Expression, purification and crystallization of N-acetyl-(R)-β-phenylalanine acylases derived from Burkholderia sp. AJ110349 and Variovorax sp. AJ110348 and structure determination of the Burkholderia enzyme

The structure of N-acetyl-(R)-β-phenylalanine acylase from Burkholderia sp. AJ110349 was determined. The three domains showed structural similarity to the corresponding domains of the large subunit of N,N-dimethylformamidase from Paracoccus sp. strain DMF, despite the quaternary structures of these holoenzymes in solution being completely different from each other.

Abstract

N-Acetyl-(R)-β-phenylalanine acylase is an enzyme that hydrolyzes the amide bond of N-acetyl-(R)-β-phenylalanine to produce enantiopure (R)-β-phenyl­alanine. In previous studies, Burkholderia sp. AJ110349 and Variovorax sp. AJ110348 were isolated as (R)-enantiomer-specific N-acetyl-(R)-β-phenyl­alanine acylase-producing organisms and the properties of the native enzyme from Burkholderia sp. AJ110349 were characterized. In this study, structural analyses were carried out in order to investigate the structure–function relationships of the enzymes derived from both organisms. The recombinant N-acetyl-(R)-β-phenylalanine acylases were crystallized by the hanging-drop vapor-diffusion method under multiple crystallization solution conditions. The crystals of the Burkholderia enzyme belonged to space group P4₁2₁2, with unit-cell parameters a = b = 112.70–112.97, c = 341.50–343.32 Å, and were likely to contain two subunits in the asymmetric unit. The crystal structure was solved by the Se-SAD method, suggesting that two subunits in the asymmetric unit form a dimer. Each subunit was composed of three domains, and they showed structural similarity to the corresponding domains of the large subunit of N,N-dimethyl­formamidase from Paracoccus sp. strain DMF. The crystals of the Variovorax enzyme grew as twinned crystals and were not suitable for structure determination. Using size-exclusion chromatography with online static light-scattering analysis, the N-acetyl-(R)-β-phenylalanine acylases were clarified to be dimeric in solution.

1. Introduction

β-Phenylalanine (3-amino-3-phenylpropionic acid, β-Phe) is a β-amino acid that is present in several bioactive molecules such as antibiotics and enzyme inhibitors (Achilles et al., 2000 ▸; Nakao et al., 2001 ▸). Enantiopure β-Phe and its derivatives are important chiral building blocks for the synthesis of pharmaceuticals. For example, the (R)-enantiomer of β-Phe is a key intermediate in the preparation of the side chain of paclitaxel, a complex diterpene isolated from the bark of Taxus brevifolia that possesses strong anticancer activity (Schiff & Horwitz, 1980 ▸). Several enzymatic methods have been proposed for the enantiomeric separation of β-Phe and its derivatives: (i) enantiomer-specific hydrolysis of the β-Phe ester using lipase (Faulconbridge et al., 2000 ▸), (ii) enantiomer-specific hydrolysis of the N-acetyl-(R,S)-β-Phe ester using chymotrypsin (Cohen & Weinstein, 1964 ▸), (iii) enantiomer-specific hydrolysis of N-phenylacetyl-β-Phe using penicillin G acylase (Soloshonok et al., 1995 ▸), (iv) enantiomer-specific hydrolysis of N-chloro­acetyl-β-Phe using porcine kidney acylase (Cardillo et al., 1998 ▸), (v) enantiomer-specific hydrolysis of N-glutaryl-β-Phe using glutaryl-7-aminocephalosporanic acid acylase (Cardillo et al., 2000 ▸) and (vi) enantiomer-specific aminotransfer from β-Phe to 2-oxo acids using β-Phe aminotransferase (Mano et al., 2006 ▸; Crismaru et al., 2013 ▸).

In a previous study, enantiomer-specific amide hydrolysis of N-acetyl-(R)-β-Phe using N-acetyl-(R)-β-Phe acylase (β-FAA) has been proposed as an alternative enzymatic method for the production of enantiopure β-Phe (Kawasaki et al., 2006 ▸; Fig. 1 ▸). To the best of our knowledge, only Burkholderia sp. AJ110349 and Variovorax sp. AJ110348 have been found to be β-FAA-producing organisms. The enantiomeric excess values for N-acetyl-(R)-β-Phe of β-FAA in cell-free extracts prepared from both organisms were >99.5%, with a high molar conversion yield (94–96%; Kawasaki et al., 2006 ▸). In our previous study, the native β-FAA produced by Burkholderia sp. AJ110349 was purified and characterized, showing that it had strict (R)-enantiomer specificity (Imabayashi et al., 2016 ▸). The β-FAA gene was cloned from the Burkholderia sp. AJ110349 genome, and recombinant expression in Eschericia coli and purification of this enzyme were attempted for further investigations, but failed because the expression level was insufficient due to the expression system used in the study. In spite of its low expression in E. coli, the recombinantly expressed protein exhibited β-FAA activity (Imabayashi et al., 2016 ▸).

Figure 1.

The (R)-enantiomer-specific amide hydrolysis of N-acetyl-β-Phe by β-FAAs.

The primary structure of β-FAA from Burkholderia sp. AJ110349 (B-β-FAA; GenBank accession No. CAS03316) has 34% sequence identity and 48% similarity to that of the large subunit of N,N-dimethylformamidase (DMFase) from Alcaligenes sp. KUFA-1 (Hasegawa et al., 1999 ▸). It was reported that the primary structures of the DMFases from Alcaligenes sp. KUFA-1 and Paracoccus sp. strain DMF have the same length as each other and 99% identical residues, and require not only the large subunit but also another small subunit for their activity (Hasegawa et al., 1999 ▸; Arya et al., 2020 ▸). From structural studies using cryo-EM and crystallography, it was revealed that the DMFase derived from Paracoccus sp. strain DMF (parDMFase) forms an (αβ)₂ heterotetramer composed of two large 84 kDa subunits and two small 16 kDa subunits (PDB entry 6lvv; Arya et al., 2020 ▸). Although the function of the small subunit remains unclear, it is most likely to play a role in structural stabilization of an enzymatically active form (Arya et al., 2020 ▸). On the other hand, B-β-FAA expressed in E. coli exhibited activity without requiring another subunit, showing that it has a different quaternary structure to the DMFases. Since the molecular mass deduced from the 760 amino-acid residues of B-β-FAA was calculated to be 81.4 kDa and activity of the native enzyme was detected in size-exclusion chromatography at positions corresponding to the estimated molecular mass of 206 kDa, B-β-FAA was thought to have a homodimeric or homotrimeric structure (Imabayashi et al., 2016 ▸).

In another study, the gene for β-FAA from Variovorax sp. AJ110348 (V-β-FAA) was cloned by the shotgun cloning method as described by Suzuki et al. (2010 ▸). The V-β-FAA gene (GenBank accession No. HW373125) encoded 779 amino-acid residues (GenBank accession No. CAS03318) and the molecular mass of the peptide was calculated to be 85.4 kDa. V-β-FAA recombinantly expressed in E. coli also exhibited the same strict (R)-enantiomer specificity as recombinant B-β-FAA (Suzuki et al., 2010 ▸). Interestingly, fewer residues are conserved between the two β-FAAs than are conserved between β-FAA and the large subunit of DMFase. The primary structure of V-β-FAA has 25% sequence identity and 37% similarity to that of B-β-FAA, while it has 28% sequence identity and 40% similarity to that of the large subunit of the DMFases from Alcaligenes sp. KUFA-1 and Paracoccus sp. strain DMF. Since both β-FAA and DMFase catalyze amide-hydrolysis reactions and have similar primary structures, the ternary structures of β-FAA and the large subunit of DMFase would have a similar fold despite their differing substrate specificities and quaternary structures. Aiming to reveal the structural basis determining the strict (R)-enantiomer specificity, we initiated structural studies of these β-FAAs.

In order to overexpress the proteins, recombinant expression of B-β-FAA and V-β-FAA was first attempted using the T7 promoter expression system in E. coli, but failed because most of the recombinantly expressed proteins formed inclusion bodies and were found in the insoluble fraction. In this study, therefore, alternative expression plasmids for β-FAAs were constructed using a cold-shock expression system. The enzymes were recombinantly expressed, purified and crystallized for structural studies. The quaternary structures of these enzymes in solution were also analyzed.

2. Materials and methods

2.1. Macromolecule production

The B-β-FAA gene (GenBank accession No. HW373123) and the V-β-FAA gene (GenBank accession No. HW373125) were amplified by PCR using genomic DNA of the individual source as a template. A primer pair with forward primer 5′-GGAACGGTACCCATATGATCACGATCACCGGATACAGCGAC-3′ and reverse primer 5′-ATTCCGGATCCTCAGTCGACCTCGGTGTGAGCGAC-3′ was used to amplify the B-β-FAA gene. The PCR product was digested with KpnI and BamHI and the obtained fragment was cloned into pUC18 vector at the corresponding restriction-enzyme sites. Another primer pair with forward primer 5′-CGGTAAGCTTAAGGAGGAACGCGGCATGACGATG-3′ and reverse primer 5′-GAGGCCGGTCTCTAGAGTTCTGAAAGGTCGG-3′ was used to amplify the V-β-FAA gene. This primer pair was designed so that the PCR fragment would contain the upstream and downstream regions of the V-β-FAA gene. The PCR product was digested with HindIII and XbaI and the obtained fragment was cloned into pUC19 vector at the corresponding restriction-enzyme sites.

In order to construct a plasmid, namely pColdI-B-His-β-FAA, that is capable of expressing amino-terminally 6×His-tagged B-β-FAA (B-His-β-FAA), the pUC18 plasmid harboring the B-β-FAA gene was digested with NdeI and BamHI and the obtained fragment was cloned into pColdI expression vector (TaKaRa) at the corresponding restriction-enzyme sites. In order to construct a plasmid, namely pColdIII-B-β-FAA-His, that is capable of expressing carboxyl-terminally 6×His-tagged B-β-FAA (B-β-FAA-His), the B-β-FAA gene was amplified by PCR using a primer pair with forward primer 5′-ATCACAAAGTGCATATGATCACGATCACCGGATACAGCG-3′ and reverse primer 5′-AGCTTGAATTCGGATCCTCAATGATGATGATGATGATGGTCGACCTCGGTGTGAGCGA-3′ and pColdI-B-His-β-FAA as a template. The PCR product was digested with NdeI and BamHI and the obtained fragment was cloned into pColdIII expression vector (TaKaRa) at the corresponding restriction-enzyme sites.

In order to construct a plasmid, namely pColdI-V-His-β-FAA, that is capable of expressing amino-terminally 6×His-tagged V-β-FAA (V-His-β-FAA), the β-FAA gene was amplified from the pUC19 plasmid harboring the V-β-FAA gene by PCR using a primer pair with forward primer 5′-CGCCATATGACGATGCAGCAGCAGAAGATC-3′ and reverse primer 5′-AGTGAATTCATGGTGCTGCGTGCTCCAGGGA-3′. The PCR product was digested with NdeI and EcoRI and the obtained fragment was cloned into pColdI expression vector at the corresponding restriction-enzyme sites. In order to construct a plasmid, namely pColdIII-V-β-FAA-His, that is capable of expressing carboxyl-terminally 6×His-tagged V-β-FAA (V-β-FAA-His), the β-FAA gene was amplified by PCR using a primer pair with forward primer 5′-­ATCACAAAGTGCATATGACGATGCAGCAGCAGAAGATCC-3′ and reverse primer 5′-TCGACAAGCTTGAATTCTCAATGATGATGATGATGATGTGGTGCTGCGTGCTCCAGGG-3′ and pColdI-V-His-β-FAA as a template. The PCR product was digested with NdeI and EcoRI and the obtained fragment was cloned into pColdIII expression vector at the corresponding restriction-enzyme sites. Macromolecule-production information is summarized in Table 1 ▸ and Supplementary Table S1.

Table 1. Macromolecule-production information.

	B-His-β-FAA	B-β-FAA-His, B-Se-β-FAA-His
Source organism	Burkholderia sp. AJ110349	Burkholderia sp. AJ110349
DNA source	Burkholderia sp. AJ110349 genomic DNA	pColdI-B-His-β-FAA
Forward primer†	5′-GGAACGGTACC CATATGATCACGATCACCGGATACAGCGAC-3′	5′-ATCACAAAGTGCATATGATCACGATCACCGGATACAGCG-3′
Reverse primer†	5′-ATTCCGGATCCTCAGTCGACCTCGGTGTGAGCGAC-3′	5′-AGCTTGAATTCGGATCCTCAATGATGATGATGATGATGGTCGACCTCGGTGTGAGCGA-3′
Cloning vector	pUC18	—
Expression vector	pColdI	pColdIII
Expression host	E. coli strain BL21 (B-His-β-FAA)	E. coli strain BL21 (B-β-FAA-His), E. coli strain B834 (B-Se-β-FAA-His)
Complete amino-acid sequence of the construct produced‡	MNHKV HHHHHHIEGRHMITITGYSDVLSAGPGETVEFKVSSKSPHPFTAELVRVIHADPNPAGPGMRFEPLGQVFSGTFASFDKPLLPGSFARVSGVPAAGSAAGLVAGARIRPTALARGDQCVMSQWNTARHAGFALLVSERGLELRLGAGTGEPPVCVLCAARLEVRWYDVWFAIDTASNRIEVGVTEVDGSVAAPVRHRTLQMLDARWRAPHSDDAADLLIGALEDGAGRRAHFNGQIEAPFVADALPSPATPAATVEYAAPRASDFSTDALYAAWDFARGIDTLKIADTTPHARHGTLQNLPTRAVRSSAWNGRERCWRTAPAHYAAIHFHDDDLHDAGWSTDFAFTVPATLKSGAYAMRLSVDGATDYLPFYVRPELGRPGAPLVFVAATYTYQAYANYARGNFDAALRDKVGRWGAYPHNPDDHPEVGLATYNLHSDGSGVMFSSRLRPMLTMRPGFLTFDDSRGSGCRHYIADSHLLDWLEHEGFSFDVVTDDDLERFGAALLEPYAAVLTGTHPEYHTAATLDALAGYKRSGGNLAYLGGNGFYWRVGRSERVPGALEVRRTEGGVRAWAAEAGEYFHALDGEYGGLWRSSARTPQQLVGVGFSSQGPFEGSHYRVLDAARSQPGGSLLKDIAGPLFGGYGLSGGGAAGFELDSTEAADGTPANVIILARSESHSAAFGPALDALLSHTATRARKTPDTLIRSEIVYYETGYGGAVFSVGSITFCGALSHNDYRNDVSTLLRNVLIRFSRDRGAQAHAVPAVAHTEVD	MNHKVHMITITGYSDVLSAGPGETVEFKVSSKSPHPFTAELVRVIHADPNPAGPGMRFEPLGQVFSGTFASFDKPLLPGSFARVSGVPAAGSAAGLVAGARIRPTALARGDQCVMSQWNTARHAGFALLVSERGLELRLGAGTGEPPVCVLCAARLEVRWYDVWFAIDTASNRIEVGVTEVDGSVAAPVRHRTLQMLDARWRAPHSDDAADLLIGALEDGAGRRAHFNGQIEAPFVADALPSPATPAATVEYAAPRASDFSTDALYAAWDFARGIDTLKIADTTPHARHGTLQNLPTRAVRSSAWNGRERCWRTAPAHYAAIHFHDDDLHDAGWSTDFAFTVPATLKSGAYAMRLSVDGATDYLPFYVRPELGRPGAPLVFVAATYTYQAYANYARGNFDAALRDKVGRWGAYPHNPDDHPEVGLATYNLHSDGSGVMFSSRLRPMLTMRPGFLTFDDSRGSGCRHYIADSHLLDWLEHEGFSFDVVTDDDLERFGAALLEPYAAVLTGTHPEYHTAATLDALAGYKRSGGNLAYLGGNGFYWRVGRSERVPGALEVRRTEGGVRAWAAEAGEYFHALDGEYGGLWRSSARTPQQLVGVGFSSQGPFEGSHYRVLDAARSQPGGSLLKDIAGPLFGGYGLSGGGAAGFELDSTEAADGTPANVIILARSESHSAAFGPALDALLSHTATRARKTPDTLIRSEIVYYETGYGGAVFSVGSITFCGALSHNDYRNDVSTLLRNVLIRFSRDRGAQAHAVPAVAHTEVDHHHHHH

^†Restriction-enzyme sites (KpnI, NdeI, BamHI) are underlined in the primers. An additional sequence coding for a His tag is indicated in bold.

^‡Amino-acid residues coded by the translation-enhancing element derived from pCold vector are underlined. His tags are indicated in bold.

E. coli BL21 cells were transformed with each expression plasmid mentioned above. The transformed cells were individually cultured at 310 K in LB medium containing 50 µg ml⁻¹ ampicillin sodium to express B-His-β-FAA, B-β-FAA-His, V-His-β-FAA and V-β-FAA-His. To produce selenomethionine-substituted B-β-FAA-His (B-Se-β-FAA-His), E. coli B834 was used as a host. E. coli B834 cells transformed with pColdIII-B-β-FAA-His were cultured at 310 K in LeMaster medium (LeMaster & Richards, 1985 ▸) containing 50 µg ml⁻¹ ampicillin sodium to express B-Se-β-FAA-His. When the OD₆₀₀ of each culture reached 0.4–0.6, it was cooled to 288 K in ice water. Isopropyl β-d-1-thiogalactopyranoside was then added to final concentrations of 0.02 and 0.2 mM to induce the expression of His-tagged B-β-FAAs and His-tagged V-β-FAAs, respectively, and the culture was grown for a further 18 h at 288 K. The cells were harvested by centrifugation and the pellets were resuspended in buffer A [50 mM sodium phosphate pH 7.0, 500 mM sodium chloride, 10%(v/v) glycerol] containing 20 mM imidazole and homogenized by sonication. Cell debris was removed at 277 K by centrifugation at 13 000g for 30 min and the supernatant was used for purification.

B-β-FAA-His was purified by nickel-affinity chromatography, strong anion-exchange chromatography and size-exclusion chromatography at 277 K. Supernatant containing B-β-FAA-His was applied onto a HisTrap HP 5 ml column (GE Healthcare) equilibrated with buffer A containing 20 mM imidazole. After the unbound proteins had been washed out with buffer A containing 40 mM imidazole, proteins were eluted with buffer A containing 300 mM imidazole. The fractions containing B-β-FAA-His were collected and the solvent was exchanged to buffer B [20 mM Tris–HCl pH 8.0, 5%(v/v) glycerol] by ultrafiltration. Next, the obtained sample was applied onto a HiTrap Capto Q 5 ml column (GE Healthcare) equilibrated with buffer B. Proteins were eluted with a linear gradient of 0.0–1.0 M sodium chloride. The fractions containing B-β-FAA-His were concentrated to approximately 8–10 mg ml⁻¹ and applied onto a Superdex 200 10/300 GL size-exclusion chromatography column (GE Healthcare) equilibrated with buffer C [20 mM Tris–HCl pH 8.0, 150 mM sodium chloride, 5%(v/v) glycerol]. Finally, purified B-β-FAA-His was concentrated to 2–4 mg ml⁻¹ by ultrafiltration. B-His-β-FAA was purified using the same procedures as used for the purification of B-β-FAA-His.

B-Se-β-FAA-His was purified by four column chromatography steps at 277 K. The procedures for the purification of B-Se-β-FAA-His were same as for B-β-FAA-His, except for an additional high-resolution anion-exchange chromatography step. The first nickel-affinity chromatography step and the second strong anion-exchange chromatography step were performed as in the purification of B-β-FAA-His described above. The fractions containing B-Se-β-FAA-His were collected and the solvent was exchanged to buffer B by ultrafiltration. The obtained sample was then applied onto a RESOURCE Q 1 ml column (GE Healthcare) equilibrated with buffer B. Proteins were eluted with a linear gradient of 0.0–1.0 M sodium chloride. After the selected fractions had been concentrated to approximately 8–10 mg ml⁻¹, B-Se-β-FAA-His was further purified by size-exclusion chromatography as for B-β-FAA-His. Finally, the purified B-Se-β-FAA-His was concentrated to 2–4 mg ml⁻¹ by ultrafiltration.

V-His-β-FAA and V-β-FAA-His were purified by four column chromatography steps at 277 K. The procedures for the purification of both V-His-β-FAA and V-β-FAA-His were the same as for B-β-FAA-His, except for an additional weak anion-exchange chromatography step. Firstly, V-His-β-FAA and V-β-FAA-His were individually purified by nickel-affinity chromatography as used for B-β-FAA-His. The fractions containing His-tagged V-β-FAAs were collected and the solvent was exchanged to buffer B by ultrafiltration. The obtained samples were individually applied onto a TOYOPEARL DEAE-650M 40 ml column (Tosoh Bioscience) equilibrated with buffer B. Proteins were eluted with a linear gradient of 0.0–1.0 M sodium chloride. The solvent of the collected fractions was again exchanged to buffer B by ultrafiltration. The obtained samples were then individually purified by strong anion-exchange chromatography and size-exclusion chromatography as used for B-β-FAA-His. Finally, the purified His-tagged V-β-FAAs were concentrated to 2–5 mg ml⁻¹ by ultrafiltration. All buffer exchanges and sample concentrations by ultrafiltration mentioned above were carried out at 277 K using Amicon Ultra-15 Centrifugal Filter Units (Merck) with a molecular-weight cutoff of 50 kDa.

2.2. Crystallization

Initial crystallization screening was performed by the hanging-drop vapor-diffusion method using 24-well VDX plates (Hampton Research) with NeXtal Tubes Classics Suite (Qiagen) at 293 K. A 2 µl droplet was prepared by mixing equal volumes of protein solution and reservoir solution and was equilibrated against 0.5 ml reservoir solution. The composition of the reservoir solution for crystallizing B-β-FAAs was optimized by varying the buffer type and pH (HEPES–NaOH pH 7.0–8.0, MES–NaOH pH 6.0–6.5, Tris–HCl pH 7.0–8.0) and the concentration of the precipitant (2.0–3.0 M ammonium formate, 1.4–2.0 M magnesium sulfate heptahydrate). The composition of the reservoir solution for crystallizing V-β-FAA was optimized by varying the buffer type and pH (MES–NaOH pH 6.0–7.0, Tris–HCl pH 7.0–8.0), the concentration of precipitant [10–40%(w/v) polyethylene glycol monomethyl ether 5000] and the concentration of salt (0–3.0 M ammonium sulfate). The protein concentration was also optimized by varying it in the range 2–5 mg ml⁻¹. The crystallization conditions under which larger crystals grew are summarized in Table 2 ▸.

Table 2. Crystallization.

	B-His-β-FAA, B-β-FAA-His, B-Se-β-FAA-His			V-β-FAA-His
Method	Hanging-drop vapor diffusion			Hanging-drop vapor diffusion
Plate type	24-well VDX plate			24-well VDX plate
Temperature (K)	293			293
Protein concentration (mg ml)−1	2.0–3.5			2.0–3.0
Buffer composition of protein solution	20 mM Tris–HCl pH 8.0, 150 mM NaCl, 5%(v/v) glycerol			20 mM Tris–HCl pH 8.0, 150 mM NaCl, 5%(v/v) glycerol
Composition of reservoir solution	A: 0.1 M HEPES–NaOH pH 7.73, 2.0–2.7 M ammonium formate	B: 0.1 M HEPES–NaOH pH 7.73, 1.5–2.0 M magnesium sulfate heptahydrate	C: 0.1 M Tris–HCl pH 7.5, 1.5–2.0 M magnesium sulfate heptahydrate	0.1 M MES–NaOH pH 6.5–6.8, 25–30%(w/v) polyethylene glycol monomethyl ether 5000, 0.2 M ammonium sulfate
Volume and ratio of drop	2 µl, 1:1	2 µl, 1:1	2 µl, 1:1	2 µl, 1:1
Volume of reservoir (µl)	500	500	500	500

2.3. Data collection and processing

Data collections were carried out on beamline BL-5A at the Photon Factory (PF), KEK, Tsukuba, Japan. The crystals were picked up from droplets using nylon loops and directly flash-cooled at 100 K in a nitrogen gas stream. Since apparent ice-ring diffraction was not observed, no additional cryoprotectant was used in this study. Diffraction data were recorded on an ADSC Quantum 315r CCD detector. The collected data were processed and scaled using HKL-2000 (Otwinowski & Minor, 1997 ▸), XDS (Kabsch, 2010 ▸) and the CCP4 suite (Winn et al., 2011 ▸).

2.4. Structure solution and refinement

Phase determination of the B-Se-β-FAA-His crystal was performed with Phenix AutoSol (Terwilliger et al., 2009 ▸). Modeling and crystallographic refinement were performed with Coot (Emsley et al., 2010 ▸) and phenix.refine (Afonine et al., 2012 ▸), respectively. Molecular replacement was performed with Phaser-MR (McCoy et al., 2007 ▸).

2.5. Molecular-weight analysis

The molecular weight of His-tagged B-β-FAAs and V-β-FAAs were analyzed by size-exclusion chromatography with online static light scattering (SEC–SLS; Wen et al., 1996 ▸). Bovine serum albumin (BSA) was used to determine the instrument calibration constant. The molecular extinction coefficients that were calculated based on the amino-acid sequence and used in this study were 1.368, 1.689 and 0.667 (mg ml⁻¹)⁻¹ cm⁻¹ for B-β-FAAs, V-β-FAAs and BSA, respectively, at 280 nm. The protein concentration of the samples was adjusted to 4.0 mg ml⁻¹, and 20 µl of the samples was injected into a Superdex 200 10/300 GL size-exclusion chromatography column equilibrated with buffer C. Protein concentration was calculated from the absorbance at 280 nm detected with a SPD-20A UV–Vis detector (Shimadzu). The static light-scattering data were measured with a DAWN HELEOS II multi-angle static light-scattering detector (Wyatt Technology). The weight-averaged molecular weight was computed from the protein concentration and the intensities of light-scattering using the ASTRA software version 6.1.7.15 (Wyatt Technology).

3. Results and discussion

3.1. Macromolecule production

B-His-β-FAA and B-β-FAA-His were expressed in E. coli BL21 and purified by three column chromatography steps. SDS–PAGE analysis showed a single band, indicating that B-His-β-FAA and B-β-FAA-His were purified to near-homogeneity (Supplementary Figs. S1a and S1b ). The final protein yields were approximately 1.5 mg from 800 ml of cultured cells. B-Se-β-FAA-His was expressed in E. coli B834 and purified by four column chromatography steps. SDS–PAGE analysis showed that B-Se-β-FAA-His was less pure than B-His-β-FAA and B-β-FAA-His (Supplementary Fig. S1c ). The final protein yields were approximately 1.0 mg from 800 ml of cultured cells. V-His-β-FAA and V-β-FAA-His were expressed in E. coli BL21 and purified by four column chromatography steps. SDS–PAGE analysis showed a single band, indicating that V-His-β-FAA and V-β-FAA-His were purified to near-homogeneity (Supplementary Figs. S2a and S2b ). The final protein yields were approximately 2.0 mg from 800 ml of cultured cells.

3.2. Crystallization

In the initial crystallization screening for B-β-FAAs, small crystals of B-His-β-FAA and B-β-FAA-His were observed with NeXtal Tubes Classics Suite solution No. 28 composed of 0.1 M HEPES–NaOH pH 7.5, 2.0 M ammonium formate and with NeXtal Tubes Classics Suite solution No. 62 composed of 0.1 M MES–NaOH pH 6.5, 1.6 M magnesium sulfate heptahydrate. By optimizing these crystallization conditions, the growth of larger crystals was observed. The reservoir compositions that gave larger crystals of B-β-FAAs were as follows: reservoir solution A (0.1 M HEPES–NaOH pH 7.73, 2.0–2.7 M ammonium formate), reservoir solution B (0.1 M HEPES–NaOH pH 7.73, 1.5–2.0 M magnesium sulfate heptahydrate) and reservoir solution C (0.1 M Tris–HCl pH 7.5, 1.5–2.0 M magnesium sulfate heptahydrate). Suitable protein concentrations were 2.0–3.5 mg ml⁻¹. Using reservoir solutions A–C, rod-shaped crystals grew to dimensions of approximately 0.2 × 0.1 × 0.1 mm in two weeks (Figs. 2 ▸ a–2 ▸ c, Supplementary Figs. S3a–S3c ). B-Se-β-FAA-His was also crystallized under the same conditions as used for the crystallization of B-β-FAA-His. B-Se-β-FAA-His crystals grew to dimensions of approximately 0.15 × 0.05 × 0.05 mm in three weeks (Supplementary Figs. S4a–S4c ).

Figure 2.

B-β-FAA-His crystals. The crystals in (a), (b) and (c) were obtained using reservoir solutions A, B and C, respectively.

In the initial crystallization screening for V-β-FAAs no crystals of V-His-β-FAA were observed, while small crystals of V-β-FAA-His were observed with NeXtal Tubes Classics Suite solution No. 92 composed of 0.1 M MES–NaOH pH 6.5, 30%(w/v) polyethylene glycol monomethyl ether 5000, 0.2 M ammonium sulfate. By optimizing the crystallization conditions, the growth of larger V-β-FAA-His crystals was observed using a reservoir solution composed of 0.1 M MES–NaOH pH 6.5–6.8, 25–30%(w/v) polyethylene glycol monomethyl ether 5000, 0.2 M ammonium sulfate. Suitable protein concentrations were 2.0–3.0 mg ml⁻¹. V-β-FAA-His crystals grew to dimensions of approximately 0.15 × 0.1 × 0.1 mm in four weeks (Supplementary Fig. S5).

3.3. Data collection and processing

The quality of each β-FAA crystal was examined by taking diffraction snapshots. The crystal of B-His-β-FAA obtained from a drop prepared with reservoir solution A diffracted to 2.75 Å resolution, while those obtained using reservoir solutions B and C diffracted to approximately 7.00 Å resolution. The crystal of B-β-FAA-His obtained with reservoir solution A diffracted to 2.45 Å resolution, while those obtained using reservoir solutions B and C diffracted to approximately 4.00 Å resolution. The crystal of B-Se-β-FAA-His obtained with reservoir solution A diffracted to 3.20 Å resolution, while those obtained using reservoir solutions B and C diffracted to approximately 7.00 Å resolution. Data sets from B-His-β-FAA and B-β-FAA-His crystals obtained with reservoir solution A were used for structural analysis. A data set from a B-Se-β-FAA-His crystal obtained with reservoir solution A was used for phase determination by Se-SAD. These B-β-FAA crystals belonged to space group P4₁2₁2, with unit-cell parameters a = b = 112.70–112.97, c = 341.50–343.32 Å. On the basis of the molecular weight and unit-cell parameters, the crystals are likely to contain two β-FAA subunits in the asymmetric unit, with a corresponding Matthews coefficient of 3.27–3.35 ų Da⁻¹ and a solvent content of 62.6–63.3% (Matthews, 1968 ▸). Crystallographic parameters for B-β-FAA crystals are listed in Table 3 ▸. Several crystals of V-β-FAA-His diffracted to approximately 3.00 Å resolution. The V-β-FAA crystals probably belonged to space group P222₁, with unit-cell parameters of approximately a = 86, b = 100, c = 180 Å. However, the V-β-FAA-His crystals appeared to grow as twinned crystals with an estimated twin fraction of approximately 0.4 (Supplementary Table S2). Although preliminary phasing of V-β-FAA-His crystals was performed by molecular rerplacement using a search model derived from a B-β-FAA model obtained as described in the following section, no molecular-replacement solution was found. An alternative structural analysis such as cryo-EM would be required to determine the structure of V-β-FAA.

Table 3. Data collection and processing.

Values in parentheses are for the outer shell.

	B-His-β-FAA	B-β-FAA-His	B-Se-β-FAA-His
Diffraction source	BL-5A, PF	BL-5A, PF	BL-5A, PF
Wavelength (Å)	1.0000	1.0000	0.9785
Temperature (K)	100	100	100
Detector	ADSC Quantum 315r	ADSC Quantum 315r	ADSC Quantum 315r
Crystal-to-detector distance (mm)	421.531	419.828	429.770
Rotation range per image (°)	0.5	0.4	0.5
Total rotation range (°)	360	180	110
Exposure time per image (s)	12.0	10.0	7.0
Space group	P41212	P41212	P41212
a, b, c (Å)	112.86, 112.86, 343.32	112.97, 112.97, 341.69	112.70, 112.70, 341.50
α, β, γ (°)	90, 90, 90	90, 90, 90	90, 90, 90
Mosaicity (°)	0.12	0.12	0.14
Resolution range (Å)	49.05–2.75 (2.83–2.75)	48.81–2.45 (2.58–2.45)	30.00–3.20 (3.26–3.20)
Total No. of reflections	1688584 (111998)	995913 (73430)	322841 (24619)
No. of unique reflections	58819 (4467)	81705 (11160)	68732 (5238)
Completeness (%)	100 (100)	99.2 (94.9)	100 (100)
Mutliplicity	28.7 (25.4)	12.2 (6.6)	4.7 (4.7)
〈I/σ(I)〉	29.9 (4.8)	18.4 (4.0)	20.2 (5.0)
R meas	0.127 (0.943)	0.107 (0.489)	0.100 (0.387)
Overall B factor from Wilson plot (Å2)	44.1	31.7	49.3

3.4. Structure solution and refinement

No structural information for any protein with a related sequence to the β-FAAs was available when we tried to solve their crystal structures. Therefore, we used the Se-SAD method. The crystal structure of B-Se-β-FAA-His was solved by the Se-SAD method using Phenix AutoSol, which gave an initial model of B-Se-β-FAA-His. Iterative cycles of rebuilding and refinement were performed in Coot and phenix.refine, respectively, giving an improved electron-density map and model (Fig. 3 ▸ a). As expected, two subunits were found in the asymmetric unit. This model was then used as a search model to solve the crystal structure of both His-tagged B-β-FAAs by molecular replacement using Phaser-MR. The B-His-β-FAA and B-β-FAA-His crystals had the same molecular packing and overall structure as the B-Se-β-FAA-His crystal. The models of B-His-β-FAA and B-β-FAA-His were improved in the same way as the B-Se-β-FAA-His model. The crystal structure of B-His-β-FAA was refined at 2.75 Å resolution to an R _work and R _free of 19.1% and 24.0%, respectively, and the crystal structure of B-β-FAA-His was refined at 2.45 Å resolution to an R _work and R _free of 23.0% and 26.5%, respectively. The crystal structure of B-Se-β-FAA-His was also refined at 3.2 Å resolution to an R _work and R _free of 17.5% and 22.8%, respectively. The coordinates and the structure factors of B-His-β-FAA, B-β-FAA-His and B-Se-β-FAA-His have been deposited in the Protein Data Bank as entries 8i5a, 8huy and 8i59, respectively. Refinement statistics are summarized in Table 4 ▸. In the crystal, β-FAA seems to exist as a dimer (Fig. 3 ▸ b). Each subunit is composed of three domains, I, II and III, as in the large subunit of parDMFase (Arya et al., 2020 ▸; Fig. 3 ▸ c). In addition, each domain has essentially the same fold as that of the large subunit of parDMFase (Figs. 3 ▸ d–3 ▸ f). More detailed structural analyses together with enzymatic analyses of mutant proteins should provide a better understanding of the structural basis that determines the strict (R)-enantiomer specificity and the reaction mechanism of β-FAA.

Figure 3.

Crystal structure of B-β-FAA. (a) 2F _o − F _c electron-density map of B-Se-β-FAA-His contoured at 1σ. (b) Overall structure of the B-β-FAA-His dimer. Subunits A and B are shown in green and magenta, respectively. Each subunit is composed of three domains: I, II and III. (c–f) Structural superposition of B-β-FAA-His on the large subunit of parDMFase (PDB entry 6lvv). B-β-FAA-His and parDMFase are shown in green and gray, respectively. Superpositions of subunit A and domains I, II and III of B-β-FAA-His on the corresponding subunit and domains of parDMFase are shown in (c), (d), (e) and (f), respectively. The r.m.s.d. values on C^α atoms of the superposed structures in subunit A and domains I, II and III are 1.134, 0.704, 1.781 and 0.723 Å, respectively. Figures were generated using PyMOL version 1.20 (Schrödinger).

Table 4. Structure refinement.

Values for the outer shell are given in parentheses.

	B-His-β-FAA	B-β-FAA-His	B-Se-β-FAA-His
PDB entry	8i5a	8huy	8i59
Resolution range (Å)	46.50–2.75 (2.82–2.75)	47.12–2.45 (2.48–2.45)	29.93–3.21 (3.25–3.21)
Completeness (%)	99.9	99.1	99.6
σ Cutoff	F > 1.34σ(F)	F > 1.34σ(F)	F > 1.34σ(F)
No. of reflections
Working set	56657 (3947)	77478 (2358)	65177 (2435)
Test set	1999 (138)	4063 (121)	3440 (115)
Final R cryst	0.191 (0.293)	0.230 (0.270)	0.175 (0.276)
Final R free	0.240 (0.362)	0.265 (0.357)	0.228 (0.341)
No. of non-H atoms
Protein	11150	11099	11017
R.m.s. deviations
Bonds (Å)	0.008	0.008	0.009
Angles (°)	1.108	0.954	1.117
Average B factors (Å2)
Protein	46.9	36.0	50.1
Ramachandran plot
Favored regions (%)	97.7	96.8	94.7
Additionally allowed (%)	2.3	3.2	5.3
Outliers (%)	0.0	0.0	0.0

3.5. Molecular-weight analyses

In a previous study, native β-FAA produced by Burkholderia sp. AJ110349 was purified and the molecular weight of the purified β-FAA in solution was analyzed by size-exclusion chromatography and estimated to be 206 kDa from the elution volume (Imabayashi et al., 2016 ▸). Since the molecular masses of native β-FAA monomers, homodimers and homotetramers were calculated to be 81.4, 162.8 and 244.2 kDa, respectively, the oligomeric state of β-FAA in solution remained unclear. In order to clarify the oligomeric state of β-FAA in solution, the molecular weights of His-tagged β-FAAs were analyzed by the SEC–SLS method (Wen et al., 1996 ▸). As a result, the weight-averaged molecular weights of B-β-FAA-His and B-His-β-FAA were calculated to be approximately 161 and 160 kDa, respectively (Fig. 4 ▸, Supplementary Fig. S6a ). The weight-averaged molecular weights of V-β-FAA-His and V-His-β-FAA were calculated to be approximately 180 and 176 kDa, respectively (Supplementary Figs. S6b and S6c ). These values are in good agreement with the theoretical mass of homodimeric His-tagged β-FAAs. On the basis of the primary structures listed in Table 1 ▸ and Supplementary Table S1, the molecular masses of homodimeric B-His-β-FAA, B-β-FAA-His, V-His-β-FAA and V-β-FAA-His were calculated to be 167.0, 166.0, 174.8 and 173.8 kDa, respectively. We concluded that β-FAA exists as a stable dimer in solution as found in the crystal.

Figure 4.

SEC–SLS analysis of B-β-FAA-His. The chromatograms of the absorbance at 280 nm and the static light scattering at 90° are scaled into arbitrary units and shown as an orange line and a blue line, respectively. The shapes of the two chromatograms are essentially the same, indicating that the detected sample was monodisperse. The weight-averaged molecular weights are shown as a black line and dots. The averaged value calculated from the intensities from 23.8 to 25.1 min was 161.4 kDa (±1.11%).

Supplementary Material

PDB reference: N-acetyl-(R)-β-phenylalanine acylase from Burkholderia sp. AJ110349, carboxy-terminally His-tagged form, 8huy

PDB reference: selenomethionine derivative of carboxy-terminally His-tagged form, 8i59

PDB reference: amino-terminally His-tagged form, 8i5a

Funding Statement

Funding for this research was provided by: Research Institute for Science and Technology of Tokyo Denki University (grant No. Q13L-01 to Ryo Natsume); JSPS KAKENHI (grant No. 15K07401 to Ryo Natsume).