Relationships between Substrate Promiscuity and Chiral Selectivity of Esterases from Phylogenetically and Environmentally Diverse Microorganisms

Abstract

Substrate specificity and selectivity of a biocatalyst are determined by the protein sequence and structure of its active site. Finding versatile biocatalysts acting against multiple substrates while at the same time being chiral selective is of interest for the pharmaceutical and chemical industry. However, the relationships between these two properties in natural microbial enzymes remain underexplored. Here, we performed an experimental analysis of substrate promiscuity and chiral selectivity in a set of 145 purified esterases from phylogenetically and environmentally diverse microorganisms, which were assayed against 96 diverse esters, 20 of which were enantiomers. Our results revealed a negative correlation between substrate promiscuity and chiral selectivity in the evaluated enzymes. Esterases displaying prominent substrate promiscuity and large catalytic environments are characterized by low chiral selectivity, a feature that has limited commercial value. Although a low level of substrate promiscuity does not guarantee high chiral selectivity, the probability that esterases with smaller active sites possess chiral selectivity factors of interest for industry (>25) is significantly higher than for promiscuous enzymes. Together, the present study unambiguously demonstrates that promiscuous and selective esterases appear to be rare in nature and that substrate promiscuity can be used as an indicator of the chiral selectivity level of esterases, and vice versa.

1. Introduction

Presently, there is a great need for suitable biocatalysts with high process performance as greener alternatives to chemical synthesis [1,2]. It is expected that up to 40% of bulk chemical synthesis processes could be substituted by enzymatic catalysis by 2020 [1]. Along with requirements of a technical nature, such as process development and optimization, it is however widely recognized that the establishment of enzymatic processes is mainly a problem of finding, optimizing, or designing new and/or better performing enzymes. Nature is a rich reservoir from where enzymes can be isolated [3,4], because they are continuously evolving as a consequence of natural selection. Promiscuous enzymes are effective for converting multiple substrates, thus, they are industrially relevant [4,5,6]. Enzymes need to also be robust and, preferably, chiral selective to reduce raw material costs in the synthesis of pure chiral compounds [1,2,4,7]. That is, they need to be able to cleave preferentially only one chiral ester when offered a racemic mixture. Is it possible to find versatile enzymes displaying prominent substrate range and stringent chiral selectivity? Evaluating this possibility was the starting point of the present study.

In this study, we are interested in investigating as model enzymes serine ester hydrolases, hereafter referred to as esterases, from the structural superfamily of α/β-hydrolases. The activity of these esterases relies mainly on a catalytic triad usually formed by Ser, Asp/Glu, and His [8]. This enzyme class was selected for a number of reasons: it is widely distributed in the environment, it has important physiological functions, it includes hydrolases that are among the most important industrial biocatalysts, and extensive biochemical knowledge has been accumulated [4,5,7].

Just focusing on those from uncultivated microorganisms discovered through metagenomic approaches, esterases with prominent chiral selectivity have been identified and their use in the kinetic resolution of a number of esters is reported. Recent examples include those preferably hydrolyzing one of the chiral esters in racemic mixtures of ibuprofen esters [9,10]; ketoprofen esters [11,12,13,14]; solketal esters [15]; esters of phenylalkyl carboxylic acids, 1,1,1-trifluoro-2-phenylbut-3-yn-2-yl acetate and 3,7-dimethyl-1,6-octadien-3-yl acetate [16,17]; methyl 3-phenylglycidate [18]; 1-phenylethyl acetate [19,20]; ofloxacin butyl ester [21]; 1-octin-3-ol, 3-chlor-1-phenyl-1-propanol, trimethylsilylbutinol, cis/trans-1,2-cyclohexanediol, and isopropylidenglycerol acetate [22]; glycidyl butyrate [23]; methyl-mandelate, glycidyl-4-nitrobenzoate, methyl-3-bromo-2-methyl propionate, methyl lactate, menthyl acetate, neomenthyl acetate, pantolactone, and methyl 3-hydroxybutyrate [22,24,25]; 1-octin-3-ol, 3-chlor-1-phenyl-1-propanol, and trimethylsilylbutinol [22]; methyl-3-hydroxy-2-methylpropionate [26]; and esters of secondary alcohols [27,28], to cite some. The advances in metagenomics techniques and screening methods have allowed the discovery of these and other selective esterases [29]. These studies exemplify that esterases with selective character occur naturally, and that their chiral preference depend on structural factors in the proximity of the active-site. However, whether the selective character of these esterases, and many others, is linked to a broad or a narrow substrate spectrum has not been investigated, due to limited substrate sets employed.

Here, we investigate the relationships between the level of substrate promiscuity and chiral selectivity of a large set of 145 phylogenetically and environmentally diverse microbial esterases, whose specific activity against 96 distinct esters that included 20 chiral esters have been recently reported [5]. We provide unambiguous experimental evidence suggesting a negative association between substrate specificity and chiral selectivity in native esterases.

2. Results and Discussion

2.1. Relationships between Substrate Promiscuity and Chiral Selectivity

We have recently described an extensive analysis of the substrate spectra of 145 phylogenetically and environmentally diverse microbial esterases [5]. Experimental data on substrate conversion (i.e., units g⁻¹ or U·g⁻¹) followed for 24 h, at pH 8.0 and 30 °C was reported for 96 distinct esters. They included esters with variation in size of acyl and alcohol groups and with growing residues (aromatic, aliphatic, branched, and unbranched), halogenated esters, sugar esters, lactones, an alkyl di-ester, and 20 chiral esters (including (R) and (S) enantiomers of menthyl acetate, N-benzyl-proline ethyl ester, methyl mandelate, ethyl 4-chloro-3-hydroxybutyrate, methyl 3-hydroxybutyrate, methyl 3-hydroxyvalerate, neomenthyl acetate, methyl and ethyl lactate, and pantolactone). By meaning of the partitioning coefficient (log p value), which reflects electronic and steric effects and hydrophobic and hydrophilic characteristics, the 96 esters do show a broad chemical and structural variability [5]. This chemical variability also characterized the chiral esters tested (Figure 1).

To find the relationships between substrate promiscuity and chiral selectivity we calculated the chiral selectivity factor for each of the 145 esterases and the 20 chiral esters tested (i.e., two enantiomers per pair) that were included in the 96-ester library (Figure 1). Selectivity factor was calculated as the ratio of specific activity (U·g⁻¹) of the preferred over the non-preferred chiral ester when both esters were tested separately [30] (see Materials and Methods). These calculations were extracted from datasets reported previously [5]. It should be mentioned that these apparent values may not correspond to true selectivity or enantiomeric factors calculated when the enzyme is confronted to a racemic mixture, because the rates of hydrolysis of the enantiomers were measured separately [30]; nevertheless, recent studies have clearly demonstrated that apparent and true selectivity values closely match each other [15]. These values were plotted against the number of esters hydrolyzed by each of the esterases (Figure 2), previously reported for each of the esterases [5].

From the 145 esterases, 40 did not show appreciable activity under assay conditions for any of the chiral esters tested. From those being active against at least one of the chiral esters (105 in total), 80 esterases were characterized by selectivity factors below a threshold of 25. Although esterases with stringent selectivity are preferred, it is commonly considered that enzymes with selectivity factor of 25 or above begin to have commercial value [31]. On the other hand, we found 25 chiral selective esterases, as judged by a selectivity factor above 25 (Figure 2 and Figure 3). Ten of them showed stringent selectivity, that is they were capable of hydrolyzing only one of the enantiomer (Figure 3). Twelve of them were characterized by selectivity factors ranging from 25.9 to 59.3, and three did show prominent selectivity factors ranging from 219 to 686 (Figure 2 and Figure 3).

As shown in Figure 2, we found a negative association between the level of substrate promiscuity, by meaning of the number of esters hydrolyzed, and the chiral selectivity factor. In further detail, according to criteria previously established [5] we considered an esterase ‘specific’ if it hydrolyses 9 esters or fewer, as ‘moderate-to-highly promiscuous’ if it hydrolyses between 10 and 42 esters, and as ‘prominently promiscuous’ if it hydrolyses 42 or more esters. None of the 25 hydrolases which showed a selectivity factor ≥25 were prominently promiscuous. Rather, they were capable of accepting 36 or fewer substrates. However, not all hydrolases converting 36 or fewer esters and acting against chiral esters were selective according to the 25-selective factor threshold. Indeed, only 25 out of 85 in total (or 29%) were selective, with different selectivity factors and chiral preferences (Figure 3). This is most likely due to the fact that the ability to selectively hydrolyze an enantiomer in a racemate may depend on the topology of the catalytic environment [5].

2.2. Occurrence of Multi Selective Esterases

Figure 3 summarizes the chiral preference of esterases fitting to the 25-selective factor threshold for each of the 20 chiral esters tested. As can be seen in Figure 3, esterases showed different level of promiscuity and preference for (R) or (S) methyl acetate, menthyl mandelate, methyl 3-hydroxybutyrate, N-benzyl-proline ethyl ester, ethyl-4-chloro-3-hydroxybutyrate, and (m)ethyl lactate. Esterases selective for (R) or (S)-pantolactone, N-benzyl-proline ethyl ester, and neomenthyl acetate were the least abundant, suggesting these chiral esters are less preferred substrates. As shown in Figure 3, we also found that 5 out of 25 esterases fitting to the 25-selective threshold did show stringent selectivity or selectivity factor higher than 25 for several chiral esters differing in chemical and structural nature. They include one being selective for methyl 3-hydroxybutyrate (R-selective) and ethyl lactate (S-preference) (EH31); one for ethyl-4-chloro-3-hydroxybutyrate (R-selective), methyl 3-hydroxybutyrate (R-selective), and ethyl lactate (S-selective) (EH47); one for methyl mandelate (S-selective), methyl 3-hydroxyvalerate (S-selective) and methyl lactate (S-selective) (EH71); and two for menthyl acetate (R-selective) and methyl lactate (S-selective) (EH72 and EH78). The other 20 esterases did show the capacity to preferentially hydrolyze only one enantiomer (Figure 3). This suggests that multi selective esterases may have a lower abundancy.

3. Materials and Methods

3.1. Source of Chemicals, Enzymes, and Datasets

All chemicals for which activity data are reported were of the purest grade available and were purchased as reported [5]. The present study used datasets of hydrolytic activity (U·g⁻¹) for 145 esterases assayed at 550 nm using 96-structurally diverse esters in 384-well plates. Reactions were followed for 24 h, at pH 8.0 and 30 °C. Datasets are available elsewhere [5].

3.2. Selectivity Factor Calculation

The chiral selectivity factor is defined as the ratio of the specific activity [30] for each enantiomer, measured separately as described previously [5]. Briefly, reaction mixture contains 5 mM N-(2-hydroxyethyl)piperazine-N′-(3-propanesulfonic acid buffer, pH 8.0, 4.5% (v/v) acetonitrile or dimethyl sulfoxide, 0.45 mM Phenol Red (used as a pH indicator), a concentration of each of the esters of 1.14 mg·mL⁻¹, and 2 µg of proteins. Reactions were allowed to proceed kinetically at 30 °C and hydrolytic activity (U·g⁻¹) calculated followed for 24 h [5]. Selectivity factor was calculated considering the preferred over the non-preferred chiral ester, whatever the preferred (R) or (S) ester.

4. Conclusions

Herein, we show the value of the systematic investigation of enzyme activity to deepen our understanding of the relationships between substrate promiscuity and chiral selectivity. By comparing the number of esters that 145 diverse esterases hydrolyze as an indicator of the substrate promiscuity level and their selectivity factors as an indicator of enantio-selectivity, we found unambiguous evidence that esterases with broad substrate spectra do commonly show low selectivity for chiral molecules. In this study, the proportion of esterases with both prominent promiscuity and selectivity approaches zero percent. By contrast, the proportion of esterases with low to moderate promiscuity but prominent selectivity was as high as 29%. This suggests that the substrate promiscuity may be used as an indicator of the selective character of esterases. Promiscuous esterases acting against multiple substrates, while at the same time being enantio-selective, appear to be rare in nature, or at least in the habitats from where the esterases herein described were isolated [5]. As these enzymes are of interest for application purposes [1,2,3,4,5,6,32], protein engineering and rational design may be needed to obtain esterases being promiscuous and selective for industrial applications. We anticipate that the possibility to transform a promiscuous but not selective esterase into an efficient enantio-selective biocatalyst would require less engineering effort because increasing the selectivity for an enantiomer may involve a reduced number of contacts close to the active sites (for a recent example see reference [33]). Conversely, increasing the substrate spectra of a selective non-promiscuous esterase would require large rearrangement of the catalytic environment which may, at the same time, result in significant reduction or even loss of enantio-selectivity. This is because non-promiscuous esterases are characterized by catalytic environments that are highly exposed and have small volumes, while an esterase for being promiscuous requires a large active site volume and lower relative solvent accessible surface area [5], that are difficult to be designed through few mutations.