Introduction

Hydroxynitrile lyases (HNLs; EC 4.1.2.X, X = 10, 11, 46, and 47) catalyze the reversible conversion of cyanohydrins into carbonyl compounds and hydrogen cyanide. In cyanogenic organisms, HNLs play a crucial role in releasing hydrogen cyanide as a defensive agent1. In industry, HNLs have been used to synthesize chiral cyanohydrins2, which are useful building blocks for the synthesis of pharmaceuticals, fine chemicals, and agrochemicals3. To date, most HNLs have been identified from cyanogenic plants4,5, although several bacteria also contain HNLs6,7,8. The structures of HNLs show similarities to those of α/β-hydrolases, oxidoreductases, carboxypeptidases, Zn2+-dependent alcohol dehydrogenases, and Mn2+-dependent cupins8,9,10,11,12. The HNL (PaHNL) from almond (Prunus amygdalus Batsch) shows high specific activity for the synthesis of (R)-mandelonitrile and strong stability; therefore, it is a valuable catalyst in industry13.

Millipedes are a highly diverged group, and more than 12000 species have been described worldwide14. Among them, polydesmid millipedes are generally cyanogenic15. When cyanogenic millipedes are alarmed, (R)-mandelonitrile stored in the reservoir of defensive glands is degraded into benzaldehyde and hydrogen cyanide. Both compounds are secreted externally through ozopores on some segments of the paired paranota16. Recently, an HNL (ChuaHNL) from the invasive swarm-forming polydesmid millipede, Chamberlinius hualienensis Wang, was isolated and characterized17,18. ChuaHNL showed not only the highest specific activity for (R)-mandelonitrile synthesis among all characterized HNLs, but also high enantioselectivity and stability at a range of temperatures and pHs. Other cyanogenic millipedes may also contain HNLs with high specific activity and stability. However, no other millipede HNLs have been identified because the genome sequences of cyanogenic millipedes are not available and ChuaHNL does not show similarity to any other proteins18.

In this study, we purified an HNL from a cyanogenic millipede and determined its cDNA sequence. Next, we designed degenerate primers to amplify cDNAs encoding HNL from cyanogenic millipedes belonging to the Paradoxosomatidae and Xystodesmidae families. All of these eleven cDNAs encoding HNL were heterologously expressed in insect cells, five cDNAs encoding HNL were able to express in Escherichia coli, and transgenic E. coli cells expressing a cDNA encoding HNL were used as whole-cell catalysts to synthesize (R)-mandelonitrile.

Results

Purification of HNL (NttHNL) from Nedyopus tambanus tambanus

We collected Nedyopus tambanus tambanus, which swarms every April at our facility, and extracted and purified NttHNL (Fig. S1). The crude extract from this millipede showed specific activity of 65.9 U/mg in the (R)-mandelonitrile synthetic reaction. Then, NttHNL was purified 38.6-fold from the extract with 20.2% yield. When the purified enzyme was separated by SDS-PAGE, it showed a single band with a molecular weight of 28800 Da (Fig. 1). The relative molecular weight of the native enzyme was determined to be 52400 Da by gel filtration, indicating that this enzyme functions as a dimer, like ChuaHNL18. NttHNL was R-selective HNL and the specific activity of purified NttHNL for (R)-mandelonitrile synthesis was 4702 U/mg, lower than that of ChuaHNL (7420 U/mg)18, but three times higher than that of PaHNL (1450 U/mg)19, which is used industrially. These results suggested that cyanogenic millipedes are a good source of HNLs with high specific activity for cyanohydrin synthesis.

Figure 1
figure 1

SDS-PAGE analysis of HNL purified from N. tambanus tambanus. HNL purified from N. tambanus tambanus was separated by 12% SDS-PAGE with Broad Range (Bio-Rad Laboratories) molecular weight markers. The full-length gel is presented in Fig. S4.

Full size image

Molecular cloning of cDNAs encoding HNLs from cyanogenic millipedes

Based on the protein sequence of NttHNL, the cDNA encoding NttHNL was cloned after 5′- and 3′-rapid amplification of cDNA ends (RACE). The deduced amino acid sequence of NttHNL was 64% identical to that of ChuaHNL, suggesting that HNLs are conserved among cyanogenic millipedes. To clone cDNAs encoding HNLs from cyanogenic millipedes collected from various locations in Japan (Table 1), degenerate primers were designed on the basis of homologous sequences between NttHNL and ChuaHNL. The degenerate primers were used to amplify partial sequences of cDNAs encoding HNLs from Nedyopus tambanus mangaesinus (NtmHNL), Oxidus gracilis (OgraHNL), and Parafontaria tonominea species complex 1 (Pton1HNL). The degenerate primers designed from the conserved sequence of the above-mentioned HNLs also amplified partial sequences of cDNAs encoding HNLs from P. falcifera (PfalHNL), P. tonominea species complex 2 and 3 (Pton2HNL and Pton3HNL, respectively), P. tokaiensis (PtokHNL), Riukiaria semicircularis semicircularis (RssHNL), and Riukiaria sp. (RspHNL). Full-length cDNAs encoding these HNLs were obtained by 5′- and 3′-RACE. Deduced amino sequences of these HNLs showed 45–66% similarities to ChuaHNL at the amino acid level, but showed no similarity to plant and bacterial HNLs. All of the newly identified HNLs from millipedes were predicted to contain glycosylation sites and secretion signal sequences (Table S1), consistent with the fact that ChuaHNL is a glycosylated protein and that HNLs are secreted into reaction chambers of defensive glands18. Amino acid sequence analyses showed that eight Cys residues were conserved among all the millipede HNLs (Fig. 2). Considering that iodoacetic acid inhibits ChuaHNL activity18, Cys residues are likely to be crucial residues in millipede HNLs.

Table 1 Millipedes and their collection sites.
Full size table
Figure 2
figure 2

Amino acid sequence alignment of HNLs from millipedes. Multiple sequence alignment was visualized by ESPript 3 (http://espript.ibcp.fr/ESPript/ESPript/). Red background shows strictly conserved residues; red letters indicate residues well conserved within a group according to a Raisler matrix; and remainder are shown in black. Residues conserved between groups are boxed. Stars indicate Cys residues conserved among millipede HNLs.

Full size image

In the phylogenetic analysis, the HNLs were separated into two branches (Fig. 3) that corresponded to the two families (Xystodesmidae and Paradoxosomatidae). The HNLs from Nedyopus, Parafontaria, and Riukiaria clustered together (Fig. 3). These results suggested that genes encoding HNL likely evolved from one ancestral gene during the evolution of polydesmoid millipedes.

Figure 3
figure 3

Phylogenetic analysis of HNLs from cyanogenic millipedes. Phylogenetic tree was constructed by the neighbor-joining method with 1000 bootstrap replicates. Bar indicates 5% divergence.

Full size image

Recombinant production of millipede HNLs in heterologous hosts

To characterize the HNLs from millipedes, recombinant HNLs were produced in heterologous expression hosts. First, we chose the baculovirus–insect cell expression system, because it has the highest similarity to millipedes in terms of the patterns and capacity of posttranslational modifications. In this system, the millipede HNLs were secreted into the medium and all of them catalyzed the asymmetric synthesis of (R)-mandelonitrile. Their concentrations ranged from 0.4 to 10.5 U/mL culture medium (Table 2). These results confirmed that the cloned cDNAs encoded functionally active HNLs.

Table 2 Heterologous expression of HNL genes in insect cells and E. coli.
Full size table

Next, we tried to express the millipede HNLs in E. coli, which is easily transformed and can be cultured in a short time for large-scale production. However, the millipede HNLs were not expressed in E. coli BL21(DE3) (Table 2), which is generally used for heterologous protein production. The millipede HNLs were found to contain conserved eight Cys residues (Fig. 2), which might be involved in the formation of disulfide bonds. Proteins that require disulfide bonds for their folding and stability have been shown to be poorly expressed, misfolded, or inactive when expressed in the cytoplasm of wild-type E. coli strains20. Therefore, we used the genetically engineered strain E. coli SHuffle T7 as the expression host. This strain constitutively expresses disulfide bond isomerase DsbC, which corrects mis-oxidized proteins20,21. In SHuffle T7, we detected the expression of NttHNL, NtmHNL, OgraHNL, Pton2HNL, and Pton3HNL. Among them, the HNL showing the highest concentration was Pton3HNL (40.3 U/mL culture medium), higher than the highest concentration in the insect cell expression system (Table 2). The specific activities of purified recombinant NttHNL, NtmHNL, OgraHNL, Pton2HNL, and Pton3HNL were 1945 U/mg, 1997 U/mg, 2741 U/mg, 3371 U/mg, and 2140 U/mg, respectively (Table 2), all of which were higher than that of PaHNL (1450 U/mg). When the same amounts of purified enzymes (10 U/mL) were used in 5 min of reaction, (R)-mandelonitrile was obtained with an enantiomeric excess of 85% (NttHNL), 88% (NtmHNL), 83% (OgraHNL), 91% (Pton2HNL), and 91% (Pton3HNL), indicating that recombinant millipede HNLs have excellent stereoselectivity.

Effect of temperature and pH on enzyme activity and stability of recombinant Pton3HNL

Since Pton3HNL showed the highest expression level in E. coli SHuffle T7, we used purified recombinant Pton3HNL to evaluate the effects of temperature and pH on enzyme activity and stability. The optimal temperature and pH ranges of the enzyme were 30–40 °C and 4.5–4.8, respectively (Fig. 4a,b). After a 1-h incubation in 10 mM potassium phosphate buffer (KPB; pH 7.0), Pton3HNL retained full activity between 15 °C and 60 °C (Fig. 4c). After a 1-h incubation at 25 °C, Pton3HNL retained activity over the pH range of 3–10.5 (Fig. 4d). The pH and temperature stabilities of the recombinant Pton3HNL were similar or better than those of ChuaHNL purified from millipedes18.

Figure 4
figure 4

Effect of temperature and pH on recombinant Pton3HNL. (a) Optimum temperature. Reaction was performed at 15–55 °C for 5 min in 300 mM sodium citrate buffer, pH 4.2. (b) Optimum pH. Reaction was performed at 22 °C for 5 min in 300 mM sodium citrate buffer (pH 2.5–5.5). (c) Temperature stability. Remaining activity was measured after incubation of Pton3HNL at various temperatures (15–95 °C) in 10 mM potassium phosphate buffer, pH 7.0, for 60 min. (d) pH stability. Remaining activity was measured after incubation of Pton3HNL at 25 °C for 60 min with the following buffers: 100 mM sodium citrate buffer (pH 3.0–6.0); 100 mM potassium phosphate buffer (pH 6.0–8.0), 100 mM Tricine-NaOH buffer (pH 8.0–9.0), and 100 mM sodium carbonate buffer (pH 9.5–10.5). Values are mean ± SD; n = 3.

Full size image

Circular dichroism (CD) was utilized as an additional tool to examine the pH and thermal stabilities of the purified recombinant Pton3HNL. The CD spectrum of Pton3HNL exhibited that the enzyme contains α-helix (222 nm) and random structures, based on a decrease in mean residue ellipticity at 200–210 nm (Fig. S2). This result is corresponding to the CD spectrum analysis using DichroWeb, which suggested that the enzyme contains 29%, 30% and 41% for α-helices, β-strands, and random structures, respectively. After a 1-h incubation in the pH range of 3.0–10.0 at 25 °C, CD spectra of Pton3HNL were converged together at the pH ranges of 3.0–9.0 (Fig. S2a). The proportion of β-strands at pH 10.0 was slightly decreased (25%) from those (30%) at pH 7.0. After 1-h incubation in KPB (pH 7.0), Pton3HNL was stable at temperature ≤70 °C and the curves converged together (Fig. S2b). In the temperature ranges of 90–100 °C, the curves were different from those at 25–70 °C, the percentages of β-strands were notably decreased (4% at 90 °C and 5% at 100 °C). The results are in agreement with that the recombinant Pton3HNL almost lost its activity after 1-h incubation at 95 °C (Fig. 4c). Taken together, the recombinant Pton3HNL has excellent stabilities toward pH and temperatures.

Synthesis of (R)-mandelonitrile using whole recombinant E. coli cells

Using whole cells as biocatalysts circumvents cell lysis and enzyme purification steps, and hence, substantially cuts costs22. The feasibility of a whole-cell reaction system to produce (R)-mandelonitrile was tested using E. coli SHuffle T7 cells expressing Pton3HNL. In initial assay conditions (pH 5.0) without organic solvents, E. coli cells catalyzed the synthesis of (R)-mandelonitrile with 49.2% enantiomeric excess (Fig. 5-a). It has been reported that low pH suppresses the spontaneous chemical reaction producing racemic mandelonitrile23. Therefore, we synthesized (R)-mandelonitrile in the pH range of 2.5–5.0. The enantiomeric excess of (R)-mandelonitrile synthesis reached 95.1% at pH 3.0 (Fig. 5a). Finally, we optimized the amount of recombinant cells to produce a maximum enantiomeric excess of 97.6% (Fig. 5b). A HPLC chromatogram of (R)-mandelonitrile synthesis under the optimized condition was shown in Fig. S3. These results indicated that Pton3HNL expressed in SHuffle T7 has the potential for industrial use.

Figure 5
figure 5

Synthesis of (R)-mandelonitrile using E. coli cells expressing Pton3HNL in aqueous solution. (a) Relationship between pH and enantiomeric excess (ee). E. coli cells harvested from 0.4 mL culture were reacted in 0.2 mL 300 mM sodium citrate buffer (pH 2.5–5.0) containing 50 mM benzaldehyde and 100 mM KCN at 22 °C for 5 min. (b) Relationship between cell amounts and ee. E. coli cells harvested from 0.4 (2×), 0.8 (4×), or 1.2 mL (6×) were reacted in 0.2 mL 0.3 M sodium citrate buffer, pH 3.0, containing 50 mM benzaldehyde and 100 mM KCN at 22 °C for 5 min.

Full size image

Discussion

Previous studies have focused on HNLs from plants and bacteria5,7, since these HNLs are valuable enzymes for the asymmetric synthesis of cyanohydrins. Recently, an HNL from the cyanogenic millipede C. hualienensis was found to have the highest specific activity toward (R)-mandelonitrile synthesis detected to date, along with excellent temperature and pH stabilities18. The results of the present study showed that HNLs are conserved among cyanogenic millipedes and that recombinant HNLs produced in E. coli have high specific activity and excellent temperature and pH stabilities. The results also showed that recombinant E. coli cells can synthesize asymmetric (R)-mandelonitrile with excellent enantiomeric excess in an aqueous solution.

To use millipede HNLs as biocatalysts, their overexpression in heterologous host is preferable due to the limited availability of millipedes in nature. The baculovirus–insect cell expression system could produce all of eleven millipede HNLs tested as functional proteins. However, growth of cells and cost of medium are inferior to E. coli and yeast systems24. The methylotrophic yeast Pichia pastoris is often utilized as heterologous host of plant HNLs25,26,27,28. The yeast could produce functional ChuaHNL18 yielding 0.9 U/mL culture. The production level was not sufficient. It is known that co-overexpression of genes encoding protein disulfide isomerase and disulfide-rich protein in P. pastoris can increase production level of disulfide-rich protein29,30. This method may improve expression of millipede HNLs in P. pastoris. Although expression of ChuaHNL was not detected in E. coli, five other millipede HNLs were successfully expressed in E. coli in this study (Table 2). The HNL showing the highest concentration in culture medium was Pton3HNL (40.3 U/mL) and the purified recombinant enzyme showed high specific activity (2140 U/mg) and strong temperature and pH stabilities (Fig. 4), like ChuaHNL. Since synthesis of cyanohydrins is usually carried out at low pH to avoid undesirable non-enzymatic reactions, the excellent stability of millipede HNLs at low pH is an attractive trait for industrial use. Glycosylation is crucial for the stability of HNLs from plants such as Prunus amygdalus and Passiflora edulis Sims27,31. ChuaHNL is a glycoprotein18 and the HNLs identified in this study have putative glycosylation sites (Table S1). Although E. coli does not produce glycosylated proteins, recombinant Pton3HNL showed high temperature and pH stabilities (Fig. 4c,d). The millipede HNLs were found to contain eight highly conserved Cys residues (Fig. 2). Considering that E. coli cells were able to produce functional millipede HNLs in the presence of DsbC, but not in its absence (Table 2), Cys residues likely form disulfide bond(s) that stabilize the protein32. Thus, the stability of millipede HNLs is likely due to their disulfide bond(s).

The synthesis of cyanohydrins is usually carried out in aqueous–organic two-phase systems with the enzyme in the aqueous phase23,33,34, or in micro-aqueous organic solvents using immobilized enzymes35. During these processes, laborious steps such as cell disruption, purification, and/or concentration are required to prepare the enzyme solution. Whole-cell biocatalysis can circumvent these laborious steps and simplify the reaction36,37. In this context, we demonstrated that E. coli cells expressing Pton3HNL were able to produce (R)-mandelonitrile with an enantiomeric excess of 97.6%, without using organic solvents (Fig. 5). Similarly, E. coli cells expressing HNL from Arabidopsis thaliana were shown to catalyze the synthesis of (R)-mandelonitrile with an excellent enantiomeric excess in a monophasic micro-aqueous reaction system38. Our whole-cell catalysis utilizing millipede HNLs reduces the need for organic solvents, thereby reducing hazardous waste and environmental pollution39.

The arthropod class Diplopoda (millipedes) contains more than 12000 described species18 and is one of the most ancient groups of terrestrial animals. Among them, polydesmoid millipedes, which comprise more than 5000 species14, generally biosynthesize (R)-mandelonitrile from L-phenylalanine40 and accumulate this compound as a cyanide source to release hydrogen cyanide as a defensive agent15. In this study, we showed that HNLs are conserved among cyanogenic millipedes (Fig. 2) and likely evolved from one ancestral HNL gene during the evolution of polydesmoid millipedes (Fig. 3). Thus, millipedes are a large and diverse source of industrial HNLs with high specific activity and strong temperature and pH stabilities.

Materials and Methods

Millipedes

Millipedes were collected from 10 locations in Honshu and Kyushu, Japan (for details see Table 1). Parafontaria. tonominea is recognized as a species complex since it represents an aggregation of very closely related forms comprising diverse evolutionary states ranging from geographic variation to full reproductive units41,42. In this study, we tentatively assigned species numbers to P. tonominea based on their collection sites (Table 1). Millipedes were reared in litter of Japanese cedar (Cryptomeria japonica D. Don) at room temperature until use.

Enzyme assay

The synthesis of mandelonitrile from benzaldehyde was analyzed by HPLC with a chiral column as described in a previous report18 with slight modifications. The enzyme samples were mixed with 200 µl 300 mM citrate buffer (pH 4.2) containing 50 mM benzaldehyde and 100 mM potassium cyanide. The mixture was incubated at 22 °C for 5 min. Then, 50 µl reactant was added to 450 µl of n-hexane:2-propanol = 85:15, mixed vigorously, and the mixture was centrifuged at 20000 g for 3 min. A 5-µl aliquot of the organic phase was analyzed as described previously40 using a UFLC Prominense Liquid Chromatograph LC-20AD equipped with a CHIRALCEL OJ-H column (particle size: 5 µm; 4.6 mm i.d. × 250 mm; Daicel Corporation, Tokyo, Japan) connected to a Prominence UV-Vis detector SPD-20A (Shimadzu, Kyoto, Japan). The separation conditions were as follows: mobile phase, n-hexane:2-propanol (85:15); flow rate, 1 mL/min. Eluted products were detected by monitoring absorbance at 254 nm.

Purification of HNL from N. tambanus tambanus

NttHNL was extracted from N. tambanus tambanus in phosphate buffered saline (PBS; 1.47 mM KH2PO4, 8.1 mM Na2HPO4, 137 mM NaCl, 2.7 mM KCl, pH 7.4) and then purified. After adding ammonium sulfate to 30% saturation, the crude extract was loaded onto a TOYOPEARL Butyl-650M column (Tosoh, Tokyo, Japan). The adsorbed proteins were eluted by a step-wise gradient of ammonium sulfate in PBS (20%, 10%, and 0% saturation). The active fractions were loaded onto a PD-10 column (GE Healthcare, Little Chalfont, UK) equilibrated with 10 mM Tris-HCl (pH 9.0). The eluted proteins were loaded onto a MonoQ 5/50 GL column (GE Healthcare), and eluted with a linear gradient of sodium chloride (0–100 mM) in the same buffer. The active fractions were pooled and concentrated using a centrifugal filtration device (Amicon Ultra 0.5 Centrifugal Filter Unit with ultracel-10 membrane; Merck Millipore, Billerica, MA, USA). The concentrated enzyme solution was applied to a Superdex 200 10/300 column (GE Healthcare) equilibrated with PBS.

Protein sequencing

The N-terminal amino acid sequencing of purified NttHNL was performed at Nippi Inc. (Tokyo, Japan) using a Procise492HT protein sequencer (Applied Biosystems, Foster city, CA, USA). The internal amino acid sequence was determined as described previously43. Briefly, after separation of purified NttHNL by 12% SDS-PAGE, the protein was visualized by Coomassie Brilliant Blue staining, excised, and then treated with trypsin (sequencing-grade modified trypsin; Promega, Madison, WI, USA). The digested peptides were analyzed using a nanoflow liquid chromatography–tandem mass spectrometry system. Desalting and separation were performed on a nanoACQUITY UPLC Symmetry C18 trap column (particle size: 5 μm; 180 μm i.d. × 20 mm, Waters, Milford, MA, USA) and an ACUITY UPLC Peptide CSH C18 nanoACQUITY column (particle size: 1.7 μm; 75 μm i.d. × 200 mm, Waters), respectively. The mobile phases were water containing 0.1% formic acid (A) and acetonitrile containing 0.1% formic acid (B). The elution profile was a linear gradient of B (10–80%) at a flow rate of 300 nL/min for 90 min. Mass spectrometry was simultaneously performed in positive electrospray ionization mode with a capillary voltage of 3 kV and cone voltage of 40 V using a SYNAPT G2-Si high-resolution quadrupole TOF MS (Waters). [Glu1]-fibrinopeptide B was used as a mass calibrant. The MS/MS spectra were analyzed using BioLynx (Waters).

Molecular cloning of cDNAs encoding NttHNL and HNLs from millipedes

Total RNA was prepared from millipedes using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA) and the Direct-zol MiniPrep kit (Zymo Research, Irvine, CA, USA). RACE-ready cDNA was synthesized using the GeneRacer Kit (Thermo Fisher Scientific) and the SMART RACE cDNA amplification kit (Clontech Laboratories, Palo Alto, CA, USA). In cDNA cloning, degenerate oligonucleotide primers were designed on the basis of N-terminal and internal amino sequences of NttHNL or consensus amino acid sequences between ChuaHNL (GenBank: BAS02094) and NttHNL. The PCRs were performed using Tks Gflex DNA polymerase (Takara, Shiga, Japan) as the Taq DNA polymerase, and the PCR products were cloned into the pCR-Blunt vector (Thermo Fisher Scientific) and sequenced. Primers for 5′-and 3′-RACE were designed from partial sequences of cDNAs encoding HNLs. The RACE products were cloned into the pCR-Blunt vector and sequenced. Finally, the coding sequences of HNLs were amplified and cloned into the pCR-Blunt vector. To avoid PCR-derived errors, more than four independent clones were sequenced. Signal peptide cleavage sites were predicted using the SignalP 4.1 Server44. Glycosylation sites were predicted using the NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/) and the NetOGlyc 4.0 Server (http://www.cbs.dtu.dk/services/NetOGlyc/). The oligonucleotide primers used in this study are shown in Table S2. Sequence data have been submitted to the DDBJ/EMBL/GenBank databases under the accession numbers LC314430 –LC314439.

Phylogenetic analysis

A phylogenetic analysis was performed for the millipede HNLs. Multiple sequence alignment of HNLs was performed using ClustalX45 with default parameters. The phylogenetic tree was constructed by the neighbor-joining method using MEGA 6 software46. The significance level for the phylogenetic tree was assessed by bootstrap testing with 1000 replications.

Production of recombinant HNLs in insect cells

cDNAs encoding HNLs from millipedes were expressed as non-tagged proteins in Sf9 cells. Coding sequences of HNLs were re-amplified by PCR and inserted into the BamHI–HindIII site of the pFastbac1 (Thermo Fisher Scientific) using the In-Fusion HD Cloning kit (Clontech Laboratories). After confirmation of the inserted DNA sequences, the plasmids were used to transform E. coli DH10BAC (Thermo Fisher Scientific) to obtain recombinant bacmids. Recombinant viruses were prepared as described previously47. Briefly, Sf9 cells maintained in Sf-900II-SFM medium (Thermo Fisher Scientific) containing 5% fetal bovine serum (FBS; Gibco BRL, Grand Island, NY, USA) at 28 °C were transfected with recombinant bacmids using X-tremeGENE 9 DNA Transfection Reagent (Roche, Basel, Switzerland). Recombinant viruses were amplified and P2 viruses were used to infect Sf9 cells; the virus preparation was stored at 4 °C. The titer of virus stocks was determined by quantitative real time PCR using ie-1-specific primers48. Viral genomic DNA was extracted from P2 virus stocks using the NucleoSpin Virus kit (Takara). Real-time PCR was performed using an ABI 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) and SYBR premix Ex Taq II (Ti RNaseH Plus; Takara) with ROX as the reference dye. The PCR amplification conditions were as follows: 95 °C for 30 s, and 40 cycles at 95 °C for 5 s and at 60° for 40 s. A melting curve analysis was performed after the assay to verify specificity. The titer of samples was calculated from the standard curve generated from serial dilutions of a sample with known titer.

To produce HNLs, Sf9 cells (1.0 × 106 cells/ml) maintained in Sf-900II-SFM supplemented with 5% FBS were infected with P2 viruses at multiplicity of infection of 1. Cells were removed by centrifugation (500 g, 4 °C, 5 min) at 72 h after infection and the culture supernatants were recovered.

Production of recombinant HNLs in E. coli

cDNAs encoding HNLs from millipedes were expressed in E. coli BL21(DE3) or E. coli SHuffle T7 (New England BioLabs). SHuffle T7 constitutively expresses a chromosomal copy of disulfide-bond isomerase to promote the correction of mis-oxidized proteins20. To express HNLs in the cytosol of E. coli, we re-amplified HNLs without the signal peptide coding sequences and cloned each one into the NdeI–HindIII site of pET28 for expression as an N-terminal His-tagged protein. After confirmation of the inserted DNA sequences, the resultant plasmids were used to transform E. coli BL21(DE3) or SHuffle T7. Transformants were inoculated into LB medium containing kanamycin (50 µg/mL) and glucose (1% w/v), and cultured at 30 °C overnight. Cultures were transferred to 3 L of TB-based autoinduction medium and cultured at 26 °C for 24 h. Then, the cells were harvested by centrifugation to yield approximately 130 g of wet cells and resuspended in 300 mL of 10 mM HEPES-NaOH (pH 8.0) containing 0.1 M NaCl and 20 mM imidazole. After disrupting cells, insoluble materials were removed by centrifugation and the supernatant was collected as the cell-free extract. The extract was loaded onto a HisTrap HP column (GE Healthcare) equilibrated with 10 mM HEPES-NaOH (pH 8.0) containing 0.1 M NaCl and 20 mM imidazole, and then HNL was eluted with a linear gradient of imidazole (20–300 mM in the same buffer). The active fractions were pooled and loaded onto a Resource Q column (GE Healthcare) equilibrated with 10 mM HEPES-NaOH (pH 8.0), and eluted with a linear gradient of sodium chloride (0–300 mM in the same buffer). Protein concentrations were determined using the TaKaRa BCA Protein Assay Kit (Takara) with bovine serum albumin as the standard.

Effect of temperature and pH on the recombinant Pton3HNL

The optimum temperature was evaluated by monitoring the (R)-mandelonitrile synthetic activity for 5 min at various temperature ranges from 15 °C to 55 °C. The optimum pH was determined by measuring the activity for 5 min in 300 mM sodium citrate buffer (pH 2.5–5.5). The thermostability was evaluated by pre-incubating the Pton3HNL at different temperatures (15 °C to 95 °C) in 10 mM KPB (pH 7.0) for 1 h, and then the residual enzyme activity was determined. The pH stability was determined by measuring the remaining activity after incubation for 1 h in buffers of varied pH (3.0–10.5). The buffers used were sodium citrate buffer (pH 3.0 to 6.0, 100 mM), KPB (pH 6.0 to 8.0, 100 mM), Tricine-NaOH buffer (pH 8.0–9.0, 100 mM), and sodium carbonate buffer (pH 9.0 to 10.5, 100 mM).

CD spectroscopy analyses

The recombinant Pton3HNL used for CD measurements had protein concentration of 0.15 mg/mL. Samples were incubated in various buffers (3.0–10.0) at 25 °C or in KPB (pH 7.0) at various temperatures (25 °C–100 °C) for 1 h. Far-UV CD spectra were collected on a JASCO J-720W circular dichroism spectropolarimeter using 2-mm quartz cuvette (Tosoh) containing 500 µL of protein solution under the following condition; temperature, 20 °C; cell length, 0.02 cm; resolution, 0.2 nm; bandwidth, 2.0 nm; sensitivity, 100 mdeg; response, 1 s; speed, 100 nm.min−1; and accumulation, 10–20. Every sample was measured in triplicate and data were analyzed using the DichroWeb49 with the K2D method50.

Production of (R)-mandelonitrile by whole recombinant E. coli cells

E. coli SHuffle cells expressing Pton3HNL were used as whole-cell biocatalysts for the synthesis of chiral cyanohydrins in aqueous buffer. After the culture as described above, cells were harvested from 0.8 mL culture to yield 35.2 ± 3.4 mg of wet cells and resuspended in 150 µl 0.4 M citrate buffer (pH 3.0). After adding 10 µl 1 M benzaldehyde and 20 µl 1 M KCN, the mixture was incubated at 22 °C for 5 min. Reaction products were extracted and analyzed by HPLC as described above. Enantiomeric excess was determined as described in a previous study43.