Introduction

Poly(aspartic acid) (PAA) is a bio-based, biocompatible, biodegradable, and eco-friendly alternative to conventional nonbiodegradable polycarboxylates, such as poly(acrylate). Studies of PAA have focused on the development of synthetic methods, structure analyses, and applications (Low et al. 1996; Freeman et al. 1996; Tang and Wheeler 2001; Ross et al. 2001; Joentgen et al. 2003; Thombre and Sarwade 2005). The thermal synthesis of PAA has been well studied, and the resultant PAA (tPAA) contains a high proportion of β-Asp units (70 %), D-Asp units (50 %), branched units, and irregular end groups as shown in Fig. 1 (Pivcova et al. 1981, 1982; Wolk et al. 1994; Matsubara et al. 1998; Nakato et al. 1998). To effectively demonstrate the functions of polymer materials, it is important that their structure-function relationship be well understood to enable application to the design of materials showing the desired functions. tPAA biodegradability is one of the most vital functions that should be taken into account in practical use. The semicontinuous activated sludge (SCAS), mini-continuous activated sludge (Mini-CAS), and modified sturm (CO2 production) tests have been used to assess tPAA biodegradability in the natural environment (Freeman et al. 1996; Tang and Wheeler 2001). However, despite the direct responsibilities of microorganisms and enzymes in the biodegradation and lifetime control of tPAA, there have been no reports of tPAA biodegradation by isolated microorganisms and enzymes until our work. To elucidate tPAA biodegradation, our research group isolated two tPAA-degrading bacteria, Sphingomonas sp. KT-1 and Pedobacter sp. KP-2, from river water and investigated their tPAA biodegradation behaviors (Tabata et al. 1999, 2000). Furthermore, two isolated enzymes, PAA hydrolase-1 and hydrolase-2, which were surmised to participate in tPAA biodegradation, were characterized and their catalytic mechanisms were examined (Tabata et al. 2001; Hiraishi et al. 2003a, 2003b, 2004, 2009, 2015; Hiraishi and Maeda 2011).

Fig. 1
figure 1

Structures of monomeric units, branched units, and irregular end groups of tPAA polymer

Full size image

As polymer structure control generally results in high functionalization and performance improvement, the creation of structure-controlled PAA is one of the most preferred ways to maximize the potential of PAA as a functional material. Recently, β-peptides have attracted interest as functional materials that exhibit the functions of α-peptides and metabolic stability. In green polymer chemistry, an active research area is the use of purified enzymes for the enzyme-catalyzed synthesis of polypeptides, by taking advantage of their substrate specificities. From the standpoint of eco-friendly and low-cost synthesis of structure-controlled PAA, enzymatic synthesis is expected to become a highly significant synthetic process. Thus, as one of the attractive applications of PAA hydrolases, we performed the enzyme-mediated synthesis of β-linked PAA (β-PAA), which is composed of only β-linkages and belongs to β-peptides, using the unique substrate recognition ability of PAA hydrolase-1 (PahZ1KP-2) from Pedobacter sp. KP-2 (Hiraishi et al. 2011).

This mini-review presents an overview of the biochemical and functional properties of PAA hydrolases. It also focuses on the enzymatic synthesis of structure-controlled β-PAA by taking advantage of the substrate specificity of PahZ1KP-2.

PAA hydrolases from Sphingomonas sp. KT-1

Tabata et al. (1999, 2000) isolated two bacteria, Sphingomonas sp. KT-1 and Pedobacter sp. KP-2, as tPAA-degrading and -assimilating microorganisms from river water. Until now, only Sphingomonas sp. KT-1 and Pedobacter sp. KP-2 have been isolated as microorganisms producing PAA hydrolases. tPAA biodegradation by Sphingomonas sp. KT-1 proceeded endogenously, and tPAA was completely degraded into Asp monomers by its soluble fraction. PAA hydrolase-1 (PahZ1KT-1) purified from the soluble fraction of Sphingomonas sp. KT-1 was the first enzyme related to tPAA metabolism (Tabata et al. 2001). The biochemical, structural, and genetic properties of PahZ1KT-1 are listed in Table 1. The PahZ1KT-1 gene encodes a signal sequence of 35 amino acids, indicating that PahZ1KT-1 is located in the periplasmic space (Hiraishi et al. 2003a). SDS-PAGE of the enzyme purified from the wild strain revealed that the molecular weight of the mature enzyme is approximately 30 kDa, in agreement with that of the amino acid sequence deduced from PahZ1KT-1 gene. Sensitivity to inhibitors and mutagenesis study indicated that PahZ1KT-1 is a Ser-type hydrolase having a Ser141 residue in the lipase box (Gly-Xaa-Ser-Xaa-Gly) as the catalytic center. Nuclear magnetic resonance (NMR) and gel permeation chromatography (GPC) analyses of the products of tPAA hydrolysis by PahZ1KT-1 indicated that the enzyme selectively cleaves the amide bonds between β-Asp units in tPAA to generate oligo(aspartic acid)s (OAA) having molecular weights of a few thousand Daltons via an endo-type process.

Table 1 tPAA- and β-peptide-degrading microorganisms and characteristics of their hydrolases
Full size table

PAA hydrolase-2 (PahZ2KT-1) was purified as a second enzyme that likely participated in the tPAA biodegradation by Sphingomonas sp. KT-1 (Hiraishi et al. 2003b). Table 1 shows several properties of PahZ2KT-1. The predicted polypeptide encodes a preprotein of 425 amino acids containing the first 21 amino acids as the signal peptide, indicating that PahZ2KT-1 is located in the periplasm of cells. The molecular mass of the deduced amino acids of PahZ2KT-1 (42,584 Da) agreed well with the value (42 kDa) determined by SDS-PAGE of the purified protein from the wild strain. PahZ2KT-1 exhibited limited activity for tPAA but was able to hydrolyze OAA. From a viewpoint of the terminal structures, tPAA has the irregular end groups (Fig. 1), while OAA has a freshly generated end groups. PahZ2KT-1 was also able to hydrolyze high-molecular-weight α-poly(L-Asp), which does not have the irregular end groups, to yield Asp monomers. These findings suggest that the irregular end groups of tPAA disturb the exo-mode hydrolysis by PahZ2KT-1. The deduced amino acid sequence of PahZ2KT-1 showed similarities to that of carboxypeptidase G2 that hydrolyzed the C-terminal glutamate moiety of folic acid, thereby supporting the abovementioned exo-mode hydrolysis by PahZ2KT-1. The deduced amino acid sequence also showed similarities to that of a putative peptidase belonging to the metallopeptidase M20/M25/M40 family in Caulobacter crescentus CB15 (63.4 % identity in 413 aa), suggesting that PahZ2KT-1 contained metal ion(s) within its active site, in agreement with the fact that the enzyme was inhibited by EDTA.

Substrate specificities of PahZ1KT-1 and PahZ2KT-1

Research of the enzymatic hydrolysis of well-defined oligomer substrates, including α- and β-tetra(L-Asp)s, demonstrated that PahZ1KT-1 is capable of hydrolyzing oligomers not smaller than trimer composed of β-linkages (Hiraishi et al. 2004). The hydrolysis of well-defined oligomers by PahZ2KT-1 indicated that this enzyme can hydrolyze both β- and α-linked Asp oligomers via an exo-type process and probably cleaves the amide bond at the C-terminus (Hiraishi et al. 2004).

Based on the biological, genetic, and functional characterization of PahZ1KT-1 and PahZ2KT-1, the following mechanisms for the microbial degradation of tPAA by Sphingomonas sp. KT-1 are proposed:

  1. 1.

    Not high-molecular-weight but low-molecular-weight tPAA is internalized into the cell.

  2. 2.

    The internalized polymer is hydrolyzed by PahZ1KT-1 via an endo-type process to generate OAA in the periplasmic space.

  3. 3.

    The resultant OAA is subsequently hydrolyzed into Asp monomers via an exo-type process by PahZ2KT-1 that is possibly located in the periplasm fraction.

  4. 4.

    The resultant monomers are utilized in the Asp metabolic process.

PAA hydrolase-1 from Pedobacter sp. KP-2

PAA hydrolase-1 (PahZ1KP-2) was purified from the soluble fraction of Pedobacter sp. KP-2 and found to be localized in the periplasm fraction (Hiraishi et al. 2009). The properties of PahZ1KP-2 are listed in Table 1. Gene analysis suggested that this enzyme contains a signal peptide sequence (41 aa), supporting its periplasmic localization in the cell. The molecular weight of the mature enzyme deduced from its gene was 30,274 Da, in accord with that determined by SDS-PAGE of the purified enzyme from wild strain. The relative molecular mass of PahZ1KP-2 as estimated by gel filtration was around 31 kDa, suggesting that PahZ1KP-2 is a monomeric enzyme. The enzyme exhibited a temperature optimum of 40 °C, and was, like PahZ1KT-1, inhibited by diisopropyl fluorophosphate (DFP) and phenylmethylsulfonyl fluoride (PMSF). This sensitivity to the two inhibitors suggested that PahZ1KP-2 is also a Ser-type hydrolase; this is strongly supported by the presence of a lipase box containing Ser125 in its deduced amino acid sequence and the complete loss of activity upon amino acid substitution of Ser125 with Ala.

Substrate specificity of PahZ1KP-2

GPC analysis of the products of hydrolysis of tPAA by PahZ1KP-2 demonstrated that products with molecular weights of approximately 4000 Da were accumulated, indicating that this enzyme selectively hydrolyzes a portion of the amide linkages in certain tPAA sequences to yield OAA (Hiraishi et al. 2009). 1H and 13C NMR analyses indicated that this enzyme specifically, but not completely, cleaves the amide bond between β-Asp units in tPAA via an endo-type process.

As stated earlier, as the tPAA molecule contains equivalent moles of D- and L-Asp units, uncommonly occurring sequences in nature [(D-Asp)-(D-Asp), (D-Asp)-(L-Asp), (L-Asp)-(D-Asp)] in addition to the (L-Asp)-(L-Asp) sequence may be formed, which may affect the cleavage of the β-β amide bonds in tPAA by PahZ1KP-2. To reveal the effects of the sequences of L- and D-Asp units on tPAA hydrolysis by PahZ1KP-2, we performed the hydrolysis of β-tri(Asp)s having all possible combinations of L- and D-Asp units by PahZ1KP-2 (Hiraishi et al. 2015). The results provided the following information of its substrate recognition mechanism (Fig. 2). The substrate-binding site of PahZ1KP-2 is composed of at least four subsites (subsites 2, 1, −1, and −2). When Asp units occupy three of the four subsites, amide bond cleavage occurs between subsites 1 and −1. Subsite 1 can recognize only the L-Asp unit, whereas the other subsites can recognize both L- and D-Asp units. PahZ1KP-2 cleaves the amide bond at the carboxyl part of the β-L-Asp unit in stereoisomeric β-tri(Asp)s. Among the dimer sequences, the (L-Asp)-(D-Asp) sequence is the most acceptable to the two central subsites.

Fig. 2
figure 2

Schematic model of substrate recognition site of PahZ1KP-2

Full size image

Functionally and structurally related enzymes to PahZ1KT-1 and PahZ1KP-2

Figure 3 shows the multi-alignment of the deduced amino acids of matured PahZ1KT-1 and PahZ1KP-2 with those of PHB depolymerases using ClustalW2 (GENETYX software). BLAST analysis revealed that the deduced amino acid sequence of matured PahZ1KP-2 had similarity to that of PahZ1KT-1 (39 % identity in 264 aa) (Hiraishi et al. 2009). Based on the ESTHER database (http://bioweb.ensam.inra.fr/esther), PahZ1KT-1 and PahZ1KP-2 are classified into α/β-hydrolase_5 family, which have an α/β-hydrolase fold in their structures and contain putative PHB depolymerase (LpqC) from Bordetella parapertussis. Alignment of the deduced amino acid sequences of PahZ1KT-1 and PahZ1KP-2 demonstrated that the residues possibly composing their catalytic triads, in which the Ser residue formed together with Asp and His residues, were highly conserved (Fig. 3).

Fig. 3
figure 3

Multi-alignment of putative amino acids of matured PAA hydrolases-1 with those of matured PHB depolymerases using ClustalW2 (GENETYX software). Sequences of PAA hydrolases-1 from Sphingomonas sp. KT-1 (PahZ1KT-1) and Pedobacter sp. KP-2 (PahZ1KP-2) and PHB depolymerases from Alcaligenes faecalis AE122 (PhaZAfaAE122) and Pseudomonas lemoignei (PhaZ2Ple) are shown. Identical and conserved amino acids are marked in black and gray, respectively. Box indicates lipase box. Proposed active site residues are marked by closed circles

Full size image

Previous sequence analyses unveiled the similarity of the amino acid sequence of PahZ1KT-1 to those of PHB depolymerases from Alcaligenes faecalis AE122 (PhaZAfaAE122) (26.5 % identity, 257 aa) and Pseudomonas lemoignei (PhaZ2Ple) (25.8 % identity, 244 aa) (Hiraishi et al. 2003a). As shown in Fig. 3, the proposed active site residues are conserved in PahZ1KT-1, PahZ1KP-2, PhaZAfaAE122, and PhaZ2Ple. PHB depolymerases are monomer enzymes having an α/β-hydrolase fold and cleave β-ester bonds in PHB via an endo-type process (Jendrossek and Handrick 2002; Hisano et al. 2006; Wakadkar et al. 2010). Early hydrolysis studies of oligo(3-hydroxybutyrate)s having well-defined sequences yielded the following information of the substrate recognition sites of PHB depolymerases: (i) the active site has four subsites (2, 1, −1, and −2), three of which should be occupied by monomer units for cleavage to occur, and (ii) for the hydrolysis to proceed, subsites 1 and −1 should be occupied by the R-3-hydroxybutyrate (R3HB) unit, whereas the other two subsites may accept both R3HB and S-3-hydroxybutyrate units (Bachmann and Seebach 1999; Hiraishi et al. 2000; Scherer et al. 2000).

Aside from the PAA-hydrolyzing enzymes (PahZ1KT-1, PahZ2KT-1, and PahZ1KP-2), five β-aminopeptidases (three BapA enzymes, one BapF enzyme, and one DmpA enzyme) are known to hydrolyze short β-peptides and β-amino-acid-containing peptides (Geueke et al. 2005, 2006; Heck et al. 2006, 2007, 2012; Geueke and Kohler 2007; Fuchs et al. 2011). Their properties are listed in Table 1. They remove the β-amino acid unit from the N-terminus of oligopeptides, amides, and esters via an exo-type process. BapA enzymes from Sphingosinicella xenopeptidilytica 3-2W4 (3-2W4 BapA) and Sphingosinicella microcystinivorans Y2 (Y2 BapA) hydrolyze β-dipeptides, as well as β-tripeptides, and prefer those with the L-configuration of the N-terminal unit (Geueke et al. 2005, 2006). DmpA from Ochrobactrum anthropi LMG7991 hydrolyzes both α- and β-peptides, but the rate of hydrolysis of α-peptides is lower than that of β-peptides (Heck et al. 2006). In the MEROPS database (Rawlings et al. 2010), β-aminopeptidases are classified into peptidase family P1 that includes aminopeptidases and self-processing proteins.

Figure 4 shows a phylogenetic tree on the basis of amino acid sequences showing relationships among PAA hydrolases-1 (PahZ1KT-1 and PahZ1KP-2), PHB depolymerases (PhaZAfaAE122 and PhaZ2Ple), and β-aminopeptidases [Y2 BapA, Ps BapA (BapA from Pseudomonas sp. MCI3434), 3-2W4 BapA, BapF (BapF from Pseudomonas aeruginosa PAO1), and DmpA]. The analysis suggests that PahZ1KT-1 and PahZ1KP-2 are related to PhaZAfaAE122 and PhaZ2Ple, but not to β-aminopeptidases. Thus, based on these functional and structural findings of PAA hydrolases-1, β-aminopeptidases, and PHB depolymerases, it is assumed that PAA hydrolases-1 share a common ancestor with PHB depolymerases rather than β-aminopeptidases.

Fig. 4
figure 4

Neighbor-joining tree showing phylogenetic relationships among PAA hydrolases-1, PHB depolymerases, and β-aminopeptidases. The scale bar represents the expected number of substitutions per amino acid position

Full size image

Application of PAA hydrolases: PahZ1KP-2-catalyzed synthesis of β-PAA

Generally, the substrate specificity of the enzymes for polymer hydrolysis is tightly linked to that for polymer synthesis. Proteases specifically recognizing α-linked polypeptides are the most commonly used for the enzyme-catalyzed synthesis of poly(amino acid)s, and the resultant poly(amino acid)s are composed of α-amino acid units (Aso et al. 1988; Uemura et al. 1990; Matsumura et al. 1999; Uyama et al. 2002; Soeda et al. 2003; Li et al. 2006, 2008). Due to the novel substrate specificity of PahZ1 enzymes, their application to enzyme-catalyzed polymerization may result in the synthesis of β-linked PAA (β-PAA), which may possess such unexpected properties as high metabolic stability, in keeping with the advantages of α-peptides (Seebach et al. 2004; Seebach and Gardiner 2008). Therefore, PahZ1 enzymes could be used in enzyme-mediated polymerization as one of the more active uses of the enzymes by taking advantage of their substrate specificities.

We reported the first enzyme-catalyzed synthesis of β-PAA using PahZ1KP-2 (Hiraishi et al. 2011). β-PAA synthesis from diethyl L-aspartate substrate was accomplished by using PahZ1KP-2 modified with poly(ethylene glycol) (PEG). Matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF MS) analysis demonstrated that the synthesized polymer produced pairs of peaks having regular peak-to-peak intervals of m/z = 143, which corresponded to the molecular mass of repeat units. The synthesized polymer was observed in the range of m/z = 750–2500 (polymerization degree 5–17), and the most abundant peaks were found at 9 to 11 degrees of polymerization. Moreover, the mass difference between each pair of major and minor peaks was 28, which was equal to that between –OCH2CH3 and –OH, indicating that the product is a mixture of ethyl ester and free carboxyl groups at the C-terminus. 1H NMR analysis revealed that synthesized polymer was composed of only β-amide linkages. Available data led us to conclude that the modification of PahZ1KP-2 with PEG improved dispersibility in organic solvents and the resultant PEG-modified enzyme was useful for the synthesis of β-PAA due to its unique substrate specificity.

Conclusions

In this mini-review, the biochemical and genetic characteristics of PAA hydrolases (PahZ1KT-1, PahZ2KT-1, and PahZ1KP-2) as well as one of their applications, namely, the enzymatic synthesis of structure-controlled β-PAA, were described. To obtain information regarding tPAA biodegradation, tPAA-degrading and tPAA-assimilating microorganisms were isolated from river water and their tPAA biodegradation behaviors were examined. Two PAA-hydrolyzing enzymes, PahZ1KT-1 and PahZ2KT-1, were purified from Sphingomonas sp. KT-1, whereas PahZ1KP-2 was purified from Pedobacter sp. KP-2. Among these enzymes, PahZ1KT-1 and PahZ1KP-2, which showed similarities in their deduced amino acid sequences, caught our interest because they possessed unique substrate recognition properties, that is, these enzymes were able to specifically, but not completely, hydrolyze the amide bond between β-Asp units. Taking into consideration that β-Asp units accounted for 70 % of the total units in tPAA, we assumed that PahZ1 enzymes would play a vital role in tPAA biodegradation in the natural environment. However, the existence of other hydrolase(s), including PahZ2KT-1, may be important for the complete tPAA biodegradation in the natural environment because of the incomplete tPAA degradation by PahZ1 enzymes.

The β-amino acid derived substructures are found as part of a wide variety of bioactive secondary metabolites, such as coenzyme A, L-carnosine, taxol, microcystin-LR, and bestatin, in bacteria, fungi, and plants (Heck et al. 2012). Moreover, β-L-Asp and β-D-Asp units are found in various proteins, including αA- and αB-crystallins, β-amyloid protein, and elastin, from diverse tissues of elderly individuals (Fujii et al. 2011). As there are no known natural peptides that are solely composed of β-amino acid units, compounds containing β-amino acid derived substructures and mixed α,β-peptides produced in nature may serve as physiological substrates for PahZ1 and PahZ2 enzymes as well as DmpA and BapA ones.

In green polymer chemistry, the isolated enzymes have seen increasing applications as catalysts for biopolymer synthesis in vitro. In vitro enzymatic polymerization offers many advantages, such as easier control of polymer structure and monomer reactivity than conventional chemical methods, as well as the enzymes themselves in part of a sustainable system. Novel enzymes capable of cleaving peptides containing β-amino acids, such as PahZ1KT-1 and PahZ1KP-2, could be useful biocatalysts for β-peptide production because of the close relationship between their substrate specificities for polymer hydrolysis and synthesis. Based on this concept, in addition to our work, there have been attempts to synthesize β-peptides in the presence of an enzyme (Heck et al. 2007), and in those cases, exo-type BapA and DmpA enzymes were used. In the future, custom-made enzymes generated by evolutionary engineering may affect the creation of high-performance β-peptides in improved in vitro systems. The development of such enzymes is expected to catalyze dramatic breakthroughs in the industrial, pharmaceutical, and agricultural fields.