Aliphatic and enantioselective amidases: from hydrolysis to acyl transfer activity


Alain Arnaud UFR de Microbiologie Industrielle et Génétique des Microorganismes, Ecole Nationale Supérieure Agronomique de Montpellier, 34060, Montpellier cedex 1, France.

1. Summary, 381

2. Amidase-producing organisms, 381

2.1 Amidases from bacteria, 381

2.2 Amidases from yeasts, 382

2.3 Amidases from fungi, 382

2.4 Amidases from plants, 382

2.5 Amidases from animals, 382

3. Amidase classification, 382

3.1 As a function of the catalytic activity, 382

3.2 As a function of the amino-acid sequence, 383

3.3 Evolutionary aspects, 384

4. Enzymatic activities and their applications, 384

4.1 Mechanism of action, 384

4.2 Amide hydrolysis, 386

4.3 Acyl transfer activity, 388

5. Conclusion, 389

6. References, 391


This paper aims at making a review with an historical perspective on amidases and the different reactions they can catalyse.

Amidases (EC are ubiquitous enzymes in the living world and can be divided into two types. The first type includes aliphatic amidases hydrolysing only short-chain aliphatic amides and the second type includes aliphatic amidases hydrolysing mid-chain amides, some arylamides, α-aminoamides and α-hydroxyamides. All of these amidases exhibit an acyl transfer activity leading to the formation of hydroxamic acids. Some of them are also able to transform diverse amides, acids, esters or nitriles in their corresponding carboxylic acids, hydroxamic acids or acid hydrazides.

The current role of amidases in the living world is not clearly defined. However, we have learnt a lot about the evolutionary aspects of these enzymes and the in vitro amidase-catalysed reactions during the last 50 years. Amidases turned out to be efficient tools for the synthesis of various compounds.


Amidases are amide bond-cleaving enzymes that have been extensively investigated during the last 20 years. They catalyse the hydrolysis of amides to carboxylates and ammonia and exist in all kingdoms of the living world.

2.1 Amidases from bacteria

Most of the currently known amidases have been found and described in bacteria. Many genera are concerned: Rhodococcus, Corynebacterium, Mycobacterium, Pseudomonas, Bacillus, Micrococcus, Brevibacterium, Nocardia, Streptomyces, Blastobacter, Arthrobacter, Alcaligenes, Helicobacter, Lactobacillus, and Methylophilus.

These different amidases exhibit various substrate specificities. As an example, Kimura partially purified a nicotinamidase from Mycobacterium avium which appeared to be absolutely specific for nicotinamide (Kimura 1959a), whereas Kobayashi and coworkers purified and characterized an amidase from Rhodococcus rhodochrous J1 with a very wide substrate spectrum (Kobayashi et al. 1993). We can indeed distinguish absolutely specific amidases from nonspecific ones but also aliphatic amidases from aryl amidases and amino-amide amidohydrolases. It is important to note that one microorganism can contain several amidases. For instance, the strain Rhodococcus sp. R312 described by Arnaud and coworkers contains a L–α–aminoamidase, an enantioselective amidase, an aliphatic amidase, and several absolutely specific amidases for nicotinamide, formamide or urea (Arnaud et al. 1976).

2.2 Amidases from yeasts

There are very few references in the literature to enzymes from yeasts which catalyse the breakdown of amides to carboxylates and ammonia. Gorr and Wagner described the degradation of some amides (acetamide, asparagine and urea) by freeze-dried preparations of Candida utilis whole cells (Gorr and Wagner 1933). Later, Brady showed that, when Candida utilis cells grew on a glucose containing medium, they were able to hydrolyse a wider spectrum of amides (Brady 1969). This author purified a protein which exhibited no activity for asparagine but which efficiently hydrolysed acetamide, propionamide, butyramide, valeramide, hexanoamide and acrylamide. Finally, Joshi and Handler purified another protein in Torula cremoris which specifically hydrolysed nicotinamide (Joshi and Handler 1962).

2.3 Amidases from fungi

Aspergillus nidulans has been the most studied fungal strain. Several amidases have been described (Hynes and Pateman 1970a): a formamidase, active with formamide and glycinamide as substrates; an acetamidase, hydrolysing aliphatic amides from 1 to 6 carbon atoms; and a wide spectrum amidase, exhibiting activity with aliphatic amides (butyramide, valeramide, hexanoamide) and with arylamides (benzamide, phenylacetamide). The expression of the genes encoding these different amidases has been widely studied (Hynes and Pateman 1970b; Hynes 1975a, b, 1978, 1979, 1980, 1982).

2.4 Amidases from plants

In 1993, Kammermeier-Steinke and coworkers partially purified a peptide amidase from oranges (Citrus sinensis L.). This protein is a highly regio- and stereo-selective enzyme which hydrolyses C-terminal amide groups in peptides or N-protected amino acids while amino-acid amides are not deamidated. From the investigation of the substrate range it is further known that the peptide amidase requires a L-configured amino acid at the C-terminus and accepts all proteinogenic amino acids with the exception of L-proline (Kammermeier-Steinke et al. 1993).

2.5 Amidases from animals

In 1949 and 1950, Bray and coworkers made enzyme preparations from animal tissues which were able to hydrolyse aliphatic amides; rat liver extracts hydrolysed acetamide and propionamide but higher activities were obtained with mid-chain amides (C6–C7) and the preparation also hydrolysed some aromatic amides. These authors concluded that they were dealing with a single enzyme which had a very broad range of substrate specificity (Bray et al. 1949, 1950).

In 1995, Ueda partially purified a porcine brain enzyme which hydrolysed anandamide (arachidonyl ethanolamide), an endogenous ligand for cannabinoid receptors (CB1, CB2) and a putative neurotransmitter (Ueda et al. 1995). This enzyme, thus leading to the formation of arachidonate and ethanolamine, was also described in rat brain (Arreaza et al. 1997). The hydrolytic enzyme is called anandamide amidase or fatty acid amidohydrolase (FAAH) or N-arachidonyl ethanolamine deacylase (EC or EC 3.5.160). Anandamide is the preferred substrate for this enzyme although it reacts with a variety of other lipid signalling molecules such as fatty acid ethanolamides or fatty acid primary amides including oleamide, a putative sleep factor. Synthetic oleamide induces sleep when injected into rats but was also shown to affect serotonergic systems and block gap junction communication in glial cells. All of these biological properties are not replicated by oleic acid (cis-9-octadecenamide) and trans-octadecenamide.


3.1 As a function of the catalytic activity

As we have seen, amidohydrolases, amide bond-cleaving enzymes, exist in both prokaryotic and eukaryotic forms. According to the categorical numbering system of EC (Schomburg and Salzmann 1991) that uses such properties as substrate specificity and physicochemical characteristics as criteria, amidohydrolases have been divided into two major types: 77 were included in the EC 3.5.1 category (EC– and 14 were placed under EC 3.5.2 (EC– Among them, one amidohydrolase class, the acylamide amidohydrolases EC (with which we are particularly concerned here), catalyse the breakdown of aliphatic acylamides to the corresponding acid and ammonia. These strict amidases have been extensively described in bacteria. Two types of such enzymes can be distinguished. The first type includes aliphatic amidases hydrolysing only short-chain aliphatic amides. The aliphatic amidases from Pseudomonas aeruginosa (Clarke 1970), Rhodococcus sp. R312 (Maestracci et al. 1984; Soubrier et al. 1992), Arthrobacter sp. J1 (Asano et al. 1982), Methylophilus methylotrophus (Wyborn et al. 1996), Helicobacter pylori (Skouloubris et al. 1997) and Bacillus stearothermophilus BR388 (Cheong and Oriel 2000) belong to this group. Acetamide, acrylamide and propionamide are the most rapidly hydrolysed substrates. The second type includes aliphatic amidases which hydrolyse mid-chain aliphatic amides and which are coupled with nitrile hydratases to be involved in microbial nitrile metabolism (see Figs 1 and 2). Propionamide, isobutyramide, valeramide, and hexanoamide are the most rapidly hydrolysed substrates. These enzymes have also been shown to be enantioselective towards several racemic amides. The aliphatic enantioselective amidases from Rhodococcus sp. R312 (Mayaux et al. 1990), Rhodococcus sp. N-774 (Hashimoto et al. 1991), Rhodococcus sp. (Mayaux et al. 1991), R. erythropolis JCM6823 (Duran et al. 1993), R. rhodochrous J1 (Kobayashi et al. 1993), R. erythropolis MP50 (Hirrlinger et al. 1996), Ps. chlororaphis B23 (Ciskanik et al. 1995) and Bacillus sp. BR449 (Kim and Oriel 2000) belong to this group.

Figure 1.

 Structural organization of the genes encoding different microbial nitrile hydratases and amidases involved in nitrile metabolism. α and β are the genes encoding the α and β subunits of nitrile hydratases; P47K, P12K and Orf1038 are opened reading frames encoding proteins whose roles are still not clearly defined

Figure 2.

 Chemical and enzymatic pathways for nitrile hydrolysis

All of these strict amidases also exhibit more or less high activity with some arylamides such as benzamide, phenylacetamide, phenylpropionamide, nicotinamide or others.

3.2 As a function of the amino-acid sequence

Using computer methods for database search and multiple alignment, statistically significant sequence similarities have been identified between several nitrilases, cyanide hydratases, β-alanine synthase and the first group of aliphatic amidases described before (acylamide amidohydrolases from Rhodococcus sp. R312 and Ps. aeruginosa, EC The results, published by Bork and Koonin in 1994, showed that all these proteins, which appear to be involved in the reduction of organic nitrogen compounds and ammonia production, exhibited several conserved motifs (Bork and Koonin 1994). One of which contains an invariant cysteine that is part of the catalytic site in nitrilases. Another highly conserved motif includes an invariant glutamic acid that might also be involved in catalysis. Regarding sequence conservation over the entire length, as well as the similarity in the reactions, we could thus consider that these enzymes constitute a definite family which points to a common catalytic mechanism.

In 1996, Chebrou and coworkers, using similar methods and cluster analysis, compared the amino-acid sequences of the second group of aliphatic amidases (EC; acylamide amidohydrolases from Rhodococcus sp. R312, Ps. chlororaphis B23, R. erythropolis JCM6823, Rhodococcus sp. and R. rhodochrous J1) with the amino acid sequence of several indole-acetamide hydrolases, 6-amino-hexanoate-cyclic-dimer hydrolases (EC and amidases from yeasts and fungi (EC 3.5). Results indicated significant homology between sequences of the enzymes, which also appear to be involved in the reduction of organic nitrogen compounds and ammonia production. All of the enzymes contain a highly conserved motif (GGSS) which may be important in the catalytic mechanism. They also contain several conserved distinct amino acids, such as glycine, aspartic acid and serine located 17, 19 and 23 amino acids downstream from the GGSS signature (Chebrou et al. 1996). Regarding sequence similarities, we could thus consider that these enzymes constitute a single amidase family.

3.3 Evolutionary aspects

As we have seen, the first group of aliphatic amidases described earlier (aliphatic amidases from Rhodococcus sp. R312, Ps. aeruginosa, Arthrobacter sp. J1, Methylophilus methylotrophus and Helicobacter pylori) seems to be descended from an ancestor in common with nitrilases. These amidases are made up of four, six or eight subunits, and have been classified as sulfhydryl enzymes.

Mainly based on various inhibitor studies, amidases belonging to the second group described earlier (aliphatic enantioselective amidases from Rhodococcus sp. R312, Rhodococcus sp. N-774, Rhodococcus sp., R. erythropolis JCM6823, R. rhodochrous J1, R. erythropolis MP50 and Ps. chlororaphis B23) came also to be regarded as sulfhydryl enzymes. Not completely satisfied with this generally accepted interpretation, Kobayashi and coworkers performed a series of site-directed mutagenesis studies on one particular amidase of R. rhodochrous J1. With regard to the presumptive active site residue Cys203, a Cys203 → Ala mutant enzyme still retained 11·5% of the original specific activity. In sharp contrast, glutamic acid substitution of Asp191 reduced the specific activity of the mutant enzyme to 1·33% of the wild-type activity. Furthermore, Asp191 → Asn substitution as well as Ser195 → Ala substitution completely abolished the specific activity. It would thus appear that, among various conserved residues residing within the signature sequence common to all enantioselective amidases, the real active site residues are Asp191 and Ser195 rather than Cys203. In as much as an amide bond (CO–NH2) in the amide substrate is not too far structurally removed from a peptide-bond (CO–NH–), the signature sequences of various amidases were compared with the active site sequences of various types of proteases (see Fig. 3). It was found that aspartic acid and serine residues corresponding to Asp191 and Ser195 of the Rhodococcus amidase are present within the active site sequences of aspartic proteinases (Kobayashi et al. 1997).

Figure 3.

 Alignment over the amidase signature sequence region of all members of the GGSS motif containing amidase family with the aspartic proteinases (Kobayashi et al. 1997). From top to bottom: R312: amidase from Rhodococcus sp. R312; Rho: amidase from Rhodococcus sp.; J1-L: amidase from Rhodococcus rhodochrous J1; N-774: amidase from Rhodococcus sp. N-774; B23: amidase from Pseudomonas chlororaphis B23; Asp-f2: the N-terminal region of calf chymosin; Asp-f1: the N-terminal region of porcine pepsin; Asp-11: the C-terminal region of porcine pepsin; Asp-12: the C-terminal region of calf chymosin; VDHAP: a vitamin D3 hydroxylase-associated protein from cockerel; FAAH: oleamide hydrolase from rat; Yeast: putative amidase from Saccharomyces cerevisiae; EI: EI enzyme from Flavobacterium; IND-P: indoleacetamide hydrolase from Pseudomonas savastanoi; IND-A: indoleacetamide hydrolase from Agrobacterium tumefaciens; IND-B: indoleacetamide hydrolase from Bradyrhizobium japonicum; ACE-O: acetamidase from Aspergillus oryzae; ACE-N: acetamidase from Aspergillus nidulans; Com: amidase from Comamonas acidovorans; Urea: urea amidolyase from Candida utilis. Arrows indicate the amino-acid residues that correspond to Asp191, Ser195 and Cys203 of the Rhodococcus rhodochrous J1 amidase examined in this sequence similarity (Chebrou et al. 1996). Residues highlighted in reverse type are conserved across both amidases and aspartic proteinases in at least four of 20 sequences

We could thus conclude in evolutionary relationships first between aliphatic amidases from the first group described before and nitrilases, and then between the amide bond-cleaving enantioselective amidases (coupled with nitrile hydratase) and the peptide bond-cleaving aspartic proteinases.


4.1 Mechanism of action

Several different classes of enzymes are known to hydrolyse some amide bonds. Many proteolytic enzymes can hydrolyse peptide or amino-acid amides (e.g. trypsin and papain). Bacterial proteinases range from those attacking only a few bonds of complex protein substrates to those with a very wide specificity. But the action of proteolytic enzymes on simple amides may be regarded as incidental to their main physiological role of splitting peptide bonds. However, these reactions have been of great value in investigations of the mechanism of enzyme catalysis, and a comparison of specificities and relative catalytic efficiencies may be illuminating in attempting to trace the evolution of enzymes attacking C–N bonds.

In order to study the mechanism of a reaction involving two substrates (amide and water) and two products (carboxylate and ammonia), the initial concentrations of the substrates in the mixture must be known. In the case of amide hydrolysis, it is difficult to study the mechanism of the reaction, one of the reactants being water. Fortunately, amide-hydrolysing enzymes are often able to transfer the acyl moiety of the amide to hydroxylamine to form hydroxamates. Consequently, Maestracci and coworkers (Maestracci et al. 1986) decided to determine, for the aliphatic amidase from Rhodococcus sp. R312, the mechanism of the acyl transfer reaction described several years before by Clarke (Clarke 1970) with the aliphatic amidase from Ps. aeruginosa. Clarke showed that both acetamide and propionamide were good amidase substrates for hydrolysis and acyl transfer reaction but surprisingly propionamide was the most rapidly hydrolysed substrate whereas acetamide was the most rapidly transferred onto hydroxylamine. Clarke also showed that the amide transferase activity extended to the related acids and esters (e.g. acetate, propionate and ethyl acetate) but the activity for the best ester substrate (ethyl acetate) was only about 1% of that for acetamide. The reactions catalysed by the amidase from P. aeruginosa were thus as follows:

inline image

No activity could be detected with mono or di-N-substituted amides or with thioacetamide where the carbonyl oxygen had been replaced by a sulphur atom.

A lot of work has been done by Clarke and colleagues on some genetic mutations that led to the altered substrate specificity. The mutant B, resulting from a N-nitroso-guanidine (NMG) treatment of the wild-type strain, was shown to exhibit higher activity with butyramide as a substrate. The mutant AI, resulting from a NMG treatment of a L10 strain (strain with constitutive high amidase activity, resistant to catabolite repression, relatively insensitive to repression by amide analogues including acetanilide), was shown to utilize acetanilide (N-phenylacetamide) as a growth substrate. The mutant B and AI amidases did not differ significantly from the wild-type A amidase in overall physicochemical properties, such as molecular weight and number of subunits, although they could be distinguished from the A amidase by the differences in their electrophoretic mobilities at pH 8·5. On the other hand, the three enzymes differed markedly in their substrate specificity. Whereas butyramide was not a substrate for the A amidase, the B and AI enzymes were able to transform butyramide for both hydrolysis and acyl transfer reactions. Whatever the enzyme, propionamide and acetamide were the best substrates for hydrolysis and acyl transfer reaction, respectively. For hydrolysis, propionamide was a 3-, 4- and 18-fold better substrate as acetamide with the A, B and AI amidases, respectively. On the contrary, acetamide was a 5-, 2·4- and 1·1-fold better substrate for the acyl transfer reaction with the A, B and AI amidases, respectively. Lastly, acetanilide was shown to be a substrate for the acyl transfer reaction on hydroxylamine catalysed by the AI amidase.

In all cases, Clarke suggested that the mechanism of action most likely involved the formation of an acyl-enzyme intermediate which was subjected to nucleophilic attack by water or hydroxylamine (Clarke 1970).

In 1986, Maestracci et al. thus determined the mechanism of action of the aliphatic amidase from Rhodococcus sp. R312 with acetamide as the substrate and hydroxylamine as the cosubstrate. These authors effectively showed that the mechanism was of Ping Pong Bi Bi type, extrapolating that the mechanism for amide hydrolysis should be the same (see Fig. 4). Amides react with the enzyme to give acyl-enzyme complexes, which then transfer acyl groups to the cosubstrate (water or hydroxylamine) to lead to formation of carboxylates or hydroxamates.

Figure 4.

 Mechanism of the acyl transfer reaction from amides to hydroxylamine (a) and the amide hydrolysis reaction (b) catalysed by the aliphatic amidase from Rhodococcus sp. R312 (Maestracci et al. 1986)

In a previous paper (Fournand et al. 1997a), we determined that the aliphatic amidase from Rhodococcus sp. R312, also named wide spectrum amidase, was additionally able to catalyse the acyl transfer reaction from short-chain amides to hydrazine (NH2NH2), thus leading to formation of acid hydrazides (RCONHNH2). The reaction mechanism was most likely the same. As regards to enantioselective amidases (GGSS signature-containing amidases), we recently showed that the mechanism of action was also of Ping Pong Bi Bi type, indicating a substituted intermediate form of the enzyme (Fournand et al. 1998).

All of these amidases catalyse the different reactions described before (amide hydrolysis, amide transfer, acid transfer, ester hydrolysis and ester transfer) but the reaction rates with acids or esters as acyl donors are much lower than those obtained with amides as acyl donors (Fournand et al. 1997a, 1998), so that only amide hydrolysis and amide transfer reactions represent a real interest for industrial applications.

4.2 Amide hydrolysis

Nitriles (cyano compounds) are important in synthetic organic chemistry as precursors providing amides and carboxylic acids by hydrolysis. Chemical hydrolysis of nitriles proceeds in two steps. In the step to amides, protons (H+) and/or metal cationic species (M+) work to activate the carbon–nitrogen triple bond, which facilitates the addition of water molecules. In the following step from amides to carboxylates, the C–N bonds are more resistant to the hydrolytic cleavage than the conventional C–O bonds of esters. For this reason, the total hydrolytic procedure requires harsh conditions such as heating at an acidic or alkaline pH. This situation has made the selective transformation difficult, especially with the molecules bearing other acid- or alkaline-labile functionalities. So far, to overcome this problem, enzyme-mediated transformations have been developed. Three classes of enzymes, nitrile hydratases, amidases and nitrilases, have been disclosed (see Fig. 2). It is needless to say that a distinctive advantage is that enzyme-catalysed hydrolysis proceeds under as mild conditions as a neutral pH and room temperature.

As an example, the nitrile hydratase from Rhodococcus rhodochrous J1 is used by the Nitto Chemistry Industry Company Ltd (Japan) to produce 30 000 tons of acrylamide annually (Yamada and Kobayashi 1996). This is said to be the first instance of successful industrial production of a commodity chemical using a microorganism. Research has continued into the application of such enzymes to produce a range of other products such as acrylic acid, nicotinamide, p-aminobenzoic acid, and the antimycobacterial agents isonicotinic acid hydrazide and pyrazinamide (Kobayashi and Shimizu 1994; Yamada and Kobayashi 1996).

Another example is the production of adipic acid (HOOC–(CH2)4–COOH) which is one of the main raw materials for the synthesis of nylon 6:6, a very important industrial polyamide. Its synthesis by chemical means requires large amounts of energy and concentrated acid. The production of this diacid from adiponitrile (NC–(CH2)4–CN) by enzyme-mediated transformation could therefore become very interesting. The different reactions involved with the combined action of nitrile hydratase and amidase are described in Fig. 5.

Figure 5.

 Pathways for the degradation of adiponitrile by the combined action of nitrile hydratase (NHase) and enantioselective amidase (Amd) from Rhodococcus sp. strains

The nitrile hydratase- and amidase-containing Rhodococcus sp. R312 strain was able to produce adipic acid from adiponitrile (Moreau et al. 1993). Mutants of this strain were isolated: one specific mutant (Ad), with a modified cell wall, exhibited a three-fold higher activity with adipamide compared with the wild type strain, and another mutant (ACV2), derived from the latter one, exhibited a 30-fold higher activity with cyanovaleric acid as substrate and a seven-fold higher activity with adipamide as substrate compared with the wild type strain.

The combined action of nitrile hydratases and amidases was also described for the enzymatic synthesis of ammonium lactate from lactonitrile (Jallageas et al. 1980), ammonium acrylate from acrylonitrile (Jallageas et al. 1980), and for the production of rhizobitoxine analogues (Marais et al. 1985).

As we have seen, the GGSS sequence-containing amidases, coupled with nitrile hydratase in nitrile metabolism, were shown to enantioselectively hydrolyse some racemic amides. So far, the substrates concerned are all aromatic compounds (most of them are derived from propionamide) which are hydrolysed in higher-value aromatic acids used in pharmaceuticals or herbicides. As an example, Ciskanik and coworkers characterized in 1995 an amidase from Ps. chlororaphis B23 which enantioselectively hydrolyses several aromatic amides, including 2-phenylpropionamide (enantiomeric excess [ee]=100%), phenylalaninamide (ee=55%), and 2-(4-chlorophenyl)-3-methylbutyramide (ee=96%), but not 2-(6-methoxy-2-naphthyl)propionamide (the amide form of naproxen) (Ciskanik et al. 1995). Hirrlinger and coworkers also characterized, one year later, an amidase from Rhodococcus erythropolis MP50, which converted racemic 2-phenylpropionamide, naproxen amide and ketoprofen amide to the corresponding S-acids with an enantiomeric excess of > 99% and an almost 50% conversion of the racemic amides. The enzyme also hydrolysed different α-aminoamides but without significant enantioselectivity (Hirrlinger et al. 1996).

An additional example is the application of the peptide amidase from oranges (Citrus sinensis L.) for the resolution of racemates. Stelkes-Ritter and coworkers (Stelkes-Ritter et al. 1997) used N-protected amino-acid amides as substrates to determine amidase kinetics for the deamidation and investigated the influence of the side chain of the amino acid on the performance of the enzymatic hydrolysis. Kinetic data do not allow firm conclusions with regard to the influence of the side chain. There was no clear preference for aromatic over aliphatic amino acids nor for charged over uncharged derivatives. This is a marked difference between this plant peptide amidase and the bacterial enantioselective amidase from Rhodococcus sp. R312, which has been shown to exhibit various substrate specificities as a function of amide charge and side-chain structure (Fournand et al. 1998). With regard to peptide amidase enantioselectivity, the enzymatic resolution of racemic N-acetyl amino-acid amides yielded N-acetyl-L-amino acids in optical purity ≥ 99% at complete conversion (Stelkes-Ritter et al. 1997). Thus, the peptide amidase from Citrus sirensis L. is an interesting biocatalyst firstly for the resolution of racemates, but also to exploit the amide function as a protecting group in peptide synthesis.

4.3 Acyl transfer activity

Acyl transfer activity of amide-hydrolysing enzymes has been described long since. Grossowicz and coworkers reported in 1950 the formation of hydroxamic acids (RCONHOH) by the enzyme-catalysed replacement of the amide groups of glutamine and asparagine with hydroxylamine. They also reported that the enzyme, prepared from cell-free extracts of Proteus vulgaris X-19, could split hydroxamic acids (Grossowicz et al. 1950). Later, in 1959, Kimura showed that an enzyme preparation purified 19-fold from Mycobacterium avium catalysed hydroxamate formation from butyric or valeric acid and hydroxylamine (Kimura 1959b). These studies thus allowed the drawing of the general reaction pathway shown in Fig. 6a.

Figure 6.

 General reaction pathway of the combined hydrolysis (water as cosubstrate) and acyl transfer activity (hydroxylamine as cosubstrate) of amide-hydrolysing enzymes. (a) State of the art at the end of the 1950s; (b) state of the art at the end of the 1990s

This original acyl transfer activity has been extensively investigating for the last 10 years with both of the amidase families described before for hydrolysis. In 1986, Thiéry and coworkers used hydroxylamine as amidase cosubstrate in addition to water in the reaction medium. In such conditions, the aliphatic amidase from Rhodococcus sp. R312 was shown to transfer acyl groups of amides, acids and esters to hydroxylamine, following a Ping Pong Bi Bi mechanism, thus leading to the formation of the corresponding hydroxamic acids and ammonia (Thiéry et al. 1986). We recently confirmed these results by a detailed study of the kinetics of this amidase overproduced in an Escherichia coli strain and we demonstrated that a competition for the acyl-enzyme complex deacylation step effectively occurred (Fournand et al. 1997a). All of these results allowed the drawing of a new general reaction pathway (Fig. 6b) with a central role played by the acyl-enzyme complex. The aliphatic amidase from Rhodococcus sp. R312, which belongs to the first group of amidases described before, was shown to catalyse the formation of short-chain aliphatic hydroxamic acids from the corresponding amides with rates higher than those obtained for amide hydrolysis and with interesting bioconversion yields (Table 1). Some trials for discontinuous synthesis of acetohydroxamic acid after amidase immobilization were performed in our laboratory and gave interesting results (Fournand et al. 1997b).

Table 1.   Bioconversion yields obtained with the aliphatic amidase from Rhodococcus sp. R312 (Fournand et al. 1997a) Thumbnail image of

In addition to these results concerning aliphatic amidases, several studies dealt with the acyl transfer activity of enantioselective amidases. Kobayashi and coworkers determined in 1993 the relative activities of the acyl transfer reactions (amide transfer, acid transfer and ester transfer) catalysed by the amidase from Rhodococcus rhodochrous J1. Only amide transfer activity was observed. Compared with that towards benzamide (100%), the activity towards propionamide was 145%. Other amides, namely acetamide (64%), butyramide (734%), isobutyramide (544%), valeramide (943%), acrylamide (542%), crotonamide (139%), nicotinamide (20%), isonicotinamide (26%) and pyrazinamide (39%) functioned as acyl donors (Kobayashi et al. 1993). Hirrlinger and Stolz recently described the use of the amidase from Rhodococcus erythropolis MP50, characterized one year before for its hydrolysis activity (Hirrlinger et al. 1996), for the production of chiral hydroxamic acids subsequently chemically transformed by Lossen rearrangement to chiral amines (Hirrlinger and Stolz 1997). The rates of acyl transfer activity of the purified amidase for the substrates acetamide, phenylacetamide, and 2-phenylpropionamide were higher than the corresponding rates for the hydrolysis reactions.

The most detailed study of the acyl transfer activity of amidase has been performed in our laboratory (Fournand et al. 1998). For this study, large amounts of the enantioselective amidase from Rhodococcus sp. R312 were produced in an Escherichia coli–T7 expression system (Bigey et al. 1999). The amidase was then tested with various amides and turned out to form a very wide range of hydroxamic acids (Table 2). The optimum working pH values were 7 with neutral amides and 8 with α-aminoamides. The kinetic constants demonstrated that the presence of a hydrophobic moiety in the carbon side chain considerably decreased the Km,amide values. Moreover, very high turnover numbers (kcat) were obtained with linear aliphatic amides, whereas branched-side-chain-, aromatic cycle- or heterocycle-containing amides were sterically hindered. In contrast, the highest affinities of the acyl-enzyme complexes for hydroxylamine were obtained with short, polar or unsaturated amides as acyl donors. Only amides and hydroxamic acids, both of which contained amide bonds, were determined to be efficient acyl donors. This amidase was also shown to be L-enantioselective towards α-hydroxy- and α-aminoamides. These results, particularly the high kcat values obtained with aliphatic amides, α-hydroxyamides and α-aminoamides, highlighted the interest in amidases for use in the production of various hydroxamic acids and showed that enzymatic synthesis of such molecules is an interesting alternative to chemical synthesis methods.

Table 2.   Kinetic constants of the acyl transfer reaction on hydroxylamine catalysed by the enantioselective amidase from Rhodococcus sp. R312 (Fournand et al. 1998) Thumbnail image of

Hydroxamic acids are known to possess high chelating properties. Some of them, particularly α-aminohydroxamic acid derivatives (Ikeda et al. 2000), are potent inhibitors of matrix metalloproteases, a family of zinc endopeptidases involved in tissue remodelling (Cawston 1996) (i.e. osteoarthritis, rheumatoid arthritis, corneal ulceration, osteoporosis, periodontitis, tumour growth, metastasis). Some other hydroxamic acids (i.e. α-aminohydroxamic acids, synthetic siderophores, acetohydroxamic acid) have also been investigated as antihuman immunodeficiency virus agents or antimalarial agents or have been recommended for treatment of ureaplasma infections and anaemia (Brown et al. 1978; Gao et al. 1995; Tsafack et al. 1995; Holmes 1996). Recently, hydroxamic acid derivatives have also been shown to act as potent peptide deformylase inhibitors (Apfel et al. 2000). Moreover, some fatty hydroxamic acids have been studied as inhibitors of cyclooxygenase and 5-lipoxygenase with potent anti-inflammatory activity (Hamer et al. 1996). Apart from these medical applications, some hydroxamic acids (particularly polymerizable unsaturated hydroxamic acids and mid-chain or long-chain hydroxamic acids) have also been extensively investigated in wastewater treatment and nuclear technology studies as a way to eliminate contaminating metal ions (Koide et al. 1987; Heitner and Ryles 1992; Lewellyn 1996).

Finally, we can not approach amidase acyl transfer activity without saying a word on the newly found acyl acceptor hydrazine (NH2NH2). So far, only water (H2O) and hydroxylamine (NH2OH) were shown to be efficient acyl acceptors for hydrolysis and acyl transfer activity. But it has been recently demonstrated that in the presence of hydrazine the aliphatic amidase from Rhodococcus sp. R312 and the enantioselective amidase from Rhodococcus rhodochrous J1 were able to catalyse a range of acid hydrazides (RCONHNH2) (Fournand et al. 1997a; Kobayashi et al. 1999). The mechanism of the reaction is most likely the same as that shown in Fig. 3a. To our knowledge, these are the first reports on the acid hydrazide synthesis through enzymatic reactions. As hydroxamic acids, acid hydrazides are potent chelating agents, and they have been extensively investigated. They are known to be tuberculostatic and antileprosy agents (Yale et al. 1953; Thuc-Cuc et al. 1961), antibacterial and antifungal compounds (Shvelashvili et al. 1980), and antitumour and anticancer agents (Adeniyi and Patel 1996). Metal chelates of polyhydrazides have also been studied to determine their utility as thermally stable fibers and films (Frazer and Wallenberger 1964).


Amidases are ubiquitous enzymes in the living world but their role is still not clearly defined. An exception has to be made for the mammalian integral membrane enzyme (which belongs to the GGSS signature-containing amidase family) responsible for the hydrolysis of a number of neuromodulatory fatty acid amides. This fatty acid amide hydrolase (FAAH) is implicated in the catabolism of the endocannabinoid anandamide and the sleep-inducing lipid oleamide, and thus participates in the induction of sleep and analgesia (Patricelli and Cravatt 2000). In prokaryotes, we also know that amidases enable cells to use some amides as both carbon and nitrogen source (Soubrier et al. 1992), but amides are rarely the sole carbon and nitrogen source available in nature. In animals and plants, one could justify the presence of amidases by the considerable importance of an amide bond in biochemistry since many C-terminal amide-containing peptides act as hormones. But what about bacteria and fungi?

Therefore, scientists have been working on amidases for about 50 years without clearly understanding their utility in nature. However they learned a lot about their catalytic capabilities. Amidases are able to catalyse various reactions in vitro such as amide hydrolysis, ester hydrolysis, hydroxamic acid hydrolysis, acid hydrazide hydrolysis, amide transfer on hydroxylamine, amide transfer on hydrazine, ester transfer on hydroxylamine, ester transfer on hydrazine, carboxylic acid transfer on hydroxylamine and carboxylic acid transfer on hydrazine. Some of these reactions are very low (with esters or carboxylic acids as acyl donors) but some others lead to high bioconversion yields. These latter reactions concern C–N bond-containing substrates (amides, hydroxamic acids, acid hydrazides). Kobayashi and coworkers even showed that the amidase from Rhodococcus rhodochrous J1 was surprisingly found to catalyse the hydrolytic cleavage of the C–N triple bond in a nitrile (R–C≡N) to form an acid and ammonium (Kobayashi et al. 1998). This amidase exhibited a Km of 3·26 mmol l–1 for benzonitrile in contrast to that of 0·15 mmol l–1 for benzamide as the original substrate, but the Vmax for benzonitrile was about 1/6000 of that for benzamide. These authors showed that the Ser195 residue plays a crucial role in the hydrolysis of nitriles as well as amides.

In other words, it took half a century to determine the specificity of amidases and to classify them as C–N bond-cleaving enzymes, with water, hydroxylamine or hydrazine (and likely ammoniac) as the deacylating cosubstrate. As we have seen, using computer methods for database search and multiple alignment, some authors (Bork and Koonin 1994; Chebrou et al. 1996; Kobayashi et al. 1997) identified statistically significant sequence similarities first between amidases not coupled with nitrile hydratases and nitrilases and second between enantioselective amidases coupled with nitrile hydratases and the peptide bond-cleaving aspartic proteinases. Both sequence studies and catalysis experiments thus lead to C–N bond-cleaving specificity for amidases. Consequently, one is entitled to wonder whether the current enzymatic activity of amidases could not result from an ancestral separate evolution of two distinct enzyme families, nitrilases and proteinases, so that the residual in vivo activity and the role of current amidases are no more clearly defined and explainable.

Whatever the case, the current amidase activity is of great interest for present and future biotechnological applications. Hydrolysis (i.e. acyl transfer on water) is valorized in chemical industry to produce adipic acid, acrylic acid, p-aminobenzoic acid, lactic acid and rhizobitoxine analogues, and in pharmaceutical industry to produce S or R enantiomers from 2-phenylpropionic acid, naproxen, ketoprofen and some others. Acyl transfer on hydroxylamine is valorized in medicine to provide aminohydroxamic acids or aminohydroxamic acid-containing peptides, acetohydroxamic acid and synthetic siderophores, and in environment to provide polymerizable monomers such as acrylohydroxamic acid and methacrylohydroxamic acid. But numerous studies are still going on and we should see in the near future new applications of amidases.