Pink-to-purple closed circle: serine hydrolases (Group 1); orangish closed square: aspartidyl hydrolase (Group 2); greenish closed diamond: metal-independent non-serine hydrolases with water molecule as a nucleophile (Group 3); blueish closed triangle: metal-dependent metallohydrolases (Group 4). Different colors are used to distinguish reaction mechanisms. No symbol label means no information is available on their catalytic reaction. Localization of CEHs is described using the following markers at the outside of PDB codes in the phylogenetic tree: Asterisks (*) denote CEHs with signal peptide for secretion and dollar signs ($) denote membrane protein CEHs. CEHs without a marker are cytosolic CEHs. The figure is prepared using PROMALS3D [185,186,187] for sequence alignment and MEGA X [188] for phylogenetic tree.

Figure 3. Representative structures of CEHs in groups based on active site are described in 3D structures. The colors in Figure 2 are applied in the cartoon structures of (A) Group 1-1, (B) Group 1-2, (C) Group 1-3, (D) Group 1-4, (E) Group 2, (F) Group 3-1, (G) Group 3-2, (H) Group 3-3, (I) Group 3-4, (J) Group 3-5, (K) Group 4-1 and (L) Group 4-2. Catalytic residues, oxyanion hole residues and important residues are depicted using a sticks model with light gray. The catalytic important water molecule is shown with a red sphere, and metal ions are shown with blue-gray spheres. The names of functional core motifs are depicted near their composing residues. The four-digit PDB codes of the models are noted in bottom-right corner of each panel. Structures are visualized using PyMOL software [192].

Figure 4. Consensus sequence blocks in the GDSX superfamily. Conserved regions in sequences of the GDSX motif-containing CEHs are described in black-outlined boxes. Assigned numbers above each box are the position of residues in CjCE2A (PDB code: 2WAA) as representative GDSX motif CEHs, and oxyanion hole-forming residues are highlighted using asterisks below the boxes. Fully conserved positions are marked by red shading, and highly conserved positions (>70%) are highlighted in yellow. The conserved residues with high similarity are in bold. PROMALS3D [185,186,187] for sequence alignment and ESPript 3 [195] for visualization were used.

Figure 5. Sequences of peptidyl-tRNA hydrolases are aligned, and functionally important areas are described. The numbers above the sequences are the residual numbers of peptidyl-tRNA hydrolase from V. cholerae (PDB 4Z86). Catalytic His and Asp are marked using asterisks, substrate-binding residues are marked using S and oxyanion hole-forming Asn residues are marked using O beneath the sequences. Fully conserved positions are marked with red shading, and highly conserved positions (>70%) are highlighted by yellow shading. The conserved residues with high similarity are in bold. PROMALS3D [185,186,187] for sequence alignment and ESPript 3 [195] for visualization were used.

Figure 6. The sequences of lactonases containing the HXHXDH motif are aligned. The numbers above the aligned sequences are the residue positions in N-acyl lactonase from Bacillus thuringiensis (PDB code: 2A7M). The position, which is close to the first and the second metal ions, is marked bottom of the alignment using 1 (the first) and 2 (the second). Fully conserved positions are marked by red shading, and highly conserved positions (>70%) are in yellow. The conserved residues with high similarity are in bold. PROMALS3D [185,186,187] for sequence alignment and ESPript 3 [195] for visualization are used.

Figure 7. Various tertiary architectures of CEHs in PDB are described using rainbow colors in the direction of the N to C terminus. The 3D models are obtained from representative structures, which contain each architecture: (A) α/β/α sandwich (PDB code: 3FAK), (B) up-down bundle (PDB code: 1LWB), (C) 6-blade propeller (PDB code: 3DR2), (D) 7-blade propeller (PDB code: 6NAU), (E) β-barrel (PDB code: 5DQX), (F) α/β/β/α 4-layer (PDB code: 4KEP), (G) α/β barrel (PDB code: 4D8L), (H) 3-solenoid (PDB code: 2NSP). Structures are visualized using PyMOL software [192].

Figure 8. Various topologies of α/β hydrolases in CEHs are described. (A) Classical α/β hydrolase fold with 1-2-4-3-5-6-7-8 order of β sheets. (B) A variant of α/β hydrolase fold with 1-3-2-4-5-6-7 order of β sheets. (C) Minimal α/β hydrolase fold with 2-1-3-4-5 order of β sheets. β sheets are described with rainbow colors (red to navy blue) by the order of β sheets, and all α helices are described with gray. Structures are visualized using PyMOL software [192].

Table 1

List of 424 Protein Data Bank (PDB) codes in carboxylic ester hydrolases (CEHs).

EC Number	Enzyme Function	PDB Codes
EC 3.1.1.-	carboxylic ester hydrolases	1C7I[19], 1C7J[19], 1ESC[20], 1ESD[20], 1ESE[20], 1QE3[19], 2A7M[21], 2WAA[22], 2WYL[23], 2WYM[23], 3DHA[24], 3DHB[24], 3DHC[24], 3DNM[25], 3F67, 3FAK[26], 3G9T, 3G9U, 3G9Z, 3GA7, 3H17[27], 3H18[27], 3H19, 3H1A, 3H1B, 3H2G[28], 3H2H[28], 3H2I[28], 3H2J[28], 3H2K[28], 3I2F[29], 3I2G[29], 3I2H[29], 3I2I[29], 3I2J[29], 3I2K[29], 3IDA[30], 3K6K, 3L1H, 3L1I, 3L1J, 3LS21[31], 3PF8[32], 3PF9[32], 3PFB[32], 3PFC[32], 3PUH[33], 3PUI[33], 3QM1[32], 3S2Z[32], 4F21[34], 4KRX[35], 4KRY[35], 4OB6[36], 4OB7[36], 4OB8[36], 5A2G[37], 5HC0[38], 5HC2[38], 5HC3[38], 5HC4[38], 5HC5[38], 5MAL2[39], 5UGQ[40], 5UNO[40], 5UOH[40], 6EHN[41], 6GRY[42]
EC 3.1.1.1	carboxylesterase	1AUO[43], 1AUR[43], 1EVQ[44], 1L7Q[45], 1L7R[45], 1R1D, 1TQH[46], 2H1I, 2HM7[47], 2R11, 3CN7[48], 3CN9[48], 3DOH[49], 3DOI[49], 3KVN[50], 4BZW[51], 4BZZ[51], 4C01[51], 4C87[52], 4C88[52], 4C89[52], 4CCW[53], 4CCY[53], 4FHZ[54], 4FTW[54], 4IVI[55], 4IVK[55], 4JGG[56], 4OU4[36], 4OU5[36], 4ROT, 4UHC[57], 4UHD[57], 4UHE[57], 4UHF[57], 4UHH[57], 4V2I[58], 4YPV[59], 5AO9[60], 5AOA[60], 5AOB[60], 5AOC[60], 5DWD, 5EGN, 5GMX[61], 5H3B[62], 6AAE[63], 6IEY[63]
EC 3.1.1.2	arylesterase	1VA43[64], 2Q0Q[65], 2Q0S[65], 3HEA3[64,66], 3HI43[66], 3IA23[67], 3T4U3, 3T523, 4ROT, 4TX1[68]
EC 3.1.1.3	triacylglycerol lipase	1CVL[69], 1EX9[70], 1HQD[71], 1I6W[72], 1ISP4[73], 1JI3[74], 1KU0[75], 1OIL[76], 1QGE, 1R4Z[77], 1R50[77], 1T2N[78], 1T4M[78], 1TAH[78], 1YS1[79], 1YS2[79], 2DSN[80], 2ES4[81], 2FX5, 2HIH[82], 2LIP[83], 2NW6[84], 2ORY[85], 2QUA[86], 2QUB[86], 2QXT[87], 2QXU[87], 2W22[88], 2Z5G[80], 2Z8X[89], 2Z8Z[89], 2ZJ6[90], 2ZJ7[90], 2ZVD[91], 3A6Z[91], 3A70[91], 3AUK, 3D2A[92], 3D2B[92], 3D2C[92], 3LIP[83], 3QMM[93], 3QZU[94], 3UMJ[95], 3W9U, 4FDM[96], 4FKB, 4FMP, 4GW3[97], 4GXN[97], 4HS9[97], 4LIP[98], 4OPM, 4X6U[99], 4X71[99], 4X7B[99], 4X85[99], 5AH1[100], 5CE5[100], 5CRI[101], 5CT4[101], 5CT6[101], 5CT9[101], 5CTA[101], 5CUR[101], 5H6B[102], 5LIP[98], 5MAL2[39], 5XPX, 6A12[103], 6CL4[103], 6FZ1[104], 6FZ7[104], 6FZ8[104], 6FZ9[104], 6FZA[104], 6FZC[104], 6FZD[104]
EC 3.1.1.4	phospholipase A2	1FAZ[105], 1KP4[105], 1LWB[106], 1QD5[107], 1QD6[107], 1FW2[108], 1FW3[108], 1ILD[109], 1ILZ[109], 1IM0[109], 5DQX
EC 3.1.1.5	lysophospholipase	1IVN5[110], 1J005[110], 1JRL5[110], 1U8U5[111], 1V2G5[111], 5TIC5[112], 5TID5[112], 5TIE5[112], 5TIF5[112]
EC 3.1.1.6	acetylesterase	2XLB[113], 2XLC[113], 3FVR, 3FVT, 3FYT, 3FYU, 4NS4
EC 3.1.1.11	pectinesterase	1QJV[114], 2NSP[115], 2NST[115], 2NT6[115], 2NT9[115], 2NTB[115], 2NTP[115], 2NTQ[115], 3UW0[116]
EC 3.1.1.17	gluconolactonase	3DR2[117]
EC 3.1.1.20	tannase	3WA6[118], 3WA7[118]
EC 3.1.1.23	acylglycerol lipase	3RLI[119], 3RM3[119], 4KE6[120], 4KE7[120], 4KE8[120], 4KE9[120], 4KEA[120], 4LHE[121], 5XKS
EC 3.1.1.24	3-oxoadipate enol-lactonase	2XUA[122]
EC 3.1.1.25	1,4-lactonase	3MSR, 3OVG
EC 3.1.1.27	4-pyridoxolactonase	3AJ3, 4KEP, 4KEQ
EC 3.1.1.29	aminoacyl-tRNA hydrolase	2PTH[123], 2Z2I[124], 2Z2J[124], 2Z2K[124], 3KJZ, 3KK0, 3NEA[125], 3OFV, 3P2J, 3TCK[126], 3TCN[126], 3TD2[126], 3TD6[126], 3V2I[127], 3VJR[128], 4DHW, 4DJJ, 4ERX, 4FNO[129], 4FOP[130], 4FOT[130], 4FYJ[131], 4HOY[130], 4IKO[130], 4JC4[129], 4JWK[130], 4JX9[130], 4JY7[130], 4LIP[98], 4LWQ, 4OLJ, 4P7B[132], 4QAJ[129], 4QBK[129], 4QD3[129], 4QT4[133], 4V95[134], 4YLY[135], 4Z86[136], 4ZXP[136], 5B6J[136], 5EKT, 5GVZ, 5IKE[136], 5IMB[136], 5IVP, 5Y98[137], 5Y9A[137], 5YL8, 5YLA, 5YN4, 5ZK0, 5ZX8[138], 5ZZV, 6A31, 6IVV, 6IX6, 6IYE, 6J93, 6JGU, 6JJ1, 6JJQ, 6JKX, 6JQT
EC 3.1.1.31	6-phosphoglucono-lactonase	1PBT, 1VL1, 3ICO[139], 3LWD, 3OC6[139], 4TM7[140], 4TM8[140], 6NAU
EC 3.1.1.32	phospholipase A1	1QD5[107], 1QD6[107], 1FW2[108], 1FW3[108], 1ILD[109], 1ILZ[109], 1IM0[109], 4HYQ[141], 5DQX
EC 3.1.1.41	cephalosporin-C deacetylase	1L7A, 1ODS[142], 1ODT[142], 1VLQ[143], 3FCY, 3M81[143], 3M82[143], 3M83[143], 5FDF[144], 5GMA[145], 5HFN[144], 5JIB[146]
EC 3.1.1.43	alpha-amino-acid esterase	1MPX[147], 1NX9[148], 1RYY[148], 2B4K[148], 2B9V[148]
EC 3.1.1.45	carboxymethylene-butenolidase	1DIN[149], 1GGV[150], 1ZI6[151], 1ZI8[151], 1ZI9[151], 1ZIC[151], 1ZIX[151], 1ZIY[151], 1ZJ4[151], 1ZJ5[151], 4P92[152], 4P93[152], 4U2B[153], 4U2C[153], 4U2D[153], 4U2E[153], 4U2F[153], 4U2G[153]
EC 3.1.1.57	2-pyrone-4,6-dicarboxylate lactonase	4D8L[154]
EC 3.1.1.61	protein-glutamate methylesterase	1CHD[13], 3SFT[155], 3T8Y[156]
EC 3.1.1.72	acetylxylan esterase	2CC0[157], 2VPT6[158], 2WAA[22], 2XLB[113], 2XLC[113], 3FCY, 3W7V[159], 4JHL[159], 4JJ4, 4JJ6, 4JKO[159], 4OAO, 4OAP, 5BN1, 5FDF[144], 5GMA[145], 5HFN[144], 5JIB[146]
EC 3.1.1.73	feruloyl esterase	5YAE[160], 5YAL[160]
EC 3.1.1.74	cutinase	4CG1[161], 4CG2[161], 4CG3[161], 4EB0, 5ZOA, 5ZNO[162]
EC 3.1.1.75	poly(3-hydroxybutyrate) depolymerase	4BRS[163], 4BTV[163], 4BVJ[163], 4BVK[163], 4BVL[163], 4BYM[163]
EC 3.1.1.81	quorum-quenching N-acyl-homoserine lactonase	2A7M[21], 2BR6[164], 2BTN[164], 3DHA[165], 3DHB[165], 3DHC[165], 4J5F[166], 4J5H[166], 5EH9[167], 5EHT[167]
EC 3.1.1.84	cocaine esterase	1JU3[168], 1JU4[168], 1L7Q[45], 1L7R[45], 3IDA[30], 4P08[169]
EC 3.1.1.85	pimeloyl-[acyl-carrier protein] methyl ester esterase	4ETW[170], 4NMW
EC 3.1.1.95	aclacinomycin methylesterase	1Q0R[171], 1Q0Z[171]
EC 3.1.1.101	poly(ethylene terephthalate) hydrolase	5XFY[172], 5XFZ[172], 5XG0[172], 5XH2[172], 5XH3[172], 5XJH[173], 5YFE[174], 5YNS[173], 6ANE[175], 6EQD[174], 6EQE[174], 6EQF[174], 6EQG[174], 6EQH[174], 6ILW[176], 6ILX[176], 6QGC[177]
EC 3.1.1.102	mono(ethylene terephthalate) hydrolase	6QG9[177], 6QGA[177], 6QGB[177]

* PDB codes with superscript and underline have extra-assigned EC numbers corresponding to any additional activities of the CEHs. (¹ = [EC 3.1.2.12]; ² = [EC 3.1.2.2]; ³ [EC 1.-.-.-]; ⁴ = [EC 3.4.21.-]; ⁵ = [EC 3.1.2.2, EC 3.1.2.14, EC 3.4.21.- and EC 3.1.2.-]; and ⁶ = [EC 3.2.1.4]). ** Representative PDB codes are highlighted using bold characters. *** PDB codes in [EC 3.1.1.84] are the same proteins with 3I2K PDB code as a cocaine esterase. However, 3I2F, 3I2G, 3I2H, 3I2I, 3I2J and 3I2K were not assigned to [EC 3.1.1.84].

Table 2

Representative PDB Code List of CEH Clans.

Clan Types	PDB Codes
Clan A	1C7I[19], 1EVQ[44], 1L7A, 1Q0R[171], 2R11, 2XLB[113], 2XUA[122], 3FAK[26], 3FCY, 3GA7, 3K6K, 4CCW[53], 4CCY[53], 4CG1[161], 4EB0, 4KRX[35], 4OB7[36]
Clan B	1CVL[69], 1ESC[20], 1EX9[70], 1JI3[74], 1KU0[75], 1QGE, 1YS1[79], 2DSN[80], 2HIH[82], 2W22[88], 3AUK, 4FDM[96], 4FKB, 4HS9[211], 4HYQ[141], 4X6U[99]
Clan C	1AUO[43], 1TQH[46], 1ZI8 [151], 3CN9[48], 3F67, 3IA2[67], 3RLI[119], 4ETW[170], 4F21[34], 4FHZ[54], 4NMW, 4U2D[153], 4UHC[57]
Clan D	3DR2[117]
Clan E	N/A